Problems With Afrikaans Word Order and Grammar

Specific language impairment in Afrikaans

Providing a Minimalist account

for problems with grammatical features and word order
Published by
LOT phone: +31 30 253 6006
Janskerkhof 13 fax: +31 30 253 6406
3512 BL Utrecht e-mail: [email protected]
The Netherlands http://www.lotschool.nl
Cover illustration: Sossusvlei, by Frenette Southwood

ISBN 978-90-78328-37-7
NUR 616
Copyright © 2007: Frenette Southwood. All rights reserved.

Specific language impairment in Afrikaans
Providing a Minimalist account
for problems with grammatical features and word order
een wetenschappelijke proeve op het gebied van de Letteren
Proefschrift
ter verkrijging van de graad van doctor

aan de Radboud Universiteit Nijmegen
op gezag van de rector magnificus prof. mr. S.C.J.J. Kortmann
volgens besluit van het College van Decanen
in het openbaar te verdedigen op maandag 5 november 2007
om 10.30 uur precies
door
Frenette Southwood
geboren op 14 december 1971
te Kaapstad
Promotores: Prof. dr. R.W.N.M. van Hout
Prof. dr. N.F.M. Corver (Universiteit Utrecht)
Manuscriptcommissie: Prof. dr. A.C.M. Rietveld

Dr. J. de Jong (Universiteit van Amsterdam)
Dr. J.B. den Besten (Universiteit van Amsterdam)
The research presented in this book was supported by the South African
National Research Foundation under grant number GUN 2069783. Any
opinion, findings, and conclusions or recommendations expressed in this
material are those of the author and do not necessarily reflect the views of the
National Research Foundation.
Vir Zoë
Table of contents
Acknowledgements i
List of tables iii
List of figures vii
List of abbreviations ix
Chapter 1: Introduction 1
Chapter 2: Characteristics and theoretical accounts of SLI 7
2.1. Introduction 7
2.2. Why study SLI? 7
2.3. Some characteristics of SLI 8
2.3.1. Some characteristics of SLI as it presents itself in
English 9
2.3.1.1. Grammatical morphemes 9
2.3.1.2. Word order related phenomena 9
2.3.1.3. (Non-)co-referential relationships 10
2.3.2. How SLI presents itself in other languages, in
terms of grammatical morphology 11
2.3.2.1. Germanic languages 11
2.3.2.2. Romance languages 16
2.3.2.3. Other languages 17
2.3.2.4. Summary: How SLI presents itself in languages
other than English 22
2.4. Theoretical accounts of SLI 22
2.4.1. The Agreement/Tense Omission Model 23
2.4.2. The Representational Deficit for Dependent
Relations Hypothesis 27
2.4.3. The Feature Deficit Hypothesis 29
2.5. Chapter conclusion 31
Chapter 3: Afrikaans and its syntactic analysis in terms of
Minimalist syntax 33
3.1. Introduction 33
3.2. Why study Afrikaans? 33
3.3. Some properties of the morphology and syntax of Afrikaans 37
3.3.1. Grammatical morphology 37
3.3.1.1. Number 38
3.3.1.2. Person 39
3.3.1.3. Gender 39
3.3.1.4. Case 40
3.3.1.5. Tense 43
3.3.1.6. Summary: Morphosyntax 48
3.3.2 Word order related phenomena 48
3.3.2.1. Placement of the finite verb 49
3.3.2.2. Question constructions 54
3.3.2.3. Passive constructions 57
3.3.2.4. Preposing and scrambling 59
3.3.2.5. Co-referential relationships 60
3.3.3. Conclusion: Some properties of the morphology
and syntax of Afrikaans 61
3.4. Why work within the framework of Minimalist syntax? 62
3.4.1. General assumptions of Minimalist syntax 63
3.4.2. Feature checking 66
3.4.3. Movement 67
3.4.4. Minimalist syntax and problem structures for
children with SLI 72
3.4.4.1. Grammatical morphemes 72
3.4.4.2. Word order related phenomena 73
3.4.4.3. (Non-)co-referential relationships 74
3.4.5. Summary: Why work within the framework of
Minimalist syntax? 77
Chapter 4: Methodology 79
4.2. Design of the study 81
4.3. Participant selection 84
4.3.1. Participants with SLI 85
4.3.2. Typically developing participants 87
4.4. The participant groups 89
4.4.1. Children with SLI 89
4.4.2. Typically developing controls 92
4.5. Collection of spontaneous language 92
4.6. Experimental tasks 93
4.6.1. Number 95
4.6.1.1. Number: Comprehension tasks 95
4.6.1.2. Number: Production tasks 98
4.6.2. Person and case 99
4.6.2.1. Person and case: Comprehension tasks 99
4.6.2.2. Person and case: Production task 100
4.6.3. Possessive case 101
4.6.3.1. Possessive case: Comprehension task 101
4.6.3.2. Possessive case: Production task 101
4.6.4. Tense 102
4.6.4.1. Tense: Comprehension tasks 102
4.6.4.2. Tense: Production task 103
4.7. Data transcription and scoring 104
4.7.1. Language sample 104
4.7.2. Experimental tasks 110
Chapter 5: Number comprehension and production 111

5.2. Results: Picture selection task: Singular/plural forms of
real words 111
5.3. Results: Judgement tasks: Correct/incorrect plural forms
of real and nonsense words 119
5.3.1. Real words requiring regular plural suffixes 120
5.3.2. Real words which should have irregular plural
forms 123
5.3.3. Nonsense words 128
5.4. Results: Sentence completion tasks: Plural forms 134
5.4.1. Real words requiring regular plural suffixes 135
5.4.2. Real words which have irregular plural forms 139
5.4.3. Nonsense words 145
5.5. Number production in the language samples 149
5.6. Discussion: Number comprehension and production 152
Chapter 6: Comprehension and production of person

and case 157

6.2. Person and case on pronouns 157
6.2.1. Results: Picture selection task: Person and case on
pronouns 157
6.2.2. Results: Judgement task: Person and case on
pronouns 161
6.2.3. Results: Sentence completion task: Person and
case on pronouns 165
6.3. se-constructions 172
6.3.1. Results: Picture selection task: se-constructions 172
6.3.2. Results: Sentence completion task: se-constructions 174
6.4. Person and case production in the language samples 176
6.4.1. Pronouns 176
6.4.2. se-constructions 181
6.5. Discussion: Person and case comprehension and production 181
Chapter 7: Tense comprehension and production 189

7.2. Results: Experimental tasks 189
7.2.1. Results: Picture selection task: Present and
past tense 189
7.2.2. Results: Judgement task – Hendiadyses: Present
and past tense 195
7.2.3. Results: Sentence completion task: Past tense
constructions 197
7.3. Discussion: Experimental tasks 204
7.4. Tense production in the language samples 207
7.5. Discussion: Spontaneous data on tense production 213
Chapter 8: What are the linguistic characteristics of
Afrikaans-speaking children with SLI? 217

8.2. The linguistic characteristics revealed by the experimental
tasks 217
8.3. The linguistic characteristics revealed by the language
samples 222
8.3.1. Error analysis of the first 100 utterances of each
language sample 222
8.3.2. Error analysis of the full first 30 minutes of each
language sample – other errors 225
8.3.3. Word order errors 233
8.4. Do the errors reveal a possible clinical marker of SLI in
Afrikaans? 238
Chapter 9: Accounts of SLI as it presents itself in Afrikaans 245

9.2. The Agreement/Tense Omission Model 245
9.2.1. Merit of the ATOM 245
9.2.2. Predictions of the ATOM for Afrikaans 251
9.2.3. Are the predictions of the ATOM for Afrikaans
borne out? 258
9.3. The Representational Deficit For Dependent Relations
Hypothesis 259
9.3.1. Merit of the RDDR 259
9.3.2. Predictions of the RDDR for Afrikaans 262
9.3.3. Are the predictions of the RDDR for Afrikaans
borne out? 263
9.4. The Feature Deficit Hypothesis 265
9.4.1. Merit of the Feature Deficit Hypothesis 265
9.4.2. Predictions of the Feature Deficit Hypothesis for
Afrikaans 266
9.4.3. Are the predictions of the Feature Deficit
Hypothesis for Afrikaans borne out? 270
9.5. Summary: Predictions of the three accounts for SLI in
Afrikaans 273
9.6. An alternative account of SLI as it presents itself in
Afrikaans 273
9.6.1. Background 273
9.6.2. Another look at the errors pertaining to
grammatical features 274
9.6.3. A Minimalist account of the errors pertaining to
grammatical features 277
9.6.4. Another look at the word order errors 285
9.6.5. The alternative account 296
Chapter 10: Conclusion 299
10.1. Summary of the present study 299

10.2. Limitations of the present study 301
10.3. Suggestions for subsequent studies of SLI in Afrikaans 303
10.4. Conclusion 304
References 305
Appendix A: Information on SLI participants’ language test

results 326
Appendix B: Letter to parents: Explanation and consent 328
Appendix C: Case-history form 331
Appendix D: Items of the experimental tasks 336
Samenvatting [Dutch summary] 351

Opsomming [Afrikaans summary] 356
Acknowledgements
My PhD years have been my most challenging ones thus far, for reasons
mostly unrelated to the PhD work itself. I am thankful towards several
people who supported and assisted me in this time. I begin by thanking
my supervisors. Roeland van Hout was kind enough to inherit this
project from the late Marco Haverkort. Marco was enthusiastic about the
project from when it was still barely more than an idea; Roeland
continued where Marco left off. I am grateful for Roeland’s extreme
patience with me and my work methods; for his explanations and re-
explanations of the statistical results; and for the skillful manner in which
he gave the project direction. I also thank Norbert Corver for his
involvement in this project, for sharing his general insights but also for
his excellent teachings on matters pertaining to syntactic analyses. I have
benefited greatly from being their student.
I am indebted to colleagues at my home department, the Department of

General Linguistics at Stellenbosch University. My departmental chair,
Christine Anthonissen, was supportive of this project from start to finish
in various ways. I appreciate her advice (it is a privilege to have a
departmental chair who invests in the development of young
researchers), her constant encouragement, and the many practical ways
in which she made it possible for me to do this research project. I also
wish to thank Johan Oosthuizen for patiently trying to teach me
(particularly Afrikaans) syntax and for help with formulation in earlier
versions of chapters 2 and 3. I also thank Christine Smit for her
administrative assistance. The contributions of Ondene van Dulm are
too numerous to be listed here. Without her, this project might never
have progressed beyond the idea stage.
I am grateful for the support of various departmental and research

assistants who helped me with administrative matters. I single out Alexa
Anthonie; her splendid attitude made it an absolute pleasure to work
with her.
I am most grateful towards the South African National Research

Foundation (NRF) for grants since 2005, and for the financial
i
contributions made by Stellenbosch University’s Research Sub-
committee A and the Dean of my home faculty, Hennie Kotzé. I would
also like to thank Hennie for granting me research leave.
At Radboud University Nijmegen, I am grateful to Frans van der Slik for

initial advice regarding data coding matters, and to Annika van der Made
for her help with the practical arrangements for the defence. At
Cambridge University, I thank Theresa Biberauer for her comments on
an earlier version of my dissertation.
One does not get many opportunities to thank one’s family in print; I
will make use of this one. I thank my parents, Carl and Henriette
Tarrant, for every opportunity they have given me to study. They
faithfully continue to think that which I do is great, even though they still
do not quite understand what it is that I do or why I am so passionate
about it. I also thank my parents-in-law for their unfailing interest.
Thanks also to my husband, Michael, for sharing me with this project
and for cheering me on. I am blessed to have you all in my life.
ii
List of tables
Table 4.1. Information on participants

Table 4.2. Summary of experimental tasks and the aspects which they
assessed
Table 4.3. Item types in the picture selection task for number
comprehension
Table 4.4. Item types in the judgement task for number
comprehension – real words requiring regular plural
suffixes
comprehension – real words requiring irregular plural
suffixes
comprehension – nonsense words
Table 4.7. Item types in the production task entailing regular plural
forms
Table 4.8. Item types in the production task entailing irregular plural
forms
Table 4.9. Item types in the production task entailing plural forms of
nonsense words
Table 4.10. Item types in the picture selection task for tense
comprehension
Table 5.1. Responses to the 11 items to which a response other than
the target or the target with reversed plurality was given -
Picture selection task: Singular/plural forms of real words
Table 5.2. Item statistics – Picture selection task: Singular/plural
forms of real words
Table 5.3. Summary of performance per group – Picture selection
task: Singular/plural forms of real words
Table 5.4. Items on which SLI and TD4 groups performed
noticeably differently - Picture selection task:
Singular/plural forms of real words
Table 5.5. Item statistics – Judgement task: Real words requiring
regular plural suffixes
Table 5.6. Summary of performance per group – Judgement task:
Real words requiring regular plural suffixes
iii
Table 5.7. Item statistics – Judgement task: Items constituting
ungrammatical plural forms of real words which should
have irregular plural forms
Items constituting ungrammatical plural forms of real
words which should have irregular plural forms
grammatical plural forms of real words which should have
irregular plural forms
Items constituting grammatical plural forms of real words
which should have irregular plural forms
Table 5.11. Item statistics – Judgement task: Ungrammatical plural
forms of nonsense words
Ungrammatical plural forms of nonsense words
Table 5.13. Items on which the three groups performed noticeably
differently – Judgement task: Ungrammatical plural forms
of nonsense words
Table 5.14. Item statistics – Judgement task: Grammatical plural forms
of nonsense words
Grammatical plural forms of nonsense words
Table 5.16. Item statistics – Production task: Plural forms of real
words requiring regular plural suffixes
Table 5.17. Summary of performance per group – Production task:
Plural forms of real words requiring regular plural suffixes
Table 5.18. Item statistics – Production task: Plural forms of real
words that should have irregular plural forms
plural forms of real words that should have irregular plural
forms
Table 5.20. Item statistics –Production task: Plural forms of nonsense
words
Plural forms of nonsense words
Table 5.22. Number of singular and plural nouns used (correctly and
incorrectly) in the language samples
Table 6.1. Item statistics – Picture selection task: Pronouns
iv
Table 6.2. Summary of performance per group – Picture selection:
Pronouns
Table 6.3. Item statistics – Judgement task: Pronouns
Table 6.4. Mean number of correct responses to each item –
Judgement: Pronouns
Table 6.5. Item statistics – Production task: Pronouns
Pronouns
Table 6.7. Responses of SLI and TD4 groups to item 18 – Sy eet
appels, maar (julle eet piesangs) – of pronoun production task
Table 6.8. Item statistics – Picture selection task: se-constructions
Table 6.9. Mean number of correct responses to each item –
Production task: se-constructions
Table 6.10. Accuracy of pronoun use by the three groups of
participants
Table 6.11. Frequency of correct appearance of pronouns, appearing
correctly where they should have
Table 6.12. Summary of performance per group – Language sample:
Spontaneous production of errors on pronouns as a
proportion of all pronoun forms
Table 6.13. Number of se-constructions used correctly and incorrectly
in the language samples
Table 7.1. Item statistics – Picture selection task: Present and past
tense
Table 7.2. Item statistics of subgroup of items consisting of past
tense forms of be and have – Picture selection task: Present
and past tense
Table 7.3. Summary of performance per group – Subgroup of items
consisting of past tense forms of be and have – Picture
selection task: Present and past tense
Table 7.4. Item statistics of subgroup of items consisting of past
tense forms containing het – Picture selection task: Present
and past tense
consisting of past tense forms containing het – Picture
Table 7.6. Item statistics – Judgement: Hendiadyses
Hendiadyses
v
Table 7.8. Summary of performance per group – Sentence
completion task: Past tense, grammatical responses
Table 7.9. Summary of performance per group – Sentence
completion task: Past tense, ungrammatical responses
Table 7.10. Present tense constructions used correctly and incorrectly
in the language samples
Table 7.11. Summary of performance per group – Production of
present tense forms in the language samples
Table 7.12. Past tense constructions used correctly and incorrectly in
the language samples
Table 7.13. Summary of performance per group – Production of past
tense forms in the language samples
Table 8.1. Overview of the difference in results between the three
groups of participants on the 15 experimental tasks
Table 8.2. Summary of performance per group – Composite index
consisting of a selection of seven experimental tasks
Table 8.3. Overview of some of the measures of the language sample
analysis
Table 8.4. Pearson’s correlation between measures of the language
sample
consisting of four measures of the language sample
analysis
Table 8.6. Frequency of verb-related errors in the language samples
of the three groups of participants
Table 8.7. Frequency of non-verb-related errors in the language
samples of the three groups of participants
Table 8.8. Summary of the frequency of errors (excluding those
related to the grammatical features number, person, case
and tense) made by the SLI children
Table 8.9. Summary of word order errors made per group
Table 8.10. Results of classification of participants into three groups
based on a selection of three experimental measures
Table 8.11. Results of classification of participants into three groups
based on a selection of two measures from the language
samples
vi
List of figures
Figure 5.1. Box plot of performance per group - Picture selection task:
Figure 5.2. Box plot of performance per group – Judgement task: Real
Figure 5.3. Box plot of performance per group – Judgement task:
Items consisting of ungrammatical plural forms of real
words which should have irregular plural forms
Items consisting of grammatical plural forms of real words
which should have irregular plural forms
Figure 5.5. Box plot of performance per group − Judgement task:
Figure 5.7. Box plot of performance per group – Production task:
Plural forms of real words requiring regular plural suffixes
Plural forms of real words that should have irregular plural
forms
Plural forms of nonsense words
Figure 6.1. Box plot of performance per group − Picture selection
task: Pronouns
Pronouns
Figure 6.3. Box plot of performance per group − Production task:
Pronouns
Figure 6.4. Box plot of performance per group − Production task: se-
constructions
Figure 7.1. Box plot of performance per group – Subgroup of items
consisting of past tense forms of be and have – Picture
vii
Hendiadyses
Figure 8.1. Box plot of performance per group – Composite score of
seven experimental tasks
four measures of the language sample analysis
viii
List of abbreviations
ACC accusative case

AdvP adverbial phrase
Agr agreement
ANOVA analysis of variance
ARW Afrikaanse Reseptiewe Woordeskattoets ‘Afrikaans
Receptive Vocabulary Test’
ASHA American Speech, Language, and Hearing Association
AST Afrikaanse Semantiese Taalevalueringsmedium ‘Afrikaans
Semantic Language Evaluation Medium’
ATOM Agreement/Tense Omission Model
C complemetiser
CA chronological age
CCH Computational Complexity Hypothesis
CHL computational system for human language
CONTR contracted
CP complementiser phrase
DAT dative case
DIM diminutive
DP determiner phrase
EPP extended projection principle
EOI Extended Optional Infinitive Hypothesis
GEN genitive case
Gr (school) grade
I/INFL inflection
INF infinitive
IP inflectional phrase
IQ intelligence quotient
LARSP Language Assessment, Remediation and Screening Procedure
LCA linear correspondence axiom
LF logical form
M mean
MLU mean length of utterance
MLUm mean length of utterance measured in morphemes
MLUw mean length of utterance measured in words
ModP modal phrase
ix
N sample size
NOM nominative case
NP noun phrase
OSV object, subject, verb
PART participle
PASS passive
PERS person
PF phonetic form
PL plural
PLS-3 Preschool Language Scales – 3
PP prepositional phrase
Q-feature question feature
RDDR Representational Deficit for Dependent Relations Hypothesis
R-expression referential expression
S subject
SC small clause
SGL singular
SLI specific language impairment
SOV subject, object, verb
Spec specifier
SVO subject, verb, object
T tense
TACL-R/III Test of Auditory Comprehension of Language – Revised
/ Third edition
TD4 typically developing 4-year-old
TD6 typically developing 6-year-old
TMT Toets vir Mondelinge Taalproduksie ‘Test for Oral
Language Production’
TOLD-P Test of Language Development – Primary
TP tense phrase
UTAH Uniform Theta-Assignment Hypothesis
V verb
VO verb, object
VP verb phrase
vP light verb phrase
VSO verb, subject, object
V-to-I verb to infinitive
λ logical form
π phonetic form
x
Introduction
Chapter 1
Introduction
Specific language impairment (SLI) is a significant impairment in the

spoken language ability of children in the absence of identifiable causal
factors or obvious accompanying factors such as neurological deficits,
mental challenges, hearing disabilities, and emotional or behavioral
problems (Leonard 1998:vi; Stark and Tallal 1981). In this study, the
characteristics of SLI as it presents itself in Afrikaans1 – a descendent of
Dutch, principally spoken in South Africa – are determined. As De Jong
(1999:i) rightly states, “in SLI, target language matters. The fact that
children are language-impaired does not spell out the symptoms that
will testify to the disorder. The appearance of SLI is also dependent on
the native language that the child is in the process of learning”. English
is the “native language” that has been studied most comprehensively by
researchers interested in SLI. The characteristics of SLI as it presents
itself in English are, therefore, comparatively well-known. This study of
SLI focuses on Afrikaans and (i) provides data useful for the assessment
and remediation of children with SLI who are acquiring Afrikaans –
such data are very scarce – and (ii) contributes to the wider field of SLI
research by providing language-specific data against which the
explanatory merit of theoretical accounts of SLI can be tested.
Only a limited number of researchers has investigated the

comprehension of grammatical morphemes by children with SLI. The
research carried out by Fellbaum, Miller, Curtiss, and Tallal (1995) is one
example; they found that English-speaking children with SLI perform
worse than their typically developing peers. This finding is supported by
one from clinical practice, namely that English-speaking children with
SLI usually fare poorly on tests of receptive language such as the
1 Afrikaans is a language with approximately 6 million mother-tongue speakers, which
amounts to 13% of the South African population. It is the mother-tongue with the
third largest speaker base in South Africa: Of the other 10 official languages, only Zulu
(24%) and Xhosa (18%) have more mother-tongue speakers (Statistics South Africa
2003:14).
1
Introduction
Grammatical Morphemes subtest of the Test for Auditory

Comprehension of Language-Third Edition (TACL-III) (Carrow-
Woolfolk 1999). A larger number of studies have been performed on the
production of grammatical morphemes by children with SLI. As
discussed in section 2.3.1, English-speaking children with SLI have been
found to be especially weak in their use of grammatical morphemes.2 It
has been established that these children demonstrate a lower percentage
of use of grammatical morphemes than do their typically developing
peers. Also, English-speaking children with SLI sometimes insert
grammatical morphemes into inappropriate contexts (see, e.g., Gopnik
1990a; Roberts and Rescorla 1995).
The question arises as to whether comprehension and production of

grammatical morphemes would be a problem for Afrikaans-speaking
children with SLI as well. Specific questions following from this broader
one are:
(i) How well are grammatical morphemes comprehended and
produced by Afrikaans-speaking children with SLI? Are all of the
morphemes related to the grammatical features number, person,
case, and tense equally difficult (or easy) to comprehend and
produce correctly, and can any error patterns be detected?
(ii) How does the comprehension and production of grammatical
morphemes by these children compare with that of their typically
developing peers and that of younger, typically developing
Afrikaans-speaking children?
Once it is known how Afrikaans-speaking children with SLI fare
compared to age-matched and younger controls, one can establish
whether the children with SLI present with a delay, a deviance, or both,
in terms of their command of grammatical morphemes.
Another question is whether predictions regarding comprehension and

production of grammatical morphemes made for Afrikaans by the
Agreement/Tense Omission Model (ATOM) (Wexler and colleagues
1995, 1996, 1998), the Representational Deficit for Dependent Relations
Hypothesis (RDDR)3 (Van der Lely 1994, 1996, 2003, 2004), and the
2 In section 2.3.2, it is shown that this also holds true for non-English-speaking children
with SLI.
3 Recently, the RDDR has been renamed the “Computational Complexity Hypothesis”
(CCH). See, for example, Van der Lely (2005).

2
Introduction
Feature Deficit Hypothesis (Gopnik 1994a) are borne out by the

Afrikaans data. If not, the question is whether an alternative,
comprehensive account of SLI as it presents itself in Afrikaans can be
proposed.
The general question to be answered by this study was: How does SLI,
which is characterised by a problem with grammatical morphemes,
present itself in Afrikaans, a morphologically impoverished language? In
order to answer this general question in a comprehensive manner – and
to ascertain whether the predictions made for SLI in Afrikaans by
current accounts of SLI are borne out by actual data – six specific
questions were posed, as set out below.
1. How are grammatical morphemes – specifically those related to the

features number, person, case, and tense – comprehended by
Afrikaans-speaking children with SLI?
2. Do Afrikaans-speaking children with SLI present with a delay, a
deviance, both, or neither in terms of their comprehension of these
morphemes?
3. How are grammatical morphemes – specifically those related to the
features number, person, case, and tense – morphologically realised
(i.e., produced) by Afrikaans-speaking children with SLI?
4. Do Afrikaans-speaking children with SLI present with a delay, a
deviance, both, or neither in terms of their production of these
morphemes?
5. The ATOM, the RDDR, and the Feature Deficit Hypothesis make
specific predictions regarding the comprehension and/or
production of grammatical morphemes by Afrikaans-speaking
children with SLI, as well as about the word order of their
utterances. Are these predictions borne out by the Afrikaans data
obtained in this study?
6. If the answer to question 5 is “partially” or “not at all”, can one
propose an adequate alternative account of SLI as it presents itself
in Afrikaans?
The dissertation is structured in the following way. In chapter 2, the

reader is introduced to the characteristics of SLI as it presents itself in
English as well as in a selection of other languages. A number of
processing and linguistic accounts of SLI have appeared in recent years.
3
Introduction
In this chapter, three current linguistic accounts are set out, namely the
ATOM, RDDR, and Feature Deficit Hypothesis.
In chapter 3, the discussion turns to Afrikaans, explicating why the way

SLI presents itself in Afrikaans should be studied and what properties
Afrikaans has that makes it an interesting language to study when
investigating SLI. A brief exposition of the syntactic framework in which
the work in this study is done (viz. Minimalist syntax) is then given,
indicating how Afrikaans is analysed in terms of Minimalist syntax.4
Several subgroups of children have been identified (see, for example,

Leonard 1991; Conti-Ramsden, Crutchley, and Botting 1997, Van der
Lely 2003; Van Daal, Verhoeven, and Van Balkom 2004); those who
participated in this study all had so-called Grammatical (G)-SLI. Chapter
4 describes the methodology used to select participants for this study, to
obtain data on aspects of the language of typically developing Afrikaans-
speaking children as well as of their counterparts with SLI, and to
analyse these data. Data were gathered in such a manner as to provide
answers to the specific research questions above. A description of the
experimental tasks that were administered to gather these data is
included in chapter 4, with the test items appearing in appendix D. A
second data source was the spontaneous language samples collected
from each child.
In chapters 5 to 7, the results of the language sample analysis and those

of the experimental tasks assessing the comprehension and production
of grammatical morphemes in the language of Afrikaans-speaking
children are presented. In chapter 8, an overview is given of the results
of the previous three chapters: The performance of (i) the participants
across experimental tasks, (ii) some participants individually on the
experimental tasks, and (iii) the three groups of participants on aspects
4 The version of Minimalist syntax set out in section 3.4 is not the most recent one;
rather, it is the one generally associated with the proposals made in Chomsky (1995a).
Footnotes are used to refer to more recent ideas and terminology. The development of
Minimalist syntax is still very much “a work in progress”. Leading ideas about, for
example, Move as a combination of Copy and Merge, internal vs. external Merge,
probes, goals, phases, and edge features, amongst many others, are still being
developed. For this reason, the present study will be couched within the somewhat
older but more “established” version of Minimalist syntax.
4
Introduction
(relevant to chapters 5 to 7) of the language samples are considered. Also

in chapter 8, the errors made by the participants – including some errors
not involving aspects focused on in chapters 5 to 7 – are discussed. The
general finding was that the Afrikaans-speaking children with SLI fared
significantly worse than their age-matched typically developing peers on
the experimental tasks; the children with SLI present like younger
typically developing ones. In terms of their spontaneous use of the
grammatical morphemes in question, the two groups of typically
developing children fared the same, with the children with SLI faring
worse than both. The children with SLI mostly made the same errors as
did the typically developing 4-year-olds; however, some errors were
unique to the children with SLI.
Chapter 9 contains a discussion on the extent to which the three

linguistic accounts correctly predict the comprehension and production
of grammatical morphemes by Afrikaans-speaking children with SLI.
This discussion comprises the answer to research question 5, which, in
short, is “no”. Then, regarding research question 6, a comprehensive
account of the characteristics of SLI as it presents in Afrikaans – more
comprehensive than the accounts provided by the theories discussed in
chapter 2 – is proposed. The proposed account essentially states that the
problem of Afrikaans-speaking children with SLI lies not with the
narrow syntax, or computational system, but with the mapping of the
syntactic component onto the phonological one.
Chapter 10 concludes the dissertation, by providing a summary of its

content and by pointing the way forward for subsequent examinations of
SLI as it presents itself in Afrikaans.
5
Characteristics and theoretical accounts of SLI
Chapter 2
2.1. INTRODUCTION
The aim of this chapter is to describe some of the characteristics of SLI

as it presents itself in a variety of languages, and to discuss some current
theoretical accounts of SLI. The general reasons for studying SLI (in any
language) are presented in section 2.2. This is followed, in section 2.3, by
a brief discussion of the characteristics of SLI as it has been found to
present itself in English and in a selection of other languages. In section
2.4, the content and merit of three theoretical accounts are presented:
the ATOM (section 2.4.1), the RDDR (section 2.4.2), and the Feature
Deficit Hypothesis (section 2.4.3).
2.2. WHY STUDY SLI?
The reasons for studying SLI are three-fold. The first two concern the
practical application of knowledge and understanding of SLI. In a study
designed to determine the prevalence of SLI in the United States of
America,5 Tomblin, Records, Buckwalter, Zhang, Smith, and O’Brien
(1997) found that an estimated 7.4% of all 5-year-olds – 8% of all boys
and 6% of all girls – have SLI. SLI is of a long-standing nature (cf.
Brinton, Fujiki, and Robinson 2005; Clegg, Hollis, Mawhood, and Rutter
2005; Gopnik 1994a): 40% of kindergarten children identified as having
SLI still exhibit significant language problems 4 to 5 years after
kindergarten (Aram and Nation 1980). This means that a number of
school-going children (possibly 3%) present with SLI. The effect of SLI
on academic activity is widely documented (see, for example, Conti-
Ramsden, Knox, Botting, and Simkin 2002; Aram and Nation 1980). The
first reason for studying SLI is that, with greater understanding of this
impairment, it might be possible to remediate these children in more
5 Participants were from two Midwestern states, but were chosen in such a way as to
represent the demographics of the larger American population.

7
effective ways, and in ways that would lead to the containment of the
effect of SLI on the cognitive activities of these children earlier in their
educational careers.
The second reason is the contribution that a study of SLI could make to
a better understanding of other, non-SLI, language problems. As
children with SLI, per definition, have no other disability, data on their
knowledge and use of language could act as baseline data for other
disabled groups who have a language impairment but also other
disabilities, such as other developmental disorders or hearing impairment
(Leonard 1998:9).
The third reason for studying SLI concerns the insights such a study
might offer into the nature of the human language faculty. The central
task of a theory of grammar could be said to be that of providing a
unified account of the properties of human language. A reasonable
assumption would then be that such a theory must also be able to
account for the language of children with SLI, seeing that SLI is, after all,
a form of human language. The study of SLI can therefore contribute to
our knowledge and understanding of the nature of human language and
may reveal shortcomings of current grammatical theory, thereby
indirectly contributing to the revision of such theory.
2.3. SOME CHARACTERISTICS OF SLI
Although some crosslinguistic data are available,6 most studies of SLI

examine data from one language only, the most common language
studied being English. What follows in section 2.3.1 is a brief discussion
of some of the characteristics of SLI as it presents itself in English. In
section 2.3.2, the characteristics of SLI as it has been reported for a
selection of other languages are presented.
6 See, amongst others, Bortolini, Leonard, and Caselli (1998); Clahsen, Bartke, and
Goellner (1997); Leonard, Sabbadini, Volterra, and Leonard (1988); Linder and
Johnston (1992); Morgan, Herman, and Woll (2007); Paradis, Crago, and Genesee
(2002).
8
2.3.1. Some characteristics of SLI as it presents itself in English
2.3.1.1. Grammatical morphemes

Children with SLI do not form a homogeneous group (Aram 1991:84-
85), neither when compared to others of the same language nor when
compared across languages. Despite their heterogeneity, one
characteristic is shared by most children with SLI. This is that they are
especially weak in their use of grammatical morphemes, with some
morphemes appearing to be disproportionately difficult to master
(Dromi, Leonard, and Shteiman 1993:760). Several studies have shown
that children with SLI exhibit a lower percentage of use of grammatical
morphemes than age-matched controls.7 Grammatical morphemes are
omitted in obligatory contexts, as in *He lick it (from Loeb and Leonard
1991:124), but are also sometimes inserted in inappropriate contexts (cf.
Gopnik 1990a; Roberts and Rescorla 1995). The example *You got a tape
recorders from Gopnik (1990a:147) illustrates such inappropriate insertion
of a grammatical morpheme, where the plural form of tape recorder is used
to refer to a single entity. In chapters 5 to 8, the comprehension and
production of grammatical morphemes by Afrikaans-speaking children
with SLI are discussed. It will be seen that these children also omit such
morphemes from obligatory contexts, substitute one morpheme with
another, and, to a lesser extent, insert grammatical morphemes
inappropriately.
2.3.1.2. Word order related phenomena

Question constructions produced by children with SLI often exhibit the
incorrect word order. Leonard (1995) found that children with SLI
exhibit less movement of the auxiliary verb when producing wh-
questions than do mean length of utterance (MLU)-matched controls.
Hence, children with SLI more frequently produce question forms such
as *What we can make? instead of What can we make? than do MLU-
matched controls. Furthermore, De Villiers, De Villiers, Roeper, and
Seymour (2001) reported that children with SLI have more problems
than typically developing peers with the production of multiple wh-
questions, such as Who ate what?. In contrast to the production of wh-
questions, no published research has been done on the production of
7 See, amongst others, Leonard, Bortolini, Caselli, McGregor, and Sabbadini (1992);
Loeb and Leonard (1991); Rice, Wexler, and Cleave (1995); Ullman and Gopnik (1994).
9
yes/no-questions by children with SLI. Even so, one would expect that
these children often do not make use of subject-verb inversion when
producing yes/no-questions; the fact that a question has been produced
would probably have to be inferred from rising intonation. In chapter 8,
some examples of question constructions with the incorrect word order
produced by the Afrikaans-speaking children with SLI are given. No
instances of multiple wh-questions occurred, but single wh- and yes/no-
questions were produced.
Regarding passive constructions, Van der Lely (1996) found that, when
compared to typically developing children who obtain similar language
test scores, children with SLI show greater difficulty in interpreting “full”
passives, such as The teddy is mended by the girl. Children with SLI are also
more likely to interpret “short” passives, such as The teddy is washed, as an
adjectival construction – which can be paraphrased as The washed teddy –
than as an agentless passive, i.e., a passive construction without the
prepositional (by) agent phrase. As Leonard (1998:59) points out, “The
[adjectival – FS] interpretation is not incorrect, of course; however, the
difference between the children suggested a possible avoidance of a
passive interpretation on the part of the children with SLI”. The
comprehension of passive constructions by Afrikaans-speaking children
was not assessed for the purposes of this study. The production of
passive constructions was considered when analysing the spontaneous
language samples; some results in this regard are presented in chapter 8.
2.3.1.3. (Non-)co-referential relationships

According to Van der Lely and Stollwerck (1997), children with SLI
experience problems in establishing (non-)co-referential relationships
between nominal expressions. For example, these children find it
difficult to determine the referent of personal and reflexive pronouns
when they have to make use of syntactic clues only. When they do not
receive any visual or semantic clues, nor have the opportunity to use
real-world knowledge, they typically have problems deciding to
what/whom the him and himself refer in constructions such as Peter is
washing him, Dad says that Peter is washing himself, and The boy says that every
soldier is washing himself (Van der Lely and Stollwerck 1997:276-278). Note,
however, that even typically developing children (some as old as 7 years)
experience problems in establishing such binding relationships (De
10
Villiers, Cahillane, and Altreuer 2006).8 The comprehension and

production of these relationships were not investigated for the purposes
of this study.
2.3.2. How SLI presents itself in other languages, in terms of

grammatical morphology
Presented here is a selection of languages other than English in which

SLI has been studied. The participants with SLI in these studies mostly
ranged from 4 to 9 years. In contrast to English, in which SLI has been
studied widely, the number of studies of SLI in these other languages is
often very limited.
2.3.2.1. Germanic languages

Dutch-speaking9 children with SLI experience difficulty with verb
morphology. Regarding number, the third-person marker -t,10 and to a
lesser extent the plural –en, may be omitted in obligatory contexts. An
example given by De Jong (2003:157) is given here as (1).
8 Interestingly, De Villiers et al. (2006) found that these children’s correct production of
constructions containing (non-)co-referential relationships surpassed their correct
comprehension of such constructions.
9 Dutch is considered by some to be an SOV language (Koster 1975; but see Zwarts
1993 for an analysis of Dutch as an underlyingly SVO language). In embedded clauses,

the finite verb occurs in the final position, following the object, as shown in example
(i); in subject-initial declarative matrix clauses, the finite verb is in the second position,
following the subject but preceding the object, as shown in (ii) (both examples from
Bastiaanse and Bol 2001:276).
(i) (Ik denk) dat de boer de koe melkt
(I think) that the farmer the cow milks
‘(I think) that the farmer milks the cow’
(ii) De boer melkt de koe
The farmer milks the cow
‘The farmer milks the cow’
Apart from inverted constructions involving the second person singular pronoun –
such as kom jij ‘come you’ – tense is indicated by means of overt morphology –
specifically, by means of suffixation – on all verbs except for the first-person present
tense singular form.
10 -t is also a second-person singular marker, as shown in jij kijkt ‘you-SGL look’.
11
(1) Target:
dan kijk mama naar kas dan kijkt mama naar de kast
then look mother at cupboard then looks mother at the cupboard
‘Then the mother looks at the cupboard’
Also, the number feature may be represented inappropriately, usually by

producing a singular verb form where the subject is in the plural form
(De Jong 2003:156). An example of such an error is given in (2), taken
from De Jong (2003:157).
(2) Target:
boeken valt boeken vallen
books falls books fall
‘The books fall’
Furthermore, the verb may appear in its infinitival form (and remain in
the sentence-final position) but without a (compulsory) auxiliary to
indicate number (and person) (De Jong 2004:274-276). Regarding tense,
past tense marking may be omitted or substituted by present tense
marking, or the verb may appear in its infinitival form (De Jong
2004:273-274). Wilsenach (2006:116) reports that Dutch-speaking
children with SLI more frequently omit auxiliary verbs and the prefix ge-
(used for the inflection of the past participle) than do age-matched
controls.
German-speaking11 children with SLI often omit the article in

determiner phrases (DPs). Alternatively, they use an incorrect gender for
the article, e.g., *die Lehrer ‘the teacher-MALE’ instead of der Lehrer ‘the
11 German reembles Dutch as far as word order is concerned. It has frequent overt
constituent movement. Amongst other movement operations, German demonstrates

object scrambling, as shown in (i) and (ii) below, which have the same “deep structure”,
i.e., theta role-related meaning.
(i) Sie hat meinem Bruder den Umschlag gegeben
she has my brother the envelope given
‘She gave my brother the envelope’
(ii) Sie hat den Umschlag meinem Bruder gegeben
she has the envelope my brother given
‘She gave my brother the envelope’
Number (singular and plural), person (first, second familiar, second formal, and third),
gender, case (nominative, genitive, dative, and accusative), and tense are indicated by
means of grammatical morphemes, usually suffixes.
12
teacher-MALE’ (Clahsen 1989:904). They also demonstrate problems with

number marking, in the sense that they fail to indicate number, as
required, on the article (Clahsen 1989:903-904). As regards case marking,
German-speaking children with SLI appear to have a binary system –
nominative and non-nominative (either dative or accusative) (Clahsen
1989:906) – even though the grammar of adult speakers contains four
cases, which are indicated by means of grammatical morphemes. In the
language of German-speaking children with SLI, case is not indicated, as
required, on the article and the adjective and the noun,12 but only on
one of the three, mostly on the article (Clahsen 1989:905). Furthermore,
when case is indicated, it is often done incorrectly, as illustrated in
example (3), from Clahsen (1989:905).
(3) Target:
und den dosse Tommel und die grosse Trommel
and the-MASCULINE-ACC big drum and the-FEMININE-NOM/ACC big
drum
‘And the big drum’
German-speaking children with SLI do occasionally omit verb inflection

– as is the case for Dutch-speaking children with SLI – but are more
likely to use ungrammatical yet inflected forms (Clahsen 1989:907;
Lindner and Johnston 1992:124). Participles appear to be inflected
correctly most of the time, with the only source of errors being so-called
strong verbs classified incorrectly as weak verbs (Clahsen 1989:907;
Clahsen and Rothweiler 1992:13). This leads, for example, to the past
participial form of the strong verb gehen ‘go’ being phonologically realised
as (i.e., receiving the sound form) *gegeht instead of as gegangen. Clahsen
(1989:908) and Clahsen and Rothweiler (1992:27) conclude that it
appears that German-speaking children with SLI do not have a general
morphological deficit as regards verb morphology; rather, these children
experience problems with subject-verb agreement specifically. According
to Clahsen (1989:909), what have previously been termed “verb
placement errors” in the language of German-speaking children with SLI
could, in fact, be viewed as secondary effects of the children’s problems
with verb inflection, specifically with subject-verb agreement: Verb-final
patterns are used with uninflected verb forms or infinitives, whereas
12Indicating case on nouns pertains to genitive case, which is indicated on singular
masculine or neuter nouns, e.g., der Mann ‘the-NOM man’ but des Mannes ‘the man’s’.
13
verb-first or verb-second is (correctly) used with inflected verbs, verbs in

the imperative form, irregular verbs, modals, and auxiliaries (Clahsen
1988:205ff).13
Lindner and Johnson (1992) compared German- and English-speaking

children with SLI and found (i) that there is a trend – which could not be
confirmed statistically – for German-speakers with SLI to show less
impairment on tests of grammatical morphology than their English-
speaking counterparts, and (ii) that the gap between the performance on
measures of morphology and vocabulary was statistically significantly
smaller for these German-speaking children than for the English-
speaking ones. Lindner and Johnson (1992:124) explained their results by
stating that children pay more attention to that which is most useful for
their current communicative needs: As basic meaning relations are
identified primarily by morphological cues in German and primarily by
word order in English, German-speaking children with SLI fare better in
terms of grammatical morphology than do English-speaking ones. This
seems to contradict Clahsen’s (1989:916) finding that German-speaking
children with SLI use word order to distinguish grammatical relations.
Swedish-speaking14 children with SLI have greater difficulty with the use
of genitive inflections, indefinite articles, and article-adjective-noun
constructions than do MLU- and age-matched controls (Leonard,
Salameh, and Hansson 2001). Both omissions and substitutions occur.
13 Note that optional infinitives are normal when children are still in the process of
acquiring non-null subject languages. The occurrence of optional infinitives per se is

therefore not an indication of SLI; what is an indication of SLI is that the grammar of
German children with SLI present with this (normal) phenomenon for a longer time
than does the grammar of typically developing German children.
14 Swedish, a North Germanic language, is verb-second, without overtly indicated
subject-verb agreement (Platzack 2001). The language differs from Dutch and German
in the sense that Swedish demonstrates an SVO word order in embedded sentences,
and not an SOV one. As stated in Hansson and Nettelbladt (1995:590), verbs, nouns,
and adjectives may appear as bare stems, and, when inflected, such inflections take the
form of suffixes, sometimes consonantal and sometimes syllabic in nature. For
instance, the stem titta ‘look’ takes the suffix -r in its present tense form (tittar), whereas
the stem bygg ‘build’ takes –er (bygger). Article-adjective-noun agreement occurs with
regard to gender, number, and definiteness – compare en stor stol ‘a big chair’ with den
stora stolen ‘the big chair’ and with att stort tåg ‘a big train’ (from Leonard, Salameh and
Hansson (2001:620,621). Nouns are inflected for plural and case (nominative and
genitive).
14
The omission of articles occurs more frequently in article-adjective-noun

target constructions than in article-noun ones.
Swedish-speaking children with SLI appear to be less proficient than age-

matched controls with regard to present tense inflection (Hansson and
Leonard 2003:364), past tense inflection, and irregular past tense forms
(Hansson and Leonard 2003:365). These children with SLI use copula,
auxiliaries, and modals (viz. the verbs that are considered to be
grammatical morphemes) less often than their MLU would predict
(Hansson 1997:209). They also experience more problems with marking
perfect tense, which is done by means of an auxiliary, than they do with
marking the simple (past and future) tenses, which is done by means of
suffixation (Håkansson 1998:319-320; Hansson and Nettelbladt
1995:593).
Problems with word order appear to be a common characteristic of the

language of Swedish-speaking children with SLI (Hansson and
Nettelbladt 1995:592). In sentences in which topicalisation occurs and in
question constructions, the obligatory inverted word order is not
consistently used (Hansson and Nettelbladt 1995:595), and the verb is
not consistently placed in the second position in negated main clauses
(Håkansson 1998:320-321). For example, a 6-year-old boy with SLI
produced the utterance in (4), where the verb ‘sleeps’ should have
occurred before ‘not’ (from Håkansson 1998:321), as is indicated in the
grammatical sentence produced by a non-impaired 3-year-old.
(4) 6-year-old with SLI: 3-year-old without SLI:

hon inte sover hon hoppar inte
she not sleep he jump not
‘She does not sleep’ ‘He does not jump’
Furthermore, Swedish-speaking children with SLI omit subjects, verbs,

and complementisers more frequently than do MLU-matched controls
(Håkansson 1998:321-322; Hansson and Nettelbladt 1995:594). This led
Hansson and Nettelbladt (1995:595) to conclude that syntax, in addition
to grammatical morphology, is a problem area for Swedish-speaking
children with SLI.
15
2.3.2.2. Romance languages

Paradis and Crago (2000:840) found that French-speaking15 children
with SLI produce significantly fewer finite verbs in obligatory contexts
than do age-matched controls. The children with SLI also produce
significantly fewer past and future tense constructions in obligatory
contexts than do age-matched controls, substituting these constructions
mostly with an infinitival verb form but also at times with the present
tense form (Paradis and Crago 2000:841; Rose and Royle 1999:87; also
see Jakubowicz 2003:189-190). Rose and Royle (1999:77) furthermore
found that French-speaking people with SLI, but not controls, have
difficulty with correcting ungrammatical sentences if these sentences
contain agreement and tense violations.
Regarding Italian,16 Bortolini, Caselli, Deevy, and Leonard (2002:79)

state that (i) most of the differences between Italian-speaking children
with SLI and MLU-matched controls revolve around function words –
an example is that the language of Italian-speaking children with SLI
demonstrates a lower percentage of articles in obligatory contexts than
does that of MLU-matched controls (Leonard, Sabbadini, Volterra, and
15 French is a Romance language with overt verb movement, although to a lower
position than that found in German, Dutch, and Swedish main clauses. French word
order is SVO (as can be seen in examples (i) and (ii)), except when the object is a
pronoun, in which case the word order is SOV (as can be seen in example (iii)). Tense,
number, person, and agreement are marked overtly. Case is indicated on pronouns.
(i) La femme voit la fille
the woman sees the girl
‘The woman sees the girl’
(ii) Les femmes voient les filles
the women see the girls
‘The women see the girls’
(iii) L'homme l'a vue
the man her saw
‘The man saw her’
16 Italian is highly inflected: As regards verbs, no bare stems are allowed; every finite
verb is marked for both person and number. Regular past tense is indicated by means
of an auxiliary verb and the past participial form of the main verb, the latter formed for
some verbs by changing the word-final –are, –ire, and –ere into –ate, –ito, or –uto
(Leonard, Sabbadini, Leonard, and Volterra 1987:240). There is overt gender and
number agreement between nouns, possessive pronouns, and adjectives (Leonard et al.
1987:236,239). Regular singular-plural distinction is made by means of a word-final
vowel change (Leonard et al. 1987:239), as in libro ‘book’ – libri ‘books’ or palla ‘ball’ –
palle ‘balls’.
16
Leonard 1988:45) – and (ii) only a few select inflections produce

differences between these two groups. Present third-person plural
inflection on verbs is one of the inflections that do pose problems for
Italian-speaking children with SLI (Leonard, Sabbadini, Leonard, and
Volterra 1987:241-242). Because Italian does not allow bare stems, these
children do not omit the inflection but rather substitute it, usually with
the present third-person singular inflection (Bortolini, Leonard, and
Caselli 1998:11, 2002:80) and less often with the infinitival form
(Bortolini et al. 1998:11). Leonard et al. (1987:244-245) furthermore
found that regular past tense marking and copulas have a low percentage
of use in obligatory contexts in the language of Italian-speaking children
with SLI.
2.3.2.3. Other languages

The remaining five languages to be reviewed here are too diverse to be
grouped together according to language family. Hebrew-speaking17
children with SLI show lower percentages of use than do age-matched
controls of the free-standing accusative case marker et and of
17 Hebrew, a Semitic language, has a rich system of overt grammatical morphology.

Verbs, nouns, and adjectives are always inflected; bare stems do not occur. Stems,
which convey core meaning, almost without exception consist of a tri-consonantal root,
to which syllabic prefixes, syllabic suffixes, and vowel infixes are added. Grammatical
distinctions are made through this affixation, through word-internal vowel changes, or
through both (Dromi et al. 1993:761,763; Leonard 1992:124). Finite verbs are inflected
for number and gender in the present tense form, and for number, gender, and person
in the past tense form (Dromi et al. 1993:763; Rom and Leonard 1990:96) as illustrated
in (i). Nouns are mostly inflected for number (specifically plural) and gender through
suffixation, as shown in (ii). Adjective-noun agreement in terms of number and gender
occurs (Dromi et al. 1993:764; Rom and Leonard 1990:97). Accusative case is marked
by means of a free-standing morpheme (viz. et) when the noun is specific (Rom and
Leonard 1990:98), as shown in (iii), from Dromi et al. (1993:764).
(i) moxer moxrim
sell-PRESENT-MASCULINE-SGL sell-PRESENT-MASCULINE-PL
maxarta mexartem
sell-2NDPERS-MASCULINE-SGL-PAST sell-2NDPERS-MASCULINE-PL-PAST
(ii) xaver xaverim
friend-MASCULINE-SGL friend-MASCULINE-PL
uga ugot
cake-FEMININE-SGL cake-FEMININE-PL
(iii) Ha-jeled roxets et ha-mexonit
the boy washes accusative-marker the car
‘The boy washes the car’
17
connectives, which take the form of prefixes (Rom and Leonard

1990:100). Substitutions of these morphemes do occur, but the most
frequent error is that of omission. Regarding the use of present and past
tense verbal inflections, Rom and Leonard (1990:100) found that
Hebrew-speaking children with SLI and age-matched controls perform
similarly on plural noun inflection and adjective-noun agreement, with
errors involving mainly substitution. However, Dromi et al. (1993:766)
found that Hebrew-speaking children with SLI fare worse with the use
of present verb inflections, present-past verb contrasts, and singular-
plural noun contrasts than do age-matched controls.
There is some limited information available on how SLI presents itself in

Inuktitut.18 Crago and Allen (1994) performed a single case study in
which the language of a 64-month-old Inuktitut-speaking child with SLI
was compared to that of a typically developing peer and a 25-month-old
MLU-matched control. They found a lower percentage of use of the
following verb-internal or noun-internal morphemes to be characteristic
of the language of the child with SLI: (i) verbal inflections to mark
person, number, and modality; (ii) verb-verb affixes to mark causative,
passive, and adverbials; and (iii) verbalisers to change verbs to nouns.
However, the child also inappropriately used –mi, which is a suffix
frequently used on nouns – and occasionally on verbs as an internal
morpheme meaning “also” – as a suffix on verbs and locatives, where it
18 Inuktitut is an Eskimo-Aleut, null-subject language. As explained by Crago and Allen
(1994:206), “It is a highly polysynthetic language with an ergative case marking system
and prolific nominal and verbal inflectional paradigms. Verbal inflection agrees with
both subject and object for four persons, three numbers, and ten verbal modalities.
Nominal inflection represents eight cases and three numbers, and the possessive
paradigm encompasses four persons and three numbers. In addition, there are over
1000 verb- and noun-internal productive morphemes that serve as nominalisers,
verbalisers, valency-changers, and modifiers. This language has no uninflected
infinitival form.” Note that there is no inflectional form for tense (Crago and Paradis
2003:100). Examples of Inuktitut utterances are found in (i) and (ii), from Crago and
Allen (2001:64,67).
(i) Qailangannginavit = Qai langa nngit gavit
come FUTURE NEGATION CAUSATIVE-2NDPERS-SGL-SUBJECT
‘You won’t come’
(ii) Nuvujaaluk paaniittuq = Nuvujaq aluk pa ani it juq
cloud big-ABSOLUTE-SGL upthere LOCATIVE be PARTICIPATIVE-3rdPERS-SGL-
SUBJ
‘The big cloud is up there’
18
can never appear in the adult grammar. Furthermore, the child at times
used overt pronouns instead of inflecting the verb. For instance, in the
two examples given by Crago and Allen (1994:210), the child used the
pronoun ivvit ‘you’ instead of inflecting the verb and locative for second-
person and subject, amongst others, which would have been the norm in
Inuktitut, seeing that it is a pro-drop language. This lack of verbal and
locative inflection19 (i.e., the use of bare stems) “contradict[s] the
polysynthetic nature of her language and [is] highly irregular” (Crago and
Allen 1994:210). Lastly, the child also did not use any passive
constructions in the 200 utterances examined. This is highly unusual, as
Allen and Crago (1993:115) found that typically developing 2- and 3-
year-old Inuktitut-speakers use passive constructions frequently.20
Greek-speaking21 children with SLI experience problems with

inflectional morphology, specifically with mapping affixes onto stems
(Dalalakis 1997). According to Dalalakis (1997:121), this is because the
children “have trouble judging where the root ends and the affix begins
which is reflected in the root boundary errors”. In an earlier study
(Dalalakis 1994), it was found that Greek-speaking children experience
problems with tense marking (making use of substitutions but not
omissions) and with pluralisation (making use of substitutions and also
omissions).22 Tsimpli and Stavrakaki (1999:71-72) found that correct
subject-verb agreement in obligatory contexts is low for second-person
singular and second-person plural. Dalalakis (1994:225) also found that
19 As stated by Crago and Allen (1994:209), the Inuktitut locative system is quite
complex: “Locatives require a verbaliser and then a verbal inflection in certain

obligatory pragmatic and semantic contexts but not in all contexts.”
20 An average of 3.7 passives per hour was recorded.
21 Greek is a Hellenic language. It is highly inflected and does not allow bare stems.
Nominals are marked for case, gender, number, and, in the case of pronouns, for
person. Verbs are marked for number, person, tense, aspect, voice, and mood
(Dalalakis 1994:217). It appears that there is still little consensus regarding the
unmarked word order in Greek. VSO has been taken to be the basic order by
Alexiadou and Anagnostopoulou (1998), but SVO orders also occur, and VOS is seen
to be an order derived via so-called short object shift (cf. Alexiadou 1999). Greek is a
pro-drop language (as shown in example (i) – from Tsimpli and Stavrakaki 1999:43),
but subject pronouns may be used emphatically for emphasis (Dalalakis 1994:217).
(i) Parakoluthisa dhialeksis
attended-1stPERS-SGL lectures
‘I attended lectures’
22 For another study on pluralisation, with similar results, see Dalalakis (1999).
19
Greek-speaking children with SLI experience problems with derivational

morphology; again, errors consist mostly of substitutions. These children
also fare more poorly than MLU- and age-matched controls in terms of
forming compound nouns (Dalalakis 1999).
Concerning production of wh-questions, Stavrakaki (2002) found that

Greek-speaking children with SLI fare significantly more poorly than
MLU-matched controls. They also have difficulty understanding
sentences containing passive constructions and negation. Furthermore,
their ability to judge whether or not a sentence is grammatical, as well as
their ability to correct ungrammatical sentences, are impaired (Dalalakis
1994:224). These children also interpret semantically reversible relative
clauses23 qualitatively differently to MLU- and age-matched controls
(Stavrakaki 2001).
Japanese-speaking24 children with SLI, when compared to age-matched

controls, experience difficulties producing morphology related to tense
and aspect (Fukuda and Fukuda 1994:163-168). The children with SLI
also fare worse than the controls in grammaticality judgements of
ungrammatical sentences where the ungrammaticality is caused by
incorrect or omitted morphemes related to case, tense, and aspect.
Furthermore, Tanaka Welty, Wanabe, and Menn (2002:185) found that
Japanese-speaking children with SLI produce fewer case particles in
23 An example of such a construction is given by Stavrakaki (2001:427): The woman is
kicking the girls that hold the man, in which the girls and the man could have been reversed.
Although this reversal would lead to the sentence having a different meaning, the
sentence would still be interpretable.
24 Japanese is agglutinative, with rich overt morphology. Regarding verbs, bare stems do
not occur. Verbs are marked for, among other things, tense, aspect, voice, negation,
and causation (Tanaka Welty, Wanabe, and Menn 2002:177), but not for number,
person, or gender; i.e., subject-verb agreement is not indicated overtly (Fukuda and
Fukuda 1994:153-155; Tanaka Welty et al. 2002:177). Passives are formed by suffixing a
passive morpheme to the verb stem (Fukuda and Fukuda 2001:309). Tense is indicated
by means of overt morphology on both verbs and adjectives, and aspect is indicated
overtly by verbal affixation followed by an auxiliary. All lexical noun phrases (NPs) are
marked for one of eight cases. Furthermore, overt and null pronouns occur (Kanno
1998:1126). A Japanese sentence illustrating some of these properties is given in (i),
from Fukuda and Fukuda (2001:309).
(i) Kazuo-ga Tatako-ni os-are-ta
Kazuo-NOM Tatako-DAT push-PASS-PAST
‘Kazuo got affected by the event that Tatako pushed him (Kazuo)’
20
obligatory contexts and that, where case particles are produced, a greater
percentage is incorrect, compared to age-matched controls. Verb
morphology is also used less, and more often incorrectly, by the children
with SLI than by MLU- and age-matched controls (Tanaka Welty et al.
2002:186-187).
Moreover, Fukuda and Fukuda (1994:162,164) found that Japanese-

speaking children with SLI fare significantly worse than age-matched
controls on the comprehension of passive constructions as well as on
correctly performing grammaticality judgements on illicit passive
constructions.
As is the case for Inuktitut, there is limited information available on how

SLI presents itself in Zulu.25 Demuth and Suzman (1997) did a single
case study on a 31-month-old boy whose language development was
regarded by his family to be delayed. When comparing his language to
that of two typically developing Zulu-speaking children – one younger
(24 months old) and the other age-matched (31 months old) – it was
found that his language was not merely delayed, but that it also deviated
in significant ways from that of the typically developing children. Unlike
the younger typically developing child, the boy with SLI did not produce
possessives and demonstratives at all. He did use nouns, with which he
used one of two prefixes. However, noun class prefixes were omitted
twice as often as in the language of the younger typically developing
25 Zulu, a Nguni language, is a pro-drop language which demonstrates very little overt
movement, and has an elaborate system of prefixes (Suzman 2002:156). As stated by
Demuth and Suzman (1997:125), Zulu is a head-initial language – “that is, nominal and
verbal modifiers follow the noun and verb respectively, and grammatical morphology is
prefixed to both nouns and verbs”. Verbs are inflected for number, tense, aspect,
agreement, negation, and mood, amongst others (Suzman 2002:158). There are 15 noun
classes, each with its own prefixes for marking possession, adjectives, subject
agreement, object agreement, and noun class number. The prefix for marking noun
class number, in fact, consists of a prefix and a pre-prefix, the latter being a vowel
which is identical to the vowel of the prefix. Demuth and Suzman (1997:125) give the
example in (i), where the noun class marker for the class-8 noun consists of the pre-
prefix i- and the prefix –zi-:
(i) Izi-cathulo za-mi ezi-bili zi-lahlek-ile
noun class 8-shoes possession 8-my numeral 8-two 1stpers-sgl-subject 8-lose-perfect aspect
‘My two shoes are lost’
Note that the hyphens above serve to demarcate affixes and stems and do not occur in
standard Zulu orthography.
21
child (Demuth and Suzman 1997:132). Subjects were prefixed with the
same two prefixes as the nouns were. Verb phrases (VPs) mostly
consisted of single verbs (not of a verb and a DP), and tense was
indicated by overgeneralising one tense marker (Demuth and Suzman
1997:130). Suzman (2002:162) studied the language of a 66-month-old
Zulu-speaking girl with SLI and found her noun morphology to be
similar to that used by a younger (41-month-old) typically developing
child, with the exception of lower use of object markers and relative
prefixes by the child with SLI. Furthermore, Suzman (2002:163-165)
reported that both of these Zulu-speaking children with SLI primarily
produced simple sentences.
2.3.2.4. Summary: How SLI presents itself in languages other

than English
It appears that, regardless of the typology of the language which they
speak, children with SLI demonstrate problems with grammatical
morphology, omitting and/or substituting grammatical morphemes
and/or function words.
2.4. THEORETICAL ACCOUNTS OF SLI
Although SLI is not a new field of study, the development of theories to

account for the characteristics of SLI is a relatively new focus in this
field. At present, accounts of SLI fall into two main groups, which may
be referred to as “linguistic accounts” and “processing accounts”.
Briefly, linguistic accounts assume a deficit in the linguistic knowledge of
the children with SLI, whereas processing accounts assume a limitation
in processing capacity. For the purposes of this study, we will focus on
linguistic accounts only,26 particularly on that of Wexler and colleagues
26 There are several types of processing accounts. Some view SLI to be the result of
limitations in information-processing capacity. This capacity can be limited in a number

of ways, for example in terms of space (memory), energy (“fuel” available to complete a
task), and speed (rate at which information can be processed) (see, e.g., Kail 1994; Kail
and Salthouse 1994; Roediger 1980). As stated by Thal (1999), the limited processing
capacity hypothesis is applicable to a wide range of tasks and operations (not all of
them language-related), but it is precisely because of its generality that the hypothesis
has difficulty explaining the exact nature of SLI. Because this study aimed to establish
and account for the characteristics of SLI in Afrikaans, an account centred around a
general processing limitation was not considered, as it was assumed that such a general
account would not be specific enough to explain and account for the exact nature of
22
(1995, 1996, 1998), Van der Lely (1994, 1996, 2003, 2004), and Gopnik
(1994a). In this section, the content of each account is given. After the
Afrikaans data have been presented in chapters 5 to 8, these three
accounts are evaluated in chapter 9, where the general merit of the
explanations they offer (for the characteristics of SLI discussed in
section 2.3) is examined. Thereafter, predictions that they make for SLI
in Afrikaans are set out, after which it is determined whether these
predictions were borne out by the Afrikaans data, i.e., whether these
accounts are adequate for the data on SLI in Afrikaans.27
2.4.1. The Agreement/Tense Omission Model
Wexler (1994) claimed that typically developing children, at least those

acquiring English, go through a stage in which they optionally use the
infinitival verb form in contexts where the finite verb form is required in
the grammar of adults. By “optionally” is meant that the use of infinitival
forms is alternated with that of finite forms, resulting in infinitival forms
appearing in contexts where the adult grammar only allows finite forms.
Rice, Wexler, and Cleave (1995) subsequently posited the Extended
Optional Infinitive Hypothesis (EOI) which states that children with SLI
remain in this early, normal stage in the development of tense marking of
verbs – either for an extended period or not progressing past it at all. In
contrast, typically developing children progress to a more advanced stage
in which finite verb forms are consistently used in obligatory contexts.
The predictions of the EOI have to be understood against the

background of Wexler’s (1994) claims regarding the use of overt tense
marking by typically developing children (see Wexler 1994:312 for a
summary): In the Optional Infinitive stage, these children sometimes
refrain from overtly marking tense by means of a grammatical
SLI in Afrikaans. Furthermore, the more “narrow”, or specific, processing accounts

(such as the Surface Hypothesis of Leonard 1989; Leonard, Eyer, Bedore, and Grela
1997) have shortcomings of a nature that makes their cross-linguistic testability
questionable (cf. Leonard 1994; Leonard and Eyer 1996).
27 The accounts in question are presented below as the authors set them out. However,
in chapter 9, they are critically commented on from the viewpoint of Minimalist syntax,
the reason being that the framework of assumptions and concepts of Minimalist syntax
are accepted for the purposes of this study; this is the framework in which the
Afrikaans data will be analysed.
23
morpheme in matrix clauses (i.e., in clauses in which tense marking is

obligatory). The morphemes by which tense is marked overtly in
English, and which are (sometimes) omitted even by typically developing
children in the Optional Infinitive stage, include the following (Rice and
Wexler 1996:1241):28 (i) the third-person singular -s, as in She sings;29 (ii)
the regular past tense morpheme -ed, as in She/They played the game; (iii) be
forms, as in She is teaching him, They are teaching him, and He was happy; and
(iv) do forms, as in They do spot him from time to time, where do is used in an
emphatic sense, or as in She does not spot him, where the auxiliary verb does
is used unemphatically.30
Wexler (1994:335) claims that children fail to mark tense overtly because
they treat the grammatical category tense (T) – i.e., the head of the tense
phrase (TP), on the split-INFL Hypothesis of Pollock (1989) – as
optional. On this view, if TP is present in a derivation, then the verb will
move from the VP to the TP so that the verb can be marked for tense.
For certain languages, such as English, it is assumed that this move takes
place covertly. Evidence for this comes from the fact that the finite verb
occurs after the adverb in a sentence such as He always brushes his teeth,
with the adverb taken to mark the left-periphery of the VP. However, if
TP is not present, the construction will be treated as an infinitival one.
This would mean that the verb does not move to TP (because TP is not
present), resulting in the verb not being marked for tense. Hence, the
grammatical tense marker is omitted in the phonological realisation, with
the verb instead displaying the infinitival form. For this reason, children
with SLI often produce utterances like *Yesterday we walk home, where the
grammatical tense marker -ed has been omitted, rather than Yesterday we
walked home.
28 Interestingly, Rice and Wexler do not include have forms, as in, for example, She had
seen the cat, where the temporal auxiliary had appears in its past tense form.
29 The plural equivalent, namely sing in They sing, also represents a finite verb form, even
though tense is not indicated overtly. In She makes him sing, the verb sing is, however, in
the infinitival form (whereas makes is in the finite form), as is sing in She likes to sing
(where likes is finite).
30 In these examples, the do forms are marked for tense and spot is in the infinitival form
throughout. As in the examples in note 29, finite and infinitival verb forms may be
found in the same clause.
24
According to Rice and Wexler (1996:1240), the term “omitted” in the

above analysis is an abbreviation for “not apparent in the surface forms,
i.e., in the phonological forms”. They stated that the absence of surface
tense markers can be attributed to the non-occurrence of the functional
category tense in a given derivation. It is thus not the case that tense is
totally absent in the grammars of children with SLI and that surface
tense markers are always omitted. Rather, on their view, these children’s
grammars allow utterances with TPs as well as utterances without TPs,
whereas the adult grammar, in this context, would accept as grammatical
only utterances with TPs. Hence the qualification “optional” in Rice et
al.’s (1995) Extended Optional Infinitive hypothesis: The tense-marked
phonological form is optional in the grammars of children with SLI – at
least for a markedly longer period than is the case for typically
developing children.
In recent years, both the Optional Infinitive and the Extended Optional
Infinitive accounts have been expanded to accommodate the observation
by Schütze and Wexler (1996) and Wexler et al. (1998) that the
inappropriate use of infinitival verb forms seems to be related to
pronoun case error by English-speaking children. Specifically, they
observed that these children’s “subject case marking in clauses with a
fully-specified INFL is essentially perfect” (Schütze and Wexler
1996:672); by contrast, all non-nominative subjects occur with optional
infinitives.31 This led to the proposal that optional infinitives can result
from the underspecification of either the tense or agreement features (or
both) in children’s grammars.32 Accordingly, this underspecification can
result in the combinations in (5) if the lexical entries in (6) are assumed
(based on Schütze and Wexler 1996:678,679):
31 Wexler and Schütze (1996:677) assume “the separation of T and Agr”, where “Agr”
refers to subject (S) agreement. In other words, a clause is taken to contain an
AgrSP(hrase) and a TP, which together replace the category inflectional phrase (IP) of
earlier theories of phrase structure. Wexler and Schütze (1996:677) further assume that
“Agr, not T, assigns/checks NOM case”.
32 It is assumed that by “underspecification” Wexler and colleagues mean that the
feature’s value is negative. They use several terms in this regard, presumably all
synonyms: “missing (or have a negative value)” (Schütze and Wexler 1996:678), and
“omitted” and “deleted” (Wexler 1998).
25
(5a) [+tense, +agr] (Nominative case assigned) he walks

(5b) [+tense, -agr] (Nominative unassignable; him walk,
default accusative surfaces) him walked
(5c) [-tense, +agr] (Nominative case assigned;
agreement invisible)33 he walk
(5d) [-tense, +agr] (Nominative case unassignable;
genitive case assigned) his walk(ing)
(6a) [+tense=present; +agr=3sg] → -s

(6b) [+tense=past] → -ed
(6c) [-tense; -agr] → Ø
Under Schütze and Wexler’s (1996:678) analysis, in the Optional

Infinitive stage, either tense or agreement may be independently missing
(or have a negative value) in finite environments. In other words,
Schütze and Wexler (1996) and Wexler et al. (1998), accepting the
assumption that agreement licenses nominative case, propose that either
tense or agreement or both may be missing in the grammars of children,
giving rise to optional infinitives. If both tense and agreement are
present (and marked), the utterance will contain a nominative subject. If
agreement is missing, nominative case will not be licensed, and the
structural subject will occur in the default case of the language, which,
for English, is taken to be the accusative case (Schütze 1999:750). This
proposal has been termed the “Agreement/Tense Omission Model”
(ATOM).34 The EOI has thus been amended to become a two-factor
33 Non-nominative subjects are said to be produced with non-agreeing verbs when

Agreement is “present but hidden” because Tense is missing (Pine, Joseph, and Conti-
Ramsden 2004:913).
34 Rispoli (2005:94) summarises the proposals as follows: When the subject pronoun
occurs in the accusative case form (e.g., when him occurs instead of he), the main verb
must lack tense/agreement marking. However, when the subject pronoun is correct,
there may be agreement, but not necessarily; i.e., if a verb lacks agreement, one cannot
predict the form of the subject pronoun. The examples in (i) and (ii) – found in Charest
and Leonard (2004:232-33) – of possible utterances according to the ATOM illustrate
Rispoli’s latter point:
(i) Third person singular context: (ii) Past tense context:
(ia) She plays +agr, +tense (iia) She played +agr, +tense
(ib) She play +agr, -tense (iib) She play +agr, -tense
(ic) Her play -agr, +/-tense (iic) Her play -agr, -tense
(iid) Her played -agr,+/-tense
26
account of SLI. For children with SLI, the following has been observed
(summary based on Lin 2006):
(i) Children with SLI make more frequent case-marking errors on
subjects than do language-matched controls.
(ii) Children with SLI (English-speaking ones, at least) use the
accusative case form of the subject where adults require the
nominative form.
(iii) Case-marking of subjects is dependent on whether or not clauses
contain an auxiliary or main verb which agrees in person and
number with the subject.
(iv) Children with SLI may leave verbs underspecified for tense and
agreement in finite contexts, resulting in the production of optional
infinitives with accusative subjects.
2.4.2. The Representational Deficit For Dependent Relations

Hypothesis
According to Van der Lely’s (1994) account of SLI, known as the

RDDR, the language deficits of children with SLI stem from a selective
impairment in establishing the structural relationship between dependent
constituents. This impairment leads to some obligatory constituent
movements being optional in the grammar of children with SLI.
Before discussing the details of Van der Lely’s account, it may be useful
to first consider what is meant by “movement” within Minimalist syntax.
The operation Move35 takes place because uninterpretable grammatical
features, specifically those associated with functional categories like T,
complementiser (C), and others, need to be checked. If these features are
not checked, the expression, according to the principle of Full
Interpretation (cf. section 3.4.1), will not be interpretable – i.e., the
derivation will crash on one or both of the interface levels. In the
sentence The children played rugby, the head T of TP has a uninterpretable
V feature which must be checked against the tense feature of played. For
this reason, played moves to the TP (covertly, in the case of English). The
T thus attracts the V feature, and with it the lexical item which contains
35 Cf. note 77.

27
that feature (in this case, played).36 In short then, feature checking is
driven by the need for (full) semantic and phonetic interpretation, and is
accomplished by means of Move. According to Van der Lely (2003:126),
“a dependent structural (syntactic) relation is formed in a sentence for
the purpose of linking and checking (matching, copying, or moving)
grammatical features associated with lexical items (or constituents)”; or,
in what Van der Lely calls “more theory-neutral terms”, “this syntactic
dependency occurs when one sentence constituent ‘looks for’ a ‘sister
constituent’ for feature checking/matching/copying”.
Van der Lely (2003:127) claims that the linguistic deficits in children with
SLI should not be ascribed to the total absence of Move, but rather to
the optionality37 of this operation in the grammar of such children. She
considers two principles involved in movement. The first, which is
obeyed by children with SLI, is that constituents only move if (i) they
have features that need to be checked, or (ii) they have features against
which those of some other constituent must be checked. Thus, a verb
with the feature [+past] will only move to the TP if the TP has a V
feature that must be checked, and not for any other reason. On the
RDDR, children with SLI experience problems in establishing the
dependent relationship between different constituents and thus in
establishing the syntactic domain in which the feature can be checked.
Therefore, these children often fail to move the constituent to the
correct syntactic domain for checking purposes. According to Van der
Lely (1996:246), the past tense feature of a verb can, in the grammar of
children with SLI, be checked either against the finite TP or against
another constituent which is marked for “time” (such as an adverb of
time). For this reason, it often seems as if the phonological realisation of
such features is optional in the language of children with SLI. Moreover,
Van der Lely (1996:246) argues that this optionality leads to the omission
of grammatical morphemes in obligatory contexts and not to their
insertion in inappropriate contexts. According to Van der Lely
(1996:246), this means that a checked tense feature will be realised
correctly in the phonological form; however, an unchecked feature will
36 See section 3.4.3 for a brief discussion of the proposal that features can be moved on
their own, without the lexical item of which they form part moving with them.
37 The reason why Move would be optional in the grammar of children with SLI is not
obvious.
28
not appear in the phonological form of the utterance and therefore the
verb will appear in the infinitival form.
The second principle involved in movement, according to Van der Lely

(2003:127), is that Move is forced if a constituent’s features have not yet
been checked. She claims that this principle is “missing” in the grammar
of children with SLI. This absence results in Move being an optional
operation in the grammar of children with SLI; accordingly, some
features are left unchecked.38 To illustrate a tense/Q-feature error and a
wh-movement error, Van der Lely (2004) gives the following example.
(7) Target:
What did colonel Mustard had What did colonel Mustard have in his
something in his pocket? pocket?
In this example, a wh-movement error called “gap-filling” (i.e., leaving

the referential phrase in the base position; Van der Lely 2004) can be
observed. Also, while do-support does occur, the past tense form had is
produced instead of have.
2.4.3. The Feature Deficit Hypothesis
A third linguistic account of SLI is Gopnik’s (1994a) Feature Deficit

Hypothesis, originally termed “the Feature Blindness Hypothesis” and
then, in revised form, the “Implicit Grammatical Rule Deficit” (Gopnik
1990a, 1990b).39 On Gopnik’s hypothesis, SLI is the result of a deficit in
38 Note that Van der Lely does not state that the absence of this second principle
implies that an item can move even though its features have already been checked.
39 In contrast to the analyses set out in sections 2.4.1 and 2.4.2, Gopnik’s analysis of
impaired language use is not limited to that of children. Gopnik (1990b, 1994a, 1994b)
presents the results of a study of a family comprising three generations as evidence for
what eventually became known as the Feature Deficit Hypothesis. She and her
colleagues administered a battery of 14 tests to the 30 family members (one
grandmother, her five children, and her 24 grandchildren), and also analysed samples of
their spontaneous language use (spoken and written). They found that 16 of the family
members (the grandmother, all three of her daughters, one of her sons, six of her 13
granddaughters, and five of her 11 grandsons) performed significantly more poorly
than the rest on the four tests evaluating syntactic-semantic abilities, but not on the
other 10 tests.
As an example, Gopnik (1990b:715) mentions that the performance of the two groups
did not differ significantly when their knowledge of possessive relations (The baby’s
29
the knowledge of rules regarding the morphological marking of a specific

class of linguistic features (which includes number, gender, person, tense,
and aspect). This deficit is claimed to result in an inability to formulate
implicit grammatical rules. For example, according to the Feature Deficit
Hypothesis, English-speaking children with SLI will not acquire the rule
that the verb ends in an -s if the subject is in the third-person singular
form.
Gopnik claims that children with SLI can compensate for the absence of
these implicit rules in two ways. The first is through rote learning
(Gopnik 1990b:715), whereby they must memorise all inflected forms.
For example, the regular past tense form played has to be learnt by
children with SLI in the same way as the irregular form bought is learnt by
all children, whether language-impaired or not. Typically developing
English-speaking children subconsciously acquire the rule that -ed is
added to the end of the verb to form the past tense and therefore only
mother vs. The mother’s baby), reflexive pronouns (He washes him vs. He washes himself), and
negative passive constructions (The car is not being pulled by the truck) was assessed.
It should be noted that other researchers (e.g., Leonard 1995) found that children with
SLI do experience significant problems with the possessive ’s in English. Also, it is not
clear why Gopnik’s participants did not demonstrate problems with reflexive pronouns,
whereas Van der Lely and Stollwerck (1997) found that children with SLI do experience
such problems. It could be that the tasks and visual materials used by the two research
teams to evaluate the knowledge of (non-)co-referential relationships differed from
each other in such a way as to influence the participants’ responses. Furthermore, it is
interesting that the example construction provided by Gopnik (1990b:715) to illustrate
the negative passive construction is reversible, and is therefore the type of passive
construction that Van der Lely and Stollwerck (1997) found to be problematic for
children with SLI. Again, the two different responses could be task-related: Gopnik
aimed to test negative passive constructions, and it could thus well be that the visual
test material, coupled with real-world knowledge, caused the reversibility of these
constructions to be of no consequence to its comprehension – in this case, it is
probably more likely for a truck to pull a car than vice versa.
Returning to Gopnik’s (1990b, 1994a, 1994b) results, the two groups of family
members did differ significantly in their abilities to provide the plural of nonsense
words (such as zat) and to alter tense (when, for example, they were requested to
complete Every day he kisses his nanny. Yesterday he _____) (see also Goad and Rebellati
1994; Gopnik 1994a; Ullman and Gopnik 1994). From the results of these studies, it
appeared that the family members with SLI could not acquire implicit rules and that,
where they did provide the correct surface form of words (such as verbs in the past
tense form and nouns in the plural form), they made use of forms that they had
memorised.
30
have to memorise irregular forms such as bought, saw, and went. By

contrast, children with SLI have to memorise every stem (such as play) as
well as every inflected form (plays, played, playing). The second way in
which children with SLI can compensate for the absence of implicit
rules, is by applying rules that were explicitly taught to them. For
example, when such children are cognitively mature enough, they can
consciously learn and apply rules like “add an –s if there is more than
one” (Paradis and Gopnik 1994:146).
According to Gopnik, the phonetic form of the utterance is not a reliable

indicator of the features involved. In other words, when it does occur,
the grammatical morpheme does not necessarily reflect the feature
usually associated with it. The phonetic form of a multimorphemic word
can thus be regarded as an unanalisable unit by a person with SLI.
Gopnik (1994a:77) gives the following example – a written utterance by a
person with SLI – to illustrate her point: *On Saturday I got up and I wash
myself and I get dress and I eat my breakfast and I watched TV all day and I went
to bed. The phonetic form of three of the verbs (got, watched, and went)
suggests that the person knows how to express past tense grammatically
in English, but the absence of past tense marking on the other three
verbs raises the question whether watched was not rather a memorised
form.40
2.5. CHAPTER CONCLUSION
In this chapter, some characteristics of SLI as it presents itself in English

and in a selection of other languages were outlined. The characteristics
include problems with grammatical morphology and movement
operations. Three theoretical accounts of these characteristics of SLI
were presented. On the first account, the ATOM, the optional presence
of a TP is proposed to be responsible for some of the problems with
grammatical morphology, specifically those related to finiteness. The
second account, the RDDR, states that children with SLI experience
problems in establishing the dependent relationship between different
constituents and thus in establishing the syntactic domain in which a
grammatical feature can be checked. On the third account, the Feature
40 The verbs got and went are irregular past tense forms and, as such, have to be
memorised by all speakers of English.

