Cognitive and Language Development in Children Chi PDF
Cognitive and Language Development in Children Chi PDF
Cognitive and Language Development in Children Chi PDF
Development in Children
This publication forms part of an Open University course ED209 Child Development.
Details of this and other Open University courses can be obtained from the Student
Registration and Enquiry Service, The Open University, PO Box 197, Milton Keynes,
MK7 6BJ, United Kingdom: tel. +44 (0)870 333 4340, e-mail [email protected]
Alternatively, you may visit the Open University website at http://www.open.ac.uk where
you can learn more about the wide range of courses and packs offered at all levels by The
Open University.
To purchase a selection of Open University course materials visit www.ouw.co.uk, or
contact Open University Worldwide, Michael Young Building, Walton Hall, Milton Keynes
MK7 6AA, United Kingdom for a brochure. tel. +44 (0)1908 858785;
fax +44 (0)1908 858787; e-mail [email protected]
Cognitive and Language
Development in Children
2 First words 61
MARGARET HARRIS
Acknowledgements 333
John Oates
Andrew Grayson
Introduction: perspectives on
cognitive and language
development
John Oates and Andrew Grayson
Contents
References 19
8
. At what stage did the two children learn the names of the objects that they
use? Did they have an understanding of the characteristics of these objects
before the names emerged, or were the names acquired before any real
understanding?
. Are Amy and Tom’s brains pre-programmed to acquire the language that
they now use so easily and flexibly, or have they had to learn it all from
others?
. How have these children developed the ability to construct sentences out of
the words that they have acquired that enable them to communicate
apparently so effortlessly with each other? What about the sentence ‘They’ve
lefted the children behind’? Has Tom copied that from someone else, or did
he make it up himself?
. What processes lie behind Amy and Tom being able to plan a trip to market,
and to organize their behaviour and language into sequences that make
their plans happen?
. At what age did Amy and Tom become capable of understanding the
thoughts and feelings of other people? How significant for their cognitive
development are their interactions with each other and with other people?
This first chapter introduces a theme that runs through much of the book; that
as development proceeds, mental representations become progressively more
abstract and freed from the specific contexts and concrete experiences from
which they spring. In this chapter the mental representations in question are
infants’ representations of categories. Also, the chapter shows that even the very
early categories formed by infants can be flexible and adaptable. The picture that
begins to emerge is that of development as an active, constructive process, and
that this can be seen from one of the earliest cognitive, meaning-making activities
of young infants – namely, category formation.
3.2 Constructivism
The work of Jean Piaget (1896–1980) made a key contribution to the field of
developmental psychology, especially in the area of cognitive development. It set
the scene for a much more scientific approach to understanding this area than
had ever been attempted before. Piaget’s theory was based firmly on the
observation of children’s behaviour. He found that there are many problem-
solving tasks in which children behave quite differently from adults. Furthermore,
he believed that children’s mental abilities pass through a regular series of stages
as development progresses.
Piaget’s model of development (Piaget, 1955, 1959 and 1973) involves the child
becoming increasingly freed up from the constraints of their own perspective and
INTRODUCTION: PERSPECTIVES ON COGNITIVE AND LANGUAGE DEVELOPMENT 15
the concrete objects around them, as their mental operations become more
abstract. In Piaget’s theory, children are at first egocentric, dominated by, or
‘centred’ on, their own perceptions, because they are still very much tied to the
concrete world and their actions within it. Similarly, Piaget held that the young
child is unable to comprehend points of view different from his or her own. A
classic example of this particular difficulty is provided by children’s responses to
the ‘three mountains’ task.
As used by Piaget, this task involves sitting a child beside a three-dimensional
model of three mountains (see Figure 1). A doll is placed at the edge of the
model, with a different view to that of the child, and the child’s task is to show the
experimenter what the doll’s view is. Piaget used various ways of doing this so
that the child would not have to describe it verbally. For example, the child has to
arrange three cardboard mountains (like the model ones), or choose one of a set
of drawings of the mountains model, drawn from different viewpoints, to indicate
what view of the mountains the doll has.
BOX 1
D B
(a) (b)
Figure 1 (a) Plan of the three mountains task; (b) view of the three mountains from
position A (Source: Cox, 1980).
The child was asked to represent the view of the mountains from his own position at A.
Then a wooden doll (about two to three cm high) was put at position C and the child
16 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
was asked to represent the view that could be seen from there. This procedure was
repeated for positions B and D. The child was then asked to move to position B (or C
or D) and to represent the view from there; in addition he was asked to represent the
view from A or other positions he had already occupied.
Piaget and Inhelder elicited three different modes of response. First, the child was asked
to reproduce the doll’s view by arranging three pieces of cardboard shaped like the
mountains. Second, the child was asked to select the doll’s view from a set of ten
pictures (each measuring 20 by 28 cm). In the third task, the child was asked to choose
one picture and then decide which position the doll must occupy to have that particular
view of the mountains.
Source: adapted from Roth, 1990, p. 99.
n
Presented in this way, the problem is usually difficult for a child of up to about 7
years old to solve correctly. A common response is for children to indicate that
the view that they themselves have is also the one that the doll has. Piaget saw
this as reflecting the inability of younger children to decentre from their own
position. His explanation was that children in this ‘pre-operational’ stage have not
developed the ability to recognize that there can be many different viewpoints on
a scene like the three mountains, their own being just one of these. Nor are they
able to hold in mind the idea that the viewpoint changes as the viewing position
changes. This process reaches its end-point in the final stage of Piaget’s
developmental model, when mental operations become ‘formal’ – when wholly
abstract thinking can be achieved and the child becomes able to reason
hypothetically and systematically. Although this book does not cover the details
of Piaget’s work, the developmental journey that he describes, from concrete to
progressively more abstract ways of thinking, is an important theme in the
account of cognitive and language development that we give.
A classic Piagetian task, the pendulum test, which is widely used in physics
teaching in many schools, is a good illustration of these points. The child’s task is
to find out what factor(s) affect the rate at which a pendulum swings to and fro.
There are several factors that could affect the rate, for example the weight of the
pendulum, the distance through which it swings, the length of the string, the force
of the initial push, the thickness of the string and so on. A child whose reasoning
capacities are still developing may vary these factors without a systematic
approach, perhaps trying various ideas more than once as a result, whereas a
child who can reason more formally can work out that varying each factor
independently to isolate the important one(s) is the most effective approach.
Piaget’s theory describes a child who is progressively elaborating a more
abstract and general capacity to tackle problems in the world, but in an essentially
solitary, independent way. This view of development was used in the second half
of the twentieth century to support the pedagogic principles of discovery
learning, which hold that it is a rich learning environment, rather than direct
teaching, that is essential for cognitive development.
INTRODUCTION: PERSPECTIVES ON COGNITIVE AND LANGUAGE DEVELOPMENT 17
framework. So, for example, when Piaget’s three mountains task was changed so
that the question asked was whether a policeman could see a robber who was
trying to hide from them (see Figure 2), children aged around 4 years who failed
the classic Piagetian task showed that, in this new situation, they could decenter
and see things from the perspective of another person. A policeman looking for a
robber made greater sense to children of this age than the more abstract notion of
another character’s view of a set of mountains. Several other experiments in the
same vein carried out around this time confirmed that children’s cognition shows
itself to be more advanced when the tasks that they tackle are situated in a
meaningful context.
BOX 2
A D
Figure 2 Plan of the experimental layout in the hiding game (Source: adapted from
Donaldson, 1978).
Source: adapted from Roth, 1990, p. 115.
n
INTRODUCTION: PERSPECTIVES ON COGNITIVE AND LANGUAGE DEVELOPMENT 19
References
Chomsky, N. (1965) Aspects of the Theory of Syntax, Cambridge, MIT Press.
Doise, W. and Mugny, G. (1981) Le Développement Social de l’Intelligence, Paris,
Interedition.
Donaldson, M. (1978) Children’s Minds, London, Fontana.
Piaget, J. (1955) The Child’s Construction of Reality, trans. by M. Cook, London,
Routledge and Kegan Paul (first published 1936).
Piaget, J. (1959) The Language and Thought of the Child, translated by M. Gabain
and R. Gabain, London, Routledge and Kegan Paul (first published 1923).
20 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Piaget, J. (1973) The Child’s Conception of the World, trans. by J. Tomlinson and A.
Tomlinson, London, Paladin.
Roth, I. (1990) Introduction to Psychology, London, Erlbaum Associates/The
Open University.
Vygotsky, L.S. (1962, 2nd edn) Thought and Language, Cambridge, MA, MIT
Press (first published 1934).
Wood, D. (1988) How Children Think and Learn, Oxford, Blackwell.
Chapter 1
Early category representations and
concepts
Paul C. Quinn and John Oates
Contents
Learning outcomes 23
1 Introduction 23
1.1 Cognitive structure 24
1.2 Classes, categories and concepts 27
8 Conclusion 57
References 57
1 EARLY CATEGORY REPRESENTATIONS AND CONCEPTS 23
Learning o u t c o m e s
After you have studied this chapter you should be able to:
1 define ‘categories’ and ‘concepts’ and discuss the differences between
them;
2 explain how category formation in infants can be studied experimentally;
3 describe and discuss the evidence that infants can form categories before
they can speak;
4 understand that infants may use certain cues to aid categorization;
5 explain what is meant by single-process and dual-process models of
categorization;
6 evaluate these models in relation to evidence from infants’ categorization of
human and non-human animals, and developmental trends;
7 consider possible links between categorizing and naming.
1 Introduction
In this first chapter we look at how very young infants begin to make sense of the
world in which they find themselves. We start by considering how important it is
in everyday life to be able to see the links between similar experiences and to
group them together. It is through this grouping that our mental world becomes
structured and that we form concepts.
When you, as an adult, see an unfamiliar thing, it is a basic reaction to try to fit it
into a category, to recognize what sort of thing it is. And much of the time this is
not really a conscious process; it is only if we come across something that is really
unfamiliar, such as a strange organism in an aquarium, that we are aware of our
categorizing. We might for example see an odd thing in a tank that might be a
seaweed, might be a worm, or might be some other sort of living being. A lot of
categorizing is tied up with language; if there is someone with us at the aquarium,
we are very likely to say ‘What is that?’ to our companion, when we see the
strange creature, to try to fit it into a category that has a name.
These observations point to the idea that much of our making sense of the
world is to do with the very rich and extensive system of categories that we all
have. These have been acquired from all sorts of different sources, from books
and other media, from schooling and from talking with other people, as well as
our own direct experience.
But what about very young babies, who cannot ask what things are, who do
not have ‘words for things’? What sense do they make of the sights and sounds
that confront them in the first weeks and months after their births? Is each
successive experience totally new, or do they begin to register the similarities
among repeated experiences of the same kinds of things?
24 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
t
Activity 1 Recognizing and categorizing
Allow about This activity will help you to think about the ways in which people form and use categories.
5 minutes
Look at the pictures in Figure 1 and see whether there are any fruits that you recognize.
Think about what types of information you used from the pictures and from your own past
experience to make this judgement. Next, decide which picture is the odd one out and again
think about what information you used to make that judgement.
Comment
Before we turn to the questions about the information you used to make these judgements,
let us first consider what was going on in this activity. These are all pictures of tropical fruit.
They are all members of a category; in fact a subcategory, ‘tropical fruit’, nested in the broader
category ‘fruit’.
In the first sentence of the activity, the word ‘recognize’ was used. ‘Recognize’ is made up of
two parts, ‘cognize’ (a word rarely used by itself), which refers to forming knowledge and ‘re’,
which suggests a repetition. If you have seen all these fruits before, if you have ‘re-cognized’
them, you have made use of your previous experience and knowledge and probably found it
quite easy to separate one from another. But to achieve that category distinction you made
use of some features of the fruits to do so: What were they? Probably they were a mixture of
visible attributes like the shape of the fruit, its inside structure and the texture of the surface,
along with characteristics like taste and smell, which you might know about.
But if you are unfamiliar with some or all of these fruits, you might still have grouped two
together and separated one out using visual features only. You might have focused on the fact
that the inside of the durian fruit (Figure 1, right ) is different from the other two fruits that
have similar insides. Thus you were categorizing the longan (Figure 1, middle) and the
rambutan (Figure 1, left) as members of one category, and the durian as a member of another
category, (although they are all members of the category, ‘tropical fruit’).
s
26 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Activity 1 should have helped to give you an awareness of the mental processes
behind ‘re-cognition’. To break this process down further, you could say that to
recognize something as a ‘thing’ that has been experienced before, as a member
of a category or class, involves two basic cognitive processes. First, the storing of
some sort of memory of an experience and, second, the comparing of a later
experience with the memory trace and registering that there is a similarity. Human
cognition thus involves the building up of an internal, mental structure that has a
relation to corresponding structure in the outside world. These mental structures
lie at the core of our psychological functioning, since life without any of the
regularity and predictability that they provide would be chaotic and highly
inefficient. Life would hardly be possible if each new encounter with the world
involved working out anew, from scratch, appropriate ways of responding and
behaving. Instead, it is a central, almost a defining feature of mental and
behavioural function that objects and events in the world are encountered in
terms of their similarities and differences; that the elements of our world are seen
as falling into many classes and subclasses. The ability to recognize new things as
being members of classes that we have experienced before gives humans a real
advantage in responding to situations rapidly, appropriately and effectively.
Across the animal kingdom there are numerous ways that categorical forms of
mental structure can be genetically programmed, as, for example, in many
animals’ reactions to potential predator-like entities. This is one of the most
effective ways in which evolution provides animals with survival mechanisms. It
is obviously a great advantage to be able to rapidly identify something as
belonging to the class ‘potential predator’ and to behave accordingly. For some
organisms, usually those that inhabit highly predictable niches, this sort of
‘cognition’ is all that is necessary. For example, a shrimp shoots backwards if
anything large looms nearby. However, other organisms need to be able to adapt
their behaviour to more varied environments, and for them it is also important to
be able to generate new classes and forms of corresponding behaviours. For
humans, this is the predominant way of classifying the world, through learning
about things and building new mental structures that aid recognition. Thus, a
central question for developmental psychologists is ‘How is mental structure
formed and elaborated?’
Clearly, once a child has acquired language, this provides them with a powerful
tool for learning about new classes. Indeed, language is in itself a vast repository
of classes. Nouns represent a multitude of objects and events, and verbs represent
a vast number of actions. The natural reaction to a child’s question ‘What is that?’
is simply to give the ‘class name’ of whatever it is that the child is asking about.
Chapter 2 focuses on the first stages of language acquisition and considers how a
child comes to understand the referential function of language and that words
may represent ‘invitations’ to form categories (Waxman and Markow, 1995). But
before infants have access to the use of language in this way, do they lack the
structures that enable them to represent objects? Clearly not, because very young
infants can quickly come to group objects based on their similarity, which is at
least an initial form of classification.
1 EARLY CATEGORY REPRESENTATIONS AND CONCEPTS 27
Finding out more about infants’ abilities to classify entities in the world, before
they start to use language as an aid, offers an opportunity to uncover the basic
cognitive processes that language builds on and to gain a better understanding of
human cognition and its development. New experimental methods, which get
around the problem that we cannot ask infants directly, have begun to uncover
some basic features of this foundational period of development.
involved in each? And are there changes in how infants’ categories develop
between the time that they are born and the emergence of language? Before we
can begin to answer these questions, however, it is necessary to investigate how
and to what extent infants are able to identify, form and store categories. This is
the subject of Sections 2 and 3.
BOX 1
most likely that the infant’s attention will be given more to the novel one in preference
to the now familiar one. This is called a ‘novelty preference’.
The familiarization/novelty-preference procedure has two stages, as you can see in
Figure 2. In Stage 1, ‘Familiarization’, the infants are shown a number of different
exemplars (members) of a single category one after another. In Figure 2, the category
chosen was ‘cats’ and so the infants are shown four different cat pictures. In Stage 2,
‘Novelty-preference’, the infants are given a preference test. In the test, a novel
member of the familiar category (in Figure 2, a cat) is shown alongside a novel member
of a novel category (in Figure 2, a dog). The results from the preference test are then
observed and interpreted. If infants look longer at the novel member of the novel
category, this is a good indicator that they have formed a representation in their
memories of the familiar category – ‘a category representation’. So what these
experiments do is play off the novelty value of a previously unseen exemplar from
within a familiar category (of which infants have just seen several exemplars) against the
novelty value of a previously unseen exemplar from a novel category. If infants give more
attention to the latter, then we can assume that they are seeing the novel exemplar
from the familiar category as ‘just another one of those’, whereas the response to the
novel category exemplar is more like ‘Aha! Here’s a new sort of thing!’
Stage 1.
Familiarization set of stimuli.
First, these are shown
to the infant, one after another.
Stage 2.
Test pair of stimuli.
Next, these are shown to the infant,
both at the same time, to see which
attracts more attention.If the infant
looks longer at the dog (the novel
exemplar of the novel category)
then we could infer that the infant
is applying a category something
like ‘cat’, based on the Novel exemplar Novel exemplar
familiarization set and possibly also of novel category of familiar category
on previous knowledge.
These experiments are usually carried out in a laboratory so that the conditions can be
well controlled. Stimuli are often shown to infants on computer screens and are
computer-controlled, so that the presentations of the images can be accurately timed.
Infants are sat at a fixed distance from the stimulus display, close enough so that their
limited ability to focus does not affect their perception.
n
30 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
t
Activity 2 Evaluating the familiarization/novelty-preference test
Allow about This activity will raise your awareness of possible flaws or limitations in the familiarization/novelty-
5 minutes preference method.
When analysing scientific studies, it is always important to look for alternative, perhaps
simpler, explanations of observations. With this in mind, think about the following question.
If infants generally look for longer at the dog in the experiment illustrated in Figure 2, are
there other possible explanations for this result, other than the one offered in Box 1? Try to
come up with two others, and make a note of your answers.
Comment
One possibility is that infants might generally prefer to look at the picture of the dog, perhaps
because it is more colourful or has a more interesting shape. A way of avoiding this sort of
problem is to take another group of infants who have not seen the familiarization stimuli, and
test them on a number of different dogs paired with different cats. If there is no prior
preference, we can be reasonably confident that any preferences found are due to the effect
of the set of images shown during the familiarization phase.
Another possibility is that young infants might not be as good as adults at distinguishing
between different cats; one might look much the same as another to the infant, but that does
not mean that they are seeing them both as members of a single category, i.e. cats. A way of
checking this out would be to show another group of infants the picture of one cat followed
by just a picture of another cat. If the infants’ attention declines as they looked at the first
picture and then increases for the next, this would be good evidence that they can
discriminate between the two within-category exemplars. These are the kinds of controls used
in well-designed experiments on infant categorization that can be used to guard against
misleading results.
It is worth noting that the results from the sorts of studies that we will be considering in this
chapter come from analysing responses made by a number of infants. There is, typically, a lot
of variation in how infants behave. Some fail to engage properly with the task and hence do
not provide data. The findings reported are based on significant trends observed across whole
groups of infant participants.
s
novel category, e.g. a triangle. This method works on the following rationale: if
infants form a category representation that recognizes similarities among the
distorted versions of the shape, then we can expect the representation to be close
to the undistorted ‘prototype’ shape (e.g. a square).
Comment
If infants had established the category ‘square’ from the set of distorted exemplars
during familiarization, which of the two test stimuli (a prototype square or a
prototype triangle) do you think they would find more interesting? (Clue: remember
that infants generally prefer looking at novel, unfamiliar things.)
Research summary 1 gives details of one of these studies, so that you can see how
the data from such research are used to find out more about infant categorization.
RESEARCH SUMMARY 1
Prototype Distortion
Good
Intermediate
Poor
Figure 3 Good, intermediate and poor prototypes and their distortions (Younger and
Gotlieb, 1988).
Next, the infants were shown a test pair that included a prototype (i.e. an undistorted
shape) of the distorted forms that they had just seen, alongside a prototype of another
form. This was carried out in order to test whether the infants were treating the
32 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
prototype of the distorted forms as more familiar than a different form. (Bear in mind
that they had not actually seen this prototype before; they had only seen a series of
distorted versions of it.) The percentage of time that infants spent looking at the novel
prototype was then observed and analysed as shown in Table 1.
Between-category comparison
percentage of time spent looking at novel prototype
Good
Intermediate
Poor
The first thing to note about these results is that in all conditions, except for the
intermediate form for 3 month olds, infants were tending to look for longer at the
novel prototype, which belonged to a class from which they had not previously seen
distorted exemplars. This is indicated by the fact that all the percentage looking times
(except one) are greater than 50 per cent. This shows that the novel prototype was
attracting more attention than the prototype of the exemplars that they had seen
before. This means that they were behaving as if the prototype of the distorted forms
that they had seen was more familiar to them (even though they had not actually seen
this undistorted prototype before). However, this preference was only reliably
greater than chance variations (shown by the asterisks) for the good form for 3 month
olds, for the good and intermediate forms for the 5 month olds and for all forms for
the 7 month olds. Clearly, the older the infants, the better they were at extracting the
prototype as a category from the distorted exemplars.
n
1 EARLY CATEGORY REPRESENTATIONS AND CONCEPTS 33
In the experiment described in Research summary 1 the novel prototype from the
novel category was preferred by most infants, indicating that they had formed a
category representation as a result of looking at the familiarization set of distorted
exemplars. Even though they had never actually seen the prototype of the familiar
category, they were, nevertheless, treating it as familiar. Their representation of
the category was sufficiently structured so as to include novel exemplars from the
same, familiar category and exclude novel exemplars from novel categories.
Figure 4 Cat and dog stimuli (Quinn et al., 1993; Eimas and Quinn, 1994).
In addition, infants of this age range who were familiarized with horses treated
novel horses as familiar, but showed greater interest in cats, giraffes, and zebras
(Eimas and Quinn, 1994). These findings indicate that young infants can form
separate representations for cats and horses each of which excludes instances of
the other category, as well as excluding exemplars of other related animal species.
Although, as we have seen, infants can group objects into categories, perhaps
they do not at first form these sorts of nested sets of categories? It is possible that
at first young infants simply recognize different classes of things. For example,
they might group cats into one category and dogs into another, but not recognize
that at a higher level, both of these categories fall within the category of ‘domestic
pets’ or the broader category of ‘animals’.
Behl-Chadha (1996) wanted to find out whether infants are able to form the
sorts of ‘categories within categories’ outlined in Table 2. To do this, she extended
the findings of studies looking at early categorization differentiation among
animal species and designed an experiment to discover whether infants could
form separate categories for particular types of furniture. Furniture was chosen as
a stimulus class because infants have a lot of exposure to furniture from birth
onwards.
In the experiment, infants of 3 and 4 months of age were first familiarized with
twelve realistic photographs of chairs (including examples of the subordinate
categories: armchairs, desk chairs, kitchen chairs, rocking chairs and stuffed
chairs). Then, they were shown novel chairs, along with non-chair furniture
(either couches, beds or tables). The infants gave more attention to these non-
chair pictures than to the novel chairs. This indicates that they had formed a
category equivalent to ‘chair’ that did not include other types of furniture. When
they were familiarized with a set of photographs of couches, they treated novel
couches as familiar, but showed novel category preferences for chairs, beds and
tables by looking at these for longer, showing that in this case they had formed
category representations for ‘couch’. This is good evidence that 3 and 4 month
olds can form individuated representations for chairs and couches as separate
categories, separate also from the categories of beds and tables. In other words,
nested within the infants’ category ‘furniture’ were further, ‘basic’ or ‘intermediate’
categories for chairs, couches, beds and tables.
Behl-Chadha (1996) also found that infants could form more global (inclusive)
category representations for broader classes of stimuli. In one experiment, 3 and 4
month olds were familiarized with photographs of mammals (deer, domestic cats,
elephants, horses, rabbits, squirrels, tigers and zebras). Then, photographs of an
animal from a novel mammal category were paired with either (1) a novel example
of a familiar mammal category (2) a non-mammalian animal (bird or fish) or (3) an
item of furniture. The infants preferred birds, fish and furniture to instances from
the novel mammal categories, but they did not prefer members of novel mammal
categories to novel members of familiar mammal categories. These findings
indicate that young infants can form a broad category representation for mammals
that includes examples from novel mammal categories that they had not seen
during familiarization, and excludes some non-mammals (birds and fish) as well as
furniture. However, what is even more interesting is that within the broad category
representation ‘animal’, they can use nesting categorization skills. In other words,
within the animal category, they can form basic category representations for
mammals.
1 EARLY CATEGORY REPRESENTATIONS AND CONCEPTS 37
Results from the same series of experiments backed up these findings, showing
that 3–4 month olds can make the same kinds of ‘nested’ distinctions the other
way round – i.e. for furniture versus animals. The results showed that infants
could form a category representation for furniture that includes beds, chairs,
couches, cabinets, dressers and tables, but excludes mammals. This evidence
suggests that young infants can form global category representations for at least
some natural (mammal) and artefact (furniture) categories and this is strong
evidence that infants can, indeed, form nested category representations. These
experimental outcomes are highly significant because they indicate quite clearly
that young infants have the cognitive abilities to group stimuli into categories that
are similar to many of the cognitive groupings that older children and adults use.