31
Deficit Hypothesis, SLI is the result of a deficit in the marking of certain

grammatical features (including number, gender, person, tense, and
aspect). In chapter 9, it will be determined whether the predictions that
these three accounts make for SLI in Afrikaans are borne out by the
Afrikaans data obtained in this study. In the next chapter, though, the
reader is introduced to the characteristics of (non-impaired) Afrikaans, as
well as to some of the devices of Minimalist syntax, which constitute the
syntactic framework within which the data of this study were analysed.
32
Afrikaans and its syntactic analysis
Chapter 3
Afrikaans and its syntactic analysis in terms of

Minimalist syntax
3.1. INTRODUCTION
As is clear from the previous chapter, the characteristics of SLI have

already been studied in several languages. The reasons why the present
study of SLI in Afrikaans is warranted are given in section 3.2. In section
3.3, the reader is introduced to some characteristics of Afrikaans
morphology and syntax. Finally, in section 3.4, some assumptions and
devices of Minimalist syntax – the syntactic framework used for the
analysis of the Afrikaans data in this study – are discussed. An indication
is also given of how (non-impaired) Afrikaans is analysed in terms of
Minimalist syntax. This serves to familiarise the reader with the Afrikaans
of adult speakers, so that a comparison can be made between the
utterances of the Afrikaans-speaking children in this study and that of
adult speakers of the language.
3.2. WHY STUDY AFRIKAANS?
Apart from the languages discussed in chapter 2, there are several others
in which SLI has been studied.41 The question arises as to why it should
be deemed necessary to perform yet another study on SLI in a language
such as Afrikaans.
There are several clinical and theoretical reasons for studying SLI as it
presents itself in Afrikaans. With regard to the clinical reasons, it should
be noted that no agreed-upon protocol exists for the identification of
SLI in Afrikaans-speaking children. Two of the obvious reasons for this
are (i) the lack of information on the precise characteristics of the
41For example, Croatian was studied by Ljubešić and Kovačević (1992), Hungarian by
Vinkler and Pléh (1995), and Spanish by Merino (1983) and Simon-Cereijido and
Gutierrez-Clellen (2007).
33
language of Afrikaans-speaking children with SLI, and (ii) the lack of

assessment instruments for Afrikaans-speaking children. To date, only
three standardised tests have been developed for use with Afrikaans-
speaking children, viz.:
(i) Toets vir Mondelinge Taalproduksie ‘Test for Oral Language
Production’ (TMT) by Vorster (1980). This test is not widely used,
amongst other reasons because it is seen as dated and because it
has poor test-retest reliability (according to the author himself).
Also, due to its age, this test is not informed by any recent
developments in the field of grammar.
(ii) Afrikaanse Semantiese Taalevalueringsmedium ‘Afrikaans Semantic
Language Evaluation Medium’ (AST) by Pretorius (1989). This test
evaluates the receptive and expressive semantic abilities of
Afrikaans-speaking children. The test comprises 17 subtests,
amongst others, ones evaluating receptive vocabulary; spatial
relations and preposition groups; synonyms; and antonyms.
(iii) Afrikaanse Reseptiewe Woordeskattoets ‘Afrikaans Receptive
Vocabulary Test’ (ARW) by Buitendag (1994). As its name implies,
this test evaluates the receptive vocabulary of Afrikaans-speaking
children.
Despite the fact that only three instruments have thus far been
developed for use with Afrikaans-speaking children, these instruments
are not at all routinely administered by speech-language therapists, even
if their client is Afrikaans-speaking and in the age range for which the
instruments have been standardised. There are two main reasons for this.
Firstly, it is common practice among South African speech-language

therapists to administer (mostly non-standardised) Afrikaans translations
of British or American English-medium tests. The norms obtained for
the English-speaking population for which these tests were developed
are then used to determine the language ability of Afrikaans-speaking
children. This is especially, but not exclusively, the case when receptive
morphological and syntactic abilities are evaluated,42 as none of the three
42 For instance, even though an Afrikaans-medium instrument, namely the ARW, is

available to evaluate receptive vocabulary, some speech-language therapists prefer to
use the Afrikaans translation of the Peabody Picture Vocabulary Test – Revised (Dunn
and Dunn 1981).
34
available Afrikaans-medium tests evaluates these abilities – apart from

one subtest of the AST which evaluates passive relations.
Secondly, as regards evaluating language production, the TMT is not the

speech-language therapist’s instrument of choice, for reasons given in (i)
above. Because no other test has been developed to evaluate the
expressive morphology or syntax of Afrikaans-speaking children, most
South African speech-language therapists make use of spontaneous
language sample analysis when evaluating these aspects of the language
of their Afrikaans-speaking clients.
Whether any of these three tests have been performed or not, the
general judgement of the speech-language therapist regarding slow
and/or abnormal development determines whether or not an Afrikaans-
speaking child is diagnosed with SLI. This general judgement is based on
the results of two or more of the following: (i) a detailed case-history
regarding language and other development; (ii) standardised testing in the
form of one or more of the above-mentioned three tests developed for
Afrikaans-speaking children; (iii) testing with non-standardised Afrikaans
translations of tests developed for English-speaking children; (iv)
informal testing; and/or (v) language sample analysis.
A study of the characteristics of SLI as they appear in Afrikaans will be

of clinical value, as it will enable speech-language therapists to (i) select
standardised tests which target relevant aspects of language, (ii) devise
informal test material which targets these aspects, (iii) analyse language
samples of Afrikaans-speaking children in a more sensible manner, and
(iv) plan remediation (including therapy activities) in a more informed
and focused manner.
A study of the characteristics of SLI as they appear in Afrikaans will,

however, also be of theoretical value. As mentioned in section 2.3.1, the
characteristics of SLI include problems with the use of grammatical
morphemes, movement operations, and the establishment of (non-)co-
referential relationships. Within the framework of Minimalist syntax
(Chomsky 1995a, 1998, 1999, 2001, 2002),43 these problems could be
43 Cf. section 3.4.

35
shown to be related;44 if so, a unified account of SLI in terms of a

problem with grammatical features may be proposed. On the one hand,
the problem might lie with the existence (or absence) of grammatical
features themselves in the grammar of children with SLI. On the other
hand, the problem might lie with the requirement that semantically
uninterpretable grammatical features need to be checked by means of
movement of lexical items or expressions45 – in order that a derivation
does not crash (Chomsky 1995a:230; Hornstein, Nunes, and Grohmann
2005:293).
Afrikaans differs from the other languages in which SLI has thus far
been studied: Unlike in the languages discussed in section 2.3 above, very
few grammatical features are realised phonologically in Afrikaans.
However, Afrikaans shows word order variation, amongst others, due to
scrambling and left dislocation.46 The limited phonological realisation of
grammatical features and the frequent overt movement of syntactic
constituents in Afrikaans make this language an interesting one to study,
from a theoretical perspective, when looking at the characteristics of SLI.
The question arises: If children with SLI experience problems realising
grammatical features correctly, how does SLI present itself in a language
in which grammatical features are, in any case, realised phonologically to
a very limited extent, but where overt movement – which is assumed in
Minimalist syntax to be driven by the need to check features – occurs
frequently?
To summarise then, it appears that a study of SLI in Afrikaans-speaking

children could be of both clinical and scientific import, contributing to
the (relatively limited) knowledge base on SLI in two ways. Firstly, such a
study could provide data useful for clinical practice, i.e., for the
identification, assessment, and treatment of Afrikaans-speaking children
with SLI. At present, no comprehensive study has been made of how
44 In order to achieve this for the problems with co-referential relationships, a so-called
feature based account of co-reference (see Oosthuizen, forthcoming, for Afrikaans)
needs to be assumed.
45 It is assumed that it is this need for checking which licenses movement (Belletti and
Rizzi 2002:33).
46 See section 3.3 for a more detailed discussion of some of the syntactic characteristics
of Afrikaans. For descriptive grammars of Afrikaans, see, e.g., De Villiers (1971);

Donaldson (1993); Ponelis (1979).
36
SLI presents itself in Afrikaans, and no relevant data have been collected.
Secondly, Afrikaans has properties (such as being morphologically
impoverished) useful for testing theoretical accounts of SLI based on
other languages. Hence, such a study could shed light on the explanatory
merit of various theoretical accounts of SLI.
Several researchers have set out to provide such linguistic accounts of

SLI based on the characteristics of SLI, amongst others those presented
in section 2.3 above (see, for example, Clahsen 1989; Gopnik 1990a,
1990b, 1994b; Rice and Wexler 1996; Rice et al. 1995; Van der Lely 1994,
1998, 2003, 2004). In these accounts (some of which are discussed in
section 2.4), reference is often made to grammatical features that are
phonologically observable in the language of non-impaired speakers but
are omitted or substituted in the language of children with SLI. The
following brief discussion provides an indication of how (non-impaired)
Afrikaans presents in terms of the phonological realisation of
grammatical features and in terms of movement and co-reference
relationships. This should give the reader some indication of what could
potentially “go wrong” in the language of Afrikaans-speaking children
with SLI.
3.3. SOME PROPERTIES OF THE MORPHOLOGY AND

SYNTAX OF AFRIKAANS
This section describes aspects of the morphology and syntax of

Afrikaans (some of which have already been referred to above).
Specifically, how Afrikaans presents in terms of grammatical
morphology, constituent movement, and co-referential relationships will
be discussed.
3.3.1. Grammatical morphology
The grammatical features number, person, case, and tense are realised
phonologically in Afrikaans. Semantic gender is also indicated
morphologically, but not grammatical gender, which is absent from the
language.
37
3.3.1.1. Number
In Afrikaans, agreement in terms of number is not phonologically
realised on verbs. As can be seen in the examples in (8) and (9), the verb
has the same form whether the subject and object are singular or plural.
(8a) (8b)
Die kind vra ’n vraag Die kind vra vrae
the child ask a question the child ask questions
‘The child is asking a question’ ‘The child is asking questions’
(9a) (9b)
Die kinders vra ’n vraag Die kinders vra vrae
the children ask a question the children ask questions
‘The children are asking a question’ ‘The children are asking questions’
With regard to nouns, there are no bound morphemes to indicate the

cardinal one, and there is no single default rule for forming the plural of
any noun. There are two regular plural suffixes, namely –e (as in katte
‘cats’, the plural form of kat) and –s (as in bekers ‘mugs’, the plural form
of beker). The decision whether a noun is pluralised by means of
suffixing –s or –e can be described as being rule-based, but there are
many rules determining which suffix is the correct one47 (cf. Donaldson
1993:69-84). There are also many exceptions to these rules. Hence, to a
certain extent, the correct plural form of a noun has to be learnt, i.e.,
speakers of Afrikaans must learn whether
(i) the plural form is an irregular one (such as kinders ‘children’, the
plural form of kind; vrouens ‘women’, the plural of vrou; volksliedere
‘national anthems’, the plural of volkslied; hemde ‘shirts’, the plural of
hemp; or berigte ‘reports’, the plural of berig);
(ii) the plural is formed through both a suffix and a vowel change (as
in stede ‘cities’, the plural of stad; or skepe ‘ships’, the plural of skip);
47 Two examples of such rules are

(i) Nouns ending in [a], [Ar], [Art], [«], [i¥u], [«l], [«m], [«n], [«r], [«rt], [i], [ir], [l«m], or
[r«m] are pluralised by means of affixing –s (Donaldson 1993:69). For instance,
the plural form of kamera ‘camera’ is kameras; that of skakelaar ‘switch’ is
skakelaars; and that of opinie ‘opinion’ is opinies.
(ii) Nouns containing the long vowels [A], [i¥«], [o], or [y] in a closed syllable are
pluralised by means of affixing –e (Donaldson 1993:72). For example, the plural
of plaas ‘farm’ is plase; that of skool ‘school’ is skole; and that of vuur ‘fire’ is vure.
38
(iii) the noun takes one of the two regular plural suffixes –s or –e, but
not the one specified by the rules for forming regular plurals. For
instance, the plural form of tenk ‘tank’ is tenks and not the expected
tenke (on analogy to wenke ‘tips’, the plural of wenk). Also, the plural
form of oom ‘uncle’ is ooms and not the expected ome (on analogy to
bome ‘trees’, the plural of boom).
3.3.1.2. Person
In Afrikaans, neither verbs nor nouns are inflected for person. As can be
seen from example (10), the verb remains in the same form, regardless of
the person of the pronoun.
(10)
Ek /Jy /Hy lees die tydskrif
I /you-SGL /he read the magazine
‘I am /You are /He is reading the magazine’
However, as can also be seen from example (10), personal pronouns

have different forms for first-, second- and third-person, as do
possessive pronouns (cf. section 3.3.1.4).
3.3.1.3. Gender
Agreement in terms of gender is not indicated in Afrikaans. In other
words, no distinction is made between semantic gender and grammatical
gender. In a language such as German, a grammatically female noun,
such as Katze ‘cat’, can refer to a male or female object (thus to a tomcat
or a female cat) and a grammatically male noun, such as Hund ‘dog’, can
refer to a male or female object (thus to a bitch or a male dog).48 By
contrast, semantic gender is often indicated overtly in Afrikaans.49
48 In German, gender distinctions are also sometimes lexically encoded – for instance,
Kater only refers to a male cat and Löwin only to a lioness.
49 Semantic gender is indicated on some nouns denoting people and (less commonly)
animals, mostly by the use of derivational suffixes. Speakers of Afrikaans have to learn
whether semantic gender is indicated (i) lexically, i.e., phonologically in the form of
separate/different words due to the occurrence of suppletion, as in ram ‘ram’ and ooi
‘ewe’; (ii) by compounding, as in leeuwyfie ‘lioness’ which is the female form of leeu ‘lion’;
(iii) not at all, as in outeur ‘author’ or ‘authoress’; or (iv) by affixation, as in kelnerin
‘waitress’ the female form of kelner ‘waiter’. Note that there are several suffixes to
indicate semantic gender, not only -in.
39
Example (11) indicates the lack of agreement in terms of grammatical

gender in Afrikaans.
(11a) (11b)
Die vriendelike50 man Die vriendelike vrou
the friendly man the friendly woman
‘The friendly man’ ‘The friendly woman’
3.3.1.4. Case
The form of the definite article in a very limited number of archaic
Afrikaans expressions indicates that Afrikaans is largely descended from
a language – specifically 17th century Dutch – in which case was realised
phonologically to a greater extent than it is currently in Afrikaans.
Examples of such archaic expressions are given in (12) and (13).
(12)
In der waarheid
in the-DAT truth
‘in fact’
(13)
Om den brode
for the-DAT bread(s)
‘in order to make a living’
Except in such archaic expressions, the form of the definite article is die
‘the’ and that of the indefinite article is ’n ‘a(n)’. In other words, case is
not realised phonologically on the article. However, it is realised on
singular personal and possessive pronouns. The pronoun system in
Afrikaans is summarised in (14). Note that, unlike in languages such as
French and Dutch, there is no weak (clitic)–strong distinction in
Afrikaans.51
50 In Afrikaans, some attributive adjectives are inflected with –e, specifically those
consisting of more that two syllables (‘n vinnige nota ‘a quick note’; dankbare mense
‘thankful people’) and monosyllabic ones ending in /d/ (‘n vreemde gevoel ‘a strange
feeling’), /f/ (stywe spiere ‘stiff muscles’), /x/ (die klere is droog ‘ the clothes are dry’, but die
droë klere ‘the dry clothes’), and [s] (snaakse dinge ‘funny things’) (Donaldson 1993:163-167).
51 For example, in Dutch, there is a distinction between the stong mijn ‘mine’, jouw
‘your’, zijn ‘his’ and haar ‘her’, on the one hand, and the clitic m’n, je, z’n and d’r, on the
other.
40
(14) Nominative Oblique Possessive

1st person singular Ek ‘I’ My ‘me’ My kar ‘my car’
1st person plural Ons ‘we’ Ons ‘us’ Ons kar ‘our car’
2nd person singular (familiar) Jy ‘you’ Jou ‘you’ Jou kar ‘your car’
2nd person plural (familiar) Julle52 ‘you’ Julle ‘you’ Julle kar ‘your car’
2nd person singular/plural U ‘you’ U ‘you’ U kar ‘your car’
(formal)
3rd person singular masculine Hy ‘he’ Hom ‘him’ Sy kar ‘his car’
3rd person singular feminine Sy ‘she’ Haar ‘her’ Haar kar ‘her car’
3rd person singular neuter Dit ‘it’ Dit ‘it’ Sy kar ‘its car ’
3rd person plural Hulle ‘they’ Hulle ‘them’ Hulle kar ‘their car’
In contrast to the prenominal possessive pronouns given in (14), the

substantively used possessive pronominal forms are indicated in (15) to
(17).
(15a) Myne /Joune /U s’n /Syne /Hare /Syne is weg

mine /yours-SGL-FAMILIAR /yours-SGL-FORMAL /his /hers /its be gone
‘Mine /Yours /Yours /His /Hers /Its is gone’
(15b) Ons s’n /Julle s’n /U s’n /Hulle s’n is weg
ours /yours-PL-FAMILIAR /yours-PL-FORMAL /theirs be gone
‘Ours /Yours /Yours /Theirs is gone’
(16a) Hy is nie by Stan se huis nie; hy is by myne /joune /u s’n /syne /hare
/syne
he be not at Stan possessive-marker house not; he be at mine /yours-SGL-
FAMILIAR /yours-SGL-FORMAL /his /hers /its
‘He is not at Stan’s house; he is at mine /yours /yours /his /hers /its
(house)’
(16b) Hy is nie by Stan se huis nie; hy is by ons s’n /julle s’n /u s’n /hulle s’n
he be not at Stan possessive-marker house not; he be at ours /yours-PL-
FAMILIAR /yours-PL-FORMAL /theirs
‘He is not at Stan’s house; he is at ours /yours /yours /theirs’
(17a) Dit is myne /joune /u s’n /syne /hare /syne

it be mine /yours-SGL-FAMILIAR /yours-SGL-FORMAL /his /hers /its
‘It is mine /yours /yours /his /hers /its’
(17b) Dit is ons s’n /julle s’n /u s’n /hulle s’n
it be ours /yours-PL-FAMILIAR /yours-PL-FORMAL /theirs
‘It is ours /yours /yours /theirs’
52 Julle and hulle may be reduced to jul and hul, respectively, regardless of their case.
41
Afrikaans has a possessive construction consisting of a DP with the

structure given in (18) (Oosthuizen and Waher 1994:21; see also Den
Besten 2006) and illustrated in (19) to (23), where the particle se indicates
possession and is to some extent equivalent to the English possessive ’s.
The form of se is invariant, regardless of the characteristics of the
antecedent possessor (which is phrasal). For instance, in (19), the
possessor is singular, whereas it is plural in (20). In (21), se is preceded by
a DP with a relative phrase and, in (22), by one with a relative clause.
Example (23) indicates that se can be used recursively.
(18) XP se NP
(19)
Pieter se idee
Peter possessive -marker idea
‘Peter’s idea’
(20)
Die kinders se storie
the children possessive-marker story
‘The children’s story’
(21)
Die man met die swart pak se sambreel
the man with the black suit possessive-marker umbrella
‘The umbrella of the man in the black suit’
(22)
Die ou motor wat daar staan se pap wiel
the old car which there stand possessive-marker flat tyre
‘The flat tyre of the old car standing there’
(23)
My oom met die kierie se vrou se suster se ... se kleinkind se dosent se lesing
my uncle with the walking-stick possessive-marker wife possessive-marker sister
possessive-marker ... possessive-marker grandchild possessive-marker lecturer possessive-
marker lecture
‘My uncle with the walking stick’s wife’s sister’s ... grandchild’s lecturer’s
lecture’
42
The particle se in Afrikaans is also used with the interrogative and relative
pronoun wie ‘who’ and, in informal speech, with the relative pronoun wat,
as illustrated in (24) to (26).
(24)
Wie se boek lees jy?
who possessive-marker book read you
‘Whose book are you reading?’
(25)
Ek ken die vrou wie se boek jy lees
I know the woman who possessive-marker book you read
‘I know the woman whose book you are reading’
(26)
Hier is die boek wat se bladsye geskeur is
here be the book that possessive-marker pages torn be
‘Here is the book of which the pages are torn’
3.3.1.5. Tense
Present tense, on the one hand, is indicated on the modal auxiliaries in
Afrikaans constructions containing (one or more of) these auxiliaries.53
As illustrated in (27), these auxiliaries co-occur with the infinitival form
of the main verb.
(27)
Ek sal /wil /kan /moet /mag baie praat
I will /want-to /can /must /may a-lot talk-INF
‘I will /want to /can /must /may talk a lot’
When modal auxiliaries do not occur, present tense is “carried” by the

main verb. Such a verb has the same form as the infinitive, regardless of
the person and number features of the subject, as can be seen in (28). Hê
‘to have’ and wees ‘to be’ are the exceptions, as shown in (29) and (30),
respectively: The present tense form of hê is het, and the past tense form
either had54 or het gehad (with gehad being the past participial form). The
53 In section 3.3.1.5, active constructions are at issue. Passive constructions are
discussed in section 3.3.2.3.

54 The had form − as in (29c) − is now rarely used (De Villiers 1971:24).
43
present tense form of wees is is, and the past tense form either was or was
gewees (with gewees being the past participial form).
(28)
Ek /Ons /Jy /Julle /Hy /Sy /Dit /Hulle /Die seun(s) praat
I /we /you-SGL /you-PL /he /she /it /they /the boy(s) talk-PRESENT
‘I /We /You /You /He /She /It /They /The boy(s) talk(s)’
(29a) Ek /Ons /Jy /Julle /Hy /Sy /Dit /Hulle het twee bene
I /we /you-SGL /you-PL /he /she /it /they have-PRESENT two legs
‘I /We /You /You /He /She /It /They have/has two legs’
(29b) Gister het ek vrede gehad
yesterday did55 I peace have-PAST PART
‘Yesterday I had peace’
(29c) Gister had ek vrede
yesterday have-PAST I peace
‘Yesterday I had peace’
(30a) Ek /Ons /Jy /Julle /Hy /Sy /Dit /Hulle is hier

I /we /you-SGL /you-PL /he /she /it /they be-PRESENT here
‘I am /We are /You are /You are /He is /She is /It is /They are here’
(30b) Gister was ek hier gewees
yesterday be-PAST I here be-PAST PART
‘Yesterday I was here’
(30c) Gister was ek hier
yesterday be-PAST I here
‘Yesterday I was here’
The present tense form of the main verb may also facultatively be used
in contexts where past tense is denoted by, for example, an adverbial
phrase (AdvP), as shown in (31). This is called the “historic present tense
form”.
(31)
Gister stap hy dorp toe
yesterday walk he town to
‘Yesterday he walked to town’
55‘have’ would be a more accurate translation for the temporal auxiliary het, but in order
to differentiate between the main verb het which is translated as ‘have’ and the temporal
auxiliary het, the latter is translated as ‘did’ throughout.
44
Past tense, on the other hand, is expressed by the obligatory temporal

auxiliary het in constructions not containing modal auxiliaries. The het co-
occurs with the past participial form of the main verb, as shown in (32).
This form resembles the infinitive, but has the prefix ge-, except in the
case of verbs beginning with the derivational morphemes be-, ge-, her-, er-,
ont-, or ver-, or another unstressed prefix, as shown in (33) (cf. Donaldson
1993:section 8.5.1).
(32)
Die man het kos gekoop
the man did food buy-PAST PART
‘The man bought food’
(33)
Sy het dit vermy /ontken /erken /begryp
she did it avoid-PAST PART /deny-PAST PART /admit-PAST PART /grasp-PAST
PART
‘She avoided/denied/admitted/grasped it’
Another exception is the second verbal element of a hendiadys. A

hendiadys in Afrikaans is a syntactic construction in which two verbal
elements are connected by means of the conjunctive particle en ‘and’ to
express a single complex idea (Roberge 1994:45). An example is loop en eet
‘walking along eating’. According to Roberge (1994:45), this second
verbal element is the main verb. Yet, this element occurs in the form
resembling the infinitival one in a past tense hendiadys, whereas the first
element can occur either in the form resembling the infinitival one or in
the ge- past participial form, as shown in (34).
(34a) (34b)
Hy het gesit en eet Hy het sit en eet
he did sit-PAST PART and eat-INF he did sit-INF and eat-INF
‘He was (sitting and) eating’ ‘He was (sitting and) eating’
When expressing past tense in constructions containing a modal

auxiliary, the use of the temporal auxiliary het and the past participial (ge-)
form of the main verb is optional. If het and the past participle are not
45
used, the main verb remains in its infinitival form. In such cases, the
modal auxiliary takes its past tense form, as can be seen in (35b).56
(35a) (35b)
Sy wil /moet /kan sien Sy wou /moes /kon sien
she want-to-PRESENT /must-PRESENT / she want-to-PAST /must-PAST /
can-PRESENT see-INF can-PAST see-INF
‘She wants to /must /can see’ ‘She wanted to /had to /could see’
The two exceptions here are the modals mag ‘may’ and durf ‘dare’, the
latter exclusively found in negative and interrogative constructions. The
past tense form of mag, namely mog, is virtually extinct. In constructions
expressing past tense, the phonological form of durf is either durf57 or –
less commonly and in combination with the temporal auxiliary het –
gedurf.58 In sentences containing these two modals, the past tense form is
formed by using the modal, the temporal het, and the past participial
form of the main verb, as shown in (36) and (37).
(36)
Ek mag dit gesien het
I may it see-PAST PART did
‘I was allowed to see it’ / ‘There is a possibility that I saw it’
56Less commonly, the ge- is bound to the modal instead of to the main verb, as shown
in (i).
(i) Sy het gewou /gekon deelneem
she did want-to-PAST PART / can-PAST PART participate-INF
‘She wanted to / could participate’
57 As in Ek het dit nie durf sê nie ‘I dared not say it’.
58 In some cases, the ge- that is to be expected on the main verb, is bound to the modal
durf, as illustrated in (ia) below. Note that whereas the absence of ge- is acceptable in
constructions such as (ib), two occurrences of ge- (one on the modal and one on the
main verb) lead to ungrammaticality, as shown in (ic).
(ia) Ek het dit nie gedurf sê nie
I did it not dare-PAST say-INF not
‘I dared not say it’
(ib) Ek het dit nie durf sê nie
I did it not dare-PAST say-INF not
(ic) *Ek het dit nie gedurf gesê nie
I did it not dare-PAST PART say-PAST PART not
46
(37)
Ek durf dit nie gesê het nie
I dare it not say-PAST PART did not
If the temporal het and the past participial form of the main verb are
used in past tense constructions containing modal auxiliaries (apart from
mag and durf), then these modals may occur in either of the two tense
forms (present or past). The past tense feature is then presumably
“carried” by the het, and not by the modal(s). The following examples
serve to illustrate this, where (38a) and (38b) are synonymous with the
past tense constructions in (35) above.
(38a) (38b)
Sy wil /moet /kan gesien het Sy wou /moes /kon gesien het
she want-to /must /can see-PAST PART she want-to-PAST /must-PAST /
did can-PAST see-PAST PART did
‘She wanted to /had to /could see’ ‘She wanted to /had to /could see’
In other words, the three sentences in (39) could have the same temporal
reference.
(39)
Ek kon dit doen
Ek kan dit gedoen het
Ek kon dit gedoen het
‘I could do it’
An interesting property of Afrikaans modal auxiliaries is that they may be

“stacked”, i.e., more than one modal auxiliary may occur in one and the
same construction, together with or independent of the temporal het, as
can be seen in (40).
(40a) (40b)
Ek sou dit moes kon doen Ek sou dit moes kon gedoen het
I will-PAST it must-PAST can-PAST I will-PAST it must-PAST can-PAST
do-INF do-PAST PART did
‘I would have had to be able to do it’ ‘I would have had to be able to do it’
47
3.3.1.6. Summary: Morphosyntax

In Afrikaans, agreement is not realised morphologically, not in terms of
number, person or grammatical gender features. Case is realised
morphologically, but only on pronouns. Present tense is not realised
morphologically, but past tense is, and in more than one way.
3.3.2 Word order related phenomena59
Afrikaans is a verb-second60 language. Since Koster (1975), verb-second

languages which demonstrate a verb-second surface word order in
matrix clauses but a verb-final one in embedded clauses have been
analysed as SOV underlyingly (Biberauer 2002:22). Because Afrikaans
patterns like Dutch (its parent) in terms of being verb-second, it has
been analysed together with Dutch as an SOV language since the 1970s
and 1980s (see, for example, Waher 1982). However, after Zwart (1993,
1994) proposed that Dutch is a SVO language and particularly since
Kayne’s (1994) proposal that all phrase structure could be seen as
underlyingly head-initial, Afrikaans VPs are increasingly considered to
have an underlying VO structure (cf. Oosthuizen 1996, 1998). For the
purposes of this study, Afrikaans is taken to be a verb-second language
which demonstrates an underlying SOV word order. Specifically, (i) for
the projections within the vP, the traditional SOV underlying order will
be adopted; and (ii) I will abstract away from the existence of object
movement within the verbal domain; but (iii) the functional layers TP
and CP will be analysed as being head-initial.
Aspects of word order pertaining to (i) the placement of the finite verb,
(ii) the derivation of question constructions, (iii) the derivation of passive
constructions, and (iv) preposing and scrambling are discussed below, as
are aspects of co-reference in Afrikaans.
59 See section 3.4.3 for a discussion on movement.

60 The verb-second phenomenon entails the obligatory occurrence of finite verbs in the
clause-second position, preceded by some clause-initial (usually phrasal) constituent
(Biberauer 2002:19).
48
3.3.2.1. Placement of the finite verb

In simple declarative sentences containing one verb, this main verb is
placed in the second position of the sentence. The sentence-initial
position can be occupied by a single word, as is illustrated in (41)61,62 and
(42), or by a multi-word phrase, as in (45).
(41)
Hy eet piesangs
he eat bananas
‘He is eating bananas’
(42)
Vandag eet hy piesangs
today eat he bananas
‘Today, he is eating bananas’
61 Tree diagrams are used to indicate the derivation of the surface word order of some
of the utterances. Where the derivation of different utterances is rather similar, the
diagram of only one is given. For instance, the derivation of (45) is not given, as it is
similar to that of (43). These tree diagrams now precede the discussion on Minimalist
syntax, which is not ideal, but which makes it possible to indicate how a construction is
derived while the construction is still under discussion.
62
In previous versions of generative syntax, it was argued that a syntactic object (a head
or a phrase) is moved as a whole; in the process, it leaves behind a “trace” of itself, i.e.,
a phonetically empty element. It is currently generally accepted, however, that
movement rather implies copying a syntactic object, i.e., duplicating the object involved
in movement (see the discussion of the Inclusiveness Condition in section 3.4.1; see
also Chomsky 2006:7). This copy then merges with some other syntactic object, with
feature checking taking place in the process. The two copies are identical, except that
the one which is (re-)merged has a fuller set of checked features. The copy which was
left behind, i.e., the one which did not merge, is (usually) deleted later, for reasons
pertaining to PF (Hornstein et al. 2005:242; also see Corver and Nunes 2007; Nunes
2004). Thus, usually only the copy occurring in the highest position created by internal
merge (cf. note 77) is pronounced (Chomsky 2006:8), because presumably this copy is
the only one which has had all its uninterpretable features checked (cf. Hornstein et al.
2005:242). So instead of a lexical element moving and leaving a trace, such a lexical item
is copied and then merges with other lexical items, with all but one copy being deleted
later. Note that Merge is recursive, and it can be applied to both lexical items and
expressions which are themselves a result of the application of Merge (Hornstein et al.
2005:209). In chapter 9, when explaining the word order errors of the participants in
the present study, I draw on this copy theory. In the tree diagrams, copies instead of
traces are entered, with deleted copies indicated by means of strike-through notation.
49
From the tree diagram in (43), it can be seen that the subject is taken to
originate in the specifier position of vP, from which it moves first to the
specifier position of TP and then to the specifier position of CP. The
verb originates in the V-position and moves from there first to v, then to
T and then to C.
(43) CP
Spec C'
hy
C TP
eet
Spec T'
hy
T vP
eet
Spec v'
hy
VP v
eet
Spec V'
DP V
piesangs eet
In (44), the verb undergoes the same movement as it did in (43), but the
subject does not undergo the final move to the specifier position of CP.
This is because this position is occupied by the adverb, which originated
adjoined to the vP from where it moved to the specifier position of CP.
In (45), op die trappe naby die biblioteek is seen as one constituent, which
would have originated adjoined to the vP, from where it would move to
the specifier position of CP. The sentence demonstrates a verb-second
surface word order, with the verb eet occupying this second position
(which is C). The subject hy is in the third position (in the specifier
position of TP to which it moved from the specifier position of the vP).
50
(44) CP
Spec C'
vandag
C TP
eet
Spec T'
hy
T vP
eet
Spec v'
hy
VP v
eet
AdvP V'
vandag
DP V
piesangs eet
(45)
Op die trappe naby die biblioteek eet hy piesangs
on the steps near the library eat he bananas
‘On the steps near the library, he is eating bananas’
In sentences containing auxiliary verbs, the auxiliaries appear in the

second position and the finite verb or past participle in the sentence-final
position, as shown in (46) which contains a temporal auxiliary, and (47)
which contains a modal one. The derivation of (47) is given in (48).
(46)
Hy het piesangs geëet
he did bananas eat-PAST PART
‘He has eaten bananas’
(47)
Hy sal piesangs eet
he will bananas eat-INF
‘He will eat bananas’
51
(48)63 CP
Spec C'
hy
C TP
sal
Spec T'
hy
T ModP
sal
Mod vP
sal
Spec v'
hy
VP v
eet
Spec V'
DP V
piesangs eet
In embedded sentences, the object appears before the verb (and before
any auxiliaries), as shown in (49), (51), and (52). In the case of modal
auxiliaries, the infinitive appears in the sentence-final position, as shown
in (51). If an embedded sentence contains the temporal auxiliary het, this
auxiliary occurs in the sentence-final position, directly preceded by the
past participle, as shown in (52).
(49)
Hy het bevestig dat hy piesangs eet
he did confirm that he bananas eat
‘He confirmed that he eats bananas’
63 Whereas this derivation is unproblematic for matrix clauses containing modal

auxiliaries, the same cannot be said of embedded clauses containing modal auxiliaries,
because there the modal needs to occur, in the surface word order, between the object
and the verb in its infinitive form: dat hy piesangs sal eet. A proposed solution to this
problem, offered by Biberauer and Richards (2004), entails that the modal does not
merge in the ModP but in the v-position. The infinitive merges in the V-position and
moves from there to the v, right-adjoining the modal, rendering piesangs eet sal eet.
Subsequent tree diagrams will follow this proposal.
52
(50) CP
Spec C'
C TP
dat
Spec T'
hy
T vP
Spec v'
hy
VP v
eet
Spec V'
DP V
piesangs eet
(51)
Hy het bevestig dat hy piesangs sal eet
he did confirm that he bananas will eat-INF
‘He confirmed that he will eat bananas’
(52)
Hy het bevestig dat hy piesangs sou geëet het
he did confirm that he bananas will-PAST eat-PAST PART did
‘He confirmed that he would have eaten bananas’
Note that where the object is [+animate] and definitely where it is

[+human], the object itself can (and, in some cases, must) be preceded
by vir64 – which Den Besten (2000:950) glosses with ‘for’ (see also
Molnárfi 2002). This is illustrated by (53), in which both Afrikaans
sentences have the same meaning.
64 Note that this vir is often found in spoken Afrikaans, but is not encouraged in written
Afrikaans.
53
(53a) (53b)
Ek sien hom /die vrouens /die pot Ek sien vir hom /vir die vrouens
/*vir die pot
I see him /the women /the pot I see for him /for the women /for
the pot
‘I see him /the women /the pot’ ‘I see him /the women /the pot’
3.3.2.2. Question constructions

The derivation of yes/no-questions in Afrikaans involves what is generally
called “subject-verb inversion”, as shown in (54) to (59). In (55), the
main verb sien moves to the head position of the complementiser phrase
(CP);65 this is in contrast to English, in which do-support is required to
derive the equivalent sentence. As in English, auxiliary verbs in Afrikaans
can also undergo subject-verb inversion, as illustrated in (55) and (56).
(54) Declarative: Interrogative:

Hy sien dit Sien hy dit?
he see it see he it
‘He sees it’ ‘Does he see it?’
In a tree representation, the derived structure of (54) would look as in

(57), where sien has moved to C (via T) and hy has moved from within
the vP to the specifier position of TP.
With a modal auxiliary, the sentence looks like (55) and has the derived
structure shown in (58). The equivalent question construction containing
the temporal auxiliary het is shown in (56).

Hy sal dit sien Sal hy dit sien?
he will it see will he it see
‘He will see it’ ‘Will he see it?’

Hy het dit gesien Het hy dit gesien?
he did it see-PAST PART did he it see-PAST PART
‘He saw it’ ‘Did he see it?’
65For simplicity’s sake, the CP is treated here as if it is a single phrase. Recently, it has been
proposed that CP is a domain, consisting of a force phrase, topicalisation phrase, focus
phrase, and finiteness phrase. Cf. Benincá and Poletto (2004); Botha (2006); Rizzi (1997).
54
(57) CP
Spec C'
C TP
sien
Spec T'
hy
T vP
sien
Spec v'
hy
VP v
sien
Spec V'
DP V
dit sien
(58) CP
Spec C'
C TP
sal
Spec T'
hy
T vP
sal
Spec v'
hy
VP v
sal
Spec V'
DP V
dit sien
In the derivation of Afrikaans wh-questions, as is the case in English, the

phrase containing the wh-element moves to the specifier position of the
CP. Where this element is in a prepositional phrase (PP), the whole PP
can move to the specifier position of the CP – an operation known as
55
pied-piping (Ross 1967; cf. (61a)) – or the wh-element can move on its
own, leaving the preposition “stranded” (cf. (61b)). In the course of the
derivation, the verb also moves into the head position of the vP and then
the head position of the TP and from there to the head position of the
CP. Again, the equivalent of do-support does not occur in Afrikaans.66

Hy sien die kind Wie sien hy?
he see the child who see he
‘He sees the child’ ‘Who does he see?’
(60) CP
Spec C'
wie
C TP
sien
Spec T'
hy
T vP
sien
Spec v'
hy
VP v
sien
Spec V'
DP V
wie sien
(61a) Declarative: Interrogative:

Hy betaal met note Met wat /Waarmee betaal hy?
he pay with notes with what /what-with pay he
‘He pays with notes’ ‘With what does he pay?’
[but: *Waar betaal hy mee?]
66As can be seen in example (61), the wh-element and the preposition often change
form when the whole PP is fronted in an interrogative structure: met wat changes to
waarmee, op wat to waarop, vir wat to waarvoor, etc. In this regard, see, for example,
Oosthuizen (2000).
56
(61b) Declarative: Interrogative:

Hy betaal met note Wat betaal hy mee?
he pay with notes what pay he with
‘He pays with notes’ ‘With what does he pay?’
[but: *Waar betaal hy mee?]
Wie (‘Who/m’)-questions are ambiguous in Afrikaans, if (i) the sentence

is in the active voice, and (ii) the post-verbal element is not clearly either
nominative or accusative. A question such as Wie sien die man? could
mean ‘Who sees the man?’ or ‘Whom does the man see?’. Likewise, Wie
jaag hulle? could mean ‘Who is chasing them?’ or ‘Whom are they
chasing?’. This ambiguity does not occur when pronouns are in the
singular form, because the form of these pronouns indicates whether
they are nominative or oblique. For example, Wie volg hom? could only
mean ‘Who is following him?’, and Wie volg hy? could only mean ‘Whom
does he follow?’.67
3.3.2.3. Passive constructions

In Afrikaans, passive constructions are formed by means of a non-
infinitival form word in the present indicative form or is (present perfect)
or was (pluperfect). Note that in the present perfect and pluperfect
tenses, the past participle form of word, namely geword, is implied (except
for some Afrikaans dialects, in which it is present in the passive
construction). See example (62).
(62a) Present indicative:

Ek word gesien deur die skare
‘I am seen by the crowd’
(62b) Present perfect:
Ek is deur die skare gesien but *Ek is gesien deur die skare
(geword) geword
‘I was /have been seen by the and *Ek is gesien geword deur
crowd’ die skare
67 Wie sien die man? ‘Whom does the man see?’ is equivalent to Vir (‘for’) wie sien die man?
and Wie jaag hulle? ‘Whom are they chasing?’ is equivalent to Vir (‘for’) wie jaag hulle?.
The vir unambiguously marks a non-subject and therefore vir–questions cannot be
ambiguous in the way their vir–less counterparts can (cf. Raidt 1969). Due to vir being a
type of accusative marker for animate objects, vir cannot occur before wat, as in *Vir
wat sien die man? ‘What does the man see?’. In fact, Vir wat sien die man is an informal way
of asking why the man sees.
57
(62c) Pluperfect:
Ek was deur die skare gesien but *Ek was gesien deur die skare
geword
‘I had been seen by the crowd’ and *Ek was gesien geword deur die
skare
As in English, the theme argument occupies the subject position (i.e., the
specifier position of TP) in Afrikaans passive constructions, that is, the
position associated with the subject in active constructions. In main
clauses, the theme argument may move on to the specifier position of
CP. This is illustrated by the example in (63) and the tree representation
in (64).
(63) Active: Passive:

Die hond eet die been Die been word geëet
‘The dog eats the bone’ ‘The bone is eaten’
(64) CP
Spec C'
die been
C TP
word
Spec T'
die been
T VPpassive
word
Spec V'passive
VP V passive
word
DP V
die been geëet
In “full” or agentive passives, the phrase expressing the theta-role of

agent forms part of a deur- (‘by’) phrase which can occur either before or
after the main verb, as illustrated in (65).
(65a) (65b)
Ek word deur hom gesien Ek word gesien deur hom
‘I am seen by him’ ‘I am seen by him’
58
3.3.2.4. Preposing and scrambling

As mentioned, Afrikaans simple main declarative sentences have a
surface verb-second word order, but subordinate clauses with an overt
complementiser like dat ‘that’ or of ‘whether’ have a surface SOV order.68
Compare the examples given in (66).
(66a) Hulle sien my69

they see me
‘They see me’
(66b) Ek dink dat hulle my sien
I think that they me see
‘I think that they see me’
(66c) Ek wonder of hulle my sien
I wonder if they me see
‘I wonder if they see me’
In linear terms, the finite verb appears in the second position in

constructions with a non-subject phrase (such as a preposed object DP,
PP, wh-phrase, AdvP, etc.) in the left-most position. In such cases, the
finite verb is followed by the subject and any other phrasal constituents
that may be present. Compare the examples given in (67) to (70).
(67)
[Daardie meisie] sien [hy] [dikwels] [by die winkel]
that girl see he often at the shop
‘That girl he often sees at the shop’
(68)
[Op die stoep] sit [die kinders]
on the veranda sit the children
‘The children are sitting on the veranda’
68 Where the sentence contains a auxiliary, the auxiliary occurs in the second position of
the sentence and the infinitive or the past participle follows the object, as shown in
Hulle sal my sien ‘They will see me’ and Hulle het my gesien ‘They saw me’, respectively.
69 Afrikaans also demonstrates an SVO surface word order in so-called parenthetical
constructions, as shown in (i), where ek dink is the parenthetical part, and hulle sien my
displays the (main clause) SVO order.
(i) Ek dink hulle sien my
I think they see me
‘I think they see me’
59
(69)
[Waar hy nou staan] het [ek] [al] [ook] gestaan
where he now stand did I already also stand-PAST PART
‘Where he stands now, I have also stood’
(70a) (70b)
Almal bring geld Vandag bring almal geld
everyone bring money today bring everyone money
‘Everyone brings money’ ‘Today everyone brings money’
Besides the fronting of a constituent to a clause-initial position (i.e., to

the specifier position of CP), Afrikaans also demonstrates scrambling in
the so-called middle field,70 as shown in (71). Here, direct objects can
precede or follow sentence adverbs, resulting in an adverb-object or
object-adverb order.
(71a) (71b)
Omdat hy gister ’n appel geëet het Omdat hy ’n appel gister geëet het
because he yesterday a apple because he a apple yesterday
eat-PAST PART did eat-PAST PART did
‘Because he ate an apple yesterday’ ‘Because he ate an apple yesterday’
3.3.2.5. Co-referential relationships

In Afrikaans, a reflexive pronoun can take one of two surface forms:
either the same form as that of the non-reflexive personal pronoun, or
that of such a pronoun plus –self. Unlike in English, for example, where
He bites him means ‘He bites a male entity other than himself’, Hy byt hom
can mean either ‘He bites a male entity other than himself’ or ‘He bites
himself’.71 However, as is the case for the English equivalent, Hy byt
homself can only mean ‘He bites himself’.
Afrikaans also has a number of “reflexive” verbs, that is, verbs which
can only take a reflexive pronoun as complement.72 Some examples are
misgis ‘miscalculate’, verbeel ‘imagine’, and vererg ‘loose one’s temper’, the
latter illustrated in (72) and (73).
70 For differences between scrambling in Afrikaans, on the one hand, and scrambling in
Dutch and German, on the other, see Molnárfi (2002).
71 When hom is preposed, a reflexive reading is no longer possible: Hom byt hy can only
mean ‘Him he bites’ and not ‘Himself he bites’.

72 See Reinhart and Reuland (1993) for inherently reflexive predicates in Dutch.
60
(72)
Hy vererg hom73 /*Hy vererg haar /*Hy vererg hulle
He looses his temper /He looses her temper /He looses their temper
(73)
Ek vererg my /Sy vererg haar
I loose my temper /She looses her temper
3.3.3. Conclusion: Some properties of the morphology and syntax

of Afrikaans
Afrikaans is a morphologically impoverished language; few grammatical

features are realised overtly. Afrikaans sentences demonstrate a verb-
second surface word order, and scrambling is possible. Unlike in English,
yes/no-questions do not require do-support. As in English, the wh-element
moves to the specifier position of CP in wh-questions. This position of
CP is occupied by the subject in declarative sentences (those without a
fronted adverb), a position the subject moves to from the specifier
position of TP after originating in the specifier position of vP.
Whereas the problems that Afrikaans-speaking children experience with

grammatical morphemes and word order will be discussed in chapters 5
to 9, we will not return to examine any problems they might demonstrate
with the comprehension or production of constructions in which co-
referential relationships occur.
In the next section, the reason for working within the framework of
Minimalist syntax is given. It will be argued that two well-documented
characteristics of SLI, namely problems with grammatical morphology
and problems with constituent movement, can both be related to a
problem with grammatical features. In order to give a precise account of
SLI as it presents itself in Afrikaans, it is necessary to use a theoretical
framework that enables one to account for the main characteristics,
which are presumed to involve problems with grammatical morphology
and problems with constituent movement, as in other languages in which
SLI has thus far been examined. One framework which enables such an
account is that of Minimalist syntax. As mentioned above, Afrikaans is
73 Increasingly, the pronoun co-occurring with these reflexive verbs take the –self form
(possibly under the influence of English), as in Hy misgis / vererg / verbeel homself.

61
an interesting language when examining SLI from the perspective of

Minimalist syntax: Although few grammatical features are overtly realised
morphologically in this language, it has many instances of overt
constituent movement. In sections 3.4.1 to 3.4.3, some of the
assumptions and devices of Minimalist syntax are discussed, and an
indication is given of how non-impaired Afrikaans is analysed in this
framework.
3.4. WHY WORK WITHIN THE FRAMEWORK OF

MINIMALIST SYNTAX?
Jakobson (1971:40), referring to aphasia, stated that “any individual case

offers highly complex and intricate linguistic material which cannot be
satisfactorily analysed without the cooperation of a linguist equipped
with all the technical means and methods of the modern science of
language”. It seems reasonable to assume that these remarks hold for
SLI as well. This section outlines a number of “technical means and
methods” of Minimalist syntax. Firstly, some general assumptions of
Minimalist syntax are set out. This is followed by a brief discussion of
two interrelated aspects of Minimalist theory, namely those involving
feature checking and movement. The question is subsequently addressed
whether such theories could offer possible explanations for the problems
experienced by children with SLI – specifically, problems with the
(phonological) realisation of morphosyntactic features, the production of
wh-questions and interpretation of passive constructions, as well as the
interpretation of (non-)co-referential relations. It will be argued that the
Minimalist theories in question provide an adequate framework for
describing and explaining the language characteristics of children with
SLI. Specifically, central to Minimalist syntax is the notion of formal (or
grammatical) features – linguistic expressions are considered to be
arrangements of features – and it seems to be the formal features that
children with SLI have difficulties with: at the level of phonology (at the
sensorimotor or articulatory-perceptual interface),74 in their
interpretation (involving the conceptual-intentional interface),75 and with
the syntactic structure of their utterances (involving movement
operations).
74See section 3.4.1.

75See section 3.4.1.
62
3.4.1. General assumptions of Minimalist syntax
A distinction should be made between the Minimalist Programme (or

Minimalism, for short) and Minimalist syntax. The Minimalist
Programme is not a theory, but a research programme. Chomsky
(1995a:1) puts this point as follows:76
This work is motivated by two related questions: (1) what are the
general conditions that the human language faculty should be
expected to satisfy? and (2) to what extent is the language faculty
determined by these conditions, without special structure that lies
beyond them? The first question in turn has two aspects: what
conditions are imposed on the language faculty by virtue of (A) its
place within the array of cognitive systems of the mind/brain, and
(B) general considerations of conceptual naturalness that have
some independent plausibility, namely, simplicity, economy,
symmetry, nonredundancy, and the like?
Chomsky, in an interview in Cheng and Sybesma (1995:32), stated that

one cannot speak of a Minimalist approach to something, as “there is no
specific Minimalist approach. There is a set of Minimalist questions”,
which implies that the Minimalist Programme is a “set of questions that
guide inquiry”. The task of the Minimalist Programme, according to
Chomsky (2001:3), is to critically examine the devices that are used to
characterise language, and to determine the extent to which such devices
can be eliminated. The general goal is to develop a principled account of
the language faculty based on general conditions of computational
efficiency and interface conditions that the language faculty must satisfy
in order to function.
Although there is no unique Minimalist approach to a specific linguistic

phenomenon, research carried out from a Minimalist perspective has led
to the development of a number of theories to account for various
syntactic phenomena (Hornstein et al. 2005:6). These theories can
collectively be referred to as “Minimalist syntax”. Thus, within
Minimalist syntax, there may be a theory of phrase structure and
movement (both involving Merge), a theory of feature checking (or
76 See also Chomsky (1999; 2000; 2002:108; 2006).

63
feature valuation), a theory of binding (involving the establishment of

(non-)co-referential relationships), etc. In fact, there may be several
competing theories for a specific phenomenon, in which case Minimalist
criteria of economy, simplicity, symmetry, elegance, etc. may guide the
selection of the “best” account. It is not clear at present precisely what
such criteria entail, nor how they enter into the evaluation of competing
theories. A brief sketch of some assumptions and devices of Minimalist
syntax will contextualise the discussion of feature checking and
movement undertaken below.
Chomsky (1995a:168) states that the language faculty consists of two

components, namely a lexicon and a computational system for human
language (CHL). The lexicon specifies the lexical items with their
idiosyncratic features. CHL derives a linguistic expression on the basis of a
selection of lexical items. Such a selection is called a “Numeration”
(Chomsky 1995a:169). It is important to note that, in the course of a
particular derivation, the devices of CHL have access only to the features
specified in the Numeration. This restriction, known as the Inclusiveness
Condition, rules out the introduction of traces, indices, bar-levels, etc.
(see, e.g., Chomsky 1995a:228). The derivation proceeds as the operation
Merge strings the lexical items together in binary fashion, and the
operation Move77 carries out the necessary movement of syntactic
objects in the expression. A general constraint on the operation
performed by Merge is that the two elements involved are themselves
left “unchanged”, a constraint known as the No-Tampering Condition
(Chomsky 2006:5).
The language is “embedded in performance systems” of two general

types: articulatory-perceptual and conceptual-intentional (Chomsky
1995a:168). Accordingly, there are two interface levels: phonetic form
(PF) at the articulatory-perceptual interface, and logical form (LF) at the
conceptual-intentional interface (Chomsky 1995a:2). A linguistic
expression of a language L is a pair (π, λ), where π is a PF representation
77 In more recent work, Move is defined as “internal Merge”, i.e., as a combination of
Copy and Merge. See, e.g., Adger (2003:section 4.2); Chomsky (2006); Hornstein
(2001:18-19); Lasnik and Uriagereka (2005:section 5.7). The term “Move” is used here
and below for the sake of simplicity. However, as has been the case previously in this
chapter, in tree diagrams, copies instead of traces are entered.
64
and λ an LF representation (Chomsky 1995b:390). Chomsky (1995b:394)

states that PF representations and LF representations are “differently
constituted”; importantly, elements interpretable at the PF interface with
the sensorimotor component are not interpretable at the LF interface
with the conceptual-intentional component, and vice versa. At some
point, the computation must split into a part forming the PF
representation and a part forming the LF representation; this point is
known as “Spell-Out”78 (Chomsky 1995b:394). At Spell-Out, the
elements relevant to PF are stripped away for eventual mapping onto the
PF representation, while the remainder continue in the computation for
eventual mapping onto the LF representation. A distinction is
accordingly made between “overt syntax”, operating before Spell-Out,
and “covert syntax”, operating after Spell-Out. We return to this
distinction in section 3.4.3.
On the basis of universal principles of grammar and associated

parameters, a language L determines an infinite set of linguistic
expressions, each a (π, λ) pair. A derivation is said to “converge” if it
produces a legitimate linguistic expression, and to “crash” if it does not
(Chomsky 1995a:171). A derivation can converge or crash at either PF or
LF, but must converge at both PF and LF if it is to converge at all
(Chomsky 1995a:171). The legitimacy of a linguistic expression is
determined by the principle of Full Interpretation, which requires that
uninterpretable features associated with certain categories must be
“checked”.79 According to Full Interpretation, only features which can
actually be interpreted at a certain level of representation are permitted at
that level, where “level of representation” refers to either Narrow Syntax,
the level of LF, or that of PF (Zwart 1993:13).80 Movement is the
78 It is assumed in recent accounts that transfer from the narrow syntactic part of the
computational system to the conceptual-intentional system is not a single event but
occurs at particular stages, following the completion of so-called phases, i.e., following
the completion of possibly νP and at CP (see, for example, Hornstein et al. 2005:347-
348; Lasnik and Uriagereka 2005:239). In short, “singular” Spell-Out has been replaced
by multiple Spell-Out (cf. Chomsky 2006:16).
79 The term “feature checking” will be used in a general, “theory-neutral” way, i.e., I
leave open the possibility that what is called “checking” here could, in fact, be
assignment, agreement, or valuation (cf. Adger 2003:167-169).
80 Although the operation of feature checking (via Move) seems to be an operation
found only in LF, the features which reach the PF interface should be interpretable at
PF.
65
mechanism which allows such feature checking to take place.81 What

follows is a discussion of feature checking and Move, respectively.
3.4.2. Feature checking
Within Minimalist syntax, the grammar is assumed to be “feature-driven”

in the sense that the movement of lexical items occurs so that feature
checking can take place. Movement, in turn, must comply with the
economy condition of Last Resort, according to which movement is only
allowed if it leads to the elimination of uninterpretable features
(Hornstein et al. 2005:293). Feature checking is “triggered by the need to
eliminate uninterpretable features from the computation” (Hornstein et
al. 2005:286,293).
Lexical items are defined as bundles of features, specifically

phonological, semantic, and formal (or syntactic) features (Chomsky
1995a:230). Retrieving a lexical item from the lexicon to form a part of a
Numeration thus implies retrieving a set of features. On the Lexicalist
view, lexical items reach the Numeration as fully inflected bundles of
features; all features (including phonological and semantic ones) are
specified (cf. Chomsky 1995a:275). As regards formal features, they serve
to describe the grammatical properties of words, and are accessible in the
course of the computation (Chomsky 1995a:230). Examples of such
features for the noun airplane are [nominal], [3rd person] and the gender
feature [-human] (Chomsky 1995a:231).
Phonological features are interpretable at PF, but not at LF, whereas

semantic features are interpretable at LF, but not at PF (cf. Chomsky
1995b:394). At the point of Spell-Out,82 the phonological features are
sent to PF and the semantic features to LF. It may be that certain formal
features are required for the morpho-phonological component, but these
formal features have to be deleted after fulfilling their function, i.e., after
playing their role in allowing the derivation to converge at PF. An
example is the formal feature [+plural] in dogs, which is associated with a
phonological feature bundle, the phoneme /z/. Once the formal and
81 If Move is taken to be an instance of Copy and Merge, it would be Merge which
allows feature checking to take place.
82 Or at every point of Spell-Out, if the proposal of multiple Spell-Out is accepted; cf.
note 78.
66
phonological features have been computed by the morpho-phonological

component, only the phonological features proceed to PF, in order that
the derivation does not crash, as the formal feature [+plural] is not
interpretable at PF (cf. Hornstein et al. 2005:291). Although the issue is
more complex in the case of the legibility of formal features at LF (cf.
Hornstein et al. 2005:291-292), it appears that such features are not
interpretable at LF either.
The question arises as to why lexical items possess uninterpretable

features at all, as this appears to be contrary to the requirements of
optimality and economy that the Minimalist Programme posits as
inherent properties of the language faculty. In the absence of a clear
answer to this question, Chomsky (2002) suggests that movement83 is the
mechanism whereby these uninterpretable features may be eliminated.
Thus, movement is the optimal solution to the problem of
uninterpretable features, the existence of which remains an unsolved
puzzle thus far. Note that, upon the matching of features, only
uninterpretable features are eliminated. Interpretable features continue in
the computation, and may even be used in subsequent feature checking
operations (Chomsky 2001).
3.4.3. Movement
In Minimalist syntax, the construction-specific movement rules (e.g.,

passive, question formation, extraposition, etc.) and category-specific
movement rules (e.g., V-to-I movement, NP-movement, wh-movement,
etc.) of earlier theories of generative grammar are replaced by Move.84
Move in Minimalist syntax essentially means “do not move anything
anywhere, unless as a last resort to prevent the derivation from
crashing”, entailing that a lexical item may only move for the purposes of
feature checking (i.e., for the purpose of eliminating uninterpretable
features). In the remainder of this section, two types of movement,
namely head movement and operator movement, will be discussed.
83Or Copy and Merge; cf. note 77.

84As stated in note 77, Move has come to be seen not as a single operation, but as “a
composite operation made up of two or more primitive processes, Copy and Merge”
(Chomsky 1993, 1995b; Hornstein et al. 2005:337). In Minimalist terms then, Merge is
preferred over Move, as the latter is a component of the former (Chomsky 1998;
Hornstein et al. 2005:337).
67
Head movement, or head-to-head movement, entails movement from

one head position to another. An example is auxiliary movement, i.e., the
movement of the auxiliary verb to a position preceding the subject in
structures such as yes/no-questions in English. Radford (1997a:106)
compares a yes/no-question in English, like Will he eat the food?, to a CP
such as … if he can eat the food, where the auxiliary will in the first sentence
occupies the same position as the complementiser if in the second.85 This
position is that of C, the head of CP. As Radford (1997a:108) notes, the
question arises as to why the auxiliary should undergo such head
movement, i.e., what makes the head movement analysis plausible here.
Bear in mind that movement must be motivated by a certain
computational need. A proposal made by Chomsky (1995a:292-294)
entails that the head C in an interrogative CP is a “strong” head and
therefore must be filled. The strong C has the power to attract the
auxiliary from the inflection (I) position.86 A brief explanation of strength
is required here. Within Minimalist syntax, it is assumed that some
categories and features are strong, i.e., they have to be checked in the
overt syntax,87 whereas others are weak and can be checked after Spell-
Out, in the covert syntax (Chomsky 1995a:233). As an illustration of
strong and weak categories/features, consider auxiliary-less yes/no-
questions in English, such as Do they like summer?. Here, the “dummy” (or
expletive) auxiliary do is required, whereas no do-insertion occurs in the
corresponding statement, They like summer. It is proposed that the
interrogative C is strong, and so must be filled (with do, initially inserted
in the T position). T, by contrast, is weak, and so need not be filled, as is
clear from They (do) like summer, where the do is only possible when it
carries emphasis. In terms of feature interpretability, C may be proposed
to have a (phonetically) uninterpretable Q-feature which must be
eliminated by moving an appropriate element into that position, e.g., a
85 The Afrikaans equivalent of this yes/no-question is Sal hy die kos eet? ‘Will he the food
eat? = Will he eat the food’ (with the infinitive in the sentence-final position). Similarly,
in … of hy die kos kan eet ‘… if he the food can eat = … if he can eat the food’, the
infinitive is sentence-final.
86 In view of more recent proposals regarding functional categories, the category labels
T and TP are used in this study instead of I and IP.

87 According to Chomsky (1995a: 233), there is a single exception here, namely “covert
merger (at the root) of a lexical item that has a strong feature but no phonological
features”.
68
lexical item with an interpretable Q-feature, such as will (as in Will they eat
the food?) or do (as in Do they like summer?).88
A distinction related to that of strong vs. weak features is one of overt

vs. covert movement. Strong features require overt movement, i.e.,
movement before Spell-Out, whereas weak features do not. In the latter
case, covert movement takes place, i.e., movement after Spell-Out,
resulting in the movement not being “visible” in the surface word order.
With regards to the overt-covert movement distinction, it has been
proposed that operations which have previously been regarded as covert
movement are, in fact, a type of overt movement, specifically overt
movement of grammatical features. In other words, it has been proposed
that features can be moved on their own, without the lexical item of
which they form part moving with them (Chomsky 1995a:264-265;
Hornstein et al. 2005:par. 9.4.2). If this proposal is accepted – in other
words, if there is no overt-covert distinction but rather a distinction
between the movement of features and the movement of categories –
then there is no reason to assume that Move takes place before Spell-
Out in certain languages and after Spell-Out in others. It is then rather
the case that as many movement operations as necessary are performed,
after which Spell-Out takes place (Hornstein et al. 2005:312).
The second type of movement to be discussed is operator movement,

i.e., the movement of an operator or operator expression. An example of
an operator is the interrogative element who in Who will you invite?, where
who is an operator expression. Such an interrogative operator triggers
auxiliary inversion in English89 (Radford 1997b:267). In this case of
operator movement, the complement of the verb in a VP moves to the
specifier position of the CP. Consider the tree diagram of the question
Who will you invite? in (74).
88 Regarding Afrikaans, one proposal for the features of C and T in interrogative

constructions is the following: In Afrikaans wh-question constructions, C has a strong
V-feature as well as a strong wh-feature, whereas T has a strong V-feature and a strong
D-feature (the latter causes the subject to move to the specifier position of TP). In
Afrikaans yes/no-question constructions, C has a strong V-feature, but not a strong Q-
feature; therefore, the specifier position of CP remains unfilled. The T in Afrikaans
yes/no-question constructions has the same features as the T in wh-question
constructions (T. Biberauer, personal communication).
89 For the derivation of an Afrikaans wh-question constructions, see (60).
69
(74) CP
Spec C'
who
C TP
will
you T'
T ModP
will
Mod νP
willi
DP ν'
you
ν VP
<agent>
invite V DP
invite who
In this tree diagram, auxiliary movement (as discussed above) is

instantiated in the movement of will from the head position of the modal
phrase (ModP), first to T and then to the head C position of CP. The
operator expression who, in turn, moves from its initial position as
complement of the verb invite in the VP to the specifier position of CP.
An explanation for why the interrogative operator expression who moves
to the specifier position of the CP is offered by Radford (1997a:135,
building on insights by Chomsky).90 As in the case of auxiliary
movement, it is proposed that the head C in question constructions
carries a strong wh-feature (cf. Chomsky 1995a:232),91 which must be
checked so as to be eliminated. It is further proposed that an
interrogative operator like who carries an interpretable interrogative
feature. On this analysis, the interrogative operator moves to the
90 I leave open here the question of how and where in the configuration the case of the
DP complement of the verb (i.e., who) is checked. See, for example, Chomsky (1995a)
and Hornstein et al. (2005:318,347) for the idea that accusative case is checked (or
specified), at least in English, under agreement with a light verb. Also see Adger
(2003:217-222).
91 The distinction, if any, between a “wh-feature” and a “Q-feature” is left open here.
For the sake of convenience, “Q-feature” is used in the case of yes/no-questions and
“wh-feature” in that of wh-questions.
70
specifier position of the CP to check the strong wh-feature of C (cf.

Chomsky 1995a:263).
As stated in section 3.4.1, the Inclusiveness Condition states that the

devices of CHL have access only to the features specified in the
Numeration. The introduction of traces, amongst other things, into a
derivation is thus ruled out (see, e.g., Chomsky 1995a:228). In Minimalist
syntax, traces are seen to be copies, where the displaced constituent and
its copy are seen to be strictly identical (Chomsky 1993, 1995a:253). This
means that a copy is not seen to be a new element introduced into the
derivation; rather, “a copy is whatever the moved element is, namely, a
syntactic object built based on the features of the numeration”
(Hornstein et al. 2005:213). In short then, in terms of a copy theory of
movement, the moved element does not leave behind a trace of itself but
the element in question is copied and then moved, resulting in a copy
being left behind. Movement creates complex syntactic objects, mainly
chains, consisting of the moved element and its copy, both with the
same set of features (Chomsky 1995a:251). In general, only one of the
copies of the chain can be spelled out, typically the head of the chain.
According to Nunes (1995, 2004), this relates to Kayne’s (1994) linear
correspondence axiom (LCA) which states that a copy cannot both
precede and follow an item Z. For example, in *Who did you see who?, the
second who is in the tail position of the chain. Given that the chain {who,
who} is one syntactic object, who both linearly precedes and follows the
verb see, resulting in ungrammaticality. In order for the syntactic
structure to be linearised, one of the copies must be deleted. This
process is called “chain reduction” (Nunes 2004), which entails the
following: Delete the minimal number of constituents of a nontrivial
chain CH that suffices for CH to be mapped into a linear order in
accordance with the LCA. The principles on which chain reduction are
based are (i) only one copy can be pronounced, and (ii) the copy with the
fewest unchecked features is pronounced (Nunes 2004).
This brief discussion of feature checking and two types of movement

within Minimalist syntax provides the framework within which the SLI
data discussed in the previous chapter will be re-examined below.
71
3.4.4. Minimalist syntax and problem structures for children

with SLI
As pointed out in section 2.3, it is well documented that children with

SLI experience problems with grammatical morphemes, with constituent
movement – the latter demonstrated by the difficulty associated with
producing various question constructions and with comprehending and
producing passive constructions – and with interpreting (non-)co-
referential relationships. In this section, these problems will be discussed
briefly against the background of the assumptions and devices of
Minimalist syntax.
3.4.4.1. Grammatical morphemes

One assumption of Minimalist syntax is that lexical items enter the
Numeration in their inflected form. This means, for example, that the
past tense form of the verb play enters the Numeration as played, and is
not the result of a syntactic merger of the retrieved items play and –ed. As
was stated in section 2.3, most children with SLI experience problems
with the phonological realisation of grammatical morphemes, with some
morphemes appearing more difficult to master than others. These
children demonstrate a lower percentage of correct use of grammatical
morphemes than age-matched controls; they sometimes omit such
morphemes from obligatory contexts (as in *two cat; cf. Oetting and Rice
1993) or insert them in inappropriate contexts (as in *You got a tape
recorders, from Gopnik 1990a, where the plural form of tape recorder is used
to refer to a single tape recorder). A possible explanation for the lower
percentage of use of grammatical morphemes could be that children with
SLI retrieve the uninflected form more often than typically developing
children do. It is, however, not clear why this should be the case. In
those cases where children with SLI insert grammatical morphemes
inappropriately, it could similarly be argued that they selected the
“incorrect” form of the word from the lexicon (i.e., the inflected form92
instead of the uninflected form), as in a tape recorders.
92 If one assumes that the lexical item is phonologically fully inflected when its enters
the Numeration, then one could argue that children with SLI retrieve the “incorrect”
form more often than do age-matched controls. However, if one assumes that the
lexical item is retrieved as a feature bundle and that each phonetically relevant feature
needs to be spelled out (or made “concrete”) in PF, then the problem has a phonetic
origin: Children with SLI could then be argued to experience more problems than age-
72
3.4.4.2. Word order related phenomena

Children with SLI do not appear to experience problems with the
concept of syntactic movement. This is clear from typical SLI utterances
like What we make? and What we can make?. From a Minimalist viewpoint,
it could be argued that the child “knows” that a wh-question is derived by
moving the wh-operator (what in the above example) to the specifier
position of the CP to check the uninterpretable [+wh]-feature of the
head C. However, what is striking about the wh-questions produced by
children with SLI, is the incorrect linear order of the subject DP and the
(auxiliary) verb, as illustrated above. In Minimalist terms, it could be
argued that these children either fail to move the (auxiliary) verb at all, or
do move it to the head T position but not from there to the head C
position.93 Possible reasons for such failures could be the following: (i)
the child has not yet acquired the relevant features that necessitate the
movement of the (auxiliary) verb; (ii) the feature make-up of the relevant
heads – e.g., the (auxiliary) verb, or T, or C – differs from that of age-
matched controls; or (iii) the child has problems establishing which copy
needs to be spelled out, with children with SLI sometimes spelling out a
“lower” one instead of the left-most one.
As was pointed out in section 2.3.1.2, children with SLI appear to have a
preference for interpreting agentless passives as adjectival constructions
(cf. Van der Lely 1996). For example, an utterance such as The teddy is
washed is usually assigned an interpretation on which is washed describes a
property of the teddy (i.e., the washed teddy), rather than one on which the
teddy is the theme undergoing some action by an unspecified agent (i.e.,
the teddy is washed by someone). There are various possible explanations for
this preference. One would be that the child does not differentiate
between a copula and a passive auxiliary verb, a distinction that would be
expressed by means of formal lexical features within Minimalist syntax.
Another possibility would be that children with SLI do not differentiate
between active and passive sentences; hence, the teddy is washed is simply
matched controls with the phonological realisation of features which are present. This
would be in contrast to the former view, where the absence or inappropriate presence
of features is at issue. At present, it is still being debated which view on “fully inflected”
is the more plausible one; for the purposes of this study, it will be assumed that the
problem of children with SLI lies in spelling out the features that are present.
93 To keep the discussion simple here, I do not refer to the various heads within the CP
domain. Cf. note 65.

73
interpreted as an active sentence. Support for this possibility may come

from the fact that such children also have difficulty with the
interpretation of “full” passives, that is, passive constructions containing
an agent phrase as in The teddy is washed by the girl. Children with SLI
usually give such sentences an active interpretation: “the teddy is the one
who is washing the girl” (Van der Lely 1996:260-261). Again, a possible
feature-based account could be that the child does not differentiate
between, say, the past participial form and the passive form of the main
verb, a distinction that is presumably expressed by means of formal
features.
A third possible explanation, which could hold for both agentless and
full passives, concerns the assignment of the appropriate theta-roles to
the nominal expressions functioning as arguments in a structure, as
pointed out by Van der Lely (2003:125). It could be that a child with SLI
overgeneralises the so-called Uniform Theta-Assignment Hypothesis
(UTAH), which holds, briefly, that identical thematic relationships
between items correspond to identical structural relationships between
such items at an underlying level of representation (Baker 1988:46). For
example, in active constructions, the Agent role is associated with the
structural subject position, i.e., the specifier of little-v (Radford
1997a:204). Hence, the child simply takes any argument occurring in the
canonical structural subject position as the one receiving the Agent theta-
role, irrespective of whether the structure in question is an active or a
passive one. It could well be that this overgeneralisation of the UTAH is
related to the second possibility mentioned above: The child fails to
distinguish between the active form of the verb (which has the property
of assigning an Agent theta-role), and the passive form (which does not
have this property). And again, this failure could be ascribed to some
problem involving the features associated with the verb forms in
question.
3.4.4.3. (Non-)co-referential relationships

Minimalist accounts of why children with SLI experience problems with
the interpretation of (non-)co-referential relationships between nominal
expressions cannot be viewed as “exclusive” to Minimalist syntax.
Rather, these accounts are to a lesser or greater extent based on the
assumptions and devices of binding theory, one of the modules of
government and binding theory (see, e.g., Chomsky 1982a, 1982b). For
74
example, the “Minimalist” binding theory set out in Hornstein et al.’s

(2005) introduction to Minimalist syntax, is in essence that presented in
Chomsky (1982a) and modified in Chomsky (1995a:chapter 1).94,95 A
brief description of the concepts of binding theory, based on the
proposals in Hornstein et al. (2005), is given below.
The theory assumes the typology of nominal expressions presented in

Chomsky (1982a:78). According to this typology, a distinction is made
between two main classes of nominal expressions. The first class
comprises overt nominal expressions, which include anaphors (i.e.,
reflexives such as himself and reciprocals such as each other), pronouns
such as he and her, and referential (R-) expressions such as the boy. The
second class comprises covert nominal expressions, which include
anaphors (specifically DP-traces), pronouns (specifically pro), and R-
expressions (specifically wh-traces).96 Covert nominal expressions also
include PRO, for which there is no overt equivalent. Two features,
namely [±anaphoric] and [±pronominal] are used to differentiate
between these seven types of nominal expressions, as shown in (75)
below (based on Chomsky 1982a:78).
(75)
Overt: [anaphoric] [pronominal] Covert:
Anaphors + - DP-traces
Pronouns - + pro
R-expressions - - wh-traces
--- + + PRO
94 See, for example, Oosthuizen (2006) and Zwart (2002) for alternative, feature-based
accounts within Minimalist syntax. See also Hornstein (2001:chapter 5) for a
movement-based theory of binding, as well as Reuland (2001).
95 This chapter first appeared as Chomsky, N. and H. Lasnik (1993). Principles and
parameters theory. In J. Jacobs, A. von Stechow, W. Sternefeld, and T. Vennemann

(eds.) Syntax: An international handbook of contemporary research. Berlin and New York:
Walter de Gruyter. pp. 506-569.
96 In Chomsky (1982b), DP-traces (or NP-traces in earlier versions of generative
syntax) are referred to as A-traces (= argument traces), and wh-traces as A΄-traces (=

non-argument traces). Note that in Minimalist syntax, traces are copies. They are not
semantically empty; their “emptiness” is a PF-property.
75
Binding Theory, as set out in Chomsky (1995a:92-110), comprises the

following three interpretive procedures:97
Given a local domain D,

A. if α is an anaphor, interpret it as co-referential with some c-
commanding phrase98 in D;
B. if α is a pronoun, interpret it as disjoint from every c-commanding
phrase in D;
C. if α is an R-expression, interpret it as disjoint from every c-
commanding phrase.
One possible explanation for the problems that children with SLI
experience with the interpretation of (non-)co-referential relationships,
could be that they have not fully acquired (any one or all of) the three
binding procedures. Another could be that aspects related to these
procedures and/or to the typology underlying them have not been
acquired. For example, it could be that children with SLI do not yet
“know” what constitutes a (local) domain99 or what the relationship of c-
command entails; or it could be that they do not differentiate between
the various types of nominal expressions in the typology in the same way
that age-matched controls do. A third explanation could be that the
agreement relation established by these children between, for example,
an anaphor and its antecedent is incorrect, i.e., that there is a problem
with the agreement between the grammatical features of an anaphor like
himself and its antecedent he in, for example, Peter said that he hurt himself.100
97 These interpretive procedures incorporate the binding principles A, B, and C
presented in Chomsky (1982b:188):

(A) An anaphor is bound in its governing category.
(B) A pronominal is free in its governing category.
(C) An R-expression is free.
98 The notion c-command is defined as follows in Hornstein et al. (2005:78):
α c-commands β iff
i. α does not dominate β;
ii. β does not dominate α;
iii. the first branching node dominating α also dominates β; and
iv. α does not equal β.
99 There are several definitions of the notion domain (see, e.g., Hornstein 2001:153;
Hornstein et al. 2005:248); see Chomsky (1995a:101-103) for what constitutes a local
domain.
100 For Afrikaans-speaking children with SLI, the fact that the sound form of reflexives
need not be distinct from their non-reflexive counterparts could add to interpretation
76
3.4.5. Summary: Why work within the framework of Minimalist

syntax?
It would appear from the brief discussion in the preceding sections that
Minimalist syntax has, at least, the potential to offer interesting
explanations for the problems that children with SLI experience with
grammatical morphology and constituent movement. More specifically, it
could be argued that Minimalist syntax makes it possible to give a unified
explanation of apparently unrelated phenomena in terms of the devices
of feature checking and movement (i.e., merge and copy, and Spell-Out).
In view of these considerations, the assumptions and devices of
Minimalist syntax will be assumed as the framework for the grammatical
analyses presented in this study.
In this chapter, some characteristics of non-impaired Afrikaans and the

derivation of non-impaired Afrikaans utterances in terms of Minimalist
syntax were given. The next chapter provides details on the methodology
by which data were obtained from Afrikaans-speaking children in order
establish what the characteristics of impaired Afrikaans are. The
Afrikaans data are presented in chapters 5 to 8.
problems as far as (non-)co-referential relationships are concerned. Note that, as stated

before, (non-)co-referential relationships were not examined in this study.
77
Methodology
Chapter 4
Methodology
4.1. INTRODUCTION
In studies of children with SLI, it is customary to compare their language

(or other) abilities to those of other children. Recently, comparisons are
increasingly made to other non-typically developing children, such as
those with Williams syndrome (cf. Reilly, Losh, Bellugi, and Wulfeck
2004), Down’s syndrome (cf. Eadie, Fey, Douglas, and Parsons 2002;
Laws and Bishop 2004; McGuckian 2004), hearing impairment (cf. Bol
and Kuiken 1990), and focal brain injury (cf. Marchman, Saccuman, and
Wulfeck 2004; Weckerly, Wulfeck, and Reilly 2004).
However, the most common study design still entails comparing children
with SLI to typically developing ones.101 In order to ascertain whether or
not the language abilities of the children with SLI are age appropriate,
comparisons are made to age-matched controls. In these comparisons,
children with SLI usually fare worse than their same aged peers. Leonard
(1998:27) mentions the following potential problem when
(chronological) age-matching only is employed. Suppose one is interested
in two aspects of the language. If one finds that the performance of
children with SLI on these two aspects differs, one might conclude either
that this difference is meaningful or that this difference reflects normal
language development. In order to decide which it is, one can ascertain
how well typically developing age-matched controls fare on these two
aspects: If they do better with one than with the other, then one can
conclude that the development of the children with SLI mirrors that of
the typically developing ones. However, if one finds that the typically
developing children fare well on both of these aspects, then one does not
know whether (i) they mastered them at the same time, or (ii) one has
been mastered for some time and the mastery of the other is a recent
101 See sections 2.3.1 and 2.3.2 for examples of studies with this design.
79
Methodology
occurrence. In order to ascertain this, the aspects need to be studied

before both are mastered.
For this reason, another group of children often included in studies of

SLI is one comprising younger typically developing ones – and some
studies include both age-matched and younger controls. These younger
controls are usually matched to the children with SLI on a particular
language measure. Often this measure is MLU, but this matching
technique is not without its problems, as discussed in section 4.2.
Another often-employed measure is performance on a standardised
language test, such as the Peabody Picture Vocabulary Test – Revised
(Dunn and Dunn 1981). As stated by Leonard (1998:30), a general
problem with using younger controls, regardless of matching criterion, is
that these typically developing children might differ from the children
with SLI in many ways, “some of which may have more to do with the
general developmental differences between the groups than with
differences pertaining to the details of language under investigation” (cf.
Plante, Swisher, Kiernan, and Restrepo 1993).
Despite the criticism against various aspects of matching and

comparison with typically developing children, the different designs
(comparison with age-matched children; with younger age-controlled
children; with language-matched children; or with both age-controlled
and language-matched children) are frequently employed by researchers
interested in the language abilities of children with SLI. The design of
the present study included two groups of typically developing children,
namely an age-matched and a younger age-controlled group. The design
is given in section 4.2, where the reasons for not employing MLU-
matching are discussed. Instead, the MLU is used as a control measure to
prevent overlap in MLU between the groups of typically developing
children. The two sections thereafter deal with the participants: Section
4.3 contains information on participant selection, whereas section 4.4
gives the details of the three participant groups which were included.
The design of the present study comprises both spontaneous language
data and data obtained by a series of specific experimental tasks. Section
4.5 gives information about the procedures used for collecting
spontaneous language. In section 4.6, the experimental protocol is
described. Lastly, the data transcription and scoring procedures for the
spontaneous and elicited language data can be found in section 4.7.
80
Methodology
4.2. DESIGN OF THE STUDY
In the present study, 15 Afrikaans-speaking 6-year-olds with SLI were

compared to 15 typically developing age-matched ones and to 15
younger typically developing ones. There were two main criteria
according to which the participants in the younger control group could
be selected. The first of these was their age. The second was their MLU:
Either their MLU could fall within a specific range or it could be
matched to those of the participants with SLI. It was decided to use both
age and MLU as selection criteria.
Regarding age, the children had to be 4-year-olds, for the following three
reasons:
(i) It was assumed that this was the age at which the children would
be cognitively mature enough to cope with the demands of the
language assessment tasks.
(ii) It was also assumed that the grammatical morphemes examined in
this study would have been acquired (to a great extent) by this
age.102 As stated by Balason and Dollaghan (2002:962), it is
believed that variability regarding grammatical morphology
declines at around age 4 in typically developing children.
(iii) The language-matched controls in some other studies (for example
Johnston, Miller, Curtiss, and Tallal 1993:974; Oetting and Rice
1993:1239; Rice 2003:72) were on average 2 years younger than the
experimental group consisting of children with SLI, even when
MLU-matching was performed (De Jong 2003:154; see also
Leonard 2000; Rice, Redmond, and Hoffman 2006:805).
The MLUs of the 4-year-olds were calculated, as well as those of the age-
matched controls, in order to ensure that the MLUs of these two groups
did not overlap, i.e., that no 4-year-old’s MLU was higher than that of
any 6-year-old, and that no 6-year-old’s MLU was lower than that of any
4-year-old. The reasons why strict pairing of each 6-year-old in the SLI
group with a (younger) MLU-matched control was not carried out are
discussed below.
102There are no relevant developmental data available for Afrikaans; therefore, this
assumption was based on the findings of researchers such as Lahey, Liebergott,
Chesnick, Menyuk, and Adams (1992) and Paul and Alforde (1993) for English-
speaking children.
81
Methodology
There is no consensus on the validity and reliability of MLU as a

spontaneous language measure. Regarding validity, MLU is regarded by
some to be a general measure of syntactic and/or grammatical
complexity (see, amongst others, Bornstein, Haynes, Painter, and
Genevro 2000:413; Brown 1973:77; Leadholm and Miller 1992:37; Miller
1981:75), but by others to be a global measure of expressive language
ability (cf. DeThorne, Johnson, and Loeb 2005; see also Leonard and
Finneran 2003:886).103 So, despite its frequent use, it remains unclear
what MLU actually reflects. Dethorne et al. (2005: 646) conclude that
MLU is best viewed as a global measure of expressive language ability,
despite its original introduction as a measure of morphosyntactic ability.
This view is supported by Eisenberg, Fersko, and Lundgren (2001).
Regarding reliability, it is generally accepted that MLU is a less accurate

measure once it reaches a certain value: at least 3 according to Klee and
Fitzgerald (1985) and Rondal, Ghiotto, Bredart, and Bachelet (1987); 3.5
to 4.0 according to Bol (1996); and 4.5 according to Blake, Quartaro, and
Onoratti (1993). Also, it appears that MLU might plateau at around age 4
(Bernstein and Tieger-Farber 1997; Chabon, Kent-Udolf, and Egolf
1982; see also Bol 2003:261-262).104 For this reason, care was taken in the
present study to limit the prominence given to MLU when choosing
younger controls for the 6-year-olds with SLI.
Brown (1973:77) claims that “almost every new kind of knowledge

increases length”. However, length and complexity (or sophistication)
are not always correlated. Consider the following two utterances from
one of the Afrikaans-speaking 6-year-olds with SLI in the present
study:105
103 A review by Williams, Dethorne, and Galvanoni (reported in Dethorne et al.