This is also strong evidence that infants are forming categories on the basis of
their experience, because evolution could hardly have provided us with innate
templates for types of furniture!
RESEARCH SUMMARY 2
(a)
(b)
The results were that both groups of infants showed a preference for the novel spatial
category, consistent with the idea that they had formed category representations for
the ‘above’ and ‘below’ relations between the dot and the horizontal bar. More recent
research using the same method indicates that by 6 or 7 months of age, infants can form
category representations for ‘between’ and ‘outside’ (Quinn et al., 2003).
n
Given that infants can categorize objects in quite sophisticated ways, the finding
that they can also categorize the ways in which objects are spatially located shows
that they have the basic cognitive building blocks (‘primitives’, as they are often
called) to form quite complex representations of the physical world. These
representations also provide infants with a foundation for learning to
communicate about different types of information: to learn the words ‘dog’, ‘table’
and ‘under’, and then say things like ‘dog under table’ if a dog is below a table. So,
clearly, before they can speak, infants have the necessary mental abilities to
construct category representations that will be useful in later development to
support their use of language (see Chapter 2).
What is a ‘cue’?
In its everyday sense, a ‘cue’ is some sort of signal that launches a sequence of
behaviour or some other response. For example, a green traffic light is a cue to
the behaviour of moving forward across a road junction. The word ‘cue’ is used in
a similar way in psychology, to refer to something that indicates or triggers
something else, such as behaviour or recognition. A cue could be in any sensory
modality; it could be a visual signal, a sound, a touch or a taste. If you were
meeting someone at a railway station, for example, spotting their hat might serve
as a visual cue to trigger your recognition of them. Or, if you were waiting for a
particular programme on the radio, the signature tune would give you an auditory
cue that the programme is about to start. In categorization, then, we often use
specific cues as indicators of which category something belongs to. Specific cues
might include the values of attributes such as colour or shape.
One strategy that has been used to identify the cue (or cues) that infants use to
form a particular category representation is to show that they form the category
representation when a particular cue is present, but not when the cue is absent.
Such a strategy has been used to find out, for example, how infants form distinct
category representations for dogs versus cats. Because the two species have
considerable perceptual overlap (both have facial features, a torso, four legs, fur
and tails) it is not immediately obvious what information infants might be using to
form and differentiate the categories.
t
Activity 3 What are the differences between dogs and cats?
Allow about This activity will help you to identify the possible cues that might be used to differentiate the category
5 minutes ‘dog’ from the category ‘cat’.
What visual features mark the difference between dogs and cats for you? In other words,
what would you look for to decide whether an animal that you saw for the first time is a dog
or a cat? You might like to try visualizing (or drawing) a ‘prototypical’ dog and cat, and see
what are the main features of your visual images and how the dog and cat differ.
40 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Comment
For many people, this is not a very easy activity. Although they assert that they have little
difficulty in distinguishing a cat from a dog, they have a lot more difficulty in saying how they do
so. This activity serves to remind us that our categorization processes are not always fully
accessible to consciousness.
s
It is possible that infants are able to make categorical distinctions on the basis of
differences in single attributes, in the pattern of correlations among attributes or
in the overall combination of attributes. For example, in the case of the visible
features of birds, beaks and feathers are reliably correlated; they almost always
co-occur. Hence, this correlation between attributes might feasibly form a basis
for categorical distinctions. Or, in the case of the ‘snake’ category, a long,
cylindrical body is a reasonable basis, although using this cue alone would
include worms, slugs and legless lizards in the category. One way of finding out
how infants select from the range of visible information is to vary systematically
the attributes shown by a category exemplar.
Whether infants use a subset of cues in order to categorize was the subject of a
series of studies by Quinn and Eimas (1996b), Spencer et al. (1997) and Quinn
et al. (2001a). These studies found that infants categorized animals when the
exemplars presented during familiarization and test trials showed only the head
region of the animal, but they failed to show the categorical distinction when only
the body region was displayed (Quinn and Eimas, 1996b; Quinn et al., 2001a).
Infants also formed category representations based on the head (and not the
body) when they were familiarized with whole cats or dogs and were then
preference-tested with a pair of hybrid stimuli: a novel cat head on a novel dog
body versus a novel dog head on a novel cat body (Spencer et al., 1997).
Examples of hybrid stimuli used in an experiment by Quinn and Eimas (1996b)
are shown in Figure 6.
So it seems that, at least in the case of dogs and cats, infants do not make use of all
the available cues for categorization, but they do tend to focus on a specific
subset. These results suggest that infants make the most use of perceptual cues in
the head region to distinguish cats from dogs and that cues from the rest of these
animals’ bodies are less important. From this set of experiments at least, we can
say that infants may select certain attributes as a basis for categorization.
1 EARLY CATEGORY REPRESENTATIONS AND CONCEPTS 41
BOX 2
n
There is evidence that salient part differences (e.g. legs versus wheels) can play a
role in signalling broader (global) category contrasts in older infants. Rakison and
Butterworth (1998) examined categorization of toy objects by 14–22-month-old
infants, using the sequential touching procedure, and found that the infants
would categorically differentiate between animals and vehicles. However, in a
subsequent experiment, when the legs were removed from the animals and the
wheels from the vehicles, the infants no longer showed the category
42 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
differentiation. This supports the idea that older infants use specific, salient
attributes to make broad contrasts between global categories of objects, for
example, using legs versus wheels to distinguish animals from vehicles.
their summary of the single-process model, ‘a representation like animal that may
begin by picking out relatively simple features from seeing and other sensory
modalities comes over time to have sufficient knowledge to permit specifying the
kind of thing something is through a single continuous and integrative process of
enrichment’ (Quinn and Eimas, 2000, p. 57). We can call this a perceptual learning
model. Computer simulations (connectionist models) of learning networks have
supported the validity of the single-process framework to some extent. It has also
been found that the same computer model could form category representations
based only on perceptual features or on arbitrary labels that classify exemplars
(Quinn and Johnson, 1997).
BOX 3
t
Activity 4 Are the single- and dual-process models competing or
complementary?
Allow about This activity will help you to understand the differences and possible similarities between these two
10 minutes models of categorization processes.
Consider these two models of category formation. Do you think that they really are
alternative, competing explanations, or might they both have some validity? One way of
approaching this activity could be to list the features of each model in a table, to help you to
see points of similarity and divergence between the two models.
46 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Comment
You might feel that perhaps the dual-process model is not so very different from the single-
process model, in that a single-process model could incorporate functions and names, which
are clearly not just straightforward perceptual features. As well, the idea of ‘function’ might
itself be seen as a form of redescription of perceptual schemas, in other words, based on
things that are visible. Similarly, we might question whether an attribute like ‘has legs’, which at
first sight seems an obvious perceptual feature might also be tied up with the use of ‘legs’ for
locomotion, a functional aspect: in other words, is the notion of ‘legs’ also tied up with these
limbs being used for walking?
s
Experiments have not yet been able to determine conclusively whether a single-
or dual-process model best describes how infants’ categories develop; the
question is still an open one. Although this specific issue has yet to be resolved,
researchers have clarified some other aspects of infants’ categorization abilities.
One of these aspects is the way in which infants’ categories are organized into
larger structures of knowledge about the world, a topic which we turn to in the
next section.
BOX 4
Object examination
Object examination is a method that has been used for studying category
representation in older infants from about 6 months of age upwards (Oakes et al.,
1991). This procedure, like the familiarization/novelty-preference method, depends on
infants gradually losing interest in successive new members of a single category and
then regaining interest in a member from a novel category.
For example, if a vehicle versus animal category contrast were being studied, infants
would be given toy vehicles (e.g. a tractor, a car, a lorry, a bus, a quad-bike and a
motorcycle) one after the other, each for a fixed number of seconds (typically 20–30).
Infants are allowed to look at and manipulate the toys. They are then given a toy animal,
for example, a horse (novel exemplar from a novel category), for the same number of
seconds. A measure of active examination (that combines the amounts of handling and
looking at each toy) is then used to see whether infants examine the novel exemplar
from the novel category significantly more than the previous one or two vehicle toys. If
we know that the infants can discriminate among the instances from the familiar
category and that they do not have a pre-existing preference for the novel category
exemplar (remember the precautions that are taken to rule out these possibilities) then
one can conclude that categorization is taking place.
n
48 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
A perceptual learning account (i.e. single-process model) has been put forward to
describe this differentiation-based pattern of findings on the early development of
object categorization (Quinn and Eimas, 1997, 2000; Quinn and Johnson, 1997,
2000; Quinn, 2002). At first, infants seem to learn global categories like mammals
and furniture based on the presence or absence of salient features. Then they go
on to differentiate these into basic-level categories like cats and dogs based on
specific values of shared features. This observation is supported by the fact that
simple connectionist networks also learn category representations in a global-to-
basic order when provided only with input features (e.g. leg length, body length)
that are measured directly from the stimuli shown to the infants (in other words,
they are basic perceptual attributes). This again lends some support to the single-
process model of category development.
Figure 7 Example of spatial relations test stimuli (based on Quinn et al., 1996).
1 EARLY CATEGORY REPRESENTATIONS AND CONCEPTS 49
t
Activity 5 Separating location and object
Allow about This activity will help you to understand how the familiarization/novelty-preference method can be
5 minutes used to address a new research question about infants’ capacities to categorize locations separately
from the objects found in those locations.
If infants are able mentally to separate an object’s position (above/below) from the shape of
the object that is in that position, which of the test stimuli shown in Figure 7 will they look at
for longer? Make a note of your answers. (Remember that infants tend to prefer novelty:
think about which of the two test stimuli would seem more novel.)
Comment
In the familiarization phase, infants became familiar with a particular shape (e.g. a dot) and a
particular location (e.g. ‘above’). If they can separate out object and position information, then,
because they have been familiarized to ‘above’, they should see the right-hand test stimulus as
novel because it shows the spatial relation ‘below’ and they will look at it for longer.
s
In contrast to the infants tested in the original ‘above’ versus ‘below’ experiment
described in Section 2, the infants tested in this new version of the categorization
task did not show a preference for the novel spatial category test stimulus, instead
they divided their attention equally across both test stimuli. This result suggests
that infants of this age do not form the abstract category representations for
‘above’ and ‘below’ separately from the particular objects showing the relation.
A follow-up experiment (Quinn et al., 1996) found that older infants, aged 6–7
months, did prefer the novel spatial relation in the test phase. This result shows that
by this age they were now able to separate their more abstract categories of spatial
location from the changes in the identity of the objects. A similar concrete-to-
abstract developmental pathway has been reported for the spatial relation
‘between’ in the age range from 6–7 months to 9–10 months (Quinn et al., 2003).
These results, taken together, support the idea that category representations of
spatial relations may be initially tied up with the objects depicting the relations, but
later become independent of the objects.
If this is true, then it should be relatively easy to modify the way that young
infants categorize non-human animals. One study found that infants who were
familiarized with a series of cats, then showed a novelty preference for a dog over a
novel cat, suggesting that in this condition they were indeed using a ‘cat’ category
that excluded dogs (Quinn et al., 1993). However, in a second condition, another
group of infants who were shown a series of dogs did not then show a novelty
preference for a cat over a novel dog. This suggests that the ‘dog’ category formed in
this second condition was an inclusive category – it was broad enough to
encompass cats as well. So in some conditions infants can form narrow categories
(e.g. a category of cats that excludes dogs) and in other conditions they can form
broad categories (e.g. a category of dogs that includes cats). This suggests that
infants do not have rigid categories for ‘cats’ and ‘dogs’ that they apply in all
conditions.
Two further experiments support the idea that these are flexible categories and
that the dog versus cat asymmetry only arose because of the particular sets of stimuli
that were used (Quinn et al., 1993; French et al., 2001). In one of these experiments
(Quinn et al., 1993), the variability of the dog stimuli was reduced (the stimuli were
chosen to be more similar to each other). In this case, 3–4-month old infants did not
include cats in the ‘dog’ category that they found. In another experiment with
infants of the same age, new sets of dog and cat pictures resulted in the ‘cat’ category
this time including ‘dogs’ (French et al., 2001). Thus, when familiarized with
exemplars that show a lot of variation in their features and overlap between
categories, infants tend to form broad, inclusive categories. But if they are
familiarized with very similar exemplars within a category and no overlap between
categories, their category representation tends to be much narrower. These
findings, taken together, indicate that these non-human animal categories were
quite fluid and determined more by the experimental condition than any prior
experience.
The studies of the categorization of cat and dog images by young infants makes
the important point that the infants seem to be forming their category
representations for non-human animals over the course of the familiarization
trials (i.e. bottom up), rather than tapping into pre-existing concepts that had
been formed prior to arriving at the laboratory (i.e. top down). If infants in the cat
versus dog categorization studies had simply been tapping into category
representations established before the experiments, then the experimental
changes should have had no effects. The fact that infant responsiveness did vary
across experiments suggests that the categories were being formed during the
familiarization experience and that the boundaries of categories can be relatively
easily affected by modifying the familiarization stimuli. Thus we can see that
infants can not only form complex category representations, but also that these
are open to change in the light of new experiences.
values of attributes (e.g. clothing, and types of clothing, upright stance and
different gaits, hair colours and different styles, speech, etc.).
A study by Quinn and Eimas (1998) familiarized 3- and 4-month-old infants
with photographs of twelve humans (male and female) in a variety of standing,
walking and running poses. Examples of the human stimuli used in the
experiment are shown in Figure 8. The infants were then preference-tested with a
novel human paired with a cat and a different novel human paired with a horse.
Most surprisingly, no novel category preferences for either the cats or horses were
observed. (There was no spontaneous preference among the infants for humans
over non-human animals, which might have mistakenly led us to believe that
infants categorized humans as different from horses and cats.) It seems unlikely
that infants cannot tell the difference between humans, and cats and horses, so
something rather odd seems to be going on.
Object As well as the categorization task, three other assessments of children’s cognitive
permanence test development were carried out each time that the children were observed: a
A test of whether
infants believe
measure of language use, a test of means–end behaviour and a test of object
that an object permanence. The language assessment comprised a combination of a diary and a
continues to exist questionnaire which the children’s mothers used to give data about how many
after it has been
words the children were using. This measure was employed to identify the point
hidden from view.
at which children showed a sudden increase in the number of names that they
used for things, the so-called ‘vocabulary spurt’, which is also called the ‘naming
explosion’. This phenomenon, which is seen to happen in many children’s
development, is discussed further in Chapter 2.
Means–end test
The means–end test assessed how well children could solve problems that
A test that involved some planning, such as using a stick to obtain an object. The test of
involves carrying object permanence assessed how aware children were of the continued existence
out one (or more)
and location of objects when they were not visible. The reason why these other
actions to achieve
some other result. assessments were carried out was to find out whether language use and
For example, categorization were truly closely linked, or whether they might simply be
pulling a cloth on developing together as part of a more general improvement in cognitive abilities.
which a toy is
placed in order to The findings of the experiment are shown in Table 4. The table shows the
get the toy. levels of correlation between the ages at which the highest levels of
categorization, ‘vocabulary spurt’, means–ends and object permanence were
achieved. It can be seen that of the only two significant correlations, the strongest
was between categorization and naming, with a significant but weaker correlation
between object permanence and vocabulary. This pattern of observations
suggests that there is a specific link between categorization and the vocabulary
spurt, and that a developmental change in categorization-related behaviour goes
along with a related change in naming. The association with the development of
object permanence suggests that other processes may also be operating, not just
the link between categorizing and the vocabulary spurt. These children became
much more interested in learning and using the names for things at the same time
as they showed a development in how they categorized objects.
Categorization – – 0.78**
It is important to note that categorization was not strongly linked with the other
cognitive measures, nor were these other measures strongly linked with each
other. This helps to rule out the possibility that what was being observed was just
a general improvement in cognitive ability in the children, but that something
more specific, linking categorization and vocabulary, was going on.
8 Conclusion
In this chapter we have shown that new experimental techniques have provided
very productive ways of gaining a window into infants’ minds and how they
operate. We have seen how infants are able to group things into categories in
remarkably flexible and sensitive ways. They are selective in the cues that they
use to categorize different sorts of things and they can, very early in their lives,
build the beginnings of complex, multi-levelled meaning structures. Having the
abilities to form these rich structures from the earliest months of life gives infants a
powerful tool to help them in making sense of the world of people and things
that surround them and provides an essential foundation for almost all aspects of
human experience.
This chapter has also explored the developmental pathway of infants’
categorization abilities and looked at how this fits in with language development.
We found that infants’ initial categories seem very much based on what things
look like and that more abstract, less immediately obvious features are only used
in categorization later on in the period of infancy. It is only towards the end of this
period, as infants begin to use their first words, that their categorizing becomes
better at handling more abstract ideas separately from the visible objects to which
they relate. And when their language vocabulary suddenly begins to race ahead,
this appears to have a basis in a further development in categorization, which
may be the point at which categories make the transition to becoming true
concepts that are richly imbued with meanings. This is the exciting time when
children are learning to communicate effectively with language and enter more
fully into their social worlds. But that is a topic for the next chapter.
Acknowledgements
The preparation of this chapter was supported by NSF Grant BCS-0096300 and
NIH Grant HD-42451.
References
Arterberry, M. E. and Bornstein, M. H. (2001) ‘Three-month-old infants’
categorization of animals and vehicles based on static and dynamic attributes’,
Journal of Experimental Child Psychology, vol. 80, pp. 333–46.
Arterberry, M. E. and Bornstein, M. H. (2002) ‘Infant perceptual and conceptual
categorization: the roles of static and dynamic stimulus attributes’, Cognition, vol.
86, pp. 1–2.
Behl-Chadha, G. (1996) ‘Basic-level and superordinate-like categorical
representations in early infancy’, Cognition, vol. 60, pp. 105–41.
Bomba, P. C. and Siqueland, E. R. (1983) ‘The nature and structure of infant form
categories’, Journal of Experimental Child Psychology, vol. 35, pp. 294–328.
58 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Deloache, J. S., Kolstad, V. and Anderson, K. (1991) ‘Physical similarity and young
children’s understanding of scale models’, Child Development, vol. 62,
pp. 111–26.
Eimas, P. D. and Quinn, P. C. (1994) ‘Studies on the formation of perceptually
based basic-level categories in young infants’, Child Development, vol. 65,
pp. 903–17.
Fantz, R. L. (1963) ‘Pattern vision in newborn infants’, Science, vol. 140, pp. 296–7.
French, R. M., Mermillod, M., Quinn, P. C. and Mareschal, D. (2001) ‘Reversing
category exclusivities in infant perceptual categorization: simulations and data’, in
Stenning, K. and Moore, J. (eds) Proceedings of the 23rd Annual Conference of
the Cognitive Science Society, pp. 307–12, Mahwah, NJ, Erlbaum.
Gopnik, A. and Meltzoff, A. N. (1987) ‘The development of categorization in the
second year and its relation to other cognitive and linguistic developments’, Child
Development, vol. 58, pp. 1523–31.
Gopnik, A. and Meltzoff, A. N. (1992) ‘Categorization and naming: basic-level
sorting in eighteen-month-olds and its relation to language’, Child Development,
vol. 63, pp. 1091–103.
Karmiloff-Smith, A. (1986) ‘From meta-processes to conscious access: evidence
from children’s metalinguistic and repair data’, Cognition, vol. 23, pp. 95–147.
Madole, K. L. and Oakes, L. M. (1999) ‘Making sense of infant categorization:
stable processes and changing representations’, Developmental Review, vol. 19,
pp. 263–96.
Mandler, J. M. (1992) ‘How to build a baby: II. Conceptual primitives’,
Psychological Review, vol. 99, pp. 587–604.
Mandler, J. M. (1997) ‘Development of categorization: perceptual and conceptual
categories’, in Bremner, G., Slater, A. and Butterworth, G. (eds) Infant
Development: recent advances, pp. 163–90, Hove, Psychology Press.
Mandler, J. M. (2000) ‘Perceptual and conceptual processes’, Journal of Cognition
and Development, vol. 1, pp. 3–36.
Mandler, J. M. and McDonough, L. (1993) ‘Concept formation in infancy’,
Cognitive Development, vol. 8, pp. 291–318.
Mandler, J. M. and McDonough, L. (1996) ‘Drinking and driving don’t mix:
inductive generalization in infancy’, Cognition, vol. 59, pp. 307–35.
Mandler, J. M., Bauer, P. J. and McDonough, L. (1991) ‘Separating the sheep from
the goats: differentiating global categories’, Cognitive Psychology, vol. 23,
pp. 263–98.
Millikan, R. G. (1998) ‘A common structure for concepts of individuals, stuffs, and
real kinds: more mama, more milk, and more mouse’, Behavioural and Brain
Sciences, vol. 21, pp. 55–100.
Oakes, L. M., Madole, K. L. and Cohen, L. B. (1991) ‘Infants’ object examining:
habituation and categorization’, Cognitive Development, vol. 6, pp. 377–92.
1 EARLY CATEGORY REPRESENTATIONS AND CONCEPTS 59
Quinn, P. C., Slater, A. M., Brown, E. et al. (2001b) ‘Developmental change in form
categorization in early infancy’, British Journal of Developmental Psychology,
vol. 19, pp. 207–18.
Rakison, D. and Butterworth, G. (1998) ‘Infants’ use of object parts in early
categorization’, Developmental Psychology, vol. 34, pp. 49–62.
Rakison, D. and Poulin-Dubois, D. (2001) ‘Developmental origin of the animate–
inanimate distinction’, Psychological Bulletin, vol. 127, pp. 209–28.
Spencer, J., Quinn, P. C., Johnson, M. H. et al. (1997) ‘Heads you win, tails you
lose: evidence for young infants categorizing mammals by head and facial
attributes’, Early Development and Parenting, vol. 6, pp. 113–26.
Waxman, S. R. and Markow, D. B. (1995) ‘Words as invitations to form categories:
evidence from 12- to 13-month-old infants’, Cognitive Psychology, vol. 29,
pp. 257–302.
Younger, B. A. and Gotlieb, S. (1988) ‘Development of categorization skills:
changes in the nature or structure of infant form categories?’, Developmental
Psychology, vol. 24, pp. 611–19.
Younger, B. A. and Fearing, D. D. (1999) ‘Parsing items into separate categories:
developmental change in infant categorization’, Child Development, vol. 70,
pp. 291–303.
Chapter 2
First words
Margaret Harris
Contents
Learning outcomes 63
1 Introduction 63
1.1 Understanding and producing words 63
2 Recognizing speech 65
2.1 Identifying speech sounds 65
2.2 Cues to word boundaries 68
7 Conclusion 94
References 95
Readings 98
Reading A: Is Dutch native English?
Linguistic analysis by 2-month-olds 98
Reading B: Developments in early lexical
comprehension: a comparison of parental report
and controlled testing 106
2 FIRST WORDS 63
Learning o u t c o m e s
After you have studied this chapter you should be able to:
1 describe the task that confronts the young child in learning first words and
explain how this compares with the adult experience of acquiring
vocabulary in an unfamiliar language;
2 describe how infants can use cues to identify word boundaries in a stream
of continuous speech;
3 outline Bruner’s arguments about the importance of the social context for
early word learning;
4 discuss ways in which reliable data can be collected on children’s early
comprehension and production of words using a variety of methods;
5 describe the typical time course of the development of children’s
comprehension and production vocabulary while being aware of
individual differences in development;
6 describe the stages in babbling;
7 understand the range of meanings that is found in early words.
1 Introduction
In the previous chapter you learned that very young babies are able to construct
and represent categories that group together similar objects. In this chapter you
will discover how young children learn about the names for objects in these
categories as well as the words used to describe actions and personal names. I
will be focusing on the first words that children say but I will also be considering
the fundamental processes and developments that lay the foundations for the
emergence of first word use.
Parents usually regard the moment when children begin to say words as very
significant but, by the time this happens, children already know a great deal both
about the way that words sound and about the way in which language is used in
familiar situations. As you have seen, they also know about object categories. All
of this knowledge is crucial as children search for the meaning of words.
t
Activity 1 Early word comprehension
Allow about This activity asks you to reflect on the strategies you use when learning new words and to compare
10 minutes your experience of word learning to that of a young child at the beginning of language development.
Imagine that you are listening to someone talking and you hear a word that you do not
recognize. Note down how you could discover what the new word meant if you could not
look it up in a dictionary or ask someone else about the word’s meaning. What difficulties
would you face?
Next, note down the difference between your experience of hearing an unfamiliar language
and what you imagine the experience of a young child to be. Remember, you already know a
lot about at least one language – your own native tongue. Think about the advantage this
gives you when you hear a new language for the first time. Then try to imagine what it would
be like if you did not know any words in any language.
Comment
At the very beginning of language development, young children do not know about the
meaning of any words so they are rather in the position that you would be in if you heard a
new language for the first time. However, you have an advantage over the young child – you
already know a lot about language from speaking your own native tongue. Young children do
not know any words in any language.
s
Speech stream The chart below summarizes the different stages that young children must go
The flow of sound through in learning about first words. You will see that there are more
produced when
people speak, requirements for the production of words than for comprehension. At the end of
made up of the chapter you will look at this chart again so that you can see the various
different aspects of early word learning that you have covered.
frequencies. If a
graphical plot of
this stream is Learning first words:stages in learning
examined, it is
often unclear
COMPREHENSION
arguing that the human capacity to learn language arises from an innate
mechanism.
Phonology Interestingly, nativist theories of language development have little, if anything,
The set of sounds to say about early word learning (although they have made some claims about
that make up the
basic building phonology that do not concern us here). Traditional nativist theories have been
blocks of speech, mainly concerned with children’s mastery of morphology and syntax, that is, the
which vary from grammatical rules for combining words into phrases. You will be reading about
one language to
another and even these aspects of language development in Chapters 3 and 4.
within languages
to some extent.
Summary of Sect ion 1
. Most infants comprehend many more words than they can produce. It is
therefore essential to consider early word comprehension before early
word production.
. Babies need to draw on more skills for word production than for word
comprehension.
. Babies do not know the meaning of any words, so their experience of
word learning differs greatly from the adult experience of learning
another language.
2 Recognizing speech
2.1 Identifying speech sounds
The first two skills that young children need in order to develop an understanding
of what words mean are the ability to recognize and remember speech sounds,
and the ability to segment words from the speech stream – in other words, to
identify where words begin and end from the flow of sounds people make when
they speak. These are the same abilities that you would use when listening to an
unfamiliar language. At first you would find it difficult to tell where one word
ends and the next begins, and if the phonology of the language is very different
from English, you may well also have problems in recognizing and remembering
the individual sounds that you hear.
Recent research has shown that babies begin learning about the language
around them even before they are born. They can hear their mother’s voice quite
clearly while still inside the womb and will respond to it (Richards et al., 1992).
Babies’ responses to speech before and just after birth have now been
investigated in a series of extraordinary experiments. These have shown that
unborn babies can actually remember and recognize some aspects of the speech
they hear.
66 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
RESEARCH SUMMARY
Recognizing voices
These dramatic findings about babies’ ability to recognize speech sounds before
birth can help to explain why they soon develop preferences for familiar speech
sounds after they are born. Babies prefer the human voice to other sounds and, a
few days after birth, according to an earlier study by DeCasper and Fifer (1980)
using the pressure-sensing dummy technique, they will increase their rate of
sucking in order to hear a recorded human voice but not to hear recorded music
or a rhythmical non-speech sound. Over the first weeks of life, this general
preference for voices over other sounds becomes more specific, and by 4 weeks
of age, infants prefer their own mother’s voice to other female voices (Mehler and
Dupoux, 1994).
t
Reading
At this point you should turn to the end of this chapter and read Reading A ‘Is Dutch native
English? Linguistic analysis by 2-month-olds’ which is the paper written by Christophe and
Morton (1998). While reading, think about the method that was used to assess the babies’
perception of the two languages. Make notes on the authors’ conclusions about the cues that
the babies were using to tell the difference between the two languages.
s
68 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Figure 1 Sound spectrogram of the phrase ‘How are you?’. Reproduced from an original
recording provided by Mark Tatham (University of Essex).
2 FIRST WORDS 69
Syllable stress
There are several possible prosodic cues to word boundaries. One is syllable
stress. Many languages, such as Italian or Greek, have a very regular pattern of
stress within words. In English, however, stress is variable as you will see in the
next activity.
t
Activity 2 Stress patterns in spoken English
Allow about This activity will help you to identify stress patterns in spoken English.
10 minutes
Say the following words aloud to yourself and underline the syllable that is stressed. For
example, telephone is stressed on the first syllable since we normally say telephone rather
than telephone. If you are not sure where the stress falls, it may help to think of stress in
terms of emphasis. You could also try saying the word in different ways, varying the stress,
and see which one sounds natural.
Comment
In typical, conversational English, around 90 per cent of content words (i.e. nouns, verbs,
adjectives, adverbs and personal names) have stress on the first syllable. You should have
found that only belong, across and guitar were exceptions to this rule. So if babies could use the
presence of a stressed syllable as a guide to the beginning of a word, they would be correct
most of the time.
Stress patterns vary from language to language and they make up an important part of a
language’s prosodic structure. As you read in the Christophe and Morton (1998) study, babies
can tell one language from another on the basis of prosody so it is perhaps not surprising to
find that they are very sensitive to stress patterns.
s
Transitional probabilities
Although syllable stress is a good cue to the beginning of a new word, it is not
always reliable. However, there are other cues available for babies to identify
word boundaries which are not based on prosody. One further example of a cue
is the probability of certain syllables appearing together. This is known as
transitional probability.
70 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
n Johnson a n d J u s c z y k ’s fi r s t s t u d y
In their first study, Johnson and Jusczyk looked at the effect of transitional
probabilities on 8 month olds. In order to consider transitional probabilities, they
had to control the infants’ prior experience of syllable sequences. To do this,
Johnson and Jusczyk invented ‘words’ by taking twelve syllables and combining
them into four sequences to make pakibu, tibodu, golatu and daropi.
There were two phases in the experiment. During the first phase – called the
familiarization phase – the infants listened to the ‘words’ repeated over and over
in random order for 3 minutes, with no pauses between them. Because there
were no pauses, what the infants heard was a long, uninterrupted string of
syllables. Furthermore, because the order of the ‘words’ was randomized, the
sequences of syllables that resulted from two words being next to each other (for
example, bu-go-la, heard only on those occasions when pakibu was followed by
golatu) were heard less frequently than the sequences of syllables that formed the
‘words’ (pa-ki-bu).
Because the sequences of syllables that formed ‘words’ occurred more often
than those that occurred as a result of one word following another, transitional
probabilities could be used to distinguish the ‘words’ from other syllable
sequences. To get some idea of what the babies were listening to, try saying this
sequence out loud, making sure that you keep the interval between each syllable
the same: pa/ki/bu/ti/bo/du/go/la/tu/da/ro/pi/go/la/tu/pa/ki/bu.
The second phase of the experiment was the test phase. In the test phase,
having listened to the syllables during the familiarization phase, the babies were
presented with all the words that they had heard and also with part-words where
syllables from two words were recombined. For example the part-word, tudaro,
was formed from the last syllable of golatu and the first two syllables of daropi.
2 FIRST WORDS 71
n Johnson a n d J u s c z y k ’s s e c o n d s t u d y
Having shown that infants could detect and remember transitional probabilities in
their first study, Johnson and Jusczyk went on to investigate the role of syllable
stress in word boundary detection. Using a similar experimental paradigm, with a
familiarization phase followed by a test phase, they manipulated stress so that
every time a part-word appeared during the familiarization phase it was stressed
on the first syllable. There were three possible outcomes for the second test
phase.
. The infants might still spend longer listening to the part-words, as in the first
experiment. This would indicate that stress was not a strong cue; that the
infants had not made use of the stress to highlight part-words.
. The infants might listen equally to words and part-words. This would imply
that both cues were equally strong.
. The infants might spend longer listening to the words than the part-words,
that is, they would show the opposite response to the one observed in the
first experiment. This would suggest that the stress cues were very strong
and had overwhelmed the transitional probabilities, that the stress on the
first syllables had effectively labelled these as ‘words’ too.
72 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
t
Activity 3 Studies of word boundary detection in babies
Allow about This activity will help you to compare the results of Johnson and Jusczyk’s two studies and think about
10 minutes the implications for babies’ detection of word boundaries.
Figure 3 shows the results that Johnson and Jusczyk obtained in their second experiment.
Compare these results with those from their first experiment shown in Figure 2 which you
looked at earlier.
. How have the infants responded to the
9
words and part-words in the second Part-word
experiment? 8 Word
. How does this compare with their
7
responses in the first experiment?
Bruner’s theory
One theory that provides a useful framework for thinking about early word
learning was put forward by Bruner (1975, 1993). His key insight was that young
children encounter language in highly familiar social contexts because people
generally talk to them about familiar events and objects. So they hear speech that
relates to events, objects and people that they are already familiar with and that
are quite likely to be present or happening along with the speech.
To illustrate Bruner’s theory, consider the following specific example of the
way that early word meanings arise from a familiar social routine. The example
comes from the diary records I kept of my daughter, Francesca, as she began to
master her first words.
74 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Francesca’s understanding of ‘Are you ready?’ grew directly out of the routine that
took place at the end of nappy changing. It is clear from this example that, by 4
months of age, she had become sufficiently familiar with this routine to be able to
predict what came next. She was then able to associate the question ‘Are you
ready?’ with being raised into a sitting position and so to anticipate this event by
raising her own head from the changing-table when she heard the familiar
question. This instance of language comprehension, which arose directly in the
context of a highly familiar social routine involving the infant and adult, is an
example of what Bruner describes in his theory.
If first words grow out of familiar social contexts it should be possible to show
that there is a close relationship between what parents say to their children and
the events that are occurring at the time. The next activity gives you an
opportunity to examine this relationship in detail.
t
Activity 4 The social context of early language experience
Allow about This will help you to examine the way in which things that are said to young children often relate
30 minutes closely to what they are doing, and how this impacts on language development.
The following is part of a transcript made in a study by myself and two colleagues
(Harris, et al., 1983). It records a conversation between a mother and her
10-month-old daughter.
When the transcript begins, the child (who is just learning to walk) is standing holding onto
the arm of a chair with one hand. The mother is sitting on the floor next to her child.
The first column of the transcript shows the time (in minutes and seconds) at which a
particular utterance was made during the session. The second column records the
mother’s utterances and activity, and the third column describes what the child was
doing and saying. If you look at the child’s column you will see that she is not yet saying
any words but she is babbling (recorded as ‘vocalizes’ on the transcript). Where an utterance
is shown on the same line as an activity, they happened at the same time.
2 FIRST WORDS 75
Look at the mother’s utterances in the transcript (numbered 1–13) and work out what she
is talking about in each case. Then draw a table as shown in the illustration below the
transcript. In your table, write each of the mother’s utterances in the appropriate
numbered row in the second column. Then, in the third column write a brief description
of the topic of each of the mother’s utterances. Be sure to include the following information:
. Is the mother responding to something that has just happened?
. Is the mother describing something that the child is doing?
. Is the mother talking about something that she is doing?
There are two example answers shown in Figure 4 to help you.
Transcript of a conversation between a Mother (M) and her 10-month-old child (C).
Utterances are shown in bold type
Time Mother Child 10 months old
Comment
In our study we found that the majority of a mother’s utterances referred to objects or events
that her child was currently attending to. The extract that you have just analysed was typical of
what we found. All of the mother’s utterances refer to something that has just happened, to
something that the child is attending to, or the utterance actively directs the child’s attention
onto a particular object. The first two utterances occur immediately after the child has
touched a teddy on the floor. Then the child picks up the teddy and puts it next to a toy emu
(which was a glove puppet). The mother immediately responds by saying, ‘And that one as
well?’. She then picks up the emu, puts her arm inside it and says, ‘That’s a very strange looking
thing’. As she says this, she makes the emu peck the child.
You will see that most of the mother’s utterances follow on from something the child had
done immediately beforehand. However, as well as taking her lead from the child, the mother
also provides supporting actions of her own that serve to make her general meaning clear.
Notice that utterances 8–11 are all concerned with the child’s attempts to walk towards her
mother, hence they are focused on the child’s current actions rather than objects. The
conversation then returns to the topic of the toy animals.
s
Bruner himself does not give many specific examples of exactly how familiarity
with the social context might help young children to understand what adults are
saying to them. It is clear, however, that different kinds of words place different
demands on young children. Like the child’s attempts at walking in the activity
above, Francesca’s understanding of ‘Are you ready?’ was not focused on a
particular object since she was learning to associate a question (‘Are you ready?’)
with an action (lifting her head in anticipation of being pulled up). However, in
many other cases, a young child has to learn that a particular personal name (such
as ‘mummy’) is associated with the presence of a particular person, or that a
2 FIRST WORDS 77
Cues to reference
The development of an understanding of associations of this kind – between
sounds and personal names or objects – is only possible if a child has some way
of deciding what an adult is talking about. There are two specific cues that seem
to be significant. These are gaze direction and pointing. Gaze direction alone can
be difficult to interpret. However, when someone turns to look at a particular
object or person, they often turn their head. Together, gaze, head turning and
pointing can provide invaluable cues to reference, that is, what someone is
referring to when they use a particular word.
t
Activity 5 Cues to reference
Allow about This activity will help you to evaluate the different cues to reference that people use.
5 minutes
Imagine that you are listening to an unfamiliar language and have begun to recognize the
sounds of a few words. Now you want to know they mean. Think about how much you could
tell from where a speaker is looking or from where they are pointing. Which is the better
cue? Can you think of any other cues that might be used? Make a brief note of your answer
before reading further.
Comment
It turns out that pointing provides a more accurate cue to reference than either gaze or head
movement alone. This is true for both infants and adults. Adults are better at working out
what another person is referring to when the other person points than when they merely turn
to look at a particular object, and infants can locate an object more accurately when someone
points at it rather than merely looking at it (Butterworth, 1998). Another cue that people
sometimes use is to touch or pick up an object that is being named. Parents sometimes pick
up and animate objects that they are naming for children.
s
Pointing
Baldwin (1995) looked at the significance of pointing in the acquisition of new
vocabulary. She studied the amount of time that infants looked at a novel object
when an adult pointed to it and compared this with the time spent looking at an
object when there was no pointing. Infants, who were as young as 10 months of
age, spent significantly longer looking at objects when an adult pointed. When an
adult named an object, as well as pointing to it, the amount of looking was even
greater, suggesting that the young child is most predisposed to look at objects that
are singled out both through pointing and through naming.
However, there appears to be an even closer relationship between pointing
and first words than that illustrated by Baldwin. In a longitudinal study, Harris, et
al., (1995a) found that the age at which children first pointed – which was around
78 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
10 months – was highly correlated with the age at which they first showed signs
of understanding the names of objects. This close relationship between pointing
and the understanding of object names is well illustrated by another example
from Francesca. Again, the example comes from the detailed diary records that I
kept.
The first object name that Francesca understood was ‘nose’. She began by
touching the nose on a toy koala bear when asked and then, the following
day, she was asked ‘Where’s mummy’s nose?’ and ‘Where’s daddy’s nose?’.
Francesca reached out and touched her parents’ noses. This first occurred
when she was just over 9 months old. The very same day she pointed at a
plant in the conservatory. This was the first time that she pointed.
This close relationship in time between first referring to objects in the world by
pointing at them and first understanding an object name suggests that these are
closely interlinked processes that may well have a common origin. For example,
when sighted children read picture books with their care-givers, there is a shared
attention to individual objects and animals through pointing. This seems to be one
way in which infants acquire object names. However, since using pointing in this
way obviously depends on a child’s vision, it is relevant to ask what happens if an
infant is blind and this activity is not accessible in the same way. Blind children have
been found to produce significantly fewer words for discrete objects than sighted
infants (Norgate, 1997) and this supports the importance of pointing in acquiring
object names.
t
Reading
At this point you should turn to the end of this chapter and read Reading B, ‘Developments in
early lexical comprehension: a comparison of parental report and controlled testing’ by
Harris and Chasin (1999). This paper considers the accuracy of parental reports of early
comprehension and compares data collected using the MacArthur checklist with data
derived from experimental testing of comprehension. As you read this paper, think about
the factors that might contribute to the accuracy of parental reports about vocabulary.
s
2 FIRST WORDS 79
Figure 5 shows the average number of words understood by boys and girls
between 8 and 16 months of age as reported by their parents who completed the
MacArthur checklist. You can see two clear patterns. Girls are generally ahead of
boys in the number of words they can understand. However, for both boys and
girls, the overall pattern of development is essentially similar. The total number of
words comprehended grows fairly slowly up to about 12 months of age. Then
there is a sudden increase in vocabulary size.
300
Number of words understood (average score)
250
Female
200
Male
150
100
50
0
8 9 10 11 12 13 14 15 16
Age (months)
Figure 5 Average numbers of words understood by boys and girls between 8 and 16 months of
age (adapted from Fenson et al., 1994, p. 74).
t
Activity 6 Why do babies sound different from adults?
Allow about This will help you to compare the physical characteristics of the adult vocal tract with those of the
15 minutes infant vocal tract and reflect on how such differences might impact on the production of speech
sounds.
Soft
palate
Soft
palate
Tongue Soft
palate
Tongue
Epiglottis Tongue
Epiglottis
Vocal Epiglottis
Vocal cords Vocal cords
cords
Larynx Larynx Larynx
Figure 7 The anatomy of the vocal tract of chimpanzees and infant and adult humans.
Look at the three pictures in Figure 7 which show the vocal tracts of an adult chimpanzee, an
infant human and an adult human. Look carefully at the size and position of the tongue and
the position of the larynx (voice box) in each picture. Next, draw and fill in a table like the
one below and use the information you have gathered to decide whether the infant is more
like an adult human or more like a chimpanzee in the structure of its vocal tract. In order to
compare the position of the tongue and larynx you may find it helpful to draw a line from the
lips to the larynx and look at the angle that is created.
Comment
The infant human’s vocal tract is not simply a miniature version of the adult human tract and,
up to the age of 3 months, it actually resembles the vocal tract of a chimpanzee more closely
than that of an adult human. The infant’s larynx is positioned high up so that the epiglottis
nearly touches the soft palate at the back of the mouth. The baby’s tongue is large in relation
to the size of the mouth, nearly filling the oral cavity, while the pharynx is very short compared
to that of an adult, allowing little room for the back portion of the tongue to be manipulated.
s
The reason for the characteristic shape of the infant vocal tract at birth becomes clear
once you realize that the main purpose of the infant’s tongue in the first weeks after
birth is to enable the strong piston-like movements that are essential for sucking.
Once the first 4 months are over, and the baby has gained in weight, sucking
becomes less of a priority. At this point, the vocal tract gradually changes into a more
adult-like form so that infants can produce the complex range of movements that
will be required for speech. Changes in the anatomy of the vocal tract are
accompanied by neural maturation of the related motor areas in the brain. Together
these developments enable infants to develop increasing control over the fine motor
movements that are essential for producing the full range of speech sounds.
150
Number of words produced (average score)
125
100
Female
75
50
Male
25
0
8 9 10 11 12 13 14 15 16
Age (months)
Figure 8 Average numbers of words produced by boys and girls between 8 and 16 months of
age (adapted from Fenson et al., 1994, p. 75).
84 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
are taken from the MacArthur study (Fenson et al., 1994). Although parents are
often unsure about children’s early understanding of words, they can usually
provide very reliable information about the new words that their children are
producing. When collecting such data it helps to provide parents with clear
instructions about what counts as a word. This is important because young
children often simplify and shorten long words (e.g. saying nana for banana or
mama for mummy), and hence such utterances, although not strictly words, need
to be counted as such.
The MacArthur data show that most children produce their first word at around
10 months and gradually produce more words over the next few months. As with
comprehension, there is a sudden increase in the rate of learning new words. This
occurs around 13 months, and at this point girls are, on average, significantly
ahead of boys in the number of words they are able to produce.
You have just read about James’s use of mummy and there which were both
context-bound. However, it is notable that James also used words in a
contextually flexible way. He used teddy to refer to one particular teddy in a
variety of different contexts (for example, when he was sitting on a teddy and
when he was pointing to the teddy’s reflection in a mirror). He also used the word
more when he was reaching into a toybox for some bricks, taking another drink
from his cup or holding out an empty bowl. Another child in our study, a girl
called Madeleine, first used the word shoes in a range of situations including
looking at pictures of shoes in a book, pointing at her own shoes and also when
holding her doll’s shoes.
In our study (Harris et al., 1988), we looked at the first ten words produced by
four different children. Table 1 shows the number of context-bound and
contextually flexible words that the children produced.
You will see from the table that all the children produced words of both types
although there was variation from child to child in the relative number of each
type. Jenny and Madeleine produced more contextually flexible words, whereas
James and Jacqui both produced more context-bound words.
You will also see that several of the words in the table – the ones in bold – were
initially used as the names for objects. This is consistent with the view that
children do not suddenly develop a ‘naming insight’ but in fact use object names
right from the beginning. Data from a study by Goldfield and Reznick (1990)
support this view. In a longitudinal study of 24 children, starting from the age of
just over 1 year, they found that half of the early words used by the children
before the vocabulary spurt were object names.
2 FIRST WORDS 87
However, this same study also found that a small number of children showed a
more gradual rate of lexical growth. These children seemed to be acquiring a
broader range of different word types, while most of the children, who were
acquiring mainly object words in their early language development, were more
likely to show the rapid increase in rate of learning. This suggests that there may
be different pathways in early language development, but a common feature for
many children is a period when learning names for things is a dominant strategy.
Issues relating to individual variability and their links to the vocabulary spurt will
be discussed further in Section 6.
First words
Earlier in this chapter you read about Bruner’s theory and saw that children’s early
experience of language occurs within the context of familiar social routines. Our
study explored the extent to which children’s first words grow directly out of their
experience (Harris et al., 1988).
You saw the first words that were produced by the children in our study in
Table 1. We had two ways of finding out about how these first words were used.
We asked the parents to keep a record of all the occasions when a child used a
word. We also filmed the children and their mothers talking and playing together
in a laboratory. These two sources of information allowed us to maintain a very
accurate record of the children’s first words so that we could see precisely how
they were using them.
We made recordings of the mothers and children in the laboratory every 2
weeks. This meant that we also had a detailed sample of the mother’s utterances
over the course of the study. We started filming when the children were 6 months
old and carried on until they were 2 years old. As soon as we were sure that a
child was using a word, we looked at the mother’s utterances that we had
recorded in the previous month. We identified all the mother’s uses of that word.
Then we found out how many times she had used the word and in how many
different ways she had used it.
We found that the children’s use of their first words bore a very close
resemblance to the mother’s use of these same words. Of the 40 words that we
studied (shown in Table 1) there were only three cases where there was no
apparent relationship between a child’s use and the mother’s use of that word in
the preceding month. Furthermore, for 33 out of the 40 words, the child’s use was
identical to the mother’s most frequent use. For example, the first word used by
James was Mummy but, rather puzzlingly, he initially restricted its use to
situations where he was holding out a toy for his mother to take. This apparently
idiosyncratic use of Mummy was explained when we found that his mother had
most commonly used this word when holding out her hand to take a toy saying,
‘Is that for Mummy?’.
88 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Table 2 gives some other examples of the children’s first words and their
mothers’ use of the same word.
You can see from these examples that hello was rather like Mummy in that the
child had a single use that corresponded exactly to the mother’s use. As in the
case of James’s use of Mummy, the corresponding mother’s use was her most
frequent use of hello when talking to her child. The other two examples are more
complex because the first use of these words is contextually flexible. Teddy is an
object name that the child first used only to refer to one particular teddy. (We
know that it was an object name rather than a personal name because the child
rapidly used it to refer to other teddies.) The mother’s use was similar because she
also used teddy to refer to a particular big teddy. No was used by Jenny in a range
of situations where an anticipated action was not going to be carried out.
Interestingly Jenny used no both to refuse an action that was requested by her
mother and also in connection with her own action. Again you can see that the
mother’s own use of this word covered a similar range.
Table 2 Examples of four children’s first words and their mothers’ use
of the same word
a precedent in the mother’s own use dropped to 45 per cent. This compared to
the figure of 93 per cent that we had found for initial uses of the first ten words.
A study by Hart (1991) also shows that the importance of direct experience
becomes much less important as children acquire more words. Hart compared
early vocabulary with later vocabulary to see whether there was a difference in
the frequency with which children heard the words that they acquire early on.
Hart found that there was a difference. Children’s first words tended to be the
ones that their parents frequently used when talking to them – on average these
words occurred 30 times in a monthly observation session. However, when the
children were 6 months older, the words that they were acquiring had typically
been used only twice in parental speech during an observation session.
Some effects of language experience do, however, appear to be of longer
duration. Children learning English have a predominance of object names in their
early vocabulary and relatively few verbs. In contrast, Gopnik and Choi (1995)
found that verbs appear earlier, and form a greater proportion of early vocabulary
in the speech of children acquiring Korean, while Tardif (1996) found that 21-
month-old children who were learning Mandarin Chinese had as many different
verbs as nouns in their vocabulary. These patterns reflect the fact that Korean-
and Mandarin-speaking mothers use relatively more verbs than English-speaking
mothers when they are talking to young children. It is thought that this is a result
of the difference in structure between Korean and Mandarin (both of which
emphasize verbs), and English (which emphasizes nouns).
t
Activity 6 Differences associated with language delay
Allow about This will help you to think about possible reasons why mothers’ speech to their children with language
5 minutes delay is different.
Read the four paragraphs above on language experience and language delay again. Note
down at least two contrasting explanations for the differences in the speech of mother’s of
children with language delay, as compared with mothers of children with typical language
development.
Comment
There are at least two possible explanations for the relationship between the mothers’ speech
and the children’s language.
One possibility is that the differences in the mothers’ speech in the two groups arose because
of differences in the language ability between the groups. In other words, the two groups of
mothers might have been talking differently because the two groups of children responded in
different ways to what they were saying.
Another possible explanation is that these differences in maternal speech were responsible, at
least in part, for the differences in the children’s language ability.
s
There is good reason to prefer the second of the explanations offered above.