2005:635) of articles on child language published in the Journal of Speech, Language, and
Hearing Research from 1999 to 2004 revealed that, of those including MLU,
approximately one in two referred to it as a global measure of language proficiency and
one in three specifically as a measure of grammatical or morphosyntactic complexity.
104 Note, however, that other researchers have found that MLU increases with an
increase in age, sometimes even after 5 years of age (cf. Hunt 1970; Loban 1976; Miller
and Chapman 1981). It appears then that there is no consensus regarding the reliability
of MLU as a measure of general language proficiency or grammatical complexity in
children older than 5.
105 See Eisenberg et al. (2001:324) for a similar example in English.
82
Methodology
(76) Target:
die baba het so hom gekou het die baba het hom so gekou
the baby did so him chew-PAST PART did the baby did him so chew-PAST PART
‘The baby chewed him like this’
(77) Target:
laat ek sal dit sommer doen laat ek dit sommer doen
let I will it just do let I it just do
or
ek sal dit sommer doen
I will it just do
‘It’s no problem; I’ll do it’
Both of these utterances are one morpheme longer than the target
utterance; however, both are also ungrammatical. The following two
ungrammatical utterances, from the same child as those above, contain
the same number of morphemes as do the target utterances.
(78) Target:
daar is hom tas daar is sy tas
there is him-OBLIQUE suitcase there is his suitcase
‘There is his suitcase’
(79) Target:
ons ma leer ons saam ons ons ma leer saam met ons
our mom learn us with us our mom learn together with us
‘Our mom is learning with us’ (i.e., she is in our class at school)
There was a third reason for not making use of MLU-pairing in this
study. This reason is related to MLU being a quantitative measure, one
which does not reveal much about the quality of what is said – as
illustrated by the above examples.106 Using MLU-pairing can lead to an
overestimation of a particular child’s grammatical abilities.107 According
to Eisenberg et al. (2001), whereas a low MLU can be interpreted as
106 See Bol (2003:260) for a brief discussion of quantitative vs. qualitative measures of
linguistic phenomena.
107 See also Johnston et al. (1993) who found that children with SLI seem to respond
with higher degrees of ellipsis than do language-matched controls to high rates of

questioning by their conversational partners.
83
Methodology
support for a diagnosis of language impairment, a higher than expected

MLU cannot be taken as evidence that no such impairment exists.
Because of these problems with MLU as a measure,108 the prominence

given to it was limited in this study. Seeing that the language-matched
controls in other studies were on average 2 years younger than the
children with SLI, sometimes regardless of whether MLU-matching was
performed (as mentioned above), it was decided to choose 4-year-olds as
younger typically developing controls. So, rather than searching for
children younger that 6 years with MLUs exactly matching those of the
6-year-olds with SLI, MLU was used as a more general measure for
inclusion in the study: Typically developing 6-year-olds could only be
included in the study if, amongst other criteria, their MLU was better
than that of the typically developing 4-year-olds. Stated differently, 4-
year-olds could act as younger controls, provided that their parents and
teachers deemed them to be typically developing in all respects, and
provided that their MLU did not surpass that of any of the typically
developing 6-year-olds.
4.3. PARTICIPANT SELECTION
Because the language investigated here is Afrikaans, the participants were

selected mostly from schools and other institutions in the Western Cape
Province, which is the South African province with the largest number
of mother-tongue speakers of Afrikaans, and where 55% of the
population speak Afrikaans as mother-tongue (Statistics South Africa
2003). Of the 45 participants, only three (all with SLI) resided outside of
this province.109 In selecting participants, care was taken to include only
speakers of so-called standard dialects of Afrikaans (as judged by the
participants’ teachers or speech-language therapists). Contact with
108 It has been suggested that productive vocabulary size replaces MLU as a matching
variable (cf. Caselli, Casadio, and Bates 1999; De Jong 1999). However, as stated by Bol
(2003:270), “not enough is known about the relation between vocabulary size and
grammar … to justify using vocabulary size instead of MLU from now on as a
matching criterion”.
109 All of the typically developing children resided in and/or attended schools in
Stellenbosch. Of the participants with SLI, three resided in Bloemfontein (in the Free
State Province), six in the northern suburbs of Cape Town, and one each in Somerset
West, Oudtshoorn, the Laingsburg region, the Uniondale region, Hartenbos, and
Witsand.
84
Methodology
speakers of other, non-standard dialects could, however, not be ruled

out.
4.3.1. Participants with SLI
In order to obtain participants with SLI, speech-language therapists at

seven government-funded institutions and in 12 private practices were
requested to identify from their case-loads all Afrikaans-speaking 6-year-
olds from monolingual Afrikaans-speaking homes who demonstrated
language problems in the absence of hearing, intellectual, socio-
emotional and neurological problems. Over a period of 21 months, 17
children who met these criteria were identified by their therapists as
possible participants.
As noted in section 3.2, there is no agreed-upon protocol for the

identification of SLI in Afrikaans-speaking children. One of the obvious
reasons for this is a lack of assessment instruments. To date, only three
tests have been developed for use with Afrikaans-speaking children,
namely the AST, the TMT, and the ARW (cf. section 3.2). All three of
these instruments were developed for use with, amongst other age
groups, 6-year-olds. Yet not all 6-year-olds in the SLI group had been
tested with (one of) these three instruments, for the reasons mentioned
in section 3.2.110
This lack of scores on standardised language tests for the children in the
SLI group is a problem, as one of the criteria for SLI is abnormally low
language test scores. Some researchers (e.g., Starke and Tallal 1981)
employ age scores.111 In such cases, one or more of the following need to
be present in order for a language problem to qualify as SLI: (i) a
receptive language age score of 6 months or more below mental or
chronological age, whichever is lower; (ii) an expressive language age
score of 12 months or more below the lower of the mental or
chronological age; or (iii) a combined language age score of 12 months
or more below the lower of the mental or chronological age. Other
authors employ as criterion language test results that are at least 1.25
110 Of the 15 children with SLI, the AST was performed with six and the ARW with
two. The TMT was not performed at all. Cf. appendix A.
111 See Plante (1998) for a critique of this practice.
85
Methodology
standard deviations below the mean for the child’s age (Leonard
2003:211).112 The fact that scores of language tests are needed in order to
diagnose SLI is a problem when working with Afrikaans-speaking
children, for two reasons. The first reason is the aforementioned scarcity
of Afrikaans-medium language tests. Starke and Tallal (1981:121) state
that, to determine the presence of SLI, “a representative set of language
tests is needed … , not one alone”. However, for certain language skills,
such as expressive morphosyntactic ability, there is not even one
Afrikaans-medium test. The second reason is that the age scores and
standard deviations obtained when administering translated English-
medium tests are not necessarily meaningful or even valid, seeing that
these scores reflect how the child would have fared had the child been
English-speaking and had the raw scores been obtained with the original,
standardised English-medium version.
Bearing in mind the absence of an agreed-upon protocol for the

identification of SLI in Afrikaans-speaking children, the judgement of
the speech-language therapists regarding slow and/or abnormal
development was used to determine whether a potential participant had
SLI. This judgement was based on a case-history and one or more of the
following: (i) standardised testing, (ii) testing with non-standardised
Afrikaans translations of tests developed for English-speaking children,
(iii) informal testing, and (iv) language sample analysis.
The parents of each of 17 potential participants were contacted to obtain

written consent for the inclusion of their child in the study; 16
consented.113 One of the 16 was initially included in the study, but it was
found that, in terms of morphosyntax, his language sample and
responses to the experimental tasks were similar to those of the best-
performing typically developing 6-year-olds. The standardised test used
by the therapist to evaluate his language, was the AST. Although his
scores on selected subtests of the AST indicated a language delay, this
test assesses semantics rather than morphosyntax. Because it was not
112Also see, amongst others, Records and Tomblin (1994).

113Only one parent was not willing to have her child participate. This parent stated that
her child had had several evaluations (amongst others, an IQ test) at that time, and that
he had been warning her that he would accompany her to assessment sessions but that
he will no longer co-operate during these sessions. For this reason, she decided to
withhold consent.
86
Methodology
clear that this boy’s language problems were at all morphosyntactic in

nature, he was subsequently excluded from the study.
Appendix B contains a copy of the letter which was sent to the parents
to explain the study and to obtain written consent.114 Included with this
letter was a case-history form which the parents were requested to
complete in order to provide background information on, amongst other
things, the child’s language development and current language abilities. A
copy of this form can be found in appendix C. After obtaining the
written consent, and checking that no information provided on the case-
history form contra-indicated inclusion in the study, arrangements were
made for a non-verbal IQ-score to be obtained,115 where such a score
had not yet been obtained. If this score was 85 or above, the child was
visited at his/her school, home, or therapy centre. During these visits, (i)
the auditory sensitivity of the participant was screened according to the
American Speech, Language, and Hearing Association’s guidelines
(ASHA 1997-2006), if no previous hearing test had been done; (ii) a
language sample was collected; and (iii) the experimental tasks (cf.
section 4.6) were performed.
4.3.2. Typically developing participants
Hereafter, participants for inclusion in the age-matched control group

were obtained by contacting four after-care centres of Afrikaans-medium
institutions and asking the staff to identify typically developing
Afrikaans-speaking 6-year-olds from monolingual Afrikaans-speaking
homes. At the same time, day-care centres were contacted and requested
114 Two participants were enrolled in a preschool for hearing-impaired children, even
though they were not hearing impaired. When this school has the capacity to do so, it
takes in normal hearing children with language problems. These children benefit from
the small class sizes, personal tuition, and regular on-site speech-language and
occupational therapy. The school is linked to a university training hospital. In order to
gain access to these children, the format and content of this letter had to meet the
requirements of the Ethics Committee of the Research Committee of the hospital. For
the sake of consistency, it was decided to use this letter (albeit reader-unfriendly) for all
45 participants.
115 I take note (i) of the problems mentioned by Krassowski and Plante (1997) involved
in testing IQ in children with SLI, and (ii) of the finding by Dethorne and Watkins
(2006) that no significant association can be observed between the non-verbal IQ of
children with language impairment and criterion measures of language.
87
Methodology
to identify all the typically developing Afrikaans-speaking 4-year-olds

from monolingual Afrikaans-speaking homes. These children would then
form the younger typically developing control group.
Only those 6-year-olds who were close to the same age as one of the
children with SLI were taken to be potential participants. Initially, all 4-
year-olds were seen as potential participants. Parental consent for
participation of these 6- and 4-year-olds was obtained via the staff of
these centres, and a case-history form was completed by the parents.
Most of the children were then visited at their school or at the childcare
centre; three 4-year-old boys and two 6-year-old boys were visited at
their homes. During these visits, hearing screening took place as it did
for the participants in the SLI group, a language sample was collected,
and the experimental tasks (cf. section 4.6) were performed.
Regarding the typically 6-year-olds, data were initially gathered from 17

children, including one girl (6 years 6 months) and one boy (6 years 1
month) whose MLUs were later discovered to be uncharacteristically low
in comparison with those of their peers, and therefore their data were
discarded. Inspection of the transcript of their language samples revealed
the following: The girl often role-played, talking on behalf of one of the
figurines with which she played. When doing so, she made use of very
short utterances. This had a negative impact on her MLU. The boy was
very excited by the toys used during language sample collection; this
caused him to frequently name a new toy as he discovered it. These one-
word utterances contributed to an artificially low MLU.
After performing the experimental tasks and calculating the MLU, the
data of three 4-year-olds were omitted; i.e., in total, data were gathered
from 18. One of the three had to be omitted because he did not perform
one of the 15 experimental tasks, namely the tense production one. He
co-operated well during the other 14 tasks and during language sample
collection, but on four occasions (on three different days) refused to
perform this particular task. The reason for this refusal was not clear, but
he became tearful during the last attempt, and therefore it was decided to
discontinue any attempts at completing the full battery of experimental
tasks and to replace this boy with another 4-year-old.
88
Methodology
All experimental tasks and language sample collection were performed

with two 4-year-old girls whose results were not included in the study.
Their MLU was 5.14, which is 0.02 higher than the lowest MLU in the 6-
year-old typically developing group. The reason why they were excluded
was not purely because of their slightly higher MLU, but also because
their parents indicated that they thought their daughters were very
advanced for their age, in terms of language development. For instance,
one of the girls started speaking at an early age (she said two words
consistently at 8.5 months) and always spoke very well compared to her
peers. Because one could not be certain that the language abilities of
these girls were, in fact, typical, they were excluded as participants.
4.4. THE PARTICIPANT GROUPS
A summary of the characteristics of the participants whose data were

included for analyses is given in table 4.1.
4.4.1. Children with SLI
Fifteen Afrikaans-speaking 6-year-olds with language problems – seven

males and eight females – were included in the study. Their specific ages
ranged from 6 years 0 months to 6 years 11 months (M = 6 years 5.3
months). These children had an MLU measured in words (MLUw)
ranging from 3.54 to 5.79 (M = 4.35).116 All 15 children had normal
hearing sensitivity. Their parents and classroom teachers reported age-
appropriate socio-emotional development and an absence of any visible
neurological deficits. Their non-verbal IQ score was 85 (or the
equivalent thereof) or above.
The speech-language therapists made the diagnosis of SLI on the basis

of the results of the following:117
(i) A detailed case-history regarding language and other development,
for all 15 participants with SLI.
(ii) Standardised testing, in the form of one of more of the above-
mentioned three tests developed for Afrikaans-speaking children,
for eight of the participants.
116 See section 4.7.1 for a discussion on the calculation of MLU in this study.
117 Information on the language test results appears in Appendix A.
89
Methodology
Table 4.1. Information on participants
history of
therapy?
Partici-
Gender
School,
MLUw
Family
school
grade
pant
Age
and
SLI
In
SLI 1 M 6:11 4.63 No Mainstream Gr 1 Yes
SLI 2 F 6:8 4.06 No Mainstream Gr 0 Yes
SLI 4 F 6:7 4.00 No For learning disabled Gr 0 Yes
SLI 6 M 6:0 4.30 Possibly Mainstream Gr 0 Yes
SLI 11 M 6:6 4.85 No Mainstream Gr 0a No
SLI 12 F 6:7 4.07 No Mainstream Gr 0a Yes
SLI 13 M 6:5 5.79 No Mainstream Gr 0a Yes
SLI 14 F 6:7 4.16 No For hearing-impaired Gr 0 Yes
SLI 15 F 6:9 4.68 No For hearing-impaired Gr 1 Yes
TD6 1 F 6:11 5.32 No Mainstream Gr 1 N/A
TD6 5 M 6:9 5.90 No Mainstream Gr 1 N/A
TD4 1 M 4:3 4.40 No Mainstream play school N/A
TD4 2 M 4:2 3.91 No At home N/A
TD4 4 F 4:0 4.42 No Mainstream play school N/A
90
Methodology

aThese children lived far from the nearest school and therefore only attended Grade 0
part-time (once or twice a week).
(iii) Testing with non-standardised Afrikaans translations of tests

developed for English-speaking children, for nine of the
participants. The Afrikaans translation of the following tests were
administered by the therapists: (i) TACL - Revised/III (Carrow-
Woolfolk 1985, 1999) on five participants; (ii) Renfrew Language
Scales (Renfrew 1995, 1997) on two participants; (iii) Preschool
Language Scale-3 (PLS-3) (Zimmerman, Steiner, and Pond 1992)
on one participant; and (iv) Test of Language Development -
Primary - Third edition (TOLD-P) (Hammill and Newcomer 1997)
on one participant.
(iv) Informal testing, for four of the participants.
(v) Language sample analysis, with the Language Assessment
Remediation and Screening Procedure (LARSP) (Crystal, Fletcher,
and Garman 1989), for one of the participants, and with informal
procedures, for three of the participants.
Fourteen of the participants with SLI were receiving speech-language

therapy at the time of the study:118 two at a preschool for children with
hearing impairment,119 linked to a university training hospital; one at the
Hearing and Speech Clinic of the same hospital; one at a school for
children with learning disabilities; and 10 from private-practising
therapists.
118 The language of the child (participant SLI12) who did not receive therapy – and
never has – was severely impaired. Her parents had been concerned about her language
development and arranged for an evaluation by a speech-language therapist; they cited
financial reasons for their decision not to commence with therapy after receiving the
results of this evaluation.
119 As stated before, these two participants were not hearing-impaired. Cf. note 114.
91
Methodology
All 15 children with SLI were reported by their speech-language

therapists to demonstrate problems with morphosyntax, but not with
pragmatics. Only one of the children in the SLI group had a possible
family history of SLI: His younger sister reportedly presented with a
language delay, but her language had not been evaluated formally and a
diagnosis of SLI had not been made.
4.4.2. Typically developing controls
Fifteen Afrikaans-speaking children (six males and nine females) aged 6

years 2 months to 6 years 11 months (M = 6 years 6.8 months) formed
the age-matched control group. Their MLUw ranged from 5.12 to 7.10
(M = 5.92).
The younger control group consisted of 15 4-year-old Afrikaans-

speaking children, of whom seven were males and eight were females.
Their ages ranged from 4 years 0 months to 4 years 7 months (M = 4
years 2.3 months) and their MLUw from 3.91 to 5.00 (M = 4.56).
According to their parents and classroom teachers, the 30 participants in

the control groups were typically developing in all respects: Their
language, intellectual, and socio-emotional development were all deemed
to be age-appropriate, and there was no evidence of any visible
neurological deficits. All 30 children exhibited normal hearing sensitivity
during hearing screening and had no previous referral to, or treatment
by, a speech-language therapist.
4.5. COLLECTION OF SPONTANEOUS LANGUAGE
All data were collected by the researcher. During language sample

elicitation, the researcher and participant played alone in a quiet room at
his/her school, care centre or home or in a quiet part of a room in which
other people were also present.120 Language sample elicitation took the
form of freeplay with toys that included (i) little figurines with
accessories such as radios, hats, mugs, and brooms; (ii) wooden building
120Three of the samples were collected with other children taking part in the
conversation: One girl with SLI did not want to participate in the study if her (typically
developing) twin sister could not accompany her to all sessions, and two 4-year-old
boys wanted a friend present.
92
Methodology
blocks; and (iii) plastic kitchen furniture. The researcher initiated the
language sampling interaction by inviting the participant to join her in
kitting out the dolls, building a house, and/or assembling the kitchen. If
the participant was quiet for extended periods, the researcher used a
variety of techniques to encourage conversation, including parallel play,
“engaged” play, self talk, making statements, and question asking (both
wh- and yes/no-questions). These questions were asked about topics
found to be suitable for discussion with preschool children, such as their
families, pets, and birthday celebrations (cf. Southwood and Russell
2004).
In the literature, there is no consensus on the preferred number of

utterances in a language sample, but Crystal, Fletcher, and Garman
(1976:87) suggest that a 30-minute interaction, usually yielding 100 to
200 utterances, will generally suffice. (However, Southwood and Russell
(2004:369) found that even 15 minutes of freeplay elicited an average of
144 utterances from typically developing Afrikaans-speaking 5-year-olds.)
Following Crystal et al. (1976), the language samples collected in this
study were each 30 minutes long.121 An audio-cassette recording was made
of each language sample collection session, using an observable recorder.
4.6. EXPERIMENTAL TASKS
During the performance of the experimental tasks, the researcher and

participant sat next to each other in the room referred to above.
Participants could rest at any stage during any session, and could also
request any particular visit to end.
The full test battery is provided in appendix D. Each task had at least
two practice items, in order to familiarise the participants with the tasks
and with what was required from them. All tasks were first performed
121 There are four exceptions: (i) a 4-year-old boy whose language sample is 15 minutes
long; (ii) the girl with SLI who was recorded with her twin sister, whose sample is
slightly longer than 30 minutes, because her hundredth utterance was only made after
more than 30 minutes of recording; (iii) a typically developing 6-year-old girl, whose
sample was 23 minutes long – her session had to be terminated slightly earlier than
planned; and (iv) a typically developing 6-year-old boy. This boy was co-operative
during the experimental tasks, but did not enjoy the language sampling activity and, 18
minutes into the activity, asked for it to be terminated.
93
Methodology
with typically developing Afrikaans-speaking 3-, 4-, 5-, and 6-year-olds,

during a pilot study, in order to ensure that test items were appropriate
and that the demands placed on the participants were realistic (cf.
Southwood 2005, 2006).
The procedures used to test the participants’ comprehension and

production of grammatical morphemes have been used with success
with young children of different languages by researchers such as
Gualmini and Crain (2002); Gualmini, Crain, and Meroni (2000);
Håkansson (2001); Hansson and Leonard (2003); Jakubowicz (2003);
Leonard et al. (2001); Loeb and Leonard (1991); Marchman et al. (2004);
Ravid, Levie, and Ben-Zvi (2003); and Van der Lely and Ullman (2001).
The general format of the tasks assessing the comprehension of

grammatical morphemes was the same, regardless of whether number,
person, case, or tense was assessed. Two types of tasks were used,
namely picture selection and acceptability judgements.122 Regarding the
general format of the production tasks, this was also the same regardless
of which grammatical feature was assessed: It took the form of sentence
completion. All tasks took the form of researcher-participant interaction
with pictures or picture sheets. Table 4.2 contains a summary of the
experimental tasks and the aspects which they assessed.
Table 4.2. Summary of experimental tasks and the aspects which they
assessed
Aspect assessed Type of task
Comprehension Production
Rega plural forms of real words Picture selection and Sentb completion
judgement
Irreg plural forms of real words Judgement Sent completion
Plural forms of nonsense words Judgement Sent completion
Person and case (on pronouns) Picture selection and Sent completion
judgement
Case (se-constructions) Picture selection Sent completion
Tense (various types of construc- Picture selection Sent completion
tions including hendiadyses)
Hendiadyses only Judgement
aReg=regular. bSent=sentence.
122 Rice, Wexler, and Redmond (1999) found that children as young as 3 are able to
perform such judgements.

94
Methodology
In general, the children were very willing participants. They were

rewarded with stickers after each task and also, for the longer task, at
random intervals during the tasks. At times, the 4-year-olds showed signs
of boredom or fatigue. When this happened, the session was interrupted
to allow the participant to rest. Often, the participant returned to his/her
class and joined the researcher later the same day. However, the 6-year-
olds often completed all the tasks in one or two sessions. They seemed
to enjoy the procedures, and, when all tasks were performed, voiced
their disappointment at not being able to continue with what many of
them called die werkies ‘the work-DIM-PL’.
4.6.1. Number
4.6.1.1. Number: Comprehension tasks

The items of the picture selection task consisted of 40 real words, 20 of
them in the singular form and 20 in the plural form. Of the 20 plurals, 10
took the regular plural suffix –e and the other 10 took the other regular
plural suffix, –s, as shown in table 4.3. (Also see section 1.1 of appendix
D for the items used in this task.)
Table 4.3. Item types in the picture selection task for number
comprehension
Monosyllabic Bisyllabic
Singular 10 10
Plural requiring –e 5 5
Plural requiring –s 5 5
When performing the task, the participant was asked, for example, Wys
vir my die honde ‘Show me the dogs’. The participant would then have to
select the correct picture out of a possible four. As is customary during
the performance of such tasks, four-picture sheets were used (in order to
reduce chance level to 0.25). In the case of Wys vir my die honde, for
example, the sheet contained (i) one picture corresponding to the
requested word (honde ‘dogs’); (ii) one corresponding to the requested
word without the relevant morpheme (hond ‘dog’); (iii) one semantically-
related distracter, in its plural form (katte ‘cats’); and (iv) one
phonologically related distracter (hande ‘hands’). The syllable structure of
the distracters in (iii) and (iv) was similar to that of the word in (i).
95
Methodology
In order to determine whether or not the participant could recognise the

correct phonological realisation of the grammatical feature number, the
participant was required to judge whether plural forms produced by the
researcher were acceptable in Afrikaans. In other words, the participant
was asked to tell the researcher when she made a mistake. The researcher
and participant looked at two-picture sheets: The first picture was that of
a singular object, whereas the second depicted more than one of that
same object. The researcher named the objects, e.g., Hier is een bessie, maar
hier is baie bessies ‘Here is one berry, but here are many berries’.
Regular (–s and –e) and irregular plural forms of both real and nonsense
words were used correctly (e.g., dasse ‘ties’, gesigte ‘faces’, or worre) and
incorrectly (e.g., *sokkiese ‘socks’, *krage ‘collars’, or *siers) by the
researcher. For purposes of statistical analysis, items of each of the plural
judgement tasks were grouped into two groups: those requiring
participants to identify a grammatical plural form correctly, and those
requiring correct identification of an ungrammatical plural form. This
was done because it was expected that the participants would treat these
two types of items differently: It is known that children have a
preference for “positive responses”, i.e., for judging items to be
grammatical rather than ungrammatical (see, e.g., Rice et al. 1999).
The judgement task assessing comprehension of real words requiring

one of the two regular plural suffixes consisted of 14 items. As shown in
table 4.4, four items were correct plural forms: Two took the regular –s
plural suffix and two the –e (e.g., foto’s ‘photographs’ and dasse ‘ties’).
Then there were four items which should have had the –s, but had either
the –e or both of the regular plural suffixes (e.g., *lepele ‘spoons’ or
*sokkiese ‘socks’). Likewise, four items should have been pluralised by –e,
but were pluralised by either –s or both –e and –s (e.g., *roks ‘dresses’ and
*slakkes ‘snails’). The remaining two items were ones which take the
regular –e, but in which the final consonant should have been voiced but
was not. Section 1.2 of appendix D contains the items of this judgement
task.
96
Methodology
Table 4.4. Item types in the judgement task for number comprehension –
real words requiring regular plural suffixes
Noun requiring Noun requiring
–e –s
With correct plural morpheme 2 2
With other regular plural morpheme 2 2
With both regular plural morphemes 2 2
Regular –e but final consonant devoiced 2 --
The judgement task assessing real words which should have irregular
plural forms had 33 items: 16 of these were ungrammatical irregular
plural forms and 17 were grammatical (cf. table 4.5). The items of this
task appear in section 1.3 of appendix D.
real words requiring irregular plural suffixes
Correct irregular plural form 17
Regular –e instead of irregular form 14
Regular –s instead of irregular form 2
Singular form instead of irregular form 0
There were 49 items in the judgement task involving nonsense words (cf.
table 4.6): 24 were words which would take the regular –s plural suffix;
24 the regular –e; and four the –e, but also requiring a change in the
pronunciation of the last consonant. Of the 49 items, 24 were presented
in their correct plural form and 25 were presented either with the
incorrect plural suffix or as a singular form. Section 1.4 of appendix D
contains these items.
nonsense words
Noun requiring Noun requiring
–e –s
With correct plural morpheme 12 12
With other regular plural morpheme 11 7
With both regular plural morphemes 0 0
Regular –e but final consonant devoiced 1 --
Singular form 1 5
97
Methodology
4.6.1.2. Number: Production tasks

When assessing the production of grammatical morphemes related to
number, the participant was shown two pictures on one sheet of paper,
given the singular form of the noun, and requested to provide the plural
form of the noun. For instance, the researcher said, Hier is een blom, maar
hier is baie ... (pointing to the rest of the flowers) ‘Here is one flower, but
here are many ...’. Real words requiring regular and irregular plural
suffixes were included, as were nonsense words (after Berko 1958).
As can be seen in section 1.5 of appendix D and table 4.7, the task
assessing production of plural forms of real words requiring one of the
two regular plural suffixes consisted of 20 items. Ten were nouns
requiring the regular plural suffix –e (such as bal ‘ball’ and koerant
‘newspaper’) and the other 10 the regular –s (e.g., oom ‘uncle’ and venster
‘window’). In each of these two groups, five items were monosyllabic
and five bisyllabic.
Table 4.7. Item types in the production task entailing regular plural forms
Requiring –s Requiring –e
Monosyllabic 5 5
Bisyllabic 5 5
There were 30 items in the tasks assessing production of irregular plural

forms of real words (cf. table 4.8): Five of these were nouns ending in –te
in the plural form (e.g., vrugte ‘fruit-PL’, the plural form of vrug) and five in
–ens (e.g., ouens ‘guys’, the plural of ou). For five items, the word-final [f]
in the singular form is voiced in the plural form, which has the suffix –e
(e.g., diewe ‘thieves’, the plural of dief); for two, the word-final [p] is
voiced (e.g., robbe “seals’, the plural of [rOp]); and for another three, the
word-final [t] is voiced (e.g., honde ‘dogs’, the plural of [hOnt]). Five items
were nouns which have the suffix –e but also require a vowel change in
the stem in the plural form (e.g., vate [fAt«] ‘barrels’ – and not vatte [fat«]
– as the plural form of vat). Lastly, five items were nouns which take the
–e plural suffix, but of which the last consonant of the stem is not
pronounced in the plural form (e.g., boë [bU¥«] ‘bows’ – and not boge
[bU¥«x«] – as the plural of boog). These 30 items occur in section 1.6 of
appendix D.
98
Methodology
Table 4.8. Item types in the production task entailing irregular plural
forms
Requiring –te 5
Requiring –ens 5
Requiring –e and voicing of final consonant 10
Noun requiring –e and vowel change in stem 5
Noun requiring –e and “dropping” of final consonant 5
Forty-eight items were included in the task involving judgement of

nonsense words (cf. table 4.9): 24 were words which would take the
regular –s plural suffix, and the other 24 the regular –e. Section 1.7 of
appendix D contains these 48 items.
Table 4.9. Item types in the production task entailing plural forms of
nonsense words
Requiring –s Requiring –e
Monosyllabic 2 24
Bisyllabic 18 0
Trisyllabic 4 0
4.6.2. Person and case
These tasks assessed the comprehension and production of pronoun

forms. As explained in sections 3.3.1.2 and 3.3.1.4, person and case are
not overtly indicated on Afrikaans nouns and verbs, but are
phonologically overt on pronouns. Because both singular and plural
forms of the pronouns were assessed, these tasks, in fact, assessed not
only person and case, but also number.
4.6.2.1. Person and case: Comprehension tasks

When assessing the comprehension of the grammatical features number
and case, the participant was first told which picture of a woman would
represent the researcher throughout the task as well as which picture of a
child would represent the participant. Thereafter, the participant was
shown a four-picture sheet, and requested to point to the picture that
matched the researcher’s utterance. For instance, the participant was
shown a sheet with pictures of people sitting at a table: one of a woman
(not the researcher) and a child representing the participant; one of two
women; one of a woman and a child (representing the researcher and the
99
Methodology
participant); and one of the child representing the participant on his/her

own. The researcher then said, Wys vir my “Julle sit by die tafel” ‘Show me
“You-PL are sitting at the table”’. The participant thus had a choice
between a picture matching ‘You-PL are sitting at the table’ (the target);
‘They are sitting at the table’ or ‘We are sitting at the table’ (where the
number of the pronoun is the same as that of the target, but the person
differs); and ‘You-SGL are sitting at the table’ (where the person of the
pronoun is the same as that of the target, but the number differs). Some
items contained pronouns in their oblique case, for example Die kat lek
hom ‘The cat is licking him’, where the choice was between Die kat lek
hom/haar/hulle/jou ‘The cat is licking him/her/them/you-SGL’. Section
2.1 of appendix D contains the 32 items used in this picture selection
task: Each person-case-number combination occurred once.
In order to determine whether or not the participant could recognise the

correct phonological realisation of the grammatical features person and
case on pronouns, the participant was required to judge whether
utterances produced by the researcher were acceptable in Afrikaans. The
researcher and participant looked at pictures of people performing
various actions (e.g., of a boy eating ice-cream), and the researcher
uttered a sentence that either was grammatical and matched the picture
(Hy eet roomys ‘He is eating ice-cream’), was ungrammatical but matched
the picture (*Hom eet roomys ‘Him is eating ice-cream’), or did not match
the picture (Hulle eet roomys ‘They are eating ice-cream’). Section 2.2 of
appendix D contains the items used in this task. Again, each person-
case-number combination occurred once, rendering 32 items. Of these,
half were the correct form for the context in which they occurred and
the other half not.
4.6.2.2. Person and case: Production task

To assess the production of person and case, the participant was shown
a picture, given a sentence in which a particular pronoun occurred, and
requested to complete a sentence in which another pronoun occurred.
For instance, pointing to a picture of a boy drinking milk, the researcher
said, Hy drink melk, maar … (pointing to the picture of a girl who is
eating a banana) ‘He is drinking milk, but ...’, or Dit is my skoene (pointing
to a picture of the shoes of the woman representing the researcher), maar
dit is … (pointing to a picture of the shoes of the child representing the
participant) ‘These are my shoes, but these are ...’. The items used in this
100
Methodology
task are listed in section 2.3 of appendix D; each person-case-number

combination had to be produced twice.
4.6.3. Possessive case
These tasks assessed the comprehension and production of se-

constructions.123 Only possessive case was assessed, as other cases are
not indicated by separate morphemes (nor by affixes) in Afrikaans.
4.6.3.1. Possessive case: Comprehension task

When assessing the comprehension of the (phonological) realisation of
possessive case, the participant was shown a picture sheet, and requested
to select one picture, out of a possible three, that matched the
researcher’s utterance. For instance, the participant was shown a sheet
with pictures of a lion (with its tail visible), another lion depicted in such
a manner that its tail is not visible, and the tail of another animal (such as
a crocodile), and the researcher said, Wys vir my die leeu se stert ‘Show me
the lion’s tail’. Section 3.1 of appendix D contains the 10 items used to
assess comprehension of the possessive morpheme se.124
4.6.3.2. Possessive case: Production task

To assess the production of se-constructions, the participant was shown
two pictures, given a se-construction matching the first picture and
requested to provide another such construction matching the second
picture. For instance, pointing to a picture of a man in a car, the
researcher said, Hier is die man se kar en hier is ... (pointing to a picture of a
girl on a bicycle) ‘Here is the man’s car and here is ...’. The 10 items used
to assess the phonological realisation of possessive case in DPs are listed
in section 3.2 of appendix D.
123 I label the se-construction as “possessive case” in Afrikaans, but it might well be a
possessive marker instead of an indication of possessive case.

124 It is debatable whether or not this task did, in fact, test comprehension of the se-
construction. If the child did perceive both nouns in, for example, die man se koerant ‘the
man’s newspaper’, and wanted to select a picture in which both a man and a newspaper
occur, the child would have had only one choice, regardless of whether or not the child
comprehended that se indicates possession.
101
Methodology
4.6.4. Tense
4.6.4.1. Tense: Comprehension tasks

When assessing comprehension of the (phonological) realisation of the
grammatical feature tense, the participant was shown a sheet with three
pictures: one in which an action is being performed, one in which the
action will still be performed, and one in which the action has been
performed. The participant was instructed to study these pictures while
the researcher produced two sentences, each matching one of the
pictures on the sheet. Hereafter, the researcher repeated one of the
sentences and asked the participant to point to the picture matching that
sentence. For instance, the participant was shown a picture sheet with
one picture of a woman preparing to mow the lawn (where the grass is
still long), one of the woman mowing the lawn (where half of the grass is
still long), and one where the woman has mowed the lawn (where all the
grass is now short). While the participant studied these three pictures, the
researcher said Kies eers net met jou oë ‘Die vrou sny die gras’ ‘First select with
your eyes only “The woman mows the lawn”’ and, after a pause, En nou
‘Die vrou het die gras gesny’ ‘And now “The woman mowed the lawn”’.
Then the participant was asked to point to one of the pictures, e.g., the
one matching Die vrou het die gras gesny.125
In section 4.1 of appendix D, the items used to assess comprehension of

the grammatical feature tense are listed. Certain past tense forms, such as
125 During the pilot study, participants were merely asked to point to the picture
matching the researcher’s utterance, i.e., the Kies eers net met jou oë ‘‘First select with your
eyes only’ instruction was not given. However, this proved problematic. When the
researcher uttered a past tense construction, participants often chose the picture in
which the action is being performed instead of the one in which action had been
performed. For instance, when shown a picture of an apple hanging on a tree, one of
the same apple in mid air, and one where the apple is already on the ground and told
Wys vir my ‘Die appel het geval’ ‘Show me “The apple fell’’’, participants often pointed to
the picture in which the apple was still falling. Even though this was not the response
targeted by the researcher, it could not necessarily be classified as being incorrect,
because (i) the picture demonstrated the action (falling) mentioned by the researcher,
whereas the picture of the apple on the ground did not, and (ii) children’s books are
often written in the past tense form but illustrated with pictures showing the actions
told about in the text still being performed. So children are probably used to matching
past tense constructions to present tense pictures. For these reasons, it was decided to
give participants two sentences first, each matching a different picture, before
requesting them to identify one picture matching the stimulus sentence.
102
Methodology
Die seuntjie sou die bal moes kon geslaan het ‘The boy would have had to be
able to hit the ball’ vs. Die seuntjie sal die bal moet kan slaan ‘The boy would
have to be able to hit the ball’, were not assessed. The reason for this is
that these forms are difficult to depict in such a manner that one can be
certain that it is the contrast between [+past] and [-past] that is assessed.
Table 4.10 indicates the types of past tense forms which were assessed.
Table 4.10. Item types in the picture selection task for tense
comprehension
Present tense form of main verb (e.g., sny ‘cut’) 4
Historic present tense form (e.g., gister val hy ‘yesterday he fell’) 2
Temporal het and ge- past participle (e.g., het opgestyg ‘took off’) 4
Temporal het and ge-less past participle (e.g., het ontvang ‘received’) 2
Past tense of be (i.e., was or was gewees) 4
Present tense of have (het) 2
Past tense of have (het gehad) 2
Modal auxiliary – past and present (e.g., sal skoon wees ‘will be clean’ 4
and sou verbrand het ‘would have burnt’)
Hendiadyses were assessed using a judgement task, where the participant

had to judge whether or not sentences such as the following were
correct: Hy het geloop en eet ‘He ate while walking’ or *Gister het die kat
heeldag staan en gemiaau ‘Yesterday the cat mewed all day long’. The 10
items of this task are found in section 4.2 of appendix D.
4.6.4.2. Tense: Production task

To assess the production of the grammatical feature tense, the
participant was shown a picture of a person or animal performing an
action, was told that this action is performed everyday, and was
requested to provide information on what the person or animal did the
day before. For instance, the participant was shown a picture of a boy
brushing his teeth and told Hierdie kind borsel elke dag sy tande. Gister, net
soos elke ander dag, ... ‘This child brushes his teeth every day. Yesterday,
just like every other day, ...’. If the participant used the historic present
tense, which would be appropriate due to the adverb gister which
indicates past tense,126 the researcher provided the temporal auxiliary het,
126 See example (31) in section 3.3.1.5.

103
Methodology
as in Hierdie kind borsel elke dag sy tande. Gister, net soos elke ander dag, het ...
The following types of verbs were included:
(i) Four main verbs which take the ge- prefix in the past participial
form – as in the borsel-example given above.
(ii) Two main verbs which do not take the ge- prefix in the past
participial form – e.g., betaal ‘pay’ in Hierdie vrou betaal elke dag die
verwer. Gister, net soos elke ander dag, ... ‘This woman pays the painter
every day. Yesterday, just like every other day, ...’.
(iii) Two be forms – e.g., Hierdie katjie is elke dag hier. Gister, net soos elke
ander dag, ... ‘This kitten is here every day. Yesterday, just like every
other day, ...’.
(iv) Two have forms – e.g., Hierdie seun het elke dag ’n nuwe maatjie. Gister,
net soos elke ander dag, ... ‘This boy has a new friend every day.
Yesterday, just like every other day, ...’.
(v) Six modal auxiliaries – e.g., Hierdie eendjie wil elke dag swem. Gister, net
soos elke ander dag, ... ‘This duckling wants to swim every day.
Yesterday, just like every other day, ...’.
(vi) Two hendiadyses – e.g., Hierdie man staan elke dag en wag vir die bus.
Gister, net soos elke ander dag, ... ‘Every day, this man stands waiting
for the bus. Yesterday, just like every other day, ...’.
4.7. DATA TRANSCRIPTION AND SCORING
4.7.1. Language sample
The utterances occurring in the first 30 minutes of each language sample

were transcribed orthographically. Hereafter, the first 100 complete and
fully intelligible utterances were identified. Following Hunt (1970:4), an
utterance was considered to be a T-unit, i.e., “one main clause plus
whatever subordinate clause and non-clausal expressions are attached to
or embedded within it”. Accordingly, want ‘because’, en toe ‘and then’, and
en dan ‘and then’ were each taken to introduce a new T-unit, as were en
‘and’ and maar ‘but’ if these two were followed by a clause containing a
verb, as shown in the examples below, where “/” indicates the start of a
new T-unit.
(80)
hy wil nie staan nie / want sy bene kan nie reguit nie
‘He cannot stand / because his legs cannot be straightened’
104
Methodology
(81)
ek het groter geword / en toe verjaar ek in die gim / en toe is ek ses
‘I grew bigger / and then I had my birthday in the gymnasium / and then I was
six’
(82)
hy maak dit alles reg / en dan werk hy en almal
‘He repairs it all / and then he and everybody else work’
(83)
want my ma werk lank / en sy het ’n nuwe werk
‘Because my mom works long hours / and she has a new job’
(84)
ons het ons besems en ons grawe vergeet
‘We forgot our brooms and our spades’
(85)
hulle het ’n werk hier naby / maar hulle wil dit nou in Bellville sit
‘They have offices close by / but they now want to move them to Bellville’
(86)
want ons kragboksie is al baie oud maar baie goed
‘Because our switch board is already very old but very good’
The following where not included in the 100 utterances:

(i) Fillers such as mm or o ‘oh’ (cf. Brown 1973:54).
(ii) Utterances containing unidentifiable material (cf. Unsworth
2005:200).
(iii) Formulaic utterances, such as ek weet nie ‘I don’t know’, kyk hier
‘look here’, wag ‘wait’, or wat’s dit? ‘what’s this?’ (cf. Unsworth
2005:201).
(iv) Exact self-repetitions (cf. Johnston 2001:158), as in (87).
(87)
maar Jani het eintlik twee / ek het net een / ek het net een
‘But Janie actually has two / I have only one/ I have only one’
(v) Exact repetitions of the conversational partner (cf. Johnston

2001:158), as in (88).
105
Methodology
(88)
Adult: ek gaan hierdie hoed vat
‘I am going to take this hat’
Child: ek gaan hierdie hoed vat
‘I am going to take this hat’
(vi) Proper names in response to wh-questions where the response

contained only the so-called queried constituent (cf. Unsworth
2005:200), as shown in (89).
(89)
Adult: in wie se klas is jy?
‘In whose class are you?’
Child: Karen
(vii) Utterances which trailed off (cf. Unsworth 2005:201), as in (90).
(90)
ons kyk nou of die…
‘We now look if the…’
(viii) Ja ‘yes’ and nee ‘no’ (and their equivalents, such as jip, uh, uh-huh,
huh-uh, OK), whether occurring (a) as an answer to a question, as in
(91); (b) as an acknowledgement of the adult’s previous utterance,
as in (92); or (c) during self-talk, as in (93) (cf. Johnston 2001:158-
159).
(91)
Adult: hou jy van kerrie?
‘Do you like curry?’
Child: ja
‘Yes’
(92)
Adult: jy het ’n baie mooi kombuis / alles is blou en wit
‘You have a very pretty kitchen / everything is blue and white’
Child: ja
‘Yes’
106
Methodology
(93)
o die ding moet so staan / nee hy moet so om kom
‘Oh, this thing must stand like this / No, it must be this way around’
The words in the first 100 complete and fully intelligible utterances were
then counted and the mean determined,127 in order to calculate the MLU.
Verbs consisting of noun+verb compounds, such as motorry ‘drive’
(literally ‘car+ride’); adjective+verb compounds, such as mooimaak
‘beautify’ (literally ‘pretty+make’); and preposition+verb compounds,
such as optel ‘pick up’ (literally ‘up+pick’), were counted as one word (see
examples (94a), (95a), and (96a), respectively), unless the verb part of the
compound occurred before the noun, adjective, or preposition, as in
(94b), (95b), and (96b-c).
(94a) dan kan ek fietsry

then can I bicycle+ride
‘Then I can cycle’
(94b) want al die ander kinders ry fiets by die skool
because all the other children ride bicycle at the school
‘Because all the other children ride their bicycles at school’
(95a) hulle wil nie skoonmaak nie

they want-to not clean+make not
‘They do not want to clean’
(95b) die pa maak heeltyd skoon
the dad make all-the-time clean
‘The dad cleans all the time’
(96a) moet hom weer teruggooi

must him again back+throw
‘Must throw him back again’
127 Several researchers have found a high correlation between MLU measured in words
(MLUw) and in morphemes (MLUm) (see, e.g., Arlman-Rupp, Van Niekerk de Haan,
and Van der Sandt-Koenderman 1976; Hickey 1991; Oetting and Rice 1993;
Thordardottir and Weismer 1998). MLUw was chosen above MLUm, as it is a simpler
process to decide what constitutes a word than it is to decide what counts as a
morpheme (cf. Hickey 1991:268). Also, as stated by Miller and Deevy (2003:1157-
1158), care had to be taken not to create a confound: Morphemes were being examined
(in both the experimental tasks and the language samples); therefore, employing MLU
measured in morphemes seemed inappropriate.
107
Methodology
(96b) hy val af
he fall off
‘He is falling off’
(96c) vang hy nou my op?
catch he now me up
‘Is it recording me now?’
Merges of dit/wat/hier ‘it/what/here’ with a preposition (as in

dit+op=daarop ‘on it’), as they appear in examples (97a), (98), and (99),128
were also counted as one word. However, if the preposition occurred
before dit, as in example (97b),129 the preposition and dit were counted as
separate words.130
(97a) moet ouma daarin klim

must granny it+in climb
‘Must granny get into it’
(97b) ek wil kyk of die seuntjie op dit kan ry
I want-to see whether the boy-DIM on it can ride
‘I want to see whether the little boy can ride on it’
(98)
hier is die pad waarop ons gaan werk
here is the road what+on we will work
‘Here is the road on which we are going to work’
(99)
so nou kan ons net goed hierin bêre
so now can we just stuff here+in away-put
‘So now we can just put stuff away in here’
Also for these first 100 complete and fully intelligible utterances, the
number of occurrences of the following was tallied:
128 dit and wat change their form when combined with a preposition: dit changes to daar-
(e.g., op dit changes to daarop) and wat changes to waar- (e.g., uit wat changes to waaruit).
In this regard, see Oosthuizen (2000).
129 Note that hier cannot be separated from its preposition in the way that dit and wat
can. For instance, one can say daarop or op dit, and waarop or op wat, but one can only say
hierop (not *op hier).
130 Only waar- forms (waarop, waarmee, etc.) occurred; there were no occurrences of
forms such as op wat, met wat, etc. in any of the 45 language samples.
108
Methodology
(i) Correct use of single nouns (e.g., my mamma gaan ’n koek bak ‘my
mommy is going to bake a cake’).
(ii) Incorrect use of single nouns (twee *juffrou ‘two teacher’).
(iii) Correct use of plural nouns (ek hou van honde ‘I like dogs’).
(iv) Incorrect use of plural nouns (daar’s twee *byls instead of byle ‘there
are two axes’).
(v) Person and case on pronouns correct (kom ons sit hom in ‘let us
put him in’).
(vi) Person and/or case on pronouns incorrect (waar’s *hom hoed?
instead of sy ‘where’s his hat?’).
(vii) se-construction correct (die pa se stoel is daar ‘the dad’s chair is
there’).
(viii) se-construction incorrect (daar’s hulle *se kos131 instead of hulle kos
‘there’s their food’ or *my sussie skool ‘my sister school’ instead of my
sussie se skool ‘my sister’s school’).
(ix) Each of the various kinds of present tense constructions correct
(ek bak koekies ‘I am baking cookies’; ek wil nog speel ‘I want to play
some more’; ek het die zebra ‘I have the zebra’; jou naels is sterk ‘your
nails are strong’).
(x) Each of the various kinds of present tense constructions incorrect
(nou moet daar nog ’n wit ding in *is instead of wees ‘now there must
still be a white thing in there’; *het jy het net een byl? instead of het jy
net een byl? ‘do you have only one axe?’; *nou’s hy daar sit instead of
nou sit hy daar ‘now he sits there’).
(xi) Use of historic present tense (toe sny ek my hier ‘then I cut myself
here’).
(xii) Each of the various kinds of past tense constructions correct (hulle
het weer afgeval ‘they fell off again’; het sy betaal? ‘did she pay?’;
hierso’s hy wat see toe was ‘here is he who went to the sea’; sy het voor
die tyd daai gehad ‘she had that beforehand’).
(xiii) Each of the various kinds of past tense constructions incorrect
(*hulle seergekry instead of hulle het seergekry ‘they were hurt’; eenkeer
*het hulle baie stout gewees instead of eenkeer was hulle baie stout (gewees)
‘once they were very naughty’; toe het Jessica ’n sakkie gekan kry
131Note that constructions such as hulle se kos are acceptable in certain dialects of
Afrikaans. However, according to their speech therapists and teachers, none of the
participants in this study spoke a dialect in which such a construction would be
acceptable.
109
Methodology
instead of toe kon Jessica ’n sakkie kry /gekry het ‘then Jessica could
get a bag’).
(xiv) Passive constructions in the past tense form (dit was deur ’n hond
gekrap ‘it had been scratched by a dog).
Correct and erroneous occurrences of morphemes or grammatical

features were not tallied from utterance 101 onwards. However, each
utterance which (i) occurred after the hundredth one but before the end
of the 30 minutes, and (ii) was in any way deviant (i.e., non-adult-like)
was identified and placed in a separate data base.
4.7.2. Experimental tasks
All responses on the comprehension and production tasks were recorded

on a score sheet. Self corrections were allowed, and the final response
was the one scored. The codes given to the various responses are
discussed in the chapters 5 to 7, where the results of each specific task
are presented.
As yet, no research has been done to establish the characteristics of SLI

as it presents itself in Afrikaans. In order to answer the specific research
questions posed in chapter 1, the comprehension and production of the
grammatical morphemes related to number, person, case, and tense were
evaluated in the language of Afrikaans-speaking 6-year-olds with SLI and
typically developing Afrikaans-speaking 4- and 6-year-olds. A language
sample was obtained from each of the participants, and a series of
experimental tasks was performed. Based on the available literature (cf.
chapter 2), it was expected that the Afrikaans-speaking children with SLI
would fare worse than their typically developing peers in terms of both
comprehension and production of grammatical features, and that the
responses of the children with SLI would also differ from those of the
younger typically developing children. The findings of the language
sample analyses and experimental tasks are presented in the next three
chapters: Chapter 5 contains the results on number comprehension and
production, chapter 6 on person and case, and chapter 7 on tense.
110
Results: Singular/Plural
Chapter 5
Number comprehension and production
5.1. INTRODUCTION
This chapter contains the results on the comprehension and production

of singular/plural distinctions made by the Afrikaans-speaking 6-year-
olds with SLI, their typically developing peers, and typically developing
4-year-olds. Two experimental tasks were used to evaluate
comprehension of singular/plural distinctions. The first was a picture
selection task, the results of which will be presented in section 5.2. The
second was a judgement task; the results of this task are found in section
5.3. Section 5.4 presents the results of the production tasks. Details on
the items of the experimental tasks related to singular/plural distinctions
are given in section 4.6.1. The way in which singular and plural forms of
nouns were used by the three groups of participants in their language
sample is presented in section 5.5. Finally, in section 5.6, the results of
this study are discussed in light of those of others focusing on plural
production by children with SLI.
5.2. RESULTS: PICTURE SELECTION TASK:

SINGULAR/PLURAL FORMS OF REAL WORDS
This task was meant to provide information on the comprehension of

the three groups of children of the distinction between the singular and
plural forms of real words. The items of this task appear in section 1.1 of
appendix D.
Responses were initially coded in such a way that it was possible to

differentiate between different types of non-targeted responses, so that
error patterns might be sought. Each response of the participants to the
40 items of this task was classified as one of the following:
(i) Correct (targeted) response.
(ii) Incorrect, semantically related noun, but singular when the target
was singular (e.g., hond ‘dog’ instead of kat ‘cat’) and plural when
111
the target was plural (e.g., reënjasse ‘raincoats’ instead of sambrele

‘umbrellas’).
(iii) Correct noun but plurality the opposite of that of the target (e.g.,
katte ‘cats’ instead of kat ‘cat’, or reënjas ‘raincoat’ instead of reënjasse
‘raincoats’).
(iv) Incorrect, phonologically related noun (e.g., hand ‘hand’ instead of
hond ‘dog’, or kastele ‘castles’ instead of sambrele ‘umbrellas’).
(v) Not usable (for instance, where the participant selected more than
one picture).
(vi) Word not known to the participant, i.e., the participant said Ek
weet nie ‘I don’t know’, or asked Wat is dit? ‘What is that?’.
(vii) No response given.
This coding scheme allowed one to see which responses did not give
direct information on the singular/plural distinction of the noun in
question. For most of the items, responses did give such information: To
these items, participants responded with either (i) or (iii), i.e., they either
gave the targeted response or selected the plural instead of the singular
or vice versa. However, to 11 of the 40 items, more than two of the 45
participants did not give responses (i) or (iii). Closer inspection of these
items could reveal whether the non-targeted responses were likely to be
due to a lack of comprehension of the singular/plural form or to a
problem with the item itself (such that the item constituted a word which
was unfamiliar to the participant). The responses to these 11 items are
summarised in Table 5.1 and discussed below.
For tjops ‘chops’ (item 4),132 only 17 of the 45 participants chose the
correct picture. Of the 28 participants who gave a non-targeted response,
12 were in the SLI group, 10 in the four-year-old (TD4) group, and six in
the typically developing 6-year-old (TD6) group. One of these 28
participants gave an unusable response; two indicated that they did not
know the word; and six chose the picture of one chop, nine the
semantically related distracter, and 10 the phonologically related
distracter. The reason for the varied responses to this item could be that
the word tjops is not well-known to young Afrikaans-speaking children;
132 For each task, the item numbers – i.e., the order in which items were presented –
appear in the table containing the item statistics, in this case table 5.2.
112
they tend to use the more generic vleisie ‘meat-DIM’ to refer to both a
chop and chops (as well as to other cuts of meat).
Table 5.1. Responses to the 11 items to which a response other than the
target or the target with reversed plurality was given - Picture
selection task: Singular/plural forms of real words
Item Response
(i) (ii) (iii) (iv) (v) (vi) (vii)
Seël ‘stamp’ 14 6 6 8 1 9 1
Teël ‘tile’ 16 3 13 11 1 1 0
Tjops ‘chops’ 17 9 6 10 1 2 0
Enkel ‘ankle’ 18 9 11 1 1 1 0
Pa’s ‘dads’ 26 1 11 7 0 0 0
Ketels ‘kettles’ 27 4 11 3 0 0 0
Handsakke ‘handbags’ 29 3 13 0 0 0 0
Vurke ‘forks’ 30 1 11 3 0 0 0
Flitse ‘flashlights’ 30 3 12 0 0 0 0
Voet ‘foot’ 32 6 7 0 0 0 0
Nes ‘nest’ 33 0 6 5 1 0 0
To item 6, enkel ‘ankle’, 18 of the 45 participants (five with SLI, two from
the TD4 group, and 11 from the TD6 group) gave a correct response
and 11 selected the plural. Nine participants chose the semantically
related distracter (arm ‘arm’), one chose the phonologically related winkel
‘shop’, and one gave no response. Only five participants explicitly
indicated that they did not know the word; however, it is assumed that
those who chose the semantically or phonologically related distracters
were also not familiar with enkel.
Item 32, teël ‘tile’, was correctly identified by 16 participants. This item
was difficult for all three groups, but fewer participants (five) in the TD6
group gave a non-targeted response than in the TD4 (13) and SLI (11)
groups. All but response type (vii) (viz. no response) occurred: Thirteen
participants chose teëls ‘tiles’, three the semantically related distracter
(plank ‘plank’), and 11 the phonologically related one (tol ‘top’). One
participant gave an unusable response and another (a 4-year-old boy)
explicitly indicated that he did not know the word teël. However, as in the
case of enkel, it is assumed that those who chose the semantically and
phonologically related distracters were also not familiar with teël.
113
Responses (i) to (vii) were all given to item 18, which was seël ‘stamp’.
Only 14 of the participants chose the correct picture. Ten participants
from the SLI group, 13 from the TD4 group, and eight from the TD6
group gave a non-targeted response. Six participants chose the picture of
stamps, six the semantically related brief ‘letter’, and eight the
phonologically related teël ‘tile’. One participant gave an unusable
response, one no response, and another nine indicated explicitly that
they did not know the word seël. The range of response types and
number of non-targeted responses could be an indication that most of
the participants (and not only the nine who explicitly said as much) were
not familiar with the word seël.
Item 22, voet ‘foot’, rendered responses (i), (ii), or (iii). All of the TD6
participants gave the correct response. Of the 13 participants who gave a
non-targeted response, eight were in the SLI group and five in the TD4
group. Seven participants chose voete ‘feet’ and six chose the semantically
related distracter (been ‘leg’). The phonologically related distracter was
hoed ‘hat’; no-one chose this. A reason for the response pattern could be
that the pictures were misleading: The picture of the leg also included a
foot. Some children possibly chose the foot (and not the whole leg) in
the picture of the leg instead of the picture of the foot on its own.
Item 17 was nes ‘nest’. Twelve participants gave a non-targeted response.

Of these, five were from the SLI group, six from the TD4 group, and
one from the TD6 group. Only one of these participants gave an
unusable response. The rest chose either the plural form (six
participants) or the phonologically closely related mes ‘knife’ (five
participants).
The thirtieth item was handsakke ‘handbags’. Sixteen participants did not
give the targeted response. Of these, eight were from the SLI group,
seven from the TD4 group, and one from the TD6 group. In total, 13 of
the non-targeted responses involved response type (iii), i.e., handsak.
Only three of the participants chose the semantically related distracter,
rugsakke ‘backpacks’.
Item 37, pa’s ‘dads’, was correctly identified by 26 participants. Of the 19

participants who gave a non-targeted response, nine each were in the SLI
and TD4 groups and only one was in the TD6 group. Eleven
114
participants chose the picture of one dad, one the semantically related
distracter (seuns ‘boys/sons’), and seven the phonologically related kaas
‘cheese’.
Ketels ‘kettles’ was item 36. Twenty-seven children gave the targeted
response, and 11 chose the picture of one kettle. Four chose the
semantically related koppies ‘cups’ and three the phonologically related
sleutels ‘keys’. Of the children who did not select the target, only one was
in the TD6 group. Another nine and eight were in the SLI and TD4
groups, respectively.
To item 33, vurke ‘forks’, mainly responses (i) and (iii) were given (30 and
11 times, respectively), but one child (a 4-year-old girl) chose the
semantically related messe ‘knives’, and three (two with SLI and one 4-
year-old) chose varke ‘pigs’.
Item 40, flitse ‘flashlights’, was the last item of this task and also the last
item for which more than two of the 45 participants did not give
responses (i) or (iii). Exactly two thirds of the participants did give the
targeted response (i), and 12 gave the singular (iii). Three participants
(one in each group) chose the semantically related ligte ‘lights’.
No clear error pattern in or across groups could be detected. For this

reason, responses to all items were recoded, making only a distinction
between correct (targeted) response and incorrect (not targeted)
response. In other words, response types (ii) to (vii) above were all coded
as being incorrect. These values were used in the analyses that follow.
The reliability of the scale of 40 items was high; Cronbach’s alpha was
.889. The statistics of the individual items are presented in Table 5.2. The
corrected item-total correlation shows how strong the item is correlated
to the total scale. The items in Table 5.2 are ordered from the item
which the participants found the most difficult to that which they found
the easiest, and the order of presentation can be seen from the item
numbers. The proportion of correct answers had a wide range: from .31
to .93. No pattern could be detected (from studying the order of items in
table 5.2) in the type of items which participants found easy or difficult.
115
Table 5.2. Item statistics – Picture selection task: Singular/plural forms of

real words
Item Item Proportion of Corrected
no. participants giving item-total
the correct answer correlation
18 Seël ‘stamp’ .31 .276
32 Teël ‘tile’ .36 .390
4 Tjops ‘chops’ .38 .402
6 Enkel ‘ankle’ .40 .467
13 Uitveër ‘ eraser’ .58 .080
37 Pa’s ‘dads’ .58 .567
34 Ma’s ‘moms’ .60 .573
36 Ketels ‘kettles’ .60 .704
20 Badprop ‘bath plug’ .64 .218
25 Blomme ‘flowers’ .64 .469
30 Handsakke ‘handbags’ .64 .595
11 Boeke ‘books’ .67 .320
26 Dokter ‘doctor’ .67 .019
40 Flitse ‘torches’ .67 .565
33 Vurke ‘forks’ .67 .452
22 Voet ‘foot’ .71 .670
31 Seekoei ‘hippo’ .71 .385
38 Pen ‘pen’ .71 .289
27 Kat ‘cat’ .73 .327
17 Nes ‘nest’ .73 .348
19 Leeus ‘lions’ .73 .503
7 Aarbei ‘ strawberry’ .76 .340
9 Oor ‘ear’ .76 .426
21 Skape ‘ sheep-PL’ .76 .571
39 Glyplanke ‘slides’ .76 .325
16 Borsels ‘brushes’ .78 .503
29 Naels ‘(finger) nails’ .78 .391
5 Spykers ‘ nails’ .78 .519
12 Baadjie ‘jacket’ .78 .458
10 Wasbakke ‘basins’ .80 .512
8 Sambrele ‘umbrellas’ .80 .273
24 Mielies ‘corn cobs’ .80 .319
116
1 Seun ‘boy’ .80 .250

3 Rusbank ‘ couch’ .80 .092
15 Visbak ‘fish bowl’ .82 .187
23 Man ‘man’ .82 .434
35 Emmers ‘buckets’ .82 .322
28 Seesterre ‘starfish-PL’ .87 .472
14 Lepel ‘spoon’ .93 .267
2 Voël ‘bird’ .93 -.019
Five of the 40 items proved to be problematic, in the sense that there

was a low correlation between them and the scale as a whole (<.20).
These items were voël ‘bird’, uitveër ‘eraser’, rusbank ‘couch’, dokter
‘doctor’, and visbak ‘fish bowl’. To the first of these, 42 of the 45
participants gave the targeted response, and the other three chose the
picture of birds. This item had a corrected item-total correlation of
-0.019, indicating that voël was very easy in relation to the other 39 items.
By contrast, uitveër, rusbank, dokter, and visbak proved to be difficult items.
Possible reasons for the poor performance on uitveër, rusbank, and dokter
could be (i) that these preschool children were not yet very familiar with
erasers, as they had not yet made use of erasers in their childcare centres
and therefore did not know the word uitveër; (ii) that these children called
a couch “bank” or “sofa” instead of “rusbank”; and (iii) that dokter
[dOkt«r]) was phonologically too similar to the phonological distracter
(dogter [dOxt«r]) to be distinguishable from it. It is not clear why the
children would find visbak difficult.
The performance of the three groups of participants on the 40 items is
portrayed in figure 5.1. As can be seen from this figure, it appears that
the TD6 group fared better than the other two. The performance of the
SLI and TD4 groups appeared comparable; however, the median of the
SLI group was lower than that of the TD4 group, and more variability
occurred in the SLI than in the TD4 group. The latter was confirmed by
Levene’s statistic of homogeneity of variance: F2,42=6.340; p=.004. The
difference appeared to be between the SLI and TD4 groups on the one
hand (both showing high variability) and the TD6 group on the other
(showing low variability).
117
35
30
Score
25
20
15
10
SLI TD4 TD6
Group
Figure 5.1. Box plot of performance per group - Picture selection task:
The descriptive statistics of the three groups of participants on the 40

items are summarised in Table 5.3. A one-way analysis of variance
(ANOVA) returned a significant outcome, which means that a difference
between the groups could be assumed (F2,42=32.259; p=.000).
Subsequent post hoc comparisons (Tukey’s HSD test; alpha=.05)
showed that the SLI and TD4 groups did not differ from each other and
that the TD6 group differed significantly from both the SLI and the
TD4 groups.
Table 5.3. Summary of performance per group – Picture selection task:

Group N Mean Standard Minimum Maximum
deviation score obtained score obtained
SLI 15 21.33 6.11400 14 32
TD4 15 22.87 4.48596 16 30
TD6 15 32.87 2.19957 28 35
Total 45 25.69 6.82516 14 35
In order to ascertain whether or not there were differences between the

SLI and TD4 groups in terms of their responses to individual items, their
responses to each of the 40 items were re-examined. The items for which
there were notable differences are given in Table 5.4. The number of
correct and incorrect responses in each group is also given. From this
118
table, it appears that, where responses to certain items differed between

the two groups, more members of the TD4 group than of the SLI group
had the item correct. Exceptions here were badprop ‘bath plug’ and
blomme ‘flowers’, to which more members of the TD4 group gave an
incorrect response.
Table 5.4. Items on which SLI and TD4 groups performed notably
differently - Picture selection task: Singular/plural forms of
real words
Item No. in SLI group No. in TD4 group
whose response was whose response was
Correct Incorrect Correct Incorrect
Voet ‘foot’ 7 8 10 5
Spykers ‘nails’ 8 7 12 3
Sambrele ‘umbrellas’ 8 7 12 3
Oor ‘ear’ 8 7 14 1
Wasbakke ‘basins’ 8 7 12 3
Uitveër ‘eraser’ 9 6 12 3
Man ‘man’ 9 6 13 2
Badprop ‘bath plug’ 11 4 5 10
Blomme ‘flowers’ 10 5 6 9
From visual inspection of scatter plots, one of these two groups did not
fare better or worse than the other on either mono- or bisyllabic words.
It also appeared that the two groups did not fare differently on plurals
formed by –e and those formed by –s. However, it did appear that the
SLI group fared worse on correctly identifying the plural forms than did
the TD4 group; in terms of correctly identifying the singular forms, the
responses of the two groups again did not appear to differ.
5.3. RESULTS: JUDGEMENT TASKS: CORRECT/

INCORRECT PLURAL FORMS OF REAL AND
NONSENSE WORDS
Three judgement tasks were performed to assess the comprehension of

number: one with real words which are supposed to take one of the two
regular plural suffixes (cf. section 5.3.1); one with real words supposed to
have irregular plural forms (cf. section 5.3.2); and one with nonsense
words (cf. section 5.3.3).
119
5.3.1. Real words requiring regular plural suffixes
The aim of this task was to establish whether or not the children with
SLI differed from typically developing ones in their ability to identify the
correct and incorrect forms of real words which require regular plural
suffixes. The 14 items of this task appear in section 1.2 of appendix D.
Participants’ responses to these items were coded in one of the following
two ways:
(i) A response was taken to be correct if a grammatical plural form
was judged as such or an ungrammatical one judged as such.
(ii) A response was regarded as incorrect if a grammatical plural form
was judged to be ungrammatical or an ungrammatical one to be
grammatical.
Participants were not requested to provide reasons for their judgements.

However, some spontaneously did so for certain items, and this was not
discouraged. The coding in (i) and (ii) were used for all of these
responses as well, regardless of what was revealed by the voluntarily
provided reason. For example, one boy with SLI said that *slakkes is
ungrammatical, and correctly so, but then provided slakkie ‘snail-DIM’ as
the alternative. This response was taken to be correct – i.e., to belong to
(i) above – even though an incorrect alternative form (at least as regards
plurality) was provided by the participant.
The four items comprising grammatical plural forms were excluded from
the analysis, because they were too easy in comparison with the other 10
items (which were all ungrammatical). The proportion of the correct
answers to the four grammatical items was between .90 and .91;
including them in the scale would have returned negative item-total
correlations. These items were foto’s ‘photographs’, bessies ‘berries’, dasse
‘ties’, and plante ‘plants’. For the remaining 10 items, Cronbach’s alpha
was .934, which indicated a high reliability for these items as a group.
Individual item statistics are presented in table 5.5, ordered from the
item which the participants found the most difficult to that which they
found the easiest. As was the case for the picture selection task, no
pattern could be detected regarding which items participants found easy
and which they found difficult.
120
Table 5.5. Item statistics – Judgement task: Real words requiring regular
plural suffixes
Item Item Correct plural Proportion of Corrected
no. form of word participants item-total
used as item giving the correct correla-
answer tion
11 *Honte Honde ‘dogs’ .36 .614
3 *Sokkiese Sokkies ‘socks’ .38 .709
9 *Brote Brode ‘loaves of
.38 .542
bread’
1 *Kars Karre ‘cars’ .44 .820
13 *Roks Rokke ‘dresses’ .44 .907
4 *Tandes Tande ‘teeth’ .47 .809
6 *Hoenderse Hoenders ‘chickens’ .47 .781
12 *Lepele Lepels ‘spoons’ .47 .753
2 *Appele Appels ‘apples’ .51 .852
8 *Slakkes Slakke ‘snails’ .53 .584
The performance of the three groups is presented in figure 5.2, from

which it appears that the TD6 group fared better than the other two. In
all groups, there was substantial variance, with the greatest found in the
SLI group: Some of the participants with SLI performed on a par with
the worse-performing 4-year-olds, and others performed as well as the
best-performing typically developing 6-year-olds. Levene’s statistic of
homogeneity of variance produced an F-value of 1.711 (df=2,42;
p=.193), which means that the variances were not statistically different.
Of note is the “artificially” good score obtained by one of the 4-year-
olds: This boy (participant 14) judged all items, whether grammatical or
ungrammatical, to be ungrammatical. Also noteworthy is the relatively
poor performance of two of the typically developing 6-year-olds (a girl
and a boy). The girl (participant 37) judged all items to be grammatical,
which is a response mode more often observed in the SLI and TD4
groups. The boy’s (participant 41) responses to the items of this task had
no pattern, frequently alternating between judging items as grammatical
and judging them as ungrammatical.
121
10
8 14
Score
2
41
0 37
SLI TD4 TD6

Group
Figure 5.2. Box plot of performance per group – Judgement task: Real
Table 5.6 presents the details of the performance of the three groups on
this judgement task involving plural forms of real words which should
take one of the two regular plural suffixes. A one-way ANOVA returned
a significant outcome, indicating that a difference between the groups
could be assumed (F2,42=12.64; p=.000). Post hoc analyses (Tukey’s HSD
test; alpha=.05) revealed that the statistically significant differences were
between the SLI and TD6 groups, and between the TD4 and TD6
groups. There was no statistically significant difference between the
mean scores of the SLI and TD4 groups.
Table 5.6. Summary of performance per group – Judgement task: Real

SLI 15 3.93 3.69298 0 10
TD4 15 1.80 2.48424 0 8
TD6 15 7.60 3.29068 0 10
Total 44 4.44 3.95173 0 10
For only one item was there a notable difference between the responses
of the SLI and TD4 groups. This was item 8, *slakkes. As mentioned
above, one of the children with SLI correctly judged this item to be
ungrammatical and then spontaneously gave slakkie ‘snail-DIM’ as
122
alternative. Of the other 14 participants with SLI, seven incorrectly said

*slakkes is grammatical. The pattern in the TD4 group was different:
Only two of these participants judged *slakkes to be ungrammatical and
the rest incorrectly judged it to be a grammatical plural form. It could be
that the 4-year-olds perceived the researcher as saying slakkies ‘snail-DIM-
PL’, which is a grammatical plural form. However, the reason why the 4-
year-olds may have heard slakkies more often than had the SLI
participants then still begs explanation.
5.3.2. Real words which should have irregular plural forms
This task was administered to ascertain whether the three groups of

children differed in terms of their ability to identify the correct irregular
plural forms of real words. The task consisted of 33 items, which are
given in section 1.3 of appendix E. Responses to these items were coded
as for the judgement task involving regular plural forms of real words.
That is,
(i) a response was taken to be correct if a grammatical plural form
was judged as such or an ungrammatical one judged as such;
(ii) a response was regarded as incorrect if a grammatical plural form
was judged to be ungrammatical or an ungrammatical one to be
grammatical.
As was the case for the judgement task involving regular plural forms of
real words, participants were not requested to provide reasons for their
judgements, but some did so occasionally. For example, one 4-year-old
boy and one boy with SLI indicated that wolwe ‘wolves’ is ungrammatical
and then said Dis jakkalse ‘These are jackal-PL’; this response was coded
as incorrect.
As explained in section 4.6.1.1 of chapter 4, items comprising

grammatical plural forms and those comprising ungrammatical ones
were analysed separately. Cronbach’s alpha was .902 for the
ungrammatical forms and .846 for the grammatical ones, which indicated
a high reliability for these two sets of items as groups. Individual item
statistics for the ungrammatical items are presented in Table 5.7, ordered
from the items for which the participants obtained the lowest mean
score to that for which they obtained the highest. As can be seen from
this table, all ungrammatical items were difficult for the participants: The
proportion of correct responses ranged from .13 to .42.
123

ungrammatical plural forms of real words which should have
no. form of word participants giving item-total
used as item the correct answer correlation
5 *Sa[x]e Sae ‘saws’ .13 .471
21 *Kruiwas Kruiwaens .536
.16
‘wheel barrows’
16 *Stofe Stowe ‘stoves’ .18 .646
24 *Kalwe Kalwers ‘calves’ .18 .536
28 *Stadde Stede ‘cities’ .18 .630
18 *Hawes Hawens ‘harbours’ .20 .646
23 *Hempe Hemde ‘shirts’ .20 .646
26 *Tafel- Tafelber[g]e .570
.20
ber[x]e ‘Table Mountains’
32 *Grafe Grawe ‘spades’ .20 .738
33 *Kra[x]e Krae ‘collars’ .20 .601
25 *Vuise Vuiste ‘fists’ .22 .434
27 *Padde Paaie ‘roads’ .22 .606
8 *Glasse Glase ‘glasses’ .24 .700
29 *Bese Beeste ‘oxen’ .27 .667
11 *O[x]e Oë ‘eyes’ .31 .573
2 *Golfe Golwe ‘waves’ .42 .251
The performance of the three groups on the 16 ungrammatical plural

forms is presented in Figure 5.3. The TD6 group appeared to have a
higher median than the other two groups (who both performed very
poorly), but there was a very large amount of variability in the
performance of these typically developing 6-year-olds: Their scores
ranged from 0 to 15 out of a possible 16. This is confirmed by Levene’s
statistic of homogeneity of variance, which was significant (F2,42=5.246;
p=.009). In the SLI group, one girl (participant 25) and one boy
(participant 28) performed better than the rest, obtaining scores
comparable to those of the typically developing 6-year-olds. No clear
reason could be found for the boy’s good performance; the girl had a
preference for judging items as ungrammatical.
124
15
28
12
9
Score
25
0
SLI TD4 TD6
Group
Figure 5.3. Box plot of performance per group – Judgement task: Items
consisting of ungrammatical plural forms of real words which
should have irregular plural forms
Table 5.8 contains the details of the performance of the three groups on
the judgement of ungrammatical items consisting of real words which
should take irregular plural suffixes. Not one of the mean scores was
higher than 40%, indicating that this was a challenging task for all
groups.
Table 5.8. Summary of performance per group – Judgement task: Items

constituting ungrammatical plural forms of real words which
SLI 15 2.73 4.00832 0 14
TD4 15 1.33 1.58865 0 4
TD6 15 6.47 4.67312 0 15
Total 45 3.51 4.20293 0 10
A one-way ANOVA returned a significant outcome, which means that a

difference between the groups could be assumed (F2,42=7.838; p=.001).
Post hoc analyses (Tukey’s HSD; alpha=.05) revealed that the statistically
significant differences were between the SLI and TD4 groups, on the
one hand, and the TD6 group, on the other. There was no statistically
125
significant difference between the mean scores of the SLI and TD4
groups. Also, there was no noteworthy difference between the responses
of the SLI and TD4 groups to any one particular ungrammatical item.
Turning to the responses of the participants to the grammatical items:

These items are ordered in table 5.9, from the items which the
participants found the most difficult to those which they found the
easiest. Individual item statistics are presented in this table; considering
the proportions of participants who gave the correct responses to the
ungrammatical forms (cf. table 5.7), the grammatical items were far easier
for the participants than were the ungrammatical ones. Also, no item was
particularly difficult: The lowest proportion of participants who gave the
correct answer to any one item was .69.

grammatical plural forms of real words which should have
no. participants giving the item-total
correct answer correlation
7 Mae ‘tummies’ .69 .635
9 Vlae ‘flags’ .69 .495
31 Lammers ‘lambs’ .73 .331
4 Brûe ‘bridges’ .76 .505
12 Nagte ‘nights’ .78 .565
10 Blaaie ‘pages’ .82 .648
17 Rûe ‘backs’ .82 .505
20 Skepe ‘ships’ .82 .628
3 Wolwe ‘wolves’ .87 .110
14 Baddens ‘baths’ .87 .473
1 Gesigte ‘faces’ .89 .152
6 Sleepwaens ‘trailers’ .91 .107
13 Gate ‘gates’ .91 .538
30 Ligte ‘lights’ .91 .565
15 Briewe ‘letters’ .93 .424
19 Berge ‘mountains’ .93 .453
22 Duiwe ‘doves’ .93 .692
126
Figure 5.4 presents the performance of the three groups on the 17

grammatical items consisting of the irregular plural forms of real words.
The groups appear to have performed very similarly, all obtaining high
mean scores. In all groups, some participants obtained the maximum
score. There was more variance in the TD6 group than in the other two,
but the difference in variance between the three groups was not
statistically significant (Levene’s test, F2,42=1.887; p=.164).
18
16
14
12
Score
10
6
20
4 28
SLI TD4 TD6

Group
Figure 5.4. Box plot of performance per group – Judgement task: Items
consisting of grammatical plural forms of real words which
Table 5.10 contains the details of the performance of the three groups
on the judgement of grammatical items consisting of irregular plural
forms. A one-way ANOVA returned a non-significant outcome, which
means that a difference between the groups could not be assumed
(F2,42=0.289; p=.750).
Table 5.10. Summary of performance per group – Judgement task: Items

constituting grammatical plural forms of real words which
SLI 15 13.80 3.56802 5 17
TD4 15 14.73 2.46306 9 17
TD6 15 14.27 2.93906 7 17
Total 45 14.50 2.94524 5 17
127
The third last item (item 31), lammers ‘lambs’, was the only one to which
the responses between the three groups differed notably. Only one of
the 4-year-olds judged this grammatical form to be ungrammatical; she
did not provide a reason for her judgement. In the TD6 group, eight
participants said that lammers is ungrammatical: Four gave no reason for
their judgement, but the other four all said that it should be lamme(r)tjies
‘lamb-DIM-PL’. One of the participants in the SLI group did not specify
why she judged lammers to be ungrammatical, but two did: One
spontaneously provided lammetjies as grammatical form and the other
provided skape ‘sheep-PL’. On this item then, the TD4 group
outperformed the other two, with the SLI group faring better than the
TD6 group. Possible reasons for this pattern could be (i) that lammers is
not known to all participants – some of those in the TD6 and SLI
groups simply did not know that word; lamme(r)tjies is the form that they
were (more) familiar with; and (ii) whether the participants in the TD4
group knew lammers or not, they had a preference for judging items as
grammatical, so their high score on this item merely reflected this
tendency.
5.3.3. Nonsense words
The aim of this task was to establish whether the children with SLI
differed from typically developing ones in their ability to identify the
correct plural forms of nonsense nouns. Its items are listed in section 1.4
of appendix D. Initially, participants’ responses to the 49 items of this
task were coded as they were for the judgement tasks involving real
words, but one more response category was added, viz. “word not
known to the participant”. This last response type was only given once,
by a boy with SLI, to the second last item (item 48), which was *sietele.
Because it occurred so infrequently, this response type was later recoded
to “incorrect”. As was the case for the judgement tasks involving plural
forms of real words, participants were not requested to provide reasons
for their judgements but some spontaneously did so occasionally. For
example, one girl with SLI indicated that donne is ungrammatical and
offered *donnetjie ‘don-DIM’ as the grammatical form; this was coded as
incorrect. One 6-year-old girl correctly judged *siemettieë to be
ungrammatical, but then provided the singular form, siemettie, instead of
siemetties. She also correctly judged *sles to be an ungrammatical plural
form of sles, but then spontaneously said that the grammatical plural
128
form should be *sles (which is what was originally said by the researcher)
− slesse would, in fact, be the grammatical plural form. The latter two
responses were still taken to be correct, even though an incorrect
alternative was provided by the participant.
As in the case of items of the judgement tasks consisting of real words,

items comprising grammatical plural forms of nonsense words and those
comprising ungrammatical ones were analysed separately. For the 25
ungrammatical items, Cronbach’s alpha was .960, indicating a very high
reliability for these items as a group. Statistics of the individual items of
the ungrammatical subgroup are presented in table 5.11, ordered from
the item for which the participants obtained the lowest mean score to
that for which they obtained the highest. This table indicates that all
items were difficult for the participants: Even the one on which they
fared the best was responded to correctly by only 38% of the
participants.
Figure 5.5 depicts the three groups’ performance on the ungrammatical

items consisting of nonsense words. The SLI and TD6 groups
performed very similarly. Also, there seemed to be a far greater degree of
variance in the SLI and TD6 groups (in both groups, some participants
obtained the minimum score and others the maximum); however, when
the responses of participant 14 (the outlier) were included in the
analyses, a non-significant Levene’s statistic was obtained (F2,42=2.414;
p=.102). Of note is the uncharacteristically good score of this participant
14, a 4-year-old boy: For this task, he judged all items, whether
grammatical or not, as being ungrammatical. This caused him to give a
correct response to every item in this subgroup, as all items were
ungrammatical.
129
Table 5.11. Item statistics – Judgement task: Ungrammatical plural forms

of nonsense words
24 *Metoekiere Metoekiers .16 .742
29 *Snuurs Snure .18 .699
22 *Sils Sille .20 .786
28 *Treens Trene .20 .582
44 *Laaps Lape .20 .403
15 *Siemettieë Siemetties .22 .571
26 *Koeps Koepe .22 .746
39 *Safs Sawwe .22 .694
21 *Suuptere Suupters .24 .733
47 *Siefaard Siefaards .24 .769
14 *[dÏrx«] [dÏrg«] .27 .655
49 *Rieks Rieke .27 .703
16 *Sles Slesse .29 .772
17 *Nollerde Nollerds .31 .741
25 *Foutemme Foutems .31 .668
35 *Meks Mekke .31 .721
43 *Sapenne Sapens .31 .767
45 *Gants Gante .31 .608
48 *Sietele Sietels .31 .641
3 *Fooms Fome .33 .693
5 *Tonke Tonkes .33 .648
46 *Oeselaar Oeselaars .33 .726
4 *Sotta Sottas .36 .662
42 *Fleeu Fleeus .36 .779
1 *Oks Okke .38 .628
130
25 14
20
15
Score
10
0
SLI TD4 TD6
Group
Figure 5.5. Box plot of performance nce per group − Judgement task:
Table 5.12 contains the details of the performance of the 45 participants

(per group) on the judgement of ungrammatical plural forms of
nonsense words. A one-way ANOVA returned a non-significant
outcome, which means that a difference between the groups could not
be assumed (F2,42=2.371; p=.106).

SLI 15 8.93 8.68057 0 25
TD4 15 3.33 6.55381 0 24
TD6 15 8.33 7.81634 0 25
Total 44 6.45 7.56201 0 25
The items for which there were noteworthy differences between the
responses of the SLI and TD6 groups are given in table 5.13. The
number of correct and incorrect responses in each group is also given.
As can be seen from this table, for two of the items to which responses
differed between the two groups, more members of the SLI group than
of the TD6 group had the item incorrect. However, for three items more
participants in the SLI group gave the correct response. No pattern in
131
terms of kind of suffixation (i.e., no suffix; –e instead of –s, or –s instead

of –e) could be detected.
Table 5.13. Items on which the three groups performed noticeably

differently – Judgement task: Ungrammatical plural forms of
nonsense words
Item No. of SLI participants No. of TD6 participants
whose response was whose response was
Correct Incorrect Correct Incorrect
*Sotta 5 10 10 5
*Oks 5 10 9 6
*Rieks 8 7 3 12
*Sietele 8 6 3 12
*Oeselaar 9 6 5 10
Turning to the grammatical items, Cronbach’s alpha was .951, indicating

a very high reliability for these 24 items as a group. Individual items’
statistics are presented in table 5.14, ordered from the most difficult to
the easiest item.
In figure 5.6, the three groups’ performance on the grammatical items

consisting of nonsense words are depicted. This performance appears to
overlap to a great extent, with a higher median and less variability in the
TD4 group than in the other two. Levene’s statistic of homogeneity of
variance was not significant (F2,42=1.404; p=.257). In all three groups,
some participants fared very well, obtaining the maximum score. The
performance of three participants warrants further comment. The first is
the 4-year-old boy who judged all items to be ungrammatical. Whereas
this rendered a perfect score on the previous subgroups of items (the
ungrammatical ones), it rendered a score of zero for the items of this
subgroup, as all of these items were, in fact, grammatical. The second 4-
year-old who performed markedly worse than the rest was a girl in
whose responses no clear pattern could be detected. The 6-year-old girl
who performed worse than her peers had a preference for judging items
as being ungrammatical, providing interesting alternatives to the items
given by the researcher.
132
Table 5.14. Item statistics – Judgement task: Grammatical plural forms of

nonsense words
Item Item Proportion of participants Corrected item-total
no. giving the correct answer correlation
7 Iptas .68 .662
8 Dese .70 .746
40 [slÏrg«] .73 .421
34 Tieme .75 .729
12 Kloke .77 .794
23 Tisse .77 .498
27 Spiperds .77 .667
6 Fekke .80 .725
9 Seeus .80 .753
18 Foukes .80 .660
19 Slake .80 .725
20 Nalle .80 .595
31 Donne .80 .613
37 Piwwe .80 .586
38 Tanders .80 .660
2 Pure .82 .734
13 Koenaards .82 .569
30 Spalaars .82 .763
32 Kuens .82 .675
10 Kwamies .84 .748
41 Kêlonniers .84 .435
11 Bokels .86 .684
33 Korrems .86 .651
36 Loese .89 .619
133
25
20
15 12
Score
10
34
0 14
SLI TD4 TD6

Group

Table 5.15 contains the details of the performance per group on the
judgement of grammatical plural forms of nonsense words. A one-way
ANOVA returned a non-significant outcome, which means that a
difference between the groups could not be assumed (F2,42=1.803;
p=.177). There was not one item for which there was a marked
difference between the responses of the three groups.

SLI 15 16.13 8.17546 1 24
TD4 15 21.00 6.42540 14 24
TD6 15 18.93 6.38600 1 24
Total 44 19.11 6.65850 1 24
5.4. RESULTS: SENTENSE COMPLETION TASKS:

PLURAL FORMS
Three production tasks were performed. In the first one, the results of
which are discussed in section 5.4.1, the participants were required to
produce the plural forms of real words which require one of the two
regular plural suffixes. The second task involved the production of
134
irregular plural forms of real words (cf. section 5.4.2), and the third one
the production of plural forms of nonsense words (cf. section 5.4.3).
5.4.1. Real words requiring regular plural suffixes
The aim of this task was to ascertain whether Afrikaans-speaking

children with SLI experience more difficulty than typically developing
children in the production of regular plural forms. The task comprised
20 items, presented in section 1.5 of appendix D. Initially, a separate
code was given to each of the following provided by the participants:
(i) A plural form with the –e suffix, correctly so.
(ii) A plural form with the –e suffix, correctly so, but also with a
deleted syllable, rendering a word that was still a real word, e.g.,
broeke ‘trousers and shorts’ instead of langbroeke ‘trousers’.
(iii) A plural form with the –e suffix, correctly so, but also with syllable
deletion, rendering a nonsense word, e.g., *boue instead of geboue
‘buildings’.
(iv) A plural form with –s as suffix instead of the targeted –e, e.g., *pops
instead of poppe ‘dolls’.
(v) A plural form with an irregular suffix instead of the targeted –e,
for example, *huiste instead of huise ‘houses’.
(vi) A singular form instead of a plural with the –e suffix, e.g.,
*langbroek instead of langbroeke.
(vii) A singular form in the diminutive instead of the plural with the –e
suffix, e.g., *poppie instead of poppe.
(viii) A singular form instead of a plural with –e suffix, but also with
syllable deletion, e.g., *bel instead of oorbelle ‘earrings’.
(ix) A plural form with the –s suffix, correctly so.
(x) A plural form with the –s suffix, correctly so, but also a diminutive
form, e.g., omies instead of ooms ‘uncles/men’.
(xi) A plural form with –e as suffix instead of the targeted –s, e.g., *ome
instead of ooms.
(xii) A plural form with an irregular suffix instead of the targeted –s,
for example, *tenktes instead of tenks ‘tanks’.
(xiii) A singular form instead of the plural with the –s suffix, e.g.,
*venster instead of vensters ‘windows’.
(xiv) Ek weet nie ‘I don’t know’, Wat is dit? ‘What is that?’, or similar.
(xv) No response given.
135
This was done in order to differentiate between the types of errors that
children made (for instance, omission of plural suffix) and the type of
plural forms for which errors were made (for instance, on a bisyllabic
word requiring a –e as plural suffix). However, some of the codes were
not used frequently. Regarding incorrect plural forms of nouns pluralised
by –e, code (ii) was used seven times, code (iii) three times, code (vii)
twice, and code (viii) only once. For incorrect forms of nouns pluralised
by –s, code (x) was used 11 times. Only on one occasion did a participant
(a 4-year-old) say that she did not know the answer (for ghoen), and one
boy with SLI gave no response on three occasions. In total, 12 of the
responses were unusable. The following four irregular plural forms were
provided instead of regular –e forms: *oorbelles (by one girl with SLI),
*geboues (a 4-year-old boy), *huiste (a 4-year-old girl), and *meste (a 6-year-
old boy). The following irregular plural forms were provided instead of
regular –s forms (six times in total): *tenkes (by two 4-year-old boys and
one 4-year-old girl), *tenktes (a 4-year-old girl), *lêert (a 4-year-old boy),
and *fliekse (a 4-year-old girl).
For further analysis, the responses were recoded: Codes (i) to (iii) and
code (ix) were taken to be correct responses and all others were taken to
be incorrect, i.e., indications that the participant was not able to produce
the correct plural form.
The reliability of the 20 items was acceptable, but not high: Cronbach’s
alpha was .731. The statistics of the individual items are presented in
table 5.16, ordered from the item which the participants found the most
difficult to that which they found the easiest. No pattern could be
detected (from studying the order of items in table 5.16) in the type of
items which participants found easy or difficult, apart from the five most
difficult items all being monosyllabic words requiring the plural suffix –s.
Two items were subsequently removed; Cronbach’s alpha then improved

slightly to .755. The first of these items was skoen ‘shoe’ (plural: skoene).
This item proved to be too easy – 44 of the 45 participants gave the
targeted response (the other one said *skoen) – and its corrected
correlation to the rest of the items was low (.058). The second item to be
removed was lêer ‘folder’. Only 11 participants gave the targeted response
to this item. Although not one of the other 34 participants (explicitly)
indicated that they did not know the word – 26 said *lêere, seven said
136
*lêer, and one gave an irregular plural form – it could indeed be that these
participants (especially the 4-year-olds) were not familiar with this item.
If this was the case, then lêer, in fact, tested pluralisation of a nonsense
word rather than a real word.
Table 5.16. Item statistics – Production task: Plural forms of real words
requiring regular plural suffixes
Item Item Proportion of Corrected item-
no. participants giving the total correlation
correct answer
11 Oom ‘uncle’ .22 .318
2 Fliek ‘movie’ .24 .314
6 Tenk ‘tank’ .24 .238
20 Lêer ‘folder’ .24 -.048
1 Ghoen ‘marble’ .38 .364
9 Gebou ‘building’ .69 .364
7 Koerant ‘newspaper’ .76 .469
14 Oorbel ‘earring’ .80 .457
16 Kersboom
.80 .133
‘Christmas tree’
5 Bottel ‘bottle’ .82 .210
8 Beker ‘mug’ .82 .466
13 Pop ‘doll’ .82 .129
15 Venster ‘window’ .87 .626
17 Lekker ‘sweet’ .87 .404
18 Mes ‘knife’ .89 .101
4 Langbroek ‘trousers’ .89 .505
12 Mandjie ‘basket’ .91 .461
3 Bal ‘ball’ .93 .169
19 Huis ‘house’ .96 .233
10 Skoen ‘shoe’ .98 .058
With the scoring altered and the two problem items removed, the
performance of the three groups was compared. This comparison is
depicted in figure 5.7. As can be seen in this figure, the median of the
TD6 group was the highest of the three. Although the SLI and TD4
groups appeared to have the same median, the variability was higher in
137
the SLI group than in the TD4 one, as confirmed by Levene’s statistic of
homogeneity of variance (F2,42=4.391; p=.020).
17.5
15.0
Score
12.5
10.0
7.5 6
5.0
21
2.5
SLI TD4 TD6
Group
Figure 5.7. Box plot of performance per group – Production task: Plural
forms of real words requiring regular plural suffixes
Table 5.17 shows the mean scores per group. These scores of the SLI
and TD4 groups were comparable (12.07 and 11.93 out of 18,
respectively). However, the minimum score of the SLI group was lower
and the maximum one higher than that of the TD4 group.
Table 5.17. Summary of performance per group – Production task: Plural

forms of real words requiring regular plural suffixes
SLI 15 12.07 3.88158 3 17
TD4 15 11.93 2.43389 7 14
TD6 15 14.73 1.48645 12 17
Total 45 12.91 3.01377 3 17
A one-way ANOVA was carried out on the responses of the 45

participants; it returned a significant outcome, which means that a
difference between the groups could be assumed (F2,42= 4.839; p=.013).
Post hoc analyses (Tukey’s HSD; alpha=.05) revealed that the difference
was not between the SLI and TD4 groups. It was the scores of the SLI
and TD6 groups, and those of the TD4 and TD6 groups, which differed.
138
This means that the TD6 group outperformed the other two groups. In
order to ascertain whether there were differences between the responses
of the SLI and TD4 groups to certain items, their responses to each of
the 18 items were examined. Such differences could be noticed for two
items, both of them monosyllabic nouns requiring the –s suffix. These
differences are discussed below.
Six of the participants with SLI and two 4-year-olds produced the correct
plural form of item 2 (fliek). Nine of the TD4 participants provided the
singular form instead of flieks. By contrast, only one of the participants
with SLI did so.
Item 6 was tenk. None of the 4-year-olds provided the correct plural
form, but six of the participants with SLI did. Four 4-year-olds produced
an irregular form as plural form (either *tenkes or *tenktes), whereas only
one girl with SLI did so (*tenktwee, literally ‘tanktwo’).
Another monosyllabic noun requiring –s which rendered different

responses from the two groups was the last one (item 20), lêer. As stated
above, this item was removed before the data were analysed statistically.
However, as was the case for tenk, none of the 4-year-olds provided the
correct plural form, but six of the participants in the SLI group did. As a
group, the 4-year-olds preferred *lêere (10 of the 15 gave this response),
whereas six of the participants with SLI produced a plural with –e instead
of –s.
In short then, regarding the production of regular plural forms of real

words, the participants with SLI performed on a par with the 4-year-
olds, and significantly worse than the typically developing 6-year-olds.
We now turn to the results of the task requiring participants to produce
irregular plural forms.
5.4.2. Real words which have irregular plural forms
The aim of this task was to assess whether children with and without SLI
differed in their ability to produce irregular plural forms, i.e., plural forms
which are “exceptions to the rule”. Responses to the 30 items (given,
with their English translations, in section 1.6 of appendix D) were coded
in such a way that it was minimally possible to see whether participants
139
(i) gave a correct response, (ii) produced a singular form, (iii) indicated
that they did not know the item or the answer to the item, (iv) gave no
response, (v) produced a plural form with the suffix –e, or (vi) produced
a plural form with the suffix –s.
There were four items to which no participant gave the correct response.
These were items 4 (weg, plural: weë), 5 (gebed, plural: gebede), 7 (hof, plural:
howe), and 8 (lid, plural: lede). Items 4, 7, and 8 were probably “nonsense
words” to the participants; a high score on these items was not expected,
as these are words rarely used or heard by young Afrikaans-speaking
children. However, the fact that not one participant could provide the
correct plural form of gebed ‘prayer’ was surprising, as most of the
children attended childcare centres of the type in which morning prayer
is an institution. One would therefore assume that at least some of them
would be familiar with the word gebede – or would at least produce
gebedjies ‘prayer-DIM-PL’. Nevertheless, all four of these items were
removed before further analyses were done. Cronbach’s alpha for the
remaining 26 items was .851. The statistics of the individual items are
presented in table 5.18, ordered from the items which the participants
found the most difficult to those which they found the easiest. As can be
seen from this table, there was an extremely wide range in the difficulty
of items, from .02 to .80.
Those items which were negatively correlated to the group as a whole

were then removed. These were items 2 (bevel), 17 (glimlag), and 21 (vat).
Two further items (9, boog, and 14, kroeg) were also removed, as very few
participants gave the correct response to them. This somewhat improved
the reliability for the remaining 21 items as a group: Cronbach’s alpha
was then .860.
With the nine problem items removed, the performance of the three
groups was compared. Figure 5.8 depicts this comparison: The TD6
group seemed to outperform the other two but also showed more
variability. This was confirmed by Levene’s statistic (F2,42=6.107;
p=.005). The median of the SLI and TD4 groups were similar.
140
Table 5.18. Item statistics – Production task: Plural forms of real words
that should have irregular plural forms
9 Boog Boë .02 .328
17 Glimlag Glimlagte .02 -.084
21 Vat Vate .02 -.053
2 Bevel Bevele .04 -.030
14 Kroeg Kroeë .04 .197
6 Sif Siwwe .13 .401
18 Kas Kaste .13 .518
24 Vraag Vrae .18 .450
29 Web Webbe
.18 .308
[vEp]
3 Gas Gaste .20 .604
19 Skyf Skywe .20 .643
16 Vlieg Vlieë .22 .407
13 Wa Waens .27 .644
22 Dief Diewe .27 .266
28 Brood Brode
.31 .273
[brU¥«t]
25 Insek Insekte .33 .513
1 Skroef Skroewe .36 .472
11 Golf Golwe .36 .581
10 Ou Ouens .38 .595
12 Vrou Vroue(ns) .38 .446
15 Bed Beddens .42 .640
27 Rob Robbe
.42 .337
[rOp]
20 Vrug Vrugte .49 .289
23 Kind Kinders .73 .245
26 Hond Honde
.80 .424
[hOnt]
30 Hand Hande
.80 .264
[hant]
141
20
15
Score
10
30
22
21
0
SLI TD4 TD6
Group
forms of real words that should have irregular plural forms
The score of three of the participants with SLI differed notably from
that of the other participants in their group: Participant 21 was a boy
with SLI. To 13 of the 21 items he gave either no or an unusable
response, or said that he did not know the answer. Participant 22 was
also a boy. He responded to all 21 items, all of his responses were usable,
and he did not say Ek weet nie ‘I don’t know’ to any item, but he
produced only two correct irregular plural forms. Participant 30, a girl
with SLI, outperformed the rest of the participants in the SLI group and
most of those in the TD4 group. The interesting aspect of this girl’s
responses is that, were she did not give the correct irregular plural form,
she on more than one occasion provided “her own” irregular form, such
as *wawe (instead of waens, the plural of wa) and *wewwe (instead of webbe,
the plural of [vEp]).
In table 5.19, the mean scores are presented per group. The mean,
minimum, and maximum scores of the SLI and TD4 groups were very
similar (means of 5.00 and 5.20 out of 21, respectively). The mean score
of the TD6 group was more than double that: 12.47. However, this
group also showed the most variability, with some participants obtaining
a score of 3 and others 18.
142
Table 5.19. Summary of performance per group – Production task: plural

forms of real words that should have irregular plural forms
SLI 15 5.00 2.17124 1 9
TD4 15 5.20 2.73078 1 10
TD6 15 12.47 4.54920 3 18
Total 45 7.56 4.77472 1 18
A one-way ANOVA returned a significant outcome, indicating that a

difference between the groups could be assumed (F2,42= 24.781; p=.000).
As was the case for the production of plural forms of real words
requiring regular plural suffixes, post hoc analyses (Tukey’s HSD;
alpha=.05) revealed that the difference for the production of irregular
plural forms was not between the SLI and TD4 groups. Rather, the
differences were between the scores of the SLI and TD6 groups, and
between those of the TD4 and TD6 groups.
This means that the TD6 group yet again outperformed the other two
groups and that there was no statistically significant difference between
the SLI and TD4 groups. In order to ascertain whether there were
differences between the responses of the latter two groups to certain
items, their responses to each of the 21 items were examined. Such
differences could be noticed for six items; some of these differences are
discussed below.
None of the participants in the SLI group and only two in the TD4
group gave the correct response (gaste) to item 3, gas. The participants
with SLI gave a comparable number of responses entailing regular –e
(i.e., *gasse) and the singular form (five and seven, respectively).
However, the participants in the TD4 group showed a preference for the
regular –e plural form (seven responses) over the singular form (two
responses).
Item 15 was bed (pronounced [bEt]). In the SLI and TD4 groups, four
participants each gave the correct plural form, beddens. More participants
in the SLI than in the TD4 group said *bedde (five vs. two), with the TD4
group showing a preference for *bette (nine, vs. two in the SLI group).
143
For the next item, item 16 (which was vlieg), only two participants with
SLI and one in the TD4 group gave the correct plural form, vlieë. The
TD4 group had a strong preference for the regular –e (*vliege): 13 of the
15 gave this response, compared to 5 participants in the SLI group. Of
the participants in the latter group, three gave *vliegs as plural form,
whereas no TD4 participant produced this form.
Both groups fared very poorly on item 19, skyf. Of the 30 participants,
only one (a girl with SLI) gave the correct plural form, skywe. The
response given by most participants with SLI was the singular form
(eight participants), whereas the response favoured by the TD4
participants was the regular –e plural without voicing of the last
consonant, *skyfe (11 participants).
The next item was vrug. Here again, more participants with SLI (nine of
them) than TD4 participants (three) gave the correct vrugte. None of the
participants with SLI gave the singular form, whereas four TD4
participants did. However, the response given by most of the TD4
participants was *vrugge (seven, compared to the five in the SLI group).
We now turn to item 12, vrou, which can be pluralised as either vroue or
vrouens. In total, 11 participants in the SLI group and 14 in the TD4
group gave a correct plural form, but the TD4 group preferred the
regular form vroue (11, compared to six in the SLI group). Also, none of
the participants in the TD4 group gave the regular –s form *vrous,
whereas three participants with SLI did.
For the production of both regular and irregular plural forms, the TD6
group outperformed the other two. Although there were no statistically
significant differences between the performance of the SLI and TD4
groups, some differences in the types of response were noticed: In terms
of regular plural forms, monosyllabic nouns requiring –s as suffix
appeared to be more difficult for the TD4 group than for the SLI one.
In terms of nouns with an irregular plural form, the 4-year-olds showed a
clear preference for replacing these forms with regular –e plural forms.
This preference was not noted for the SLI group; rather, in total, this
group used the regular –s plural form 32 times, compared to the nine
times by the TD4 group.
144
5.4.3. Nonsense words
This tasked was aimed at ascertaining whether the Afrikaans-speaking

children with SLI differed from typically developing ones in terms of
their ability to form the plural of words which are novel to them. The
task comprised 48 items, given in section 1.7 of appendix D.
Participants’ responses were coded similarly to the ones for the real
words requiring regular plural suffixes (cf. section 5.4.1), but extra codes
were added to indicate whether (i) the nonsense word was replaced by a
real word (as in *kere ‘times’ instead of knure, the plural of knuur); or (ii)
the nonsense word was replaced by another nonsense word (as in *knude
as the plural form of knuur).
Cronbach’s alpha for the 48 items of this task was .945, indicating a high
reliability for the items as a group. The statistics of the individual items
are presented in table 5.20, ordered from the most difficult to the easiest
item. There was a wide range in the difficulty of the items, but not one
of them was particularly easy: The proportion of participants giving the
correct response ranged from .02 to .62.
Table 5.20. Item statistics – Production task: Plural forms of nonsense

words
Item Item Correct Proportion of Corrected
no. plural form participants item-total
of word used giving the correct correlation
as item answer
38 [stÏrx] [stÏrg«] .02 .112
35 [lÏrx] [lÏrg«] .04 .198
16 Lif Liwwe .07 .301
5 Tef Tewwe .09 .262
21 Lerke Lerkes .11 .254
24 Saan Sane .16 .133
43 Tiese Tieses .16 .278
7 Knuur Knure .22 .200
11 Latoewier Latoewiers .24 .458
14 Gol Golle .27 .162
36 Keen Kene .27 .549
13 Beeu Beeus .29 .462
145
22 Diemounier Diemouniers .29 .475

34 Wor Worre .29 .567
20 Luur Lure .33 .560
33 Ges Gesse .33 .479
46 Wis Wisse .33 .534
15 Kreeu Kreeus .36 .365
26 Fant Fante .36 .519
28 Stiel Stiele .36 .419
29 Hal Halle .36 .709
30 Goom Gome .36 .465
18 Sook Soke .38 .534
25 Woek Woeke .38 .584
40 Dies Diese .38 .485
47 Pygter Pygters .38 .667
19 Klaat Klate .40 .510
32 Fiender Fienders .40 .646
39 Fasel Fasels .40 .592
41 Mek Mekke .40 .596
17 Dissem Dissems .42 .471
42 Loet Loete .42 .613
48 Assa Assas .42 .564
44 Beel Bele .44 .474
8 Toelem Toelems .47 .510
1 Apoenaar Apoenaars .47 .470
2 Foten Fotens .49 .563
3 Biesaard Biesaards .49 .467
27 Kolla Kollas .49 .746
37 Kiemaard Kiemaards .49 .555
45 Kottel Kottels .49 .572
9 Swelaar Swelaars .51 .675
6 Lienkert Lienkerts .53 .655
23 Waken Wakens .53 .602
31 Banalie Banalies .53 .720
4 Lil Lille .60 .511
10 Pekkerd Pekkerds .60 .617
12 Slofie Slofies .62 .495
146
Figure 5.9 portrays the comparison of the performance of the three

groups, indicating again that the TD6 group appeared to fare better than
the other two, and that the SLI and TD4 groups fared similarly. More
variability may seem to occur in the SLI and TD4 groups than in the
TD6 group, but Levene’s statistic was not significant (F2,42= 2.928;
p=.065). Of note is the performance of one typically developing 6-year-
old: She fared markedly worse than the other participants in her group.
In fact, her performance was worse than that of many of the participants
in the SLI and TD4 groups. She gave only six correct responses; her
other 42 responses all consisted of singular forms. In two cases, she
replaced the nonsense word with a real word: *klier ‘gland’ instead of
knure as the plural of knuur, and *stoel ‘chair’ instead of stiele as the plural
form of stiel.
40
30
20
Group
10
33
0
SLI TD4 TD6
Group
Table 5.21 contains more detail on the scores, presented per group. As
was the case for the production of irregular plural forms of real words,
the mean, minimum, and maximum scores of the SLI and TD4 groups
were very similar. Again, the mean score of the TD6 group was more
than double that of either of the other two groups.
147
Table 5.21. Summary of performance per group – Production task: Plural

SLI 15 11.07 10.48446 1 32
TD4 15 13.93 10.41610 2 33
TD6 15 27.27 7.38209 6 36
Total 45 17.42 11.73861 1 36
A one-way ANOVA returned a significant outcome, indicating that a

As was the case for the production of plural forms of real words, post
hoc analyses (Tukey’s HSD; alpha=.05) revealed that the differences
were between the scores of the SLI and TD6 groups, and between those
of the TD4 and TD6 groups. There was no statistically significant
difference between the mean scores of the SLI and TD4 groups.
However, on some of the items, there were differences between the

responses of these two groups, which warrant discussion. The first of
these items was item 2, foten. A similar number of participants in the SLI
and TD4 groups (four and five, respectively) gave the correct answer.
However, as a group, the participants with SLI favoured the singular
form (nine of them said *foten), whereas five of the TD4 participants
gave this response. None of the SLI participants gave the singular form
of a real word, but two 4-year-olds did: Both said *foto ‘photograph’. A
similar (but not identical) pattern was observed for the groups’ responses
to waken (which is phonologically similar to foten): Six participants in the
SLI and eight in the TD6 group gave the correct response, wakens. Again,
nine participants with SLI produced the singular form, compared to six
in the TD4 group.
None of the participants with SLI gave the correct response (fante) to
item 26, fant; four 4-year-olds did. Five participants in each group gave
the singular form. Three participants with SLI said fants, whereas only
one 4-year-old gave this response. The participants’ poor response to
this item is somewhat surprising: Because olifant ‘elephant’ is a word
commonly known to young Afrikaans-speaking children, it was expected
that at least the 4-year-olds would say fante, on analogy to olifante.
148
An item to which no correct response was given by any participant in the

two groups was lerg [lÏrx], item 35. The correct response would have
been [lÏrg«]. The response given most often by both groups was the
singular form. Four participants with SLI gave unusual irregular forms
(one girl said [lÏrxt«] and one [lÏr«], and one boy said [lÏrk«] and one
[lÏrts]), whereas no 4-year-old did so. A similar pattern was observed for
the responses to item 38, [stÏrx], of which the correct plural form would
be [stÏrg«]. No participant gave this correct response; the groups’
preference was again for the singular form (eight participants with SLI
and seven 4-year-olds), and again some children in the SLI group (three
girls) but none in the TD4 group gave an unusual irregular form:
[stÏrd«], [stÏr«] which is a real word meaning “stars”, and [stÏrt«],
which is a real word meaning “tails”.
5.5. NUMBER PRODUCTION IN THE LANGUAGE

SAMPLES
The use of singular and plural nouns in the first 100 complete and fully
intelligible utterances of each language sample (i.e., those utterances used
to calculate the MLU) was examined. In total, the 45 participants used
3743 nouns in these 100 utterances – 18.8% of all words used were
nouns: The SLI group used 1116 (17.1% of their 6521 words were
nouns), the TD4 group 1150 (which amounted to 16.8%), and the TD6
group 1477 (16.6%). The difference in variance between the three groups
was statistically significant (Levene’s test, F2,42=5.164; p=.010). Of the
3743 nouns, 3043 (81.3%) were supposed to be in their singular form
and 700 (18.7%) in their plural form. Table 5.22 contains a per-group
summary of the use of nouns in the language samples.
The proportion of plural forms was calculated per participant in order to

analyse the differences between the three groups. Of the three groups,
the SLI group used the highest proportion of plural forms (.21),
followed by the TD4 (.19) and TD6 groups (.15). A one-way ANOVA
returned a significant outcome, which means that a difference between
the groups could be assumed (F2,42=4.525; p=.017). Subsequent post hoc
comparisons (Tukey’s HSD; alpha=.05) showed that the TD6 and SLI
149
groups differed from each other, and that the TD4 group differed from
neither of the other two groups.
Table 5.22. Number of singular and plural nouns used (correctly and
incorrectly) in the language samples
SLI TD4 TD6
Singular form correct 880/881 972/974 1188/1188
99.9% 99.8% 100%
Singular form incorrect 1/881 2/974 0/1188
0.1% 0. 2% 0%
Plural form correct
-e 62/69 60/61 109/109
89.9% 98.4% 100%
–s 71/78 77/79 119/119
91.0% 97.5% 100%
Irregular 74/88 33/36 59/61
84.1% 91.7% 96.7%
Total 207/235 170/176 287/289
88.1% 96.6% 99.3%
Plural form incorrect
-e 7/69 1/61 0/109
10.1% 1.6% 0%
–s 7/78 2/79 0/119
9.0% 2.5% 0%
Irregular 14/88 3/36 2/61
15.9% 8.3% 3.3%
Total 28/235 6/176 2/289
11.9% 3.4% 0.7%
None of the groups showed a strong preference for –s forms over –e

forms or vice versa. There was a slight tendency for some children in the
SLI group to replace –e forms with –s ones (as in byls instead of byle) but
not vice versa. Surprisingly, irregular forms had a higher frequency of
occurrence in the language samples of the SLI group than in those of the
two typically developing groups. Also, a higher proportion of the SLI
group’s plural forms consisted of irregular ones (.37). The frequency of
occurrence of irregular forms was the lowest in the TD6 group’s
language samples; however, the proportion of irregular forms was similar
for the two typically developing groups: .19 for the TD4 group and .21
for the TD6 one. As expected, the group with SLI made more errors on
150
these irregular forms than did either of the other two groups. This was
also the case for errors on –e and –s forms.
Some types of errors were made by both children with SLI and typically
developing ones. These include devoicing the final consonant of the
stem in the plural form (which is how the stem is pronounced in its
singular form). This occurred in the case of *hoete (singular: hoed ‘hat’). It
also occurred on *goete ‘things/stuff’ (singular: ding ‘thing’), but voicing
the final consonant of the stem (as in *goede) would still not have
rendered a grammatical plural form. Either goed, goeters, or dinge would be
the grammatical form. Another error made by typically developing
children and those with SLI was using the regular plural suffix –e instead
of –te when producing the plural form of kas ‘cupboard’. In two cases, a
plural instead of a singular form was used: A 4-year-old boy said hier ’n
*grafies instead of hier is ’n grafie ‘Here is a spade-DIM’ and a boy with SLI
said dan hy dan wil dit hom *koppe sit hy instead of dan wil hy dit op sy kop sit
hy ‘Then he wants to put it on his head he’.
However, some error types were only made by the children with SLI.
These include:
(i) using the regular –s instead of the –e, as in *byls instead of byle
‘axes’ and *honds instead of honde ‘dogs’;
(ii) using both regular suffixes when only the –e should have occurred,
as in *handes instead of hande ‘hands’ and *koppes instead of koppe
‘heads’;133,134
(iii) idiosyncratic devices for forming plurals. When asked whether he
has any siblings, one boy with SLI replied net ’n sussie en twee sussies
‘just a sister-DIM and two sister-DIM-PL’. He had, in fact, (only)
two sisters, so it is assumed that net ’n sussie en twee sussies was his
133 For most nouns, the singular form is the most frequently used and thus the most
frequently heard one. There are some exceptions though: Oë ‘eyes’, hare ‘hair-PL’ or
hande ‘hands’, for instance, are likely to be heard more often than their singular
counterparts. In the case of *handes then, it could well be that the most frequently heard
form of the noun is pluralised, which would render a correct plural form for most other
nouns.
134 Here it is interesting to note that, in Dutch, language change has resulted in plural
forms such as eier and kinder being “repluralised”, resulting in a “double” plural (eieren
and kinderen) which is now regarded as the grammatical plural form. The “double
plurals” produced by the Afrikaans-speaking children with SLI thus have correlates in
other, “non-impaired” languages.
151
formulation for net twee sussies ‘only two sister-DIM-PL’. This same
boy said groot hond en twee honde ‘big dog and two dogs’ when asked
whether he had any dogs; from the context, it emerged that he had
two dogs of which one was big and one small. It appears then that
this boy makes use of a type of parataxis to indicate plurality.
Only the singular and plural forms occurring in the first 100 utterances
were tallied. However, the rest of the 30 minutes of each language
sample was also examined for errors involving plurality. The errors made
by both the typically developing children and those with SLI − those
discussed above − all occurred in the rest of the 30 minutes as well. In
addition, the following ones were made:
(i) Three children with SLI used the regular –e instead of the –s: *arme
instead of arms ‘arms’ and *leeue instead of leeus ‘lions’.
(ii) One boy with SLI said lammetjies se *sterde which should have been
lammetjies se sterte ‘lamb-DIM-PL possessive marker tails’. This boy also
said *lammetjiesterts instead of lammetjiesterte ‘lamb-DIM tails’.
(iii) Three typically developing children used the incorrect form of
what should have been an irregular plural form: A 4-year-old boy
said *glasse instead of glase ‘glasses’ (singular: glas); one typically
developing 6-year-old boy said *Saterdags instead of Saterdae
‘Saturdays’ (singular: Saterdag); another said *diewens instead of diewe
‘thieves’ (singular: dief).
5.6. DISCUSSION: NUMBER COMPREHENSION AND

PRODUCTION
Where performance between groups differed on the experimental tasks –

and only for the judgement tasks assessing irregular forms of nonsense
words and grammatical irregular forms of real words, there was no
statistically significant difference between groups – the pattern which
emerged was that the typically developing Afrikaans-speaking 6-year-olds
fared better than both the 6-year-olds with SLI and the typically
developing 4-year-olds; the latter two groups performed similarly.135 This
pattern was partly observed for the language sample results: The typically
135Overall, all three groups of children performed worse on the task involving the
production of the plural form of nonsense words than they did on those involving real
words. This was also found by Dalalakis (1999).
152
developing 6-year-olds outperformed their peers with SLI, but the

performance of the 4-year-olds did not differ from either of the groups
of 6-year-olds.
As was the case in the present study, Dromi et al. (1993:766) found that
Hebrew-speaking children with SLI used significantly fewer plural forms
than did age-matched controls in elicited production. The performance
in this regard of the children with SLI did not differ significantly from
that of (younger) MLU-matched controls. Similarly, Leonard et al.
(2001:629) found that Swedish-speaking children with SLI (aged 4 years
3 months to 5 years 7 months) fared significantly worse than age-
matched controls in terms of elicited production of plural forms, with no
significant differences between the production of plural forms by the
children with SLI and MLU-matched controls.
Similar results were obtained by Conti-Ramsden and Hesketh (2003),

who found no significant difference between the performance of two
groups of children on a plural elicitation task, namely English-speaking
children with SLI (aged 4 years 4 months to 5 years 10 months) and 2- to
3-year-old controls (matched for language age with the children with
SLI). Furthermore, Conti-Ramsden (2003:1032) found that these
children with SLI fared significantly worse than age-matched controls on
this task.
The results of the present study also offer support for those of Dalalakis
(1999), who found that Greek-speaking children with SLI (ages 5 years 8
months to 17 years 7 months) performed significantly worse than age-
matched controls, and that there was no significant difference between
the performance of these controls and language-matched ones; in other
words, in contrast to the findings of the present study, the participants
with SLI also performed significantly worse than their language-matched
controls in elicited production of plural forms. The reason for this
discrepancy in findings could be related to plural forms in Greek being
morphologically more complex than they are in Afrikaans, as nouns are
also marked for gender and class in Greek, but not in Afrikaans.
Several other studies also found that children with SLI performed
similarly to younger (mostly MLU-matched) controls. One example is
the 1990 study by Rom and Leonard. They compared the language
153
samples of 4- to 5-year-old Hebrew-speaking children with SLI to those

of younger typically developing children matched on a morpheme-per-
utterance measure. They also found no difference in the correct use of
plural nouns between the two groups of children.
The accuracy of plural production by the children with SLI in their

spontaneous language samples was 88% in the present study; that of the
younger controls was 97%. This compares well with the 5-year-old
children with SLI in a study by Oetting and Rice (1993): These children
produced plurals at 90% accuracy (as did MLU-matched controls).
However, this was not the case in their elicited production data. Here the
children with SLI experienced more difficulty than the controls with the
plural form of infrequently pluralised nouns. Similarly, Rice and Oetting
(1993) found that 4- to 6-year-olds with SLI demonstrated relatively high
levels of accuracy in the pluralisation of nouns. Similar to what was
found in the present study, there was no significant difference between
the performance of children with SLI and younger children matched for
language age, in terms of their elicited production of regular and irregular
plural forms. The children with SLI had a mean chronological age of 5
years 0 months and a mean language age of 3 years 0 months, whereas
the control group’s mean age was 2 years 10 months.
In contrast to the present study, others found no difference between the

three groups: children with SLI, age-matched controls, and younger
typically developing children. One example is Blake, Myszczyszyn, and
Jokel (2004), who found no significant difference between the correct
spontaneous use of plural forms of nouns by children with SLI (aged 5
years 1 month to 9 years 8 months) and that of age-matched controls
and controls matched in terms of expressive language score. Roberts and
Rescorla (1995) compared the spontaneous production of noun
morphology of 4-year-olds with an expressive language delay to that of
age-matched and MLU-matched controls. They found that the three
groups of children did not differ in terms of their production of plurals.
Lastly, Rom and Leonard (1990) compared the language samples of 4- to
5-year-old Hebrew-speaking children with SLI to those of younger
typically developing children matched on a morpheme-per-utterance
measure. They found no difference in the correct use of plural nouns
between the two groups of children.
154
There is consensus in the literature that children with SLI perform on a

par with younger typically developing controls; this was also found in the
present study, but only for the elicited data. However, the findings on
comparisons with typically developing age-matched controls are mixed:
Some studies concluded that children with SLI fare worse (as was the
case in the present study) and others not.
Turning to the research questions posed in chapter 1 (specifically

questions 1 to 4), it appears that Afrikaans-speaking children with SLI
present with a delay in their comprehension and production of
grammatical morphemes related to the singular/plural distinction.
Although the responses of the SLI and 4-year-old groups to certain items
of the experimental tasks differed, no pattern could be detected in these
differences. Therefore, based on the results of the experimental tasks
alone, one would conclude that the language of Afrikaans-speaking
children with SLI is merely delayed. However, analyses of the
spontaneous language samples of these two groups revealed that the SLI
group made even more errors than did the 4-year-olds, but also that the
SLI groups made errors not found in the language of the typically
developing children. Therefore, the Afrikaans-speaking children with SLI
do not present with merely a language delay; their language is also to
some extent deviant.
155
Results: Person and case
Chapter 6
Comprehension and production of person and case
6.1. INTRODUCTION
The comprehension of person and case on personal and possessive

pronouns was tested with two different experimental tasks, namely
picture selection – the results of which are given in section 6.2.1. – and a
judgement task – the results of which are presented in section 6.2.2. In
section 6.2.3, the results of the production task eliciting person and case
on pronouns are presented. Section 6.3 contains the results of the
comprehension (section 6.3.1) and elicited production tasks (section
6.3.2) involving se-constructions. Thereafter, the way in which pronouns
(section 6.4.1) and se-constructions (section 6.4.2) were used by the
participants in their language samples will be discussed. Section 6.5
compares the results of this study to those of others on pronouns and
possessive constructions.
6.2. PERSON AND CASE ON PRONOUNS
6.2.1. Results: Picture selection task: Person and case on

pronouns
This task aimed to establish how Afrikaans-speaking children with SLI

fare in comparison with typically developing ones in terms of the correct
identification of pronouns. The 32 items of this task are given in section
2.1 of appendix D. Each response given to these items was classified as
either correct or incorrect. Initially, a more complex classification system
was used – one which made it possible to detect error patterns, i.e., to
establish which grammatical (number or person) and/or semantic
(gender) feature the children found difficult – but most errors were made
very infrequently, and no error patterns could be detected. For this
reason, the coding system was simplified.
157
The reliability of the 32 items was high; Cronbach’s alpha was .830. The
statistics of the individual items are presented in table 6.1, ordered from
the item which the participants found the most difficult to that which
they found the easiest.
Table 6.1. Item statistics – Picture selection task: Pronouns

I- Item English translation Proportion Corrected
tem of partici- item-total
no. pants giving correla-
correct tion
answer
10 Sy oor is af It’s ear is off .11 .203
9 Die seuntjie sien hulle The boy sees them .38 .108
22 Ouma vryf dit Grandma is stroking it .42 .369
6 Dit is ons speelgoed These are our toys .49 .117
12 Hulle wys ’n prentjie They are showing a .53
.200
picture
29 Die seuntjie sien julle The boy sees you-PL .60 .161
16 Dit is julle speelgoed These are your-PL toys .62 .347
4 Dit is hulle speelgoed These are their toys .64 .322
2 Julle wys ’n prentjie You-PL are showing a .67
-.103
picture
17 Hy sit by die tafel He is sitting at the .71
.319
table
28 Die koerant lê The newspaper is lying .71
.503
daarop on it
14 Die kat lek haar The cat is licking her .73 .297
18 Ons wys ’n prentjie We are showing a .76
.338
picture
21 Dit lê op die grond It is lying on the .76
.478
ground
1 Die reën val op julle The rain is falling on .78
.384
you-PL
17 Die kat lek my The cat is licking me .78 .666
31 Die reën val op hulle The rain is falling on .78
.563
them
30 Ek staan langs die I am standing next to .80
.254
tafel the table
7 Die seuntjie sien ons The boy sees us .82 .471
19 Die voël sit op haar The bird is sitting on .82
.459
her
25 Die kat lek hom The cat is licking him .82 .495
158
23 Sy sit by die tafel She is sitting at the .82

.435
table
26 Die reën val op ons The rain is falling on us .84 .177
32 Die voël sit op my The bird is sitting on .84
.513
me
5 Dit is sy roomys It is his ice-cream .84 .399
11 Die voël sit op jou The bird is sitting on .84
.450
you-SGL
3 Dit is haar roomys It is her ice-cream .87 .430
20 Die kat krap jou The cat is scratching .89
.485
you-SGL
8 Dit is my roomys It is my ice-cream .91 .336
15 Dit is jou roomys It is your-SGL ice- .91
.135
cream
13 Die voël sit op hom The bird is sitting on .93
.267
him
24 Jy staan op die stoel You-SGL are standing .93
.445
on the chair
The response to item 2 (Julle wys ’n prentjie ‘You-PL show a picture’) had a
negative correlation to the 32 items as a whole. This indicated that this
item somehow did not test what it was supposed to. Whereas the item-
total correlation of the other three items involving julle – viz. items 1 (Die
reën val op julle ‘The rain is falling on you-PL’), 16 (Dit is julle speelgoed
‘These are your-PL toys’), and 29 (Die seuntjie sien julle ‘The boy sees you-
PL’) – was low, it was not negative. Furthermore, the four items involving
julle were not particularly difficult for the participants: 60 to 78% of them
gave the correct response. Therefore, all items involving julle were
included in the statistical analyses, also the one with a negative item-total
correlation.
The performance of the three groups of participants on the 32 items is

portrayed in figure 6.1. From this figure, it appears that the TD6 group
fared better than the other two. The maximum scores of the SLI and
TD4 groups seemed similar; however, their minimum scores differed
greatly (the lowest appeared in the SLI group, as expected) and there was
more variability in the SLI group than in the other two groups. Levene’s
statistic of homogeneity of variance confirmed this: F2,42=8.036; p=.001.
159
35
30
Score
25
20
15
10
SLI TD4 TD6
Group
Figure 6.1. Box plot of performance per group − Picture selection task:
Pronouns
The performance of the three groups of participants are summarised in

Table 6.2. The mean scores of the SLI and TD4 groups were similar
(20.07 and 22.73 out of 28, respectively), whereas the mean of the TD6
group (27.33) was higher than that of the other two. A one-way
ANOVA returned a significant outcome, which means that a difference
between the groups could be assumed (F2,42=10.599; p=.000). Post hoc
analyses (Tukey’s HSD; alpha=.05) revealed that this difference was
groups.
Table 6.2. Summary of performance per group – Picture selection task:

Pronouns
SLI 15 20.07 6.19293 10 29
TD4 15 22.73 3.01109 18 28
TD6 15 27.33 3.15474 20 31
Total 45 23.38 5.24096 10 31
No statistically significant difference was found between the SLI and

TD4 groups. In order to ascertain whether there were differences in the
types of responses given by these two groups, their responses to each of
the 32 items were examined. There were two items for which there were
160
notable differences between the responses of these two groups. The first
was item 12 (Hulle wys ’n prentjie ‘They are showing a picture’): The SLI
group obtained a higher score (eight participants selected the correct
picture as opposed to only four in the TD4 group). The error made most
frequently by both groups (eight participants in the TD4 and four in the
SLI group gave this response) was selecting the phonologically similar
Julle wys ’n prentjie ‘You-PL are showing a picture’. No noticeable
difference was seen in the other items entailing hulle: Dit is hulle speelgoed
‘These are their toys’ (item 4), Die seuntjie sien hulle ‘The boy sees them’
(item 9), and Die reën val op hulle ‘The rain is falling on them’ (item 31).
The second item for which the responses of the two groups differed
noticeably was item 28 (Die koerant lê daarop ‘The newspaper is lying on
it’). Here, the SLI group again outperformed the TD4 group: 11 vs. six
correct responses. Again, the groups had the same most-given incorrect
response: Three participants in the SLI group and six in the TD4 one
chose Die koerant lê op jou ‘The newspaper is lying on you-SGL’. As was
the case for hulle, no clear difference could be found in the responses of
the two groups on the other items containing the third person singular
neuter pronoun. To Sy oor is af ‘His ear is off’ (item 10, where sy refers to
a mug), only one participant in each group gave the correct response; the
types of incorrect responses were similarly distributed between the
groups. To Dit lê op die grond ‘It is lying on the ground’ (item 21) and
Ouma vryf dit ‘Grandma is stroking it’ (item 22), the groups also
responded in similar ways.
6.2.2. Results: Judgement task: Person and case on pronouns
Sentences with correct and others with incorrect pronominal forms were
presented to participants in order to assess whether they could identify
correct vs. incorrect use of such forms. Participants’ responses to the 32
items of this judgement task received one of the following two codes
(items themselves and their English translations are given in section 2.2
of appendix D):
(i) A response was taken to be correct if a grammatical form was
judged as such or an ungrammatical one judged as such.
(ii) A response was regarded as incorrect if a grammatical form was
judged to be ungrammatical or an ungrammatical one to be
grammatical.
161
Initially, four extra codes were included. However, each of these four
codes were used only once; therefore, the relevant four responses were
subsequently recoded as either correct or incorrect, simplifying the
coding system. Participants were not asked to give any reason for their
judgements. However, when they spontaneously did provide such
reasons, this was not discouraged.
The reliability for the 32 items as a group was not very high: Cronbach’s
alpha was .771. Individual item statistics are given in Table 6.3; items are
ordered from the most difficult to the easiest. As can be seen from this
table, two items − item 7 (Die koppies staan bo dit ‘The cups are standing
above it’)136 and 26 (Sy hare is af ‘Its hair is gone’) − had a negative
correlation to the items as a whole. Of these, one (item 7, the one with
the lower mean score) was removed. Cronbach’s alpha for the remaining
31 items was then .794.
Figure 6.2 depicts the performance of the three groups: The TD6 group
again seemed to perform better than the other two. However, there was
also more variability is this group than in the other two: Some of the
participants obtained a perfect score whereas others performed worse
than some of the 4-year-olds. The difference in the variability in the
three groups was, however, not significant (Levene’s statistic of
homogeneity of variance, F2,42=637; p=.534).
136 Interestingly, most errors on this item were made by the typically developing 6-year-
olds. The picture was one of a broom cupboard. In the cupboard, a broom stood on
the bottom shelf and two cups on the top shelf. The researcher said Die koppies staan bo
dit ‘The cups are standing above this/it’ with emphasis on dit, while pointing to the
broom. However, Die koppie staan bo ‘The cups are on top’ would also have been a
grammatical description of the picture. It could be that the typically developing 6-year-
olds thought that the dit was a type of trick word here (seeing that the absence of dit
also renders a grammatical sentence) and therefore incorrectly judged Die koppies staan
bo dit as being ungrammatical.
162
Table 6.3. Item statistics – Judgement task: Pronouns

the correct answer correlation
20 Dit is *hulle tuin (julle) .18 .133
13 Die skaap jaag *ons (hulle) .27 .193
18 Dit is *hulle kar (ons) .27 .537
29 *Jou slaap (jy) .29 .612
30 Die vark kyk na *sy (haar) .29 .549
24 Die hond krap *ek (my) .31 .575
31 *Hy lees boeke (julle) .31 .453
10 Dit is *haar roomyse (julle) .36 .406
3 *Sy is wakker (hy) .40 .563
12 Die skoenlapper sit op *haar .42
.172
(jou)
17 Ouma hou *haar vas (dit) .42 .504
15 Die hond spring oor *julle .49
.487
(ons)
5 Dit is *sy hoed (my) .51 .380
2 Dit is *hom hoed (sy) .51 .250
22 Die perd dra *jou (haar) .53 .420
4 *Sy staan in die hoek (dit) .60 .235
7 Die koppies staan bo dit .71 -.271
9 Die hond spring oor hulle .80 .216
6 Die koei jaag julle .84 .241
28 Die hond spring oor julle .84 .042
26 Sy hare is af .87 -.057
32 Die roomys val op my .89 .070
11 Sy staan .91 .226
8 Die krap knyp hom .91 .208
1 Hulle lees boeke .93 .243
16 Die eend swem langs hom .93 .038
14 Ons lees boeke .96 .177
19 Dit is haar roomys .96 .078
21 Die hond krap jou .96 .202
23 Ek is wakker .96 .128
27 Die koei jaag ons .96 .078
25 Dit is jou roomys .98 .123
163
32
30
28
26
24
Score
22
20
18
16
14
SLI TD4 TD6
Group

Pronouns
Table 6.4 summarises the details of the performance of the three groups.
The mean score of the SLI and TD4 groups was almost identical, as
were their maximum and minimum scores. A one-way ANOVA
returned a significant outcome, indicating that a difference between the
groups could be assumed (F2,42=20.455; p=.000). Post hoc analyses
(Tukey’s HSD; alpha=.05) revealed that the significant differences were
between the SLI and TD6 groups, and between the TD4 and the TD6
groups.
Table 6.4. Mean number of correct responses to each item – Judgement

task: Pronouns
Group N Mean Standard Minimum score Maximum
deviation obtained score obtained
SLI 15 17.73 3.28344 15 25
TD4 15 17.47 2.79966 14 24
TD6 15 24.33 3.82971 17 31
Total 44 19.84 4.57243 14 31
There was no statistically significant difference in the mean scores of the

SLI and TD4 groups. Differences in their responses could be found for
only two items, both of them ungrammatical ones. The first was *Dit is
sy hoed ‘This is his hat’ (item 5, which should have been Dit is my hoed
‘This is my hat’). Here, eight children in the TD4 group gave the correct
164
response compared to four in the SLI group. All other participants in

both groups said that *Dit is sy hoed was correct, without providing
reasons for their judgements.
A similar response pattern occurred for *Die perd dra jou ‘The horse is
carrying you’ (item 22, which should have been Die perd dra haar ‘The
horse is carrying her’). Nine participants in the TD4 group but only five
in the SLI group gave the correct response. Again, all participants who
said that *Die perd dra jou is correct refrained from providing a reason for
their judgement.
6.2.3. Results: Sentence completion task: Person and case on

pronouns
This task was performed in order to ascertain whether the elicited

production of pronouns by Afrikaans-speaking children with and
without SLI differed. The 64 items of this task (with their English
translations) are listed in section 2.3 of appendix D. Initially, when
coding responses, 24 different codes were used in such a way that it was
possible to detect any patterns in the errors made by the children.
However, most of the error categories occurred very infrequently, and
no pattern could be detected in the errors made by the three groups. For
these reasons, the coding system was simplified, making a distinction
only between correct and incorrect responses.
The reliability of the items was high: Cronbach’s alpha was .941.
Individual item statistics are provided in table 6.5, where items are
ordered from the most difficult to the easiest. The items had a wide
range of difficulty: The proportion of participants giving the correct
answer ranged from .00 to .98. In general, the items requiring the
production of julle and dit/sy/daar- were the most difficult ones and those
requiring the production of ek/my the easiest.
165
Table 6.5. Item statistics – Production task: Pronouns

I- Item Proportion of Corrected
tem participants item-total
no. giving correct correla-
answer tion
11 Die hond lek my en die seuntjie lek (dit) .00 .000
52 Dié apie spring oor my en dié apie spring (daarin .00
.000
/in dit)
13 Die baba kyk vir my en die hond kyk (daarvoor .02
.256
/daarna /vir dit)
30 Ek is skoon, maar (dit is vuil) .02 .111
40 Dié kat krap my en dié kat krap (dit) .02 -.086
45 Jy is groot, maar (dit is klein) .02 .032
8 Dit is haar bene en dit is (julle bene) .07 .410
57 Dit is haar tande maar dit is (sy tande) .13 .213
46 Dit is sy hond en dit is (julle honde) .16 .537
61 Dié seun spuit hom nat en dié seun spuit (julle) .16
.603
nat
28 Die vark lek hom en die honde lek (julle) .18 .586
62 Hy sien ’n skaap, maar (julle sien ’n koei) .20 .665
9 Die skoenlapper sit op jou en die skoenlappers sit .20
.410
(op julle)
22 Dit is haar nek en dit is (sy nek) .22 .293
53 Dié bal hop op jou en dié bal hop (op julle) .24 .496
18 Sy eet appels, maar (julle eet piesangs) .31 .547
1 Die hond lek haar en die katte krap (hulle) .42 .317
17 Die hond sit langs my en die kat sit (langs haar) .51 .617
49 Sy spring tou, maar (ons praat op die foon) .51 .447
26 Sy drink water, maar (ons eet vrugte) .53 .478
55 Dié seun stamp haar en dié seun stamp (hulle) .53 .442
20 Die voël sit op my en die voëls sit (op hulle) .56 .519
23 Die hond lek my en die kat krap (haar) .56 .426
38 Dit is my glas en dit is (sy glas) .56 .257
5 Dit is haar swembroek en dit is (sy trui) .58 .390
14 Die hond lek my en die baba lek (hom) .58 .321
31 Die kat krap hom en die honde lek (ons) .60 .506
15 Die baba kyk vir my en die hond kyk (vir hom) .62 .415
42 Dié perd sien my en dié perd sien (haar) .62 .527
58 Dié vark sien hom en dié vark sien (ons) .64 .587
166
29 Dié hond krap hom, maar dié hond lek (jou) .71 .510
54 Hy voer die eekhoring, maar (sy voer die voëls) .71 .435
56 Dit is sy boek en dit is (hulle boeke) .71 .492
47 Die hond kyk vir jou en die perd kyk (vir hom) .71 .572
50 Dié slang seil oor my en dié slang seil (oor haar) .71 .546
27 Dit is hulle musse en dit is (ons skoene) .73 .581
34 Dit is haar tasse en dit is (ons tasse) .73 .659
6 Dié hond sit langs hom, maar dié hond sit (langs .76
.612
jou)
10 Dit is sy lepel en dit is (jou mes) .76 .346
12 Dit is ons hare en dit is (hulle hare) .76 .485
36 Dié bul skop haar en dié bul skop (hom) .76 .481
44 Jy staan op die tafel, maar (hy staan op die stoel) .76 .499
2 Die voël vlieg oor hulle en die vlieër vlieg (oor ons) .78 .662
7 Hy eet pizza, maar (sy eet ’n stokkielekker) .78 .569
32 Dit is my hand en dit is (haar hand) .78 .511
33 Dié hoender pik hom, maar dié hoender pik (jou) .78 .531
37 Dit is sy kar en dit is (haar bal) .80 .602
51 Die voël sit op my en die haas sit (op my) .80 .441
39 Dié emmer val op my en dié emmer val (op hulle) .82 .525
60 Die vliegtuig vlieg oor hom en die helikopter vlieg .82
.567
(oor ons)
16 Hy staan, maar (jy sit) .84 .444
59 Dit is sy oë en dit is (jou hare) .84 .384
19 Die eekhoring sit langs my en die hond sit (langs .87
.474
my)
21 Sy hang in die lug, maar (hy sit op die grond) .87 .271
35 Ons staan, maar (hulle sit) .87 .410
48 Hy lê, maar (jy staan) .89 .363
63 Die koei jaag my en die perd jaag (my) .89 .463
24 Dié kat krap my en dié kat krap (my) .93 .311
3 Hy eet ’n roomys, maar (ek eet ‘n appel) .93 .477
4 Ons eet koek, maar (hulle eet roomys) .93 .178
41 Hy eet waatlemoen, maar (ek eet koek) .93 .139
43 Dit is my voet en dit is (my voet) .93 .414
64 Dié perd spring oor hom, maar dié perd spring (oor .93
.139
jou)
25 Dit is my romp en dit is (my broek/romp) .98 .060
167
Six problematic items were identified; all six entailed the production of
dit or the phonetically altered form thereof (daar-). The first was Die hond
lek my en die seuntjie lek (dit) ‘The dog is licking me and the boy is licking
(it)’ (item 11). Not one participant gave the targeted response. The same
was noted for Dié apie spring oor my en dié apie spring (daarin /in dit) ‘This
monkey is jumping over me and this monkey is jumping over (it)’ (item
52): Not one correct response was given, with the consequence that the
correlation to the rest of the items was .000.
To the following three items, only one of the 45 participants gave the
correct response, and the correlation of each of these items to the rest
was very low: Ek is skoon, maar (dit is vuil) ‘I am clean, but (it is dirty)
(item 30); Dié kat krap my en dié kat krap (dit) ‘This dog is scratching me
and this dog is scratching (it)’ (item 40); and Jy is groot, maar (dit is klein)
‘You are big, but (it is small) (item 45). Die baba kyk vir my en die hond kyk
(daarvoor /daarna /vir dit) ‘The baby is looking at me and the dog is
looking at (it)’ (item 13) also rendered only one correct response, but its
correlation to the scale was slightly better (although still poor).
These six items were removed before further analyses were done.
However, because they all entailed the production of dit or daar-, it was
decided to remove the other two items of the dit/daar- set as well. The
poor response of the participants to these six items, compared to their
responses to the other items of the task, was taken to be an indication
that the production of dit/daar- cannot be elicited successfully by using
the procedure in question – rather than that the children could not
produce dit/daar- forms. Also, although their responses to the remaining
two dit/daar- items were better, these responses were still poor. These
two items were Dit is haar nek en dit is (sy nek) ‘This is her neck and this is
(its neck)’ (item 22, 10 correct responses) and Dit is haar tande maar dit is
(sy tande) ‘These are her teeth and these are (its teeth)’ (item 57, six
correct responses). The removal of these eight items made hardly any
difference to Cronbach’s alpha, which was then .942.
The performance of the three groups on the remaining 56 items is

depicted in figure 6.3; that of the TD6 group appeared to be the highest
of the three. The other two groups seemed to fare similarly and to have
the same median. Variability appeared higher in the SLI group than in
the TD4 group; however, Levene’s statistic of homogeneity of variance
168
indicated that the difference in variability between the groups was not
significant (F2,42=1.414; p=.255).
60
50
40
Score
30
20
10
21
0
SLI TD4 TD6
Group
Figure 6.3. Box plot of performance per group − Production task:

Pronouns
Table 6.6 summarises the performance of the three groups. This table
also shows that the mean scores of the SLI and TD4 groups were
comparable (32.53 and 33.27 out of 56, respectively), but that more
variability occurred in the SLI group: The minimum score of the SLI
group was lower and the maximum score higher than that of the TD4
group. Comparing the minimum and maximum scores of the SLI and
TD6 groups, one sees that the highest scoring participant in the SLI
group fared almost as well as that of the TD6 group, but that the score
of the worst performing participant in the SLI group was far lower than
that of the TD6 group.

Pronouns
SLI 15 32.53 12.60310 4 52
TD4 15 33.27 7.60138 16 42
TD6 15 45.60 8.66685 29 56
Total 45 37.13 11.38300 4 56
169
A one-way ANOVA was carried out; it returned a significant outcome,

which means that a difference between the groups could be assumed
(F2,42=8.314; p=.001). Post hoc analyses (Tukey’s HSD; alpha=.05)
revealed that this difference was not between the SLI and TD4 groups.
It was the scores of the SLI and TD6 groups, and those of the TD4 and
TD6 groups, which differed.
Despite there being no statistically significant difference between the SLI

and TD4 groups, the types of responses they gave differed for nine
items. These differences are discussed below.
To item 5, Dit is haar swembroek en dit is (sy trui) ‘This is her swimming
costume and this is (his jersey)’, the majority (eight) of the participants in
the TD4 group gave the correct answer, whereas only five participants
with SLI did. The incorrect response given most by both groups was
*hom trui, where the pronoun has the incorrect case but is otherwise
correct. On item 38, Dit is my glas en dit is (sy glas) ‘This is my glass and
this is (his glass)’, the 4-year-olds also outperformed the participants with
SLI: Six participants with SLI gave the correct response, whereas eight 4-
year-olds did. The incorrect response most frequently given by both
groups was *hom glas, where the case of the pronoun in incorrect.
Item 10, Dit is sy lepel en dit is (jou mes) ‘This is his spoon and this is (your-
SGL knife)’, also proved to be significantly more difficult for the SLI
group than for the TD4 group. Only one participant in the TD4 group
did not give the correct response (she said sy mes ‘his knife’), whereas
eight participants with SLI gave the incorrect answer. Their answers
included sy mes ‘his knife’ (two participants), haar mes ‘her knife’ (four),
hom mes ‘him knife’ (one) and one unusable response. Item 29, Dié hond
krap hom, maar dié hond lek (jou) ‘This dog is scratching him, but this dog is
licking (you)’, was also more difficult for the SLI group than for the TD4
one: The majority (eight) of the participants with SLI gave the incorrect
response whereas the majority of the 4-year-olds (12) gave the correct
response.
To item 23, Die hond lek my en die kat krap (haar) ‘The dog is licking me
and the cat is scratching (her)’, nine of the participants in the TD4 group
and only six in the SLI group gave the correct response. The incorrect
response given most frequently (three times) by the SLI group was jou
170
‘you-SGL’, whereas that given most frequently by the TD4 group (also
three times) was to replace the pronoun with a DP.
A reversed pattern was observed for item 17, Die hond sit langs my en die
kat sit (langs haar) ‘The dog is sitting next to me and the cat is sitting (next
to her)’: The SLI group outperformed the TD4 one, with the majority
(eight) of the participants in the SLI group giving the correct response.
No single incorrect response was favoured by the SLI group, but the
TD4 had a preference for replacing haar with a DP – six of the TD4
group members gave this response.
This pattern was also observed for item 31, Die kat krap hom en die honde
lek (ons) ‘The cat is scratching him and the dogs are licking (us)’: The
majority (10) of the participants with SLI gave the correct response
whereas the majority of the 4-year-olds (also 10) gave an incorrect one.
Most of the 4-year-olds (nine) said my en jou ‘me and you-SGL’; only one
said my ‘me’. The same occurred on item 49, Sy spring tou, maar (ons praat
op die foon) ‘She is skipping, but (we are talking on the phone)’: Nine
participants with SLI but only three 4-year-olds gave the correct
response. Most of the 4-year-olds (10) replaced ons ‘we’ with ek en jy ‘I
and you-SGL’. The majority of the participants in both groups gave an
incorrect response to item 26, Sy drink water, maar (ons eet vrugte) ‘She is
drinking water by (we are eating fruit)’. The incorrect response most
commonly given by both groups was to replace the ons with ek en jy ‘I
and you-SGL’; however, this response was given twice as frequently by
the 4-year-olds than by the participants with SLI (10 vs. five times).
On item 55, Dié seun stamp haar en dié seun stamp (hulle) ‘This boy is pushing
her and this boy is pushing (them)’, the SLI group fared better than the
TD4 group: Eight participants with SLI and five 4-year-olds gave the
correct response. The most common incorrect response by both groups
was replacing the pronoun with a phrase such as die oom en die tannie ‘the
uncle and the auntie=the man and the woman’.
Item 9 was Dié skoenlapper sit op jou en dié skoenlappers sit (op julle) ‘This
butterfly is sitting on you and these butterflies are sitting (on you-PL)’.
Both groups fared poorly on this item: No participant in the TD4 group
gave the correct response; only three in the SLI group did. The most
common incorrect response differed between the two groups. Most
171
(seven) of the participants in the SLI group said op haar ‘on her’, whereas
the response favoured by the TD4 group was to replace the pronoun
with a phrase such as die dogtertjies ‘the girls’ or haar en haar ‘her and her’
(eight participants gave this response).
Both groups also fared poorly on item 18, Sy eet appels, maar (julle eet
piesangs) ‘She is eating apples, but (you-PL are eating bananas)’. Only three
participants with SLI and one in the TD4 group gave the correct answer.
Of note is the difference between the types of incorrect answers
provided. These are given in table 6.7. Whereas the incorrect responses
of the SLI group were varied, the response given most by the TD4
group was an unusable one.
Table 6.7. Responses of SLI and TD4 groups to item 18 – Sy eet appels,
maar (julle eet piesangs) – of pronoun production task
Response type SLI group TD4 group
Correct response 3 1
Hulle ‘they’ or ons ‘we’ 5 3
Jy ‘you-SGL’ 2 1
Unusable response, such as jy en hy ‘you-SGL 2 10
and him’
Hy ‘he’ or sy ‘she’ 2 0
No response 1 0
6.3. se-constructions
6.3.1. Results: Picture selection task: se-constructions
This picture selection task was performed in order to assess whether

Afrikaans-speaking children with and without SLI differ in their ability to
correctly comprehend se-constructions. The task had 10 items, which are
given in section 3.1 of appendix D. Each response given to the 10 items
of this picture selection task was classified as one of the following:
(i) Correct picture selected.
(ii) Picture selected matches the first noun of the item (e.g., selected
the picture of a snake without its tongue visible in response to die
slang se tong ‘the snake’s tongue’).
172
(iii) Picture selected matches the second noun of the item (e.g., selected
the picture of a newspaper without its “owner” in response to die
man se koerant ‘the man’s newspaper’).
(iv) Response not usable.
The response given most was (i) – a total of 83% responses were correct.
Response types (iii) and (iv) were only given four times and twice,
respectively (out of a total of 450 responses).
The reliability of the items was low; Cronbach’s alpha was .327. Also,
two of the items correlated negatively with the items as a group and
another four correlated very weakly with the items as a group, as can be
seen in table 6.8. This means that six of the ten items were highly
problematic. The highest correlation of the other four was .507.
Furthermore, to seven of the 10 items, a correct response was given by
almost all participants. This all indicates that this was not a well-designed
task; it did not test what it was supposed to. The rest of the discussion of
the items, as well as that of the participants’ responses, is therefore given
only for the sake of interest; no conclusions regarding SLI as it presents
itself in Afrikaans should be drawn based on this task.
Table 6.8. Item statistics – Picture selection task: se-constructions

Item Item English translation Proportion of Corrected
no. participants item-total
giving correct correla-
answer tion
10 Die man se koerant The man’s .47 .282
newspaper
7 Die vrou se hoed The woman’s hat .53 .058
8 Die seuntjie se kar The boy’s car .60 .507
2 Die slang se tong The snake’s tongue .91 .239
1 Die man se hand The man’s hand .93 -.214
4 Die motor se wiel The car’s wheel .93 -.145
9 Die hond se been The dog’s bone .93 .235
5 Die hond se mandjie The dog’s basket .96 .056
3 Die baba se bottel The baby’s bottle 1.00 .000
6 Die vrou se tande The woman’s teeth 1.00 .000
173
The performance of the three groups of participants on nine of the 10

items was very similar. The only exception was item 10 (die man se koerant
‘the man’s newspaper’). Here, the majority (10) of the typically
developing 6-year-olds selected the correct picture compared to five and
six participants in the SLI and TD4 groups, respectively. The incorrect
picture selected most often by all three groups (nine participants each in
the SLI and TD4 groups and five in the TD6 group) was that of a
newspaper on its own.
6.3.2. Results: Sentence completion task: se-constructions
This task aimed to assess the production of se-constructions by the three

groups of Afrikaans-speaking children. Like the task assessing the
comprehension of se-constructions, this task had 10 items; they are listed
in section 3.2 of appendix D, together with their English translations. To
each response, one of the following five codes was given:
(i) Correct se-construction produced.
(ii) Only one noun was produced, even after prompting by the
researcher (e.g., die boom ‘the tree’ or blare ‘leaves’ instead of die
boom se blare ‘the tree’s leaves’).
(iii) ge ([x«]) was produced instead of [s«].
(iv) Participant indicated that (s)he did not know the answer.
(v) No response given.
Eighty five per cent of all responses were correct. Response type (iii) was
only given by one girl with SLI, but she gave this response to all 10
items. Response types (iv) and (v) were also used infrequently (three
times and once, respectively), all by two children. One of these was a 4-
year-old girl who once said that she did not know the answer − to item 4,
Hier is die blom se blare en hier is (die boom se blare) ‘Here are the flower’s
petals and here are the (tree’s leaves)’. The other child was a boy with
SLI. He also said that he did not know the answer to item 4, and also not
to item 7, Hier is die teddie se maag en hier is (die pop se hare) ‘Here is the
teddy’s tummy and here is the (doll’s hair)’. He gave no response to item
2, Hier is die seuntjie se bed en hier is (die meisie se bed) ‘Here is the boy’s bed
and here is (the girl’s bed)’.
174
The reliability of the 10 items was high, with Cronbach’s alpha being
.874. However, as can be seen from table 6.9, nine of the 10 items were
very easy for the participants.
Table 6.9. Mean number of correct responses to each item – Production

task: se-constructions
10 Hier is die seun se tandepasta en .36 .198
hier is (die vrou se tandeborsel)
9 Hier is die baba se hoed en hier is .84 .636
(die ma se handsak)
7 Hier is die teddie se maag en hier .87 .645
is (die pop se hare)
2 Hier is die seuntjie se bed en hier .89 .507
is (die meisie se bed)
1 Hier is die man se kar en hier is .91 .657
(die meisie se fiets)
4 Hier is die blom se blare en hier is .91 .702
(die boom se blare)
5 Hier is die hond se poot en hier is .91 .746
(die kat se stert)
6 Hier is die meisie se appel en hier .91 .792
is (die seun se piesang)
3 Hier is die eendjie se dam en hier .96 .826
is (die voëltjie se hok)
8 Hier is die ma se trui en hier is .96 .826
(die vrou se broek)
Item 10 was disproportionately difficult compared to the other nine

items. Also, its item-total correlation was markedly lower than that of the
others. Furthermore, not one of the typically developing 6-year-olds gave
the correct response to this item. By contrast, 12 typically developing 4-
year-olds and four children with SLI did. Because item 10 proved to be a
problem item in more than one respect, it was omitted from statistical
analyses. This raised Cronbach’s alpha to .913.
More than or close to 90% of participants gave the correct answer to the
remaining nine items. The performance of the three groups is depicted in
175
figure 6.4. The ceiling effect is an indication that the task as a whole was
too easy to differentiate between the three groups. Note, however, that
there were some participants (15, 18, and 28) who fared markedly worse
than the rest of their group. Levene’s statistic showed a significant
difference in variance between the groups (F2,42=9.109; p=.001), but this
result was rendered because no variability occurred in the TD6 group (all
of its members obtained a score of 9).
10.00
8.00
6.00 28
Score
4.00 15
2.00
0.00 18
SLI TD4 TD6

Group
Figure 6.4. Box plot of performance per group − Production task: se-
constructions
As expected, a one-way ANOVA returned a non-significant outcome,

which means that a difference between the performance of the three
groups could not be assumed (F2,42=3.014; p=.06).
6.4. PERSON AND CASE PRODUCTION IN THE

LANGUAGE SAMPLES
6.4.1. Pronouns
The first 100 complete and fully intelligible utterances of each language
sample were examined for correct and incorrect use of personal and
possessive pronouns. In total, 3523 (98.4%) were used correctly. Table
6.10 summarises the use of such pronouns by the three groups of
176
participants. As can be seen from this table, possessive case posed the
biggest challenge for the 4-year-olds and the children with SLI; both
groups fared better in terms of nominative and oblique case. In table
6.11, a breakdown of these is given for each specific pronoun. From this
table, it can be seen that specifically the possessive case of the third-
person singular masculine pronoun was problematic for the SLI and
TD4 groups, but not so for the TD6 one.
Table 6.10. Accuracy of pronoun use by the three groups of participants

Group Person
Nominative Oblique Possessive Total
SLI 694/700 117/125 140/162 951/987
TD4 784/786 124/125 169/181 1077/1092
TD6 1083/1085 192/194 220/221 1495/1500
Total 2561/2571 433/444 529/564 3523/3579
Table 6.11. Frequency of correct appearance of pronouns, appearing

where they should have
Group Person
1st sgl 1st pl 2nd sgl 2nd pl 3rd sgl 3rd sgl 3rd sgl 3rd pl
Ek Ons Jy Julle Hy Sy Dit Hulle
SLI 237a 60/61 61/62 9 132/134 20 82/83 93/94
Nom
TD4 259/260b 78 35 2 201 50/51 115 44

TD6 344 209 42 4 171/172 123/124 120 70
My Ons Jou Julle Hom Haar Dit/ Hulle
daar
SLI 7 6 3 0/1 40/44 6 46/47 9/11
Oblique
TD4 14 5 5 0 29 4 57/58 10
TD6 21 14 8 0 39 14 81/83 15
My Ons Jou Julle Sy Haar Sy Hulle
Possessive
SLI 66 14/15 9/11 2 8/27 6 16 19

TD4 88 14 10 1 20/30 10/11 16 10/11
TD6 125 28 8 2 31 14/15 2 10
To- 1161 428 181 20 671 247 535 280
tal /1161 /428 /184 /22 /702 /248 /538 /282
aA single figure indicates that the correct form appeared in all obligatory contexts and
that the pronoun in question was never inserted or replaced in inappropriate contexts.
bThe first figure is the number of correct occurrences; the second in the number of
obligatory contexts plus the number of inappropriate insertions and replacements.
177
Table 6.12 shows the mean scores per group in terms of incorrect
pronoun forms as a proportion of all pronouns used in the language
sample. The mean score of the TD6 groups was the highest of the three
and that of the SLI one the lowest. More variation appeared in the SLI
group than in the other two. Levene’s statistic confirmed that the
difference in variation between groups was significant (F2,42=15.820;
p=.000).
Table 6.12. Summary of performance per group – Language sample:

Spontaneous production of errors on pronouns as a
proportion of all pronoun forms
SLI 15 3.33 4.65304 0 13.56
TD4 15 1.71 2.40198 0 6.67
TD6 15 .38 .61716 0 2.04
Total 45 1.81 3.21433 0 13.56
A one-way ANOVA was carried out; it returned a significant outcome,

indicating that a difference between the groups could be assumed
(F2,42=3.528; p=.038). Post hoc analyses (Tukey’s HSD; alpha=.05)
revealed that the difference was only between the SLI and TD6 groups.
Overall, a very small percentage (1.6%) of errors regarding personal and

possessive pronouns was made in the first 100 utterances of the language
samples. These errors are discussed below.
The first error type was the omission of a pronoun in an obligatory

context. All three groups made this error, but the typically developing 6-
year-olds only omitted oblique dit, as in example (100), whereas there was
no specific pattern to pronoun omission by the other two groups. An
example of an omission by a boy with SLI is given in (101).
(100) Target:
juffrou het gesê ons kan vir juffrou het gesê ons kan dit vir
iemand by die huis gee iemand by die huis gee
teacher did say-PAST PART we can for teacher did say-PAST PART we can
someone by the house give it for someone by the house give
‘Teacher said we can give it to someone at home’
178
(101) Target:
een ander kat by Nicole nou né een ander kat by Nicole né
naam is Pienkie sy naam is Pienkie
one other cat at Nicole now hey one other cat at Nicole hey
name is Pienkie his name is Pienkie
‘One cat at Nicole hey his name is Pienkie’
The second main error type was the substitution of a pronoun, either by
a definite article or by another pronoun. Only the 4-year-olds used an
article instead of a pronoun, as shown in (102). Both the TD4 and SLI
groups at times used an incorrect pronoun. The most common
substitution in both groups was that of the possessive sy with hom, as in
*hom hoed ‘him hat’ instead of sy hoed ‘his hat’. Only one substitution in
the TD4 group, that given in (103) below, was not of the hom-sy type.
The SLI group, however, made a wider range of substitutions; one of
them is given in (104).
(102) Target:
hulle bly in die eie huis hulle bly in hulle eie huis
they live in the own house they live in their own house
‘They live in their own house’
(103) Target:
is dit hy maatjie? is dit sy maatjie?
is this he friend-DIM is this his friend
‘Is this his friend?’
(104) Target:
’n hand vashou met hy hande vashou met hom
a hand fast-hold with he hands fast-hold with him
‘hold hands with him’
All of the substitution errors made by the TD4 group involved case (as
opposed to person, number, and/or gender), and, as mentioned above,
all of these errors occurred on what should have been sy ‘his’. The
feature that the SLI group most frequently had incorrect was also case.
Like the TD4 group, this group made no errors related to number
and/or gender. However, a total of four errors related to person were
made. In the TD6 group, one girl made one error related to case,
substituting nominative sy ‘she’ with haar ‘her’ in a comparative
179
construction. One boy made one error related to person, substituting hy

‘he’ with jy ‘you-SGL’.
The third main error type regarding pronouns was their inappropriate
insertion. The TD6 group did not make this error, but the SLI and TD4
groups did. An example of the inappropriate insertion of a pronoun by a
girl with SLI is given in (105).
(105) Target:
hierso is jou klere jou hierso is jou klere
here are your-SGL clothes your-SGL here are your-SGL clothes
/you-OBLIQUE-SGL
‘Here are your clothes’
As mentioned above, only the pronouns occurring in the first 100

complete and fully intelligible utterances were tallied. The occurrence of
pronouns in the rest of the 30 minutes of each language sample was,
however, also examined for pronoun errors. The TD6 group made no
errors, and the only ones made by the TD4 group were those of
substitution. Again, most substitutions were of the hom-sy kind, except
for two. One involved the substitution of possessive hom with hy, and the
other the substitution of nominative hy with nominative sy, as in (106).
(106) Target:
hy het dit gevang [re. a baby girl] sy het dit gevang
he did it catch-PAST PART she did it catch-PAST PART
‘She caught it’
The SLI group also made errors of substitution, of hom-sy and of other
types. This group furthermore made errors of omission – one of them
given in (107) – as well as errors of insertion – one given in (108).
(107) Target:
ek kon gedoen het ek kon dit gedoen het
I can-PAST do-PAST PART did I can-PAST it do-PAST PART did
‘I could have done it’
180
(108) Target:
ons ma leer ons saam ons ons ma leer saam met ons
our mom learn we/us with us our mom learn with us
‘Our mom is also learning with us’ (i.e., our mom is also in our class)
6.4.2. se-constructions
The occurrence of se-constructions in the first 100 complete and fully

intelligible utterances of the language samples was examined. In total, in
these 45000 utterances, se-constructions (whether correct or incorrect)
were produced 99 times: 41 times by the SLI group, 21 times by the TD4
group, and 37 times by the TD6 group. Of these 99 se-constructions, 96
(97%) were grammatical. Table 6.13 contains a per-group summary of
the use of se-constructions in the 45 language samples.
Table 6.13. Number of se-constructions used correctly and incorrectly in

Group Grammatical Ungrammatical
SLI 38/41 95% 3/41 5%
TD4 21/21 100% 0/21 0%
TD6 37/37 100% 0/37 0%
Total 96/99 97% 2/98 3%
Both typically developing groups made no errors on se-constructions.