When we sampled the speech of the two sets of mothers, both groups of children,
who were then 16 months old, were producing similar speech. Evidence of
differences among the children did not appear until several months later. At 2
years of age, the slower developers were still producing single words but the
typical developers were producing utterances that were several words long. The
groups also differed in their vocabulary size, with the typical developers
producing significantly more words than the slower developers. Because these
differences between the groups did not emerge until several months after the
mothers’ speech was sampled, it seems unlikely that the speech style of the
mothers was influenced by the speech of their children, but rather that the
maternal speech style was influencing the language development. These findings
can therefore be seen as evidence that the close tying of maternal speech to the
current social context is an important factor in early language development.
However, the issue of the causes and effects of language delay is complex. For
example, it is possible that the 16-month measures of children’s speech did not
capture some aspects of the children’s speech that were differentially affecting the
mothers’ behaviour. It is also possible that the children with language delay were
less able to make use of their mothers’ speech to them. What we can be sure of is
that there is an interactive relationship between mothers’ speech and their
children’s language development.
2 FIRST WORDS 91
300
+ 1 SD
Female
200
Male
150
_ 1 SD
100
50
0
8 9 10 11 12 13 14 15 16
Age (months)
Figure 9 Amount of variation in the numbers of words understood between the ages of 8 and
16 months by boys and girls (Fenson et al., 1994, pp. 74–5).
150
Number of words produced (average score)
+ 1 SD
125
100
Female
75
50
Male
25
_ 1 SD
0
8 9 10 11 12 13 14 15 16
Age (months)
Figure 10 Amount of variation in the numbers of words produced between the ages of 8 and
16 months by boys and girls (Fenson et al., 1994, pp. 74–5).
Although not all children will be producing words by 18 months of age, most will
comprehend a considerable number of words at this age. This suggests that the
first thing to look for when a child is not yet talking is vocabulary comprehension.
If comprehension is poor relative to the norms, this could well be a sign of
language delay. Another indication that can be used in younger children is
babbling. The stages of babbling that you read about in Section 4 are fairly similar
2 FIRST WORDS 93
7 Conclusion
At the beginning of this chapter, I outlined a number of essential steps that are
necessary for children to understand and produce their first words. You have now
learned about each stage in this complex process and I want to end by asking you
to reflect on the wide range of skills that come together when young children
begin to build up a vocabulary.
Look again at the chart of component skills that you saw in Section 1.
When you consider all the skills that come together in learning first words, you
can begin to appreciate what an extraordinary achievement it is when children
first comprehend and produce words. And that, as you will see in Chapter 4, is
only the beginning of a learning process that continues for many years until
children become fully competent in their native language.
2 FIRST WORDS 95
References
Baldwin, D. A. (1995) ‘Understanding the link between joint attention and
language’, in Moore, C. and Dunham, P. J. (eds) Joint Attention: its origins and
role in development, pp. 131–58, Hillsdale, NJ, Erlbaum.
Barrett, M. D. (1986) ‘Early semantic representations and early word usage’, in
Kuczaj, S. A. and Barrett, M. D. (eds) The Development of Word Meaning, pp. 313–
34, New York, Springer-Verlag.
Barrett, M. D., Harris, M. and Chasin, J. (1991) ‘Early lexical development and
maternal speech: A comparison of children’s initial and subsequent uses of
words’, Journal of Child Language, vol. 18, pp. 21–40.
Bates, E., Bretherton, I., and Snyder, L. (1988) From First Words to Grammar:
individual differences and dissociable mechanisms, Cambridge, Cambridge
University Press.
Bruner, J. S. (1975) ‘The ontogenesis of speech acts’, Journal of Child Language,
vol. 2, pp. 1–19.
Bruner, J. S. (1993) Child’s Talk: learning to use language, New York, Norton.
Butterworth, G. (1998) ‘What is special about pointing in babies?’, in Simion, F.
and Butterworth, G. (eds) The Development of Sensory, Motor and Cognitive
Capacities in Early Infancy, pp. 63–88, Hove, Psychology Press.
Christophe, A. and Morton, J. (1998) ‘Is Dutch native English? Linguistic analysis
by 2-month-olds’, Developmental Science, vol. 1, pp. 215–19.
DeCasper, A. J. and Fifer, W. (1980) ‘Of human bonding: newborns prefer their
mothers’ voices’, Science, vol. 208, pp. 1174–6.
DeCasper, A. J. and Spence, M. J. (1986) ‘Prenatal maternal speech influences
newborns’ perception of speech sounds’, Infant Behavior and Development, vol.
9, pp. 133–50.
DeCasper, A. J., Lecanuet, J.-P., Busnel, M.-C., Granier-Deferre, C. and Maugeais,
R. (1994) ‘Fetal reactions to recurrent maternal speech’, Infant Behavior and
Development, vol. 17, pp. 159–64.
Dore, J. (1978) ‘Conditions for the acquisition of speech acts’, in Markova, L. (ed.)
The Social Context of Language, pp. 87–111, New York, John Wiley.
Elman, J., Bates, E., Johnson, M., Karmiloff-Smith, A., Parisi, D. and Plunkett, K.
(1996) Rethinking Innateness: a connectionist perspective on development,
Cambridge, MA, MIT Press.
Fenson, L., Dale, P., Resnick, S., Bates, E., Thal, D. and Pethick, S. J. (1994)
‘Variability in early communicative development’, Monographs of the Society for
Research in Child Development, vol. 59, pp. 1–73.
Goldfield, B. A. and Reznick, J. S. (1990) ‘Early lexical acquisition: rate, content,
and the vocabulary spurt’, Journal of Child Language, vol. 17, pp. 171–83.
96 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Readings
Reading A: Is Dutch native English? Linguistic analysis
by 2-month-olds
Anne Christophe and John Morton
Abstract
A variant of the non-nutritive habituation/dishabituation sucking method was
used to test 2-month-old English infants’ perception of languages. This method
tests for the spontaneous interest of the baby to a change in the stimulus. English
and Japanese were clearly discriminated. The difference between French and
Japanese was equally clearly not of interest to babies using this procedure, the
babies behaving as though both languages were classified simply as ‘foreign’. In
order to further specify babies’ representation of native and foreign language, we
used Dutch, which shares a number of suprasegmental features with English. The
results from our last 2 experiments indicate that a portion of our 6–12 week-old
babies consider Dutch as native, suggesting that we tapped in a transition period
where the babies are still refining the suprasegmental specification of their native
language.
One of the most important tasks for a new-born infant is to learn its native
language. The majority of babies grow up in a multi-lingual environment and
must learn some characteristics of their mother tongue as early as possible so as to
distinguish it from other languages. This is a particularly crucial ability, since
infants could not possibly learn the syntax of a language (that is, discover the
regularities shared by a number of sentences) if they worked on a database
containing sentences from several different languages (Mehler et al., 1994).
It has been shown that newborns can discriminate between their mother tongue
and a foreign language. Mehler, Jusczyk, Lambertz et al. (1988) found that 4-day-
old French infants discriminate between French (their mother tongue) and
Russian stimuli, showing a preference for their native language (see also Moon et
al. 1993, for equivalent results with English and Spanish). In addition, newborns
are able to discriminate between utterances in two foreign and unfamiliar
languages (Mehler and Christophe, 1995; Nazzi, Bertoncini, and Mehler, 1998).
Most of these studies have been replicated successfully using speech which has
been low-pass filtered at 400 Hz. Under these conditions, prosodic features such
as intonation and rhythm are preserved, whereas most phonemic information is
missing. It is therefore probable that babies’ ability to discriminate between
languages is based on a representation of speech prosody. It is very likely that the
infant’s preference for their native language comes from their having learned its
prosody in utero. However, we still do not know the precise nature of the
prosodic representation that babies use to classify languages.
Babies of 2 months of age behave slightly differently from newborns. They still
discriminate between their native language and other languages but they fail to
show any recovery of interest when switched from one foreign language to
another. Thus, Mehler et al. (1988) showed that while 2-month-old American
2 FIRST WORDS 99
babies were able to discriminate between English (their mother tongue) and
Italian, they did not discriminate between French and Russian. A possible
interpretation of this counter-intuitive result is that, while newborns still attempt
to analyse in detail any speech sample they are exposed to, 2-month-old infants
have sufficient knowledge of their mother tongue to be able to filter out any
foreign language as being not relevant.
Hesketh, Christophe and Dehaene-Lambertz (1997) developed a variant of the
contingent sucking response method which has the advantage that it can be used
both with new-borns and with 2-month-old infants and can be used with
extended segments of speech. With this technique, 2-month-old English babies
distinguished clearly between English and Japanese. It is this technique which we
used to explore the infants’ abilities further.
Method
The method for the Hesketh et al. experiment will first be briefly described (see
Hesketh et al., 1997, for details). The other experiments to be reported used the
same technique apart from changes in language. The stimuli consisted of 80
sentences, half in English, half in Japanese, between 15 and 21 syllables long.
These were recorded by four female native English speakers and four female
native Japanese speakers respectively. Speakers were naı̈ve as to the aim of the
experiment and were instructed to read as naturally as possible. Ten sentences
from each speaker were selected and matched for syllabic length (17.8 syllables)
and duration (3.1 seconds). Each infant underwent two changes in stimulation,
one experimental (language) change, the other control (or speaker) change. The
key measure was the difference between these two changes.
Half the babies received the experimental change first and the control change
second. In addition, the order of presentation of languages and of speakers was
counter-balanced across subjects. This yielded eight conditions. In each of the
three phases the baby heard sentences from two speakers with the idea of making
speaker change mundane.
Subjects were seated in a car seat placed in a sound-proofed chamber and offered
a standard (steam sterilised) pacifier. One experimenter, out of view of the baby
and deaf to the stimuli, checked that the pacifier stayed in the baby’s mouth
throughout the experiment. A second experimenter monitored the experiment on
the computer outside the chamber. The computer recorded the pressure of the
infant’s sucks via an analogue-digital card (NIDAQ), detected the sucking
responses and delivered the sentences through a ProAudio 16 sound board
according to the reinforcement schedule (see below). The computer also saved
both the moment and amplitude of each suck as well as the stimuli triggered by
the sucks. Hesketh et al. (1997) reported that the number of sentences triggered
was a cleaner measure than the number of sucking responses. Only this measure
will be reported here.
The experiment started with a short period without stimulation (about 30 secs) to
settle the infants. The first phase of the experiment then began, during which
infants heard sentences in either English or Japanese contingently upon their
high-amplitude (HA) sucks. After a short ‘shaping’ phase, three HA sucks were
required to trigger each sentence (such that there was less than one second
100 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
between two consecutive sucks). There was an ISI of at least 600 ms between
consecutive sentences. When reaching the end of an ISI period after presentation
of one sentence, the program looked back to see if HA sucks had occurred
recently: any sequence of three HA sucks such that the last one occurred within
the last 600 ms was used to instantly trigger a new sentence. This procedure
ensured fluent presentation of sentences in case of sustained sucking activity.
Within each phase of the experiment, the order of presentation of the sentences
was quasi-random for each baby.
A switch in stimulation occurred after a predefined habituation criterion had been
met. For two consecutive minutes the infant’s HA sucking rate had to be less than
80% of the maximum sucking rate from the beginning of the experiment. Each
phase of the experiment lasted at least 5 full minutes. Sixteen babies aged
between 6–12 weeks participated in the study, mean age 8 weeks 6 days. Subjects
were randomly assigned to one of the eight conditions prior to testing.
To assess the effect of the experimental manipulation, two kinds of analyses were
performed on the data: ANOVAs and non-parametric tests. For each baby we
counted the number of sentences triggered during the two minutes before and
after the experimental (language) switch. The difference between these two
values gives us a measure of dishabituation to the language shift. The equivalent
measure was computed for the control (speaker) switch. The difference between
these two dishabituation scores represents a discrimination index for each baby:
whenever this value is positive, the baby reacted more to the language change
than to the speaker change. These values are shown in Figure 1 (left hand
column).
A Wilcoxon signed ranks test showed that the median of the discrimination index
for the number of sentences triggered was significantly above zero (Z = 3.4,
p < 0.001). In the ANOVAs, the dependent measure was the dishabituation scores
for the Experimental and Control switches. There was one within-subject factor
(Experimental vs Control switch) and two between-subject counterbalancing
factors, Order (experimental switch first, versus control switch first), and
Language (English first vs Japanese first). There was a main effect of the
Experimental factor (F (1,12) = 11.6, p < 0.01), no significant effect of any of the
counterbalancing factors, and no interactions between the Experimental and
counterbalancing factors.
The distribution of the discrimination index can be seen in the second column of
Figure 1. The infants gave no indication of being more interested in language
change than in speaker change (Wilcoxon, Z<1; ANOVA: F (1,12) <1). This result
is significantly different from the results of the experiment with English and
Japanese. In an ANOVA contrasting the distribution of discrimination indices in
these experiments, F (1,28) = 8.32, (p < 0.01).
The lack of interest shown by 2-month-olds in the differences between foreign
languages is in line with previous work. Paired with Nazzi et al.’s (1998)
confirmation that newborn infants can discriminate between sentences belonging
to two foreign languages (with the same experimental technique as here), this
result confirms the developmental trend already described. Our best
interpretation is that 2-month-old infants have enough knowledge of the
properties of their native language to be able to filter out foreign input as being
irrelevant to their language learning. In that case, both French and Japanese
would simply be classified as ‘foreign’ and would not be analysed to a sufficient
depth to allow the differences to be detected.
The results of the first two experiments immediately pose a new question: how
specified is the 2-month-olds’ representation of their mother-tongue? What do
they consider native, and what do they filter out as being foreign? To answer this
question, we picked a language which shares with English a number of prosodic
properties. Dutch, like English, has vowel reduction, complex syllabic structure,
and the same sort of word stress as English. These factors lead to both English and
Dutch as being heard as stress-timed (Cutler et al., 1997). In fact, Dutch and
English have already been shown to be rather similar to babies’ ears: Nazzi et al.
(1998) demonstrated that French newborns do not distinguish between Dutch
and English filtered sentences.
Discussion
Using the modified contingent sucking response we have shown that English
2-month-olds discriminate English from Japanese but not French from Japanese.
Given that this habituation–dishabituation technique measures infants’ interest in
changes in auditory stimulation, it allows us to evaluate their spontaneous
partitioning of perceptual space into categories. In the present case, the results
suggest that babies form two major categories, one for English, which could be
termed ‘native’ or ‘mother tongue’, and one, with French and Japanese, of ‘foreign
languages’.
In the last 2 experiments of this paper, we studied English infants’ perception of
Dutch, a language that is prosodically very similar to English. We contrasted
Dutch to English and to Japanese. If English babies treat Dutch as native, they
should not be able to discriminate between English and Dutch, but should readily
distinguish Dutch from Japanese; in contrast, if they have already set up their
native category such that Dutch is excluded, they should distinguish between
Dutch and English but ignore the difference between Dutch and Japanese, both
of which would be in the category foreign. Both experiments gave marginally
significant results, indicating that some English babies consider Dutch as native
but others do not. The former would distinguish Dutch from Japanese but not
from English; the latter group would distinguish Dutch from English but not from
Japanese. If we tested babies in both conditions, we predict that whenever one
condition works the other would not. What factors may account for this individual
variation? The most obvious candidate is age. At one month, all infants might
104 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
regard Dutch as native, whereas by four months they might all have excluded it.
Further, we would expect early exposure to different languages to affect the
speed of setting up a tight specification of native but it is unlikely to be a factor in
our experiments since we selected the babies to come from monolingual English
households.
Eventually we will need to distinguish between environments where second
languages are addressed to the infant from those where second languages are
present but not directly addressed. The second case might accelerate the
definition of native whereas the first case, true bilingualism, might lead to
confusion. Recent experiments by Bosch and Sebastian (1997) showed that by
four months of age, bilingual Spanish/Catalan babies already behaved differently
from monolingual babies (either Spanish or Catalan). Monolingual babies
oriented faster to their mother tongue than to English. In contrast, bilinguals
orient to Spanish or to Catalan significantly more slowly than to English. Is this
because of confusion? Apparently not, since in more recent and still unpublished
work, these authors showed that, although Spanish and Catalan are close, both
monolingual and bilingual 4-month-olds can discriminate between them. Of
course the gap between these 4-month-olds and our 2-month-olds is enormous
and it could be that at 2 months bilingual babies are confused. At any rate, it has
become clear that, from birth, infants work hard at learning what language is
native.
Acknowledgements
The work reported in this paper was assisted by a grant from the Human Frontiers
Science Programme, the Human Capital and Mobility Programme, and the
European Science Foundation. We especially want to thank Sarah Hesketh, Jon
Bartrip and Sarah Minister for their help in recruiting and testing subjects.
Mehler, J., & Christophe, A. (1995). Maturation and learning of language in the
first year of life. In M. S. Gazzaniga (Ed.), The Cognitive Neurosciences: A
handbook for the field (pp. 943–954). Cambridge, Mass.: MIT Press.
Mehler, J., Dehaene-Lambertz, G., Dupoux, E., & Nazzi, T. (1994). Coping with
linguistic diversity: The infant‘s viewpoint. In J. L. Morgan & K. Demuth (Eds.),
Signal to Syntax: Bootstrapping from speech to grammar in early acquisition (pp.
101–116). Mahwah, New Jersey: LEA.
Mehler, L., Jusczyk, P. W., Lambertz, G., Halsted, G., Bertoncini, J., & Amiel-Tison,
C. (1988). A precursor of language acquisition in young infants. Cognition, 29 ,
143–178.
Moon, C., Cooper, R., & Fifer, W. (1993). Two-day-olds prefer their native
language. Infant Behavior and Development, 16, 495–500.
Nazzi, T., Bertoncini, J., & Mehler, J. (1998). Language discrimination by
newborns: Towards an understanding of the role of rhythm. Journal of
Experimental Psychology: Human Perception and Performance, 24, 1–11.
Source: Christophe, A. and Morton, J. (1998) ‘Is Dutch native English?
Linguistic analysis by 2-month-olds’, Developmental Science, vol. 1,
pp. 215–19.
106 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Abstract
Six children were studied from the age of 0;6 to 1;6 in order to chart their
developing comprehension vocabularies from the first to the 100th word.
Observational data were used in the first instance to identify newly
comprehended words and then controlled testing was carried out for each word to
confirm and expand the observational data. Comprehension of words was
divided into four categories – object names, context-bound object words, action
words and personal names. The relative frequency of the different categories of
word was found to change with the size of the comprehension vocabulary as
personal names decreased in importance and both object names and action
words became increasingly more common. There was considerable variation
among the six children especially in the proportion of object names and action
words that they understood but vocabulary composition became highly stable
between 60 and 100 words.
Introduction
This paper compares data on early lexical comprehension derived from parental
report with that from systematic experimental testing. A major source of data on
the composition of early comprehension vocabularies comes from Fenson, Dale,
Reznick, Bates, Thal & Pethick (1994) who describe an extensive sample of
children whose vocabulary was assessed with the MacArthur Communicative
Development Inventories (Infant Scale). These data are cross-sectional rather than
longitudinal but they do provide important evidence about the first 50 words that
children understand.
Pooling data across subjects, Fenson et al. found that the words understood by
the youngest children (aged 0;8) were the names of people or were related to
games or routines. At 50 words, the main category was nouns (comprising
household items, animal names, toys, clothing, food and drink, body parts,
furniture and rooms) which accounted for 48% of items. The other categories
were games and routines (20%), actions words (16%), personal names (10%) and
sounds (6%). By 1;4 – when mean comprehension vocabularies were reported as
192 words – 52% of words understood were nouns, 19% were verbs, 9% were
words stemming from games and routines and 3% were personal names. (For a
complete list see Fenson et al. (1994) table 16.)
The use of the MacArthur Communicative Development Inventory for the
assessment of early comprehension vocabulary has been questioned since it relies
exclusively on parental report. Although there is a long tradition of using parental
report for reliable assessment of production it is less clear that this can provide
equally reliable assessment of early comprehension since it is often difficult to
determine from observation alone whether a child understands a word or is,
2 FIRST WORDS 107
Method
Participants
Six children took part in the study, four boys (Ben, Andrew, Sebastian and
George) and two girls (Katherine and Katy). All the children were first born and
English was the only language spoken in the home. At the time of the first
observation the children were 0;6 (range 0;5.24–0;6.06). Observation continued
until the children were 2;0 although all children had attained a comprehension
vocabulary of at least 100 words by 1;6. Data for the early production and
comprehension of these children have been reported in Harris, Yeeles, Chasin &
Oakley (1995a) and Harris, Barlow-Brown & Chasin (1995b).
Procedure
Assessment of comprehension
Full details of the procedure used to assess comprehension can be found in Harris
et al. (1995a). Briefly, three different sources of evidence were used in the first
instance. These were parental diary records, home observation (supplemented by
videotaping) and a comprehension checklist which contained the most
commonly understood words organized into categories (e.g. toys, food and drink,
people, games, actions). The checklist was a modified version of the one used by
Benedict (1977).
108 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Personal name Unique name for people, family Lamby – toy lamb used as
pets, favourite toys comforter
Dylan – family cat
Object name Corresponding to Nelson’s (1973) Cat – family cat, novel picture of
‘general nominal’ category; only cats
including words that were Nose – teddy’s nose, own nose,
understood in at mother’s nose
least two different behavioural
contexts
Context-bound Object words that were Bird – when indoors, looks out of
object word understood in only one window to garden
behavioural context Car – waves on hearing word or
sound of car
Action word All words or phrases that were Down – squats down on haunches
associated with actions rather Lunch – goes to kitchen and
than with objects attempts to climb into high chair
Assessment of production
The development of production was monitored through diary records, maternal
interviews, home observations and video recordings. Details of the procedure are
set out in Harris et al. (1995a). Briefly, a vocalization was counted as a word if it
was reported in the diary record and observed either during home observation or
in a videorecording. If there was no maternal report, three observations were
required before a vocalization was counted. Unlike comprehension, controlled
testing was only carried out if there was some ambiguity about the range of
contexts in which the child produced a word as, for example, where a diary entry
and an observation were not identical.
2 FIRST WORDS 109
Results
The proportion of words comprehended in each of the four categories was
calculated for vocabulary sizes of 20, 60 and 100 words. The proportion of words
in each category was found to change as the size of comprehension vocabulary
increased (see Table 2). At the 20-word level, the number of personal names and
object names was almost equal and together they accounted for two thirds of the
total. The remaining two categories – context-bound object words and action
words – each made up about one sixth of the total. At 60 words, the proportion of
object names and action words had both increased while there was a marked
decrease in the proportion of personal names. By the 100-word level the
proportion of personal names had decreased even further and there was a
corresponding rise in the proportion of object names. The proportion of action
words remained the same as at the 60-word level as did the proportion of context-
bound object words which was identical at all three points.
Table 2 Mean percentage (and range) of words in each category in relation to size
of comprehension vocabulary
Individual data for the six children reflected the overall pattern. The number of
object names understood by each child generally increased with the size of their
comprehension vocabulary. However there was considerable individual variation
in the number of different categories of word understood at each stage. A
110 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Discussion
Our data support the finding of earlier studies that the proportion of different
classes of words comprehended changes with vocabulary size (Benedict, 1977;
Bates, Bretherton & Snyder, 1988; Gunzi, 1993; Fenson et al., 1994). Personal
names figured prominently in early vocabulary but they made a relatively smaller
contribution as the total number of words understood by the children increased.
This decrease in the importance of personal names occurred as the proportion of
both object names and action words increased. The proportion of context-bound
object words remained stable throughout the period of development.
At 60 words, the mean proportion of object names was 40%. This is very
comparable to Benedict’s (1977) data for 50 words but somewhat lower than the
48% reported by Fenson et al. (1994). The mean proportion of action words at the
same vocabulary size was 27% which was considerably lower than the
proportions reported by both Benedict and Gunzi for 50 words but very similar to
the total proportion reported by Fenson et al. for action words plus words related
to games and routines (both of which were classified as action words in the
present study). Our data suggest that some early words that appear to be object
names are, in fact, context-bound object words and that a parental checklist may,
therefore, over-represent the number of object names. There is also some
suggestion from our data that parents over-estimate the number of words that
children understand since even the most precocious child that we tested attained
a score that was under the 75th percentile on the Fenson et al. norms; and there
2 FIRST WORDS 111
Contents
Learning o u t c o m e s
After you have studied this chapter you should be able to:
1 describe the developmental origins of specialist areas of the cerebral cortex;
2 evaluate arguments for and against the notion of innate modularity;
3 discuss the case for and against innate specification of cortical function;
4 identify some special functions of the prefrontal cortex in children’s
cognitive development;
5 outline some aspects of the relationship between brain development and
cognitive and language development.
6 In addition you should have:
. enhanced your understanding of principles of self-organization;
. further developed your understanding of plasticity;
. reinforced your understanding of nativism and constructivism;
. extended your knowledge of research methods.
1 Introduction
This chapter is about the development of the human brain in infancy, and about
how changes in the physical structure of the brain can be related to children’s
cognitive and language development. In the course of this chapter we will
address the following questions.
. What is the basic course of pre- and postnatal brain development?
(Section 2)
. How do different areas of the brain come to perform different functions?