Three members of the SLI group each made one error: One girl inserted
the se inappropriately (in *hulle se kos ‘their possessive marker food’); and
two boys omitted the se (in *hy’s by my sussie skool instead of hy’s by my
sussie se skool ‘he’s at my sister possessive marker school’, and in *die man kop
draai instead of die man se kop draai ‘the man possessive marker head turns’).
In the rest of the 30 minutes of the language samples, there was only one
error related to se: The boy with SLI who said *hy’s by my sussie skool, also
said *kyk die ystervarks maag ‘look at the porcupine-s tummy’.
6.5. DISCUSSION: PERSON AND CASE

COMPREHENSION AND PRODUCTION
The task testing the comprehension and production of se-constructions

proved problematic. The comprehension task had problems with item
reliability, amongst others, and the production task was too easy,
181
resulting in no difference being observable between the three groups. No

conclusion can therefore be reached about the comprehension or elicited
production of this construction by the three groups of participants. This
finding was surprising, given that the production task used in the present
study was very similar to that employed by Leonard et al. (1997) with
English-speaking children (the SLI ones aged 3 years 7 months to 5 years
9 months) and by Leonard et al. (2001) with Swedish-speaking children
(the SLI ones aged 4 years 3 months to 5 years 7 months). They found
that the children with SLI had significantly lower scores than age- and
also MLU-matched controls, with all errors being omissions of the
possessive morpheme.
The spontaneous production data also revealed very few differences

between the groups in terms of se production. When studying the use of
14 grammatical morphemes in the spontaneous language samples of 100
typically developing 4-year-olds, Balason and Dollaghan (2002) found
that there were more than three obligatory contexts for the occurrence
of the possessive ’s in only 8 of the samples. Similar to the finding of the
present study, this morpheme occurred 100% of the time in these
obligatory contexts (i.e., there was no omission of possessive ’s in the
samples of the 4-year-olds). It appears then that this is an infrequently
occurring morpheme in the language of young English-speaking
children, but when it does occur, it is always used correctly (also see
Leonard et al. 1997:747). Whereas not as infrequently occurring in the
Afrikaans data, both typically developing groups used the se-construction
correctly whenever it did occur.
Turning now to the production of pronouns: Case studies reveal that at

the time of mastery of personal pronouns, English-speaking children’s
MLU was 4.13, at 2 years 10 months (Oshima-Takane 1992) and 4.7, at 2
years 1 month (Schiff-Myers 1983). Anderson (1998) employed more
(Spanish-speaking) participants and found substantial variability with
respect to the age at which certain pronouns were mastered: Some
children in the 2-year-old group demonstrated mastery whereas some
aged 3 years 0 or 1 month did not. However, all children older that 3
years 1 month demonstrated mastery. These ages are well below those of
the children who participated in the present study. Whereas the
Afrikaans-speaking typically developing 6-year-olds demonstrated
182
(almost) full mastery of personal and possessive pronouns, the 4-year-

olds did not.
Roberts and Rescorla (1995) found that 4-year-olds with an expressive

language delay fared significantly worse than age-matched controls in the
spontaneous production of pronouns. Some studies show that children
with SLI substitute nominative pronouns with object ones at a higher
rate than do typically developing age-matched children (Moore 2001).
However, researchers are divided on whether this is the case for MLU-
matched controls: Moore (2001) found no difference between the
children with SLI and MLU-matched controls, whereas Wexler et al.
(1998) and Loeb and Leonard (1991) found that children with SLI do
make more such errors than the MLU-controls. Specifically, the latter
group of researchers (1991:344) found that children with SLI who were 4
years 0 months to 5 years 0 months, when compared to typically
developing MLU-matched controls, produced fewer subject case-marked
pronouns. In other words, the children with SLI produced more case
errors than the younger typically developing ones. Furthermore, those
children (SLI and MLU-matched) who produced a number of case
errors, produced few (if any) gender errors, and those who produced a
number of gender errors tended to produce fewer case errors. As almost
no gender errors were made by the three groups of participants in the
present study, it is not possible to state whether this would be the case
for Afrikaans-speaking children as well. However, what can be stated is
that, in the present study, there was only one substitution of a
nominative pronoun with an oblique one. The reason why not even the
Afrikaans-speaking children with SLI made this error could be that the
participants in the present study were older than most in other studies,
and that this error no longer occurred in their language. Alternatively, it
could merely be that this error type is not one made by Afrikaans-
speaking children. The feature for which most error were recorded was
indeed case, but possessive pronouns − not nominative ones − were
being replaced by oblique ones. The TD6 group did not make errors of
this kind, but both other groups did.
Studies also indicate that children with SLI and younger typically
developing children make more errors with she than with he (cf. Moore
1995, 2001; Rispoli 1994, 1998a). In the present study, this was not the
case. In fact, the reverse was observed: The Afrikaans-speaking children
183
(TD4, TD6, and those with SLI) hardly ever made any error on sy ‘she’
or haar ‘her’; the most common error made (by the 4-year-olds and the
children with SLI), in both elicited and spontaneous production, was the
substitution of sy ‘his’ by hom ‘him’, as in *wat is hom naam? ‘what is him
name?’.
It is particularly noteworthy that julle ‘you-PL’ had a very low rate of

occurrence in the language samples of all three groups. The children also
very often avoided the use of julle in the experimental task assessing
production (by, for instance, saying jy en sy ‘you-SGL and she’) or replaced
julle with hulle ‘they/them’ (the latter also frequently occurred in the
picture selection task). It is unlikely that this frequent replacement of julle
with hulle can be attributed to a lack of perspective-taking, as Ricard,
Girouard, and Gouin Décarie (1999) found that 2-year-olds were already
able to understand the perspective of the speaker, the hearer, and the
non-addressed. More likely is the Rispoli explanation:137 hulle is recalled
instead of julle, because these two forms are phonologically highly similar
and possibly also because pronouns with the form h- (hulle, haar ‘her’, hy
‘he’, hom ‘him’) are slightly more common than those with the form j- (jy
‘you-SGL’, jou ‘your-SGL’, julle ‘you-PL’).
The results of the language sample analysis support those of Moore

(1995) who found that English-speaking 5-year-olds with SLI made more
pronoun errors than did typically developing 5-year-olds, in a production
task similar to the one used during language sample elicitation in the
present study. Also similarly to what was found in the present study,
Moore (2001) found that English-speaking children (language-impaired
and typically developing) made no errors with the person of pronouns
and almost none with the number. Errors of case and gender were made,
137 On Rispoli’s Paradigm Building Hypothesis (1994, 1998a, 1998b, 1999), accessing a
word in the pronoun paradigm is guided by the grammatical features of the word as
well as the phonological features of the paradigm. If the correct pronoun cannot be
accessed immediately, the child’s mental “search” is influenced by factors such as the
number of cells in the paradigm filled by that specific word, the phonological structure
of the word, and the phonetic prominence of the word. This hypothesis predicts, for
example, that English-speaking children, if they cannot immediately recall she, will
replace she with her, amongst other reasons because her occurs more times in the
pronoun paradigm than does she (a prediction supported by the results of Rispoli 1998a
and Moore 1995). Under such circumstances, them may also be replaced by they, due to
the large similarities in the phonological form of these two words.
184
but typically developing children fared significantly better in terms of

correct production of case than did those who were language-impaired.
By contrast, Moore’s (1995) typically developing children made no errors
of case, whereas this was the most frequent error made by the children
with language impairment. One reason Moore (1995:69) offers for the
high proportion of case errors in the language of children with SLI is
that case, unlike person, semantic gender, or number, does not correlate
with differences in the real world. Case depends on the linguistic context
of the pronoun, not on the real gender or number of the referent(s).
Rispoli (2005:108) warns that errors in pronoun form – specifically

errors of case – do not occur merely because children’s grammars are
still immature; even with a very immature grammar, errors can be
minimised if children do not attempt to act more ambitiously than that
for which they have the capacity. By this, Rispoli means that errors can
be limited if children (i) use only those cells of the pronoun paradigm for
which entries already exist and which are therefore easily accessible, and
(ii) use nouns and demonstratives instead of those pronouns which are
difficult to access. As the experimental tasks revealed that Afrikaans-
speaking 6-year-olds with SLI fared worse than typically developing
peers in terms of production of correct pronoun forms, the results of the
spontaneous language samples – which indicate relatively high levels of
accuracy even by the SLI group – could possibly be explained in terms of
the strategy outlined by Rispoli: The children with SLI mainly used only
easily accessible pronouns and substituted the rest with nouns.
Finestack, Fey, and Catts (2006) analysed the narratives of young

English-speaking school-going children with and without SLI. This
analysis was done in terms of their use of complete pronominal
referencing (in other words, their use of pronouns where it is clear to
whom or what the pronoun is referring). It was found that this measure
was not sufficiently sensitive to differentiate between the two groups. A
similar finding was made in the present study: Apart from the error
relating to hom-sy ‘him-his’, there were no large (quantitative) differences
between the children with and without SLI in terms of their accuracy of
pronoun production in spontaneous language. However, when picture
selection, judgement, and sentence completion tasks were employed (i.e.,
when elicited rather than spontaneous pronoun use was examined),
185
personal pronouns was indeed a sensitive enough measure to

differentiate between 6-year-olds with and those without SLI.
Pronoun acquisition in English has been studied extensively and for a

long time (see, e.g., Cooley 1908 and Bain 1936), but there is still no
comprehensive theory which can explain all the facts, especially not the
use of pronoun case (Schütze 1999:754). By contrast, data on the
development of the pronominal system in child speakers of other
languages are scarce (Anderson 1998:394). However, it is accepted by
some that, in the case of non-agglutinative languages, pronouns need to
be learnt word by word, as there are no general rules guiding the
acquisition of the form of the pronoun (Rispoli 1994, 1998a, 1998b).
Pronouns are particularly difficult for young children to master, seeing

that their form changes according to the context in which they appear.
The form of the pronoun is affected by the number and gender of the
referents, by who the speaker and hearer is, and by the grammatical role
of the pronoun in the sentence (Moore 2001:208; Ricard et al. 1999:68-9;
also see Campbell, Brooks, and Tomasello 2000). Due to their deictic
characteristics, pronouns furthermore cannot be modelled to children
without potentially causing confusion; children need to listen to the
pronouns in the language of other speakers in order to become familiar
with the appropriate use thereof (Cole, Oshima-Takane, and Yaremko
1994:115).
Turning to the Afrikaans data, it appears that also Afrikaans-speaking

children with SLI fare worse than their typically developing peers in
terms of realising number and case on pronouns. Regarding research
questions 1 to 4 (cf. p.3), Afrikaans-speaking children with SLI present
with a delay in their comprehension and production of grammatical
morphemes related to person and case on pronouns. Although the
responses of the SLI and 4-year-old groups to certain items of the
experimental tasks differed, no pattern could be detected in these
differences. Therefore, based on the results of the experimental tasks
alone, one could conclude that the language of Afrikaans-speaking
children is merely delayed. However, analyses of the spontaneous
language samples of these two groups reveal that the SLI group made
186
some errors not found in the language of the typically developing

children. Therefore, the Afrikaans-speaking children with SLI do not
present with merely a language delay; their language is also to some
extent deviant.
187
Results: Tense
Chapter 7
Tense comprehension and production
7.1. INTRODUCTION
Two tasks were used to test comprehension of past and present tense
constructions. The first, discussed in section 7.2.1, was a picture
selection task. The second tested only comprehension of hendiadyses
(such as geloop en eet ‘ate while walking’), the results of which are
presented in section 7.2.2. Section 7.2.3 contains the results of the
production task, which entailed sentence completion. In section 7.3, the
results of the experimental tasks are discussed in light of what was found
by other researchers for elicited production of verb morphology. The
spontaneous production of past and present tense constructions, as
found in the language samples of the participants, is reported in section
7.4 and discussed in section 7.5.
7.2. RESULTS: EXPERIMENTAL TASKS
7.2.1. Results: Picture selection task: Present and past tense
In order to compare the comprehension of the present/past tense

distinction of Afrikaans-speaking children with SLI to that of typically
developing ones, a picture selection task consisting of 24 items was
performed. The items of this task are given in section 4.1 of appendix D,
together with their English translations. Initially, an elaborate coding
scheme consisting of five categories was used, but, as three of the four
error categories were used infrequently, this scheme was simplified:
Responses were recoded as either correct or not correct.
The reliability of the 24 items was low; Cronbach’s alpha was .562.
Individual item statistics are presented in table 7.1, with items ordered
from those to which most participants gave an incorrect response to
those to which most gave a correct response. As can be seen from this
table, the items of this task had a wide range of difficulty: The
189
Results: Tense
proportion of participants who gave the correct answer ranged from .18
to .98.
Table 7.1. Item statistics – Picture selection task: Present and past tense
no. giving correct correlation
answer
21 Die hond sal skoon wees na sy bad .18 .151
2 Die baba was vuil .24 .311
17 Die koek sou verbrand het .27 .108
8 Die posman was hier gewees .33 .319
14 Die meisie het ’n stokkielekker gehad .38 .497
24 Hy het die stukkende gloeilamp vervang .38 .054
11 Die teddie het twee oë gehad .40 .245
23 Die vrou moes lank wees om te kon bykom .47 .158
3 Die potlood was lank gewees .49 .201
5 Die vrou sny die gras .53 .126
12 Die vrou het kos gekoop .58 .085
20 Gister val die voëltjie uit sy nes .60 .043
22 Die appel val .60 .162
6 Die boot het gesink .62 .405
13 Die seun verf die heining .67 .211
10 Die melk was op .69 .155
9 Gister klim ek op die dak .73 .158
15 Die baba het dit ontvang .76 -.062
18 Die man moet fiks wees om die berg te kan
.76 .102
uitklim
4 Die vliegtuig het opgestyg .78 .045
7 Die pop het twee arms .87 .148
16 Die koeldrank het omgeval .87 .169
19 Die seun het twee ballonne .89 .108
1 Die ballon bars .98 .194
As can also be seen from this table, each individual item correlated
poorly with the items as a group. One item had a negative correlation to
the rest, but even of those which had a positive one, only four had a
correlation stronger than .300. These four did not contain the same kind
190
Results: Tense
of verb. Two were the past tense form of be, one the past tense form of
have, and another the temporal het with the past participial form of the
main verb.
In order to improve the reliability of the scale, (i) some items were
subsequently removed – specifically those which did not differentiate
between the three groups of participants, as well as those which rendered
inexplicable response patterns; and (ii) the remaining items were grouped
into two groups, according to the type of verb occurring in the item.
The first subgroup consisted of six items, all past tense forms of be (i.e.,
was and was gewees) and have (het gehad). Cronbach’s alpha was .646.
Although this did not indicate high reliability, it was an improvement on
the .561 obtained for all 24 items. The new individual statistics for these
six items are presented in table 7.2.
Table 7.2. Item statistics of subgroup of items consisting of past tense

forms of be and have – Picture selection task: Present and past
tense
no. participants item-total
giving correct correlation
answer
2 Die baba was vuil .24 .311
8 Die posman was hier gewees .33 .319
11 Die teddie het twee oë gehad .40 .245
3 Die potlood was lank gewees .49 .201
10 Die melk was op .69 .155
Figure 7.1 depicts the performance of the three groups on these six
items. From this figure, it appears that the TD6 group fared better than
the other two. Unlike the pattern observed for the other experimental
tasks discussed thus far, there was less variability in the SLI group than
in the other two. However, Levene’s statistic indicated that the variability
between groups was not significant (F2,42=.843; p=.438).
191
Results: Tense
6.00
5.00 24
4.00
Score
3.00
2.00
1.00
0.00
1 2 3
Group

consisting of past tense forms of be and have – Picture selection
task: Present and past tense
The mean scores of the three groups are given in table 7.3. That of the
45 participants together was low: 2.53 out of a possible 6. A one-way
ANOVA returned a non-significant outcome, which means that no
Unlike most of the experimental tasks discussed thus far then, the TD6
group did not outperform the other two on this subgroup of items;
neither did the SLI group show more variability than the other two.

consisting of past tense forms of be and have – Picture selection
task: Present and past tense
SLI 15 1.87 1.45733 0 5
TD4 15 2.53 1.64172 0 6
TD6 15 3.20 1.89737 0 6
Total 45 2.53 1.72679 0 6
Even though no statistically significant difference was found between the

responses of the SLI and TD4 groups, some differences existed for two
192
Results: Tense
of the six items. To item 3 − Die potlood was lank gewees ‘The pencil was
long’ – the majority of the SLI group’s responses (nine) was incorrect,
whereas the TD4 group mostly gave correct responses (eight). Die baba
was vuil ‘The baby was dirty’ (item 2) was difficult for all three groups,
but more so for the SLI group than for the other two. No participant in
the SLI group gave the correct answer (12 of them pointed to the picture
where the baby is still dirty); in the TD4 and TD6 groups, six and five
participants, respectively, selected the correct picture.
A second grouping of items consisted of het forms, whether het as a main

verb or the temporal auxiliary het. Only one item containing het was
omitted from this subgroup, viz. item 11 (Die teddie het twee oë gehad),
because it had a negative item-total correlation (when compared to the
others in this subgroup). The reliability of this subgroup was slightly
lower than that of the first one: Cronbach’s alpha was 6.11. The new
individual item statistics for the seven items are given in table 7.4.
Table 7.4. Item statistics of subgroup of items consisting of past tense

forms containing het – Picture selection task: Present and past
tense
12 Die vrou het kos gekoop .58 .352
6 Die boot het gesink .62 .127
4 Die vliegtuig het opgestyg .78 .517
7 Die pop het twee arms .87 .395
16 Die koeldrank het omgeval .87 .446
19 Die seun het twee ballonne .89 .209
In Figure 7.2, the performance of the three groups on this subgroup of

seven items is depicted. It appears that the TD6 group fared better than
the other two and that most variability was found in the SLI group,
which is a more “typical” result (considering those of the other
experimental tasks discussed thus far) than that of the first subgroup.
However, Levene’s statistic indicated that the variability between groups
was not significant (F2,42=.315; p=.731).
193
Results: Tense
7.00
6.00
5.00
Score
4.00
3.00
2.00 33
1.00 16
1 2 3
Group

Table 7.5 contains the mean scores of the three groups. A one-way
ANOVA returned a significant outcome, indicating that a difference
between the groups could be assumed (F2,42=4.578; p=.016). Post hoc
analyses (Tukey’s HSD; alpha=.05) revealed that the differences were
groups. In this sense, this subgroup rendered a result which matched that
of the other experimental tasks better than did the subgroup of items
involving the past tense form of be and have.

SLI 15 4.53 1.72654 1 7
TD4 15 4.47 1.30201 3 6
TD6 15 5.93 1.43759 2 7
Total 45 4.98 1.61652 1 7
A difference could be detected for only one individual item of the seven,
namely Die boot het gesink ‘The boat sank’ (item 6): To this item,
194
Results: Tense
marginally more of the SLI group’s responses (eight) were incorrect than
those of the TD4 group (seven); in other words, the SLI group mostly
gave an incorrect response, whereas the TD4 group mostly gave a
correct one.
7.2.2. Results: Judgement task – Hendiadyses: Present and past

tense
The picture selection task included two hendiadyses, neither of which

was included in the subgroups subsequently formed. However,
hendiadyses were also assessed using a judgement task. The responses to
the 10 items of this task (given in section 4.2 of appendix D, with their
English translations) were coded as follows:
(i) A response was taken to be correct if a grammatical form was
judged as such or an ungrammatical one judged as such.
(ii) A response was regarded as incorrect if a grammatical form was
judged to be ungrammatical or an ungrammatical one to be
grammatical.
Individual item statistics are given in table 7.6.
Table 7.6. Item statistics – Judgement task: Hendiadyses

no. giving correlation
correct answer
3 Die seuntjie het staan en gehuil tot hy sy .13
.663
kombersie gekry het
2 Die baba het lê en gespeel .13 .723
9 Hulle het heeldag sit en gespeel .18 .623
4 Die man het loop en gedink .20 .463
7 Gister het die kat heeldag staan en gemiaau .22 .709
giving incorrect
answer
5 Hy het geloop en eet .09 .685
10 Die honde het gister heeldag gelê en slaap .11 .637
8 Hulle het gesit en werk .16 .724
6 Die man het gister koerant gesit en lees .16 .724
1 Hulle het heeldag gestaan en praat .33 .544
195
Results: Tense
The reliability for the 10 items as a group was a problem, as indicated by

a high negative Cronbach’s alpha (-.973). This was because participants
said “yes” or “no” fairly consistently, independently of the test item.
Grammatical items were then recoded, with the code being swopped
(i.e., every correct response to these items was awarded zero and every
incorrect response 1). Cronbach’s alpha then increased to .894, which
indicated high reliability.
Figure 7.3 indicates that all three groups showed great variability: The
scores of the TD6 group ranged from 0 to 7 out of 10, whereas some
participants in the SLI group obtained zero and others the maximum
score. Levene’s statistic of homogeneity of variance was not significant
(F2,42=.553; p=.579), indicating that the variance in each of the three
groups can be taken to be equal.
28
10
14
25
8
6
Score
SLI TD4 TD6

Group
Hendiadyses
Three participants performed markedly better than their group. The first
two were a boy and girl with SLI. The boy (participant 28) judged all
items to be ungrammatical. Because of the recoding of responses, he
obtained a perfect score. This was not his response mode on any of the
other judgement tasks: Whereas he had a preference for judging items to
be ungrammatical on these tasks, he did not judge all of them as such.
There was no clear pattern to the girl’s (participant 25) responses. One 4-
196
Results: Tense
year-old boy also fared far better than the other participants in his group.
On two of the four other judgement tasks, this boy judged all items to be
ungrammatical. Here he showed a preference for this type of response,
but he (correctly) judged the last item, Die honde het gister heeldag gelê en
slaap ‘The dogs lay sleeping all day long yesterday’, to be grammatical.
Before the recoding of responses, the number of correct responses was

71, 78, and 76 out of a possible 150 for the SLI, TD4, and TD6 groups,
respectively. That is a mean score of 4.73, 5.50, and 5.07 for the three
groups, respectively. The performance of all groups was therefore at or
close to chance level. Table 7.7 contains the mean scores after recoding
of responses. A one-way ANOVA returned a non-significant outcome,
which means that no difference between the mean scores (after
recoding) of the three groups could be assumed (F2,42=.145; p=.865).
Item-by-item inspection of the responses of the three groups
furthermore indicated that there were no noteworthy differences
between the groups on or across items.

Hendiadyses
SLI 15 1,47 3,13657 0 10
TD4 15 1,67 2,43975 0 9
TD6 15 1,67 2,59119 0 7
Total 45 1,71 2,68516 0 10
7.2.3. Results: Sentence completion task: Past tense

constructions
The aim of this task was to establish whether or not Afrikaans-speaking

children with SLI perform age-appropriately as regards their production
of various types of past tense constructions. The task had 18 items, given
in section 4.3 of appendix D. However, if the participant did not give the
targeted response (for instance, if a non-targeted auxiliary verb or the
historic present tense was used), the researcher prompted the participant
by giving the auxiliary verb in part or in full, in order to see whether the
participant could then produce the rest of the targeted construction. So a
participant could potentially have given 36 responses to this task.
197
Results: Tense
Initially, the response (the unprompted and, where applicable, the

prompted one) of each participant to each item was placed in one of 52
categories. A summary of these categories is given below.
Without the researcher providing any part of the modal auxiliary or the
temporal het, the participant:
(i) gave the correct past tense construction;
(ii) replaced het with was ‘be-PAST’ (e.g., *was hy ook gekraak elke dag
instead of het hy ook elke dag gekraak);
(iii) used the past participle but omitted het (e.g., *’n blom gepluk instead
of het ‘n blom gepluk);
(iv) used the het but replaced the past participle with the infinitive (e.g.,
*het elke dag swem instead of het hy elke dag geswem);
(v) used the het but inappropriately used the prefix ge- with the past
participle (e.g., *het sy die verwer gebetaal instead of het sy die verwer
betaal);
(vi) used het twice and replaced the past participle with the infinitive
(e.g., *het hy ’n appel eet het instead of het hy ’n appel geëet);
(vii) gave a historic present tense construction (e.g., eet hy ’n appel).
After the researcher provided the modal auxiliary or the temporal het in
full or in part, the participant:
(viii) gave the correct past tense construction;
(ix) used het (het thus occurred twice) and replaced the past participle
with the infinitive;
(x) inappropriately used the prefix ge- with the past participle;
(xi) (still) used the historic present tense;
(xii) used het but replaced the past participle with the infinitive (after a
modal auxiliary was provided by the researcher) (e.g., *kon sy buite
speel het instead of kon sy buite gespeel het).
Other categories included:

(xiii) participant gave a bare stem as response (e.g., sit en slaap instead of
het hy gesit en slaap);
(xiv) participant gave only a PP/AdvP/DP as response (e.g., laat
instead of was hy laat);
(xv) the response was not usable;
(xvi) participant indicated that (s)he did not know the answer;
198
Results: Tense
(xvii) participant repeated the researcher’s utterance (i.e., the stimulus)

verbatim;
(xviii) participant gave no response.
Instead of comparing all responses (up to 36 of them per participant)

across the 52 response categories, these categories were grouped so that
it was possible to compare the three groups of participants in terms of
the following:
A. The number of targeted (unprompted) responses given.
B. The number of grammatical past tense constructions, whether
exactly like the target or not, given without prompting (including
response type A) — for example, het hy skool toe gegaan instead of
moes hy skool toe gegaan het; het hy geslaap instead of het hy gesit en slaap;
and wou sy buite speel instead of kon sy buite gespeel het.
C. The number of targeted responses given after prompting (i.e.,
after a part of or the full auxiliary was given by the researcher).
D. The number of acceptable past tense constructions, whether
exactly like the target or not, given after prompting (including
response type C).
E. The number of historic present tense constructions which were
direct equivalents of the targeted construction − naturally, these
would all be without prompting by the researcher − e.g., pluk sy ’n
blom instead of het sy ’n blom gepluk; verstaan sy alles instead of het sy
alles verstaan; or moet hy skool toe gaan instead of moes hy skool toe
gegaan het.
F. The total number of historic present tense constructions, whether
direct equivalents of the targeted construction or not (including
response type E) − again, these would all be without prompting by
the researcher − e.g., speel sy buite instead of kon sy buite gespeel het; or
wil hy skool toe gaan instead of moes hy skool toe gegaan het.
G. The number of highly idiosyncratic errors.
H. The number of past tense constructions in which het occurred, but
the past participial (ge-) form was replaced by an infinitival one,
e.g., *het sy ’n blom pluk instead of het sy ’n blom gepluk.
I. The number of other errors – excluding error types G and H –
before any prompting.
J. The number of other errors – excluding error types G and H –
after prompting by the researcher.
199
Results: Tense
Response types A to F thus represented grammatical responses (whether

targeted or not) and types G to J ungrammatical ones. The performance
of the three groups in terms of grammatical responses is presented in
table 7.8. As can be seen from this table, the TD6 group gave the most
targeted responses (A). A one-way ANOVA returned a significant
outcome, indicating that a difference between the groups could be
assumed (F2,42=13.631; p=.000). Post hoc analyses (Tukey’s HSD;
alpha=.05) revealed that the significant differences where between the
SLI and TD6 groups, and between the TD4 and TD6 groups.
In terms of grammatical past tense constructions (B), the TD6 group

also outperformed the other groups. Base on the results of one-way
ANOVA (F2,42=15.238; p=.000), a differnce between groups could be
assumed. Post hoc analysis (Tukey’s HSD; alpha=.05) again indicated the
significant differences being between the SLI and TD6 groups, and
between the TD4 and TD6 groups.
In the SLI and TD4 groups, six and five participants, respectively, did
not produce any grammatical past tense constructions (A or B),
compared to one in the TD6 group. Because the TD6 group gave a
higher number of grammatical past tense constructions without
prompting, less prompting was, of course, necessary for this group, and
therefore their mean scores for response types C and D were lower than
those of the other two groups – as expected. For response type C (which
was targeted responses given after prompting), a difference between the
three groups could be assumed (based on the outcome returned by a
one-way ANOVA; F2,42=3.444; p=.041), but post hoc analyses with
Tukey’s HSD (alpha=.05) did not indicate which groups differed from
which. Regarding response type D (which was acceptable past tense
constructions, whether exactly like the target or not, given after
prompting), differences between the groups could again be assumed
(based on the outcome returned by a one-way ANOVA; F2,42=4.537;
p=.016). As was the case for response types A and B, post hoc analyses
(Tukey’s HSD; alpha=.05) again indicated that the significant differences
were between the SLI and TD6 groups, and between the TD4 and TD6
groups.
Response types E and F entailed historic present tense constructions.

For these response types, the 4-year-olds had the highest mean score,
200
Results: Tense
indicating that this group preferred this type of construction to indicate

past tense. However, a one-way ANOVA returned a non-significant
outcome, which means that a difference between the groups could not
be assumed (F2,42=0.734; p=.486, for response type E; F2,42=2.159;
p=.128, for response type F).
Table 7.8. Summary of performance per group – Sentence completion

task: Past tense, grammatical responses
RTa Group N Mean SDb Min. Max. Levenec ANOVAc
score score F p F p
A SLI 15 2.78 3.326 0 10 2.354 .107 13.631 .000
TD4 15 1.73 2.520 0 9
TD6 15 8.00 4.472 0 16
B SLI 15 4.40 5.082 0 14 2.905 .066 15.238 .000
TD4 15 2.07 3.261 0 11
TD6 15 11.20 5.480 0 17
C SLI 15 2.53 2.875 0 8 .460 .634 3.444 .041
TD4 15 4.60 2.414 1 9
TD6 15 2.47 2.264 0 6
D SLI 15 3.27 3.173 0 9 .851 .434 4.537 .016
TD4 15 5.73 2.576 2 11
TD6 15 2.93 2.550 0 7
E SLI 15 3.93 4.543 0 17 .011 .989 .734 .486
TD4 15 5.80 4.144 0 14
TD6 15 4.47 4.340 0 15
F SLI 15 6.93 4.906 1 19 .822 .447 2.159 .128
TD4 15 10.40 6.220 0 21
TD6 15 6.67 5.260 1 17
aRT=response type. bSD=standard deviation. cdf=2,42.
It appears then that the TD6 group gave statistically significantly more
targeted responses than the other two groups, and also produced more
untargeted but grammatical constructions than the other two groups
(whether prompting by the researcher occurred or not).
Turning to the ungrammatical past tense constructions – response types

G to J: The most errors occurred after prompting by the researcher (J),
as can be seen from table 7.9. This could be explained as follows: If a
participant did not give the targeted construction, the researcher used
prompting in an attempt to elicit the targeted construction. This
construction was presumably not given without prompting because the
participant could not produce the construction. Now, with prompting,
201
Results: Tense
the participant was “forced” to attempt to produce the construction, and

this resulted in an ungrammatical form of the construction being
produced. A possible reason for the low mean number of errors
produced before prompting (I), is that, if participants did not know the
targeted construction, they could provide an alternative (grammatical)
construction. For instance, if the past tense form of het ‘have’ was not
known to participants, they would not have said het hy ’n nuwe maatjie
gehad ‘he had a new friend’ in response to item 16, Hierdie seuntjie het elke
dag ’n nuwe maatjie. Gister, net soos elke ander dag, … ‘Every day, this boy has
a new friend. Yesterday, just like every other day, …’. They could use
another grammatical construction, such as het hy ’n nuwe maatjie ‘he has a
new friend’ (the historic present tense form) or het hy weer met ’n maatjie
gespeel ‘he again played with a friend’. However, when prompted, the
participants did not have this freedom of choice and then a deviant or no
response was likely to be given.
Table 7.9. Summary of performance per group – Sentence completion

task: Past tense, ungrammatical responses
RT. Group N Mean SDb Min. Max. Levenec ANOVAc
score score F p F p
G SLI 15 2.33 2.526 0 8 10.987 .000 5.589 .007
TD4 15 0.47 0.743 0 2
TD6 15 0.67 1.234 0 4
H SLI 15 0.33 0.617 0 2 16.605 .000 3.128 .054
TD4 15 0.07 0.258 0 1
TD6 15 0.00 0.000 0 0
I SLI 15 1.40 1.844 0 5 7.101 .002 2.203 .123
TD4 15 0.73 1.163 0 4
TD6 15 0.40 0.737 0 2
J SLI 15 8.73 6.713 1 22 2.281 .115 3.543 .038
TD4 15 7.67 9.092 0 30
TD6 15 2.20 5.321 0 21
aRT=response type. bSD=standard deviation. cdf=2,42.
One-way ANOVA was performed on each of these types of

ungrammatical responses to determine whether the differences in the
mean scores of the three groups were significant. The results of these
statistical analyses are also given in table 7.9. Differences between groups
could be assumed to exist for response types G and J. In each case, post
hoc analyses with Tukey’s HSD (alpha=.05) showed the differences to
be between the SLI and TD6 groups. For response type G, there was a
202
Results: Tense
further difference between the TD6 and TD4 groups. The one-way
ANOVA returned a non-significant outcome for response types H and I,
which means that, for these response types, no difference between the
groups could be assumed.
Based on the outcomes of a one-way ANOVA, differences between

groups could be assumed for response type G. As expected, the highest
number of “odd” errors (G) occurred in the SLI group. The mean scores
of the other two groups were comparable on this measure. Post hoc
analyses with Tukey’s HSD (alpha=.05) confirmed that the differences
were between the SLI and TD6 groups, and between the TD4 and TD6
groups.
The only other response type for which group differences could be
assumed was type J, i.e., total number of errors – excluding error types G
and H – after prompting by the researcher. Post hoc analyses revealed
that the difference was between the SLI and TD6 groups. Here, as was
expected, the SLI group produced a significantly higher number of
ungrammatical past tense constructions than did the TD6 group.
However, almost the same number of these constructions was produced
by the SLI and TD4 groups (cf. table 7.9), and the participant who fared
worst in the TD4 group made more errors than the worst-faring
participant with SLI.
In terms of replacing the past participial (ge-) form with an infinitival one
while still using het (response type H), the three groups did not differ
significantly: All groups obtained a very low score, and this response was
never given by any member of the TD6 group. Furthermore, no
significant difference could be assumed between groups for the total
number of ungrammatical responses before prompting by the researcher
(response type I); the mean scores of all groups were again very low.
In short, the SLI and TD4 groups made a comparable number of errors
after prompting by the researcher, and the SLI group made more
idiosyncratic errors than did the other two groups.
203
Results: Tense
7.3. DISCUSSION: EXPERIMENTAL TASKS
Acceptability judgements performed on hendiadys did not render any

significant difference between the SLI, TD4, and TD6 groups. Picture
selection showed that the TD6 group outperformed the other two, if
only a subgroup of items (those involving het) were considered.
However, these results should be interpreted with caution: The
judgement task comprised only 10 items, and participants mostly gave
one type of response to all items, whether grammatical or not. The
validity of the 24 items of the picture selection task as a whole was
questionable. The reported results are based on a selection of only seven
of these 24 items; it could thus be that the results are an artefact of the
instruments used rather than a true reflection of the ability of these three
groups of children to correctly comprehend present vs. past tense
constructions.
As stated, it was found that Afrikaans-speaking typically developing 6-

year-olds outperform both 6-year-olds with SLI and typically developing
4-year-olds in the picture selection (tense comprehension) task; the latter
two groups perform similarly. This was also the general pattern for the
task eliciting production of past tense forms: The children with SLI
performed like typically developing 4-year-olds, but made more
idiosyncratic errors than did the 4-year-olds and age-matched controls.
The findings of the present study are only partly supported by most
others on the elicited production of past tense forms. Several authors
found that English-speaking children with SLI are outperformed by
both age- and MLU-matched controls. In a study by Leonard, Miller,
Rauf, Charest, and Kurtz (2003), 4- to 6-year-old children with SLI
performed significantly worse than both age-matched controls and those
matched for MLU on a task eliciting the production of regular –ed past
tense forms: Fewer past tense forms were produced in obligatory
contexts by the children with SLI, whereas there was no statistically
significant difference between the performance of the other two groups.
Oetting and Horohov (1997) found that, in elicited production, 6-year-

olds with SLI were outperformed by age-matched controls as well as
MLU-matched controls in terms of regular past tense forms (the
difference being primarily for infrequently inflected verbs). The children
204
Results: Tense
with SLI also overgeneralised regular past tense marking on irregular

verbs more than did either of the control groups.
Like Rice et al. (1995) and Rice and Wexler (1996), Rice, Wexler, and
Herschberger (1998:1421) reported that 5-year-olds with SLI were
outperformed by both the age-matched and younger (approximately 3-
year-old) MLU-matched controls in terms of spontaneous and probed
use of regular past tense marking, regular third-person singular marking,
be forms, and probed use of do forms. These forms were mastered by the
typically developing 5-year-olds (they used these morphemes in 90% or
more of the obligatory contexts). The 3-year-olds used these forms in 45
to 70% of the obligatory contexts and the children with SLI only in 25 to
48%. Similar results were obtained when these children with SLI were 6-
year-olds: They were again outperformed by the age-matched and
younger controls. The morphemes in question were mastered by the
typically developing children at 4 years of age, whereas the children with
SLI only mastered them after 7 years (Rice et al. 1998:1421).
Remaining with older children, Marchman, Wulfeck, and Weismer (1999)

and Marchman et al. (2004:209) found that, in terms of the elicited
production of regular past tense forms, children with SLI (mean age 8
years 9 months) performed similarly to age-matched controls; the most
frequently occurring error was zero-marking. In terms of number of
errors, the two groups also fared equally well on a task eliciting the
production of irregular past tense forms. However, the distribution of
errors across the two groups of children was not similar: The error type
made most frequently by both groups was suffixation; however, the SLI
group produced significantly more zero-marked forms than did the
typically developing controls. Whereas the pattern of (irregular) past
tense marking was the same for the younger and older members of the
control group (they had a preference for suffixation), the younger
children with SLI preferred zero-marking, whereas the older ones more
commonly made use of suffixation (Marchman et al. 2004:210).
Van der Lely and Ullman (1996) found that older children with SLI (aged
9 years 3 months to 12 years 10 months) produced significantly fewer
regular and irregular past tense forms than did language-matched
controls when presented with real and nonsense words. Unmarked
forms constituted 78% and 80% of their regular and irregular verbs.
205
Results: Tense
In contrast to these studies, and more like the results of the present one,
Conti-Ramsden and Hesketh (2003) found no significant difference
between the performance of the following two groups of children on a
past tense elicitation task: children with SLI (aged 4 years 4 months to 5
years 10 months) and 2- to 3-year-old controls (matched for language age
of the children with SLI). However, one of the authors’ conclusions is
that their task (very similar to the one used in the present study) was too
difficult for the young children and, therefore, possibly did not obtain
optimum data from them (Conti-Ramsden and Hesketh 2003:259). Note,
however, that Conti-Ramsden (2003:1032) found that these children
with SLI performed significantly more poorly than age-matched controls
on this task.
These general findings (children with SLI faring more poorly than age-
and/or MLU-matched controls) have been replicated for children who
speak languages other than English as well. Hansson and Leonard (2003)
examined the elicited use of verb morphology by Swedish-speaking 4- to
5-year-olds with SLI. They found that these children used present tense
inflection and irregular past tense forms as proficiently as did age-
matched and younger, MLU-matched controls. However, the children
with SLI were outperformed by the other groups in terms of use of
copulas and regular past tense inflections. These findings replicated
those of Hansson, Nettelbladt, and Leonard (2000), who made use of
spontaneous language samples. Note, however, that these findings
contrast with those of Håkansson (2001), who found no statistically
significant difference in terms of elicited production of tensed verb
forms between Swedish-speaking children with SLI (aged 4 years 0
months to 6 years 3 months) and controls matched on language age.
For Hebrew, Dromi et al. (1993:766) found that children with SLI used
significantly fewer past tense forms than did age-matched and also
(younger) MLU-matched controls in response to experimental items.
Bortolini et al. (2002), in search of a clinical marker for SLI in Italian,

found that 4- to 7-year-old children with SLI performed significantly
worse than age-matched controls on an elicitation task targeting the
production of present tense third-person plural marking on verbs.
206
Results: Tense
In conclusion, the results of the present study are supported in part by

those of previous studies on the elicited production of present or past
tense constructions by children with SLI.
7.4. TENSE PRODUCTION IN THE LANGUAGE

SAMPLES
The use of present and past tense constructions in the first 100 complete
and fully intelligible utterances of each language sample was examined.
The number of times the following were used correctly and incorrectly
were tallied: (i) present tense forms of main verbs, modal auxiliaries, have
forms, be forms, and hendiadyses; (ii) historic present tense forms; (iii)
past tense forms of modal auxiliaries, have forms, be forms, and
hendiadyses; (iv) past tense constructions consisting of the temporal het
and the past participial form (both the ge- form and the one resembling
the infinitive); and (v) passive constructions in the past tense form.
Tables 7.10 and 7.11 summarise the results. Table 7.11 contains the
statistics based on the mean scores of the informants. Note that not all
participants produced the constructions under investigation, which
explains why the degrees of freedom are smaller in some cases, for
example for hendiadyses.
Table 7.10. Present tense constructions used correctly and incorrectly in

SLI TD4 TD6 Total
Main verbs 458/469 370/378 511/513 1339/1360
97.7% 97.9% 99.6% 98.5%
Modals 243/259 337/340 282/284 862/883
93.8% 99.1% 99.3% 97.6%
Have forms 64/68 115/116 74/74 253/258
94.1% 99.1% 100% 98.1%
Be forms 305/343 332/341 319/321 956/1005
88.9% 97.4% 99.4% 95.1%
Hendiadyses 4/4 18/18 17/18 39/40
100% 100% 94.4% 97.5%
Total 1074/1143 1172/1193 1203/1210 3449/3546
94.0% 98.2% 99.4% 97.3%
Present tense verb forms were used with a high degree of accuracy
(88.9% or more) by all three groups. For all types of present tense forms
207
Results: Tense
combined, a significant difference between groups for the proportion of

accurate forms could be assumed (based on the outcomes of a one-way
ANOVA; F2,42=7.742; p=.001). Post hoc analyses (Tukey’s HSD; p=.05)
indicated these differences to be between the SLI group and the TD4
group, and between the SLI group and the TD6 group. However, no
differences could be assumed between groups in terms of the level of
accuracy with which main verbs were produced (one-way ANOVA;138
F2,42=2.435; p=.100); have forms were produced (F2,42=2.840; p=.07); or
hendiadyses were produced (F2,16=0.421; p=.663). Difference between
groups could, however, be assumed for the production of modal
auxiliaries (F2,42=3.625; p=.035) and be forms (F2,41=5.744; p=.006). In
the case of modal auxiliaries, post hoc analyses (Tukey’s HSD;
alpha=.05) did not reveal which groups differed from which, but as
regards be forms, the difference was between the two groups of 6-year-
olds. The results of the statistical analyses are summarised in table 7.11.
Table 7.11. Summary of performance per group – Production of present

tense forms in the language samples
CTa Grb N Mc SDd Min Max Levene ANOVA
score score F (df) p F (df) p
Main verbs SLI 15 .977 .028 .92 1.00 6.222 .004 2.435 .100
TD4 15 .979 .021 .95 1.00 (2,42) (2,42)
TD6 15 .993 .018 .94 1.00
Modals SLI 15 .919 .142 .60 1.00 15.359 .000 15.238 .035
TD4 15 .992 .016 .95 1.00 (2,42) (2,42)
TD6 15 .989 .028 .92 1.00
Have forms SLI 15 .873 .289 .00 1.00 15.330 .000 3.444 .070
TD4 15 .998 .007 .97 1.00 (2,42) (2,42)
TD6 15 1.00 .000 1.00 1.00
Be forms SLI 14 .904 .109 .64 1.00 1.606 .000 4.537 .006
TD4 15 .968 .065 .75 1.00 (2,41) (2,41)
TD6 15 .993 .017 .94 1.00
Hendiadyses SLI 2 1.00 .000 1.00 1.00 2.132 .151 .734 .663
TD4 7 1.00 .000 1.00 1.00 (2,16) (2,16)
TD6 10 .950 .158 .50 1.00
Total SLI 15 .940 .062 .80 1.00 14.421 .000 2.159 .001
TD4 15 .981 .024 .90 1.00 (2,42) (2,42)
TD6 15 .993 .009 .98 1.00
aCT=Type of past tense construction. bGr=group. cM=mean. dSD=standard deviation.
138 For correct present tense form divided by total number of present tense forms.
208
Results: Tense
For all three groups, most errors were ones of omission. However, other
types of errors also occurred. One made by one participant in each of
the TD4 and SLI groups was the inappropriate insertion of a be form, as
in example (109). Another was the inappropriate insertion of a main
verb, which yields a doubling pattern, as exemplified in (110).
(109) Target:
nou’s jy op die perdjie ry nou ry jy op die perdjie
now-be-CONTR you-SGL on the now ride you-SGL on the horsie
horsie ride
‘Now you are riding on the horsie’
(110) Target:
die’s al die mense wat kom by ons kom die’s al die mense wat by ons kom
kuier kuier
these-be-CONTR all the people who these-be-CONTR all the people who
come at us come visit at us come visit
‘These are all the people who are coming to visit us’
Other errors were only made by the participants with SLI. These include
(i) substituting a have form with a be one; (ii) substituting a be form with a
have one; (iii) substituting a be form with a modal auxiliary; (iv)
inappropriately inserting a modal auxiliary, as in (111), where a doubling
pattern is again yielded; and (v) using the incorrect form of the infinitive,
as in (112) and (113).
(111) Target:
gaan hulle hamers gaan nou kry gaan hulle hamers nou kry
will their hammers will now get will their hammers now get
‘Will now get their hammers’
(112) Target:
dit moet hier in die kas is dit moet hier in die kas wees
this must here in the cupboard is this must here in the cupboard be
‘This must be here in the cupboard’
(113) Target:
sy gaan ons nie nou weer te pla nie sy gaan ons nie nou weer pla nie
she will us not now again to bother not she will us not now again bother not
‘She will not bother us again now’
209
Results: Tense
All error types occurring on present tense verb forms in the first 100
utterances of the language samples also occurred in the remainder of the
30 minutes. Table 7.12 contains a summary of the occurrence of past
tense forms in the first 100 utterances of the language samples.
Table 7.12. Past tense constructions used correctly and incorrectly in the
language samples
SLI TD4 TD6 Total
Modals 0/0 1/1 16/16 17/17
--% 100% 100% 100%
Have forms 3/3 4/4 6/9 13/16
100% 100% 66.7% 81.3%
Be forms 10/11 14/21 54/55 78/87
90.9% 66.7% 98.2% 89.7%
Hendiadyses 1/1 2/2 3/3 6/6
100% 100% 100% 100%
Het ge- 47/59 73/75 121/122 241/256
79.7% 97.3% 99.2% 94.1%
Het but no ge- 2/3 2/2 9/9 13/14
66.7% 100% 100% 92.9%
Passive forms 2/3 2/3 6/6 10/12
66.7% 66.7% 100% 83.3%
Total 65/80 98/108 215/220 378/408
80.0% 90.7% 97.7% 92.6%
From table 7.12, it can be seen that the group with SLI used fewer past
tense forms than either of the two typically developing groups, but that
the performance of the latter two groups were also not similar: The 6-
year-olds used more than double the number of past tense constructions
than the 4-year-olds. Also, although the two typically developing groups
both fared well in terms of accuracy, the 6-year-olds still outperformed
the 4-year-olds. A one-way ANOVA returned a significant outcome,
which means that a difference between the groups in terms of accurate
production of all types of past tense constructions combined could be
assumed: F2,41=4.012; p=.026. Post hoc analyses (Tukey’s HSD; p=.05)
showed these differences to be between the SLI and TD6 groups.
In terms of specific types of past tense constructions, differences

between groups could be assumed for het ge- forms only (one-way
ANOVA; F2,40=5.921; p=.006). Post hoc analyses (Tukey’s HSD; p=.05)
210
Results: Tense
revealed that these differences were between the SLI and TD4 groups
and between the SLI and TD6 groups. The results of the statistical
analayses are summarised in table 7.13.
Table 7.13. Summary of performance per group – Production of past

tense forms in the language samples [results of statistical
anaysis left out when number of partipants producing relevant
data was too low]
CTa Grb N Mc SDd Min. Max. Levene ANOVA
score score F (df) p F (df) p
Modals SLI 0 . . . .
TD4 1 1.00 . 1.00 1.00
TD6 9 1.00 .000 1.00 1.00
Have forms SLI 1 1.00 . 1.00 1.00 8.174 .021 .551 .597
TD4 3 1.00 0.00 1.00 1.00 (1,8) (2,8)
TD6 7 7.61 .418 0.00 1.00
Be forms SLI 7 .964 .094 0.75 1.00 15.430 .000 2.704 .086
TD4 9 .777 .363 0.00 1.00 (2,26) (2,26)
TD6 13 .980 .069 0.75 1.00
Hendiady- SLI 1 1.00 . 1.00 1.00
ses TD4 2 1.00 .000 1.00 1.00
TD6 2 1.00 .000 1.00 1.00
Het ge- 13 .720 .421 0.00 1.00 28.243 .000 5.921 .006
15 .977 .061 0.80 1.00 (2,40) (2,40)
15 .996 .016 0.94 1.00
Het but no SLI 3 ..667 .577 0.00 1.00 19.600 .001 1.225 .350
ge- TD4 2 1.00 .000 1.00 1.00 (2,7) (2,7)
TD6 5 1.00 .000 1.00 1.00
Passsive SLI 3 .667 .577 0.00 1.00 22.250 .001 1.282 .329
forms TD4 2 .750 .354 0.50 1.00 (2,8) (2,8)
TD6 6 1.00 0.00 1.00 1.00
Total SLI 14 .777 .311 0.00 1.00 12.026 .000 4.012 .026
TD4 15 .925 .146 0.50 1.00 (2,41) (2,41)
TD6 15 .978 .047 0.83 1.00
aCT=Type of past tense construction. bGr=group. cM=mean. dSD=standard deviation.
As expected, the participants with SLI made the most errors of the three
groups. One reason that one might be tempted to offer for the low
number of past tense constructions in the language of children with SLI
is that they prefer to use the historic present form instead of the het ge-
form. However, an examination of the language samples revealed that
the group for which most historic present tense forms occurred was, in
fact, the TD6 one (38 times), despite the fact that this group also used
211
Results: Tense
more than double the number of het ge- forms than either of the other
two groups. Historic present tense forms occurred 21 times in the first
100 utterances of the 4-year-olds and only four times in those of the
participants with SLI.
In terms of types of errors, some were made only by children in the SLI
group. These were: (i) omission of the past participle, as in example
(114); (ii) omission of the ge- of the past participle, as in (115); and (iii)
omission of temporal het, as in (116).
(114) Target:
hulle altwee het op ’n blou bed hulle altwee het op ’n blou bed
geslaap/gelê
they both did on a blue bed they both did on a blue bed sleep/lie-
PAST PART
‘They both slept/lay on a blue bed’
(115) Target:
soom haar kou het so haar gekou het
so her chew did so her chew-PAST PART did
‘Chewed her like this’
(116) Target:
hulle seergekry hulle het seergekry
they sore-get-PAST PART they did sore-get-PAST PART
‘They got hurt’
In the remainder of the 30 minutes, the past tense of have was produced
as gehê het instead of as gehad het, but only in the typically developing 6-
year-old group. The following two utterances in the remainder of the 30
minutes were the only ones of their kind. They are given here for the
sake of interest. Utterance (117) was produced by a girl with SLI and
(118) by a 4-year-old girl.
(117) Target:
want hulle het al paar keer shock het want hulle het al ’n paar keer geshock
because they did already few time because they did already few time
shock did shock-PAST PART
‘Because they have already shocked themselves a few times’
212
Results: Tense
(118) Target:
het jy nie gehê sy moet skree nie? wou jy nie hê sy moet skree nie
did you not have-PAST PART she want-to-PAST you-SGL not have-INF
must scream not she must scream not
or
wou jy nie gehad het sy moet skree
nie
want-to-PAST you-SGL not have-PAST
PART she must scream not
‘Did you not want her to cry?’
7.5. DISCUSSION: SPONTANEOUS DATA ON TENSE

PRODUCTION
In the present study, all present tense verb forms were used correctly by
all three groups to a high degree. In terms of correct production of past
tense forms, the participants with SLI fared worse than the other two
groups overall, but still demonstrated relatively high levels of accuracy
(on average, almost 80% or more). This is in contrast to what Oetting
and Horohov (1997) found for their 6-year-olds with SLI. These
children’s levels of accuracy in terms of (regular) past tense marking in
their spontaneous language use was only 72%, with the age-matched
controls faring better but the language-matched ones worse. Beverly and
Williams (2004) also found that English-speaking children with SLI with
an MLU of less than 3 produced the present tense forms of be
significantly more in obligatory contexts than did MLU-matched
controls in spontaneous language samples. By contrast, the 4-year-olds in
the present study still fared better than the children with SLI, in terms of
the production of het ge- forms. This is in accord with what Balason and
Dollaghan (2002) found. They studied the use of 14 grammatical
morphemes in the spontaneous language samples of 100 typically
developing 4-year-olds and also saw high levels of accuracy. Regular past
tense occurred in 95% of obligatory contexts (compared to 97.3% in the
present study), and irregular past tense in 87%. Regarding contractible
and uncontractible copula be, the percentage of occurrence in obligatory
contexts was 88% and 93%, respectively; that of contractible auxiliary be
was 85%. In the present study, be forms occurred correctly in 97.4% of
the times (in present tense form) in the spontaneous language of the 4-
year-olds; the only single error was one of insertion into an inappropriate
context.
213
Results: Tense
The results of the spontaneous language production task lend support to

that of Paradis and Crago (2000), who also found that children with SLI
are outperformed by both control groups. They determined that French-
speaking 7-year-olds with SLI used finite verbs in obligatory contexts in
a language sample to a high degree (88-89%), but still performed
significantly worse than age-matched controls (89-99.5%). As regards the
correct choice of tense (present, past, of future) in obligatory contexts,
the results of the children with SLI did not differ significantly from those
of the control group for present tense, but the children with SLI did fare
significantly more poorly for past and future tense. Paradis and Crago
(2001:263) found that these same children with SLI were also
outperformed in terms of production of past tense by MLU-matched (3-
year-old) controls.
Other researchers who found that children with SLI are outperformed
by MLU-matched controls include Loeb and Leonard (1991). They made
use of spontaneous language production supplemented, where necessary,
by responses to probes, and found that children with SLI (4 years 0
months to 5 years 0 months) made more limited use of copula is,
auxiliary is, and the third person singular –s than did MLU-matched
controls (2 years 11 months to 3 years 4 months).
Hansson and Nettelbladt (1995) found that 5-year-old Swedish-speaking

children with SLI made more errors related to verbs (i.e., errors of tense,
on auxiliaries, and on modals) in spontaneous language use than did
MLU-matched controls. Interestingly – and not totally unlike the case in
the present study – not one of the six aspects related to verb
morphology rendered a significant difference between the two groups
when studied on its own, but when all six were combined, the verb-
related errors made by the group with SLI were significantly more than
those made by the MLU-matched controls.
However, other researchers have found that children with SLI fared
similarly to younger, MLU-matched controls. Rom and Leonard (1990)
compared the language samples of 4- to 5-year-old Hebrew-speaking
children with SLI to those of younger typically developing children
matched on a morpheme-per-utterance measure. They found no
difference in present tense inflection and also not in past tense inflection.
Hansson (1997) compared the verb usage in spontaneous language
214
Results: Tense
samples of Swedish-speaking children with SLI to that of younger,

MLU-matched controls. She found that the children with SLI used a
higher proportion of lexical verbs, as opposed to copulas, modals and
the auxiliary (Hansson 1997:209) − the latter two were frequently
omitted from obligatory contexts. These children also tended to use a
higher proportion of non-finite verb forms than their MLU-matched
controls. However, simple verb forms (in other words, present and past
tense forms) were used equally well by the two groups.
Blake et al. (2004) found that there was no significant difference between
the correct use of regular past tense forms by children with SLI (aged 5
years 1 month to 9 years 8 months) and that of language-matched
controls (matched in terms of expressive language score) in spontaneous
language samples. They found the same for the children with SLI and
age-matched controls. However, in terms of irregular past tense forms,
the SLI group fared significantly worse than the age-matched controls
(but not worse than the language-matched ones). As a group, the
children with SLI overregularised a greater number of different irregular
verbs than did either control group.
Roberts and Rescorla (1995) compared the spontaneous production of

verb morphology of 4-year-olds with an expressive language delay to that
of age-matched and MLU-matched controls. They found significant
differences between the production of infinitival forms, irregular past
tense forms, auxiliaries (including modals), do auxiliaries, and copulas by
the children with SLI and the age-matched controls, with the children
with SLI faring worse (as expected); no such difference was found
between the children with SLI and their MLU-matched controls. For all
three groups, errors of addition were the least frequent (with no
difference between groups as regards this type of error). Substitutions
occurred more frequently (here the children with SLI and the MLU-
matched controls performed similarly and were outperformed by the
age-matched controls). The most frequently occurring error type was
that of omission: Again, there was no statistically significant difference
between the frequency of occurrence of this error type when comparing
the SLI and MLU-matched groups, but these two were outperformed by
the age-matched controls.
215
Results: Tense
In the literature, there is no consensus on whether children with SLI are

outperformed by their typically developing peers and by younger
typically developing children. The general finding of the present study
was that the 6-year-olds with SLI were indeed outperformed by the
typically developing 6-year-olds but not by the typically developing 4-
year-olds, in terms of comprehension and elicited production of tense.
However, in terms of spontaneous production, both typically developing
groups fared better than the children with SLI.
These results render the following answer to research questions 1 to 4

(cf. p. 3): Afrikaans-speaking children with SLI present with a delay in
their comprehension and production of grammatical morphemes related
to tense, but these children also make significantly more highly
idiosyncratic errors than younger typically developing Afrikaans-speaking
children. Therefore, it appears that the language of Afrikaans-speaking
children is delayed but also somewhat deviant.
216
Characteristics of SLI in Afrikaans
Chapter 8
What are the linguistic characteristics of Afrikaans-

speaking children with SLI?
8.1. INTRODUCTION
This chapter gives an overview of the errors made by the language-

impaired and typically developing Afrikaans-speaking children. In section
8.2, the performance of the three groups of participants across
experimental tasks is compared, and the performance of some of the
individual children is discussed. In addition, in section 8.3.1, the error
types found in the spontaneous language samples are compared across
groups. This enables one to observe general response patterns by the
three groups of participants. Furthermore, in section 8.3.2, additional
information on the analyses of the language samples is presented: Errors
not necessarily related to the grammatical features number, person, case,
and tense (and therefore not discussed in the previous three chapters) are
discussed here. These include word order errors: The types of word
order errors made as well as those not made by the three groups of
participants are considered. In section 8.4, the results of discriminant
analysis are presented, where it is shown that performance on a selection
of experimental tasks succeeds better in classifying the 45 participants
correctly into the three groups (SLI, typically developing 4-year-old,
typically developing 6-year-old) than does a selection of measures from
the language samples. Also in this section, the issue of a possible clinical
marker of SLI in Afrikaans is discussed.
8.2. THE LINGUISTIC CHARACTERISTICS REVEALED

BY THE EXPERIMENTAL TASKS
In total, 15 experimental tasks were performed, each of them assessing

the comprehension or production of the singular/plural distinction,
pronouns, or tense. An overview of these tasks is presented in table 8.1.
An indication is given of whether or not there was a statistically
significant difference between (i) the performance of the three groups;
217
(ii) the mean scores of the SLI children and those of the typically
developing 4-year-olds; (iii) the mean scores of the SLI children and
those of their typically developing same-aged peers; (iv) the mean scores
of the two typically developing groups; and (v) the degree of variance
occurring in the three groups of participants.
Table 8.1. Overview of the difference in results between the three groups
of participants on the 15 experimental tasks
Task Experimental task Difference between
no. groups
variance across
Difference in
TD4, TD6
SLI, TD4
SLI, TD6
Overall
groups
1 PicSela: Sgl/pl Yes No Yes Yes Yes
2 Jdgmb: RWc, incorrect regd pl Yes No Yes Yes No
3a Jdgm: RW, incorrect irrege pl Yes No Yes Yes Yes
3b Jdgm: RW, correct irreg pl No -- f -- -- No
4a Jdgm: NWg, incorrect irreg pl No -- -- -- No
4b Jdgm: NW, correct irreg pl No -- -- -- No
5 SenComh: RW, reg pl Yes No Yes Yes Yes
6 SenCom: RW, irreg pl Yes No Yes Yes Yes
7 SenCom: NW, pl Yes No Yes Yes No
8 PicSel: Pronouns Yes No Yes Yes Yes
9 Jdgm: Pronouns Yes No Yes Yes No
10 SenCom: Pronouns Yes No Yes Yes No
11 PicSel: se-constructions Excluded due to low reliability of
items (Cronbach’s alpha .327)
12 SenCom: se-constructions No -- -- -- Yes
13a PicSel: Past tense of be and No -- -- -- No
have
13b PicSel: Past tense forms with Yes No Yes Yes No
het
14 Jdgm: Hendiadyses No -- -- -- No
15a SenCom: Targeted past tense Yes No Yes Yes No
constructions (unprompted)
15b SenCom: Total grammatical Yes No Yes Yes No
past tense constructions
(unprompted)
15c SenCom: Targeted past tense Yes ?i ? ? No
218
constructions (prompted)
15d SenCom: Total grammatical Yes No Yes Yes No
past tense constructions
(prompted)
15e SenCom: historic present No -- -- -- No
constructions equivalent to
target (unprompted)
15f SenCom: Total grammatical No -- -- -- No
historic present constructions
(unprompted)
15g SenCom: highly idiosyncratic Yes No Yes Yes Yes
errors related to past tense
constructions
15h SenCom: Het without ge- No -- -- -- Yes
15i SenCom: Total errors related No -- -- -- Yes
to past tense constructions
(before prompting)
15j SenCom: Total errors related Yes No Yes No No
to past tense constructions
(after prompting)
aPicSel=picture selection task. bJdgm=judgement task. cRW=real words. dReg=regular.
eIrreg=irregular. fBecause there were no statistically significant differences between the
three groups, post hoc analyses were not considered. gNW=nonsense words.
hSenCom=sentence completion task. iAlthough there was a statistically significant
difference between the three groups, this difference was not strong enough to show up
in post hoc testing.
As can be seen from table 8.1, the children with SLI fared worse than
their typically developing peers on 15 of the 26 aspects measured by the
experimental tasks. For all of these 15, the children with SLI performed
on a par with the typically developing 4-year-olds. In addition, as could
also be seen from the box plots in chapters 5 to 7, the general pattern
was that, where the variability differed between the three groups, the SLI
group showed the most intra-group variability. This variance was
statistically significant for nine of the 26 aspects measured by the
experimental tasks.
In an attempt to establish whether there was one (or more) general

factor responsible for the differentiation among the three groups, factor
analysis (principal component, varimax rotation) was performed. Several
solutions were considered, including solutions where the many scores
219
obtained on the sentence completion task assessing production of past

tense constructions were not taken into consideration. In all the different
solutions, the first factor after (varimax) rotation turned out to be stable
with consistent high loadings (>.50) for the following seven
experimental task:
(i) The picture selection task assessing comprehension of the
singular/plural distinction.
(ii) The sentence completion task assessing production of regular
plural forms of real words.
(iii) The sentence completion task assessing production of irregular
(iv) The sentence completion task assessing production of plural
forms of nonsense words.
(v) The picture selection task assessing comprehension of pronouns.
(vi) The judgement task assessing comprehension of pronouns.
(vii) The sentence completion task assessing production of pronouns.
When considering the composite score on these seven tasks (by adding
the z scores), the difference between that of the three groups can be
portrayed as in figure 8.1: The SLI and TD4 groups appeared to perform
similarly and the TD6 group better than the other two groups. Most
variability seemed to occur in the SLI group, with some children in this
group performing as well as their typically developing peers and others
worse than the 4-year-olds. However, the difference in variance between
the groups was not significant (Levene’s test; F2,42=2.007; p=.147).
Table 8.2 contains the details of the performance of the three groups on
the composite index. A one-way ANOVA returned a significant
outcome, which means that a difference between the mean scores of the
groups could be assumed (F2,42=30.662; p=.000). Post hoc analyses
(Tukey’s HSD; alpha=.05) revealed that the statistically significant
differences were between the SLI and TD4 groups, on the one hand, and
the TD6 group, on the other. There was no statistically significant
difference between the mean scores of the SLI and TD4 groups.
220
10.00
5.00
0.00
sum_zfac1
-5.00
-10.00
21
-15.00
SLI TD4 TD6

Group

seven experimental tasks

consisting of a selection of seven experimental tasks
SLI 15 -3.47 4.91243 -13.54 4.64
TD4 15 -2.57 2.40113 -6.45 2.25
TD6 15 6.04 3.24329 -1.51 9.34
Total 45 0.00 5.62460 -13.54 9.34
Considering the performance of the individual participants on this

composite index, two of the children with SLI fared markedly worse
than the rest of their group: One was a boy, participant 21 (SLI-6),139
whose composite score was -13.54; the other was a girl, participant 20
(SLI-5), whose score was -11.00. These scores were noticeably worse
than the worst one in the TD4 group, which was -6.45 – obtained by a
girl, participant 4 (TD4-4). Another two children with SLI performed
139Participant codes and additional information on participants can be found in table

4.1. Alternatively, consult table 8.8, in which codes for the participants with SLI are also
given.
221
markedly better than their group: Again, one was a boy, participant 26
(SLI-11), with a composite score of 4.64, and the other a girl, participant
25 (SLI-10), with a score of 3.83. These two scores were higher than the
lowest four in the TD6 group, illustrating the high degree of variability
found in the SLI group.
8.3. THE LINGUISTIC CHARACTERISTICS REVEALED

BY THE LANGUAGE SAMPLES
A 30-minute long language sample was collected from each participant.

The first 100 complete and fully intelligible utterances in each sample
were analysed for errors pertaining to correct and incorrect occurrences
of (i) singular and plural forms of nouns, (ii) pronouns, (iii) se-
constructions, and (iv) various types of past and present tense
constructions. The results of some of these analyses are presented in
section 8.3.1. However, the whole first 30 minutes of each language
sample was examined for errors other than those mentioned above, for
instance, for errors pertaining to word order or the inappropriate
insertion or omission of a determiner. The results of this examination are
given in section 8.3.2, together with an indication of the type of errors –
specifically those related to word order – which did not occur.
8.3.1. Error analysis of the first 100 utterances of each language

sample
The language samples were examined for the correct occurrence and the
substitution, incorrect insertion, and omission of those aspects assessed
by the experimental tasks. Table 8.3 gives an overview of a selection of
those measures which produced differences between the groups,
specifically (i) the proportion of plural forms which were produced
correctly; (ii) the proportion of pronouns produced correctly; (iii) the
proportion of present tense constructions produced correctly; and (iv)
the number of past tense forms vs. present tense forms.
In general, the SLI group fared worse than the TD6 group. However, in
contrast to the pattern found for the experimental tasks, the SLI group
was also, at times, outperformed by the TD4 group. The two typically
developing groups fared similarly. Again, the most variability was found
222
in the SLI group, with some children faring as well as the typically
developing ones.
Table 8.3. Overview of some of the measures of the language sample

analysis
Measure Difference between groups Difference in
Over- SLI, SLI, TD4, variance
all TD4 TD6 TD6 across groups
Propa correct plurals Yes No Yes No Yes
Prop correct pronouns Yes No Yes No Yes
Prop correct present tense Yes Yes Yes No Yes
Present tense vs. past tense Yes No Yes Yes No
aProp=proportion
At times, there was no score for a particular child for a certain measure,
simply because the child did not attempt the construction in question.
Despite the challenge posed by low frequency of occurrence (or absence)
of some of the measures, it was possible to establish that there was a
correlation between the performance of the groups on the four measures
given in table 8.3. The significant correlations (2-tailed) as well the others
are indicated in Table 8.4.
Table 8.4. Pearson’s correlation between measures of the language sample

Measure Propa plural Prop Prop present Past vs.
pronouns present
Prop plural -- .216 .562 .330
p=154 p=.000 p=.027
Prop pronouns .216 -- .426 .215
p=154 p=.004 p=.157
Prop present .562 .426 -- .236
p=.000 p=.004 p=.118
Past vs. present .330 .215 .236 --
p=.027 p=.157 p=.118
aProp=proportion
The difference between the three groups in terms of their composite

scores on these four measures (by summing their z scores) is portrayed
in figure 8.2. Unlike the case for the composite score of the experimental
tasks, the SLI group appeared to fare worse than both typically
developing groups, with the latter two performing similarly. Again, most
variability appeared to occur in the SLI group, with some children in this
223
group performing better than the best-performing, and others worse

than the worst-performing, typically developing ones. In this case, the
difference in variance between the groups was significant (Levene’s test;
F2,42=9.311; p=.000).
2.50
0.00
tot_corpus
-2.50
-5.00
-7.50
SLI TD4 TD6

Group
Figure 8.2. Box plot of performance per group – Composite score of four
measures of the language sample analysis
In table 8.5, the details of the performance of the three groups on the
composite index for the language sample analysis are given. A one-way
ANOVA returned a significant outcome, indicating that a difference
between the mean scores of the groups could be assumed (F2,42=4.268;
p=.021). Post hoc analyses (Tukey’s HSD; alpha=.05) revealed that the
statistically significant differences were between the SLI group, on the
one hand, and the two typically developing ones, on the other. Based on
the outcome of a one-way ANOVA, no significant difference between
the TD4 and TD6 groups could be assumed. This pattern differs from
the one for the composite score of the experimental tasks: There, the
SLI children fared similarly to the TD4 children.
224

consisting of four measures of the language sample analysis
SLI 15 -1.26 3.03358 -6.56 3.04
TD4 15 0.66 1.63232 -3.83 2.36
TD6 15 0.60 0.88641 -1.05 2.48
Total 45 0.00 2.20091 -6.56 3.04
8.3.2. Error analysis of the full first 30 minutes of each language

sample – other errors
From the above, it appears that the Afrikaans-speaking children with SLI
fared on a par with the younger typically developing ones on the
experimental tasks, but worse than both groups of typically developing
children in terms of correct spontaneous production of the grammatical
morphemes related to number, person, case, and tense. In this section,
other errors made in the language sample are discussed. The first set of
errors is verb-related. A summary of the errors and their frequency of
occurrence in the first 30 minutes of the language samples are given in
table 8.6, with illustrative examples following the table.
Table 8.6. Frequency of verb-related errors in the language samples of

the three groups of participants
Error type Error made by
SLI TD4 TD6
Errors involving infinitives 29a [12]b 14 [10] 11 [6]
Omission/insertion of main verbs 15 [6] 6 [5] 1 [1]
Omission/insertion of main het 2 [2] 0 [0] 0 [0]
Omission of verb particle 12 [6] 2 [1] 2 [2]
Other verb-related errors (difficult to 18 [7] 0 [0] 0 [0]
classify /idiosyncratic)
aThis figure indicates the number of times the error occurred in the 30 minutes of
language sample. bThe figure in square brackets indicates how many children in that
group made the error.
Errors on infinitives included the incorrect form of het ‘have’ and wees
‘be’; the omission of an infinitive, as in example (119); the omission of a
part of the infinitival structure, as in (120), where the te of om te probeer
swem ‘to try swim’ has been omitted; and the inappropriate insertion of a
225
part of the infinitival structure, as in (121), where om should not have

occurred.
(119) Target:
hoe kan hy ons goeters? hoe kan hy ons goeters steel?
how can he our stuff how can he our stuff steal
‘How can he steal our stuff?’
(120) Target:
ek het alles gedoen om probeer ek het alles gedoen om te probeer
swem swem
I did everything do-PAST PART I did everything do-PAST PART
infinitive-complementiser try swim infinitive-complementiser to try swim
‘I did everything to try swim’
(121) Target:
mens hoef nie om te betaal nie mens hoef nie te betaal nie
one have-to not infinitive-complementiser one have-to not to pay not
to pay not
‘One does not have to pay’
Errors on infinitives were made by all three groups of children.

However, the number of errors made by the SLI group was more than
double that made by the TD4 group, although only two more children in
the SLI than in the TD4 group made this error. Of the 10 children in the
TD4 group who made this error, six made it only once and the
remaining four twice each. By contrast, of the 12 children with SLI who
made this error, only four made it only once; the others made it two to
three times each, with two children – participants 21 (SLI-6) and 29
(SLI-14) – each making five such errors.140
The second error type related to the inappropriate omission or insertion

of a main verb. In example (122), the verb gaan ‘go’ is omitted, and, in
(123), reën ‘rain’ occurs twice.
140 Table 8.8 contains the information on which child with SLI made which types of
error.
226
(122) Target:
ek saam ek gaan saam
I with I go with
‘I’m going with’
(123) Target:
nou reën hulle nat reën nou reën hulle nat
now rain they wet rain now rain they wet
‘Now they are getting wet in the rain’
As was the case for errors on infinitives, a comparable number of

children in the SLI and TD4 groups inserted or omitted a main verb, but
the number of errors made by the SLI group was more than double that
made by the TD4 group. The same pattern emerged as before: Of the
five children in the TD4 group, four made the error once only. Three of
the six children with SLI made the error only once and one made it
twice. However, one boy – participant 21 (SLI-6) – made it four times
and one girl – participant 27 (SLI-12) – seven times. Het ‘have’ as a main
verb was also omitted and inserted inappropriately, but only by the SLI
group and only twice: once each by two boys.
The omission of part of a particle-verb is illustrated in example (124),

where the op of oppas ‘look after’ has been omitted. This error was made
almost exclusively by the SLI group. Of the six children from this group
who made this error, most made it only once, but one child each made
the error twice, three times, and four times.
(124) Target:
so hy pas vir my boetie so hy pas vir my boetie op
so he look for my brother-DIM so he look for my brother-DIM after
‘So he looks after my brother’
Only the children with SLI made verb-related errors which were highly
idiosyncratic and/or difficult to classify. By nature, this category of
errors is a particularly diverse one. Examples (125) to (127) are aimed at
illustrating this diversity. In (125), the target construction could be either
a passive one – ek word deur ’n volstruis daar op my hand gepik ‘I am pecked
there by an ostrich on my hand’ – or an active one – ’n volstruis het my daar
op my hand gepik ‘an ostrich pecked me there on my hand’.
227
(125)
ek word ’n volstruis het daar op my hand gepik
I be-PASS-PRESENT a ostrich did there on my hand peck-PAST PART /PASS PART
The intended meaning of (126) and (127) is not clear. For this reason,
only a gloss, and not a target construction, is provided.
(126)
dan vat hy ’n kinders maak
then take he a children make
(127)
hy wil net luister tog wat het hy gesticker vat
he want-to just listen just what did he sticker-PAST PART take
Half of the 18 errors were made by one boy, participant 21 (SLI-6). Two
other boys made two and three errors each. For the remainder of the SLI
participants who made such an error, each made it only once in their 30
minute language sample.
The second set of errors to be considered here is non-verb-related.