(Sections 3 and 4)
. Are different areas of the brain innately pre-specified to take on their
respective cognitive functions? (Sections 3 and 4)
. How can cognitive function be related to the development of brain
structure? (Section 5)
. How much can be learned about cognitive development from studying
brain development? (Sections 4 and 5)
. How is it possible to study brain function? (All sections)
We start in Section 2 by taking a brief look at some basic features of brain
development in the child before birth and in the first few months after birth. A key
idea in this section is that some brain development goes on postnatally,
maximizing the opportunities for brain and cognitive development to be
influenced by both genes and environment, in interaction with each other.
116 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Cerebral cortex In Section 3 we examine the idea that different areas of the cerebral cortex (see
The layer of cells
on the outer Figure 1, p. 118) specialize in performing different cognitive functions. One way
surface of the of thinking about this ‘functional specialization’ within the human brain is in
forebrain; only terms of cognitive modules. Cognitive modules are hypothetical constructs that
found in
mammals, and
help to map what we know about brain function (what the brain does) to brain
particularly well structure (its physical make-up). Section 3 considers the debate about how the
developed in cognitive functions performed by such modules might develop. Do they unfold
humans.
from a predefined genetic blueprint, or are they dependent upon the child’s
interactions with the environment? We conclude that these cognitive functions
develop according to epigenetic principles, with genes and environment
inextricably linked. Section 4 considers this argument in detail in relation to
language development.
Epigenesis In Section 5 we look at how psychologists relate knowledge about children’s
Development cognitive development to knowledge about the developing structure of the
by means of
interaction
cortex. This section focuses on the prefrontal cortex (see Figure 1, p. 118) because
between genes of its importance within the field of developmental psychology. The prefrontal
and their cortex has received considerable attention from psychologists because of its
environment.
involvement in higher mental functions, such as the planning and initiation of
actions, and the inhibition of irrelevant behaviour (so called ‘executive functions’;
see Chapter 5).
The chapter as a whole is about the relatively new field of ‘developmental
cognitive neuroscience’ (Johnson, 1997). This interdisciplinary field has emerged
partly as a result of new and improved ways of investigating brain structure and
Neuroimaging
The use of various function (such as advances in neuroimaging) which have led to significant
technologies for progress in understanding how brains are constructed during development.
the non-invasive These methods will be described and illustrated throughout the chapter. Their
measurement of
brain activity,
importance for the psychological study of child development is that they allow us
aimed at to investigate the relationship between the development of brain structure on the
specifying the one hand and cognitive and behavioural development on the other.
functions of
different brain
regions. Also 1.1 Innateness and epigenesis
referred to as
‘brain mapping’. An important aim of this chapter is to address the question of whether the
development of brain structure (its physical make-up) and function (what it does)
are innately pre-specified, or whether the brain develops in response to the
environment of the child. It is worth spending a few minutes thinking about what
‘innate’ means before we begin the discussion proper.
The term ‘innate’ is rather ill defined. Sometimes it is used interchangeably with
the phrase ‘genetically determined’ to refer to an aspect of development being
exclusively determined by genes. In this sense of the term it refers to ‘nothing that
exists in the natural world, except for genes themselves’ (Johnson, 1997, p. 8).
Genes always interact with their environment, so they can never be the sole
determining factor in development. At the most basic level, the way that genes
operate is affected by their local molecular and biochemical environment. There
is no aspect of development that is purely genetically determined, so ‘innateness’,
in this strong sense, is not a useful term.
3 BRAIN AND COGNITIVE DEVELOPMENT 117
t
Activity 1 Thinking about innateness
Allow about This activity is a gentle introduction to some of the complex issues you will encounter concerning
10 minutes genetic and environmental influences on child development.
Think of ways in which someone you know well is psychologically similar to their parents.
You might focus on personality traits, skills, interests, habits and so on. Write down a list of
two or three examples.
For each example, think about the following questions and make brief notes.
. What factors have caused this person to be like this? Has this trait been developed
through experience, or has it been inherited from their parents?
. Was it inevitable that this person would develop in this way, or could other choices or
experiences have produced different outcomes?
. At what age was this similarity first evident?
118 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Comment
At the very least this activity should highlight some of the complexities (perhaps the
impossibility) of trying to separate ‘genetic’ and ‘environmental’ influences on human
development. If children grow up to be like their parents in some ways, is this because they
learned to be like this from them (an ‘external’ factor) or because they inherited these
similarities from them (an ‘intrinsic’ factor)?
s
Dorsolateral Parietal
region lobe
Prefrontal
cortex
Visual
cortex
Temporal
lobe
The sequence of events that goes towards building human brains is very similar to
that observed in other mammals, but the timescale over which these events occur
is significantly more extended. This slower timescale has two major
consequences. First, there is a prolonged period of postnatal development during
which the later stages of brain development can be influenced by interaction with
the outside environment. If the timescale of human brain development were
quicker, and babies were born with more mature brains, then much more of their
brain development would take place in the relatively limited environment of the
mother’s womb.
Second, the more delayed the general time course of the development of a
species, the larger the relative volume of the later developing areas of the brain.
In humans, the slowed rate of development is associated with a relatively larger
volume of cerebral cortex, and an especially large prefrontal cortex (see Figure 1).
The prefrontal cortex is considered by most investigators to be critical for many
high-level cognitive abilities (Milner, 1982; Goldman-Rakic, 1987; Fuster, 1989)
which you can read about in Section 5, and in Chapter 5 of this book. It is
significant that the region of brain that undergoes most postnatal development,
in interaction with the rich, external environment, is the region most closely
associated with high-level cognitive abilities such as the planning and execution
of complex sequences of behaviour.
t
Activity 2 The prenatal environment
Allow about This activity explores what psychologists mean by the term ‘environment’, and emphasizes that
10 minutes environmental influences on child development operate before birth.
Imagine for a minute the prenatal environment of the child (the mother’s womb). In what
ways do you think it is similar to the child’s postnatal environment? In what ways is it
different? Write a list of characteristics under the headings as suggested below. See Figure 2
for an example to start you off.
Comment
‘The environment’ in developmental psychology means ‘everything outside the child’. So,
postnatally, children’s environments include all aspects of their physical and social worlds.
Clearly this is considerably richer in many ways than the restrictive environment of the
mother’s womb. The sample answer in Figure 2 indicates that the postnatal environment is
much more variable, even at a simple physiological level. In addition, in the mother’s womb the
child cannot move easily, and has very limited access to information from the senses (which
are also at an early stage of development). However, it is important to realize that the
mother’s womb is an environment, so even prenatal development cannot be put down
exclusively to the unfolding of a genetic blueprint. Look back at the work of DeCasper and
Spence (1986) and DeCasper et al. (1994) in Chapter 2 (Section 2) of this book to see
examples of how the environment impacts on the prenatal child.
s
circumstances are right, pass the signal down the neuron’s axon and on to other
neurons. The pattern of branching of dendrites is important, because it affects the
amount and type of signals the neuron receives. The points of communication
between neurons are called synapses (see Figure 3). Synapses begin to form in the
brain in the early weeks of gestation. The generation of synapses occurs at
different times in different areas of the cortex.
Synapses
connecting
other
neurons
Neuron
cell body
Axon
Myelin sheath
around axon
Synapse
Figure 3 (a) A diagram of a neuron showing an axon (much shortened) with axon terminal
(synapse) and dendrites with synaptic connections to other neurons; (b) a more realistic picture of
the density of a dendritic tree in a human neuron (adapted from Stewart, 1991).
Figure 4 Dendrites in the visual cortex of human infants. As the dendritic trees of neurons
extend, so the overall connectivity in the brain increases (adapted from Conel, 1939–63).
2.3 Plasticity
Plasticity is an inherent property of the developing brain. At birth the brain
appears to be highly adaptable or ‘plastic’. If one area is damaged, perhaps
through a localized brain injury, or a stroke, other brain regions can take over the
processing from the damaged region. In adult life, the brain is considerably less
plastic, and localized brain damage is much harder to overcome.
Brain growth involves a process of increasing specialization in the sense that
tissue and cells become more differentiated in their structure and functioning as
development progresses. Sometimes this increased specialization is referred to as
a ‘restriction of fate’ because at the outset of its life a cell might take on any
number of forms and functions, but as development progresses, and as it assumes
its mature form and connectivity, its ‘fate’ becomes increasingly specified.
Plasticity simply represents the state of not yet having achieved specialization at
some level. As an example, consider a piece of tissue from the cerebral cortex that
may not yet have developed its specialization for processing a certain category of
information when a neighbouring region is damaged. The same developmental
mechanisms that would have ensured specialization for one type of processing
may now bias the tissue towards the type of processing normally undertaken by
its damaged neighbour. Thus in many instances, atypical patterns of brain
specialization in developmental disorders may reflect the action of normal
developmental processes following some earlier disturbance to the typical
developmental pathway. Identifying and understanding the mechanisms
underlying specialization, particularly in postnatal life, remains one of the major
challenges for developmental cognitive neuroscience.
124 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
This section will examine how such modules develop. Are they genetically
pre-specified, simply unfolding as the child matures? Or are the functions and
underlying structures of modules influenced by interactions with the
environment?
t
Activity 3 The shared environment
Allow about This activity encourages you to reflect on the way in which psychologists reason about genetic and
20 minutes environmental influences on development. It also aims to get you thinking about how the different
functional areas of the brain might develop.
The fact that most typically developing children and adults end up with the same sort of
perceptual, cognitive and motor functions performed by the same regions of cortex seems
to suggest that the brain develops according to a genetic blueprint (in the same way that
other features of the human body develop in similar ways in different people on the basis of a
genetic blueprint). But such reasoning is not really sound.
To understand the problem with this reasoning it is necessary to consider the shared
environment in which people develop. At one level it would seem that people’s
environments are unique to each individual. But if you take a few steps back and look at the
bigger picture, you should realize that typically developing children and adults live in very
similar environments. Spend a few minutes thinking about the following questions and write
down your answers before reading further.
1 In what ways can one person’s environment differ from the environment of another?
2 In what ways can it be said that people share a common environment?
Comment
People’s environments vary in all sorts of ways, and it should not be difficult to list just a few of
these. However, at a more global level there are also many ways in which it can be said that
people’s environments are alike. For example, typically developing people are surrounded by
very similar perceptual worlds, with similar physical constraints (owing to the laws of physics
and biochemistry) on what they perceive and what they do. They live in a world of shared
objects that behave in certain predictable ways. It is necessary to think at this more global level
to understand why similar cortical structures and functions do not necessarily imply that genes
are the only factor involved in brain development. Genetic determination may play a part, but
they may also be the result of a shared environment. It is more likely that the cortical
structures and functions that typically developing adults have in common are a result of
epigenetic processes which involve the interaction of genetic and environmental factors.
3 Some people experience atypical patterns of development; for example people with
motor disabilities such as cerebral palsy, or sensory disabilities such as hearing loss or
visual impairment. Can you think of ways in which people with these kinds of disabilities
might not share a similar environment with typically developing children and adults?
What implications might this have for our understanding of brain structure and function
in atypical development?
3 BRAIN AND COGNITIVE DEVELOPMENT 127
Comment
People with different kinds of sensory disabilities operate in a world in which they rely on
different kinds of sensory input. For example, people with an autism spectrum disorder may
live in rather distinctive sensory worlds (Williams, 1996). Children with motor disabilities may
not act in and on their environment in the same way as typically developing children. If the
epigenetic view is correct we might expect such factors to have an effect on brain structure
and function. In other words, if differences in brain structure and function are detected when
comparing typically and atypically developing people, we should not assume that these
differences are the cause of the disability; they might result from the disability.
s
Comment
Phylogenesis is to do with the development of a species. Ontogenesis is to do with
the development of an individual member of a species. When discussing the role of
the environment in development, it is important to be clear about whether one is
talking ‘phylogenetically’ (that is, about the development of a species through
evolution over many generations of individuals) or ‘ontogenetically’ (about the
development over one lifespan of one or more members of the species from the
same evolutionary stage).
BOX 1
n
3 BRAIN AND COGNITIVE DEVELOPMENT 129
t
Reading
Now read Reading A ‘The fragmented mind’ from What Infants Know by Mehler and Dupoux
(1994). This excerpt further illustrates the notion of cognitive modularity by drawing on
some clinical case studies. As you read the descriptions of the cases, consider how they
relate to Fodor’s model of the mind.
s
Fodor’s ideas on innately specified modularity have been hugely influential
among cognitive developmental theorists, but they have been vigorously
contested. One critic is Karmiloff-Smith (1992), who accepts the basic notion of
encapsulated cognitive modules, but who questions whether such modules are
innately pre-specified. The more highly pre-specified a cognitive system is in its
infancy, she argues, the less creative and flexible it can become as it develops. She
suggests that because human minds are remarkably creative and flexible, a high
level of innate specificity in brain function is unlikely.
Working within a broadly constructivist framework, Karmiloff-Smith proposes
that modules are the product of development; that the human mind becomes
modular as a result of development. A relatively limited number of innately pre-
specified ‘constraints’ on the kind of information that is dealt with by the different
parts of the infant brain would, she reasons, be enough to produce modular adult
minds. ‘I argue for innately specified predispositions that are more epigenetic
than Fodor’s nativism.’ (Karmiloff-Smith, 1992, p. 5). You can read more about her
perspective, and the type of evidence that supports it, in Reading B.
t
Reading
Now read Reading B ‘Is the initial architecture of the infant mind modular?’ from Beyond
Modularity by Karmiloff-Smith (1992). This summarizes Karmiloff-Smith’s objections to
Fodor’s nativist views, and sets out some features of her constructivist alternative. While
reading, focus on the notions of domain-specificity and domain-generality, and on her argument
that the brain becomes modular as a result of development.
s
What does the wider research evidence have to say on the matter of innate
modularity (Fodor) versus modularization (Karmiloff-Smith)? As ever the picture
is complex and incomplete. Some researchers (for example, Rakic, 1988) argue
that the structural differentiation of the cortex could be explained in terms of a
molecular and genetic specification (just as other organs in our body, such as the
heart, lungs and liver, are differentiated one from another, without particular
reference to interactions with the environment). If, once differentiated, the
various structures of the cortex always took on the same cognitive functions in the
developing child’s mind, then this would support Fodor’s position.
However, the balance of evidence (O’Leary, 1989; Elman et al., 1996; Katz
and Shatz, 1996; Johnson, 1997) suggests that for most regions of the cortex
differentiation of function results from an epigenetic system. That is,
differentiation results from an interaction between molecular and genetic factors
130 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
on the one hand and environmental factors on the other. This view holds that
neural activity, itself the product of both biological and environmental factors, is
the key to understanding differentiation of function in the cortex. Before birth,
this neural activity is probably largely a spontaneous and intrinsic process, driven
Neural activity by genetic and molecular factors (though note our earlier reminder that the
The electro- mother’s womb provides an ‘environment’ in which the baby can, for example,
chemical activity
of neurons as they hear). After birth, however, neural activity is substantially influenced by sensory
transmit and motor experience as well as by intrinsic factors.
information from What evidence can be cited to support this view? Petersen et al. (1990) used
one to another.
positron emission tomography (PET) (see Box 2) to study the responses of native
English speaking adults to written stimuli in the form of (a) English words, (b)
‘pseudowords’ that obeyed English spelling rules (for example floop and toglo),
(c) nonsense strings of letters (such as pxqlo) and (d) letter-like forms (false
fonts). They found that a specific region of the left visual cortex only responded
to the English words and to the pseudowords. It is implausible to suggest that
native English speakers are genetically programmed to develop an area in the
brain that will respond to the shapes of letters, but only when they are grouped
together to form English words or words that follow the rules of English spelling
(note that the activation was in visual areas, so the findings are unlikely to be to
do with the ‘pronounceability’ of the different types of letter strings). A much
better explanation of this finding would be that experience of a particular
language environment has influenced the development of this specialist area of
cortical function. The fact that the location of this specialist area is shared by
native English speakers suggests that there are similar genetic and molecular
processes working in harmony with a similar language environment, to produce
a similar structural and functional outcome.
There are many more studies that provide evidence for the effect of experience
on cortical structure and function. One example is Neville’s (1991) work using
scalp recorded event-related potentials (ERPs) with congenitally deaf participants
(see Box 2). For these participants she found that regions of the temporal lobe
(see Figure 1, p. 118) which in typical development respond to auditory, or multi-
modal input (input from more than one of the senses) had become dominated by
responses to visual input. This suggests that the function of these regions had
been influenced by the distinctive sensory experiences of the participants.
BOX 2
n
132 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
A further example, also from Neville and her colleagues (Mills et al., 1993) shows
how, with experience, certain types of processing are performed by progressively
smaller (more localized) regions of cortex. For instance, data from scalp-recorded
ERPs suggest that processing of known words and control stimuli is initially
spread over a relatively large area of cortex. This processing narrows to an area
over the left temporal lobe only when the child’s vocabulary reaches about 200
words, irrespective of maturational age. We will look at further evidence relating
to whether modularization occurs as the brain develops when we discuss the
specific case of language in Section 4.
Comment
The area of the brain shown by Neville and colleagues (Mills et al., 1993) to process
some aspects of word recognition appears to change at a certain point in vocabulary
development, rather than at a certain age. How can we interpret these findings? If
this ‘narrowing’ of function to the left temporal lobe happened for all children at a
similar age, it could be argued that this was due to an innate ‘pre-programmed’
specification of function that was simply unfolding with maturity. Indeed this could
still be the case if it happened at different ages, given that children mature at different
speeds. The fact that it happens at different ages for different children, and that it
relates to something so dependent on experience (vocabulary development)
suggests that the environment is playing a part and supports the modularization
position.
3.2 Self-organization
Section 3.1 provided some evidence to support the view that the environment
plays a part in the development of the structures and functions of the brain. If this
is the case, how might it happen? What processes might be involved in this gene–
environment interaction that we are invoking? One answer lies with the concept
of self-organization.
Self-organization occurs when structure emerges in response to a system’s
dynamic interactions with an environment. All stages of brain development
involve an element of self-organization (Keslo, 1995; Johnson, 1997). Initially the
neural system is relatively undifferentiated (randomly organized) but, as a result
of small adaptive changes, an order begins to emerge among the elements of the
system. In the brain, synaptic adjustment rules (such as the Hebb rule; see Box 3)
can lead to ordered connection patterns that in turn lead to structured behaviours.
In arguing that cognitive modules emerge as a result of development, Karmiloff-
Smith (1992) is proposing that the brain is a self-organizing system. The child
interacts with its environment, and as a result particular neural structures and
modules emerge. Self-organization is regarded by many researchers as a
fundamental characteristic of the brain (Changeux et al., 1984; von der Malsburg,
1995).
3 BRAIN AND COGNITIVE DEVELOPMENT 133
BOX 3
n
Self-organization occurs in systems with a large number of degrees of freedom;
that is, in complex systems where there are seemingly endless possibilities for
how things might ‘turn out’. In such systems it is difficult to see how all the
elements can be following a predetermined plan. For example, in the case of
brain development, the cerebral cortex alone contains some 100,000,000,000,000
(also written as 1014) synapses. It is unlikely that the genes can (in any direct way)
encode the full information necessary to generate this level of complexity
according to a predetermined plan (Elman et al., 1996). This is another argument
in favour of an epigenetic view of brain development, and in favour of Karmiloff-
Smith’s argument that modularization happens as a product of development.
There are many well-studied examples of self-organization in the physical and
biological sciences. For example, in snowflakes and in many other crystals, a
complex structure emerges from the apparently random addition of new material
(see Figure 6). Simple local rules of where new material can be added result in
intricate global patterns appearing in the final crystal. Processes of self-
organization can explain many complex patterns in nature, from the shape of
storm clouds to the spots on a leopard (Goldfield, 1995).
134 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Figure 6 A snowflake is a complicated, non-random shape that looks ‘designed’, yet it arises
from a very simple process; there is no design for a snowflake built into water molecules. When an
ice crystal starts to form it initially takes a simple hexagonal shape. With the crystal suspended in
humid, cold air, water molecules join on more easily at the corners, starting to form a star. Alternate
melting and freezing then creates more corners where molecules join, leading to the crystal
becoming more and more complex.
4 Language
This section extends the argument from Section 3. It looks in more detail at
evidence relating to the question of whether specialist brain functions are
genetically predetermined, or the result of a more epigenetic system in which
genes and environment interact. We will focus on the case of language. Of all
human abilities, language has been regarded as the most ‘biologically special’
(Johnson, 1997, p. 137). It is in this domain that the strongest nativist case for
innateness of function has been made, notably by Chomsky (1965, for example).
Chomsky’s ideas have been further developed in the persuasive arguments of
Pinker (1994), which are the topic of this section of the chapter.
communities and of the non-native signers, are actually doing no more than every
child does – they are reinventing language anew as they use it. He sees this as
compelling evidence for a language instinct:
... children actually reinvent [language] generation after generation – not
because they are taught to, not because they are generally smart, not
because it is useful to them, but because they just can’t help it.
(Pinker, 1994, p. 32; emphasis as in original)
The second element of Pinker’s case is drawn from Chomsky’s own ‘argument
from the poverty of the input’ (Pinker, 1994, p. 42). The core of this argument is
that children produce language that they can never have heard before, indicating
that they are not ‘learning’ or imitating it, but generating it anew. Pinker gives the
example of how children are able to produce grammatically correct questions,
even though they will not have heard these questions asked before, and even
though the rule that they have to follow is comparatively complex, and one that
they are unlikely to have learned.
t
Activity 4 How are questions formed?
Allow about This activity will help you to reflect on Pinker’s second piece of evidence for the innateness of
10 minutes language – the fact that children formulate sentences correctly, even when the rules they are using
are complex. Pinker’s evidence is based on Chomsky’s classic illustrations of how children appear to
know the deep structure of their language.
1 Turn the following statement into a question: ‘A unicorn is in the garden’. Can you
devise a rule that you could follow to convert a statement like this into a question?
Comment
You should have come up with something like ‘Is a unicorn in the garden?’ or ‘Is there a
unicorn in the garden?’ The rule that you followed might have been ‘take the first ‘‘is’’ and
move it to the beginning of the sentence, and add a question mark’. (Whether or not you
added ‘there’ is not relevant to this activity.)
2 Now formulate a question from the following sentence: ‘A unicorn that is eating a
flower is in the garden’. Write your question down. Does your original rule still work?
Comment
Your original rule would produce the non-grammatical (in fact nonsensical) question ‘Is a
unicorn that___ eating a flower is in the garden?’. This is clearly not correct. This time you have
to take the second ‘is’ and move it to get ‘Is (there) a unicorn that is eating a flower in the
garden?’.
s
Pinker’s and Chomsky’s observation is that children have no difficulty in
formulating such questions, even though the rule they have to follow is not
superficially easy, and despite probably never having heard such a construction
before. ‘Surely not every child learning English has heard Mother say Is the doggie
138 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
that is eating the flower in the garden?’ (Pinker, 1994, pp. 41–2). According to the
nativists, children achieve the right grammatical form for the question because
they have an innate grasp of the deep structure of language and its units of
meaning.
The third element of Pinker’s argument is related to the second. Pinker asserts
that many languages around the world use auxiliary verbs and that these
languages move the auxiliary to the beginning of the sentence to formulate a
question. (The ‘is’ in the sentences in Activity 4 is an English auxiliary verb.) If you
think about it, there are limitless ways in which a question might be formulated.
Why not turn the sentence backwards, or exchange the first and last words? So
why have all these languages hit upon the same way of doing this? ‘It is as if
isolated inventors miraculously came up with identical standards for typewriter
keyboards or Morse code or traffic signals.’ (Pinker, 1994, p. 43). Pinker’s answer
to this question is that it is not miraculous, and it is not a coincidence; rather, it
reflects a commonality in the structure of the human brain.
The final and fourth part of Pinker’s argument brings us back to the specific
topic of this chapter; that is, to studies of brain and cognitive development. He
argues that there is ‘an identifiable seat [for language] in the brain, and perhaps
even a special set of genes that help wire it into place’ (Pinker, 1994, p. 45). We
will look in more detail at this claim, but now from the perspective of the wider
developmental cognitive neurosciences literature.
through deafness) can lead to other cortical regions performing the same
functions.
On balance, the evidence from neuroscientific studies points to the conclusion
that several cortical regions are capable of supporting the development of
language, but does not support the hypothesis of full equipotentiality. For
example, Neville et al. (1998) used the fMRI technique (see Box 2, Section 3.1) to
examine the brain regions that are involved in language processing in deaf and
hearing participants. When the deaf participants viewed American Sign Language
(ASL) sentences (see Figure 7b) they showed activation in language areas in the
left hemisphere, as did the hearing participants on reading English sentences (see
Figure 8a). However, the deaf participants (who were all native ASL signers) also
displayed a level of activation in the right hemisphere that was not observed for
the hearing participants (compare Figure 7b with Figure 8a).
Figure 7 Images of left and right hemispheric Figure 8 Images of left and right hemispheric
activation in response to American Sign Language activation in response to written English
(adapted from Neville et al., 1988). (adapted from Neville et al., 1988).
t
Activity 5 Equipotentiality:evaluating the evidence
Allow about This activity encourages you to think about the implications of Neville et al.’s pattern of findings.