These errors are summarised in table 8.7. Some illustrative examples are
provided below.
In terms of omitting the subject, a similar number of children in the SLI

and TD4 groups made this error, but the errors in the SLI group were
almost three times as many as those in the TD4 group. In the SLI group,
one boy – participant 21 (SLI-6) – was responsible for 12 of the 29
errors, two girls made three errors each, another five children made two
errors each, and one girl omitted the subject once.
Five of the six children with SLI who omitted the object did so only
once. The girl who made this error three times – participant 29 (SLI-14)
– did not omit the subject once.
Prepositions were omitted, inserted, and substituted with other

prepositions by all three groups of participants, but less so by the TD4
group than by the other two. The two 6-year-old groups had almost
the same number of children making this error, but, collectively, the
228
Table 8.7. Frequency of non-verb-related errors in the language samples

of the three groups of participants
Error type Error made by
SLI TD4 TD6
Omission of subject141 29a [9]b 10 [7] 1 [1]
Omission of object 8 [6] 2 [2] 0 [0]
Omission of single noun 5 [4] 0 [0] 0 [0]
Omission of complementiser 1 [1] 0 [0] 1 [1]
Omission/insertion/substitution of 36 [10] 12 [5] 19 [8]
prepositions
Omission/insertion of determiners 70 [13] 26 [8] 7 [5]
Omission/insertion of nie 10 [8] 1 [1] 2 [2]
Other omission 13 [10] 1 [1] 4 [2]
Other insertion 1 [1] 0 [0] 1 [1]
Other non-verb-related error 33 [9] 9 [8] 3 [3]
(difficult to classify /idiosyncratic)
aThis figure indicates the number of times the error occurred in the 30 minutes of
language sample. bThe figure in square brackets indicates how many children in that
group made the relevant error.
10 children in the SLI group made this error almost twice as often as did
the 8 children in the TD6 group. One boy and one girl – participants 26
(SLI-11) and 29 (SLI-14) – were responsible for seven and nine of the 36
errors, respectively. Two boys – participants 16 (SLI-1) and 21 (SLI-6) –
made five errors each, and the rest of the six children with SLI made one
or two errors each.
An example of the substitution of one preposition with another is *in die

plaas ‘in the farm’ instead of op die plaas ‘on the farm’. An example of the
inappropriate insertion of a preposition is given in (128).
141Due to the nature of the conversation – freeplay with frequent comments on the
objects present and the actions being performed with them – children from all three
groups at times made use of elliptical utterances, particularly ones from which the
subject was omitted. An example would be where a child says gaan nou hierdie een vat
‘going to take this one now’ while he reaches for another wooden block. These
subjectless utterances were not included here, not even those of the one boy with SLI –
participant 24 (SLI-9) – who had a very strong preference for such subjectless
utterances over ones containing a subject.
229
(128) Target:
dit werk nie so by hierso nie dit werk nie so hierso nie
it work not so by here not it work not so here not
‘It does not work that way here’
As was the case for prepositions, determiners were omitted and inserted
inappropriately by all three groups of participants, but more children in
the SLI than in the other two groups made this error, and the error was
made a disproportionately high number of times by the children with
SLI. Of the 70 errors made by this group in total, 23 were made by one
boy – participant 21 (SLI-6) – and another 11 by another boy –
participant 24 (SLI-9). A girl – participant 29 (SLI-14) – made nine; two
boys each made six – participants 16 (SLI-1) and 26 (SLI-11); two girls
made four each; and the other seven children made either one or two.
Only two children in the SLI group did not make any errors related to
determiners. Examples of the inappropriate insertion of determiners by
the children with SLI is *’n goeters ‘a stuff’ instead of goeters ‘stuff’, as well
as the one in (129).
(129) Target:
jy soek ook ’n ene? jy soek ook ene?
you look-for also a one you look-for also one
‘Do you also want one?’
Although the children who omitted or inserted nie inappropriately mostly

did so only once, far more children with SLI than typically developing
ones made this error. An example of the omission of nie is given is
(130),142 and one of inappropriate insertion is given in (131).
(130) Target:
ons het visse nie ons het nie visse nie
we have fish-PL not we have not fish-PL not
‘We do not have fish’
142 This utterance would not have been ungrammatical had the child meant “It’s fish we
don’t have (but all the other animals are here)”. However, this utterance was produced
in response to a question by the researcher: En het julle visse op die plaas, kinders? ‘And do
you have fish on the farm, children?’. There was no indication that the child meant to
say anything other than “No, we do not have fish”.
230
(131) Target:
hulle wil nie skoonmaak nie hier nie hulle wil nie hier skoonmaak nie
they want-to not clean-make not here not they want-to not here clean-make not
‘They do not want to clean here’
What is termed “other omission” in table 8.7 entailed the omission of

lexical items or phrases in which no clear pattern could be detected;
diverse errors involving omission were grouped together under this
rather uninformative label. More children with SLI than typically
developing ones made this type of error, although six of the 10 children
with SLI who did make this error did so only once. A boy and girl –
participants 21 (SLI-6) and 29 (SLI-14) – made three and two errors
each, respectively. Examples of such omissions are given in (132) to
(134) below.
(132) Target:
*kom die kos hier kom die kos
come the food here come the food
‘Here comes the food’
(133) Target:
ons nie kan eet nie kan ons nie groot as ons nie kan eet nie kan ons nie
kan word nie grootword nie
we not can eat not can we not big if we not can eat not can we not
can become not big-become not
‘If we cannot eat, we not cannot grow up’
(134) Target:
is ons toe toe is ons toe
be we closed then be we closed
‘Then we were closed’
In contrast to the difficult to classify and/or idiosyncratic verb-related

errors which were made by only the children with SLI, the non-verb-
related errors were made by all three groups of participants. However,
the SLI group made far more of them than did the two groups of
typically developing children. A total of nine such errors were made
collectively by eight children in the TD4 group, whereas nine children
with SLI made 33 in total. Of these, 11 were made by one boy,
participant 21 (SLI-6). A girl and boy – participants 24 (SLI-9) and 27
231
(SLI-12) – made five errors each, another boy – participant 16 (SLI-1) –

made three, and the other five made one or two such errors each. As was
said for the verb-related errors, this category of errors is highly diverse
by nature. A number of examples is given in (135) to (138), in an attempt
to illustrate the types of errors classified as highly idiosyncratic and/or
otherwise difficult to classify.
(135) Target:
ons het ons honde te né ’n binnehond ons honde is huishonde en
`n buitehond buitehonde
we have our dogs to hey a inside-dog our dogs be house-dogs and outside-
a outside-dog dogs
‘Our dogs are inside and outside dogs’
(136) Target:
’n rooietjie hoedjie a rooi hoedjie
a red-DIM hat-DIM a red hat-DIM
‘A red hat’
(137) Target:
daar gaan hy daai in ’n fiets in daar is ’n fiets in
there go/will he that in a bicycle in there be a bicycle in
‘There is a bicycle in there’
(138) Target:
nou gaan ek aan koffies nou gaan ek die koffie vat
now go/will I on coffees now will I the coffee take
‘I am going to take the coffee now’
Table 8.8 contains a summary of the errors made by the SLI group, other
than those discussed in the previous three chapters. In total, 13 of the 15
members of this group made such errors. The two girls who did not –
participants 17 (SLI-2) and 20 (SLI-5) – did, however, make errors
pertaining to person, number, case, or tense in their spontaneous
language production.
232
Table 8.8. Summary of the frequency of errors (excluding those related to

the grammatical features number, person, case and tense)
made by the SLI children
Error type Participant
16 18 19 21 22 23 24 25 26 27 28 29 30
SLI10
SLI12
SLI13
SLI14
SLI15
SLI11
SLI3
SLI4
SLI6
SLI7
SLI8
SLI9
SLI1
Infinitive 1 2 3 5 2 3 1 2 3 1 1 5
Main V 1 4 2 1 1 7
Main het 1 1
Prepositional V 1 4 2 1 3 1
Other V- 1 9 2 1 1 3 1
related
Omit S 2 1 3 12 2 2 2 2 3
Omit O 1 1 1 1 1 3
Omit N 1 1 2 1
Omit C 1
Preposition 5 2 5 1 2 1 7 2 9 2
Determiner 6 1 1 23 2 1 11 4 6 1 1 9 4
Nie 1 1 1 1 1 1 1 3
Omit other 1 1 3 1 1 1 1 1 1 2
Other insertion 1
Other non- 3 2 11 1 5 2 2 5 2
verb-related
8.3.3. Word order errors
The discussion now turns to the word order errors which occurred in the
language samples of the 45 participants. All three groups of participants
made word order errors, but not all types were made by all groups.
Only the typically developing 6-year-olds produced utterances in which

the subordinate conjunction omdat ‘because’ was treated as a co-ordinate
conjunction, similar to its synonym want ‘because’. Three such utterances
occurred, illustrated by example (139). However, no co-ordinate
conjunctions were followed by a subordinate word order.
233
(139) Target:
dis omdat ek tel nie die hoef op nie dis omdat ek nie die hoef optel nie
it-is-CONTR because I pick not the hoof it-is-CONTR because I not the hoof
up not up-pick not
‘It’s because I do not pick the hoof up’
Relative clauses with an incorrect (verb-second or SVO) surface word

order occurred in the language of all three groups of children. An
example of such an utterance is provided in (140). In (141), the same
incorrect word order is shown.
(140) Target:
en hierdie is sy mamma wat jy het en hierdie is sy mamma wat jy daarso
daarso het
and this is his mommy that you have and this is his mommy that you there
there have
‘And this is his mommy that you have there’
(141) Target:
dat hy kan sy fietsie ry dat hy sy fietsie kan ry
that he can his bicycle-DIM ride that he his bicycle-DIM can ride
‘That he can ride his bicycle’
Main clauses with an SOV surface word order (the order found in
embedded clauses) also occurred – as shown in example (142) – but only
in the language of two children with SLI. Only one instance of VSO
occurred, in the language of participant 21 (SLI-6); this utterance is given
in (143).
(142) Target:
hulle TV kyk hulle kyk TV
they TV watch they watch TV
‘They are watching TV’
(143) Target:
vryf hy die been en ’n pappa hy vryf die been van pappa
rub he the leg and a daddy he rub the leg of daddy
‘He is rubbing daddy’s leg’
234
The 4-year-olds as well as the children with SLI appeared to have

problems with adverb placement. Examples of utterances with the
incorrect word order in which adverbs occur, are (144) and (145).
(144) Target:
hierdie al goed al hierdie goed
these all stuff all these stuff
‘All these stuff’
(145) Target:
sy’s ’n dinasaur ook sy’s ook ’n dinasaur
she-be-CONTR a dinasaur as well she-be-CONTR also a dinasaur
‘Among other things, she is a dinasaur’ ‘She too is a dinasaur’
Word order errors related to the order of adverbs or AdvPs also

occurred, but not in the language of the 4-year-olds. Examples (146) and
(147) illustrate this type of error.
(146) Target:
mens staan langs die poot anders mens staan langs die poot anders kan
miskien kan hy op jou voet trap hy miskien op jou voet trap
one stand next-to the paw otherwise one stand next-to the paw otherwise
maybe can he on your foot step can he maybe on your foot step
‘One stands next-to the paw, otherwise he can maybe step on your foot’
(147) Target:
maar nou in die groter skool het ek maar nou het ek in die groter skool
begin begin
but now in the bigger school did I but now did I in the bigger school
start start
‘But now I’ve started big school’
A third type of error occurring in utterances containing adverbs was

related to the word order following fronted adverbs or AdvPs. Such
errors were found in the language of all three groups and are illustrated
in (148) and (149).
(148) Target:
laas jaar ek was by ’n ou plaas laas jaar was ek op ’n ou plaas
last year I be-PAST by a old farm last year be-PAST I on a old farm
‘Last year I was on an old farm’
235
(149) Target:
toe ons daar kom ek het nie eers toe ons daar kom het ek nie eers
geskrik vir hulle nie geskrik vir hulle nie
when we there come I did not even when we there come did I not even
get-a-fright-PAST PART for them not get-a-fright-PAST PART for them not
‘When we came there, I was not even frightened by them’
The children with SLI and the 4-year-olds made errors in the word order
of wh-questions. Examples are given in (150) and (151). The wh-element
was fronted, but subject-verb inversion did not take place. Utterances
with a SwhV or VwhS word order did not occur in the data. One
utterance, from the language sample of a girl with SLI, contained a wh-
question in which the subject and verb had the correct surface word
order, but in which the adverb occurred in the incorrect position. This
utterance is given in (152).
(150) Target:
watte dit is? wat is dit?
what this is what is this
‘What is this?’
(151) Target:
hoekom ding kan nie trap nie? hoekom kan die ding nie trap nie?
why thing can not pedal not why can the thing not pedal not
‘Why can the thing not pedal?’
(152) Target:
hoekom weer werk ons net so bietjie? hoekom werk ons weer net so
bietjie?
why again work we just such bit why work we again just such bit
‘Why do we again just work a little bit?’
Other word order errors, ones which are difficult to classify in terms of
misplaced elements, also occurred, mostly in the language of children
with SLI. Two examples are given here, in (153) and (154).
236
(153) Target:
en hulle meet om hulle op die lorrie en hulle meet hulle om op die lorrie
te gaan te gaan
and they measure infinitive-complemen- and they measure them infinitive-
tiser they on the truck to go complementiser on the truck to go
‘And they measure them to go onto the truck’
(154) Target:
ons babatjies ons by hier kan kies ons babatjies kan ons by hierdie kies
our baby-DIM-PL we by here can our baby-DIM-PL can we by there
choose choose
‘Our babies we can choose to match these’ [=we can choose figurines (ones
which match these pieces of toy furniture) to be our babies]
The 4-year-olds and the children with SLI made word order errors in
utterances containing particle-verbs, i.e., verbs consisting of a verbal
stem and a particle belonging to the category noun, preposition, or
adverb. Examples (155) and (156) contain such utterances.
(155) Target:
ek sal ry fiets ek sal fietsry
I will ride bicycle I will bicycle-ride
‘I will ride bicycle’
(156) Target:
daar val af die een daar val die een af
there fall off the one there fall the one off
‘There the one falls off’
It appears then that a range of word order errors were produced, but that
not all three groups produced all types of errors. Table 8.9 contains a
summary of the types of word order errors and the groups by which they
were made. As can be seen from this table, a word order error which was
unique to the SLI group was that of main clauses with a surface SOV or
VSO order.
237
Table 8.9. Summary of word order errors made per group

Made by
SLI TD4 TD6
Errors related to group group group
Treating omdat like want Yes
Relative clauses with SVO Yes Yes Yes
Main clauses with SOV Yes
Main clauses with VSO Yes, once
Adverb placement in utterance Yes Yes
Order of adverbs or AdvPs Yes Yes
Word order after fronted adverbs or AdvPs Yes Yes Yes
Wh-questions Yes Yes
Other, more difficult to classify Yes Yes
Verb-particle Yes Yes
8.4. DO THE ERRORS REVEAL A POSSIBLE CLINICAL

MARKER OF SLI IN AFRIKAANS?
Another approach, that of discriminant analysis, was taken to compare

the performance of the three groups of participants across experimental
tasks. The aim was to ascertain which combination of experimental tasks
would result in the most accurate classification of the 45 participants into
the three groups (SLI, TD4, or TD6). For these analyses, all
experimental tasks were included, using the stepwise procedure to
include and exclude the task results in the discriminant analysis
(probability F entry .05; removal .10). The combination of the following
three tasks was selected as the best to place the participant into one of
the three groups (SLI, TD4, or TD6):
(i) The picture selection task assessing comprehension of the
singular/plural distinction.
(ii) The judgement task involving what should have been regular
(iii) The number of errors – excluding highly idiosyncratic ones and
past tense constructions with het but where the past participial (ge-)
form was replaced by an infinitival one – after prompting by the
researcher, on the sentence completion task assessing production
of past tense forms.
238
Table 8.10 shows the results in terms of a classification table. This table
makes a distinction between the actual group membership and the
predicted group membership.
Table 8.10. Results of classification of participants into three groups based

on a selection of three experimental measures
Actual group Predicted group membership Total
SLI TD4 TD6
SLI 7 5 3 15
TD4 4 11 0 15
TD6 0 1 14 15
Total 11 17 17 45
This analysis classified almost all TD6 group members as belonging to

that group. The participant misclassified as a typically developing 4-year-
old was participant 33 (TD6-3), who, interestingly, had the second
highest MLU of all participants. Problems occurred in differentiating
between the members of the TD4 and SLI groups, as could be expected
considering the separate analyses of the experimental tasks. Nevertheless,
the majority of the TD4 group was classified as such, with only four
members of this group misclassified (all four of them as children with
SLI). The general pattern observed for the experimental tasks was that
the SLI and TD4 groups obtain similar average scores but that the range
of scores in the SLI group is larger than that in the TD4 group. This is
reflected in the difficulty that this discriminant analysis had with the
correct classification of the members of the SLI group: Only seven were
classified correctly, five were seen to be typically developing 4-year-olds
and another three to be typically developing 6-year-olds. This latter
misclassification (that of children with SLI as being typically developing)
is interesting from a clinical point of view, because these children with
SLI are classified as having no language problems. The SLI children
classified as belonging to the TD6 group are participants 25 (SLI-10), 26
(SLI-11), and 27 (SLI-12); the composite score of all three of them had a
positive value.
Discriminant analysis was also performed with the various measures of

the spontaneous language samples – but only with measures related to
the grammatical features number, person, case, and tense. As in the case
of the experimental tasks, the aim was to ascertain what combination of
239
measures would result in the most accurate classification of the 45

participants. Table 8.11 shows the results of the stepwise discriminant
analysis. A combination of the following two measures was selected as
the best to place the participant into one of the three groups (SLI, TD4,
or TD6):
(i) The proportion of correct present tense constructions out of all
present tense constructions.
(ii) The number of past tense forms vs. present tense forms.
Table 8.11. Results of classification of participants into three groups based

on a selection of two measures from the language samples
Actual group Predicted group membership Total
SLI TD4 TD6
SLI 8 5 2 15
TD4 1 10 4 15
TD6 0 3 12 15
Total 9 18 18 45
This analysis was slightly less successful at correctly classifying

participants in terms of their actual group membership than was the
similar analysis involving the experimental tasks. Based on the selected
two measures of the language sample, two thirds (30) of the 45
participants were classified correctly compared to 32 by the similar
analysis involving the experimental tasks.
The group most often misclassified was the SLI one: Seven of its
members were deemed to be typically developing. This again confirms
that most variability occurred in the SLI group: Some of its members
performed as well as typically developing 6-year-olds. The two
participants with SLI who were classified as belonging to the TD6 group
were not the same ones as those classified as such by the experimental
tasks. This emphasises the importance of using a combination of
spontaneous and elicited data when diagnosing SLI in Afrikaans-
speaking children.
One of the two participants classified as a TD6 group member by the

two measures of the language sample was participant 20 (SLI-5), who
obtained a composite score of 2.15. Recall that she was one of the two
children with SLI who fared worst on the composite score pertaining to
240
the experimental tasks, but that she was one of the two children who
made no errors in her language sample except those related to the
grammatical features number, person, case, and tense. The other child
was also a girl – participant 17 (SLI-2) – one who made almost no errors
in her spontaneous language sample (the other girl who does not feature
at all in table 8.8). Her MLU was one of the lower ones (it fell within the
bottom third of her group). This could lead one to think that she made
use of short utterances in an attempt to avoid problem structures and, by
doing so, increased the accuracy of her utterances. As mentioned by
Blake et al. (2004:31), the fact that children with SLI sometimes differ
from controls in terms of correct morphology when comparisons are
made based on elicited production but not when based on spontaneous
production could simply be due to avoidance, in their spontaneous
language use, of unfamiliar forms by children with SLI. This could be the
case for this girl with SLI. However, none of the four children with
MLUs lower than hers appeared to use these strategies. The other child
who fared poorly on the composite score pertaining to the experimental
tasks – participant 21 (SLI-6) – also fared worst on the composite
pertaining to measures of the spontaneous language sample: He obtained
a score of -6.56. Two other children also fared poorly: participant 22
(SLI-7), with a score of -6.13, and participant 26 (SLI-11), with a score of
-5.08.
Because the average scores of the TD4 and TD6 groups did not differ
significantly, it is understandable that some of these groups’ members
were classified as belonging to the other. What is of interest is that one
typically developing 4-year-old was classified as language-impaired. This
was participant 6 (TD4-6), whose MLU was also the second lowest of all
TD4 participants. However, based on the selection of seven
experimental tasks, her score was average compared to that of the rest of
the TD4 group.
Interestingly, the 6-year-old who had the lowest composite score on the
two measures of the language sample – participant 33 (TD6-3) – had the
second highest MLU of all participants. So, although she made more
errors than the rest of her group, she also produced longer utterances
than most of her group.
241
From the discriminant analysis and language sample analysis, it appears

that a combination of experimental and spontaneous data differentiates
successfully between children with and without SLI, to a great extent.
Considering only spontaneous production might lead to underdiagnosis,
because it is, at least in theory, possible for children with SLI to avoid
certain structures in their spontaneous language use. Elicited production
should therefore also be used when diagnosing an Afrikaans-speaking
child as SLI. As stated by Blake et al. (2004:38), differences between
spontaneous and elicited production tasks make it unlikely that a
morphological measure based on spontaneous speech alone will be
useful in diagnosing SLI (see also Bedore and Leonard 1998). Whereas
elicited production tasks pose their own special difficulties for children
with SLI, Blake et al. (2004:39) state that they may also be better at
detecting subtle deficits in older children with SLI.
Three of the seven measures discussed in this chapter are related to the
production of verbs. It appears then that one should consider the elicited
production of past tense forms and the spontaneous production of
present and past tense forms in the search for a clinical marker of SLI in
Afrikaans. According to Rice et al. (1998:1412), such a marker is “a
linguistic form, or principle that can be shown to be characteristic of
children with specific language impairment”. Rice and Wexler (1996)
identified finiteness, or tense marking, as a sensitive and specific clinical
marker of SLI in English (see also Marchman et al. 1999).
The number of highly idiosyncratic and/or difficult to classify verb-

related errors in spontaneous production appears to differentiate very
accurately between Afrikaans-speaking children with and without SLI.
However, including “difficult to classify /idiosyncratic errors” as part of
a clinical marker could be problematic in practice: This category is one of
exclusion rather than inclusion – in order to ascertain whether a child
made such an error, one would first have to establish what is meant by
“classifiable errors” before one will be able to deem any error “difficult
to classify”. The fact that difficult to classify errors are included when
considering a clinical marker of SLI in Afrikaans is not a problem per se
– Blake et al. (2004) also found such errors, which they called “odd”, to
be characteristic of the language of their participants with SLI. Rather,
the practicalities of classifying errors as difficult-to-classify are the
problem.
242
Bortolini et al. (2002:90-91) state that the notion ‘clinical marker’ can be
interpreted in two ways. The first is that the marker represents a clear
symptom of SLI and also a particular cause for this symptom. The
second, a weaker interpretation, is that the clinical marker is
representative of the symptom without assuming that the symptom
reflects a single cause. Conti-Ramsden and Hesketh (2003:252) argue for
a third interpretation, namely that a clinical marker (or risk marker)
represents a symptom, but that no assumption is made about whether
the marker reflects a single cause or that this symptom alone identifies
the disorder. “On the contrary, it is assumed that the risk marker is more
likely to be used in combination, to complement information available”
(Conti-Ramsden and Hesketh 2003:252).
If a composite consisting of the seven measures discussed above is taken

to be a clinical marker of SLI in Afrikaans, then ‘clinical marker’ should
here be given the interpretation of Conti-Ramsden and Hesketh: It may
be a useful risk marker when used together with other information that a
clinician has on the child, but it does not necessarily reflect a particular
cause for the symptom(s) which they represent.
In general, the Afrikaans-speaking children with SLI fared on a par with

typically developing 4-year-olds and worse than typically developing 6-
year-olds on experimental tasks assessing the comprehension and
production of grammatical morphemes related to the features number,
person, case, and tense. In terms of spontaneous production of
morphemes related to these grammatical features, the two typically
developing groups fared similarly, with the children with SLI being
outperformed by both. A similar pattern was observed for other errors
found in the spontaneous language samples. The general observation
regarding variability was that most variability occurred in the SLI group.
Discriminant analysis and language sample analysis revealed that a

combination of six measures would probably differentiate very
successfully between Afrikaans-speaking children with and without SLI.
A composite of these measures could possibly act as a clinical marker,
although further investigation is required in this regard; the composite
could possibly be simplified.
243
As yet, no explanation has been attempted as to why SLI presents itself

the way it does in Afrikaans. In chapter 9, we see that current linguistic
accounts do not offer adequate explanations for the way in which SLI
presents itself in Afrikaans. An alternative, more comprehensive account
of SLI as it presents itself in Afrikaans is therefore given, based on the
data obtained in the present study.
244
Accounts of SLI in Afrikaans
Chapter 9
Accounts of SLI as it presents itself in Afrikaans
9.1. INTRODUCTION
In chapter 2, the content of three linguistic accounts of SLI was

discussed, namely that of the ATOM, RDDR, and Feature Deficit
Hypothesis. In the present chapter, the merit of these accounts will be
evaluated and the predictions that they would make for SLI in Afrikaans
will be discussed. Then it will be shown that these three accounts of SLI
do not, in fact, account for the Afrikaans data presented in this study. An
alternative account of SLI as it presents itself in the data of this study is
then offered. In short, this account entails that Afrikaans-speaking
children with SLI seem to have an intact and well-functioning
computational system, but that they experience problems with mapping
of syntactic objects (functional morphemes and chains) onto sound
representations.
9.2. THE AGREEMENT/TENSE OMISSION MODEL
9.2.1. Merit of the ATOM
Wexler (1994:335) made two related claims that were subsequently

incorporated into the ATOM. The first is that a verb will move to TP if
there is a TP available in the derivation, and the second that the verb will
stay in its base-generated infinitival form if there is no TP. These claims
need to be reconsidered, however, for two reasons.
The first reason is that, in Minimalist syntax, the verb is assumed to enter
the Numeration in its inflected form; the verb does not receive its
inflection by merging with some grammatical morpheme during
derivation. For example, when the verb walked enters the Numeration, it
already has the [past] feature and the relevant phonological features
associated with the past tense suffix –ed. The tense feature of the verb is
semantically interpretable, and need not be checked and eliminated in
245
order to prevent the derivation from crashing (cf. Hornstein et al.

2005:295). By contrast, the tense feature of the head T of TP is assumed
to be semantically uninterpretable, which means that it has to be checked
and eliminated in the course of deriving the LF-representation of the
sentence (cf. Biberauer and Roberts 2005; Legate 2002:3).143 Thus, if
there is no TP, the verb will not move, neither overtly nor covertly. This
does not imply, however, that the verb will necessarily occur in its
infinitival form in its base-generated position: In this position, the verb
can have any one of the tense forms that it has available in the lexicon.
In other words, without a TP in the structure, the verb will only occur in
its infinitival form if it has entered the Numeration in that form.
In short then, a verb may occur in its finite form irrespective of whether
the structure contains a TP with the appropriate head T. Evidence for
the absence of TP should thus not focus on the inappropriate
occurrence of infinitival forms, but rather on the absence of overt verb
movement in languages in which such movement does occur. If there is
no TP, then the finite verb, when it does occur, simply cannot move to
such a category. Moreover, the fact that in some languages, such as
Afrikaans, the finite and infinitival form of the verb are not
distinguishable144 is also a reason for not focusing on the inappropriate
occurrence of infinitival forms as evidence for the absence of the TP (cf.
section 3.3.1.5).
The second reason for reconsidering Wexler’s (1994:335) proposal

concerns the fact that it does not clarify whether, and why, movement
takes place overtly or covertly: It is simply assumed that the verb moves
to the T if a TP is present in the derivation. Recall, however, that in
some languages, such as English, the verb does not move overtly out of
the VP before Spell-Out; it does move covertly, though, to check the T’s
143 If one views T as having the semantically interpretable feature, with V having the
uninterpretable one (which is possible in principle), then my argumentation against the
ATOM prediction regarding the form of unraised verbs does not hold. However, this is
generally not assumed, particularly not recently under the Probe-Goal system, where it
is crucial that the higher category has an unvalued feature if it is going to act as a probe.
144 The present tense and historic past tense forms of finite verbs are the same as the
infinitival form, e.g., stap ‘walk’ and om te stap ‘to walk’; the past tense form of finite
verbs are not, e.g., stap ‘walk’ and gestap ‘walk-PAST PART=walked’.
246
tense feature after Spell-Out145 (cf. section 3.4.3). In such languages then,
feature checking is postponed until after Spell-Out (cf. Hornstein et al.
2005:47). Suppose now that the verb is retrieved from the lexicon in its
infinitival form, as Wexler (1994:329) implies is the case for English.146
Suppose further that TP is indeed selected. If the verb is moved overtly
to T, it will be phonologically realised with the appropriate tense
morphology. By contrast, if the verb is covertly moved to T – which is
generally assumed to be the case in English – the verb will be
phonologically realised in its infinitival form. Therefore, the appropriate
or inappropriate form of the verb can be explained with reference to
whether movement takes place overtly or covertly. More specifically,
given the independently required distinction between overt and covert
movement, the occurrence of the inappropriate infinitival form of the
verb (in both typically developing children and those with SLI) can be
explained without reference to the absence/presence of TP. The
difference between these two groups of children could then be ascribed
to some overt “tense-lowering operation” – where the tense morpheme
and the verb merges in the “other” direction, by means of the tense
morpheme in I lowering onto the verb in V – which is acquired and
successfully used by typically developing children but not by children
with SLI. The merit of such an account would be that it does not require
a specific assumption about the presence/absence of TP; however, it
would require an assumption about tense lowering operations, and it
would face the well-known objections against lowering operations (cf.
Chomsky 1982b:55,256-7). Alternatively, Wexler and colleagues could
reformulate their proposal as follows for English, adopting the overall
presence of TP: T can be either (i) specified for tense (and an extended
projection principle (EPP)-feature), yielding tensed verbal forms after
covert raising of V to T; or (ii) unspecified/underspecified for tense
(while still having an EPP-feature), requiring the presence of an
infinitival verbal form (i.e., the default form). Nevertheless, as Wexler’s
proposal stands at present, it does not incorporate the distinction
145 In English, it is possible to say John often kisses Mary, but not *John kisses often Mary.
The verb kisses can thus not move overtly from the V to the T. However, in languages
such as French, overt verb movement from V to T occurs, as illustrated by the
following example, where the verb is embrasse: Jean embrasse souvent Marie.
146 According to Wexler (1994:329), where the verb appears without the –s (as in Mary
play baseball), it is the infinitive and not the verb stem which is produced.
247
between overt and covert movement, which raises questions about its
merit.147
A further criticism against the ATOM concerns the claim that one would
not find utterances such as Her walks in the language of children with
SLI, where subject-verb agreement and tense are indicated overtly, and
the subject pronoun has accusative case. If the subject is taken to
originate in the specifier position of the VP (or νP) and there is no TP
present in the derivation, then the subject cannot move to the TP. If the
subject does not move, one would expect DPs with any case to occur in
the subject position of the sentence, because nominal expressions (for
example, pronouns in English) are assumed to be retrieved from the
lexicon with their case. Because there is no TP to check whether the case
of the DP occurring in the subject position is, in fact, the correct case,
such DPs can occur in what Wexler calls the “default” case, which is
taken to be accusative for English. Because the case feature of the DP
can receive a phonetic interpretation,148 it follows that the occurrence of
a DP with the incorrect, accusative case will not cause the derivation to
crash at PF. In fact, examples such as Her walks do occur relatively
frequently in the language of children with SLI (cf. Pine, Rowland,
Lieven, and Theakston 2002).
Despite the above criticisms, 149 and although not explicitly stated as such
by Rice, Wexler, and colleagues, the ATOM – interpreted within a
147 Also see the comments in section 3.4.3 on the overt-covert distinction possibly
becoming a defunct one (cf. Hornstein et al. 2005:312).

148 It could, of course, also be that the subject does carry the correct abstract case
(NOM), but that this case is given the incorrect sound form, i.e., that an error occurs
with the mapping of the morphological information onto the phonological form.
149 I also take note of (i) the criticism by Rispoli (1999, 2002, 2005) against the ATOM;
(ii) Charest and Leonard’s (2004) indication that the proposals of the ATOM need to
be altered in order to account for their empirical findings; (iii) Joseph, Pine, and Conti-
Ramsden’s (2002) and Pine, Rowland, Lieven, and Theakston’s (2002) finding that there
are relatively frequent exceptions to the predictions of the ATOM; and (iv) Pine et al.’s
(2002:273) comment that the range of situations in which the predictions of the ATOM
can be tested is limited. Pine et al. (2004: 913) word the latter criticism more strongly:
“… the ATOM derives much of its power from the fact that it is actually extremely
difficult to test”. A related criticism is that the “invisibility” of agreement (and
subsequent assignment of nominative in the case of [-tense, +agr]) is unfalsifiable,
which could lead to the whole proposal being seen as somewhat stipulative. I
furthermore take note of Lin’s (2006) suggestion that an alternative account – one on
248
Minimalist framework – seems to account for the following

characteristics of SLI:150
(i) The use of inappropriate tense markers, as in Who carry her bag or
Which door did it creacked? (from Van der Lely 2004). If the tense
feature of the TP is not checked by that of the verb, then it is
possible for the verb to have the finite form where the infinitival
form would have been required and vice versa, the reason being
that the verb does not enter into a tense agreement relation with
TP.
(ii) The inappropriate use of accusative case on subject pronouns.
There are, however, characteristics of SLI for which the ATOM does
not seem to offer an account. The first is the lower percentage of use of
grammatical morphemes. As noted above, verbs are retrieved from the
lexicon with their grammatical morphology. The absence of a TP should
thus not influence the percentage of use of grammatical morphology,
whether on verbs or any other category.151
The ATOM can also not account for all word order problems in
question constructions. The tree diagrams in (157) and (158) indicate
that, even without a TP – i.e., with no T for the verb and no specifier
position of TP for the subject to move to – question constructions
should still demonstrate the correct surface word order for wh-question
constructions and yes/no ones containing a modal auxiliary.152 However,
as indicated in (159), modal-less yes/no-question constructions will not be
grammatical – for instance, *You wash the car? or *Wash you the car? should
in theory occur instead of Do you wash the car?. Seeing that there is no T
which nominative case is assigned by an interpretable mood feature on T – is adopted

to account for the findings that children with SLI mark tense better than they do
agreement.
150 See section 2.3 for a discussion of these characteristics.
151 Note that this criticism does not apply to Wexler’s original proposal, but to my
revised interpretation of the ATOM-model, which assumes that verbs enter syntax
together with their inflection.
152 One could also argue that there is indeed a TP present in (157) and (158). More
specifically, ModP might be taken to be identical to TP for the reason that modals in
English (unlike in languages such as Dutch and Afrikaans) are always [+tense]; i.e.,
there are no infinitival forms such as to will/to can/to must in English.
249
under which the “dummy” (or expletive) auxiliary do can initially occur,
yes/no-question constructions without do-support are to be expected.153
(157) CP
Spec C'
which car
C ModP
will
Mod vP
will
Spec v'
you
v VP
wash
V DP
wash which car
(158) CP
Spec C'
C ModP
will
Mod vP
will
Spec v'
you
v VP
wash
V DP
wash the car
153 See section 3.4.3 for a brief discussion on English question constructions requiring
do-support. Yes/no-question constructions in which auxiliaries such as have and be should
have occurred are also expected to be ungrammatical, as these auxiliaries are also seen
as elements which are merged in the T – and therefore do not have to move to the T
(cf. Lightfoot 1979, 1991, 1999; Roberts 1985, 1993, 2007).
250
(159a) (159b)
CP or CP
Spec C' Spec C'
C vP C vP
wash
Spec v' Spec v'
you you
v VP v VP
wash wash
V DP V DP
wash the car wash the car
The ATOM also does not offer an account for the problems in
establishing (non-)co-referential relationships and interpreting passive
constructions, but these fall outside of the scope of the ATOM.
9.2.2. Predictions of the ATOM for Afrikaans
Recall that, in Afrikaans, the present tense form of all verbs (with the
exception of hê ‘have’ and wees ‘be’) resembles the form of the infinitive,
that is, a “bare” stem without any realised tense/infinitival marker.154 For
example, loop ‘walk’ occurs in the same form in Hy loop elke dag ‘He walks
every day’ and Hy dink daaraan om te loop ‘He is thinking of walking’. In
sentences expressing present tense and containing one or more modal
auxiliaries, the tense is indicated on (the first of) these auxiliaries and not
on the main verb; the latter appears in the infinitival form. Only in
sentences expressing past tense does the phonological form of most
main verbs155 differ from that of the infinitival form; this difference
concerns the presence of the prefix ge- which is used to form the past
participial form. For example, the past participle of loop ‘walk’ is geloop, as
in Hy het gister geloop ‘He walked yesterday’. However, in such sentences,
specifically those not containing modal auxiliaries, past tense is not
indicated on the main verb, but on the (obligatory) temporal auxiliary het.
Where such sentences do contain modal auxiliaries, past tense is
See section 3.3.1.5 for the various phonological forms of the verbs hê and wees.
154
155That is, of those which do not contain the derivational prefixes be-, ge-, er-, her-, ont-
and ver-, or unstressed aan-, agter-, deur-, om-, onder-, oor-, and voor- (cf. Donaldson
1993:section 8.5.1).
251
indicated either on (one or more of) these modals156 or with the temporal
auxiliary het, in which case the modal auxiliary may occur in either its
present tense or its past tense form. For example, Hy wil geloop het, Hy wou
geloop het, and Hy wou loop, could all have the same temporal reference.
For Afrikaans then, the ATOM would predict that verbs may occur in
their infinitival form in the language of children with SLI. This
prediction does not seem to be linguistically significant, however: The
present tense verbs must also appear in a form resembling the infinitival
form in the language of both typically developing children and adult
speakers of Afrikaans. Therefore, this prediction of the ATOM is not
testable as far as present tense constructions in Afrikaans are concerned.
However, the claim made by the ATOM that verbs cannot always move
to check the tense feature of the TP (seeing that the TP is not always
present) is a potentially significant one. On this claim, the sentences of
Afrikaans-speaking children with SLI should have a grammatical word
order, if one assumes that (i) Afrikaans is SOV underlyingly, and (ii) the
subject is initially merged in the specifier position of vP. If there is no TP
to which the subject can move, then the subject will still move to the
specifier position of CP. The verb will move from the V position to C
(instead of moving from V to T and then from T to C), and the object
will remain in situ, arguably having its (accusative) case checked under
agreement with small v at LF (i.e., after the direct object has moved to
the specifier position of vP). As indicated in (160), this will still render a
grammatical surface SVO word order in matrix clauses. Also, in
embedded clauses, the (grammatical) surface SOV word order should
still be rendered, regardless of whether or not the TP is present. This is
illustrated in (161).
156 The modal then has a different form, e.g., moes ‘had to’ instead of moet ‘must’, or kon
‘could have’ instead of kan ‘can’.

252
(160) CP
Spec C'
hy
C vP
eet
Spec v'
hy
VP v
eet
DP V
piesangs eet
(161) CP
Spec C'
C vP
dat
Spec v'
hy
VP v
eet
DP V
piesangs eet
Also, Afrikaans yes/no-questions (both those containing a modal auxiliary

and those without a modal), as well as wh-questions, should demonstrate
the grammatical surface word order, regardless of the presence or
absence of the TP. This is illustrated in (162), (163), and (164),
respectively.
One prediction that the ATOM then makes is that the utterances of
Afrikaans-speaking children with SLI should demonstrate correct word
order in main clauses.
253
(162) CP
Spec C'
C vP
sal
Spec v'
hy
VP v
sal
DP V
dit sien
(163) CP
Spec C'
C vP
sien
Spec v'
hy
VP v
sien
DP V
dit sien
(164) CP
Spec C'
wie
C vP
sien
Spec v'
hy
VP v
sien
DP V
wie sien
Regarding subject pronoun case, it is unclear what the ATOM predicts

for the language of Afrikaans-speaking children with SLI. The reason for
this is that it is not obvious, on Schütze and Wexler’s (1996:671) analysis,
254
what the default case would be in Afrikaans.157 Schütze and Wexler

(1996:671) claim that English differs from a language like German, in the
sense that the default case in English is accusative, whereas it is
nominative in German. They present the following English (165) and
German (166) examples to illustrate this point.
(165a) Me/*I, I like beans

(165b) Who did it? – Me/*I
(165c) Me/*I too
(165d) It’s us/*we
(165e) What? Me/*I cheat on you? Never!
(166a) Was? Ich/*Mich dich betrügen? Nie!

‘What? I/Me cheat on you? Never!’
(166b) Der, den habe ich gesehen158
‘He, him I saw’
(166c) Lass mich dein/*deinen Liebhaber sein
‘Let me-ACC be your-NOM/your-ACC lover’
(166d) Der/?Dem Hans, mit dem spreche ich nicht mehr
‘Hans-NOM/DAT, with him I do not speak anymore’
The Afrikaans equivalents of the English examples in (165a-c) and (165e)

are given in (167) below. An Afrikaans equivalent of example (165d) is
not given – but an example with the first person singular is given in its
stead – as we and us have the same sound form in Afrikaans, namely ons.
(167a) *My/Ek, ek hou van boontjies

‘Me/I, I like beans’
(167b) Wie het dit gedoen? – *My/Ek
‘Who did it? – Me/I’
(167c) *My/Ek ook
‘Me/I too’
(167d) Dis ek/*my
‘It’s I/me’
157 In fact, it is not clear what should be understood under ‘default case’ in general
under Wexler’s analysis.

158 In this construction, hanging topic left dislocation occurs. However, a construction
with contrastive left dislocation, as shown in (i), is also acceptable.

(i) Den Hans, den kenne ich nicht
‘Hans-ACC, him I don’t know’
255
(167e) Wat? *My/Ek jou verkul? Nooit nie!

‘What? Me/I cheat on you? Never!’
On the one hand, it would appear from (167) that Afrikaans differs from
English in terms of default case. The default case in English is claimed to
be accusative, but the same cannot be claimed for Afrikaans. On the
other hand, Afrikaans also appears to differ from German, for which the
default case is claimed to be nominative. As regards example (166a),
Afrikaans and German are the same. However, Afrikaans resembles
English and not German as regards constructions similar to that given in
example (166b), as shown in (168).159,160
(168a) *Ek/My, my het hy geslaan

‘I/Me, me he hit’
(168b) *Sy/Haar, haar het hy gesien
‘She/Her, her he saw’
When considering the examples in (167) and (168), it appears that the
default case in Afrikaans is not obviously either nominative or
accusative. However, one could argue that the left-most ek/my in (168a)
and sy/haar in (168b) are fronted, with the second instance being a
repetition; i.e., the ek/my and sy/haar are not base-generated in a left
peripheral topic position.161 If this is the case, then the examples in (168)
should receive less emphasis when considering default case than those in
159 The Afrikaans equivalents of examples (166c-d) do not provide evidence as to the
default case of Afrikaans: (166c) translates as Laat my jou liefling wees, where the second-
person pronoun will be jou regardless of its case. However, a sentence such as Laat ek
jy/*jou wees en jy ek/*my ‘Let me be you and you be me’ clearly shows that predicate
nominals bear nominative case in Afrikaans. In (166d), Afrikaans would not permit an
article before the proper noun, and even if it did, the form of the article would be die,
regardless of the case of the DP die Hans.
160 As mentioned in note 158, it should be possible to have a contrastive left dislocated
constituent with accusative case in German. Mich, mich hat er geschlagen ‘Me, me he hit’
should therefore also be acceptable (similar to Mij, mij heeft ie geslagen being acceptable in
Dutch). If this is the case, then Afrikaans resembles both English and German in this
respect.
161 Compare English wh-question constructions such as Who (yes) who did it?, Dutch ones
such as Wie (ja) wie heeft het gedaan?, and Afrikaans ones such as Wie (ja) wie het dit gedoen?,
where the left-most wh-element is not base-generated in this left-most position but is
rather a repeated wh-element.
256
(167), in which case the default case in Afrikaans, based on (167), could
be taken to be nominative.
In short, it is not entirely clear what the ATOM would predict in terms
of subject pronoun case in the language of Afrikaans-speaking children
with SLI. However, recall the above criticism against the claim that
utterances such as Her walks – with the subject pronoun in the accusative
case – should not occur in the language of children with SLI: Seeing that
there is nothing against which the case of the retrieved pronoun can be
checked, subject pronouns should be able to enter the Numeration with
any and all case features available in the grammar. The prediction is then
that utterances such as those in (169) should occur in the language of
Afrikaans-speaking children with SLI. However, the case feature on the
pronoun is uninterpretable and thus in need of checking in order to keep
the derivation from crashing. It is therefore not clear why a derivation
rendering *Hom/*Sy sien my would, in fact, converge.162
(169a) Hy sien my
He-NOM see me
‘He sees me’
(169b) *Hom sien my
Him-ACC see me
‘Him sees me’
(169c) *Sy sien my
His-GEN see me
‘His sees me’
Finally, the ATOM makes no prediction regarding the production and

comprehension of (non-)co-referential relationships and passive
constructions; these aspects fall outside of the scope of the ATOM.
Hence, it is not possible to predict on the basis of this proposal whether
Afrikaans-speaking children with SLI will experience difficulties when
compared to their typically developing peers.
162 That is, given the ATOM, it is not clear. However, an explanation for this is given in
section 9.6.3.
257
9.2.3. Are the predictions of the ATOM for Afrikaans borne out?
As stated above, the ATOM would predict that Afrikaans verbs may
occur in their infinitival form in the language of children with SLI. As
these verbs must also appear in a form resembling the infinitival one in
the language of both typically developing children and adult speakers,
this prediction was judged to be linguistically insignificant. The only
present tense verb forms which may potentially have been interesting,
were those of have and be, for which the infinitival form and the inflected
form differ. Unlike what was predicted by the ATOM, the Afrikaans-
speaking children with SLI did not produce utterances such as (170) and
(171), where have and be occur in their infinitival form. Interestingly, one
girl with SLI did the reverse: She used the inflected form of be (i.e., is)
instead of the infinitival one (wees), as illustrated in (172).163
(170) Target:
sy hierdie mannetjie hê sy het hierdie mannetjie
she this figurine have-INF she have this figurine
‘She has this figurine’
(171) Target:
hy hier wees hy is hier
he here be-INF he be here
‘He is here’
(172) Target:
moet daar nog ’n wit ding in is daar moet nog ’n wit ding in wees
must there still a white thing in is there must still a white thing in be
‘There must still be a white thing in there’
A second possible prediction that the ATOM made for Afrikaans was
that utterances of Afrikaans-speaking children with SLI will demonstrate
the correct surface word order, under the assumption that XPs can move
to the specifier position of CP and V can move to C in main clauses.
This was not always the case. Simple declaratives with SOV and
embedded sentences with SVO occurred, as shown in (173) and (174),
respectively.
163This girl did not produce wees at all in the 30 minute language sample. She did
produce grammatical utterances containing is; however, she also at times omitted is
from obligatory contexts.
258
(173) Target:
hy vir ons jok hy jok vir ons
he for us lie he lie to us
‘He is lying to us’
(174) Target:
seker maar daai wit hondjie wat se seker maar daai wit hondjie wat
naam is Nuschka se naam Nuschka is
probably just that white doggie whose probably just that white doggie
name be Nuschka whose name Nuschka be
‘Probably that white doggie whose name is Nuschka’
As stated above, it is not entirely clear what the ATOM would predict in
terms of subject pronoun case in the language of Afrikaans-speaking
children with SLI, as it is not clear what the default case is in Afrikaans.
However, as stated, seeing that there is nothing against which the case of
the retrieved pronoun can be checked, subject pronouns could possibly
be expected to occur in any and all cases available in the lexicon. This
prediction was not borne out by the Afrikaans data: No utterances such
as *Hom sien my ‘Him sees me’ or *Sy sien my ‘His-GENITIVE sees me’
were found. However, three children with SLI did make errors on
subject pronouns. Unlike what the ATOM would predict, these errors
were not related to case, but rather were errors of omission, such as
those shown in (175).
(175) Target:
sien nie die bedde nie ek(?) sien nie die beddens nie
see not the beds not I(?) see not the beds not
‘I(?) do not see the beds’
9.3. THE REPRESENTATIONAL DEFICIT FOR

DEPENDENT RELATIONS HYPOTHESIS
9.3.1. Merit of the RDDR
As mentioned in section 2.4.2, Van der Lely (2003:126, 2004) stated that
she works within Chomsky’s (1995a) Minimalist Programme in her
analysis of the language problems of children with SLI, although the
RDDR is not “tied to” this programme. Perhaps then, one could view
the ongoing revision of the RDDR (and now the Computational
259
Complexity Hypothesis), first proposed in 1994, as an attempt to keep

abreast with developments in Minimalist syntax. However, Van der Lely
seems to adopt a rather eclectic approach when deciding which
principles of Minimalist syntax to incorporate into the RDDR. For
example, she provides the following explanation for the problems that
children with SLI experience with the interpretation of (certain types of)
passive constructions: “The SLI children seem to have a specific
problem ... with the representation of the movement of the internal
argument to the subject position where it receives case and its thematic
role” (Van der Lely 1996:267-8). As regards case, one can argue that, in
1996, Van der Lely was simply working within an older version of
syntactic theory. For this reason, she refers to case as being assigned,
rather than being checked, as has been proposed since Chomsky (1993).
However, the same cannot be said of her statement that the internal
argument moves in order to receive a thematic role: Since the earliest
proposals presented within the principles and parameters approach, it
has been assumed that arguments receive their thematic roles in the
sentence positions in which they are initially generated, specifically
before any movement operations are performed (Haegeman 1994:310;
O’Grady 1997:289). Thus, the claim that arguments cannot receive
certain thematic roles because they do not undergo movement, does not
concur with assumptions about theta-role assignment in Minimalist
syntax (nor with such assumptions in previous theories within the
principles and parameters approach). On more than one occasion, Van
der Lely states that the RDDR offers an explanation for the problems
children with SLI experience with the interpretation of passive
constructions (see, e.g., Van der Lely 1996:267, 2003:127, 2004). This
statement, however, only holds true if one accepts a key assumption
which is contrary to that of the syntactic framework in which she claims
to work. If not, the RDDR does not fully explain the problems these
children have with the interpretation of passive constructions.
Furthermore, Van der Lely and Stollwerck (1997) claim that the RDDR
can explain why children with SLI experience problems in understanding
(non-)co-referential relationships. Typically, such children find it difficult
to correctly interpret personal and reflexive pronouns in constructions
where syntactic knowledge – and not (only) semantic clues or real-world
knowledge – is required to determine the referents of such pronouns, as
in Is Mowgli tickling him?, Baloo Bear says that Mowgli is tickling himself, and
260
The boy says that every monkey is washing himself (Van der Lely and Stollwerck
1997:275). In short, one needs knowledge of the (non-)co-referential
relationships between personal pronouns, reflexive pronouns, and DPs
to understand to what/whom they (can) refer (cf. section 3.4.4.3). On
the RDDR, children with SLI find the interpretation of sentences like
those above difficult because these children fail to establish the syntactic
relationships required for the relevant (non-)co-referential
interpretations.164
The RDDR does seem to provide an account of the problems that

children with SLI have with the production of passive constructions and
other constructions involving Move (such as those involved in question
formation). As noted in section 2.4.2, Van der Lely (1996:246) claims
that the RDDR can explain (i) why the phonological realisation of
grammatical features is optional in the language of children with SLI, and
(ii) why this optionality involves omission of such features in obligatory
contexts rather than insertion in inappropriate contexts. On her
explanation, the tense feature is phonologically realised if it has been
checked; if it is not checked, it will be absent from the phonological form
of the utterance and the verb will accordingly appear in its infinitival
form. It is indeed the case that children with SLI optionally realise
features phonologically (in other words, that they sometimes pronounce
an obligatory morpheme and at other times not). However, Van der
Lely’s explanation is incompatible with the Minimalist assumption that
lexical items enter the Numeration in their inflected form. Given this
assumption, the phonetic realisation of grammatical features is not
dependent on Move, and the verb will only appear in the infinitival form
if it entered the Numeration in that form. If the verb entered the
Numeration in the finite form, that form (and therefore the obligatory
grammatical morpheme) will appear, regardless of whether or not Move
occurs. In short, if the form of the verb is not a reliable indicator of
whether movement has occurred, then the RDDR does not provide an
adequate explanation for the omission of grammatical morphemes in the
164 It appears that this explanation for the problem with the interpretation of (non-)co-
referential relationships is not related to the optionality of the second principle that Van
der Lely (2003:127) claims to be involved in movement (cf. section 2.4.2). What is at
issue here is the “complexity of syntactic dependent relationships between
constituents” (Van der Lely and Stollwerck 1997:282).
261
language of children with SLI.165 The reason why the RDDR’s

explanation is not adequate is because the verb’s phonological form is
not dependent on the application of Move.
Moreover, it is not the case that children with SLI only (optionally) omit
grammatical morphemes; they also sometimes insert such morphemes
into inappropriate contexts, as shown in the *You got a tape recorders
(Gopnik 1990a:147) example given in chapter 2. The RDDR, in its
current form, offers no explanation for the latter phenomenon.
9.3.2. Predictions of the RDDR for Afrikaans
According to Van der Lely (2003:127), the operation Move is optional in

the grammar of children with SLI. Accordingly, a first consequence of
the RDDR would be that the utterances of Afrikaans-speaking children
with SLI should demonstrate varying word order. Specifically, in simple
declaratives, the surface SOV order should occur frequently, as this is the
order that will result if no movement takes place. Similarly, wh-questions
in which the wh-phrase did not move from its original underlying
position should also occur, rendering forms such as *Jy wat doen? ‘You
what do?’ instead of Wat doen jy?. However, other surface word orders
should, at least in principle, occur equally frequently. For example,
sentences with the surface SVO main clause order, such as Ek eet druiwe
‘I eat grapes’, should occur, with the verb eet moved to the left, but the
object druiwe remaining in its initial underlying position.
Secondly, in terms of the RDDR, Afrikaans-speaking children with SLI

should have difficulties understanding and producing passive
constructions – although, as pointed out above, Van der Lely’s reasons
for why children with SLI experience problems with these constructions
are not entirely clear. A third consequence of the RDDR would be that
Afrikaans-speaking children with SLI should experience problems in
interpreting and producing constructions in which co-referential
relationships occur.
165 I acknowledge that this criticism of RDDR is highly “theory-internal”, i.e., strictly
related to the perspective of Chomsky’s (1995a) Minimalist Programme. Some of Van

der Lely’s proposals might be compatible with a Distributive Morphology approach,
where functional heads are spelled out late in the derivation, i.e., after syntactic
computation.
262
Recall that Van der Lely (1996:246) states that the RDDR can explain (i)
why children with SLI appear to use grammatical morphemes optionally,
and (ii) why this optionality involves omission of such morphemes in
obligatory contexts and not insertion in inappropriate contexts.
Although the reasons she provides for these claims are open to criticism,
as pointed out above, her prediction would be that Afrikaans-speaking
children with SLI sometimes inappropriately omit grammatical
morphemes but do not insert them into inappropriate contexts.
9.3.3. Are the predictions of the RDDR for Afrikaans borne out?
The prediction of the RDDR that Afrikaans-speaking children with SLI

should produce utterances with varying word order was not borne out by
the data. Simple declaratives almost exclusively demonstrated the surface
verb-second order (and not the SOV one which would occur if no
movement takes place). Furthermore, wh-questions in which the wh-
phrase did not move from its original underlying position did not occur.
In other words, wh-movement does not appear to apply optionally in wh-
question constructions produced by Afrikaans-speaking children with
SLI. Question constructions (whether of the wh-kind or not) mostly
demonstrated a grammatical word order, as shown in (176) and (177).
(176)
hoekom gaan daai een so?
why go that one so
‘Why is that one going like that?’
(177)
soek jy ’n bed?
look-for you a bed
‘Would you like a bed?’
Question constructions with a verb-second order did occur, but not ones
containing a wh-element. An example of such (acceptable) forms is given
in (178). Where question forms were not grammatical, the
ungrammaticality lay mainly with the omission of constituents or single
words, rather than with their position in the construction – as shown in
(179) and (180), the meaning of the latter example being unclear.
263
(178)
ek kan maar dit afhaal?
I can just it off-take
‘It’s OK for me to take this off?’
(179) Target:
*waar die yskas dan? waar is die yskas dan?
where the fridge then where be the fridge then
‘Where is the fridge then?’
(180)
kan ek nie dit ander mense nie?
can I not this/it other people not?
Regarding the prediction in terms of the RDDR that Afrikaans-speaking

children with SLI have difficulties understanding and producing passive
constructions, this is difficult to assess, seeing that passives occurred at a
very low rate in the language of all three groups of children, even in that
of the typically developing 6-year-olds. In total, only one “full” passive
construction was found in the first 30 minutes of the 45 language
samples (22.5 hours in total).
As stated above, a third consequence of the RDDR would be that

Afrikaans-speaking children with SLI experience problems in
interpreting and producing constructions in which co-referential
relationships occur. As was the case for passive constructions, these
constructions were not tested by the experimental tasks, so evidence
needs to be sought in the language samples. Like passive constructions,
constructions in which co-referential relationships occur were not
produced frequently – eight times in total in the first 30 minutes of the
language samples: twice by typically developing 6-year-olds; twice by 4-
year-olds; and four times by children with SLI (in total, seven children
produced a construction in which co-referential relationships occurred).
Of these eight constructions, not one contained a –self form (as in Hy
trek homself aan ‘He dresses himself’); all eight were mekaar ‘each other /
one another’ forms as in (181), produced by a typically developing 6-
year-old girl. One of these, produced by a boy with SLI and given in
(182), was ungrammatical.
264
(181)
en ek het lanklaas gesien hoe gee hulle vir mekaar ’n soentjie
and I did not-for-a-long-while see-PAST PART how give they for each-other a
kiss-DIM
‘And it has been a long while since I saw them giving each other a kiss’
(182) Target:
as daar nog ’n kas is dan moet ons as daar nog ’n kas is dan moet ons
hom ook op mekaar sit hom ook op die ander sit
if there another cupboard is then if there another cupboard is then
must we him also on each other put must we him also on the other put
‘If there is another upboard, we must put him on the others too’
A fourth prediction of Van der Lely for the language of Afrikaans-

speaking children with SLI was that grammatical morphemes would
sometimes be omitted inappropriately but not inserted into inappropriate
contexts. This prediction was not borne out by the Afrikaans data.
Although errors of omission were far more frequent than those of
insertion, the latter did also occur.
9.4. THE FEATURE DEFICIT HYPOTHESIS
9.4.1. Merit of the Feature Deficit Hypothesis
Recall that on the Feature Deficit Hypothesis, SLI is the result of a

deficit in the knowledge of rules regarding the morphological marking of
a specific class of linguistic features, and that this deficit is claimed to
result in an inability to formulate implicit grammatical rules. The
characteristics of SLI that the Feature Deficit Hypothesis seems to
account for are the following:
(i) The insertion of grammatical morphemes in inappropriate
contexts. This could occur when the incorrect memorised form
(the one with phonologically realised grammatical morphology
instead of the one without) is selected.
(ii) A lower percentage of use of grammatical morphemes. This could
occur if the memorised form without the phonologically realised
grammatical morphology is more often selected than the one with
the relevant morphology.
265
However, Gopnik’s hypothesis seemingly fails to account for the

following characteristics of the language of persons with SLI, because an
explanation for these characteristics of SLI falls outside of the scope of
her account:
(i) Word order problems in question constructions.
(ii) Problems in establishing (non-)co-referential relationships.166
(iii) Difficulty in interpreting passive constructions, which Gopnik
(1990b) did not find to be problematic for her participants with
SLI.
A further potential criticism against Gopnik’s hypothesis concerns the

increased burden on lexical acquisition in the case of people with SLI.
For English-speaking people with SLI, memorising every stem as well as
each of its inflected forms (e.g., walk, walked, walking, walks) should, at
least in principle, be possible, given the relatively impoverished system of
morphological inflection in English. However, relying on rote learning to
compensate for a lack of implicit grammatical rules would place an
intolerably high burden on the lexical memorisation by speakers of
languages with a significantly higher number of inflected forms for
lexical items.167 This raises questions about the cross-linguistic
generalisability of Gopnik’s hypothesis.
9.4.2. Predictions of the Feature Deficit Hypothesis for

Afrikaans
In terms of Gopnik’s Feature Deficit Hypothesis, there are at least five

predictions which could be made about the language of Afrikaans-
speaking children with SLI. Firstly, such children should find it difficult
to consistently use the correct singular vs. plural forms of nouns; hence
they would have to rely on memorisation to determine the plural form of
each noun. To a certain extent, this is what typically developing
Afrikaans-speaking children have to do as well: There is no single default
rule for forming Afrikaans plurals, unlike for English, where “add –s to
the singular form” will result in a high level of accuracy. Furthermore,
166 Note that, if pronouns are taken to be essentially spelled out feature bundles, then
problems in establishing (non-)co-referential relationships might also be related to the
difficulties children with SLI experience with the features of pronominal forms.
167 For example, according to Pinker (1994:127), Greek has about 350 and Turkish up
to 2 million forms for each verb.

266
there are many exceptions to the pluralisation rules in Afrikaans (cf.

section 3.3.1.1). In contrast to nouns, number is not indicated through
affixation on personal and possessive pronouns, but rather through a
difference in the way they are spelled out phonologically. The resulting
phonological forms have to be memorised by all speakers of Afrikaans,
because, as in English, there is no generic rule for forming the plural
counterpart of a singular personal or possessive pronoun (i.e., the “add
–s to the singular form” rule cannot be applied to these pronouns).
Because all speakers of Afrikaans must memorise all plural forms of
pronouns and many plural forms of nouns, Gopnik’s hypothesis makes
no interesting prediction for the language of Afrikaans-speaking children
with SLI as regards their use of plural forms. The prediction it does
make is that Afrikaans-speaking children with and without SLI should
demonstrate more or less the same level of accuracy in using plural
forms.
A second prediction of the Feature Deficit Hypothesis concerns

grammatical gender. As this is not overtly expressed in Afrikaans, it is
predicted that the language of Afrikaans-speaking children with SLI
should not demonstrate any problems in this regard.168
The third prediction concerns the grammatical expression of person and

case. In Afrikaans, verbs are not inflected for person; however, person
and case are phonologically indicated on personal and possessive
pronouns – through variation in sound, not through affixation. On
Gopnik’s hypothesis then, Afrikaans-speaking children with SLI should
at times use the incorrect form of the pronoun (e.g., julle ‘you-PL’ instead
of hulle ‘they’). However, this same prediction should hold for typically
developing Afrikaans-speaking children as well. In short, because the
168 As noted in section 3.3.1.3, in Afrikaans, semantic gender is indicated on some
nouns denoting people and (less commonly) animals, mostly by the use of derivational
suffixes (cf. note 49). Whether semantic gender is indicated phonologically in the form
of suppletion, affixation, compounding, or not at all, has to be learnt by all speakers of
Afrikaans. On Gopnik’s hypothesis, Afrikaans-speaking children with SLI may at times
use the noun denoting a male entity where a noun denoting a female one should have
been used and vice versa. However, as there is no single implicit rule for expressing
semantic gender in Afrikaans, the same prediction should hold for typically developing
Afrikaans-speaking children. Again, on the Feature Deficit Hypothesis, Afrikaans-
speaking children with and without SLI should demonstrate the same level of accuracy
in using nouns denoting semantic gender.
267
phonological form of Afrikaans personal and possessive pronouns

cannot be derived in terms of a default grammatical rule, all children
need to memorise the correct phonological form of these pronouns.
The fourth prediction relates to the phonological realisation of tense. As

mentioned in section 3.3.1.5, tense is not phonologically indicated on
(the vast majority of) main verbs in Afrikaans. The present tense form of
the main verb is the same as the infinitival form.169 Past tense can be
indicated by using any of the following:
(i) The temporal auxiliary het and the past participial form of the main
verb, as can be seen in (183) below.
(ii) The past tense form of the modal auxiliary/-ies and the infinitival
form of the main verb, as in sou slaap ‘would have slept’ and sou
moes kon slaap ‘would have had to be able to sleep’.
(iii) The past tense form of the modal auxiliary/-ies, the temporal
auxiliary het and the past participial form of the main verb, as in sou
geslaap het ‘would have slept’ and sou moes kon geslaap het ‘would have
had to be able to sleep’.
(iv) The present tense form of the modal auxiliary, the temporal
auxiliary het and the past participial form of the main verb, as in sal
geslaap het ‘would have slept’ and the less common sal moet kan
geslaap het ‘would have had to be able to sleep’.
(v) The present tense form of the main verb, which is phonologically
identical to the infinitival form, in contexts where past tense is
denoted by an AdvP, as can be seen in (184).
(183)
Toe ons daar gekom het, het hulle geslaap
when we there come-PAST PART did did they sleep-PAST PART
‘When we came there, they slept’
(184)
Toe ons daar kom, slaap hulle
when we there come sleep they
Note that (183) and (184), which are paraphrases of each other, can both
also be paraphrased as (185) and (186).
169 Compare slaap ‘sleep’ and verkies om te slaap ‘prefer(s) to sleep’.
268
(185)
Toe ons daar gekom het, slaap hulle
when we there come-PAST PART did sleep they
(186)
Toe ons daar kom, het hulle geslaap170
when we there come did they sleep-PAST PART
On Gopnik’s hypothesis, Afrikaans-speaking children with SLI should

experience difficulty acquiring the past tense rules in (i) to (v) above.
Utterances such as the following should therefore appear in the language
of these children.171
(187) Target:
Hulle slaap Hulle het geslaap
they sleep they did sleep-PAST PART
‘They sleep’ ‘They slept’
(188) Target:
Hulle sal slaap Hulle sou slaap
they will sleep- INF they will-PAST sleep- INF
‘They will sleep’ ‘They would have slept’
(189) Target:
Hulle het slaap Hulle het geslaap
they did sleep-INF they did sleep-PAST PART
‘They slept’ ‘They slept’
(190) Target:
Hulle geslaap Hulle het geslaap
they sleep-PAST PART they did sleep-PAST PART
‘They slept’ ‘They slept’
170 Besides the meaning indicated by the English translation, this sentence can also have
the meaning ‘When we came there, we discovered that they had (previously) been
sleeping’.
171 Examples (187) and (188) are ungrammatical on the intended meaning; examples
(189) and (190) are ungrammatical regardless of intended meaning.

269
Finally, as noted above, Gopnik’s Feature Deficit Hypothesis does not

seem to offer any account for the difficulties children with SLI
experience with the production and interpretation of question
constructions, (non-)co-referential relationships, and passive
constructions. Accordingly, it is not clear whether, on this hypothesis,
Afrikaans-speaking children with SLI should experience difficulties in
this regard when compared to their typically developing peers.
9.4.3. Are the predictions of the Feature Deficit Hypothesis for

Afrikaans borne out?
There were at least five predictions made about the language of

Afrikaans-speaking children with SLI, in terms of Gopnik’s Feature
Deficit Hypothesis. The first was that such children should find it
difficult to consistently use the correct singular vs. plural forms of nouns;
hence, they would have to rely on memorisation to determine the plural
form of each noun. Recall that, due to the nature of the Afrikaans plural
system, the actual prediction is that Afrikaans-speaking children with and
without SLI should demonstrate more or less the same level of accuracy
in using plural forms. This prediction was not borne out by the
Afrikaans data: The 6-year-olds with SLI were outperformed by those
without SLI.
The second prediction concerned grammatical gender. As such gender is

not encoded in Afrikaans, it was predicted that the language of
Afrikaans-speaking children with SLI should not demonstrate any
problems in this regard. No task specifically assessed grammatical
gender; therefore, it is not possible to state whether or not this
hypothesis was supported.
The third prediction concerned the grammatical expression of person

and case: Afrikaans-speaking children with and those without SLI should
at times use the incorrect form of the pronoun. This prediction was not
fully borne out by the Afrikaans data. The typically developing 6-year-
olds fared better than those with SLI and better than the 4-year-olds in
terms of both elicited and spontaneous production of pronouns. Certain
errors of case (on both pronouns and se-constructions) were made by the
6-year-olds with SLI but not by those without SLI.
270
The fourth prediction related to the phonological realisation of tense:

Afrikaans-speaking children with SLI should experience difficulty
acquiring past tense rules, and utterances such as *Hulle slaap (meaning
‘They slept’) instead of Hulle het geslaap ‘They slept’ should therefore
appear in their language. However, no evidence of this error type was
found in the spontaneous language samples. Furthermore, it was not
possible to assess whether such forms occurred inappropriately in the
experimental task or not: Such forms did indeed occur frequently, but
had to be taken as being appropriate, as the researcher provided the
adverb gister ‘yesterday’, which then permitted the use of the historic
present tense, as in Gister slaap hulle ‘Yesterday they slept’. Likewise, the
inappropriate occurrence of utterances such as *Hulle sal slaap instead of
Hulle sou slaap was not determinable by the experimental tasks, and there
was no evidence of this error type in the spontaneous language of any of
the groups.
It was also predicted that utterances such as *Hulle het slaap instead of
Hulle het geslaap would occur. This error type was not made by any of the
4-year-olds and occurred only once in the language samples of the
typically developing 6-year-olds: A girl produced the utterance given in
(191).
(191) Target:
dit was toe ek vir jou wag het dit was toe ek vir jou gewag het
it was when I for you-SGL wait-INF did it was when I for you-SGL wait-PAST
PART did
‘It was when I waited for you’
By contrast, this error was made four times in the first 100 utterances of
language samples of the children with SLI and another 14 times in the
remainder of the 30 minutes. Two examples are given in (192) and (193),
the latter concerning a particle-verb.
(192) Target:
ons het kyk ’n puppet show ons het ’n puppet show gekyk
we did watch-INF a puppet show we did a puppet show watch-PAST
PART
‘We watched a puppet show’
271
(193) Target:
daai klein kleintjie het hy heel melk daai klein kleintjie het hy al sy melk
opdrink? opgedrink?
that little little-one did he whole milk that little little-one did he all his milk
up-drink-INF up-drink-PAST PART
‘That little one, did he finish all his milk?’
The omission of the temporal auxiliary het occurred only in the language
of the children with SLI: six times in the first 100 utterances and 11
times in the remainder of the 30 minutes. Examples of such errors are
given in (194) and (195).
(194) Target:
hy jy bed gesteel hy het jou bed gesteel
he you-SGL-NOM bed steal-PAST PART he did your-SGL bed steal-PAST PART
‘He stole your bed’
(195) Target:
jy moet nie geloop nie jy moet/moes nie geloop het nie
you-SGL must not walk-PAST PART not you-SGL must/must-PAST not
walk -PAST PART did not
‘You were not supposed to walk’ / ‘You should not have walked’
Finally, on Gopnik’s Feature Deficit Hypothesis, it is not clear whether

Afrikaans-speaking children with SLI should experience difficulties in
terms of the production and interpretation of passive constructions,
(non-)co-referential relationships, and question constructions when
compared to their typically developing peers. As stated in section 9.3.3,
passive constructions and those containing co-referential relationships
had a very low rate of occurrence in the language samples of all three
groups of children, making it difficult to draw any conclusion regarding
these constructions. In terms of question constructions, the children
with SLI made errors not found in the language of the typically
developing 6-year-olds. However, these errors mostly related to the
omission of words or constituents, and definitely not to grammatical
morphemes.
272
9.5. SUMMARY: PREDICTIONS OF THE THREE

ACCOUNTS FOR SLI IN AFRIKAANS
Research question 5 asked whether or not the predictions made for the
language of Afrikaans-speaking children with SLI by some current
theoretical accounts of SLI are borne out by the Afrikaans data obtained
in this study. Support was found neither for the three predictions of the
ATOM nor for the four of the RDDR. Of the five made by the Feature
Deficit Hypothesis, one was partly borne out.
Apart from the fact that the accounts of SLI examined here do not make
useful predictions for the language of Afrikaans-speaking children with
SLI, the children with SLI in the present study made errors in their
spontaneous language production which fall outside the scope of these
accounts. Therefore, it appears that there is a need for an alternative
account of SLI as it presents itself in Afrikaans. In the next section, an
attempt is made to address this need.
9.6. AN ALTERNATIVE ACCOUNT OF SLI AS IT

PRESENTS ITSELF IN AFRIKAANS
9.6.1. Background
A general assumption was that Minimalist syntax has the potential to

offer interesting explanations for the problems that children with SLI
experience with grammatical morphology and constituent movement. In
chapter 3, it was said that the assumptions and devices of Minimalist
syntax could make it possible to give a unified explanation of apparently
unrelated phenomena (i.e., problems with grammatical morphemes,
word order, question and passive constructions, and co-reference) in
terms of the devices of feature checking and movement – in the sense
that movement is driven by the need for feature checking in order to
prevent a derivation from crashing. In this section, we investigate
whether this is indeed the case: Can an account for the language
problems of the Afrikaans-speaking children with SLI be offered in
terms of these assumptions and devices of Minimalist syntax?
Before examining whether or not this is possible, another look is taken at

exactly what needs to be accounted for. In section 9.6.2, the error
273
patterns regarding the grammatical features number, person, case, and

tense are discussed separately. The purpose of this discussion is to
establish exactly what it is that a theoretical account of SLI has to
account for, i.e., how SLI presents itself in Afrikaans in terms of
problems with these four features. Then, in section 9.6.3, an attempt is
made at offering an explanation for these errors in term of the devices of
Minimalist syntax. The utterances with ungrammatical word order
produced by the Afrikaans-speaking children with SLI are re-examined
in section 9.6.4. In section 9.6.5, we consider whether the problems
regarding the morphological expression of grammatical features and
those pertaining to word order are explainable as two types of
manifestation of one underlying problem in the grammar of Afrikaans-
speaking children with SLI.
9.6.2. Another look at the errors pertaining to grammatical

features
In terms of the comprehension of the grammatical feature number,

there were differences between the three groups of children, but no clear
pattern could be detected in these differences (except that the typically
developing 6-year-old group fared well and the other two groups almost
equally poorly). An account of SLI as it presents itself in Afrikaans does
not have to explain these differences, given the small number of items
involved and the fact that the children with SLI and the 4-year-olds
performed similarly to a great extent. The comprehension of number by
the children with SLI could therefore be merely delayed, but probably
not deviant.
In terms of the elicited production of plural forms of both real and

nonsense words, it appeared that the children with SLI presented with a
delay: There were differences between the SLI and typically developing
4-year-old groups on certain items, but, in general, the children with SLI
and the typically developing 4-year-old group omitted the plural
morpheme a similar number of times, and also replaced the targeted
plural morpheme with another one a similar number of times.
The same cannot be said for the spontaneous production of plural

forms: Here, only the children with SLI omitted the plural marker (but in
total only twice), and they replaced one plural morpheme with another
274
far more frequently than did either of the two typically developing
groups. Mostly irregular plural forms were replaced by regular ones. This
pattern cannot be seen as one of delay; rather, the use of plural forms by
the Afrikaans-speaking children with SLI appears to deviate from the
norm. An account of SLI in Afrikaans needs to explain this deviation.
Turning to the grammatical feature case: On the elicited production of

pronouns, the children with SLI made more errors than did either of the
two groups of typically developing children, with the 4-year-olds faring
worse than the 6-year-olds. In spontaneous production of pronouns, the
same pattern was seen: The children with SLI fared the worst and the
typically developing 6-year-olds the best in terms of correct realisation of
pronoun case. The majority of the errors involved substituting
possessive sy (‘his’) with hom (‘him’). The errors made by the children
with SLI on se-constructions were so few that one should not have to
account for them. However, the fact that the children with SLI are more
delayed than even the 4-year-olds in terms of correctly realising pronoun
case needs to be accounted for.
The proportion of errors related to person in the elicited language of the

children with SLI was approximately double that of the 4-year-olds. This
needs to be accounted for. In terms of spontaneous production of
person on pronouns, almost no errors occurred.
Regarding the elicited production of past tense forms, the SLI and
typically developing 4-year-old groups performed similarly in terms of
grammatical past tense constructions (whether it was the targeted
construction or another one which was produced); the typically
developing 6-year-old group performed better than the other two. There
was no difference in the frequency of use of the historic present tense
form by the three groups. The only notable difference between the
groups was that the SLI group made more idiosyncratic errors than the
other two; there was no significant difference in the number of other
errors between the three groups. In terms of spontaneous production of
past tense forms, the children with SLI fared worse than the other two
groups in terms of correct production of het ge- forms. Mostly, the errors
involved the omission of the temporal het, but the past participle was also
omitted at times (and at other times only the ge- of the past participle was
omitted). Interestingly, the children with SLI fared similarly to the
275
typically developing 6-year-olds – and thus better than the typically

developing 4-year-old group – in terms of the correct production of the
past tense forms of be. Here, the typically developing 4-year-old group
omitted the be and sometimes used is/het gewees ‘be-PRESENT/did be-PAST
172
PART’ instead of was (gewees) ‘be-PAST (be-PAST PART)’. So there are two
aspects that need to be accounted for: the fact that het ge- forms were
problematic only for the children with SLI and not for the two groups of
typically developing children; and the fact that the children with SLI
presented like same-aged typically developing ones in terms of the
production of past tense forms of be, whereas the 4-year-olds made
errors on these constructions.
Regarding the spontaneous production of present tense forms, the

children with SLI omitted modal auxiliaries at times, whereas the two
typically developing groups did not. Also, the children with SLI omitted
present tense be forms more than three times as often as the 4-year-olds.
An account of SLI in Afrikaans should ideally address these differences.
In summary, an adequate account of SLI as it presents itself in Afrikaans

needs to offer explanations for why Afrikaans-speaking children with
SLI
(i) use the incorrect plural morpheme far more frequently than do
younger and age-matched typically developing children;
(ii) make more errors pertaining to case and person of pronoun than
do younger and age-matched typically developing children;
(iii) make more idiosyncratic errors and more errors on het ge- forms
when producing past tense constructions than do younger and
age-matched typically developing children;
(iv) omit modal auxiliaries at times, whereas typically developing
(younger and same-aged) children do not;
(v) present like same-aged typically developing children in terms of
the production of past tense forms of be, whereas 4-year-olds
make errors on these constructions;
172 It should be noted though that speakers of some dialects of Afrikaans accept is gewees
and het gewees in main clauses, as in Ons is/het gister daar gewees. It is mostly het gewees that
was used by the 4-year-olds, and always in main clauses. By contrast, is gewees and was
gewees are unacceptable in subordinate clauses (apparently for all speakers of Afrikaans),
but not het gewees: Toe ons gister daar gewees het vs. *Toe ons gister daar gewees is/was. Gewees
is/was/het was not used in subordinate clauses by any of the participants.
276
(vi) omit present tense be forms more than three times as often as 4-
year-olds.
9.6.3. A Minimalist account of the errors pertaining to

grammatical features
Before turning to explanations for the errors pertaining to grammatical

features in the language of Afrikaans-speaking children with SLI, a brief
discussion is required of which features are present in the lexicon. Recall
that lexical items are defined as bundles of features (Chomsky
1995a:230), specifically phonological, semantic, and formal (or syntactic)
features, and that retrieving a lexical item from the lexicon to form a part
of a Numeration thus implies retrieving a set of features. On the
Lexicalist view, lexical items reach the Numeration as fully inflected
bundles of features; all features (including phonological and semantic
ones) are specified (cf. Chomsky 1995a:275). To be more specific,
morphological forms are created in the lexicon. That is, an inflectional
feature A is associated with a root B via some morphological rule
applying in the lexicon, and this complex word receives a sound
representation before it enters the syntactic derivation as a lexical item.
Stated differently, the complex word [X+infl] forms a word that is part
of the Numeration, which constitutes the input for the syntactic
derivation.
An alternative view, one which is compatible with Minimalist syntax, is

that of Distributed Morphology (cf. Halle and Marantz 1993). According
to this approach, in short, lexical items are retrieved from the lexicon
and all their formal features are added to them before they reach the
narrow syntax. After the derivation has taken place, phonological
features are added to the items at PF. Specifically, all morphological
structures are built in the syntax (as opposed to in the lexicon). The
nodes that are manipulated in the syntactic derivation are of two types,
namely functional morphemes and roots. Functional morphemes are
composed exclusively of non-phonetic features, such as [past], [present],
[plural], or [singular], or the feature(s) that make up the determiner node
of the definite article the. Roots comprise the open class, i.e., lexical,
vocabulary (e.g., car or sit). In the PF component of the grammar, these
functional morphemes receive phonological representations (i.e., sound
form) in the process of Vocabulary Insertion. This process involves the
277
selection of a vocabulary item from a set of possible Spell-Outs of this

functional morpheme. Stated differently, Vocabulary Insertion is a
process that provides phonological content to functional heads, which
are assumed to be feature bundles without phonological content in the
syntactic derivation. For example, in the case of Vocabulary Insertion of
an English past tense form, a vocabulary item must be selected from the
competing items in (196).
(196)
T[past] Å> -t/__ {lent, sent, …} (=specific past tense form)
T[past] Å> -ø/__ {hit, quit, …} (=specific past tense form)
T[past] Å> -ed (=default past tense form)
According to Distributed Morphology, vocabulary items compete

according to specificity, so that the most highly specified item is inserted.
That is, more specified items (e.g., irregular ones) can block less specified
ones (e.g., regular ones) (cf. Embick 2007). When Vocabulary Insertion
takes place, the vocabulary item (i.e., the functional morpheme) is
adjoined to the root.173
Note this competition between vocabulary items during the process of

Vocabulary Insertion, i.e., the process taking place in the PF-component
of the grammar, where these functional morphemes receive a
phonological representation. Thus, Vocabulary Insertion assigns
phonological content to the syntactic nodes. When one vocabulary item
(e.g., –t) “wins” this competition, it prevents the other vocabulary items
from being inserted. So, when –t appears as T[past] in the context of the
root lend, it is at the expense of the default case, which has the form –ed.
According to Embick and Marantz (2007), blocking in the sense of a
competition for the expression of syntactic or semantic features is
limited to the insertion of the phonological exponents of such features
(where “exponent” refers to the phonological expression of a
morpheme; cf Embick and Noyer 2001) at terminal nodes from the
syntax. Blocking therefore does not occur at the word level or above,
173To be more precise, the vocabulary item (or the functional morpheme) adjoins to
the root+category-defining head (e.g., small v). However, for the sake of simplicity, the
vocabulary item representing the functional morpheme will be taken to attach to the
root (e.g., V) directly.
278
and there is no competition between grammatical and ungrammatical

structures.
Bearing this in mind, if explanations are to be given for (i) to (vi) of

section 9.6.2, the following could be suggested:
(i) Where an incorrect plural form is produced, the correct feature

bundle has been selected from the lexicon, but, at PF, this feature
bundle was spelled out incorrectly, i.e., it received the incorrect (in
terms of non-SLI grammar) sound form. In terms of Distributed
Morphology, one could say that the “incorrect” vocabulary item
was inserted, which means that –s rather than –e (or vice versa) is
selected as the exponent (or sound form) for the abstract plural
feature. These incorrectly spelled out, but nevertheless
interpretable, plural forms will not cause a derivation to crash in
the semantic component, seeing that the feature [+plural] is still
present. This feature is merely realised phonologically in a way that
is non-adult-like. The problem here lies not with the semantic or
syntactic features – after all, the meaning is conveyed adequately
and the derivation is not caused to crash – but with the
phonological features, which could point to a problem at Spell-
Out in PF.
(ii) A similar explanation can be given for the case errors found on
pronouns: The correct feature bundle is selected from the lexicon
– one which would be spelled out as sy by adult speakers of
Afrikaans – and, after undergoing the necessary movement
operations in the computational component, this feature bundle is
initially, incorrectly, spelled out as hom by Afrikaans-speaking
children, and for a prolonged period by those with SLI. These
children thus do not select a pronoun with incorrect case; rather,
the third-person singular masculine possessive pronoun has the
form hom in the grammar of these children. It is clear from the
Afrikaans pronoun paradigm why the children would prefer hom
over sy:174 Sy is the nominative form of the third-person singular
feminine pronoun. The oblique and possessive forms of this
pronoun are both haar. Hy is the nominative form of the third-
174 See (14) in section 3.3.1.4 for the Afrikaans pronoun paradigm.
279
person singular masculine pronoun. However, its oblique and

possessive forms differ: The oblique is hom, similar in sound form
to the feminine haar (to a certain extent), but the possessive form
is the same as the nominative form of the third-person singular
feminine pronoun, viz. sy. By substituting sy ‘his’ with hom,
paradigm leveling takes place. Again, the problem appears to lie
not with the semantic or syntactic features themselves, but with
the phonological form given to the feature bundle at Spell-Out.
Note that the sound form which the child chooses still belongs to
the (personal) pronominal paradigm. In a sense then, the mapping
of the incorrect sound form is “local”, by which is meant that the
sound form is chosen from a specific paradigm of “pronominal”
sound forms. From this, one might conclude that the child has
difficulties in identifying the right sound form (out of a larger set
of forms) which maps onto a feature structure (e.g., [third-person;
sgl; feminine]). An alternative explanation for the occurrence of
incorrect forms of personal pronouns might thus be that the
Vocabulary Insertion rules are not yet as fixed, or as robustly
organised, as the Vocabulary Insertion rules in the grammar of
(older) typically developing Afrikaans-speaking children.
A possible reason why children with SLI make errors pertaining to

person without causing the derivation to crash – such as dit is my
hand instead of dit is haar hand – is the following: Person, together
with number and gender, is a so-called phi-feature. This particular
phi-feature is semantically interpretable and therefore does not
need to be checked for LF purposes. It could thus be that the
children with SLI select a pronoun with incorrect person without
causing the derivation to crash in the semantic component.
Similarly, the derivation will not crash in the phonological
component, as the feature in question is interpretable.
Alternatively, the correct feature is selected but it is spelled out at
PF in a non-adult-like way, i.e., it is given a different sound form
to that in the adult grammar.
(iii) The relative scarcity of generative literature on the grammar of

Afrikaans auxiliaries makes it difficult to explain the errors on het
ge- past tense forms and the omission of modal auxiliaries by
280
Afrikaans-speaking children with SLI. The errors on het ge- forms

were of three types: At times, het was omitted; at times, ge- of the
past participle was omitted; and at times, the past participle was
omitted as a whole. Each of these errors will be accounted for
below. Before turning to these accounts, one should notice that
there are various formal means of expressing past tense in
Afrikaans (cf. section 3.3.1.5). In terms of Distibuted Morphology,
this could again be a case where there is a set of competing forms
from which the child has to choose, and where a non-adult-like
form “wins” the competition.
Where het is omitted in a sentence containing a modal auxiliary –

e.g., Hy sou gestap het ‘He would have walked’ – such omission
should not cause the derivation to crash, given that the tense
feature is carried by the modal and that the modal can move to
check the V feature of the T. However, the past participle is not
selected by this modal, but by het. Therefore one needs to assume
that het (i.e., the node representing the temporal auxiliary) is
somehow present in the syntactic structure. This het then receives
no sound form at Spell-Out, possibly because past tense property
is already expressed by the modal (sou, in this case). Also, if there is
no modal auxiliary in the sentence – as in Hy het gestap ‘He walked’
– then the derivation should crash if the het is omitted, given that
the T would have an unchecked V feature. However, these
derivations do not crash in the case of children with SLI;
therefore, an alternative explanation is required. One possibility
could be that het is initially present in the derivation, is copied and
moved to the T (yielding Hy het gestap het), but that at the point of
Spell-Out no copy of het (not even the left-most one of the chain
{het, het}) receives sound form. In terms of Distributed
Morphology, one would argue that the syntactic node representing
the temporal auxiliary is not assigned an exponent (or sound
form), due to Vocabulary Insertion (i.e., the mapping of a sound
form onto a node carrying certain features) not taking place. This
results in the node remaining silent at PF. A possible reason why a
child with SLI would find it acceptable to leave het phonologically
empty (or silent) could be related to Kayne’s (1993) proposal that
‘have’ could be seen as ‘be’ plus an incorporated preposition (i.e.,
‘have’=PrepositionDAT/LOCATIVE+‘be’), along the lines of Benveniste
281
(1966). As both ‘be’ and the abstract preposition are semantically

poor, it could be that the child does not provide the complex form
[P+‘be’] with phonological contents at Spell-Out.
Where ge- is omitted, one could argue that the feature bundle
which in the adult grammar matches the sound form of the past
participle is indeed selected from the lexicon, but that, at Spell-
Out, this feature bundle receives a sound form matching that of
the infinitive in the adult grammar. In terms of Distributed
Morphology, a non-adult-like vocabulary item would have “won”
the competition between the verbal forms (finite form, infinitival
form, participial form). This explanation is plausible, seeing that
there are indeed competing past participial forms: those which
resemble infinitival ones (such as onthou ‘remember’ – het onthou
‘remembered’) and those with ge- (such as bou ‘build’ – het gebou
‘built’). The Afrikaans-speaking child with SLI has to identify one
past participial form from a set of competing potential candidates,
and the correct form is then not necessarily identified.
However, even if one assumes that the incorrect verb form (i.e.,
the infinitive instead of the past participle) was selected from the
lexicon, this incorrect selection should not cause the derivation to
crash, as neither the infinitive nor the past participle has features
which check any feature of the T; the T’s features are checked by
the temporal auxiliary het. However, a past participial form
presumably has a feature which causes it to be selected as the
complement of het. If an infinitive is selected instead of a past
participle, then a selectional feature of het is not checked. Given
that, despite this (apparently) unchecked feature, the derivation
does not crash, one could assume that the feature is, in fact,
checked: A feature bundle resembling that of a past participle was
selected from the lexicon by the child with SLI; at Spell-Out, this
feature bundle is realised in a way which differs morphologically
from the way in which the adult speaker of Afrikaans would
realise it. Again, one could argue that the problem in the grammar
of the Afrikaans-speaking child with SLI lies not in the syntactic
representation, but at Spell-Out; i.e., at the point where the
syntactic representation is mapped onto a phonological
representation.
282
Where a past participle is omitted as a whole – as in Hulle altwee het

op ’n blou bed geslaap – there are possible semantic implications, as
the sentence now no longer necessarily conveys the intended
meaning. If the absence of the past participle is ascribed to non-
selection from the lexicon, this would mean that the selectional
feature of het which requires a past participle complement would
remain unchecked. This should cause the derivation to crash, but
the derivation, in fact, converges. Therefore, it is argued that the
past participle is indeed selected from the lexicon but is then, for
some or other reason, not spelled out in the phonological
component.175 This would mean that the complement feature of
het is checked in the course of the syntactic derivation, which
explains why the derivation converges.
(iv) The omission of modal auxiliaries (as in OK, nou die kinders eet. Hoe
moet hulle eet? ‘OK, now the children (must) eat. How must they
eat?’) is problematic for two reasons. The first is that the element
which carries tense is not present, so there is no verbal element
which checks the V feature of the T – a similar problem to one
discussed above in connection with the omission of het. Again,
one could propose that the modal is initially present in the
derivation, is copied and moved to the T, and that, at Spell-Out,
both copies of the modal are then deleted instead of only the
lowest one. The question arises as to why both copies can be
deleted. This is not possible in the grammar of adult speakers of
Afrikaans; typically, the head of the chain (i.e., the left-most copy)
is spelled out, seeing that this is the one which has had most of its
features checked. One could propose that this principle (viz.
phonologically realising the copy which has entered into most
checking relations) is not yet known to Afrikaans-speaking
children with SLI. For these children then, other considerations
are involved in deciding which copy to spell out. However, the
“decision” as to which copy to spell out might be so complex
(given that it is not a principled decision) that these children might
175 One reason could be that the problems which the child with SLI experience with
selecting the correct verbal form are of such a magnitude that the child opts for not
mapping the terminal node onto a sound form at all.
283
opt for omitting the sound form of all copies, resulting in none of
the copies being spelled out phonologically.
The second problem pertaining to the omission of modal

auxiliaries is that the meaning of the sentence might no longer be
clear if a modal is omitted. Unlike the omission of the temporal
auxiliary het – which does not lead to an interpretation problem,
due to the fact that the “past” interpretation is arguably
recoverable from the past participle verb; cf. *Hy geslaap vs Hy het
geslaap – the omission of a modal auxiliary could make the
intended meaning of the sentence unclear. For example, by
omitting a modal auxiliary like sal ‘will’, moet ‘must’, mag ‘may’, or
kan ‘can’, it is not clear exactly what the child intends to ask with
*Ons met die ding werk?.
(v, vi) The fact that the Afrikaans-speaking children with SLI are age-
appropriate in terms of their production of the past tense forms of
be, but frequently omit the present tense form of be, could be
explained as follows: A structure expressing a proposition
minimally consists of a subject argument and a predicate, and the
smallest construction by which such a proposition can be
expressed is a so-called small clause. A small clause does not allow
for more than one argument: It consists of a subject to which a
specific attribute is given (e.g., John handsome in I find John handsome).
Where the be is omitted, a construction similar to a small clause is
rendered – the be is implied and its omission does not alter the
intended meaning of the sentence. However, if one wants to
convey the idea that a subject previously had a specific attribute
but no longer has it, the verb can no longer be implied. For
instance, one can omit the be in *Ek hier ‘I here’ and still convey
the intended meaning, namely “I am here”. However, if one wants
to convey that one had been somewhere (as in Ek was hier (gewees)
‘I had been here’), a phonologically realised verb is required. On
this proposal then, is ‘be-PRESENT’ can be left phonologically
empty without compromising meaning, but was (gewees) ‘be-PAST’
must be expressed phonologically in order to convey the intended
meaning.
284
In conclusion, it appears that, in the grammar of Afrikaans-speaking

children with SLI, the abstract feature representation and the movement
operations are intact. In other words, the language problems that these
children experience do not seem to reside in the computational system
underlying the construction of a syntactic representation. That is, SLI-
children are able to build a convergent LF-representation. Their problem
seems to lie at Spell-Out, where the feature bundles need to receive
sound form: Either a deviant (i.e., non-adult-like) sound form is given, or
the features are not spelled out phonologically at all.
9.6.4. Another look at the word order errors
Recall that all three groups of participants made word order errors, but
not all types of errors were made by all groups, and that the only error
type which appeared unique to the SLI group was that of SOV and VSO
in non-embedded clauses. In this section, we will revisit some of the
examples given in section 8.3.3. Specifically, it will be argued that these
problems with word order relate to the Spell-Out of chains, i.e., to the
“decision” as to which copy will be realised in the PF-representation. At
times, the head of a chain is spelled out (this is not discussed here, seeing
that this renders the correct, adult-like word order); at times, an
intermediate copy is spelled out (e.g., the copy of a verb in T is spelled
out in T instead of the copy in C); and at yet other times, the syntactic
constituent is spelled out in situ (i.e., the “lowest” copy receives sound
form). Though not given in section 8.3.3, there were also examples in the
Afrikaans data of more than one copy receiving sound form, and these
examples will also be discussed here.
Let us begin by considering the derivation of examples (142), (143), and

(154) – repeated here as (197), (199), and (201), respectively – presented
in (198), (200) and (202). From (198), it appears that the child with SLI
does not experience problems with the movement operations required
for feature checking. Rather, it appears that the problem lies at Spell-
Out, where an incorrect copy receives sound form. More specifically, in
(198), the copy of kyk surfaces phonologically in the head position of
small v. It is assumed that the verb is finite (and as such specified for the
grammatical features tense and agreement). As such, the verb also
undergoes movement to C (via T), rendering a verb-second pattern. In
other words, the computational rules of this Afrikaans-speaking child
285
with SLI work in the same way as they do in adult Afrikaans. The only
difference is that, in contrast to adult Afrikaans, the head of the chain is
not spelled out here, but rather a lower copy. In (198), it is indicated that
it is an intermediate copy (the one in v) which is spelled out; however,
one could also argue that it is the lowest copy (the one in V) which
receives sound form.176
(197) Target:
hulle TV kyk hulle kyk TV
they TV watch they watch TV
‘They are watching TV’
(198) CP
Spec C'
hulle
C TP
kyk
Spec T'
hulle
T vP
kyk
Spec v'
hulle
VP v
kyk
DP V
TV kyk
A boy with SLI produced utterance (199), which was a statement but has
the surface word order of a yes/no-question.
(199) Target:
vryf hy die been en ’n pappa hy vryf die been van pappa
rub he the leg and a daddy he rub the leg of daddy
‘He is rubbing daddy’s leg’
176Alternatively, one could argue that TV kyk (or then TV-kyk) is, in fact, a compound
verb, in which case the analysis of (197) would be similar to that of (210).
286
As shown in (200), the verb vryf moved from the VP to the v, the T, and
then the C. The left-most copy of vryf was then spelled out (correctly
so), whereas that of hy – which moved from the specifier position of vP,
to that of TP, and then to that of CP – was not. As in the case of (198),
an intermediate copy thus received sound form.
(200) CP
Spec C'
hy
C TP
vryf
Spec T'
hy
T vP
vryf
Spec v'
hy
VP v
vryf
DP V
vryf
die been
en ’n pappa
In (201), ons babatjies is focalised, meaning that the modal auxiliary kan
‘can’ needs to move to the C – seeing that focalisation requires subject-
verb inversion in Afrikaans. This presumably took place, as shown in
(202), as did the other required movement operations, after which the
left-most and intermediate copies of kan did not receive sound form at
Spell-Out; the lowest copy did, rendering an incorrect word order.
(201) Target:
ons babatjies ons by hier kan kies ons babatjies kan ons by hierdie kies
our babies-DIM we by here can choose our babies-DIM can we by there
choose
‘Our babies we can choose to match these’ [=we can choose figurines (ones
which match these pieces of toy furniture) to be our babies]
287
(202) CP
Spec C'
ons
babatjies C TP
kan
Spec T'
ons
T vP
kan
Spec v'
ons
VP v
kan kies
AdvP V'
by hier
DP V
kies
As stated in chapter 8, “other” word order errors – ones which are

difficult to group in terms of misplaced elements –occurred, mostly in
the language of children with SLI. An example of such an error is
repeated here as (203).
(203) Target:
en hulle meet om hulle op die lorrie en hulle meet hulle om op die lorrie
te gaan te gaan
and they measure infinitive-complemen- and they measure them infinitive-
tiser they on the truck to go complementiser on the truck to go
‘And they measure them to go onto the truck’
In (203), a left-most copy (of hulle) was deleted incorrectly – presumably

because similar information is realised phonologically elsewhere in the
utterance: Here an infinitival clause is introduced by the infinitive
complementiser om. The subject of the infinitival clause is PRO – not an
“ordinary” pronoun, because this subject cannot receive case. In the
adult grammar, there is a co-referential relationship between the second
hulle (of which the referent is ostriches) and the PRO. In the child’s
utterance, this co-referential relationship is still expressed, because it is
hulle – and not some other pronoun such as sy ‘she’ – which is in the
subject position of the infinitival clause. In other words, instead of
288
“silent” PRO, the features of PRO are realised phonologically as hulle –

hulle because of the co-referential relationship in question – and the
“real” hulle then does not receive sound form, because its features are all
realised phonologically elsewhere in the utterance. Stated differently,
there is competition between two proniminal forms, namely between an
empty form PRO and a phonologically overt form hulle. In this case, the
incorrect form “won” the competition.
On these derivations it would appear that Afrikaans-speaking children

with SLI do not experience a problem with the syntactic computation in
terms of Move (or Copy and Merge), but that the problem lies with
spelling out the correct copies that constitute a chain. Whereas the left-
most copy is typically spelled out in the adult grammar and (usually) all
lower copies deleted, these children sometimes delete left-most copies
and spell out lower (intermediate or right-most) ones.
When considering the other types of word order errors made by the
children with SLI (those which were also made by one or both of the
groups of typically developing children), a similar observation is made.
To illustrate this point, the derivation of some of the example utterances
given in chapter 8 – and repeated here for the sake of convenience – is
presented below. The tree diagram in (205) proposes that all the
necessary movement operations occurred but that the left-most copy of
was was then not spelled out; an intermediate copy was spelled out.
(204) Target:
laas jaar ek was by ’n ou plaas laas jaar was ek by ’n ou plaas
last year I be-PAST by a old farm last year be-PAST I by a old farm
‘Last year I was on an old farm’
The utterance in (206) is a wh-question construction. The wh-element has

the correct surface position, but so-called subject-verb inversion
apparently did not take place. However, from (207), it appears that a
copy of the auxiliary did, in fact, move to the complementiser position,
but that this copy then did not receive phonological form at Spell-Out;
an intermediate copy did (or the lowest one, if one argues that it is the
copy in V which received sound form).
289
(205) CP
Spec C'
laas jaar
C TP
was
Spec T'
ek
T VP
was
SC V
was
D PP
ek
by ’n ou plaas
(206) Target:
hoekom ding kan nie trap nie? hoekom kan die ding nie trap nie?
why thing can not pedal not why can the thing not pedal not
‘Why can the thing not pedal?’
(207)177 CP
Spec C'
hoekom
C TP
kan
Spec T'
ding
T vP
kan
Spec v'
ding
VP v
kan trap
Spec V'
DP V
trap
177For the sake of simplicity, the nie’s are not indicated in this derivation. For a
proposal regarding derivations containing nie, see Oosthuizen (1998).
290
Another question construction produced by a child with SLI is presented

in (208). Here, the verb werk does occur to the left of the subject ons, that
is, an inverted order can be observed. However, in the surface form, the
adverb weer occurs between the wh-element and the verb, which causes
this main clause not to demonstrate the required verb-second word
order. From (209), it appears that the finite verb moved to C (via T)
(rightly so), and that the adverb weer was base-generated in the specifier
postion of TP (acting as a type of sentential adverb), which means that
the subject could not move into that position. What is at issue here, it
that the intermediate copy of the finite verb (the one in T), instead of the
left-most one, received sound form.
(208) Target:
hoekom weer werk ons net so bietjie? hoekom werk ons weer net so
bietjie?
why again work we just such bit why work we again just such bit
‘Why do we again just work a little bit?’
(209) CP
Spec C'
hoekom
C TP
werk
Spec T'
weer
T vP
werk
Spec v'
ons
VP v
werk
DP V
net so werk
bietjie
291
In the grammar of adult speakers of Afrikaans, particle-verbs have the linear

order particle+verb (e.g., houtkap or afval),178 but in the grammar of children,
it appears that the compound can also be verb+particle (i.e., kaphout and
valaf). When verb movement takes place to the T for feature checking
purposes, the whole particle-verb (kaphout or valaf) is raised and not just the
verbal stem (kap or val).179 One could say that pied piping takes place here,
whereas something similar to preposition stranding (say, “particle
stranding”) should have occurred. This would explain the word order in
utterances such as (210). As can be seen from (211), it appears that
unnecessary material has been copied and/or incorrect copies spelled out.
(210) Target:
daar val af die een daar val die een af
there fall off the one there fall the one off
‘There the one falls off’
(211) CP
Spec C'
daar
C TP
val af
Spec T'
die een
T VP
val af
SC V
val af
DP Particle
die een af
178 Compounds in Afrikaans are head-final. Consider, for example, plakboek ‘scrap
book’ N = plak ‘paste’ V + boek ‘book’ N, and mooimaak ‘beautify’ V = mooi ‘pretty’ A +
maak ‘make’ V.
179 In constrast to the adult speaker of Afrikaans, the child with SLI moves the entire
complex verb [V+Preposition] to T and subsequently to C. It is possible that the child

is confused by the existence of other verbal forms consisting of a verb and a
preposition (although in the order [Preposition+V]) which undergo movement as a unit
to C. Two examples of sentences containing such verbal forms are given in (i) and (ii):
(i) Paul onderhandel met die smouse
‘Paul negotiates with the hawkers’
(ii) Susan oorskat haar vermoëns
‘Susan overestimates her abilities’
292
In (212) and (214), wat and dat both receive sound form in the head
position of the CP. Dat originates in this position, whereas wat is moved
here. In these two constructions, it appears that the child moves the
finite verb to T overtly, whereas this movement would not occur overtly
in the grammar of adult speakers of Afrikaans. This movement of the
finite verb to T renders an embedded verb-second word order,180
because in these two cases, the left-most copy of the finite verb is indeed
the one which is spelled out.
(212) Target:
en hierdie is sy mamma wat jy het en hierdie is sy mamma wat jy daarso
daarso het
and this is his mommy that you have and this is his mommy that you there
there have
‘And this is his mommy that you have there’
(213) DP
DP CP
sy
mamma Spec C'
C TP
wat
Spec T'
jy
T vP
het
Spec v'
jy
VP v
het
AdvP V'
daarso
DP V
wat het
180Other utterances containing relative clauses also point to the movement of finite V
to T by Afrikaans-speaking children with SLI; see the utterance in (174).
293
(214) Target:
dat hy kan sy fietsie ry dat hy sy fietsie kan ry
that he can his bicycle-DIM ride that he his bicycle-DIM can ride
‘That he can ride his bicycle’
(215) CP
Spec C'
C TP
dat
Spec T'
hy
T νP
kan
Spec ν'
hy
VP ν
kan ry
DP V
sy fietsie ry
It appears then that one can account for most of the word order errors
in terms of Minimalist syntax: Movement operations (mostly) occur as
they should, rendering a fully grammatical (i.e., adult-like) derivation
before the point of Spell-Out. However, at Spell-Out, some copies
which were supposed to receive sound form do not, and others which
were supposed to be left phonologically empty are, in fact, spelled out.
This leads to the proposal that Afrikaans-speaking children with SLI do
not experience problems in the computational component, but in
mapping grammatical features onto sound form.
It could, of course, be argued that the Afrikaans-speaking children with

SLI indeed demonstrate problems with movement operations. In other
words, it could be the case that there is not merely a problem spelling
out a moved element, but that such element, in fact, does not undergo all
the necessary movement operations. However, the former proposal –
that the child copies the element and then fails to spell it out at PF – is
more attractive, amongst other reasons because of the occurrence of the
utterance in (118), repeated here as (216).
294
(216) Target:
want hulle het al paar keer shock het want hulle het al paar keer geshock
because they did already few time because they did already few time
shock did shock-PAST PART
‘Because they have shocked themselves a few times already’
From this utterance, it can be seen that the temporal auxiliary het did, in
fact, move to the required position, but that two copies of het were
spelled out at PF, rendering an ungrammatical utterance. Similarly,
responses containing two hets were given to some items of the sentence
completion task assessing the production of tense – for example, het eet
het ‘did eat did’ in response to Hierdie beer kan elke dag heuning eet. Gister, net
soos elke ander dag, … ‘This bear can eat honey every day. Yesterday, just
like every other day, ...’. Other relevant responses to this task were het sy
alles staan het ‘did she everything (under)stand did’ (instead of het sy alles
verstaan) and het sy ’n blom gepluk het ‘did she a flower pick did’ (instead of
het sy ’n blom gepluk).
Other examples of utterances (apart from those involving the temporal

auxiliary het) in which both copies of a chain received sound form are
(105), (110) and (111), repeated here as (217) to (219). In (217), a
personal pronoun is spelled out twice (yielding doubling within the
DP);181 in (218) the first member of a hendiadys is spelled out twice; and
in (219), it is a modal auxiliary which receives sound form in two
positions in the utterance.
(217) Target:
hierso is jou klere jou hierso is jou klere
here are your-SGL clothes your-SGL here are your-SGL clothes
/you-OBLIQUE-SGL
‘Here are your clothes’
(218)
die’s al die mense wat kom by ons kom die’s al die mense wat by ons kom
kuier kuier
these-be-CONTR all the people who these-be-CONTR all the people who
come at us come visit at us come visit
‘These are all the people who are coming to visit us’
One could argue that the copy of jou to the right of klere is in the complement
181
position of klere, as it would be in the construction klere van jou.

295
(219) Target:
gaan hulle hamers gaan nou kry gaan hulle hamers nou kry
will their hammers will now get will their hammers now get
‘Will now get their hammers’
9.6.5. The alternative account
As has been said in the previous two sections, it appears that most of the
errors made by the Afrikaans-speaking children with SLI regarding
grammatical morphemes and word order are indeed related to
grammatical features: It is proposed here that the problem does not
principally lie with the checking of grammatical features (i.e., with the
movement operations required for feature checking), but with spelling
out these features at PF. Stated differently, the computational
component of Afrikaans-speaking children with SLI seems to be intact;
the mapping of the syntactic information onto phonological form
appears to be defective. Furthermore, the concept of ‘competition
between available (sound) forms’ seems to play a role in the language
problems demonstrated by these children.
In chapter 3, it was stated that, ideally, an account of SLI in Afrikaans

should provide a comprehensive explanation for the observed
characteristics of SLI in Afrikaans. It is possible to make use of the
assumptions and devices of Minimalist syntax – specifically those related
to Spell-Out – to account for (almost) every error made by the children
with SLI in the present study, and the proposed account has a certain
descriptive and explanatory power: It can describe the errors and relate
them to a certain part of the grammar, viz. the mapping between abstract
syntactic information and phonology.
Research question 6 asked if one can propose an alternative account182 of

SLI as it presents itself in Afrikaans. The answer to this question is “yes”:
One can propose that Afrikaans-speaking children with SLI have
problems assigning the correct sound form to a constellation of
grammatical features. Although no conclusive reasons can be provided
for why the Afrikaans-speaking children with SLI have problems with
Spell-Out at PF, it is not inconceivable that the PF component is
182 That is, other than the ATOM, RDDR, or Feature Deficit Hypothesis.
296
involved in errors of this nature, seeing that it is the component in which

variation is expected:
It is not surprising … to find a degree of variation in the PF

component, and in aspects of the lexicon: Saussurean arbitrariness
(association of concepts with phonological matrices), properties of
grammatical formatives (inflection, etc.), … . Variation in the overt
syntax or LF component would be more problematic, since
evidence could only be quite indirect. A narrow conjecture is that
there is no such variation.
(Chomsky 1995a:169-170)
Chomsky (1995a:7) furthermore states that variation of language is

“essentially morphological in character, including the critical question of
which parts of a computation are overtly realized”. Also, if children with
SLI are indeed seen to experience difficulties in organising alternative
competing forms, one might expect that this would lead to errors in
Spell-Out at PF.
On the proposed account, it appears to be possible to predict which

errors will not occur in the language of Afrikaans-speaking children with
SLI. It appears that the choice of which copy to spell out is at least in
part based on the input. Copies are not spelled out if such spell-out leads
to an utterance containing adjacent words which are not usually found
adjacent to each other in the input. More specifically, the deviant word
order patterns are typically those that are attested in the language of adult
speakers of Afrikaans, although in different structural contexts. For
example, in the language of Afrikaans-speaking children with SLI, an
embedded word order (with the verb pronounced in the clause-final
position) is used in a main clause, and main clause word order is found in
embedded contexts. These children “select” a word order pattern out of
the set of possible word order patterns that are found in the language.
Let us consider a few examples here to illustrate this point. The word
order in (197), hulle TV kyk, is often heard in embedded clauses, such as
(hulle sê) dat hulle TV kyk. Utterance (199), vryf hy die been en ’n pappa has
the surface word order of a yes/no-question construction and is thus also
often heard in the input. The kan kies in (201) is the modal-infinitive
word order often heard in embedded clauses, such as (Dit is) wat ons by
297
hier(die) kan kies. In utterance (204), the ek was by ’n ou plaas of the laas jaar
ek was by ’n ou plaas has the grammatical word order of a simple
adverbless declarative sentence – a word order which occurs frequently
in the input. A similar case could be made for ding kan nie trap nie in
hoekom ding kan nie trap nie? (206); weer werk ons net so bietjie in hoekom weer
werk ons net so bietjie? (208); jy het daarso (sy mamma) in en hierdie is sy mamma
wat jy het daarso (212); and hy kan sy fietsie ry in dat hy kan sy fietsie ry (214).
Utterance (210), daar val af die een, contains a particle-verb which is often
heard separated in the input, for example in hy val af. Lastly, parts of het al
paar keer shock het (216) are heard in hulle het al paar keer (ge)shock en dat
hulle al paar keer (ge)shock het.
It is proposed that utterances such as Ons waarom dit doen do not and will
not appear in the language of Afrikaans-speaking children with SLI due
to the fact that ons waarom and waarom dit doen are not sequences often
encountered in the input.183 The same can be said for declarative
utterances with an OSV word order, such as Hom hulle sien.
In conclusion, the account of SLI as it presents itself in Afrikaans is the

following: Afrikaans-speaking children with SLI
(i) (mostly) find movement operations unproblemic;
(ii) can render derivations up to the pre-Spell-Out level which
resemble those found in the adult speaker’s grammar;
(iii) at times spell out the incorrect copies, if this spelling out leads to
“local” linearisations also found in the input;
(iv) at times spell-out an abstract feature ([singular], [tense], etc.) in a
deviant (i.e., non-adult-like) manner.
183Both ons waarom and waarom dit doen are encountered in the input – in, for example,
Nou wonder ons waarom dt gebeur het ‘Now we are wondering why it happened’ and
Waarom dit doen as jy iets anders kan doen? ‘Why do this if you can do something else?’,
respectively, but infrequently.
298
Conclusion
Chapter 10
Conclusion
10.1. SUMMARY OF THE PRESENT STUDY
From a theoretical point of view, SLI is an interesting phenomenon:

The study of SLI – specifically, cross-linguistic data on the manner in
which it presents itself in typologically diverse languages – may provide
insight into the nature of the human language faculty, amongst other
things. However, for persons with SLI, the phenomenon is probably far
less interesting. Its longstanding nature and the reach of its influence –
which stretches beyond the linguistic (cf. Brinton et al. 2005; Voci,
Beitchman, Brownlie, and Wilson 2006) – has the potential to impact
negatively on such people’s quality of life.
It is in this context that another study on SLI was undertaken:

(i) Afrikaans, the language in which SLI was studied here, has
properties which are interesting for testing the predictions of
existing theoretical account of SLI.
(ii) No relevant data on SLI in Afrikaans had previously been
gathered. This lack of information limits the effectiveness of the
diagnosis and remediation of SLI in Afrikaans-speaking children.
The general aim of the study was to provide an adequate account of SLI
as it presents itself in Afrikaans. To this end, the language of Afrikaans-
speaking 6-year-olds with SLI was compared to that of typically
developing 4- and 6-year-olds. Specifically, the comprehension and
production of grammatical morphemes related to the grammatical
features number, person, case, and tense were evaluated, both with
experimental tasks and in spontaneous language samples. Furthermore,
errors of word order in spontaneously produced utterances were studied.
In brief, the first research questions asked how Afrikaans-speaking
children with SLI present in terms of their comprehension and
production of grammatical morphemes related to number, person, case,
and tense. Do these children differ from typically developing Afrikaans-
299
Conclusion
speaking children? If so, do Afrikaans-speaking children with SLI

present with a delay, a deviance, or both in terms of their comprehension
and production of grammatical morphemes?
The data of this study revealed that Afrikaans-speaking children with SLI
present like younger typically developing ones on the experimental tasks.
By contrast, in terms of the spontaneous production of morphemes
pertaining to these grammatical features, the Afrikaans-speaking children
with SLI fared worse than both typically developing groups. The
children with SLI mostly made the same types of errors as the younger
ones on morphemes related to these grammatical features. In general,
these errors included the omission, inappropriate insertion, and
substitution of grammatical morphemes. However, some errors were
unique to the children with SLI. For instance, only the children with SLI
omitted the main verb het and doubled the temporal auxiliary het.
Regarding word order errors, some were made by all three groups of
children (such as producing relative clauses with a surface SVO word
order), others only by the children with SLI and the younger typically
developing ones (such as moving particle-verbs as a whole), and yet
others only by the children with SLI (such as using a surface SOV or
VSO word order in main clauses). Therefore, it appears that the language
of Afrikaans-speaking children with SLI is not merely delayed, but also
somewhat deviant.
Another research question was whether or not predictions (in terms of

the comprehension and production of grammatical morphemes) made
for Afrikaans by some current theoretical accounts of SLI are borne out
by the Afrikaans data obtained in this study. Support was found neither
for the three predictions of the ATOM nor for the four of the RDDR.
Of the five made by the Feature Deficit Hypothesis, one was partly
borne out. The answer to this research question thus appears to be “no”.
The last research question concerned the possibility of proposing an

alternative, comprehensive account of SLI as it presents itself in
Afrikaans. Here, the answer is “yes”. The account proposed here is that
the problems that Afrikaans-speaking children experience with
grammatical morphemes and word order are related to problems at
Spell-Out at PF: Either certain grammatical features are given a sound
form different to that found in the adult speaker’s language, or certain
300
Conclusion
copies of a movement chain in the pre-Spell-Out derivation receive no

sound form at all at Spell-Out, whereas other copies are spelled out
twice. It was shown that the language problems that children with SLI
have could be seen to be localised (principally) in that part of the
grammar which concerns the mapping between syntax and phonology.
Furthermore, for these children, difficulties seem to arise specifically in
contexts where there is more than one potential Spell-Out candidate
available (i.e., where there is competition between sound forms that may
realise a functional category, or competition between various copies for
Spell-Out). As such, the proposed account has a measure of predictive
force: Errors are expected to occur in the formal realisation of
morphosyntactic information, especially in those contexts where two or
more potential sound forms (or Spell-Out positions) are in competition
with each other.
10.2. LIMITATIONS OF THE PRESENT STUDY
The account which has been proposed does not have very strong
predictive power, in the sense that it will not be able to predict exactly
what errors any particular Afrikaans-speaking child with SLI will make.
As stated in chapter 2, children with SLI are known to constitute a
heterogeneous group (Aram 1991:84-85). That was the case for the
Afrikaans-speaking children with SLI who participated in this study as
well: Not all of them found the same morphemes problematic, and,
where two or more children did find one type of morpheme
problematic, not all of them necessarily made the same type of error.
The diversity of errors made by the children with SLI in this study was
such that it could potentially have a negative effect on the predictive
power of any account proposed for SLI in Afrikaans.
The diversity of errors can partly be attributed to the fact that only 15
children with SLI were included in this study. Had it been possible to
gather a larger corpus of impaired language, generalisation and error
prediction might have been improved. As mentioned in chapter 4, these
15 children were identified over a period of 21 months. Speech
therapists were contacted on a continual basis, and they thoroughly
examined their case loads for possible participants. Many Afrikaans-
speaking 6-year-olds with language problems were identified by the
therapists. However, very few of them had SLI: Many, in addition to
301
Conclusion
having language problems, had apraxia, were mentally challenged, and/or

came from bilingual homes. It thus proved more difficult than expected
to obtain Afrikaans-speaking participants with SLI.
A related reason for not all children with SLI making errors on the same
aspects of language could have to do with the lack of an agreed-upon
protocol for the identification of SLI in Afrikaans-speaking children and
the limited standardised instruments available for diagnostic purposes.
The speech-language therapists were requested to refer only those
children who demonstrated problems with grammatical morphology
and/or syntax. However, the therapists had limited means with which to
diagnose delay and/or deviance of a morphosyntactic nature. It could be
that – despite their speech-language therapists diagnosing them with
grammatical SLI – some of them could have had (minimal) symptoms of
grammatical SLI while being more impaired as regards semantics or
pragmatics. If the 15 participants with SLI did not have exactly the same
type of SLI, then one would expect that which they find problematic and
that which they find easy to differ across participants. However, given
the heterogenity of SLI populations, it might have been difficult to
detect clear error patterns for the group as a whole even if the children in
this study did indeed all have grammatical SLI.
Another reason for the diverse error patterns across participants with
SLI could be that some of the children with SLI had overcome at least
some of their problems related to grammatical morphemes and word
order by the time of this study. The children with SLI were all 6 years
old, and all but one of them had been receiving speech-language therapy
when they took part in this study. Had their language been studied when
they were younger (and thus more impaired), a more uniform pattern of
errors across participants might have been found. However, including
younger children with SLI would have meant including even younger
typically developing children as controls. The younger controls in this
study were 4 years old. It is doubtful whether reliable data would have
been obtained from 2- or 3-year-olds. Initally, the experimental tasks
were indeed performed with some 3-year-olds. This proved to be a very
trying experience for researcher and participant alike: Despite frequent
rewards, the experimental tasks were not sufficiently captivating to hold
the attention of the 3-year-olds for the required length of time. This
resulted in frequent and long rest periods (sometimes as little as five
302
Conclusion
items were responded to between rest periods). While most of them

wanted to participate, the stop-start manner in which their data were
gathered (sometimes with frequent repetition of both instructions and
individual items before responses were given) led to the quality of the
responses of the 3-year-olds being questionable. Because it seemed
reasonable that 4-year-olds were the youngest children from which data
for this study could be gathered reliably, the participants with SLI had to
be 6-year-olds.184
10.3. SUGGESTIONS FOR SUBSEQUENT STUDIES OF

SLI IN AFRIKAANS
In a sense, this study is “pioneering work”: Apart from the data collected
in this study, no relevant data on SLI in Afrikaans-speaking children
exist. This means that there are obviously many areas of SLI in Afrikaans
which still require examination. One could, amongst many other things,
study the narrative ability, the use of figurative language, the possible
omission of functional categories, the verb system, etc. of Afrikaans-
speaking children with SLI. However, the discussion here will be limited
to future studies on the morphosyntactic aspects which were the focus
of the present study.
One suggestion is that data be gathered from the full age range of
preschool children with SLI (from 2 to 6 years), in an attempt to gain
insight into the development of grammatical morphemes in the language
of Afrikaans-speaking children with SLI. Because of the problems in
executing experimental tasks with very young children, spontaneous
language samples may be gathered, especially considering the relative
success with which typically developing children were identified as such
in the present study by means of discriminant analysis performed with
various measures of the spontaneous language samples. Note that
developmental data do not yet exist for typically developing Afrikaans-
speaking children. Therefore, the proposed study would have to include
typically developing children as well, in order to enable comparison
between impaired and non-impaired language development in Afrikaans.
Recall from chapter 4 that the language-matched controls in some other studies were
184
on average 2 years younger than the experimental group with SLI (cf. De Jong
2003:154; Johnston et al. 1993:974; Oetting and Rice 1993:1239; Rice 2003:72).
303
Conclusion
Such data should then allow for a systematic search for a clinical marker
of SLI in Afrikaans. The language samples of the impaired and non-
impaired children can be examined for the percentage of correct use in
obligatory contexts of grammatical morphemes related to number,
person, case, and tense, as well as to functional categories.
With a larger corpus, more systematic error patterns across the language-
impaired children may well occur. In this case, it might be possible to
propose an account of SLI which has the power to predict which errors
are likely to occur in the language of any particular Afrikaans-speaking
child with SLI.
10.4. CONCLUSION
Throughout this dissertation, it was argued that two well-documented

characteristics of SLI, namely problems with grammatical morphology
and problems with constituent movement, can both be related to
problem(s) with grammatical features, also in Afrikaans. This was shown
to be the case: The errors made by the Afrikaans-speaking children with
SLI could indeed be explained in terms of Spell-Out of grammatical
features at PF. The study made a contribution to localising the problem
children with SLI experience: This problem appears not to lie in the
computational system (i.e., not with Merge or Move), but with syntax-
phonology mapping, where children with SLI have difficulties choosing
the correct form for Spell-Out, which may be related to the fact that
these Spell-Out forms are not (yet) stored in as organised a manner as
they are in the adult lexicon.
304
References
REFERENCES
Adger, D. 2003. Core syntax. A minimalist approach. New York: Oxford

University Press.
Alexiadou, A. 1999. On the properties of some Greek word order
patterns. In A. Alexiadou, G. Horrocks, and M. Stavrou (eds.)
Studies in Greek syntax. Dordrecht: Kluwer. pp. 46-65.
Alexiadou, A. and E. Anagnostopoulou. 1998. Parametrizing Agr: Word
order, verb-movement and EPP-checking. Natural Language and
Linguistic Theory 16:491-539.
Allen, S. and M.B. Crago. 1993. Early acquisition of passive morphology
in Inuktitut. In E. Clark (ed.) Proceedings of the 24th Stanford Child
Language Research Forum. Stanford: Center for the Study of
Language and Information. pp. 112-123.
American Speech and Hearing Association. 1997-2006. Hearing
screening. Available online at http://www.asha.org/public/
hearing/testing/
Anderson, R.T. 1998. The development of grammatical case distinctions
in the use of personal pronouns by Spanish-speaking
preschoolers. Journal of Speech, Language, and Hearing Research
41:394-406.
Aram, D.M. 1991. Comments on specific language impairment as a
clinical category. Language, Speech, and Hearing Services in Schools
22:84-87.
Aram, D.M. and J.E. Nation. 1980. Preschool language disorders and
subsequent language and academic difficulties. Journal of
Communication Disorders 13:159-170.
Arlman-Rupp, A.J.L., D. Van Niekerk de Haan, and M. Van der Sandt-
Koenderman. 1976. Brown’s early stages: Some evidence from
Dutch. Journal of Child Language 3:267-274.
Bain, R. 1936. The self-and-other words of a child. The American Journal of
Sociology 41:767-775.
Baker, M.C. 1988. Incorporation: A theory of grammatical function changing.
Chicago: University of Chicago Press.
305
References
Balason, D.V. and C.A. Dollaghan. 2002. Grammatical morpheme

production in 4-year-old children. Journal of Speech, Language, and
Hearing Research 45:961-969.
Bastiaanse, R. and G. Bol. 2001. Verb inflection and verb diversity in
three populations: Agrammatic speakers, normally developing
children, and children with specific language impairment (SLI).
Brain and Language 77:274-282.
Bedore, L.M. and L.B. Leonard. 1998. Specific language impairment and
grammatical morphology: A discriminant function analysis.
Journal of Speech, Language and Hearing Research 41:1185-1192.
Belletti, A. and L. Rizzi. 2002. Editor’s introduction: Some concepts and
issues in linguistic theory. In N. Chomsky. On nature and language.
Cambridge: Cambridge University Press. pp. 1-44.
Benincá, P. and C. Poletto. 2004. Topic, focus and V2: Defining the CP
sublayers. In L. Rizzi (ed.) The cartography of syntactic structures,
Volume 2: The structure of CP and IP. New York: Oxford University
Press.
Benveniste, E. 1966. Problèmes dem linguistique générale. Paris: Gallimard.
Berko, J. 1958. The child’s learning of English morphology. Word
14:150-177.
Bernstein, D.K. and E. Tieger-Farber. 1997. Language and communication
disorders in children. Fourth edition. Boston: Allyn and Bacon.
Beverly, B.L. and C.C. Williams. 2004. Present tense be use in young
children with specific language impairment: Less is more. Journal
of Speech, Language, and Hearing Research 47:944-956.
Biberauer, T. 2002. Verb second in Afrikaans: Is this a unitary
phenomenon? Stellenbosch Papers in Linguistics 34:19-69.
Biberauer, T. and M. Richards. 2004. True optionality: When the
grammar doesn’t mind. Paper presented at the Minimalist
Theorizing Workshop, Bloomington, 26 June 2004.
Biberauer, T. and I. Roberts. 2005. Subjects, tense and verb-movement
in Germanic and Romance. Paper presented at the 5th Asian
GLOW Meeting, Jawaharlal Nehru University, Delhi, 5-7
October 2005.
Blake, J., D. Myszczyszyn, and A. Jokel. 2004. Spontaneous measures of
morphosyntax in children with and without specific language
impairment. Applied Psycholinguistics 25:29-41.
306
References
Blake, J., G. Quartaro, and S. Onoratti. 1993. Evaluating quantitative

measures of grammatical complexity in spontaneous speech
samples. Journal of Child Language 20:139-152.
Bol, G.W. 1996. Optional subjects in Dutch child language. In C. Koster
and F. Wijnen (eds.) Proceedings of the Groningen assembly on language
acquisition. Groningen: Centre for Language and Cognition
Groningen. pp. 125-133.
Bol, G.W. 2003. MLU-matching and the production of morphosyntax in
Dutch children with specific language impairment. In Y. Levy
and J. Schaeffer (eds.) Language competence across populations. Toward
a definition of specific language impairment. Mahwah: Lawrence
Erlbaum Associates. pp. 259-271.
Bol, G. and F. Kuiken. 1990. Grammatical analysis of developmental
language disorders: A study of the morphosyntax of children
with specific language disorders, with hearing impairment and
with Down’s syndrome. Clinical Linguistics and Phonetics 4:77-86.
Bornstein, M.H., O.M. Haynes, K.M. Painter, and J.L. Genevro. 2000.
Child language with mother and with stranger at home and in the
laboratory: A methodological study. Journal of Child Language
27:407-420.
Bortolini, U., M.C. Caselli, P. Deevy, and L.B. Leonard. 2002. Specific
language impairment in Italian: The first steps in search for a
clinical marker. International Journal of Language and Communication
Disorders 37:77-94.
Bortolini, U., L.B. Leonard, and M.C. Caselli. 1998. Specific language
impairment in Italian and English: Evaluating alternative
accounts of grammatical deficits. Language and Cognitive Processes
13:1-20.
Botha, M.C. 2006. Die interne struktuur van die komplementeerder-
sisteem in Afrikaans [The internal structure of the complementiser
system in Afrikaans]. Master’s thesis, Stellenbosch University.
Brinton, B., M. Fujiki, and L.A. Robinson. 2005. Life on a tricycle: A
case study of language impairment from 4 to 19. Topics in
Language Disorders. Language Disorders and Learning Disabilities: A
Look Across 25 years 25:338-352.
Brown, R. 1973. A first language: The early stages. Cambridge: Harvard
University Press.
307
References
Buitendag, M.M. 1994. Afrikaanse reseptiewe woordeskattoets [Afrikaans

receptive vocabulary test]. Pretoria: Human Sciences Research
Council.
Campbell, A.L., P. Brooks, and M. Tomasello. 2000. Factors affecting
young children’s use of pronouns as referring expressions. Journal
Carrow-Woolfolk, E. 1985. Test for auditory comprehension of language -
Revised. Austin: Pro-Ed.
Carrow-Woolfolk, E. 1999. Test for auditory comprehension of language - Third
edition. Austin: Pro-Ed.
Caselli, C., P. Casadio, and E. Bates 1999. A comparison of the transition
from first words to grammar in English and Italian. Journal of
Child Language 26:69-111.
Chabon, S., L. Kent-Udolf, and D. Egolf. 1982. The temporal reliability
of Brown’s mean length of utterance measure with post-stage V
children. Journal of Speech and Hearing Research 25:124-128.
Charest, M.J. and L.B. Leonard. 2004. Predicting tense: Finite verb
morphology and subject pronouns in the speech of typically-
developing children and children with specific language
impairment. Journal of Child Language 31:231-246.
Cheng, L. and R. Sybesma. 1995. Language is the perfect solution. An
interview with Noam Chomsky. Glot International 1:1, 31-34.
Chomsky, N. 1982a. Some concepts and consequences of the theory of government
and binding. Cambridge: MIT Press.
Chomsky, N. 1982b. Lectures in government and binding: The Pisa lectures.
Second edition. Dordrecht: Foris Publications.
Chomsky, N. 1993. A minimalist program for linguistic theory. In K.
Hale and S. Keyser (eds.) The view from Building 20: Essays in
linguistics in honor of Sylvian Bromberger. Cambridge: MIT Press. pp
1-52.
Chomsky, N. 1995a. The minimalist program. Cambridge: MIT Press.
Chomsky, N. 1995b. Bare phrase structure. In G. Webelhuth (ed.)
Government and binding theory and the minimalist program. Oxford:
Blackwell. pp. 383-439.
Chomsky, N. 1998. Minimalist inquiries: The framework. MIT Occasional
Papers in Linguistics 15. Cambridge: MIT Press.
Chomsky, N. 1999. Derivation by phase. MIT Occasional Papers in
Linguistics 18. Cambridge: MIT Press.
308
References
Chomsky, N. 2000. Minimalist inquiries: A framework. In R. Martin, D.

Michaels, and J. Uriagereka (eds.) Step by step. Essays on minimalist
syntax in honor of Howard Lasnik. Cambridge: MIT Press. pp. 89-
156.
Chomsky, N. 2001. Beyond explanatory adequacy. Manuscript, MIT.
Chomsky, N. 2002. On nature and language. Cambridge: Cambridge
University Press.
Chomsky, N. 2006. Approaching UG from below. Unpublished manuscript,
MIT.
Clahsen, H. 1988. Normale und gestörte Kindersprache. Linguistische
Unterzuchen zum Erwerb von Syntax und Morphologie [Normal and
disordered child language. Linguistic examinations on the
acquisition of syntax and morphology]. Amsterdam: Benjamins.
Clahsen, H. 1989. The grammatical characterization of developmental
dysphasia. Linguistics 27:897-920.
Clahsen, H., S. Bartke, and S. Goellner. 1997. Formal features in
impaired grammars: A comparison of English and German SLI
children. Journal of Neurolinguistics 10:151-171.
Clahsen, H. and M. Rothweiler. 1992. Inflectional rules in children’s
grammars: Evidence from German particles. In G. Booij and J.
Van Marle (eds.) Yearbook of morphology 1992. Dordrect: Kluwer.
pp. 1-34.
Clegg, J., C. Hollis, L. Mawhood, and M. Rutter. 2005. Developmental
language disorders – a follow-up in later adult life. Cognitive,
language and psychosocial outcomes. Journal of Child Psychology and
Psychiatry 46:128-149.
Cole, E.B., Y. Oshima-Takane, and R.L. Yaremko. 1994. Case studies of
pronoun development in two hearing-impaired children: Normal,
delayed or deviant? European Journal of Disorders of Communication
29:113-129.
Conti-Ramsden, G. 2003. Processing and linguistic markers in young
children with specific language impairment (SLI). Journal of Speech,
Language, and Hearing Research 46:1029-1038.
Conti-Ramsden, G., A. Crutchley, and N. Botting. 1997. The extent to
which psychometric tests differentiate subgroups of children
with specific language impairment. Journal of Speech, Language, and
309
References
Conti-Ramsden, G. and A. Hesketh. 2003. Risk markers for SLI: A study

of young language-learning children. International Journal of
Language and Communication Disorders 38:251-63.
Conti-Ramsden, G., E. Knox, N. Botting, and Z. Simkin. 2002.
Educational placements and National Curriculum Key Stage 2
test outcomes of children with a history of specific language
impairment. British Journal of Special Education 29:76-82.
Cooley, C.H. 1908. A study of the early use of self-words by a child.
Psychological Review 15:339-357.
Corver, N. and J. Nunes. 2007. The copy theory of movement.
Amsterdam/Philadelphia: John Benjamins.
Crago, M.B. and S.E.M. Allen. 1994. Morphemes gone askew: Linguistic
impairment in Inuktitut. McGill Working Papers in Linguistics
10:206-215.
Crago, M.B. and S.E.M. Allen. 2001. Early finiteness in Inuktitut: The
role of language structure and input. Language Acquisition 9:59-
111.
Crago, M. and J. Paradis. 2003. Two of a kind? The importance of
commonalities and variation across languages and learners. In Y.
Levy and J. Schaeffer (eds.) Language competence across populations.
Toward a definition of specific language impairment. Mahwah: Lawrence
Erlbaum Associates. pp 97-110.
Crystal, D., P. Fletcher, and M. Garman. 1976. The grammatical analysis of
language disability. London: Edward Arnold.
Crystal, D., P. Fletcher, and M. Garman. 1989. The grammatical analysis of
language disability. Second edition. London: Edward Arnold.
Dalalakis, J. 1994. Developmental language impairment in Greek. McGill
Working Papers in Linguistics 10:216-227.
Dalalakis, J. 1997. Developmental language impairment (DLI) and
dimunitive formation in Greek. Boston University Conference on
Language Development Proceedings 21:115-123.
Dalalakis, J. 1999. Morphological representation in specific language
impairment: Evidence from Greek word formation. Folia
Phoniatrica et Logopaedica 51:20-35.
De Jong, J. 1999. Specific language impairment in Dutch: Inflectional
morphology and argument structure. Doctoral dissertation,
University of Groningen.
De Jong, J. 2003. Specific language impairment and linguistic
explanation. In Y. Levy and J. Schaeffer (eds.) Language competence
310
References
across populations. Toward a definition of specific language impairment.

Mahwah: Lawrence Erlbaum Associates. pp. 151-170.
De Jong, J. 2004. Grammatical impairment: An overview and a sketch of
Dutch. In L. Verhoeven and H. Van Balkom (eds.) Classification of
developmental language disorders. Theoretical issues and clinical implications.
Mahwah: Lawrence Erlbaum Associates. pp. 261-281.
Den Besten, H. 2000. The slaves language in the Dutch Cape Colony
and Afrikaans vir. Linguistics 38:949-971.
Den Besten, H. 2006. The origins of the Afrikaans pre-nominal
possessive system(s). In L.L. Thornburg and J.M. Fuller (eds.)
Studies in Contact Linguistics: Essays in honor of Glenn G. Gilbert. New
York: Peter Lang Publishers. pp. 103–124.
Demuth, K. and S. Suzman. 1997. Language impairment in Zulu. Boston
University Conference on Language Development Proceedings 21:124-135.
DeThorne, L.S., B.W. Johnson, and J.W. Loeb. 2005. A closer look at
MLU: What does it really measure? Clinical Linguistics and Phonetics
19:635-648.
DeThorne, L.S. and R.V. Watkins. 2006. Language abilities and
nonverbal IQ in children with language impairment:
Inconsistency across measures. Clinical Linguistics and Phonetics
20:641-658.
De Villiers, M. 1971. Die grammatika van tyd en modaliteit [The grammar of
time and modality]. Cape Town: A.A. Balkema.
De Villiers, J., J. Cahillane, and E. Altreuer. 2006. What can production
reveal about Principle B? In K.U. Deen, J. Nomura, B. Schulz,
and B.D. Schwartz (eds.) The Proceedings of the Inaugural
Conference on Generative Approaches to Language
Acquisition–North America. Honolulu, Hawaii. University of
Connecticut Occasional Papers in Linguistics 4:89-100.
De Villiers, P., J. De Villiers, T. Roeper, and H.N. Seymour. 2001.
Assessing what every 5-year-old should know. Paper presented at
the ASHA Annual Convention, New Orleans, 16 November
2001.
Donaldson, B.C. 1993. A grammar of Afrikaans. Berlin: Mouton de
Gruyter.
Dromi, E., L.B. Leonard, and M. Shteiman. 1993. The grammatical
morphology of Hebrew-speaking children with specific language
impairment: Some competing hypotheses. Journal of Speech and
311
References
Dunn, L.M. and L.M. Dunn. 1981. Peabody Picture Vocabulary Test - Revised.
Bloomington: Pearson Assessments.
Eadie, P.A., M.E. Fey, J.M. Douglas, and C.L. Parsons. 2002. Profiles of
grammatical morphology and sentence imitation in children with
specific language impairment and Downs syndrome. Journal of
Speech, Language, and Hearing Research 45:720-732.
Eisenberg, S.L., T.M. Fersko, and C. Lundgren. 2001. The use of MLU
for identifying language impairment in preschool children: A
review. American Journal of Speech-Language Pathology 10:323-242.
Embick, D. 2007. Blocking effects and analytic/synthetic alternations.
Natural Language and Linguistic Theory 25:1-37.
Embick, D. and A. Marantz. 2007. Architecture and blocking. Available
online at http://www.ling.upenn.edu/~embick/2block.pdf.
Embick, D. and R. Noyer. 2001. Movement operations after syntax.
Linguistic Inquiry 32:555-595.
Fellbaum, C., S. Miller, S. Curtiss, and P. Tallal. 1995. An auditory
processing deficit as a possible source of SLI. Boston University
Conference on Language Development Proceedings 19:204-215.
Finestack L.H., M.E. Fey, and H.W. Catts. 2006. Pronominal reference
skills of second and fourth grade children with language
impairment. Journal of Communication Disorders 39:232-248.
Fukuda, S.E. and S. Fukuda. 1994. Developmental language impairment
in Japanese: A linguistic investigation. McGill Working Papers in
Linguistics 10:150-177.
Fukuda, S. and S.E. Fukuda. 2001. The acquisition of complex predicates
in Japanese specifically language-impaired and normally
developing children. Brain and Language 77:305-320.
Goad, H. and C. Rebellati. 1994. Pluralization in specific language
impairment: Affixation or compounding? McGill Working Papers
in Linguistics 10:24-40.
Gopnik, M. 1990a. Feature blindness: A case study. Language Acquisition
1:139-164.
Gopnik, M. 1990b. Feature-blind grammar and dysphasia. Nature
344:715.
Gopnik, M. 1994a. Impairments of syntactic tense in a familial language
disorder. McGill Working Papers in Linguistics 10:67-80.
Gopnik, M. 1994b. The family. McGill Working Papers in Linguistics 10:1-4.
312
References
Gualmini, A. and S. Crain. 2002. Why no child or adult must learn De

Morgan’s laws. Boston University Conference on Language Development
Proceedings 26:243-254.
Gualmini, A., S. Crain, and L. Meroni. 2000. Acquisition of disjunction
in conditional sentences. Boston University Conference on Language
Development Proceedings 24:367-378.
Haegeman, L. 1994. Introduction to government and binding theory. Second
edition. Cambridge and Oxford: Blackwell.
Håkansson, G. 1998. Language impairment and the realization of
finiteness. Boston University Conference on Language Development
Håkansson, G. 2001. Tense morphology and verb-second in Swedish L1
children, L2 children and children with SLI. Bilingualism: Language
and Cognition 4:85-99.
Halle, M. and A. Marantz. 1993. Distributed morphology and the pieces
of inflection. In K. Hale and S.J. Keyser (eds.) The view from
Building 20. Cambridge: MIT Press. pp. 111-176.
Hammill, D.D. and P.L Newcomer. 1997. Test of language development -
Primary. Third edition. Texas: Pro-Ed.
Hansson, K. 1997. Patterns of verb use in Swedish children with SLI: An
application of recent theories. First Language 17:195-217.
Hansson, K. and L.B. Leonard. 2003. The use and productivity of verb
morphology in specific language impairment: An examination of
Swedish. Linguistics 41:351-379.
Hansson, K. and U. Nettelbladt. 1995. Grammatical characteristics of
Swedish children with SLI. Journal of Speech and Hearing Research
38:589-598.
Hansson, K., U. Nettelbladt, and L. Leonard. 2000. Specific language
impairment in Swedish: The status of verb morphology and word
order. Journal of Speech, Language, and Hearing Research 43:848-864.
Hickey, T. 1991. Mean length of utterance and the acquisition of Irish.
Journal of Child Language 18:553-569.
Hornstein, N. 2001. Move! A minimalist theory of construal. Malden and
Oxford: Blackwell.
Hornstein, N., J. Nunes, and K.K. Grohmann. 2005. Understanding
minimalism. Cambridge: Cambridge University Press.
Hunt, K.W. 1970. Syntactic maturity in school children and adults.
Monographs of the Society for Research in Child Language Development
(35)1. Serial no. 134.
313
References
Jakobson, R. 1971. Studies on child language and aphasia. The Hague and
Paris: Mouton.
Jakubowicz, C. 2003. Computational complexity and the acquisition of
functional categories by French-speaking children with SLI.
Linguistics 41:175-211.
Johnston, J.R. 2001. An alternate MLU calculation: Magnitude and
variability effects. Journal of Speech, Language, and Hearing Research
44:156-164.
Johnston, J.R., J.F. Miller, S. Curtiss, and P. Tallal. 1993. Conversations
with children who are language impaired: Asking questions.
Journal of Speech and Hearing Research 36:973-978.
Joseph, K., J. Pine, and G. Conti-Ramsden. 2002. Case and agreement
errors in children with SLI. Paper presented at the Joint
Conference of IASCL-SRCLD, Madison, 16-21 July 2002.
Kail, R. 1994. A method of studying the generalized slowing hypothesis
in children with specific language impairment. Journal of Speech and
Kail, R. and T. Salthouse. 1994. Processing speed as a mental capacity.
Acta Psychologica 86:199-225.
Kanno, K. 1998. The stability of principles in second-language
acquisition: Evidence from Japanese. Linguistics 36:1125-1146.
Kayne, R.S. 1993. Toward a modular theory of auxiliary selection. Studia
Linguistica 47:3-31.
Kayne, R.S 1994. The antisymmetry of syntax. Cambridge: MIT Press.
Klee, T. and M.D. Fitzgerald. 1985. The relation between grammatical
development and mean length of utterance in morphemes.
Koster, J. 1975. Dutch as an SOV language. Linguistic Analysis 1: 111-136.
Krassowski, E. and E. Plante. 1997. IQ variability in children with SLI:
Implications for use of cognitive referencing in determining SLI.
Journal of Communication Disorders 30:1-9.
Lahey, M., J. Liebergott, M. Chesnick, P. Menyuk, and J. Adams. 1992.
Variability in children’s use of grammatical morphemes. Applied
Psycholinguistics 13:373-398.
Lasnik, H. and J. Uriagereka. 2005. A course in minimalist syntax.
Foundations and prospects. Malden: Blackwell.
Laws, G. and D.V.M. Bishop. 2004. Verbal deficit in Down’s syndrome
and specific language impairment: A comparison. International
Journal of Language and Communication Disorders 39:423-451.
314
References
Leadholm, B.J. and J.F. Miller. 1992. Language sample analysis: The
Wisconsin guide. Milwaukee: Wisconsin Department of Public
Instruction.
Legate, J.A. 2002. Phases in “Beyond Explanatory Adequacy”.
Manuscript, MIT.
Leonard, L. 1989. Language learnability and specific language
impairment. Applied Psycholinguistics 10:179-202.
Leonard L.B. 1991. Comments on specific language impairment as a
clinical category. Language, Speech, and Hearing Services in Schools
22:84-87.
Leonard, L.B. 1992. Specific language impairment in three languages:
Some cross-linguistic evidence. In P. Fletcher and D. Hall (eds.)
Specific speech and language disorders in children. London: Whurr
Publishers. pp. 118-126.
Leonard. L.B. 1994. Some problems facing accounts of morphological
deficits in children with specific language impairment. In R.V.
Watkins and M.L. Rice (eds.). Specific language impairments in
children. Baltimore: Paul H. Brookes. pp. 91-105.
Leonard, L.B. 1995. Functional categories in the grammars of children
with specific language impairment. Journal of Speech and Hearing
Research 38:1270-1283.
Leonard, L.B. 1998. Children with specific language impairment. Cambridge
and London: MIT Press.
Leonard, L.B. 2000. Specific language impairment across languages. In
D.V.M. Bishop and L.B. Leonard (eds.) Specific language impairment
in children: Causes, characteristics, intervention and outcome. Philadelphia:
Taylor and Francis Inc.
Leonard, L.B. 2003. Specific language impairment: Characterizing the
deficit. In Y. Levy and J. Schaeffer (eds.) Language competence across
populations. Toward a definition of specific language impairment. Mahwah:
Lawrence Erlbaum Associates. pp. 209-231.
Leonard, L.B., U. Bortolini, M.C. Caselli, K.K. McGregor, and L.
Sabbadini. 1992. Morphological deficits in children with specific
language impairment: The status of features in the underlying
grammar. Language Acquisition 2:151-179.
Leonard, L.B. and J.A. Eyer. 1996. Surface properties of grammatical
morphology and morphological deficits in children with specific
language impairment. In J. Morgan and K. Demuth (eds.) Signal to
syntax. Hillsdale:Lawrence Erlbaum. pp. 233-247.
315
References
Leonard, L.B., J.A. Eyer, L.M. Bedore, and B.G. Grela. 1997. Three
accounts of the grammatical morpheme difficulties of English-
speaking children with specific language impairments. Journal of
Speech, Language, and Hearing Research 40:741-753.
Leonard, L.B. and D. Finneran. 2003. Grammatical morpheme effects
on MLU: “The same can be less” revisited. Journal of Speech,
Leonard, L.B., C.A. Miller, L. Rauf, M. Charest, and R Kurtz. 2003.
Surface forms and grammatical functions: Past tense and passive
particle use by children with specific language impairment. Journal
Leonard, L.B., L. Sabbadini, J.S. Leonard, and V. Volterra. 1987. Specific
language impairment in children: A cross-linguistic study. Brain
and Language 32:233-252.
Leonard, L.B., L. Sabbadini, V. Volterra, and J.S. Leonard. 1988. Some
influences on the grammar of English- and Italian-speaking
children with specific language impairment. Applied
Leonard, L.B., E-K. Salameh, and K. Hansson. 2001. Noun phrase
morphology in Swedish-speaking children with SLI. Applied
Lightfoot, D. 1979. Principles of diachronic syntax. Cambridge: CUP.
Lightfoot, D. 1991. How to set parameters: Arguments from language change.
Cambridge: MIT Press.
Lightfoot, D. 1999. The development of language: Acquisition, change and
evolution. Oxford: Blackwell.
Lin, Y-A. 2006. Reflection on the Agreement and Tense Omission
Model of specific language impairment: A corpus-based study.
Paper presented at the Newcastle/Durham Post-graduate
Conference in Linguistics, Newcastle, 25 June 2006.
Linder, K. and J. Johnston. 1992. Grammatical morphology in language-
impaired children acquiring English or German as their first
language: A functional perspective. Applied Psycholinguistics 13:115-
129.
Ljubešić, M. and M. Kovačević. 1992. Some insights into specific
language impairment in Croatian. Scandinavian Journal of Logopedics
and Phoniatrics 17:37-43.
Loban, W. 1976. Language development: Kindergarten through grade twelve.
Urbana: University of Illinois Press.
316
References
Loeb, D.F. and L. Leonard. 1991. Subject case marking and verb
morphology in normally developing and specifically language-
impaired children. Journal of Speech and Hearing Research 34:340-346.
Marchman, V.A., C. Saccuman, and B. Wulfeck. 2004. Productive use of
the English past tense in children with focal brain injury and
specific language impairment. Brain and Language 88:202-214.
Marchman, V.A., B. Wulfeck, and S.E. Weismer. 1999. Morphological
productivity in children with normal language and SLI: A study
of the English past tense. Journal of Speech, Language, and Hearing
Research 42:206-219.
McGuckian, M. 2004. The grammatical morpheme deficit in children
with hearing impairment, children with Down's syndrome and
children with specific language impairment. Doctoral
dissertation, University of Ulster.
Merino, B. 1983. Language development in normal and language
handicapped Spanish-speaking children. Hispanic Journal of
Behavioral Sciences 5:379-400.
Miller, J. 1981. Assessing language production in children. Baltimore: University
Park Press.
Miller, J.F. and R.S. Chapman. 1981. Research note. The relation
between age and mean length of utterance in morphemes. Journal
of Speech and Hearing Research 24:154-161.
Miller, C.A. and P. Deevy. 2003. A method for examining productivity of
grammatical morphology in children with and without specific
language impairment. Journal of Speech, Language, and Hearing
Research 46:1154-1166.
Molnárfi, L. 2002. Focus and antifocus in modern Afrikaans and West-
Germanic. Linguistics 40:1107-1160.
Moore, M.E. 1995. Error analysis of persons by normal and language-
impaired children. Journal of Communication Disorders 28:57-72.
Moore, M.E. 2001. Third person pronoun errors by children with and
without language impairment. Journal of Communication Disorders
34:207-226.
Morgan, G., R. Herman, and B. Woll. 2007. Language impairment in
sign language: Breakthroughs and puzzles. International Journal of
Language and Comminication Disorders 42:97-105.
Nunes, J. 1995. The copy theory of movement and linearization of
chains in the minimalist program. Doctoral dissertation,
University of Maryland.
317
References
Nunes, J. 2004. Lineration of chains and sideward movement. Cambridge and

London: MIT Press.
Oetting, J.B. and J.E. Horohov. 1997. Past-tense marking in children
with and without specific language impairment. Journal of Speech,
Oetting, J.B. and M.L. Rice. 1993. Plural acquisition in children with
specific language impairment. Journal of Speech and Hearing Research
36:1236-1248.
O’Grady, W. 1997. Semantics: The analysis of meaning. In W. O’Grady,
M. Dobrovolsky, and F. Katamba (eds.) Contemporary linguistics:
An introduction. London and New York: Longman. pp. 269-312.
Oosthuizen, J. 1996. Minimalisme en woordordevariasie in Afrikaanse
vraagsinne [Minimalism and word order variation in Afrikaans
question sentences]. Stellenbosch Papers in Linguistics 29:71-98.
Oosthuizen, J. 1998. The final nie in Afrikaans negative sentences.
Stellenbosch Papers in Linguistics 31:61-93.
Oosthuizen, J. 2000. Prepositions left and right in Afrikaans. Stellenbosch
Papers in Linguistics 33:67-90.
Oosthuizen, J. 2006. Towards explaining reflexivity with minimalist
means. Paper presented at the Annual Conference of the
Linguistic Society of South Africa, Durban, 7 July 2006.
Oosthuizen, J. (forthcoming). Verpligte refleksiwiteit in Afrikaans: ’n
Minimalistiese verkenning [Compulsory reflexivity in Afrikaans:
A minimalist exploration]. To appear in Stellenbosch Papers in
Linguistics.
Oosthuizen, J. and H. Waher. 1994. On the syntax of the se-construction
in Afrikaans. Stellenbosch Papers in Linguistics 28:21-43.
Oshima-Takane, Y. 1992. Analysis of pronominal errors: A case study.
Paradis, J. and M. Crago. 2000. Tense and temporality: A comparison
between children learning a second language and children with
SLI. Journal of Speech, Language, and Hearing Research 43:834-837.
Paradis, J. and M. Crago. 2001. The morphosyntax of specific language
impairment in French: An extended optional default account.
Language Acquisition 9:269-300.
Paradis, J., M. Crago, and F. Genesee. 2002. Object clitics as a clinical
marker of SLI in French: Evidence from French-English
bilingual children. Boston University Conference on Language
Development Proceedings 27:638-649.
318
References
Paradis, M. and M. Gopnik. 1994. Compensatory strategies in familial

language impairment. McGill Working Papers in Linguistics 10:142-
149.
Paul, R. and S. Alforde. 1993. Grammatical morpheme acquisition in 4-
year-olds with normal, impaired, and late-developing language.
Pine, J.M, K.L. Joseph, and G. Conti-Ramsden. 2004. Do data from
children with specific language impairment support the
Agreement/Tense Omission Model? Journal of Speech, Language,
and Hearing Research 47:913-23.
Pine, J., C. Rowland, E. Lieven, and A. Theakston. 2002. Testing the
agreement/tense omission model: Why the data on children’s use
of non-nominative subjects count against the ATOM. Paper
presented at the Joint Conference of IASCL-SRCLD, Madison,
16-21 July 2002.
Pinker, S. 1994. The Language Instinct: How the Mind Creates Language. New
York: Harper Collins.
Plante, E. 1998. Criteria for SLI: The Stark and Tallal legacy and beyond.
Journal of Speech, Language, and Hearing Research 41:951-957.
Plante, E., L. Swisher, B. Kiernan, and M.A. Restrepo. 1993. Language
matches: Illuminating or confounding? Journal of Speech and
Platzack, C. 2001. The vulnerable C-domain. Brain and Language 77:364-
377.
Pollock, J-Y. 1989. Verb movement, UG and the structure of IP.
Linguistic Inquiry 20:365-425.
Ponelis, F.A. 1979. Afrikaanse sintaksis [Afrikaans syntax]. Pretoria: J.L.
Van Schaik.
Pretorius, A. 1989. Die Afrikaanse semantiese taalevalueringsmedium [The
afrikaans semantic language evaluation medium]. Pretoria: A.
Pretorius.
Radford, A. 1997a. Syntax. A minimalist introduction. Cambridge:
Cambridge University Press.
Radford, A. 1997b. Syntactic theory and the structure of English. Cambridge:
Cambridge University Press.
Raidt, E.H. 1969. Die gebruik van vir in objekskonstruksies [The use of
vir in object constructions]. Tydskrif vir Geesteswetenskappe 9:30-49.
Ravid, D., R. Levie, and G.A. Ben-Zvi. 2003. The role of language
typology in linguistic development: Implications for the study of
319
References
language disorders. In Y. Levy and J. Schaeffer (eds.) Language

competence across populations. Toward a definition of specific language
impairment. Mahwah: Lawrence Erlbaum Associates. pp. 171-193.
Records, N. and J.B. Tomblin. 1994. Clinical decision making:
Describing the decision rules of practicing speech-language
pathologists. Journal of Speech and Hearing Research 37:144-156.
Reilly, J., M. Losh, U. Bellugi, and B. Wulfeck. 2004. “Frog, where are
you?” Narratives in children with specific language impairment,
early focal injury, and Williams syndrome. Brain and Language
88:231-247.
Reinhart, T. and E. Reuland. 1993. Reflexivity. Linguistic Inquiry 24:657-
720.
Renfrew, C.E. 1995. Renfrew language scales – Word finding vocabulary test.
Bicester: Winslow Publications.
Renfrew, C.E. 1997. Renfrew language scales – Bus story. Bicester: Winslow
Publications.
Reuland, E. 2001. Primitives of binding. Linguistic Inquiry 32:439-492.
Ricard, M., P.C. Girouard, and T. Gouin Décarie. 1999. Personal
pronouns and perspective taking in toddlers. Journal of Child
Language 26:681-697.
Rice, M.L. 2003. A unified model of specific and general language delay:
Grammatical tense as a clinical marker of unexpected variation.
In Y. Levy and J. Schaeffer (eds.) Language competence across
populations. Toward a definition of specific language impairment. Mahwah:
Lawrence Erlbaum Associates. pp. 63-95.
Rice, M.L. and J.B. Oetting. 1993. Morphological deficits on children
with SLI: Evaluation of number marking and agreement. Journal
of Speech and Hearing Research 36:1249-1257.
Rice, M.L., S.M. Redmond, and L. Hoffman. 2006. Mean length of
utterance in children with specific language impairment and in
younger control children shows concurrent validity and stable
and parallel growth trajectories. Journal of Speech, Language, and
Rice, M. and K. Wexler. 1996. Towards tense as a clinical marker of
specific language impairment in English-speaking children.
Rice, M., K. Wexler, and P. Cleave. 1995. Specific language impairment
as a period of extended optional infinitive. Journal of Speech and
320
References
Rice, M., K. Wexler, and S. Herschberger. 1998. Tense over time: The
longitudinal course of tense acquisition in children with specific
language impairment. Journal of Speech, Language, and Hearing
Research 41:1412-1431.
Rice, M.L., K. Wexler, and S.M. Redmond. 1999. Grammaticality
judgments of an extended optional infinitive grammar: Evidence
from English-speaking children with specific language
impairment. Journal of Speech, Language, and Hearing Research 42:943-
961.
Ricard, M., P.C. Girouard, and T. Gouin Décarie. 1999. Personal
pronouns and perspective taking in toddlers. Journal of Child
Rispoli, M. 1994. Pronoun case overextensions and paradigm building.
Rispoli, M. 1998a. Patterns of pronoun case error. Journal of Child
Rispoli, M. 1998b. Me and my: Two different patterns of pronoun case
errors. Journal of Speech, Language, and Hearing Research 41:385-393.
Rispoli, M. 1999. Case and agreement in English language development.
Rispoli, M. 2002. Theory and methods in the study of the development
of case and agreement: A response to Schütze. Journal of Child
Rispoli, M. 2005. When children reach beyond their grasp: Why some
children make pronoun case errors and others don’t. Journal of
Child Language 32:93-116.
Rizzi, L. 1997. The fine structure of the left periphery. In L. Haegeman
(ed.) Elements of grammar: Handbook in generative syntax. Dordrecht:
Kluwer. pp. 281-337.
Roberge, P. 1994. On the origins of the Afrikaans verbal hendiadys.
Stellenbosch Papers in Linguistics 28:45-81.
Roberts, I. 1985. Agreement parameters and the development of English
modal auxiliaries. Natural Language and Linguistic Theory 3:21-58.
Roberts, I. 1993. Verbs and diachronic syntax. Dordrecht: Kluwer.
Roberts, I. 2007. Diachronic syntax. Oxford: Oxford University Press.
Roberts, J. and L. Rescorla. 1995. Morphological acquisition and SLI:
Evidence from children with expressive language delay. Boston
University Conference on Language Development Proceedings 19:475-486.
321
References
Roediger, H. 1980. Memory metaphors in cognitive psychology. Memory

and Cognition 8:231-246.
Rom, A. and L.B. Leonard. 1990. Interpreting deficits in grammatical
morphology in specifically language-impaired children:
Preliminary evidence from Hebrew. Clinical Linguistics and
Phonetics 4:93-105.
Rondal, J.A., M. Ghiotto, S. Bredart, and J. Bachelet. 1987. Age-relation,
reliability and grammatical validity of measures of utterance
length. Journal of Child Language 14:433-446.
Rose, Y. and P. Royle. 1999. Uninflected structure in familial language
impairment: Evidence from French. Folia Phoniatrica et Logopaedica
51:70-90.
Ross, J. 1967. Constraints on variables in syntax. Doctoral dissertation,
MIT.
Schiff-Myers, N. 1983. From pronoun reversals to correct pronoun
usage: A case study of a normally developing child. Journal of
Speech and Hearing Disorders 48:385-394.
Schütze, C.T. 1999. Different rates of pronoun case error: Comments on
Rispoli (1998). Journal of Child Language 26:749-55.
Schütze, C. and K. Wexler. 1996. Subject case licensing and English root
infinitives. Boston University Conference on Language Development
Simon-Cereijido, G. and V.F. Gutierrez-Clellen. 2007. Spontaneous
language markers of Spanish language impairment. Applied
Southwood, F. 2005. A comparison of the responses to three
comprehension and three production tasks assessing the
morpho-syntactic abilities of Afrikaans-speaking preschoolers.
Per Linguam 21:36-59.
Southwood F. 2006. An investigation of the morpho-syntactic abilities of
Afrikaans-speaking preschoolers. South African Linguistic and
Applied Language Studies 24:35-55.
Southwood, F. and A.F. Russell. 2004. Comparison of conversation,
freeplay, and story generation as methods of language sample
elicitation. Journal of Speech, Language, and Hearing Research 47:366-
376.
Stark, R.E. and P. Tallal. 1981. Selection of children with specific
language deficits. Journal of Speech and Hearing Disorders 46:114-122.
322
References
Statistics South Africa. 2003. Census 2001. Census in brief. (No. 03-02-02-
2001). Pretoria: Statistics South Africa.
Stavrakaki, S. 2001. Comprehension of reversible relative clauses in
specifically language impaired and normally developing Greek
children. Brain and Language 77:419-431.
Stavrakaki, S. 2002. A-bar movement constructions in Greek children
with SLI: Evidence for deficits in the syntactic component of
language. In E. Fava (ed.) Clinical Linguistics: Theory and applications
in speech pathology and therapy. Amsterdam and Philadelphia: John
Benjamins. pp. 131-153.
Suzman, S.M. 2002. Morphological accessibility in Zulu. In E. Fava (ed.)
Clinical Linguistics: Theory and applications in speech pathology and
therapy. Amsterdam and Philadelphia: John Benjamins. pp. 154-
174.
Tanaka Welty, Y., J. Watanabe, and L. Menn. 2002. Language production
in Japanese preschoolers with specific language impairment:
Testing theories. In E. Fava (ed.) Clinical Linguistics: Theory and
applications in speech pathology and therapy. Amsterdam and
Philadelphia: John Benjamins. pp. 175-193.
Thal, D.J. 1999. Book review: L.B Leonard, Children with specific
language impairment. Journal of Child Language 26:491-504.
Thordardottir, E.T. and S.E. Weismer. 1998. Mean length of utterance
and other language sample measures in early Icelandic. First
Language 18:1-32.
Tomblin, J., N. Records, P. Buckwalter, X. Zhang, E. Smith, and M.
O’Brien. 1997. Prevalence of specific language impairment in
kindergarten children. Journal of Speech, Language, and Hearing
Research 40:1245-1260.
Tsimpli, I.M. and S. Stavrakaki. 1999. The effects of a morphosyntactic
deficit in the determiner system: The case of a Greek SLI child.
Lingua 108:31-85.
Ullmann, M. and M. Gopnik. 1994. The production of inflectional
morphology in hereditary specific language impairment. McGill
Working Papers in Linguistics 10:81-118.
Unsworth, S. 2005.Child L2, adult L2, child L1: Differences and similarities: A
study on the acquisition of direct object scrambling in Dutch. LOT
Dissertation Series 119. Utrecht: LOT.
Van Daal, J. L. Verhoeven, and H. Van Balkom. 2004. Subtypes of
severe speech and language impairments: Psychometric evidence
323
References
from 4-year-old children in the Netherlands. Journal of Speech,

Van der Lely, H.K.J. 1994. Canonical linking rules: Forward versus
reverse linking in normally developing and specifically language-
impaired children. Cognition 51:29-72.
Van der Lely, H.K.J. 1996. Specifically language impaired children and
normally developing children: Verbal passive vs. adjectival
passive sentence interpretation. Lingua 98:243-272.
Van der Lely, H.K.J. 1998. SLI in children: Movement, economy and
deficits in the computational-syntactic system. Language
Acquisition 7:161-192.
Van der Lely, H.K.J. 2003. Do heterogeneous deficits require
heterogeneous theories? SLI subgroups and the RDDR
hypothesis. In Y. Levy and J. Schaeffer (eds.) Language competence
across populations. Toward a definition of specific language impairment.
Mahwah: Lawrence Erlbaum. pp. 111-133.
Van der Lely, H.K.J. 2004. Evidence for and implications of a domain-
specific grammatical deficit. In L. Jenkins (ed.) The genetics of
language. Oxford: Elsevier. pp. 117-145.
Van der Lely, H.K.J. 2005. Grammatical specific language impairment
(G-SLI): Identifying and characterising the G-SLI group.
Fréquence 17:13-20.
Van der Lely, H.K.J. and L. Stollwerck. 1997. Binding theory and
grammatical specific language impairment in children. Cognition
62:245-290.
Van der Lely, H.K.J. and M. Ullman. 1996. The computation and
representation of past-tense morphology in normally developing
and specifically language impaired children. Boston University
Conference on Language Development Proceedings 20:816-827.
Van der Lely, H.K.J. and M. Ullman. 2001. Past tense morphology in
specifically language impaired children and normally developing
children. Language and Cognitive Processes 16:177-217.
Vinkler, Z. and C. Pléh. 1995. A case of a specific language impaired
child in Hungarian. In M. Kovačević (ed.) Language and language
communication barriers: Research and theoretical perspectives in three
European languages. Zagreb: Croatian University Press. pp. 131-
158.
324
References
Voci, S.C., J.H. Beitchman, E.B. Brownlie, and B. Wilson. 2006. Social
anxiety in late adolescence: The importance of early childhood
language impairment. Journal of Anxiety Disorders 20:915-930.
Vorster, J. 1980. Toets vir mondelinge taalproduksie [Test for oral language
production]. Pretoria: South African Institute for Psychological
and Psychometric Research.
Waher, H. 1982. The position of the finite verb in Afrikaans. Stellenbosch
Papers in Linguistics 8:51-78.
Weckerly, J., B. Wulfeck, and J. Reilly. 2004. The development of
morphosyntactic ability in atypical populations: The acquisition
of tag questions in children with early focal lesions and children
with specific language impairment. Brain and Language 88:190-201.
Wexler, K. 1994. Finiteness and head movement in early child grammars.
In D. Lightfoot and N. Hornstein (eds.) Verb movement. New
York: Cambridge University Press. pp. 305-350.
Wexler, K. 1998. Very early parameter setting and the unique checking
constraint: A new explanation of the optional infinitive stage.
Lingua 106:23-79.
Wexler, K., C. Schütze, and M. Rice. 1998. Subject case in children with
SLI and unaffected controls: Evidence for the Agr/Tns omission
model. Language Acquisition 7:317-344.
Wilsenach, C. 2006. Syntactic processing in developmental dyslexia and in specific
language impairment. A study on the acquisition of the past participle
construction in Dutch. LOT Dissertation Series 128. Utrecht: LOT.
Zimmerman, I.L., V.G. Steiner, and R.E. Pond. 1992. PLS-3: Preschool
language scale-3. San Antonio: The Psychological Corporation.
Zwart, C.J-W. 1993. Dutch syntax. A minimalist approach. Doctoral
dissertation, Rijksuniversiteit Groningen.
Zwart, J-W. 1994. Dutch is head initial. The Linguistic Review 11:377-406.
Zwart, J-W. 2002. Issues relating to a derivational theory of binding. In
S.D. Epstein and T.D. Seely (eds.) Derivation and explanation in the
Minimalist Program. Oxford: Blackwell.
325
Appendices
APPENDIX A: INFORMATION ON SLI PARTICIPANTS’
LANGUAGE TEST RESULTS (AST, ARW, TMT)
AST
Inclusion/Ex-
Multiple word
Synonymy [E]
S A T
Concept for-
mation [R,E]
Comparative
relations [R]
Word deini-
Prond [R,E]
Vocabulary
L CAa R M
clusion [R]
tions [E]bc
meanings
I W T
[R,E]
[R]b
1 67 - - - - - - - - 83e -
2 76 - - - - - - - - <68 -
3 62 - - - - - - - - - -
4 80 60- - - - - - - - - -
65
5 76 54- 48- - 60- 36- 51-71 - - - -
59 56 62 47
6 72 - - - - - - - - - -
7 72 54- <36 - - - - 36- 48- - -
59 47 59
8 57 - - - - - - - - - -
9 77 - - - - - - - - - -
10 77 - - - - - - - - - -
11 62 54- 36- 36- 60- 48- 36-50 - - - -
59 47 41 62 50
12 67 60- 36- - - - 56- 72- - -
65 47 65 77
13 61 48- 36- - - - <36 60- - -
53 47 68
14 75 - - - - - - - - -
15 77 - - - - - - - - -
aCA=chronological age at the time of testing, measured in months. bR=receptive.
cE=expressive. dPron=pronouns. eTest scores given as age equivalents, in months.
327
Appendices
APPENDIX A (continued)
TESTS OTHER THAN AST, ARW, AND TMT
TACL-R/III
aggregate
Renfrew Informal testing Other
SLI
CA
GrMb
Word
EPSc
Only
Cla
1 67 -- -- -- Age-ap- Word order anduse of --

propriate grammatical morphemes
sentence deviant
lenth
2 76 -- -- -- -- Ed use of plurals, --
pronouns problematic. Re
langf at least 8 months
delayed; both R and E
age-inappropriate. Over-
use circumlocution.
3 62 51- 41- 53- -- -- --
54 44 55
4 80 -- -- -- 39, Word E4 lang possibly at same --
finding level as R.
problems
5 76 -- -- -- -- -- MLU=5.7
6 72 34 <36 <36 -- -- LARSP 24-
30
7 72 -- -- -- -- -- --
8 57 -- -- -- -- -- PLS-3 R
58, E 48
9 77 -- -- -- -- -- TOLD-P R
72, E 54
10 77 66 -- E lang 60 --
11 62 -- -- -- -- -- MLU=4.4
12 67 -- -- -- -- -- --
13 61 -- -- -- -- -- --
14 75 57 48 48 -- E lang 57 --
15 77 102 66 78 -- E lang 57. Poor E vocab.g --
Over-use circumlocution.
Omit plurals, diminutives.
Confuse personal pron.h
aWord Cl=word classes. bGr M=grammatical morphemes. cEPS=elaborated phrases and
sentences. dE=expressive. eR=receptive. fLang=language. gVocab=vocabulary. hPron=pronouns.
328
Appendices
APPENDIX B: LETTER TO PARENTS: INFORMATION AND

CONSENT
INFORMATION AND INFORMED CONSENT DOCUMENT
TITLE OF THE RESEARCH PROJECT: Specific language impairment in

Afrikaans
MAIN RESEARCHER: Frenette Southwood
ADDRESS: Room 518, Arts Building, c/o Merriman Avenue and Van
Ryneveld Street, Stellenbosch OR Department of General Linguistics,
Stellenbosch University, Private Bag X1, Matieland, 7602
DECLARATION ON BEHALF OF PARTICIPANT:

I, THE UNDERSIGNED, ………………………………..………….(name)
[ID No: …………..……] in my capacity of ……………………………….. of the
participant [ID No: ………………..] of ………………………………..…
…….………………………………..………….…………………… (address)
A. HEREWITH CONFIRM THE FOLLOWING:
1. The participant was invited to partake in the above-mentioned research project
undertaken by the Department of General Linguistics, Arts Faculty,
Stellenbosch University.
2. The following aspects were explained to me:
2.1 Aim: The aim of the project is to establish how language-impaired Afrikaans-
speaking children understand and use the following: grammatical morphemes
(such as the –e or –s which indicates plural), interrogative sentences (specifically
those starting with wh-question words such as wat, wie or wanneer), passive
sentences (such as die baba word/is gebad of die koeldrank is deur die pa ingeskink),
and sentences in which so-called constituent movement took place, in other
words sentences with word order changes (such as gister het ons geswem – where
het ons geswem occurs instead of ons het geswem). A great deal of research has been
done on these aspects with English-speaking children, but there is a serious lack
of information on the language of Afrikaans-speaking children with language
impairment. Without information on how the language of Afrikaans-speaking
language-impaired children looks, it is very difficulat to evaluate and remediate
the language of these children effectively.
2.2 Procedures: There are three main activities which will be performed with 15
language-impaired Afrikaans-speaking 6-year-olds, 15 typically developing
Afrikaans-speaking 6-year-olds, and 15 younger typically developing Afrikaans-
speaking children. The first main activity is to perform a hearing screening test
in order to ensure that the child has normal hearing. This is a short, non-
invasive and completely painless procedure during which soft sounds are played
to the child through head phones and he/she has to indicate when he/she
hears the sounds. This activity will take about 10 minutes. The second main
329
Appendices
activity entails the collection of a language sample. This will be done by

conversing with the child while playing with toys such as a broken doll’s house,
little figurines, Lego, puppets, etc. This activity will take about 30 minutes. The
third main activity entails the performance of language tasks. These tasks are
aimed at testing the comprehension and use of certain aspects of language. For
the comprehension test, the following would be a typical task: The child is
introduced to a puppet and told that the puppet is still in the process of
acquiring Afrikaans. The child is then requested to correct the puppet each time
he says something wrong. Then the child and the puppet jointly look at pictures
which are sometimes labeled/described correctly and sometimes incorrectly by
the puppet. For testing the use of specific aspects of language, the following
would be a typical task: The researcher and the child look at pictures which the
researcher starts to label/describe. The child then completes the sentence
initiated by the researcher. The completion of these tasks requires a certain
degree of concentration. To avoid fatigue, these tasks will be performed over
two or more sessions of 30 minutes or longer.
The interaction during the second and third main activities will be audio-
recorded. The child’s utterances on the cassettes will then be transcribed so that
the researcher has a written version of them. After completion of the study, the
cassettes will be locked away in the researcher’s office. Should you prefer that
the cassettes of your child be destroyed or should you be interested in copies of
the cassettes of your child, please inform the researcher accordingly.
2.3 Risks: Participation in this study does not pose any unusual risks for you or the
child. The activities do not include physical activities associated with a high risk
for injuries.
2.4 Possible benefits: The child’s participation in the study will not benefit the
child in any direct manner.
2.5 Confidentiality: Any information obtained about the child will be treated as
strictly confidential. The results of the study will be written up in the form of
articles and will be submitted for publication in a scientific journal (such as the
South African Journal of Communication Disorders). The results will also be
written up in the form of a doctoral dissertation. As is customary in scientific
reporting, all results will be presented in such a manner that the child will by no
means be identifiable.
2.6 Access to findings: Should you request it, you will receive a report on the
results of the child’s hearing screening test, language sample collection and
language tests. Please note that the researcher will not proceed with the
language sample collection and language tests if the child fails the hearing
screening test. In such a case, you will still receive the results of the hearing
screening test, together with the names of audiologists practising in the area, so
that you can arrange for a diagnostic hearing test, should you wish to do so.
2.7 Voluntary participation/refusal/withdrawal: The child is under no
obligation to participate in this study. Participation is voluntary. Refusal to
participate will in no way influence your or the child’s current or future
330
Appendices
treatment at this institution. If you consent to participation, you and/or the

child may still at any time indicate that participation will be terminated (reasons
for the decision need not be provided); the request will be honoured, and all
activities will be terminated immediately.
3. The above information was explained to me in Afrikaans by Frenette
Southwood and I am proficient in this language. I was offered an opportunity
to ask questions and all questions were answered satisfactorily.
4. Neither I nor the child has been coerced to consent to participation and I and
the child understand that I and/or the child can withdraw consent at any stage,
without penalisation.
5. Participation in the project will not lead to any additional costs for me.
B HEREWITH CONSENT VOLUNTARILY TO THE POTENTIAL

PARTICIPANT TAKING PART IN THE ABOVE-MENTIONED
PROJECT.
Signed/confirmed at .......................................... (place) on .......................20...... (date)
.................................................... ..................................................
Signature or right thumb print of Signature of witness
representative of participant
DECLARATION BY RESEARCHER:
I, Frenette Southwood, declare that
• I explained the information in this document to .........................................
................................. (name of the representative of the participant);
• she/*he was encouraged and given sufficient time to ask me any questions;
• this conversation took place in Afrikaans and that no interpreter was used.
Signed at .......................................... on .............................................. 20......