10 minutes
Look carefully at the findings reported for Neville et al. (1998) (see Figures 7 and 8). You will
see that they found that both hearing and deaf participants in their study showed activation in
language areas of the left hemisphere when processing language stimuli in their native
language (ASL for deaf participants, written English for the hearing participants). In addition,
the level of right hemisphere activation was higher for the deaf participants than for the
hearing participants. What implications do these findings have for the hypothesis of
equipotentiality? Make brief notes before reading further.
140 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Comment
Bearing in mind that these findings are from just one study (and you should therefore exercise
caution in interpreting them) they indicate the following:
1 The language areas of the left hemisphere are prepared in some way to process language-
related information, irrespective of the form of the language. Both signed and written
sentences were processed by the classic language areas of the left hemisphere.
2 Some areas of the right hemisphere had been ‘captured’ for language processing in the deaf
participants, perhaps because of the role of visuo-spatial information in sign languages.
Both these findings suggest some kind of bias towards language processing in areas of the left
hemisphere, but also that other areas in the right hemisphere are capable of supporting
language-related information processing. Note also the right hemisphere activation for deaf
participants when reading English sentences (Figure 8b).
s
Neville and her colleagues (Neville et al., 1992) also noted different patterns of
cortical response (as measured by ERPs) in hearing and deaf participants whilst
reading grammatically significant words in sentences, but noted similar patterns
Semantic of response across the two groups when they were reading semantically
To do with important words. So perhaps the question of how biologically special language is
meaning.
may require different answers for different aspects of language processing and
acquisition. Neville et al.’s findings indicate that grammatical aspects of language
may be more sensitive to experience than semantic aspects, because hearing and
deaf participants (who have a different experience of language) responded in
different ways to the grammatical information. After summarizing a decade of
their research, Neville and Bavelier conclude that the evidence supports the
hypothesis that ‘there are constraints on the organisation of the neural systems
that mediate formal language ... however, it is clear that the nature and timing of
sensory and language experience significantly affect the development of the
language system of the brain’ (Neville and Bavelier, 2001, p. 283).
Further evidence to support the hypothesis that different areas of the cortex are
capable of supporting language processing comes from studies of children with
localized brain damage (focal lesions) that happened either before or during
birth. In a sample of children aged between 3 and 9 years, Reilly and colleagues
(Reilly et al., 1998) found that the group of participants with a focal lesion
performed worse on a series of language tasks than the group of typically
developing control participants (see Research summary 1). However, the children
with a focal lesion showed a pattern of catching up on these measures, then
lagging behind at the next stage of language development, then catching up
again. The crucial points here are: first, functional recovery from a focal lesion
appears to be an ongoing process in childhood, and second, there is an
implication that functions affected by the original damage to a localized area of
the cortex were taken over in later development by undamaged areas of the
cortex.
3 BRAIN AND COGNITIVE DEVELOPMENT 141
RESEARCH SUMMARY 1
In study children were asked to tell a story from a 24-page picture book about a boy, a
frog and a dog. The researchers aimed to examine the effects of focal lesions which
happen before 6 months of age on children’s subsequent narrative language skills. The
study involved 62 children, of which thirteen had right-hemisphere damage (RHD),
eighteen had left-hemisphere damage (LHD) and 31 were typically developing
(neurologically intact) control participants. All the RHD and LHD children had focal
lesions in one hemisphere only, which had occurred before 6 months of age. The age
range of participants was 3 years and 5 months (3;5) to 9 years and 4 months (9;4), with
a mean age of just over 6 years. The participants were also subdivided into two age
levels: younger than 5 years (seven LHD, four RHD and eleven controls) and older than
5 years (eleven LHD, nine RHD and 20 controls). This gave the researchers an
experimental design that allowed them to test hypotheses relating to age (comparing
the younger group with the older group), neurological status (comparing the children
with focal lesions with the neurologically intact group), lateralization (comparing RHD
with LHD), and interactions among all these factors. This is known as a factorial design.
1.50
Proportion of morphological errors
1.25 Control
LHD
1.00 RHD
Regression
lines
0.75
LHD
0.50 RHD
0.25 Control
0.00
2 4 6 8 10
Age in years
The researchers collected both quantitative and qualitative data. The quantitative data
involved systematic measurements of various features of the narrative account of the
story produced by each child.
The researchers found that the children with brain injury showed some delays in
grammatical and narrative skills when compared to their control group. But the data
also showed ‘clear evidence of change and development on all fronts’ (Reilly et al., 1998,
p. 358). For example, Figure 9 plots a measure of grammatical errors (‘morphological
errors’) on the vertical axis against age on the horizontal axis. Each child is represented
by a marker within the graph. The higher up the marker is, the more errors the child
made. The ‘regression’ lines (or ‘lines of best fit’) for the three groups, all of which slope
downwards from left to right, show that in all groups (LHD, RHD and control) older
142 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
children did, on average, better than younger children. Note how the lines converge
with age; at 4 years of age there is quite a gap between the performance of the control
group, on the one hand, and the LHD and RHD groups on the other. By 6 years of age
this gap in performance had narrowed, and by 9 years of age it had narrowed even
further.
The qualitative data were used to analyse the nature of any grammatical errors. From
this analysis the researchers were able to conclude that for the brain injured group ‘the
raw data suggest ... a delay in morphological development, but there is no evidence here
for a qualitative deviation from the normal pattern’ (Reilly et al., 1998, p. 360). The
younger children in the LHD and RHD groups made significantly more errors than their
age-matched controls, but the errors that they made were ‘normal’ errors.
n
Like Reilly and colleagues (1998), Stiles and Thal (1993) also studied children with
focal lesions that had occurred either to the left or right hemisphere before 6
months of age. The children were examined longitudinally, and were found to
have language delays irrespective of which hemisphere had been damaged.
Surprisingly, though, the most significant delays in word comprehension were
found among the right hemisphere lesion group. In adults one would expect to
see this kind of problem when an area of the left hemisphere becomes damaged.
This suggests that the regions responsible for language acquisition are not
necessarily the same as those used for language processing in the adult.
The evidence presented in this section does not in itself rule out the
proposition that the left hemisphere is innately ‘pre-wired’ for language. It only
shows that other areas of the brain can ‘do’ language, if necessary. However, there
is a proposition that fits the evidence a little more securely than this. This is that
the architecture and connectivity of the left temporal lobe (the region most clearly
related to language processing in adults) biases the system towards building
neural modules that specialize in language processing in these regions. These
regions may be slightly better suited to processing rapidly changing information
than other areas of the cortex, and so can develop more efficient computational
properties for processing language. However, the regions are not the only place
where language processing can find a home. There is sufficient plasticity in the
developing child’s cortex to enable other regions to take on these functions,
perhaps with some loss of efficiency. ‘While regions of the left temporal lobe may
be best suited to language processing, they are not critical since language can
develop in a close to normal way without [them]’ (Johnson, 1997, p. 141). Overall,
therefore, the evidence from developmental cognitive neuroscience supports a
compromise position somewhere between equipotentiality on the one hand, and
genetically predetermined language-specific cortical regions on the other.
3 BRAIN AND COGNITIVE DEVELOPMENT 143
t
Activity 6 Evidence for and against a language ‘instinct’
Allow about This activity encourages you to review the debate about ‘the language instinct’, and to weigh up the
30 minutes evidence for and against the nativist approach.
Review the nativist arguments of Chomsky and Pinker in Section 4.1 in the light of the
evidence presented in Section 4.2, together with the evidence presented in Section 3. Does
the evidence support or refute their position? As usual, the answer is probably not clear cut.
To help you carry out your review, construct a list of points in favour of the nativist stance,
and points against it. Structure your list as shown in Figure 10, remembering to cite evidence
wherever possible (we have provided some examples to get you started).
. In typically developing adults, areas in the left temporal lobe deal with
many language-specific functions. Damage to these areas in adulthood
can seriously disrupt key language functions.
. It is likely that the left temporal lobe becomes the ‘seat of language’
because of slight advantages in its capacity for processing complex,
rapidly changing information, rather than because it has an innate
capacity to process language.
. Broadly speaking, the neuroscientific evidence supports an epigenetic
view of the development of language-specific functions in the adult
brain.
in EEG responses recorded from over the prefrontal cortex were correlated with
the ability to respond successfully over longer delays in delayed response object
permanence tasks; the more activation there was in the prefrontal cortex, the
longer the infant’s response could be delayed and still be successful.
A second source of evidence is work on cognitive development in children
with phenylketonuria (PKU). Even when treated, this inborn condition can cause
Phenyl- a reduction in the levels of a neurotransmitter, dopamine, in the prefrontal cortex.
ketonuria Neurotransmitters are chemicals that enable neurons to communicate with each
An inborn
condition which other across their synapses, so a reduction in the level of a neurotransmitter will
affects the infant’s affect the efficiency of neural connections. In these cases, reductions in dopamine
ability to process levels in the DLPC result in impairments on tasks thought to involve parts of the
a protein
(phenylalanine)
prefrontal cortex such as object permanence and object retrieval tasks. The
and which can affected infants show no impairment on tasks thought to depend on other regions
lead to severe of cortex.
learning
difficulties if
untreated by a 5.3 The prefrontal cortex and the acquisition of
dietary
intervention. new skills
Neuro- An alternative perspective on the role of the prefrontal cortex in cognitive
transmitters development has been proposed by several researchers who have suggested
Chemicals that are that the region may play a role in organizing other parts of cortex (for example,
released by
neurons and Thatcher, 1992), and that it plays a critical role in the acquisition of new
which affect the knowledge and skills. According to this view, regions of the prefrontal cortex play
transmission of an important role in the early stages of the development of a new skill. As mastery
information
between neurons.
of the skill increases, so involvement of the prefrontal cortex decreases, and other
specialist areas of cortex take over. The processing resources of the prefrontal
cortex are then switched to some other newly developing skill. In contrast to the
perspective presented in Section 5.2 this view leads to the prediction that the
involvement of the prefrontal cortex in a particular task or situation will decrease
with increased experience or skill in the domain.
Two recent lines of evidence are consistent with the view that regions of
prefrontal cortex play a key role during the earlier stages of skill acquisition
during infancy. Johnson and colleagues (1998) studied infants with localized
damage to parts of their cortex in a visual attention task. They found that only
infants with damage in the prefrontal regions of cortex were impaired on the task.
In adults one would expect those with damage to parietal regions (see Figure 1,
p. 118) to show such an impairment. Similarly, a recent study involving visual
attention and eye movement planning in infants showed suggestive evidence of
prefrontal cortex involvement (Csibra et al., 1998). The same effect was not
observed in adults. You can read about Johnson et al.’s research in more detail in
Research summary 2.
148 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
RESEARCH SUMMARY 2
n
The findings from the Johnson et al. (1998) study, and from Csibra et al. (1998)
are consistent with the view that the cortical regions that are crucial for a
particular ability change with the stage of acquisition of that ability. Perhaps there
is greater prefrontal activity on visual attention tasks in infancy than in adulthood
because infants are still at an early stage of developing visual attention skills. By
adulthood the skills have been learned and have ‘moved’ to a different, more
specialized area of the cortex.
3 BRAIN AND COGNITIVE DEVELOPMENT 149
t
Activity 8 The role of the prefrontal cortex
Allow about This activity will help you to review the material presented in Section 5 of this chapter.
15 minutes
Summarize the evidence that supports each of the two perspectives presented in Section 5
on prefrontal cortex development in infancy. The two perspectives are:
1 Various aspects of the cognitive and behavioural development of infants can be related
to structural changes in the prefrontal cortex.
2 The prefrontal cortex is specifically associated with the early stages of the development
of a new skill, and its involvement decreases with the development of that skill.
s
6 Conclusion
This chapter has introduced you to the field of developmental cognitive
neuroscience. It has described some aspects of early brain development, and
discussed notions of innate modularity and modularization. Particular attention
has been paid to language and to the prefrontal cortex. The chapter has taken an
epigenetic view of brain development, arguing that children’s brain structures and
functions develop out of potent interactions between intrinsic (for example,
genetic) and external (environmental) factors. It is evident from the material
presented in this chapter that the major challenge for researchers in the field of
developmental cognitive neuroscience is to find ways of enhancing our
understanding of the complex relationship between brain structures and
functions.
Further reading
Elman, J. L., Bates, E. A., Johnson, M. H. et al. (1996) Rethinking Innateness: a
connectionist perspective on development, Cambridge, MA, MIT Press.
Johnson, M. H. (1997) Developmental Cognitive Neuroscience, Oxford, Blackwell.
Nelson, C. A. and Luciana, M. (2001) Handbook of Developmental Cognitive
Neuroscience, Cambridge, MA, MIT Press.
References
Bell, M. A. (1992) ‘A not B task performance is related to frontal EEG asymmetry
regardless of locomotor experience’, in Rovee-Collier, C. (ed.) Proceedings of the
VIIIth International Conference on Infant Studies 15 (Special ICIS Issue), p. 307,
Miami Beach, FL: Infant Behavior and Development.
Bell, M. A. and Fox, N. A. (1992) ‘The relations between frontal brain electrical
activity and cognitive development during infancy’, Child Development, vol. 63,
pp. 1142–63.
Casey, B. J., Trainor, R. J., Orendi, J. L. et al. (1997) ‘A developmental functional
MRI study of prefrontal activation during performance of a go-no-go task’,
Journal of Cognitive Neuroscience, vol. 9, pp. 835–47.
Conel, J. L. (1939–63) The Postnatal Development of the Human Cerebral Cortex,
Vols I–VI, Cambridge MA, Harvard University Press.
Changeux, J.-P. (1985) Neuronal Man: the biology of mind, New York, Pantheon.
Changeux, J.-P., Heidman, T. and Patte, P. (1984) ‘Learning by selection’, in
Marler, P. and Terrace, H. S. (eds) The Biology of Learning, pp. 115–33, Berlin,
Springer-Verlag.
3 BRAIN AND COGNITIVE DEVELOPMENT 151
Chomsky, N. (1965) Aspects of the Theory of Syntax, Cambridge, MA, MIT Press.
Cowan, M. W., Fawcett, J. W., O’Leary, D. M. and Stanfield, B. B. (1984)
‘Regressive events in neurogenesis’, Science, vol. 225, pp. 1258–65.
Csibra, G., Tucker, L. A. and Johnson, M. H. (1998) ‘Neural correlates of saccade
planning in infants: a high-density ERP study’, International Journal of
Psychophysiology, vol. 29, pp. 201–15.
DeCasper, A. J. and Spence, M. J. (1986) ‘Prenatal maternal speech influences
newborns’ perception of speech sounds’, Infant Behavior and Development,
vol. 9, pp. 133–50.
DeCasper, A. J., Lecanuet, J-P., Busnel, M-C., Granier-Deferre, C. and Maugeais, R.
(1994) ‘Fetal reactions to recurrent maternal speech’, Infant Behavior and
Development, vol. 17, pp. 159–64.
Diamond, A. (1985) ‘Development of the ability to use recall to guide action, as
indicated by infants’ performance on AB’, Child Development, vol. 56, pp. 868–83.
Diamond, A. and Goldman-Rakic, P. S. (1986) ‘Comparative development of
human infants and infant rhesus monkeys of cognitive functions that depend on
prefrontal cortex’, Neuroscience Abstracts, vol. 12, p. 274.
Diamond, A. and Goldman-Rakic, P. S. (1989) ‘Comparison of human infants and
infant rhesus monkeys on Piaget’s AB task: evidence for dependence on
dorsolateral prefrontal cortex’, Experimental Brain Research, vol. 74, pp. 24–40.
Elman, J. L., Bates, E. A., Johnson, M. H., Karmiloff-Smith, A., Parisi, D. and
Plunkett, K. (1996) Rethinking innateness: a connectionist perspective on
development, Cambridge, MA, MIT Press.
Fodor, J. A. (1983) The Modularity of Mind: an essay on faculty psychology,
Cambridge, MA, MIT Press.
Fox, N. A. and Bell, M. A. (1990) ‘Electrophysiological indices of frontal lobe
development’, in Diamond, A. (ed.) The Development and Neural Bases of Higher
Cognitive Functions, pp. 677–98, New York, New York Academy of Sciences.
Fuster, J. M. (1989, 2nd edn) The Prefrontal Cortex, New York, Raven Press.
Goldfield, E. C. (1995) Emergent Forms: origins and early development of human
actions and perception, Oxford, Oxford University Press.
Goldman-Rakic, P. S. (1987) ‘Development of cortical circuitry and cognitive
function’, Child Development, vol. 58, pp. 601–22.
Hebb, D. O. (1949) The Organisation of Behavior: a neuropsychological
approach, New York, Wiley.
Huttenlocher, P. R. (1990) ‘Morphometric study of human cerebral cortex
development’, Neuropsychologia, vol. 28, pp. 517–27.
Johnson, M. H. (1997) Developmental Cognitive Neuroscience, Oxford, Blackwell.
Johnson, M. H. and Tucker, L. A. (1996) ‘The development and temporal
dynamics of spatial orienting in children’, Journal of Experimental Child
Psychology, vol. 63, pp. 171–88.
152 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Johnson, M. H., Tucker, L. A., Stiles, J. and Trauner, D. (1998) ‘Visual attention in
infants with perinatal brain damage: evidence of the importance of anterior
lesions’, Developmental Science, vol. 1, pp. 53–8.
Karmiloff-Smith, A. (1992) Beyond Modularity: a developmental perspective on
cognitive science, Cambridge, MA, MIT Press.
Katz, L. C. and Shatz, C. J. (1996) ‘Synaptic activity and the construction of cortical
circuits’, Science, vol. 274, p. 1133–38.
Keslo, S. (1995) Dynamic Pattern: the self-organization of brain and behavior,
Cambridge, MA, MIT Press.
Kwong, K. E., Belliveau, J. W., Chesler, D. A. et al. (1992) ‘Dynamic magnetic
resonance imaging of human brain activity during primary sensory stimulation’,
Proceedings of the National Academy of Sciences, vol. 89, pp. 5675–9.
Lewkowicz, D. J. and Turkewitz, G. (1981) ‘Intersensory interaction in newborns:
modification of visual preferences following exposure to sound’, Child
Development, vol. 52, pp. 827–32.
Maurer, D. (1993) ‘Neonatal synesthesia: implications for the processing of speech
and faces’, in de Boysson-Bardies, B., de Schonen, S., Jusczyk, P. et al. (eds)
Developmental Neurocognition: speech and face processing in the first year of life,
pp. 109–24, the Netherlands, Kluwer.
Mehler, J. and Dupoux, E. (1994) What Infants Know, Oxford, Blackwell.
Meltzoff, A. N. and Borton, R. W. (1979) ‘Intermodal matching by human
neonates’, Nature, vol. 282, pp. 403–4.
Mills, D. M., Coffey, S. A. and Neville, H. J. (1993) ‘Language acquisition and
cerebral specialization in 20-month-old infants’, Journal of Cognitive
Neuroscience, vol. 5, pp. 326–42.
Milner, B. (1982) ‘Some cognitive effects of frontal-lobe lesions in man’,
Philosophical Transactions of the Royal Society of London, vol. 298, pp. 211–26.
Nelson, C. A. and Bloom, F. E. (1997) ‘Child development and neuroscience’,
Child Development, vol. 68, pp. 970–87.
Neville, H. J. (1991) ‘Neurobiology of cognitive and language processing: effects
of early experience’, in Gibson K. R. and Peterson A. C. (eds) Brain Maturation
and Cognitive Development: comparative and cross-cultural perspectives,
pp. 355–80, New York, Aldine de Gruyter.
Neville, H. J. and Bavelier, D. (2001) ‘Variability of developmental plasticity’, in
McClelland J. L. and Siegler, R. S. (eds) Mechanisms of Cognitive Development:
behavioral and neural perspectives, pp. 271–87, London, Lawrence Erlbaum
Associates.
Neville, H. J., Mills, D. and Lawson, D. (1992) ‘Fractionating language: different
neural subsystems with different sensory periods’, Cerebral Cortex, vol. 2,
pp. 244–58.
Neville, H. J., Bavelier, D., Corina, D. et al. (1998) ‘Cerebral organisation for
language in deaf and hearing subjects: biological constraints and effects of
3 BRAIN AND COGNITIVE DEVELOPMENT 153
Readings
Reading A: ‘The fragmented mind’ from
What Infants Know
Jacques Mehler and Emmanuel Dupoux
A, a musical score; B, the same score in braille; C, a transcription of the braille score with the letters
used. The patient can play the score presented in braille, but has great difficulty reading E, which
represents ... ‘‘le père’’ [the father]. He reads: ‘‘Le ... Pa ... le ... ta ... le frère [the brother] ... la ... le par
... il y a [there is] pé.’’ (Signoret et al., 1987).
The example of another patient, also a professional musician, bears out this
theory. It concerns a right-handed conductor who had led the La Fenice orchestra
in Venice and the La Scala orchestra in Milan for many years. At sixty-seven, a
stroke injuring his left cerebral hemisphere caused the loss of all linguistic
aptitude, and he was unable to speak for the next six years. On the other hand,
his musical capacities remained almost intact. He was incapable of naming notes,
but could sing or play them. He made such rapid progress that, despite his
inability to communicate with an orchestra except by gestures, he continued to be
a completely professional conductor capable of detecting the most subtle musical
errors. His interpretation of Verdi’s Nabucco was especially appreciated by critics.
The lesion resulting from his first stroke therefore affected only his left
hemisphere, the seat of language. Six years later, a second vascular accident partly
destroyed his right hemisphere. Unfortunately, this time he did not recover and
his death prevented our testing his musical capacities. Other cases of lesions
suppressing the aptitude for music confirm, however, that musical ability has its
seat in the right hemisphere of the brain.
As indicated above, commissurotomy cases illustrate the dissociation that exists
between language and other mental functions, for in these patients we can
present data to one hemisphere only. For example, we can arrange for the image
of a word to reach the left hemisphere only. Subjects can read it and even point to
the corresponding object. But this is possible only if they use their right hand.
This hand is directly linked to the left hemisphere, in other words, to the one that
‘‘saw’’ the word.
If, on the contrary, we present a word to the opposite hemisphere, subjects are
unable to read it, and even claim not to have seen anything. However, they can
locate the object, with their left hand. In these cases, the left hemisphere, which is
responsible for language, has not ‘‘seen’’ the word presented, and subjects are
therefore unable to formulate a verbal response. To summarize, if we analyze the
subjects’ behavior in terms of what they perceive and produce with their right
hemispheres, we conclude that they have a very diminished linguistic capacity
similar to that of certain aphasics. Yet if we analyze their behavior in terms of
what they control with their left hemisphere, we conclude that they are in full
possession of their linguistic capacities.
156 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
How is this possible? How can one both be and not be aphasic? In fact, this
discrepancy between mental faculties in the same person is only partial and
manifests itself only in very specific experimental situations. Under normal
conditions, both hemispheres function in a coherent fashion, for they have many
means of communicating with another. For example, eye movements permit the
duplication of visual data in both hemispheres, and information can also circulate
through the intermediary of motor activity. For instance, we have observed a
subject give an incorrect verbal response, raise his eyebrows, and then
immediately correct himself. No doubt the right hemisphere, which knew the
correct answer, prompted the raised eyebrows, which informed the left
hemisphere of its error and led him to change his response. It is also possible that
such phenomena sometimes occur in healthy individuals.
Be that as it may, these surprising observations clearly show that it is
counterproductive to see our mental faculties as an inseparable whole. In the light
of neuropsychological studies, our psychological mechanism seems rather to be
composed of independent and autonomous faculties like the perception of faces
and of language. And language itself is not all of one piece either. Very specific
afflictions illustrate this. Some patients, for example, lose the use of conjunctions
but not of nouns and adjectives, or even the use of a whole semantic group like
fruits and vegetables, or trees. An apparently simple and homogeneous task
therefore calls upon many distinct and specialized submechanisms, which can
break down while the rest of the system continues to function normally.
Commonsense psychology, as we clearly see, cannot explain this type of
phenomenon. Nor is the perspicacity forged by years of experience of any use.
We therefore must assume that a part of our cognitive system is divided into
functional units, or modules, which are responsible for a given aptitude and
operate in an autonomous fashion oblivious to what is happening elsewhere in
the system. The more research advances, the less our intelligence seems to
resemble a whole with indistinct equipotential parts, rather like a bowl of jello,
and the more it seems to be subdivided into a great number of functions which
have a certain autonomy. Gall’s intuition has proven to be right. As we already
mentioned, it was taken up again in more modern form by Jerry Fodor.
According to Fodor’s formulation, we can compare a function, or module, to a
physical organ. Our body is made up organs which have their specific functions
and which, through their interaction, contribute to the functioning of the whole.
However, modules do not exchange fluids or energy, but information. Thus each
module constitutes an abstract organ which we could describe, at least in a rough
approximation, in terms of concepts borrowed from computer science
(representation, structure of data, manipulation of symbols, etc.). In a sense, then,
a module is closer to a computer program than to a heart or liver. However, like
its physiological counterpart, it has a specific function: it only processes a
fragment of the information circulating through the whole psychological
mechanism and can only use rather restricted, predetermined channels of
communication. Conscientious, but rather limited, it is an expert in its field.
Furthermore, just as in the case of physical organs, the growth and organization
of cognitive modules is guided by genetic program specific to the species.
And finally, as the example of language shows, a module is not distributed
3 BRAIN AND COGNITIVE DEVELOPMENT 157
throughout the entire brain, but involves a specific nervous structure, circuit, and/
or cortical area.
From this perspective, we can no longer imagine that our psyche is controlled by
a central intelligence. The data used by each module are limited and its field of
operation is restricted. A so-called superior intelligence is no longer necessary. It
is the system as a whole that exhibits a behavior we can term intelligent, not its
parts taken separately, and not just one part in particular.
How many modules are there? How are they organized? According to one of
Fodor’s hypotheses, there are many specialized systems which are consolidated
into large modules, each specialized in the rapid and automatic processing of one
type of data (e.g., language). Each module delivers information to a central
processing unit which compares the different entries as well as all the other
knowledge available to the organism, making it possible to elaborate the long-
term planning of actions.
It is quite obvious that we have not yet been able to explore this extremely
complex modular architecture in detail. We are still taking the first baby steps
in a discipline which is centuries behind the natural or exact sciences. This
functionalist representation of the psychological mechanism is above all a
guideline: it enables us to formulate hypotheses and submit them to experimental
verification. The actual existence of a particular system or a particular exchange of
data is not a metaphysical postulate. The beauty and perfection of the theory
matter less than its ability to raise questions that experiments can then proceed to
answer.
Contents
1 Introduction 165
References 198
Reading 201
Reading A: An extract from The Articulate Mammal 201
4 THE DEVELOPMENT OF CHILDREN’S UNDERSTANDING OF GRAMMAR 165
Learning o u t c o m e s
After you have studied this chapter you should be able to:
1 describe the key stages in children’s grammatical development;
2 understand Chomsky’s theory of ‘Universal Grammar’;
3 show familiarity with methodological approaches to studying children’s
grammatical development;
4 offer a critical account of single and dual route explanations of grammatical
development;
5 compare and contrast nativist and empiricist accounts of grammatical
development.
1 Introduction
If you have ever tried to learn a foreign language as a teenager or an adult, you
will know that one of the most complex aspects is understanding and
remembering its grammatical rules. Learning vocabulary is a relatively simple
affair, but knowing how to combine those words to form legitimate sentences that
are an accurate representation of what you want to communicate is a more
sophisticated skill. Moreover, the fact that languages can differ greatly from each
other in their grammatical rules further emphasizes the complexity of the task. So,
consider how much more difficult it would be if you were never explicitly taught
what these rules were, and were left to discover them for yourself. Think also
about how difficult this task would be if you were not already fluent in a language
of your own. And yet, this is exactly the position that infants are in during the first
few years of life.
In Chapter 2, ‘First words’, you read about how children come to acquire their
spoken vocabulary. In this chapter the focus is on how children achieve the
apparently insurmountable task of acquiring grammatical understanding. We will
Nativists
Theorists who
consider the nature of spoken language and how it is constructed before looking
argue that at methodological approaches to this area of research. We will then examine two
development is approaches to explaining how children acquire knowledge about the grammar of
primarily driven
their first language. The two approaches reflect and further develop the long-
by innate (inborn)
constraints. standing debate over the fundamental source of human knowledge, be it
linguistic or non-linguistic, between nativists and empiricists.
Empiricists
Theorists who What you will also notice while reading this chapter is how children’s
argue that grammatical development further illustrates an underlying theme of this book:
development is that cognitive development is characterized by a progression from a piecemeal
primarily driven
by factors in the knowledge of the environment to an organized, systematic understanding that
environment. imposes structure on what is experienced.
166 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Spoken language
Phonology Grammar
The structure
of speech sounds
Morphology Syntax
The structure The structure
of words of sentences
BOX 1
2.1 Phonology
Phonology and phonotactics are terms that are used to refer to the knowledge that
a speaker possesses about the sound patterns of his or her native language. For
example, in English, speakers readily distinguish sounds like [b] as in ‘bat’ from
sounds like [p] as in ‘pat’, whereas Arabic speakers ignore this distinction.
t
Activity 1 Phonological distinctions
Allow about This activity will help you appreciate the subtlety of the phonological distinctions that are made in
5 minutes understanding language.
Say ‘pig’ and ‘big’ out loud, paying attention to the very beginning of each word, and what
your mouth and throat have to do to produce these sounds. In what way are they the same,
and what is the crucial difference?
Comment
For both words, the mouth, lips and tongue are in the same position. The main difference is
that the /p/ in ‘pig’ is ‘voiceless’ (i.e. the vocal chords do not move as the initial sound is made)
but for the /b/ in ‘big’ voicing starts sooner. If you put your hand on your throat while saying
just the words you will notice this more easily. Other sounds, such as [t] and [d], and also [k]
and [g], differ in the same way. Although the difference between these two sounds is subtle,
the impact that it has on the meaning of an utterance is huge (e.g. ‘big’ and ‘pig’ simply differ in
whether the first brief sound in the word is voiced or voiceless, but they have very different
meanings!).
s
168 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Similarly, other languages such as Thai, recognize other subtle distinctions which
English speakers ignore; the aspirated [ph], as in ‘pit’, versus the unaspirated [p] as
in ‘spit’, are treated as the same sound by English speakers, but signal a difference
in meaning to speakers of Thai. To notice the difference between the aspirated
Aspirated and unaspirated [p], try saying ‘pit’ while holding your hand in front of your face;
These are you should feel a burst of air on your hand. Now do the same as you say ‘spit’.
phonemes that
are accompanied You will notice that the burst of air is absent, or at least a great deal weaker.
by a ‘puff of air’ as Likewise, there are combinations of sounds that English speakers would regard
they are spoken. as ‘foreign’. For example, the /p/ sound cannot be followed by the /s/ sound at
the beginning of a word in English but it is permitted in French. So, in the English
word ‘psychology’, the letter ‘p’ is silent, but in the French word ‘psychologie’ the
/p/ and the /s/ sounds are both pronounced. Speakers who are not careful in
articulating these subtle cues, or listeners who do not pay attention to them, are
likely to be misunderstood or to misunderstand the communications of others.
Indeed, as you saw in Chapter 2, much of language development during the first
year of life is concerned with mastering the mechanics of speech production, or
deciphering the sound patterns of the language that dominate the infant’s acoustic
environment.
2.2 Morphology
Morphology is the term used to refer to the knowledge a speaker possesses
regarding the manner in which new words can be created from existing words or
other meaningful units of language. There are many aspects to morphology, so
we will introduce you to just two here for the purposes of illustration. The first of
these is known as compounding. This is the combinatorial capacity in a language
whereby two existing words are glued together to form a new word. For
example, in English the nouns ‘lady’ and ‘bird’ can be combined to form the new
word ‘ladybird’. Sometimes, as with this example, the meaning of the new word is
not predictable from its origins. In other cases, the meaning of the newly created
word is more transparent, like ‘lighthouse’.
Young children frequently demonstrate sensitivity to morphology in the way
that they invent their own compound words that are meaningful to them. For
example, one child that we know spontaneously invented the noun ‘moregranny’
Grammatical to refer to one of her grandmothers, and to differentiate her from her other
morpheme ‘granny’. The logic of ‘granny’ and ‘moregranny’ is undeniable and appealing, and
A ‘morpheme’ is a all the more remarkable because this shows us that children are not merely
unit of speech that
has meaning. A imitating what they hear adults saying: they are generating their own ideas about
grammatical the rules that govern how language is constructed.
morpheme is a Language users exploit another combinatorial device in which existing words
unit of speech that
modifies the
are joined together with other sounds that do not have meaning in-and-of
meaning of the themselves. These other sounds, referred to as grammatical morphemes, modify
word to which it the meaning of the words with which they combine: the morpheme /s/ in the
is added, such as
word ‘boys’ indicates plurality (more than one boy); the morpheme /ed/ in the
‘ed’ when added
to the end of a word ‘kicked ’ indicates that the action was performed in the past. The use of
word like ‘kick’. grammatical morphemes such as word endings in this way is known as inflection.
4 THE DEVELOPMENT OF CHILDREN’S UNDERSTANDING OF GRAMMAR 169
English-speaking children begin to use plurals early in their third year. Most
children will also have used some past tense verbs by about 30 months of age
(see Table 1, Section 3.3). In English, plural and past tense forms are marked
either by a process of adding an ending (dog ? dogs, walk ? walked) or through
some change of the word stem (man ? me n, see ? sa w). The process whereby
an ending is added is known as suffixation. This is the regular (typical) way of
forming plurals and past tenses in English. Other changes are called irregular
(though do note that some nouns, for example ‘deer’, and verbs, for example ‘hit’,
do not change their form at all in the plural or the past tense). Although a wider
range of nouns and verbs are inflected through suffixation, irregular plurals and
past tenses are commonly used in everyday adult language. Indeed, irregular
plural and past tense forms are amongst the earliest acquired by young children.
Again, it can be inferred that children have a developing sensitivity to
inflections from their early utterances. That is, initially children produce plural
and past tense verb forms perfectly. However, towards the end of their third year
they start to make errors like these:
My teacher holded the baby rabbits and we patted them.
I love cut-upped egg.
(Pinker, 1995, p. 109)
Instances where children incorrectly add /s/ to the end of a word to indicate
plurality (e.g. ‘sheeps’), or apply the ‘add /ed/’ rule for indicating that something
has happened in the past are known as over-regularization errors. These types of
error continue for several years. However, such errors are relatively rare, with
over 95 per cent of irregular past tense verb use continuing to be perfect (Pinker,
1995; see also Figure 2). The incidence of over-regularization then decreases as
children get older before gradually returning to 100 per cent performance in
adulthood. The reason behind this ‘U-shaped development’, as it is termed by
psychologists, is linked to children’s increasing sensitivity to morphological rules
(see Figure 2). That is, initial ‘perfect’ production of past tense verbs is likely to be
due to children simply copying the speech that they hear. The period of over-
regularization suggests that children have inferred a rule about how to indicate
that things happened in the past – ‘add /ed/’ to the end of the verb – which they
then have to learn to apply appropriately. We will discuss different explanations
of how children learn to do this later in the chapter.
170 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
100 100
Figure 2 Graphs showing the percentage of irregular verb errors (a) observed in the speech of
Adam (one of the children from Brown’s study, 1973), and (b) made by a computer model that is
learning how to produce irregular verbs (Plunkett and Marchman, 1993). Note that in both cases
the overall rate of correct responses is high, but both are characterized by a small ‘dip’ in
performance over time and as vocabulary increases. This is the so-called ‘U-shaped’ pattern of
development.
2.3 Syntax
Syntax is the term used to refer to the knowledge that a speaker possesses
regarding the manner in which words can be combined to form sentences. In
English, syntax provides the information needed to determine who did what to
whom. For example, English speakers have no difficulty distinguishing the
meanings of the two sentences:
(1) John gave Mary the flowers.
(2) Mary gave John the flowers.
In the first sentence, John does the giving whereas in the second sentence it is
Mary. The meanings of the words by themselves do not yield this information. It is
necessary to pay attention to their position in the sentence. It might be tempting
on the basis of this simple observation to conclude that syntax is synonymous
with word order. For example, we might infer that the noun that occurs first is
always the subject of the sentence. However, it is relatively easy to show that
syntax is more complicated than this. In the sentence,
(3) Mary was given the flowers by John.
the subject of the sentence is the last noun mentioned, yet it is not difficult to
understand that John did the giving. Figuring out who does what to whom
requires the listener to be sensitive to the cues that indicate the type of sentence
used. The language learner must decipher these cues – in this case, whether it is a
passive (as in example (3) above) or an active sentence (as in example (1)). In
general, this means paying attention to the structure of the sentence.
One of the achievements of grammarians has been to discover the kinds of
structures listeners must recognize when understanding sentences. For example,
4 THE DEVELOPMENT OF CHILDREN’S UNDERSTANDING OF GRAMMAR 171
passive sentences are easily identified by the presence of the word ‘by’. However,
sometimes the structural cues are far more subtle. In Chapter 3, Activity 4, you
considered the example,
(4) A unicorn is in the garden.
and how easy it is to turn this into a question by moving the word ‘is’ to the front
of the sentence. However, you also noted that this simple ‘rule’ does not apply
when a sentence such as (5) is converted into a question.
(5) A unicorn that is eating a flower is in the garden.
used to describe the event <Mary chasing John> but hears the sentence,
(7) John chases Mary.
used to describe the event <John chasing Mary>, then there is good reason to
suppose that syntactic cues (word order) are important in the language. On the
other hand, if the child hears the sentence,
(8) John chases Mary-da.
used to describe the event <John chasing Mary> but hears the sentence,
(9) John-da chases Mary.
used to describe the event <Mary chasing John>, then there is good reason to
suppose that morphological cues are important in the language (the suffix ‘da’ is
indicating the object of the sentence: the person or thing that is on the receiving
end of the action in the sentence). This solution to the problem presupposes that
172 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
children can remember the word sequences or identify the inflections on the
words and then work out how they relate to the events described by the speaker.
An alternative solution is that young children already ‘know’ that languages can
be syntactically or morphologically oriented. That is, they are born with an
implicit understanding that languages can signal who did what to whom through
syntactic or morphological devices. In Chapter 3 you were introduced to the idea
of language development being underpinned by an innate predisposition to
acquire it; a position advocated by the work of Noam Chomsky, and further
developed by Steven Pinker. With respect to acquiring grammatical
understanding, Chomsky (1965) argued that humans have an innate knowledge
of potential language structures, which he refers to as Universal Grammar.
Universal Grammar contains a set of constraints on language processing that can
be switched on or off through exposure to spoken language. Languages which
exploit morphological cues ‘switch on’ the morphological constraints, whilst
those which exploit syntactic cues switch on the syntactic constraints.
Chomsky argued that this kind of innate knowledge is necessary because
children are rarely presented with coherent, grammatically complete speech that
maps directly onto things happening in their immediate environment. For
example, speakers might say a variety of different things in the context of <John
chasing Mary> such as ‘Look’ or ‘Not him again’, or the conversation may be
about something that happened in the past that the child knows nothing about.
How are children to know which are the important parts of language to focus on?
Chomsky’s answer to this question is that it is Universal Grammar that directs the
child’s linguistic attention to the right kind of information.
t
Reading
At this point you should read Reading A, which is an extract from The Articulate Mammal by
Jean Aitchison. This reading offers a simple introduction to Chomsky’s notion of a Universal
Grammar which we will return to later in this chapter.
s
4.00
2.50
III 2.75 Some specific uses of the verb ‘to be’, ‘the’, ‘a’, irregular past
tenses of verbs, use of -’s to indicate ownership
IV 3.5 Regular past tense of verbs (add /ed/), use of the ‘third
person’
BOX 2
Systematic transcription
Systematic transcription is a process whereby periods of speech are recorded and then
transcribed. The recording almost invariably includes excerpts of speech from the
child’s conversational partner, thereby providing a context in which to interpret the
child’s utterances. The continuous collection of data (sometimes over several hours)
offers the investigator a representative sample of the child’s language. Indeed, recording
in different situations allows the investigator to determine how speech varies from one
context to another. The cost of this enriched source of data is the time taken to collect
and transcribe the material. For example, a single hour’s talk can take many hours to
transcribe.
There are two ways that speech can be transcribed. The first of these is phonetic
transcription which involves representing each utterance using a special phonetic
alphabetic. This differs from standard alphabets, which may use the same letter symbol
to represent more than one sound (e.g. in English the letter ‘c’ can represent the sound
[k] or [s]), or more than one symbol to represent the same sound (e.g. in English the
letters ‘f’ and ‘ph’ both represent the sound [f ]). Moreover, each language only uses a
proportion of all the sounds that it is possible to produce. In contrast, the phonetic
alphabet has one symbol each for all the sounds that could possibly be produced by the
human voice and also distinguishes between sounds that a speaker of a single language
may not recognize (see Section 2.1). This makes it ideal for transcribing the talk of
young children whose early utterances can include ‘foreign’ sounds.
Many researchers opt instead for the second way, that is producing orthographic
transcriptions. This form of transcription glosses over the peculiarities of the child’s
pronunciation, and identifies words from adult language that the investigator believes
4 THE DEVELOPMENT OF CHILDREN’S UNDERSTANDING OF GRAMMAR 177
that the child is attempting to produce. The addition of video-recording methods and
digital enhancements provides investigators with even richer sources of information to
interpret and analyse children’s utterances. For example, it is now possible to view on a
computer screen a digital video recording of a child simultaneously with a transcription
of what the child or his or her conversational partner is saying.
However, systematic transcription records do not necessarily offer an accurate
reflection of what children know about language. After all, children may know a lot
more about language than they actually say. Also, given the naturalistic situation in
which the data are typically collected, it can be very difficult to assess what could be
contributing to a child’s linguistic behaviour. Take, for example, the case of a child who
produces an especially complex utterance, potentially indicative of a sophisticated
knowledge of language. Was it, in fact, only produced because of something an adult has
just said or done, or because of the presence of other objects and events at the time?
Consequently, investigators have turned to more experimental methods to assess
children’s comprehension of linguistic forms, which offer the investigator more control
over the situation in which the children’s talk occurs.
n
Receptive Relatively little work has been done to determine the grammatical abilities
language present in one year olds’ receptive language. In general, the ability to
The language that
a person actually
comprehend language is better developed than the ability to produce language.
understands, but For example, children are unlikely to produce the word ‘dog’ unless they are able
that may not to understand it. Consequently, there is good reason to believe that the
necessarily feature
developmental profile of speech presented in this section may underestimate the
in their own
language use. grammatical abilities of the one year old.
Experimental procedures have been developed to enable investigators to
better assess young children’s potential to understand grammatical rules. One
such procedure is the inter-modal preferential looking task. This procedure was
first developed by Spelke (1979) who used it to investigate infants’ perceptual
abilities. She showed infants two events (e.g. something being dropped or
hands clapping) but only played the sound associated with one of those events
(e.g. clapping). The children were found to spend more time looking at the event
that was consistent with the sound that they heard, their extended gaze
suggesting that they correctly associated the sound with the event. This technique
has since been adapted to analyse very young children’s understanding of
grammar and how this might facilitate the acquisition of new words (see Research
summary 1).
178 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
RESEARCH SUMMARY 1
Transitive verb
A verb that must
refer to an object
in a sentence. For
example the verb
‘to hit’ is transitive
– it is not
legitimate just to
say ‘I am hitting’.
It is necessary to
say what is being
hit. Figure 4 Experimental set-up of the preferential looking task (from Naigles and Kako,
Intransitive verb 1993, p. 1670).
A verb that does The child was shown actions where one actor had an effect on something (causative
not require its
object to be actions) and actions that had no effect on another thing (non-causative actions). So the
specified in a child saw, for example, footage of a duck making a rabbit bend over (a causative action)
sentence. For while both animals were making arm movements (the non-causative actions). This
example, ‘to read’
is intransitive as it
scene was paired with a nonsense verb (e.g. ‘gorp’) which was presented in one of two
is possible to say sentences. One sentence, ‘The duck is gorping the bunny’, implied that the nonsense
‘I am reading’. It is word was a transitive verb (one that requires a corresponding ‘object’). The
not necessary to
say what is being
other sentence, ‘The duck and the bunny are gorping’, implied that the nonsense word
read. was an intransitive verb (one that does not require an object).
Then the actions were shown separately. On one screen the duck was shown bending
the rabbit over; on the other screen the duck and the rabbit were shown just moving
their arms (see Table 2).
4 THE DEVELOPMENT OF CHILDREN’S UNDERSTANDING OF GRAMMAR 179
Naigles found that the children looked at the causative actions, such as the duck forcing
the rabbit to bend over, for longer than they did at the other screen when they heard
the phrase ‘The duck is [word]-ing the bunny!’. They looked at the non-causative
actions, such as the duck and rabbit making arm movements, for longer when they
heard the phrase ‘The duck and the bunny are [word]-ing!’. These results were taken as
evidence that children are able to use their sensitivity to grammatical forms to learn
about verb meanings.
Table 2 Structure of the stimulus video tapes for the verb ‘gorp’
The duck is forcing the rabbit Look! The duck is gorping Black screen
into a bending position; both the bunny!
are making arm gestures
Black screen Look! The duck is gorping Duck is forcing the rabbit
the bunny! into a bending position;
both are making arm
gestures
Duck is forcing the rabbit Look! The duck is gorping Duck is forcing the rabbit
into a bending position; both the bunny! into a bending position;
are making arm gestures both are making arm
gestures
Duck is forcing the rabbit Oh! They’re different now! The duck and rabbit are
into a bending position making arm gestures
Duck is forcing the rabbit Where’s the gorping now? The duck and rabbit are
into a bending position making arm gestures
Duck is forcing the rabbit Find gorping! The duck and rabbit are
into a bending position making arm gestures
Age 2;3: Play checkers. Big drum ... A bunny-rabbit walk ...
Age 2;6: What that egg doing? No, I don’t want to sit seat ...
Age 2;9: Where mommy keep her pocket book? Show you something funny ...
Age 3;0: I going wear that to wedding ... Those are not strong mens ... You dress
me up like a baby elephant ...
Age 3;2: So it can’t be cleaned? I broke my racing car. I want to have some
espresso ... Can I put my head in the mailbox so the mailman can know
where I are and put me in the mailbox? ...
(Pinker, 1994, pp. 269–70)
t
Activity 2 R e fl e c t i n g o n g r a m m a t i c a l d e v e l o p m e n t
Allow about This activity will help you to reflect on the ways in which children’s grammatical competence changes
5 minutes over time.
Look back at the examples of Adam’s speech. What do you notice about the way that Adam
forms his utterances over this period?
Comment
The most obvious thing to note is the change in the length of Adam’s utterances. His attempts
to combine words are not always grammatical, but they do increase in length over time. The
early utterances omit function words (‘the’, ‘a’, ‘on’, etc.) as expected. He attempts to
formulate questions quite early on in his multi-word utterances although ‘is’ is often omitted.
He also makes an interesting error – ‘mens’, where he correctly inflects the noun ‘man’, but still
adds the conventional /s/ ending to indicate that there is more than one. The later utterances
are more in line with adult speech (‘I want to have some espresso’) although they are not
always error free (‘... the mailman can know where I are ...’). The main source of error
throughout is from omitting words, rather than getting them wrong. You can see that simply
by inserting the missing words the sentences become grammatical: the actual order of the
words that are produced is correct, suggesting that Adam shows some grasp of syntax.
s
Even when they make errors, children still show sensitivity to correct forms, even
though their own productions may be flawed.
Child: You readed some of this too ... She readed all the rest.
Parent: She read the whole thing to you, huh?
Child: Nu-uh, you read some.
Parent: Oh, that’s right, yeah. I readed the beginning of it.
Child: Readed? (annoyed surprise) Read! (pronounced ‘red’).
Parent: Oh, yeah, read.
Child: Will you stop that, Papa?
(Pinker, 1995, p. 119)
4 THE DEVELOPMENT OF CHILDREN’S UNDERSTANDING OF GRAMMAR 181
Although children are continuously bombarded during their waking hours with a
stream of richly structured linguistic input, they seem to assimilate the structures
of their native language without explicit instruction or correction (Brown and
Hanlon, 1970). The ease with which all this seems to occur belies the underlying
complexity of the skill acquired. As is often the case with many complex abilities,
a lot can be learned from observing the order in which parts of the skill are
acquired, and tracking down the rare mistakes that individuals make in exercising
their skills. This is precisely what will be discussed in the next two sections. The
discussions in these sections will help you to evaluate the extent to which it can
be said that language acquisition is driven by innate cognitive mechanisms (like
Universal Grammar) or experience of a linguistic environment.
record, the rule system is ‘blocked’ and the memorized (irregular) form is
produced. For example, in an attempt to inflect the word ‘go’, the language user
will discover that the memory system contains ‘went’, which blocks the
application of the ‘add /ed/’ rule, preventing the error ‘goed’.
t
Activity 3 Dual route theory
Allow about This activity will enable you to assess how well you understand dual route explanations of inflection.
5 minutes
Assume that a boy’s ‘memory system’ contains only the following irregular past tense verb
forms: went, broke and held. What does the dual route theory predict the child will say when
he is required to produce the past tense of the following words: go, want and eat?
Comment
According to the dual route model, he would be expected to say ‘went’ rather than ‘goed’
because he has a representation of this irregular form in his memory system. He should say
‘wanted’ because if there is no corresponding irregular representation in the memory system
the ‘add /ed/’ rule will be applied. However, the boy will say ‘eated’ instead of ate, because
there is no stored representation of the irregular form, and so the ‘add /ed/’ rule will, once
again, be applied in this case.
s
According to this account, the development of the memory system is critical to the
proper operation of the inflectional process. In order to identify an irregular
inflection such as ‘went’ or ‘men’, the irregular forms need to be stored and
retrievable from memory. Failure to retrieve exceptions from memory will result
in a failure of blocking and hence an over-regularization error. Memory failure is
most likely due to lack of knowledge of particular words. Hence, irregular words
that only rarely occur in speech are most susceptible to over-regularization. With
greater experience (as the child gets older) memories for rare types of inflection
are consolidated and errors diminish. Dual route theory suggests that the
inflection that occurs across the greatest number of words is established as the
default rule (Marcus et al., 1992); for example -ed is the most common type of
past tense inflection in English, so it becomes the ‘rule’.