(place) (date)
......................................................... ...................................................
Signature of researcher Signature of witness
* Delete if not relevant
IMPORTANT MESSAGE TO THE REPRESENTATIVE OF THE

PARTICIPANT:
Dear Representative of the participant
Thank you very much for your and the participant’s involvement in this study.
Should an emergency situation occur due to this research, or you need any further
information about the project, at any stage during the duration of the project,
please contact Frenette Southwood on telephone number 0826631132.
331
Appendices
APPENDIX C: CASE-HISTORY FORM
INFORMATION FORM
PLEASE PROVIDE THE INFORMATION IN THE SPACES PROVIDED AND,

WHERE APPLICABLE, CIRCLE THE APPROPRIATE ANSWER. PLEASE NOTE
THAT IS FORM CONSISTS OF 4 PAGES AND THAT THE FORM NEEDS TO BE
SIGNED ON PAGE 4. THANK YOU VERY MUCH FOR YOUR TROUBLE.
Date on which the form was completed:
Information about you (the person completing the form):

Name:
Relationship to the child (e.g., mother, custodian, grandfather):
Contact telephone number:
Mother-tongue:
Information about your household:

Number of adults in your household:
Number of children in your household:
Age, gender, and mother-tongue of all the children in your household
[including the child who is possibly going to participate in the study]
Age Gender Mother-tongue
1.
2.
3.
4.
5.
6.
Age, gender, and mother-tongue of all the adults in your household
[Draw a cross next to the parent(s)/custodian(s) of the child who is possibly
going to participate in the study]
Age Gender Mother-tongue
1.
2.
3.
4.
5.
6.
332
Appendices
Which of the following languages do the adults speak to each other?

Afrikaans English Other (specify):
Which of the following languages do the children speak to each other?
Which of the following languages do the adults speak to the children?
Which of the following languages do the children speak to the adults?
Information about the child who is possibly going to participate in the

study:
Name:
Date of birth:
Number of older brothers: Number of younger brothers:
Number of older sisters: Number of younger sisters:
Mother-tongue:
Other languages which the child can speak relatively well:
Medical and developmental information about the child who is possibly

going to participate in the study:
Is the child taking any medication at present? Yes No
If so, which medicine?
what for?
Has the child ever had ear infection/middle ear infection? Yes No
If so, how many times in the left ear?
how many times in the right ear?
how long did one episode last on average?
was the child treated for this? Yes No
If the child was treated,
when?
how?
by whom?
was the treatment succesful? Yes No
At what age did the child start to crawl?

At what age did the child start to walk?
Are you /the child’s teacher concerned about the child’s intellectual
development? Yes No
If so, why?
333
Appendices
Does the child have any of the following?

Epilepsy? Yes No
Cerebral palsy? Yes No
Any brain injury? Yes No
Any physical disability? Yes No
Any mental disability? Yes No
Any other chronic condition? Yes No
(If so, specify):
Information about the language and hearing of the child who is possibly
going to participate in the study:
How well does the child hear according to the parent(s):
Good Adequately Poorly
Does the class teacher suspect that the child has
a hearing problem? Yes No
Has the child ever had a hearing test? Yes No
If so, what was the result?
Is the child’s hearing the same from day to day? Yes No
If not, describe:
Does the child experience problems locating the direction

from which sounds come? Yes No
Does it appear that the child gives preference to one ear
while listening? Yes No
If so, to which ear?
At what age did the child say his/her first word?
What was this first word?
Approximately what number of words did the child say at 18 months?
At what age did the child start using 2-word sentences (e.g., Nog sap; Mamma
weg)?
Is the child’s languae at all times comprehensible for

non-family members? Yes No
If not, why not?
How does the child’s language development compare to that of the child’s
older siblings when they were as old as the child is now?
Better The same Poorer
Why do you say so?
334
Appendices
Is the child’s parent(s) concerned about the child’s language

development? Yes No
If so, why?
Would you regard the child’s language development as “normal”

at this moment? Yes No
If not, why not?
Were the child’s parent(s) at any stage concerned about

his/her language development? Yes No
If so, when?
why?
Can the child read? Yes No
Is the child a library member? Yes No
If so, how regularly does the child visit the library?
what number of books does the child take home per visit?
Are story books read to the child? Yes No
If so, by whom?
how regularly?
does the child enjoy this?
Are stories told to the child? Yes No
If so, by whom?
how regularly?
does the child enjoy this?
Has the child ever been referred to or received treatment from any of the
following? If so, when and what for?
Occupational therapist: Yes No
Physiotherapist: Yes No
Hearing therapist: Yes No
Child psychologist: Yes No
Ear, Nose and Throat specialist: Yes No
Neurologist: Yes No
Audiologist: Yes No
Paediatrician: Yes No
Remedial teacher: Yes No
Speech therapist: Yes No
335
Appendices
Information about the child’s main caregivers:

Female caregiver Male caregiver
Relation to the child (e.g.,
mother/grandfather/custodian)
Does the person live in the same
house as the child?
Maritial status
Highest school qualification
Highest post-school qualification
Does the person work at present?
Occupation
Any other information you regard as relevant:
I, ........................................................................, (full names and surname) [ID

number: ...................................................] in my capacity as parent/custodian
(delete what is not applicable) of the participant .....................................
............................... (child’s full names and surname) hereby declare (i) that I
provided the information on this form out of my own free will, (ii) that I was
informed in Afrikaans by Frenette Southwood that the information on this
form will be treated as strictly confidential, (iii) that I was informed in Afrikaans
by Frenette Southwood that I am under no obligation to answer any question
which I deem to be inappropriate, too personal and/or offending, and (iv) that
I was informed in Afrikaans by Frenette Southwood that I will receive a copy
of this form.
...................................................................................... .......................
Signature of the person who completed the form Date
336
Appendices
APPENDIX D: ITEMS OF THE EXPERIMENTAL TASKS

For ease of reference, the items do not appear in random order (i.e., the order
in which they were presented) in this appendix.
1. NUMBER
1.1. Number comprehension (Real words, regular plurals) – Picture selection

Target Singular/plural Semantically Phonologically
form of target related distrator related distrator
Practise items
honde ‘dogs’ hond ‘dog’ katte ‘cats’ hande ‘hands’
arms ‘arms’ arm ‘arm’ vingers ‘fingers’ wurms ‘worms’
Test items
flitse ‘flash-lights’ flits ‘flash-light’ ligte ‘lights’ fietse ‘bicycles’
skape ‘sheep’ skaap ‘sheep’ koeie ‘cows’ skepe ‘ships’
blomme ‘flowers’ blom ‘flower’ blare ‘leaves’ dromme ‘drums’
vurke ‘forks’ vurk ‘fork’ messe ‘knives’ varke ‘pigs’
boeke ‘books’ boek ‘book’ balle ‘balls’ koeke ‘cakes’
handsakke‘handbags’ handsak ‘handbag’ rugsakke ‘backpacks’ handskoene ‘gloves’
seesterre ‘starfish-PL’ seester ‘starfish- dolfyne ‘dolphins’ teestelle ‘tea sets’
SGL’
glyplanke ‘slides’ glyplank ‘slide’ sandbakke ‘sand pits’ wipplanke ‘see-saw’
wasbakke ‘basins’ wasbak ‘basin’ toilette ‘toilets’ washlappe ‘face
cloths’
sambrele ‘umbrellas’ sambreel ‘umbrella’ reënjasse ‘raincoats’ kastele ‘castles’
leeus ‘lions’ leeu ‘lion’ voëls ‘birds’ luis ‘louse’
naels ‘nails’ nael ‘nail’ arms ‘arms’ pels ‘fur’
pa’s ‘dads’ pa ‘dad’ seuns ‘boys’ kaas ‘cheese’
ma’s ‘moms’ ma ‘mom’ meisies ‘girls’ maan ‘moon’
tjops ‘chops’ tjop ‘chop’ wors ‘sausage’ chips ‘chips’
borsels ‘brushes’ Borsel ‘brush’ rekkies ‘elastics’ wortels ‘carrots’
emmers ‘buckets’ emmer ‘bucket’ grafies ‘(little) swemmers
spades’ ‘swimmers’
mielies ‘corn cobs’ mielie ‘corn cob’ appels ‘apples’ kieries ‘canes’
ketels ‘kettles’ ketel ‘kettle’ koppies ‘cups’ sleutels ‘keys’
spykers ‘nails’ spyker ‘nail’ hamers ‘hammers’ suiker ‘sugar’
kat ‘cat’ katte ‘cats’ muis ‘mouse’ skat ‘treasure’
nes ‘nest’ neste ‘nests’ mes ‘knife’ veer ‘feather’
pen ‘pen’ penne ‘pens’ kwas ‘paint brush’ pan ‘pan’
oor ‘ear’ ore ‘ears’ neus ‘nose’ boor ‘drill’
voet ‘foot’ voete ‘feet’ been‘leg’ hoed ‘hat’
seekoei ‘hippo’ seekoeie ‘hippos’ walvis ‘whale’ seekat ‘octopus’
visbak ‘fish bowl’ visbakke ‘fish voëlhok ‘bird cage’ wasbak ‘basin’
bowls’
rusbank ‘couch’ rusbanke ‘couches’ wipplank ‘see-saw’ TV ‘TV’
337
Appendices
badprop ‘bath plug’ badproppe ‘bath waslap ‘face cloth’ takbok ‘reindeer’
plugs’
aarbei ‘strawberry’ aarbeie’strawberries’ roomys ‘ice-cream’ pastei ‘pie’
seun ‘boy’ seuns ‘boys’ man ‘man’ been ‘leg’
man ‘man’ mans ‘men’ vrou ‘woman’ pan ‘pan’
seël ‘stamp’ seëls ‘stamps’ brief ‘letter’ teël ‘tile’
teël ‘tile’ teëls ‘tiles’ plank ‘plank’ tol ‘top’
voël ‘bird’ voëls ‘birds’ vis ‘fish’ kool ‘cabbage’
lepel ‘spoon’ lepels ‘spoons’ klitser ‘egg beater’ sleutel ‘key’
uitveër ‘eraser’ uitveërs ‘erasers’ potlood ‘pencil’ ruitveër ‘wiper’
enkel ‘ankle’ enkels ‘ankles’ arm ‘arm’ winkel ‘shop’
baadjie ‘jacket’ baadjies ‘jackets’ kortbroek ‘shorts’ paadjie ‘narrow path’
dokter ‘doctor’ dokters ‘doctors’ verwer ‘painter’ dogter ‘girl’
1.2. Number comprehension (Real words, regular plurals) – Judgement

Practise items
Een bank, twee banke One couch, two couches
*Een tafel, baie tafele One table, many tables
Test items
Een das, baie dasse One tie, many ties
Een plant, baie plante One plant, many plants
Een foto, baie foto’s One photograph, many photographs
Een bessie, baie bessies One berry, many berries
*Een kar, twee kars One car, many cars
*Een rok, twee roks One dress, two dresses
*Een hoender, twee hoendere One chicken, two chickens
*Een appel, baie appele One apple, many apples
*Een slak, baie slakkes One snail, many snails
*Een tand, baie tandes One tooth, many teeth
*Een sokkie, twee sokkiese One sock, two socks
*Een beker, baie bekerse One mug, many mugs
*Een hond ([hOnt]), baie honte One dog, many dogs
*Een brood ([bro:t]), baie brote One loaf of bread, many loaves of bread
1.3. Number comprehension (Real words, irregular plurals) – Judgement

baddens ‘baths’
sleepwaens ‘trailers’
*kruiwas ‘wheelbarrows’
lammers ‘lambs’
*kalwe ‘calves’
*hawes ‘harbours’
gate ‘holes’
skepe ‘ships’
*stadde ‘cities’
*glasse ‘(drinking) glasses’
berge ‘mountains’
338
Appendices
*Tafelber[x]e ‘Table Mountains’

vlae ‘flags’
*krage ‘collars (of shirts)’
blaaie ‘pages’
*padde ‘roads’
brûe ‘bridges’
mae ‘tummies’
rûe ‘backs’
*oge ‘eyes’
*sage ‘saws’
gesigte ‘faces’
ligte ‘lights’
nagte ‘nights’
*vuise ‘fists’
*bese ‘oxen’
*hempe ‘shirts’
briewe ‘letters’
duiwe ‘doves’
wolwe ‘wolves’
*golfe ‘waves’
*grafe ‘spades’
*stofe ‘stoves’
1.4. Number comprehension (Nonsense words) – Judgement

Een iepta, twee ieptas [ipta] as in [«¥Uma] ‘granny’
*Een sotta, twee sotta [sOta] as in [bOla] ‘(hair) bun’
Een ou wat spal, so een spalaar, twee spalaars [spalAÉr] as in [lerAÉr] ‘reverend’
*Een ou wat oesel, so een oeselaar, twee [us«lAÉr] as in [«¥i«nAÉr] ‘owner’
oeselaar
Hierdie vrou is baie koen. So een koenaard, [kunAÉrt] as in [l«¥iAÉrt] ‘lazy person’
twee koenaards
*Hierdie seun is baie sief. So een siefaard, twee [sifAÉrt] as in [lafAÉrt] ‘coward’
siefaard
Een tonke, twee tonkes [tONk«] as in [lEnt«] ‘spring’
Een fouke, twee foukes [f«¥Uk«] as in [dAm«] ‘lady’
Een seeu, twee seeus [si¥U] as in [li¥U] ‘lion’
*Een fleeu, twee fleeu [fli¥U] as in [spri¥U] ‘starling’
Een bokel, twee bokels [bAÉk«l] as in [tAÉf«l] ‘table’
*Een sietel, twee sietele [sit«l] as in [bOt«l] ‘bottle’
Een korrem, twee korrems [kOr«m] as in [v«r«m] ‘worm’
*Een foutem, twee foutemme [v«¥Ut«m] as in [bU¥«d«m] ‘bottom’
Een kuen, twee kuens [ky¥«n] as in [ly¥«n] ‘lie’
*Een sapen, twee sapenne [sAp«n] as in [bAk«n] ‘beacon’
Een tander, twee tanders [tand«r] as in [hUnd«r] ‘chicken’
*Een suupter, twee suuptere [sypt«r] as in [slaxt«r] ‘butcher’
339
Appendices
*Een nollerd, twee nollerde [nOl«rt] as in [l«¥ip«rt] ‘leopard’

Een spiperd, twee spiperds [sp«p«rt] as in [stand«rt] ‘standard’
Een kwamie, twee kwamies [kwAÉmi] as in [stori] ‘story’
*Een siemettie, twee siemettieë [simEti] as in [famili] ‘family’
*Een metoekier, twee metoekiere [mEtukiÉr] as in [juv«liÉr] ‘jeweller’
Een kêlonnier, twee kêlonniers [kÏlOniÉr] as in [pas«siÉr] ‘passenger’
*Een laap, twee laaps [lAÉp] as in [xAÉp] ‘yawn’
Een slaak, twee slake [slAÉk] as in [smAÉk] ‘taste’
Een puur, twee pure [pyÉr] as in [myÉr] ‘wall’
*Een snuur, twee snuurs [snyÉr] as in [skyÉr] ‘barn’
Een dees, twee dese [des] as in [les] ‘last’
*Een treen, twee treens [tren] as in [sten] ‘stone’
*Een foom, twee fooms [fom] as in [bom] ‘tree’
Een klook, twee kloke [klok] as in [spok] ‘ghost’
*Een mek, twee meks [mEk] as in [hEk] ‘gate’
*Een saf, twee safs [saf] as in [blaf] ‘bark (of dog)’
Een pif, twee piwwe [p«f] as in [r«f] ‘ridge’
Een slerg, twee [slÏrg«] [slÏrx] as in [bÏrx] ‘mountain’
*Een derg, twee [dÏrx«] [slÏrx] as in [bÏrx] ‘mountain’
*Een sil, twee sils [t«l] as in [p«l] ‘pill’
Een tis, twee tisse [t«s] as in [f«s] ‘fish’
Een nal, twee nalle [nal] as in [bal] ‘ball’
*Een gant, twee gants [xant] as in [kant] ‘side’
Een tek, twee tekke [tEk] as in [nEk] ‘neck’
*Een sles, twee sles [slEs] as in [flEs] ‘flask’
Een tiem, twee tieme [tim] as in [kim] ‘germ’
*Een riek, twee rieks [rik] as in [brik] ‘brake’
*Een ok, twee oks [Ok] as in [rOk] ‘dress’
Een don, twee donne [dOn] as in [tOn] ‘ton’
*Een koep, twee koeps [kup] as in [trup] ‘troop’
Een loes, twee loese [lus] as in [hus] ‘cough’
1.5. Number production (Real words, regular plurals) – Sentence completion

Practise items
Hier is een sleutel, maar hier is baie (sleutels) Here is one key, but here are many (keys)
Hier is een rok, maar hier is baie (rokke) Here is one dress, but here are many (dresses)
Test items
Hier is een huis, maar hier is baie (huise) Here is one house, but here are many
(houses)
Hier is een bal, maar hier is baie (balle) Here is one ball, but here are many (balls)
Hier is een skoen, maar hier is baie (skoene) Here is one shoe, but here are many (shoes)
Hier is een pop, maar hier is baie (poppe) Here is one doll, but here are many (dolls)
340
Appendices
Hier is een mes, maar hier is baie (messe) Here is one knife, but here are many
(knives)
Hier is een Kersboom, maar hier is twee Here is one Christmas tree, but here are two
(Kersbome) (Christmas trees)
Hier is een koerant, maar hier is twee Here is one newspaper, but here are two
(koerante) (newspapers)
Hier is een langbroek, maar hier is baie Here is one pair of trousers, but here are
(langbroeke) many (pairs of trousers)
Hier is een oorbel, maar hier is twee Here is one earring, but here are two
(oorbelle) (earrings)
Hier is een gebou, maar hier is baie (geboue) Here is one building, but here are many
(buildings)
Hier is een lêer, maar hier is baie (lêers) Here is one folder, but here are many
(folders)
Hier is een oom, maar hier is baie (ooms) Here is one uncle/man, but here are many
(uncles/men)
Hier is een ghoen, maar hier is baie (ghoens) Here is one marble, but here are many
(marbles)
Hulle kyk een fliek, maar hulle kyk twee They are watching one movie, but they are
(flieks) watching two (movies)
Hier is een tenk, maar hier is twee (tenks) Here is one tank, but here are two (tanks)
Hier is een venster, maar hier is twee Here is one window, but here are two
(vensters) (windows)
Hier is een mandjie; hier is twee (mandjies) Here is one basket, here are two (baskets)
Hier is een bottel, maar hier is baie (bottels) Here is one bottle, but here are many
(bottles)
Hier is een piesang, maar hier is baie Here is one banana, but here are many
(piesangs) (bananas)
Hier is een lekker, maar hier is baie (lekkers) Here is one sweet, but here are many
(sweets)
1.6. Number production (Real words, irregular plurals) – Sentence completion

Hier is een bed, maar hier is twee Here is one bed, but here are two (beds)
(beddens)
Hier is een vrou, maar hier is baie Here is one woman, but here are many
(vroue/vrouens) (women)
Hier is een wa, maar hier is twee (waens) Here is one wagon, but here are two (wagons)
Hier is een ou, maar hier is baie (ouens) Here is one guy, but here are many (guys)
Hier is een kind, maar hier is baie (kinders) Here is one child, but here are many
(children)
Hier is een vat, maar hier is baie (vate) Here is one vat, but here are many (vats)
Hier is een hof, maar hier is twee (howe) Here is one court (of law), but here are two
(courts of law)
Hier is een lid, maar hier is baie (lede) Here is one member, but here are many
(members)
Hy gee een bevel, maar hulle gee baie He is giving one command, but they are
(bevele) giving many (commands)
341
Appendices
Hy sê een gebed, maar hulle sê baie He is saying one prayer, but they are saying
(gebede) many (prayers)
Hier is een kroeg, maar hier is twee (kroeë) Here is one pub, but here are two (pubs)
Hier is een weg, maar hier is twee (weë) Here is one route, but here are two (routes)
Ht vra een vraag, maar hulle vra baie (vrae) He asks one question, but they ask many
(questions)
Hier is ’n pyl en boog. Hier is een boog, Here is a bow and arrow. Here is one bow,
maar hier is twee (boë) but here are two (bows)
Hier is een vlieg, maar hier is baie (vlieë) Here is one fly, but here are many (flies)
Hier is een kas, maar hier is twee (kaste) Here is one cupboard, but here are two
(cupboards)
Hier is een vrug, maar hier is baie (vrugte) Here is one piece of fruit, but here are many
(pieces of fruit)
Hier is een gas, maar hier is baie (gaste) Here is one guest, but here are many (guests)
Hier sien ons een glimlag, maar hier sien Here we see one smile, but here we see many
ons baie (glimlagte) (smiles)
Hier is een insek, maar hier is baie (insekte) Here is one insect, but here are many (insects)
Hier is een dief, maar hier is twee (diewe) Here is one thief, but here are two (thieves)
Hier is een sif, maar hier is twee (siwwe) Here is one sieve, but here are two (sieves)
Hier sien ons ’n gesnyde waatlemoen. Hier Here we see a sliced watermelon. Here is one
is een skyf, maar hier is baie (skywe) slice, but here are many (slices)
Hier is een golf, maar hier is baie (golwe) Here is one wave, but here are many (waves)
(of sea)
Hier is een skroef, maar hier is baie Here is one screw, but here are many (screws)
(skroewe)
Hier is een hond, maar hier is twee (honde) Here is one dog, but here are two (dogs)
Hier is een brood, maar hier is baie (brode) Here is one loaf of bread, but here are many
(loaves of bread)
Hier is een hand, maar hier is baie (hande) Here is one hand, but here are many (hands)
Hier is een rob, maar hier is twee (robbe) Here is one seal, but here are two (seals)
Hier is een web, maar hier is twee (webbe) Here is one web, but here are two (webs)
1.7. Number production (Nonsense words) – Sentence completion

Een assa, twee ... (asas) [asa] as in [«¥UmA] ‘granny’
Een kolla, baie .... (kollas) [kOla] as in [bOla] ‘(hair) bun’
Kyk, hierdie ou sweel. Hy is ’n swelaar, en hier is [swelAÉr] as in [lerAÉr] ‘reverend’
’n klomp ... (swelaars)
Kyk, hierdie ou is besig om te apoen. Hy is ’n [apunAÉr] as in [«¥i«nAÉr] ‘owner’
apoenaar, en hier is twee ... (apoenaars)
Hierdie man is baie bies. Hy is ’n regte biesaard, [bisAÉrt] as in [l«¥iAÉrt] ‘lazy person’
en hier is twee ... (biesaards)
Hierdie meisie is baie kiem. Sy is ’n regte [kimAÉrt] as in [lafAÉrt] ‘coward’
kiemaard], en hier is twee ... (kiemaards)
Een lerke, baie ... (lerkes) [lÏrk«] as in [pÏrsk«] ‘peach’
Een [tiese], baie ... (tieses) [tis«] as in [dAm«] ‘lady’
Een beeu, twee ... (beeus) [bi¥U] as in [li¥U] ‘lion’
342
Appendices
Een kreeu, baie ... (kreeus) [kri¥U] as in [spri¥U] ‘starling’

Een fasel, twee ... (fasels) [fAÉs«l] as in [tAÉf«l] ‘table’
Een kottel, baie ... (kottels) [kOt«l] as in [bOt«l] ‘bottle’
Een dissem, twee ... (dissems) [d«s«m] as in [bes«m] ‘broom’
Een toelem, twee .... (tulems) [tul«m] as in [bU¥«d«m] ‘bottom’
Een foten, baie .... (fotens) [fot«n] as in [lyÉj«n] ‘lie’
Een waken, baie ... (wakens) [vAk«n] as in [lAk«n] ‘sheet (of
bedding)’
Een fiender, baie ... (fienders) [find«r] as in [hUnd«r] ‘chicken’
Een pygter, twee ... (pygters) [p«¥ixt«r] as in [slaxt«r] ‘butcher’
Een pekerd, twee ... (pekerds) [pEk«rt] as in [l«¥ip«rt] ‘leopard’
Een lienkert, baie ... (lienkerds) [liNk«rt] as in [stand«rt] ‘standard’
Een slofie, twee ... (slofies) [slofi] as in [stori] ‘story’
Een banalie, baie ... (banalies) [banAÉli] as in [famili] ‘family’
Een latoewier, twee ... (latoewiers) [latuwiÉr] as in [juv«liÉr] ‘jeweller’
Een dimoenier, baie ... (dimouniers) [dEmuniÉr] as in [pas«siÉr] ‘passenger’
Een saan, baie ... (sane) [sAÉn] as in [mAÉn] ‘moon’
Een klaat, baie ... (klate) [klAÉt] as in [kl«¥It] ‘clod’
Een luur, baie ... (lure) [lyÉr] as in [fyÉr] ‘fire’
Een knuur, twee ... (knure) [knyÉr] as in [skyÉr] ‘barn’
Een beel, twee ... (bele) [bel] as in [kel] ‘throat’
Een keen, twee ... (kene) [ken] as in [sten] ‘stone’
Een goom, baie ... (gome) [xom] as in [bom] ‘tree’
Een sook, twee ... (soke) [sok] as in [spok] ‘ghost’
Een tef, twee ... (tewwe) [tEf] as in [hEf] ‘knife handle’
Een lif, baie ... (liwwe) [l«f] as in [r«f] ‘ridge’
Een sterg, baie ... (ster[g]e) [stÏrx] as in [bÏrx] ‘mountain’
Een lerg, baie ... (ler[g]e) [lÏrx] as in [bÏrx] ‘mountain’
Een til, twee ... (tille) [t«l] as in [br«l] ‘glasses’
Een wis, twee … (wisse) [v«s] as in [f«s] ‘fish’
Een sal, baie ... (salle) [sal] as in [bal] ‘ball’
Een fant, twee ... (fante) [fant] as in [kant] ‘side’
Een mek, baie ... (mekke) [mEk] as in [nEk] ‘neck’
Een ges, twee ... (gesse) [xEs] as in [nEs] ‘nest’
Een dies, twee ... (diese) [dis] as in [kis] ‘side of oral cavity’
Een skiel, twee ... (skiele) [skil] as in [vil] ‘wheel’
Een wor, baie ... (worre) [vOr] as in [tOr] ‘large beetle’
Een gol, twee ... (golle) [xOl] as in [bOl] ‘(plant) bulb’
Een loet, baie ... (loete) [lut] as in [fut] ‘foot’
Een woek, baie ... (woeke) [vuk] as in [buk] ‘book’
343
Appendices
2. PERSON AND CASE
2.1. Person and case comprehension – Picture selection

Practise items
Hy staan in die hoekie He is standing in the corner
Dit is haar skoen It is her shoe
Test items
Ek staan langs die tafel I am standing next to the table
Dit is my roomys It is my ice-cream
Die voël sit op my The bird is sitting on me
Die kat lek my The cat is licking me
Jy staan op die stoel You-SGL are standing on the chair
Dit is jou roomys It is your-SGL ice-cream
Die voël sit op jou The bird is sitting on you-SGL
Die kat krap jou The cat is scratching you-SGL
Hy sit by die tafel He is sitting at the table
Dit is sy roomys It is his ice-cream
Die voël sit op hom The bird is sitting on him
Die kat lek hom The cat is licking him
Sy sit by die tafel She is sitting at the table
Dit is haar roomys It is her ice-cream
Die voël sit op haar The bird is sitting on her
Die kat lek haar The cat is licking her
Dit lê op die grond It is lying on the ground (where it is a newspaper)
Sy oor is af It’s ear is off (where it is a cup)
Die koerant lê daarop The newspaper is lying on it
Ouma vryf dit Granma is stroking it
Ons wys ’n prentjie We are showing a picture
Dit is ons speelgoed These are our toys
Die reën val op ons The rain is falling on us
Die seuntjie sien ons The boy sees us
Julle wys ’n prentjie You-PL are showing a picture
Dit is julle speelgoed These are your-PL toys
Die reën val op julle The rain is falling on you-PL
Die seuntjie sien julle The boy sees you-PL
Hulle wys ’n prentjie They are showing a picture
Dit is hulle speelgoed These are their toys
Die reën val op hulle The rain is falling on them
Die seuntjie sien hulle The boy sees them
2.2. Person and case comprehension – Jugement

Ek is wakker I am awake
*Dit is sy hoed (my) It is my hat
Die roomys val op my The ice-cream is falling on me
*Die hond krap ek (my) The dog is scratching me
*Jou slaap (jy) You are sleeping
Dit is jou roomys It is your-SGL ice-cream
344
Appendices
*Die skoenlapper sit op haar (jou) The butterfly is sitting on you

Die hond krap jou The dog is scratching you
*Sy is wakker (hy) He is awake
*Dit is hom hoed (sy) It is his hat
Die eend swem langs hom The duck is swimming next to him
Die krap knyp hom The crab is biting him
Sy staan She is standing
Dit is haar roomys It is her ice-cream
*Die vark kyk na sy (haar) The pig is looking at her
*Die hond dra jou (haar) The dog is carrying her
*Sy staan in die hoek (dit) It is standing in the corner
Sy hare is af Its hair is gone (where it is a broom)
Die koppies staan bo dit The cups are above it (where it is a broom)
*Ouma hou haar vas (dit) Grandma is holding it (where it is her
knitting)
Ons lees boeke We are reading books
*Dit is hulle kar (ons) It is our car
*Die hond spring oor julle (ons) The dog is jumping over us
Die koei jaag ons The cow is chasing us
*Hy lees boeke (julle) You-PL are reading books
*Dit is haar roomyse (julle) It is your-PL ice-creams
Die hond spring oor julle The dog is jumping over you-PL
Die koei jaag julle The cow is chasing you-PL
Hulle lees boeke They are reading books
*Dit is hulle tuin (julle) It is their garden
Die hond spring oor hulle The dog is jumping over them
*Die skaap jaag ons (hulle) The sheep is chasing them
2.3. Person and case production – Sentence completion

Practise items
Sy eet ’n roomys, maar (hy eet ’n appel) She is eating an ice-cream, but (he is eating an apple)
Dit is sy koffie en dit is (haar melk) This is his coffee and this is (her milk)
Test items
Hy eet ’n roomys, maar (ek eet ’n appel) He is eating an ice-cream, but (I am eating an
apple)
Dit is my romp en dit is (my This is my skirt and this is (my trousers/skirt)
broek/romp)
Die eekhoring sit langs my en die hond The squirrel is sitting next to me and the dog is
sit (langs my) sitting (next to me)
Dié kat krap my en dié kat krap (my) This cat is scratching me and this cat is
scratching (me)
Hy staan, maar (jy sit) He is standing, but (you-SGL are sitting)
Dit is sy lepel en dit is (jou mes) This is his spoon and this is (your-SGL knife)
Dié hond sit langs hom, maar dié hond This dog is sitting next to him, but this dog is
sit (langs jou) sitting (next to you-SGL)
Dié hond krap hom, maar dié hond lek This dog is scratching him, but this dog (is
(jou) licking you-SGL)
345
Appendices
Hy eet pizza, maar (sy eet ’n He is eating pizza, but (she is eating a lollipop)
stokkielekker)
Dit is my hand en dit is (haar hand) This is my hand and this is (her hand)
Die hond sit langs my en die kat sit (langs The dog is sitting next to me and the cat is
haar) sitting (next to her)
Die hond lek my en die kat krap (haar) The dog is licking me and the cat is scratching
(her)
Sy hang in die lug, maar (hy sit op die She is hanging in the air, but (he is sitting on
grond) the ground)
Dit is haar swembroek en dit is (sy trui) This is her swimming costume and this is (his
jersey)
Die baba kyk vir my en die hond kyk (vir The baby is looking at me and the dog is
hom) looking (at him)
Die hond lek my en die baba lek (hom) The dog is licking me and the baby is licking
(him)
Ek is skoon, maar (dit is vuil) I am clean, but (it is dirty)
Dit is haar nek en dit is (sy nek) This is her neck and this is (its neck) (re a
bottle)
Die baba kyk vir my en die hond kyk The baby is looking at me and the dog is
(daarvoor /daarna /vir dit) looking (at it)
Die hond lek my en die seuntjie lek (dit) The dog is licking me and the boy is licking (it)
Sy drink water, maar (ons eet vrugte) She is drinking water, but (we are eating fruit)
Dit is hulle musse en dit is (ons skoene) These are their woolen hats and these are (our
shoes)
Die voël vlieg oor hulle en die vlieër vlieg The bird is flying over them and the kite is
(oor ons) flying (over us)
Die kat krap hom en die honde lek (ons) The cat is scratching him and the dogs are
licking (us)
Ons eet koek, maar (hulle eet roomys) We are eating cake, but (they are eating ice-
cream)
Dit is ons hare en dit is (hulle hare) This is our hair and this is (their hair)
Dié voël sit op my en dié voëls sit (op This bird is sitting on me and these birds are
hulle) sitting (on them)
Die hond lek haar en die katte krap The dog is licking her and the cats are
(hulle) scratching (them)
Sy eet appels, maar (julle eet piesangs) She is eating apples, but (you-PL are eating
bananas)
Dit is haar bene en dit is (julle bene) These are her legs and these are (your-PL legs)
Die skoenlapper sit op jou en die The butterfly is sitting on you and the
skoenlappers sit (op julle) butterflies are sitting (on you-PL)
Die vark lek hom en die honde lek (julle) The pig is licking him and the dogs are licking
(you-PL)
Hy eet waatlemoen, maar (ek eet koek) He is eating watermelon, but (I am eating cake)
Dit is my voet en dit is (my neus) This is my foot and this is (my nose)
Die voël sit op my en die haas sit (op my) The bird is sitting on me and the rabbit is
sitting (on me)
Die koei jaag my en die perd jaag (my) The cow is chasing me and the horse is chasing
(me)
346
Appendices
Hy lê, maar (jy staan) He is lying down, but (you-SGL are standing)
Dit is sy oë en dit is (jou hare) This is his nose and this is (your-SGL hair)
Dié perd spring oor hom, maar dié perd This horse is jumping over him, but this horse
spring (oor jou) is jumping (over you-SGL)
Dié hoender pik hom, maar dié hoender This chicken is pecking him, but this chicken is
pik (jou) pecking (you-SGL)
Hy voer die eekhoring, maar (sy voer die He is feeding the squirrel, but (she is feeding
voëls) the birds)
Dit is sy kar en dit is (haar bal) This is his car and this is (her ball)
Dié slang seil oor my en dié slang seil This snake is slithering over me and this snake
(oor haar) is slithering (over her)
Dié perd sien my en dié perd sien (haar) This horse sees me and this horse sees (her)
Jy staan op die tafel, maar (hy staan op You-SGL are standing on the table, but (he is
die stoel) standing on the chair)
Dit is my glas en dit is (sy glas) This is my glass and this is (his glass)
Die hond kyk vir jou en die perd kyk (vir The dog is looking at you-SGL and the horse is
hom) looking (at him)
Dié bul skop haar en dié bul skop (hom) This bull is kicking her and this bull is kicking
(him)
Jy is groot, maar (dit is klein) You are big, but (it is small)
Dit is haar tande maar dit is (sy tande) These are her teeth but these are (its teeth)
(where it is a comb)
Dié apie spring oor my en dié apie spring This monkey-DIM is jumping over me and this
(daarin /in dit) monkey-DIM is jumping into (it)
Dié kat krap my en dié kat krap (dit) This cat is scratching me and this cat is
scratching (it)
Sy spring tou, maar (ons praat op die She is skipping with a rope, but (we are talking
foon) on the phone)
Dit is haar tasse en dit is (ons tasse) These are her suitcases and these are (our
suitcases)
Die vliegtuig vlieg oor hom en die The aeroplane is flying over him and the
helikopter vlieg (oor ons) helicopter is flying (over us)
Dié vark sien hom en dié vark sien (ons) This pig sees her and this pig sees (us)
Ons staan, maar (hulle sit) We are standing, but (they are sitting)
Dit is sy boek en dit is (hulle boeke) This is his book and these are (their books)
Dié emmer val op my en dié emmer val This bucket falls on me and this bucket falls
(op hulle) (on them)
Dié seun stamp haar en dié seun stamp This boy is pushing her and this boy is pushing
(hulle) (them)
Hy sien ’n skaap, maar (julle sien ’n koei) He sees a sheep, but (you-PL see a cow)
Dit is sy hond en dit is (julle honde) This is his dog and these are (your-PL dogs)
Dié bal hop op jou en dié bal hop (op This ball will bounce on you and this ball will
julle) bounce (on you-PL)
Dié seun spuit hom nat en dié seun spuit This boy is squirting him and this boy is
(julle) nat squirting (you-PL)
347
Appendices
3. CASE (POSSESSIVE)
3.1. Case (possessive) comprehension – Picture selection

Practise items
Wys vir my die leeu se stert Show me the lion’s tail (vs. a lion – depicted in such a manner that its tail
is not visible – vs. a tail of another animal)
Die meisie se skoene The girl’s shoes (vs. a girl without shoes vs. boys’ shoes)
Test items
Die man se hand The man’s hand (vs. a man with his hands behind his back vs. a
woman’s hand)
Die hond se mandjie The dog’s basket (vs. a dog vs. a picnic basket)
Die motor se wiel The car’s wheel (vs. a car depicted in such a manner that the
wheels are not visible vs. a bicycle wheel)
Die baba se bottel The baby’s bottle (vs. a baby vs. a wine bottle)
Die hond se kos The dog’s food (vs. a dog vs. a plate of cooked food)
Die vrou se tande The woman’s teeth (vs. a woman vs. a dog’s teeth)
Die seuntjie se kar The boy’s car (vs. a boy vs. a real car)
Die man se koerant The man’s newspaper (vs. a newpaper vs. a man)
Die vrou se hoed The woman’s hat (vs. a woman vs. a man’s hat)
Die slang se tong The snake’s tongue (vs. a snake vs. a smiley face’s tongue)
Die hond se been The dog’s bone (vs. a dog vs. a leg)
3.2. Case (possessive) production – Sentence completion

Practise items
Hier is die seuntjie se reënjas en hier is (die Here is the boy’s raincoat and here is (the girl’s umbrella)
dogtertjie se sambreel)
Hier is die skoenlapper se vlerk en hier is (die Here is the butterfly’s wing and here are (the bug’s legs)
gogga se pote)
Test items
Hier is die blom se blare en hier is (die Here are the flower’s petals and here are (the
boom se blare) tree’s leaves)
Hier is die meisie se appels en hier is (die Here are the girl’s apples and here is (the boy’s
seun se piesang) banana)
Hier is die seuntjie se bed en hier is (die Here is the boy’s bed and here is (the girl’s bed)
meisie se bed)
Hier is die man se kar en hier is (die meisie Here is the man’s car and here is (the girl’s
se fiets) bicycle)
Hier is die hond se poot en hier is (die kat Here is the dog’s paw and here is (the cat’s tail)
se stert)
Hier is die man se trui en hier is (die vrou Here is the man’s jersey and here are the
se broek) (woman’s trousers)
Hier is die seun se tandepasta en hier is Here is the boy’s tooth paste and here is the
(die vrou se tandeborsel) (woman’s tooth brush)
Hier is die baba se hoed en hier is (die ma Here is the baby’s hat and here is (the mother’s
se handsak) hand bag)
Hier is die teddie se maag en hier is (die Here is the teddy’s tummy and here is (the doll’s
pop se hare) hair)
348
Appendices
Hier is die eendjie se dam en hier is (die Here is the duck’s pond and here is (the bird’s
voëltjie se hok) cage)
4. TENSE
4.1. Tense comprehension – Picture selection

Practise items
Die kos was baie There was a lot of food
Die meisie het oor die The girl jumped over the fence
heining gespring
Test items
Die vrou sny die gras The woman mows /is mowing the lawn (contrasted with pictures
of The woman mowed the lawn and The woman will mow the lawn)
Die ballon bars The balloon bursts /is bursting (contrasted with pictures of The
balloon burst and The balloon will burst)
Die seun verf die The boy paints /is painting the fence (contrasted with pictures of
heining The boy painted the fence and The boy will paint the fence)
Die appel val The apple falls /is falling (contrasted with pictures of The apple fell
and The apple will fall)
Die man moet fiks The man must be fit to be able to climb the mountain (i.e., he is
wees om die berg te still busy climbing the mountain; contrasted with pictures of The
kan uitklim man had to be fit to climb the mountain, i.e., he has completed the climb,
and The man will climb the mountain)
Die hond sal skoon The dog will be clean after its bath (i.e., it is still busy bathing;
wees na sy bad contrasted with pictures of The dog was clean after its bath (but now its
dirty again) and The dog is clean after its bath)
Die vrou moes lank The woman had to be tall to reach (i.e., she has already retrieved
wees om te kon the object from the very high shelf; contrasted with a picture of
bykom The woman must be tall to reach /The woman will have to be tall to reach,
i.e., she is still trying to retrieve the object) and The woman is fat)
Die koek sou The cake would have burnt (i.e., the cake was saved in time;
verbrand het contrasted with pictures of The cake is burning and The cake burnt)
Die seun het twee The boy has two balloons (contrasted with pictures of The boy had
ballonne two balloons, i.e., one burst/flew away, and The boy has no balloons)
Die pop het twee The doll has two arms (contrasted with pictures of The doll had two
arms arms, i.e., one arm broke off, and The doll has one arm and one leg)
Die meisie het ’n The girl had a lollipop (i.e., she ate it and now only the stick is left;
stokkielekker gehad contrasted with pictures of The girl has a lollipop and The girl will have
a lollipop, i.e., she is busy buying one)
Die teddie het twee The teddy had two eyes (i.e., one eye broke off; contrasted with
oë gehad pictures of The teddy has two eyes, and The teddy has two eyes but only one
arm)
Gister val die voëltjie Yesterday the bird fell out of its nest (contrasted with pictures of
uit sy nes The bird is falling out of its nest and The bird is still in its nest)
Gister klim ek op die Yesterday I climbed onto the roof (contrasted with pictures of I am
dak climbing onto the roof and I will climb onto the roof
Die potlood was lank The pencil was long (contrasted with The pencil is still long and The
349
Appendices
gewees pencil is being shortened)

Die posman was hier The postman was here (i.e., the letter has been delivered;
gewees contrasted with The postman is here and The postman will still come)
Die melk was op The milk was finished (but they went to buy some more;
contrasted with The milk is finished and The milk will be finished, i.e., at
present there is little milk left)
Die baba was vuil The baby was dirty (contrasted with The baby is still dirty and The
baby will be dirty, i.e., something is about to fall on it)
Die vrou het kos The woman bought food (contrasted with The woman buys food and
gekoop The woman will buy food)
Die vliegtuig het The aeroplane took off (contrasted with The aeroplane is taking off
opgestyg and The aeroplane is still on the ground)
Die koeldrank het The cooldrink fell over (contrasted with The cooldrink falls over and
omgeval The cooldrink will fall over)
Die boot het gesink The boat sank (contrasted with The boat is sinking and The boat will
sink)
Hy het die stukkende He changed the broken light bulb (contrasted with He is changing the
gloeilamp vervang broken light bulb and He still has to change the broken light bulb)
Die baba het dit The baby received it (contrasted with The baby is receiving it and The
ontvang baby will receive it)
4.2. Tense comprehension – Judgement

Die man het gister koerant gesit en lees Yesterday the man read the newspaper
Die honde het gister heeldag gelê en slaap Yesterday these dogs slept all day long
Hulle het heeldag gesit en werk They sat and worked all day
Hy het geloop en eet He ate while walking
Hulle het gestaan en praat They stood and talked
*Die baba het lê en gespeel The baby played while lying down
*Gister het die kat heeldag staan en gemiaau Yesterday the cat mewed all day long
*Die man het loop en gedink The man thought while walking
*Hulle het heeldag sit en gespeel They sat playing all day long
*Die seuntjie het staan en gehuil tot hy sy The little boy cried until he found his little
kombersie gekry het blanket
4.3. Tense production – Sentence completion

Practise items
Hierdie apie dra elke dag ’n pak. Gister, net soos Every day this monkey wears a suit. Yesterday, just
elke ander dag, (het hy ’n pak gedra) like every other day, (he wore a suit)
Hierdie kind borsel elke dag sy tande. Gister, net Every day this child brushes his teeth. Yesterday,
soos elke ander dag, (het hy sy tande geborsel) just like every other day, (he brushed his teeth)
Test items
Hierdie wurm eet elke dag ’n appel. Gister, Every day this worm eats an apple.
net soos elke ander dag, (het hy ’n appel Yesterday, just like every other day, (he ate
geëet) an apple)
Hierdie vrou pluk elke dag ’n blom. Gister, Every day this woman picks a flower.
net soos elke ander dag, (het sy ’n blom Yesterday, just like every other day, (she
350
Appendices
gepluk) picked a flower)

Hierdie venster breek elke dag. Gister, net This window breaks every day. Yesterday,
soos elke ander dag, (het dit gebreek) just like every other day, (it broke)
Hierdie muur kraak elke dag. Gister, net This wall cracks every day. Yesterday, just
soos elke ander dag, (het dit gekraak) like every other day, (it cracked)
Hierdie kind moet elke dag skool toe gaan. This child must go to school every day.
Gister, net soos elke ander dag, (moes hy skool Yesterday, just like every other day, (he had
toe gaan) /(moet/moes hy skool toe gegaan het) to go to school)
Hierdie kind moet elke dag sy skoene This child must clean his shoes every day.
skoonmaak. Gister, net soos elke ander dag, Yesterday, just like every other day, (he had
(moes hy sy skoene skoonmaak) /(moet/moes to clean his shoes)
hy sy skoene skoongemaak het)
Die bye maak altyd baie heuning, so hierdie The bees always make a lot of honey, so this
beer kan elke dag heuning eet. Gister, net bear can eat honey every day. Yesterday, just
soos elke ander dag, (kon hy heuning eet like every other day, (he could eat honey)
/kon hy heuning geëet het)
Dit reën nooit hier nie, so hierdie meisie kan It never rains here, so this girl can play
elke dag buite speel. Gister, net soos elke outside every day. Yesterday, just like every
ander dag, (kon sy buite speel /kon sy buite other day, (she could play outside)
gespeel het)
Hierdie eendjie wil elke dag swem. Gister, Every day this duckling wants to swim.
net soos elke ander dag, (wou hy swem Yesterday, just like every other day, (he
/wou hy geswem het) wanted to swim)
Hierdie baba wil elke dag bottel drink. This baby wants to drink bottle every day.
Gister, net soos elke ander dag, (wou hy Yesterday, just like every other day, (he
bottel drink /wou hy bottel gedrink het) wanted to drink bottle)
Hierdie seuntjie het elke dag ’n nuwe This boy has a new friend every day.
maatjie. Gister, net soos elke ander dag, (het Yesterday, just like every other day, (he had
hy ’n nuwe maatjie gehad) a new friend)
Hierdie man het elke dag ’n seer nek. Gister, This man has a sore neck every day.
net soos elke ander dag, (het hy ’n seer nek Yesterday, just like every other day, (he had
gehad) a sore neck)
Hierdie kat is elke dag hier. Gister, net soos This cat is here every day. Yesterday, just
elke ander dag, (was hy hier) like every other day, (it was here)
Hierdie man is elke dag laat. Gister, net soos This man is late every day. Yesterday, just
elke ander dag, (was hy laat) like every other day, (he was late)
Hierdie vrou betaal elke dag die verwer. This woman pays the painter every day.
Gister, net soos elke ander dag, (het sy die Yesterday, just like every other day, (she paid
verwer betaal) the painter)
Hierdie kind verstaan elke dag alles. Gister, Every day, this child understands everything.
net soos elke ander dag, (het sy alles Yesterday, just like every other day, (she
verstaan) understood everything)
Hierdie oupa sit elke dag en slaap. Gister, Every day, this grandpa sits and sleeps.
net soos elke ander dag, (het hy gesit en Yesterday, just like every other day, (he sat
slaap) and slept)
Hierdie man staan elke dag en wag vir die Every day, this man stands and waits for the
bus. Gister, net soos elke ander dag, (het hy bus. Yesterday, just like every other day, (he
vir die bus gestaan en wag) stood and waited for the bus)
351
Samenvatting
Samenvatting
‘Specific language impairment’ in het Afrikaans. Een

minimalistische verklaring voor problemen met grammaticale
kenmerken en woordvolgorde.
Specific language impairment (SLI) is een ernstige stoornis bij kinderen

in het verwerven van gesproken taal, zonder dat er aanwijsbare
oorzaken zijn zoals neurologische gebreken, mentale afwijkingen,
gehoorproblemen, emotionele problemen of gedragsproblemen
(Leonard 1998:vi; Stark en Tallal 1981). De stoornis heet in het
Nederlands ESM: ernstige spraak- en taalmoeilijkheden. De kenmerken
van SLI in het Engels en in verschillende andere talen zijn bekend. Het
gaat om problemen met (i) de grammaticale morfologie, (ii)
vraagconstructies en passieve constructies en (iii) constructies met co-
referentiële relaties. In dit onderzoek werden de kenmerken van SLI
zoals die voorkomen in het Afrikaans – dat afstamt van het Nederlands
en dat hoofdzakelijk in Zuid-Afrika wordt gesproken – vastgesteld.
Afrikaans is een morfologisch arme taal (slechts weinig grammaticale
kenmerken worden fonetisch gerealiseerd), met variatie in de
woordvolgorde die onder meer bepaald wordt door scrambling en
linksdislokatie.
De algemene vraag die het onderhavige onderzoek moet beantwoorden

luidt: hoe komt SLI, een taalstoornis die gekenmerkt wordt door
problemen met grammaticale morfemen, tot uiting in een morfologisch
arme taal als het Afrikaans? Om deze algemene vraag afdoende te
beantwoorden en om na te gaan of recente verklaringsmodellen voor
SLI tot de juiste voorspellingen leiden voor de verzamelde Afrikaanse
gegevens, zijn er zes meer specifieke vragen geformuleerd:
1. Hoe worden grammaticale morfemen – in het bijzonder de
morfemen die betrekking hebben op de kenmerken getal, persoon,
naamval en tijd – begrepen door Afrikaans sprekende kinderen
met SLI?
353
Samenvatting
2. Hebben Afrikaans sprekende kinderen met SLI een achterstand

en/of een afwijking in hun begrip van bovengenoemde
morfemen?
3. Hoe worden grammaticale morfemen – in het bijzonder de
morfemen met betrekking tot de kenmerken getal, persoon,
naamval en tijd – morfologisch gerealiseerd (dat wil zeggen
daadwerkelijk geproduceerd) door Afrikaans sprekende kinderen
met SLI?
4. Hebben Afrikaans sprekende kinderen met SLI een achterstand
en/of afwijking in de daadwerkelijke productie van deze
morfemen?
5. Recente linguïstische verklaringsmodellen voor SLI zijn het
Agreement/Tense Omission Model (ATOM), de Representational
Deficit for Dependent Relations (RDDR) en de Feature Deficit
Hypothesis. Zij doen specifieke voorspellingen voor Afrikaans
sprekende SLI kinderen over zowel het begrip en/of de productie
van grammaticale morfemen als over de woordvolgorde. Worden
deze voorspellingen bevestigd door de data uit het Afrikaans die in
dit onderzoek werden verkregen of juist niet?
6. Als het antwoord op vraag 5 slechts “gedeeltelijk” of zelfs
“helemaal niet” luidt, is er dan een alternatieve verklaring te
formuleren voor de manier waarop SLI in het Afrikaans tot uiting
komt?
Om de eerste vier vragen te beantwoorden werden het begrip en de

productie van de grammaticale morfemen voor getal, persoon, naamval
en tijd van de volgende drie groepen Afrikaans sprekende kinderen
beoordeeld: 15 zesjarige kinderen met SLI, 15 zich normaal
ontwikkelende zesjarige kinderen en 15 zich normaal ontwikkelende
vierjarige kinderen. De te analyseren data waren afkomstig van (i) een
reeks van experimentele taken, met behulp van ‘picture selection’ en
‘grammaticality judgement’, om het taalbegrip te beoordelen, en een
zinnen-aanvultaak, om de productie te beoordelen, en (ii) spontane
taalproductiegegevens die voor ieder kind verzameld zijn.
Over het algemeen scoorden de zesjarige kinderen met SLI duidelijk

slechter op de experimentele taken dan de zich normaal ontwikkelende
leeftijdsgenoten; de kinderen met SLI presteerden op het niveau van de
zich normaal ontwikkelende vierjarige kinderen. Er kon geen algeheel
354
Samenvatting
verschil gevonden worden tussen de zesjarige groep met SLI en de zich

normaal ontwikkelende vierjarige groep. De linguïstische kenmerken van
SLI in het Afrikaans van zesjarige kinderen die in de experimentele taken
aan het licht kwamen zijn de volgende:
(i) een frequenter gebruik van het incorrecte meervoudsmorfeem dan
de zich normaal ontwikkelende jongere kinderen en de zich
normaal ontwikkelende leeftijdsgenoten;
(ii) meer fouten met betrekking tot naamval en persoon van het
voornaamwoord dan de zich normaal ontwikkelende jongere
kinderen en de zich normaal ontwikkelende leeftijdsgenoten;
(iii) meer idiosyncratische fouten en meer fouten bij het ge-vormen in
het gebruik van de verleden tijd dan de zich normaal
ontwikkelende jongere kinderen en de zich normaal
ontwikkelende leeftijdsgenoten;
(iv) het soms weglaten van modale hulpwerkwoorden, waar kinderen
met een normale ontwikkeling (zowel de jongere kinderen als de
kinderen van dezelfde leeftijd) deze hulpwerkwoorden niet
weglaten;
(v) weglating van de tegenwoordige tijd bij de be-vormen, meer dan
drie keer zo vaak als door de vierjarige kinderen.
In het spontane gebruik van de grammaticale morfemen voor getal,

persoon, naamval en tijd presteerden de kinderen met SLI slechter dan
de beide andere groepen kinderen. De kinderen met SLI maakten
voornamelijk dezelfde fouten als de zich normaal ontwikkelende
vierjarige kinderen, hoewel sommige fouten uniek waren voor de
kinderen met SLI. Voor wat betreft de fouten in de woordvolgorde
maakten de kinderen met SLI dezelfde soort fouten als de vier- en
zesjarige kinderen van de vergelijkingsgroepen. Twee soorten fouten
bleken echter uniek voor kinderen met SLI: zij maakten zowel
hoofdzinnen met een SOV woordvolgorde als hoofdzinnen met een
VSO volgorde.
Drie recente linguïstische verklaringsmodellen voor SLI (ATOM,

RDDR, en Feature Deficit Hypothesis) werden beoordeeld op de
geldigheid van de voorspellingen die ze doen voor SLI in het Afrikaans.
Van de in totaal twaalf voorspellingen die vanuit deze drie invalshoeken
werden geformuleerd, werd er slechts één gedeeltelijk bevestigd. Het
antwoord op onderzoeksvraag 5 was dan ook een pertinent “nee”. Nog
355
Samenvatting
afgezien van het feit dat op basis van deze modellen geen bruikbare
voorspellingen konden worden opgesteld over het taalgebruik van
Afrikaans sprekende kinderen met SLI, bleek uit het huidige onderzoek
bovendien dat kinderen met SLI fouten maakten in hun spontane
taaluitingen die buiten het bereik vielen van bovengenoemde modellen.
Er dient een alternatieve verklaring gevonden te worden voor SLI, zeker
voor de wijze waarop deze stoornis zich in het Afrikaans manifesteert.
Op grond van de fouten die Afrikaans sprekende kinderen met SLI

maken in de grammaticale kenmerken in zowel de experimentele taken
als in de geanalyseerde spontane taal en op grond van de fouten in de
woordvolgorde in de spontane taal, kan worden gesteld dat (i) in de
grammatica van de Afrikaans sprekende kinderen met SLI de abstracte
representaties van grammaticale kenmerken en de
verplaatsingsbewerkingen in tact zijn, en dat (ii) het probleem schuilt in
de Spell-Out, d.w.z. het moment dat de grammaticale informatie moet
worden verklankt of, anders geformuleerd, moet worden gekoppeld aan
de fonetische vorm (=PF). Voor de fouten die betrekking hebben op de
grammaticale vormen lijkt het dat ‘niet volwassen’ fonologische vormen
worden gekoppeld aan een verzameling van kenmerken (‘features’), of
dat de kenmerken in het geheel niet fonetisch worden gerealiseerd.
Dergelijke Spell-Out fouten blijken met name voor te komen in een
context waar meer dan één mogelijke klank-vormkoppeling aanwezig is
(bijvoorbeeld in het geval van de morfologie voor verleden tijd, waarbij
er meer dan één manier is om de verleden tijd in het Afrikaans uit te
drukken.) Voor wat betreft de fouten in de woordvolgorde bleken
Afrikaans sprekende kinderen met SLI een probleem te hebben om te
bepalen welke vormen moeten worden gerealiseerd en welke vormen
moeten worden weggelaten in de syntactische berekeningen. Soms
werden alle vormen weggelaten (geen enkele vorm wordt uitgedrukt in
een klankpatroon), soms werd een vorm meer dan één keer verklankt en
soms werd een incorrecte vorm verklankt (een correcte vorm volgt
overigens nooit een incorrecte bij meerdere verklankingen).
In deze dissertatie wordt beargumenteerd dat twee bekende kenmerken

van SLI, namelijk problemen met de grammaticale morfologie en
problemen met constituentverplaatsing, teruggevoerd kunnen worden op
problemen met grammaticale kenmerken (‘features’). Dat geldt dus ook
voor het Afrikaans: de fouten die gemaakt werden door Afrikaans
356
Samenvatting
sprekende kinderen met SLI kunnen inderdaad worden herleid tot de

Spell-Out van grammaticale kenmerken in de zogeheten PF. Daarmee
levert dit onderzoek een bijdrage aan de lokalisatie van de problemen die
kinderen met SLI ervaren. Het probleem zit niet in het computationele
systeem (alwaar de zinnen worden berekend op grond van de
bewerkingen ‘Merge’ or ‘Move’), maar daar waar syntaxis en morfologie
gekoppeld worden aan de verklanking. Kinderen met SLI slagen er
onvoldoende in de juiste vormen te selecteren ofwel te kiezen uit het
lexicon voor de Spell-Out. Dit zou verband kunnen houden met het feit
dat de Spell-Out vormen (nog) niet zo gestructureerd en toegankelijk zijn
opgeslagen als in het normale volwassen lexicon.
357
Opsomming
Opsomming
Spesifieke taalgestremdheid in Afrikaans. ’n Minimalistiese

verklaring vir probleme met grammatikale kenmerke en
woordvolgorde.
Spesifieke taalgestremdheid (STG) dui op ’n ernstige stoornis in kinders

se verwerwing van gesproke taal in die afwesigheid van identifiseerbare
oorsaaklike faktore of voor die hand liggende meegaande faktore soos
neurologiese afwykings, verstandelike gestremdheid en gehoor-,
emosionele en gedragsprobleme (Leonard 1998:vi; Stark en Tallal 1981).
Die eienskappe van STG in Engels en in verskeie ander tale is
welbekend, en sluit probleme met die volgende in: (i) grammatikale
morfologie, (ii) vraag- en passiefkonstruksies en (iii) konstruksies waarin
ko-referensiële relasies voorkom. In hierdie studie is die eienskappe van
STG soos dit voorkom in Afrikaans – ’n taal wat histories en
grammatikaal verwant is aan Nederlands en wat oorwegend in Suid-
Afrika gepraat word – bepaal. Afrikaans is ’n morfologies verarmde taal
waarin min grammatikale kenmerke fonologies gerealiseer word; verder
toon Afrikaans opvallende woordvolgorde-variasie, onder andere weens
scrambling en links-dislokasie.
Die algemene vraag wat hierdie studie wou beantwoord, was: Hoe
presenteer STG – wat gekenmerk word deur ’n probleem met
grammatikale morfeme – in Afrikaans, ’n morfologies verarmde taal?
Om hierdie algemene vraag afdoende te kan beantwoord – en om vas te
stel of resente verklarings vir STG akkurate voorspellings maak oor die
ingesamelde Afrikaanse data – is ses spesifieke vrae gestel:
1. Hoe word grammatikale morfeme – spesifiek dié wat verband hou
met die kenmerke getal, persoon, kasus en tempus – begryp deur
Afrikaanssprekende kinders met STG?
2. Presenteer Afrikaanssprekende kinders met STG in terme van hul
begrip van hierdie morfeme met ’n agterstand en/of ’n afwyking?
3. Hoe word grammatikale morfeme – spesifiek dié wat verband hou
met die kenmerke getal, persoon, kasus en tempus – morfologies
359
Opsomming
gerealiseer (d.i., geproduseer) deur Afrikaanssprekende kinders met

STG?
4. Presenteer Afrikaanssprekende kinders met STG in terme van hul
produksie van hierdie morfeme met ’n agterstand en/of ’n
afwyking, of nie een van die twee nie?
5. Die Kongruensie/Tempus-weglatingsmodel (Agreement/Tense
Omission Model; ATOM), die Selektiewe Gestremdheid in
Konstituentverplasing (Representational Deficit for Dependent
Relations; RDDR), en die Kenmerkagterstand-hipotese (Feature
Deficit Hypothesis; FDH) maak spesifieke voorspellings oor die
begrip en/of produksie van grammatikale morfeme deur
Afrikaanssprekende kinders met STG. Word hierdie voorspellings
ondersteun deur die Afrikaanse data wat in hierdie studie
ingesamel is?
6. Indien die antwoord op vraag 5 “gedeeltelik” or “glad nie” is, kan
’n toereikende alternatiewe verklaring vir STG soos dit in
Afrikaans presenteer, gegee word?
Om vrae 1 tot 4 te beantwoord is drie groepe Afrikaanssprekende

kinders se begrip en produksie van grammatikale morfeme wat verband
hou met die kenmerke getal, persoon, kasus en tempus getoets: 15 6-
jariges met STG, 15 tipies-ontwikkelende 6-jariges, en 15 tipies-
ontwikkelende 4-jariges. Die toetsing het die vorm aangeneem van (i) ’n
reeks eksperimentele take (naamlik prentseleksie- en
grammatikaliteitsoordeeltake om begrip te toets, en sinsvoltooiingstake
om produksie te toets) en (ii) die versamel van ’n spontane taalmonster
van elke deelnemer.
Oor die algemeen het die kinders met STG beduidend swakker gevaar as
die tipies-ontwikkelende 6-jariges in die eksperimentele take; die kinders
met STG het soos jonger tipies-ontwikkelendes gepresenteer. Geen
algemene verskil kon gevind word tussen die kinders met STG en die 4-
jariges se verskillende response op items nie. Die volgende linguistiese
eienskappe van STG in 6-jarige Afrikaanssprekendes is deur die
eksperimentele take aan die lig gebring:
(i) meer frekwente gebruik van ’n verkeerde meervoudsmorfeem as
jonger en ouderdomsgepaarde tipies-ontwikkelende kinders;
360
Opsomming
(ii) meer foute rakende die begrip en produksie van kasus en persoon
op voornaamwoorde as jonger en ouderdomsgepaarde tipies-
ontwikkelende kinders;
(iii) meer idiosinkratiese foute en meer foute met het ge-vorme waar
verledetydskonstruksies geproduseer is as jonger en
ouderdomsgepaarde tipies-ontwikkelende kinders;
(iv) sporadiese weglating van modale hulpwerkwoorde in teenstelling
met tipies-ontwikkelende kinders (jonger asook
ouderdomsgepaard) waar sulke weglatings nie voorkom nie;
(v) weglating van teenwoordigetydsvome van wees, meer as drie maal
soveel as soortgelyke weglatings deur tipies-ontwikkelende 4-
jariges.
In teenstelling met die bogenoemde bevindinge het die kinders met STG
swakker gevaar as beide groepe tipies-ontwikkelende kinders in terme
van hul spontane gebruik van die grammatikale morfeme wat verband
hou met die kenmerke getal, persoon, kasus en tempus. Die kinders met
STG het meestal dieselfde tipe foute gemaak as die tipies-ontwikkelende
4-jariges; sommige foute was egter uniek aan die kinders met STG. Wat
woordvolgorde betref, het die kinders met STG meestal dieselfde tipe
foute as óf die 4-jariges óf die 6-jariges óf beide groepe gemaak. Twee
tipe foute was egter uniek aan die kinders met STG: die produksie van
hoofsinne met ’n SOV-woordvolgorde asook hoofsinne met ’n VSO-
woordvolgorde.
Drie resente linguistiese verklarings vir STG (die ATOM, RDDR, en

FDH) is krities ondersoek aan die hand van die voorspellings wat uit
elkeen volg in verband met STG in Afrikaans. Uit die 12 voorspellings
wat in totaal deur hierdie drie verklarings gemaak word, is slegs een deur
die data ondersteun, en ook net gedeeltelik. Die kort antwoord op
navorsingsvraag 5 was dus “nee”. Afgesien daarvan dat hierdie
verklarings nie bruikbare voorspellings vir die taalgebruik van
Afrikaanssprekende kinders met STG maak nie, het hierdie kinders ook
foute in hul spontane taalproduksie gemaak wat buite die bestek van
hierdie verklarings val. Daar is dus duidelik ’n behoefte aan ’n
alternatiewe verklaring vir STG soos dit in Afrikaans presenteer.
Op grond van die foute met die realisering van grammatikale kenmerke
wat in sowel die eksperimentele take as die spontane taalmonsters
361
Opsomming
voorgekom het, asook die woordvolgordefoute in die spontane

taalmonsters, is daar twee voorstelle gemaak: (i) die abstrakte
kenmerkvoorstelling en skuifbewerkings in die grammatika van
Afrikaanssprekende kinders met STG is intakt, maar (ii) hierdie kinders
se probleem lê by Uitspel, dit wil sê, daar waar die grammatikale inligting
verklank moet word of, anders gestel, gekoppel moet word aan die
fonetiese vorm (=PF). Aangaande die foute wat verband hou met die
betrokke grammatikale vorme blyk dit dat óf ’n “nie volwasse”
klankvorm aan die kenmerkbundel gegee word óf die kenmerke geensins
fonologies uitgespel word nie. Hierdie Uitspelfoute blyk veral voor te
kom in kontekste waar meer as een kompeterende klankvorm
teenwoordig is (byvoorbeeld in die geval van verledetydsmorfologie,
waar daar meer as een manier is om verledetyd in Afrikaans uit te druk).
Wat betref die woordvolgordefoute blyk dit dat Afrikaanssprekende
kinders met STG probleme ervaar om te bepaal watter vorme om uit te
spel en watter om weg te laat: soms word alle vorme weggelaat (met
geeneen wat ’n klankvorm ontvang nie), soms word meer as een kopie
verklank, en soms word ’n verkeerde kopie (nie die mees linkse een nie)
uitgespel.
In hierdie proefskrif is daar geargumenteer dat twee goed

gedokumenteerde eienskappe van STG, naamlik probleme met
grammatikale morfeme en probleme met konstituentverplasing, albei
teruggevoer kan word na probleme met grammatikale kenmerke, ook in
Afrikaans. Daar is aangetoon dat dit wel die geval is: die foute wat die
Afrikaanssprekende kinders met STG gemaak het, kan inderdaad
verklaar word in terme van die uitspel van grammatikale kenmerke op
die vlak van PF. Die studie lewer ’n bydrae tot die lokalisering van die
probleme wat kinders met STG ervaar: die algemene probleem is
klaarblyklik nie in die komputasie-sisteem gesetel nie (met ander woorde,
betref nie die sintaktiese bewerkings Saamvoeg (“Merge”) of Skuif nie);
die probleem lê eerder by die sintaksis-fonologie-koppelvlak. Kinders
met STG ervaar probleme om die korrekte klankvorm vir Uitspel te kies
uit die leksikon, wat daarmee mag saamhang dat hierdie vorme (nog) nie
so gestruktureerd en toeganklik gestoor is soos in die normale volwasse
leksikon nie.
362
Curriculum vitae
Curriculum Vitae
Frenette Southwood obtained an undergraduate degree in speech therapy

and audiology in 1993 and a Masters degree in Linguistics (specialising in
Clinical Linguistics) in 1995, both from Stellenbosch University (South
Africa). She practised as a speech-language therapist and audiologist in
South Africa and England before becoming a lecturer in the Department
of General Linguistics at Stellenbosch University in 2000, where she
teaches undergraduate and postgraduate courses in, amongst others,
Clinical Linguistics, Second Language Acquisition, Phonology, and
Intercultural Communication.
363

Problems With Afrikaans Word Order and Grammar

Uploaded by

Copyright:

Available Formats

Problems With Afrikaans Word Order and Grammar

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Problems With Afrikaans Word Order and Grammar

Uploaded by

Copyright:

Available Formats

Specific language impairment in Afrikaans

Providing a Minimalist account

Cover illustration: Sossusvlei, by Frenette Southwood

Copyright © 2007: Frenette Southwood. All rights reserved.

een wetenschappelijke proeve op het gebied van de Letteren

ter verkrijging van de graad van doctor

Manuscriptcommissie: Prof. dr. A.C.M. Rietveld

Chapter 2: Characteristics and theoretical accounts of SLI 7

Chapter 5: Number comprehension and production 111

5.1. Introduction 111

Chapter 6: Comprehension and production of person

6.1. Introduction 157

Chapter 7: Tense comprehension and production 189

7.1. Introduction 189

8.1. Introduction 217

Chapter 9: Accounts of SLI as it presents itself in Afrikaans 245

9.1. Introduction 245

Chapter 10: Conclusion 299

10.1. Summary of the present study 299

Appendix A: Information on SLI participants’ language test

Samenvatting [Dutch summary] 351

I am indebted to colleagues at my home department, the Department of

I am grateful for the support of various departmental and research

I am most grateful towards the South African National Research

At Radboud University Nijmegen, I am grateful to Frans van der Slik for

Table 4.1. Information on participants

ACC accusative case

Specific language impairment (SLI) is a significant impairment in the

Only a limited number of researchers has investigated the

Grammatical Morphemes subtest of the Test for Auditory

The question arises as to whether comprehension and production of

Another question is whether predictions regarding comprehension and

(CCH). See, for example, Van der Lely (2005).

Feature Deficit Hypothesis (Gopnik 1994a) are borne out by the

1. How are grammatical morphemes – specifically those related to the

The dissertation is structured in the following way. In chapter 2, the

In chapter 3, the discussion turns to Afrikaans, explicating why the way

Several subgroups of children have been identified (see, for example,

In chapters 5 to 7, the results of the language sample analysis and those

(relevant to chapters 5 to 7) of the language samples are considered. Also

Chapter 9 contains a discussion on the extent to which the three

Chapter 10 concludes the dissertation, by providing a summary of its

Characteristics and theoretical accounts of SLI

The aim of this chapter is to describe some of the characteristics of SLI

2.2. WHY STUDY SLI?

represent the demographics of the larger American population.

2.3. SOME CHARACTERISTICS OF SLI

Although some crosslinguistic data are available,6 most studies of SLI

2.3.1. Some characteristics of SLI as it presents itself in English

2.3.1.1. Grammatical morphemes

2.3.1.2. Word order related phenomena

2.3.1.3. (Non-)co-referential relationships

Villiers, Cahillane, and Altreuer 2006).8 The comprehension and

2.3.2. How SLI presents itself in other languages, in terms of

Presented here is a selection of languages other than English in which

2.3.2.1. Germanic languages

1993 for an analysis of Dutch as an underlyingly SVO language). In embedded clauses,

Also, the number feature may be represented inappropriately, usually by

German-speaking11 children with SLI often omit the article in

constituent movement. Amongst other movement operations, German demonstrates