BOX 3
Connectionism
Connectionist networks are computer models loosely based on the principles of neural
information processing. They are sometimes referred to as neural networks. It is only
necessary to have a basic knowledge of how connectionist models work in order to
understand what they can tell us about cognitive processes (such as language
understanding and memory) and cognitive development.
A connectionist model consists of ‘units’, which can be thought of as roughly analogous
to neurons. A typical network is organized as follows. A set of units is designated as
input units. These can be thought of as equivalent to human sensory neurons which are
activated by a stimulus from the environment. Another set of units is designated as
output units. It might be helpful to think of these as the equivalent of motor neurons,
which produce a behaviour. In between the input and output units are various
configurations of hidden units. These can be thought of as equivalent to all the neurons
which are part of the ‘black box’ of cognition; everything that goes on between input
(stimulus) and output (response).
A connectionist network needs to be able to accept an input (a stimulus), make various
computations, and produce an output (a response). The job of a connectionist network
when it is being used as a model of human information processing, is to accept a stimulus
and to learn to make the response that a person would make to that stimulus. For
example, if a connectionist network were being built to model a child’s developmental
mastery of the English past tense it would be provided with the stimuli of English verb
stems. It would need to learn to respond to each verb with the appropriate past tense
form. Its behaviour would be observed as it learned (developed) and then compared to
the course of development that can be observed in children.
n
Plunkett and Juola (1999) trained a connectionist network on some 5,000 different
English inflections. They found that, like children, their network went through an
initial period of producing nouns and verbs correctly, which was then followed
by a period of intermittent over-regularization. They also found that the network
‘learned’ plural inflections before past tense inflections. If you look back at
Table 1, you will see that this is consistent with the results of Brown’s (1973) study
of children’s grammatical development. Given the small scale of these networks
compared to a human brain (see Chapter 3, Section 2.1), the proposal that a single
system might memorize all of the inflections in the language does not seem so
unlikely after all.
Single route theory differs from dual route theory in its account of why children
begin to make over-regularization errors. The explanation is this. As more and
more inflections are stored in memory, competition for memory resources
increases. This competition between words results in interference effects, where
words that are very similar to each other are easily confused. For example, ‘fan’
and ‘man’ sound very similar to each other but one has a regular plural inflection
(‘fans’) and the other has an irregular inflection (‘me n’). In the case of the
irregular word ‘man’, if there are lots of regular words stored that have a similar
4 THE DEVELOPMENT OF CHILDREN’S UNDERSTANDING OF GRAMMAR 185
phonological structure (e.g. ‘fan’, ‘ban’, ‘van’ and ‘span’) then ‘man’ may be
accidentally treated as if it too has a regular inflection (‘mans’).
This is exactly what Rumelhart and McClelland (1987) found when they trained
their connectionist network on regular and irregular verb inflections. Early in
training, the network only had a few verbs to memorize and so it managed to
inflect the verbs correctly. However, as its vocabulary grew, competition for
network resources intensified and over-regularization errors were found. With
further training, the irregular verbs became more robustly stored in the network
and were able to resist the interference effects of regular neighbours.
t
Activity 4 Computational modelling
Allow about This activity will encourage you to reflect on the advantages of computational modelling as a
5 minutes technique in research into child development.
Why would you want to model a developmental process? What would be the aim of such a
model?
Comment
The aim of any model is to test specific predictions about the mechanisms that underlie the
developmental process in question. It is one thing to propose a model on paper, but quite
another thing to build a computational model that actually has to function. For one thing it
forces researchers to be explicit about how all the different structures and processes that they
are proposing work. If the finished model shows patterns of behaviour that are similar to those
shown by children, then perhaps the structures and processes built into the model are similar
to those going on in the developing child. ‘Perhaps’ should be emphasized because even if two
systems (in this case a child and a computer model) produce the same behaviour, that does
not necessarily mean they are doing it in the same way. If you search your memory for a piece
of information, and a computer searches its memory for the same piece of information, the
output from you and the computer may be the same, but the search process may have been
completely different.
s
t
Activity 5 Predicting errors from the models
Allow about This activity will give you another opportunity to test your understanding of single and dual route
10 minutes accounts.
As discussed earlier, single route theories and dual route theories differ in the types of
mistakes that they predict children will make when learning about inflections. Here are two
types of child-like inflection errors [with the correct form of the verb in brackets]:
Can a dual route account explain both these types of error? Can a single route account
explain both these types of error?
Comment
Sentence (a) is an example of an over-generalization error, and hopefully you remembered
that both theories can explain this sort of mistake. Sentence (b) is an example of what
happens when a regular word takes an irregular inflection. While very rare, these do occur in
children’s speech. Dual route theory predicts that regular words are not susceptible to
interference effects from irregular words in the way illustrated by sentence (b). So errors like
‘lank’ should not occur. However, single route theory predicts that interference effects will
impact on children’s production of both regular and irregular words. That is, just as regular
inflections interfere with the production of irregular words (just as ‘fold’ becomes ‘folded’ in
the past tense, so ‘hold’ wrongly becomes ‘holded’), so should irregular words interfere with
the production of regular inflections (just as ‘sink’ becomes ‘sank’, so ‘link’ wrongly becomes
‘lank’, instead of ‘linked’). Consequently, this kind of evidence favours single route theory.
s
The two theories also differ in their account of how children generalize inflections
to new words. For example, given the novel word ‘wug’, children and adults
would most likely use the word ‘wugs’ as its plural form. Dual route theory
explains this creative use of inflections in terms of the operation of the rule, and
there being no entry in memory to block the application of the rule. Single route
theory explains generalization in terms of the similarity of the new word to
existing words in the language (‘mug’, ‘slug’, ‘hug’, etc.).
There is controversy here as to what the language user does with words that are
not similar to any other word stored in memory. This situation is particularly
poignant in the case of children with relatively small vocabularies. Dual route
theory predicts that the word will simply be inflected according to the default
rule, whereas single route theory predicts that an educated guess will be made on
4 THE DEVELOPMENT OF CHILDREN’S UNDERSTANDING OF GRAMMAR 187
the basis of a weighted similarity to all the other words in the individual’s
vocabulary. Experimental evidence provided by Marchman (1997) has examined
just this question and favours the single route account (see Research summary 2).
RESEARCH SUMMARY 2
Comment
So far, we have been careful not to relate either the dual or single route theories to
either nativist or empiricist positions. This is because both theories are simply about
specifying the nature of the cognitive mechanism that might account for the patterns
of development observed in young children. This debate can be seen as separate
from discussions of whether the mechanism is believed to be innately pre-
programmed or to develop in response to experience with the linguistic
environment. However, historically ‘nativists’ have aligned themselves with the dual
route theories (note that Pinker is an advocate of dual route theory) and
‘empiricists’ (such as Marchman) have supported the single route position. The basic
distinction is that dual route theorists generally propose that the rule mechanism is
innate whereas single-route theorists contest this view. However, it is entirely
possible to present a view in which a dual route system might arise as the result of
learning.
188 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
RESEARCH SUMMARY 3
Szagun found a much wider variety of errors in children learning German than that
reported by Marcus et al. (1995) with little evidence of the ‘s’ plural dominating in the
way they suggested. The children’s development in using the different plural inflections
was consistent with the occurrence of those inflections in their parents’ speech. The
frequency of the children’s use of the different inflections was the same as that of their
parents by the time they were 3 years and 8 months old. Finally, the children made
errors right from the beginning of their use of plural forms. These findings are
consistent with a single route account of German inflectional morphology.
Williams
4.5 Evidence for dual route theory from studies of
Syndrome developmental disorders
A non-hereditary
syndrome caused Another way of testing the strength of a theory is to see how well it can account
by a chromosomal for instances where the production of a behaviour is known to be problematic. So
abnormality. far you have seen that the pattern of typical development in English and other
Brain and physical
development can languages seems to favour the single route account. However, Pinker (1991) has
be affected, and argued that instances of the breakdown in usage of inflections favours the dual
symptoms can route account.
include poor co-
ordination, some
He cites two developmental disorders, Williams Syndrome and specific
muscle weakness, language impairment (SLI) where researchers have discovered different patterns
possible heart of behaviour for regular and irregular past tense word forms. For example, some
defects,
studies of children with Williams Syndrome (Bellugi et al., 1990) show that these
occasional kidney
damage and high children are able to produce regular forms of past tense verbs, but their
calcium levels. production of irregular past tense verbs is more problematic. In contrast, other
Development is studies of participants with SLI (e.g. Gopnik and Crago, 1991) show that irregular
generally delayed
in these children. forms are relatively preserved. For example, one 10-year-old child with SLI wrote
the following in his school notebook:
Specific
language
impairment Monday 12th September
(SLI) On Saturday I watch TV and I watch plastic man and I watch football. On
A specific and Sunday I had pork and potato and cabbage.
severe difficulty in
acquiring some (Gopnik and Crago, 1991, p. 39)
aspect of
language In this extract the regular verb ‘to watch’ is not marked for past tense, although
development,
despite no the irregular verb ‘to have’ is produced correctly. Gopnik and Crago noted that
apparent general the children were significantly more likely to make mistakes on regular verbs than
cognitive or on irregular ones. Moreover, there was evidence that when the teacher corrected
neurological
impairment.
past tense errors of this kind, the children did appear to learn the correct form in
that case, but were unable to generalize the ‘add /ed/’ rule to other regular verbs.
This pattern of results reflects a double dissociation between regular and
irregular inflection. Double dissociation refers to evidence that two cognitive
processes appear to be unrelated to each other. This is because not only does one
group of people show ability at A but impairment B (a single dissociation), but
another group of people exists who show ability at B but impairment at A. The
190 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Subject- Notice that both sentences (10) and (11) contain the necessary ingredients for the
auxiliary correct formulation of questions. The child has simply failed to place the auxiliary
inversion verb ‘is’ in front of the subject (‘Daddy’ or ‘Mummy’). Notice that sentences (10)
The change in
word order that is
and (11) stripped of their respective Wh-words constitute legitimate statements.
required to make Perhaps the first attempts to produce Wh-questions build upon the ability to make
a statement (e.g. statements? At some point, however, the child must discover the importance of
‘Mummy is
subject-auxiliary inversion for question formation. How is this achieved?
going’) into a
question (e.g.
‘What is Daddy
eating?’), which
5.1 What, where, why and who: asking questions
involves putting about syntax
the auxiliary verb
before the subject A relatively straightforward and uncontroversial answer is that children must pay
of the sentence attention to the practices of question formation in their language. They are given
(in this case, ample opportunity to discover that Wh-words are often followed by a relatively
Daddy).
small set of function words such as ‘is’, ‘has’ or ‘does’, and that these function
words are related to the noun that follows (‘Where are the boys going?’).
Uncovering these regularities is presumably just like discovering the regularities
of past tense and plural formation: looking for patterns in the language.
Unfortunately, this supposedly straightforward answer disguises a theoretical
hornet’s nest!
The idea that a limited set of function words immediately follow after a Wh-
word is not strictly true. For example, (12) is a perfectly legitimate question
beginning with a Wh-word, but is followed by a noun, (13) does not have an
auxiliary verb, and (14) is not even a question.
(12) What colour is the grass?
(13) What type of plane flies to New York in less than 4 hours?
(14) What Stephen did is of no consequence.
The language learner has to cope with all these different usages of ‘what’.
Furthermore, a little imagination reveals that the range of possible ‘What ...’
sentences is almost endless. Chomsky (1965) proposed that speakers can produce
and understand an indefinite number of sentences because they have mastered a
set of grammatical rules for combining words and morphemes in their language.
He also suggested that these rules cannot be extracted from the speaker’s
linguistic environment because there are simply too many possibilities to choose
between. Therefore, he argued, the language learner must have some innate
knowledge of the rules of language. This knowledge cannot be specific to a
particular language, so this is why he called it Universal Grammar (see Section 2).
Language learning therefore consists of finding evidence in the linguistic
environment that will enable the learner to select from the range of potential
grammatical rules that are made available by Universal Grammar.
4 THE DEVELOPMENT OF CHILDREN’S UNDERSTANDING OF GRAMMAR 193
In fact, the verb ‘give’ seems to allow this kind of paraphrase (the technical term is
dative alternation). Likewise, the verb ‘tell’ permits dative alternations:
(17) The man told the story to the boy.
(18) The man told the boy the story.
On the basis of exposure to these kinds of alternations, one might suppose that
children infer a rule that would enable them to exploit this pattern creatively.
However, care is required to avoid over-generalizing the rule. For example,
although (19) is perfectly legitimate, (20) is not:
(19) The man reported the accident to the police.
(20) The man reported the police the accident.
Despite the fact that the verbs ‘tell’ and ‘report’ have similar meanings, one allows
dative alternation and the other does not. Children need to be able to tune their
grammar to these exceptions, rather like they must notice that ‘went’ is the past
tense of ‘go’. In fact, children make over-generalizations of dative alternations
(Pinker, 1989) just like they over-regularize word endings. The problem facing
children in this case, however, is somewhat more serious. With word endings,
children will almost certainly hear counter-evidence that might persuade them to
give up their mistaken hypothesis – if they think that the past tense of ‘go’ is
‘goed’, they will eventually hear ‘went’. However, in the case of a mistaken dative
alternation, children will never hear a sentence that directly contradicts their
hypothesis (unless their parents produce negative evidence – see Box 4). They
will only hear correct renditions of the unalternated form.
BOX 4
Negative evidence
Brown and Hanlon (1970) argued that parents do not seem to correct the grammatical
errors of their children in a systematic fashion. They are more likely to correct the
factual errors or meaning of their children’s language. Explicit attempts to tell the young
language user that they have made a mistake – ‘No you said that wrongly’ – are very
rare indeed. Language researchers therefore assumed that any theory of language
acquisition needs to be able to deal with an environment that does not provide the
learner with negative evidence. This assumption has not gone unquestioned. For
example, Demetras et al. (1986) suggested that indirect or implicit feedback to the child
may occur when a grammatical mistake is made. Indirect feedback might take the form
194 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
account. If these categories exist at all in the mind of the child, they are emergent
properties of the learning process, rather than prerequisites for it.
So, according to this approach, how do children extend their knowledge to
new situations and new sets of words? It is not possible for children to memorize
all the utterances they hear for later reproduction. However, a possible
explanation for their creativity with language is analogy; children may initially
memorize some of the utterances they hear and then adapt them as prototypes for
the construction of new utterances. In fact, one of the earliest accounts of
children’s grammatical development (Braine, 1963, 1976) comes very close to
adopting this position (see Box 5 for a discussion of early ‘diary’ methods of
studying language development).
Braine (1963) examined his son Martin’s early word combinations and found
that they exhibited clear ordering patterns. Some words only ever occurred in one
position in the utterance (initial or final). He called these words Pivot words. Other
words were free ranging. He called these Open words. Of course, most words can
occur in a wide range of positions in a sentence. However, Martin may have
noticed that some words vary more in their distribution than others. For example,
nouns move around a lot in the sentence (subject versus object position) whilst
verbs tend to stay put. Did Martin observe his parents and other adults producing
speech where the position of the Open words in sentences varied more than Pivot
words and imitate these patterns to structure his own sentences?
BOX 5
n
196 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
6 Conclusion
This chapter has introduced you to the complex nature of learning grammar as an
infant, and to various theoretical explanations of how acquisition of grammar
might be achieved. These explanations can be aligned to broadly nativist and
empiricist accounts of development. Chomsky’s Universal Grammar and Pinker’s
dual route explanation of inflection use suggest that children are born with innate
cognitive mechanisms that enable them to acquire grammatical competence on
exposure to the linguistic environment. In contrast, single route theories of
inflection use and distributional accounts of syntax development suggest that
children can acquire representations of grammatical forms and how to use them
simply from exposure to the linguistic environment.
The balance of evidence presented in this chapter favours the more empiricist
accounts. Single route theories appear to offer a more credible account of
development across languages, and distributional accounts of syntax have the
support of evidence from modelling studies. This theory is also consistent with
observations of children’s language production and the nature of parental speech
to children. However, that is not to say that the idea of there being some form of
innate predisposition to acquire language has been rejected. Theoreticians are in
broad agreement that the homo sapien comes equipped with specialized
machinery that permits the acquisition of language. The days of tabula rasa
198 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
empiricism are long gone. Discussions are concerned with identifying the nature
of the skills that the child brings to language learning, and whether those skills are
specifically related to language, or part of a more general cognitive ability. The
idea of a Universal Grammar, a detailed, innate knowledge of all types of
grammatical rules, is still being considered and assessed, as is the idea of children
having the ability to extract grammar from their environment as a result of the
nature of their neural systems.
Throughout this chapter you should also have gained a sense of how methods
of studying children’s grammatical development have developed over time. Early
diary studies were superseded by the use of recording equipment and commonly
agreed methods of systematic transcription. However, as these methods only have
the potential to inform discussions of expressive language use, experimental
techniques have been used to assess the nature of children’s expressive and
receptive language abilities (for example, see Research summary 2 and Research
summary 1 respectively). Finally, advances in cognitive science have led to the
development of connectionist models that enable theorists to test the viability of
their theories regarding the cognitive architecture needed to realize acquisition of
grammar. As methods continue to develop in response to the theories,
psychologists will move closer to a resolution to debates regarding the nature and
extent of innate knowledge in children’s grammatical development.
Further reading
Aitchison, J. (1998, 4th edn) The Articulate Mammal: an introduction to
psycholinguistics, London, Routledge.
References
Bellugi, U., Bihrle, A., Jernigan, D., Trauner, D. and Dougherty, S. (1990)
‘Neuropsychological, neurological and neuroanatomical profile of Williams
Syndrome’, American Journal of Medical Genetics, vol. 6, pp. 115–25.
Bishop, D. V. M. (1997) ‘Cognitive neurospsychology and developmental
disorders: uncomfortable bedfellows’, The Quarterly Journal of Experimental
Psychology, vol. 50, pp. 899–923.
Braine, M. (1963) ‘On learning the grammatical order of words’, Psychological
Review, vol. 70, pp. 323–48.
Braine, M. D. S. (1976) ‘Children’s first word combinations’, Monographs of the
Society for Research in Child Development, vol. 41, Chicago, IL, University of
Chicago Press.
Brown, R. (1973) A First Language: the early stages, Cambridge, MA, Harvard
University Press.
4 THE DEVELOPMENT OF CHILDREN’S UNDERSTANDING OF GRAMMAR 199
Reading
Reading A: An extract from The Articulate Mammal
Jean Aitchison
We are therefore dealing with ‘a system of unifying principles that is fairly rich in
deductive structure but with parameters to be fixed by experience’ (Chomsky
1980: 66). The interlocking nature of the system will ensure that minor alterations
will have multiple consequences: ‘In a tightly integrated theory with a fairly rich
internal structure, change in a single parameter may have complex effects, with
proliferating consequences in various parts of the grammar’ (Chomsky 1981: 6).
In particular, ‘a few changes in parameters yield typologically different languages’
(Chomsky 1986: 152). This whole idea has become known as the ‘principles and
parameters’ or ‘P and P’ approach ...
Children have relatively little to do in this new system: ‘We view the problem of
language acquisition as ... one of fixing parameters in a largely determined
system’ (Chomsky 1986: 151). Indeed, many of the old rules which children had
to learn just appear automatically, because the principles underlying them are
there already. ... If this minimal effort by the child is correct, then it makes sense
to think of the language system as a ‘mental organ’, which grows mainly by itself,
in the same way that the heart grows in the body. Chomsky is increasingly
concerned to understand the principles which underlie this growth.
Contents
1 Introduction 207
5 Conclusion 228
References 228
5 EXECUTIVE FUNCTIONS IN CHILDHOOD: DEVELOPMENT AND DISORDER 207
Learning outcomes
After you have studied this chapter you should be able to:
1 provide a working definition of executive function;
2 evaluate the developmental significance of executive function;
3 describe the development of one aspect of executive function in children;
4 identify some key implications of executive dysfunction.
1 Introduction
One way of thinking about the developmental story that we have told so far in
this book is in terms of the progressive organization of children’s behaviour and
Behaviour experience.
The public side of Chapter 1 discusses ways in which infants organize their understandings of the
human life: world in terms of categories. Chapter 2 shows how this organizational framework
anything that a
person does that
underpins, and is developed by, language. Chapter 3 examines ways in which
is observable by a brain development can be seen in terms of increasing levels of organization in
third party. both structure and function. The theme is extended further in Chapter 4 by
Experience looking at how language itself develops into an organized system of linguistic
The private side structures. Why is this developmental trajectory towards organization in the
of human life:
growing child of such interest to psychologists?
internal cognitive
and emotional One answer to this question is that organization of behaviour and experience is
processes that a necessary pre-condition for intelligent, aware, planned action. It is a pre-
cannot be condition for children having control of their own actions. If a child (or indeed a
observed from the
outside. person of any age) has no way of organizing their understanding of the world
around them, then they will be unable to behave appropriately or effectively.
Imagine, for a moment, being unable to comprehend that the next person you
meet is a person, that they have things in common with other people, and that
they differ in significant ways from things that are not people. Alternatively
imagine knowing that people differ in significant ways from each other, and from
things that are not people, but not being able to control your actions in a way that
acknowledges these differences. In both cases there would simply be no basis for
anything other than chaotic behaviour.
The term ‘executive function’ (EF) is used to refer to the set of high-level
cognitive functions that enable people to plan, initiate and carry through goal-
directed behaviour in an organized and ‘thought out’ way. This definition reflects
the contrast between high-level cognitive functions such as integration, synthesis,
planning, organizing and so forth with more basic, ‘low-level’ cognitive functions
such as processing auditory, visual and tactile sensations. This distinction maps
onto the model proposed by Fodor in Chapter 3, Section 3, which sees the lower-
level cognitive functions as being organized in a modular fashion in such a way as
to feed the higher-level, and more global ‘executive’ functions.
208 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
then the children must copy the action that follows. If she says only ‘do this’ then
the children must stay still and not copy her.
Debbie’s mother says ‘Simon says, do this’ and sits on the floor. All the children
sit down as quickly as they can. She then says ‘Simon says, do this’, and jumps up
again. All the children follow. Immediately she says ‘do this’ and sits down again.
Half the children sit down. The other half put their hands over their mouths and
giggle. The children who have sat down quickly realize their mistake and jump to
their feet again.
The behaviour of the children who mistakenly sat down may be attributed to a
number of factors. They may not have listened to or understood the instructions
of the game, for example. But first and foremost their behaviour can be attributed
to a failure of executive function. Their behaviour was automatic and
momentarily beyond their conscious control. They acted first, and thought later.
Note that within seconds of sitting down (perhaps even in the same moment as
they sat down) the children probably realized their mistake. Conversely, the
children who ignored the adult model were able to resist what was most likely a
strong urge to sit down. Their behaviour was guided by a greater level of
‘executive’ control.
There are several things to note about this example that will help you to relate
the abstract concepts that will be covered in this section of the chapter to
everyday experiences (see Box 1).
BOX 1
t
Activity 1 Slips of action
Allow about This activity illustrates the distinction between automatic, habitual behaviour on the one hand, and
10 minutes consciously planned behaviour, on the other. It will enable you to use this distinction to analyse
familiar everyday mistakes.
Identify examples from your own experience of ‘slips of action’, when something that you
meant to do was overridden by automatic, habitual behaviour. One example might be dialling
a familiar telephone number, only to realize when the person answers that you had meant to
telephone someone else.
Comment
Here are some examples that we came up with:
. finding yourself taking the turning that you usually take every day on your way to work,
when you were meant to be driving somewhere else;
. writing last year’s date on a cheque on 2nd January;
. going upstairs to fetch something, getting distracted by something else, and going back
downstairs without the thing that you went to get in the first place;
5 EXECUTIVE FUNCTIONS IN CHILDHOOD: DEVELOPMENT AND DISORDER 211
. clicking the ‘OK’ button on the computer to ignore, as usual, a warning when this time you
should have pressed ‘Cancel’. You realize this as your finger clicks the mouse button. Your
whole assignment has been deleted!;
. making coffee for four people. One of them does not take milk. You intend to leave one
cup black, but you end up putting milk in all of them.
Everyday mistakes such as these appear to result from strongly automated behaviours taking
over from a conscious plan to do something else. They are examples of momentary failures of
executive function. The mistakes children make in the ‘Simon says’ game also result from a
strongly automated set of behaviours (copying the adult again and again) overriding conscious
planning.
s
It is important to note that for an act to be automatic it does not necessarily need
to be simple, such as copying an adult putting one hand in the air. Driving a car is
a complex skill but it is one which can become almost automatic if it is sufficiently
well learned. So, what does distinguish controlled actions from automatic actions?
Controlled actions involve three important features:
1 The execution of novel as distinct from familiar action sequences.
2 Making a choice between alternative responses in opposition to the
execution of a single action sequence.
3 The execution of acts that are accessible to consciousness, as opposed to
those that are not.
Again you will see that the ‘Simon says’ example fits with at least two of these
three features. We have already noted how the actions of the successful children
are at least momentarily accessible to consciousness (they are aware of what they
are doing), when the decision as to whether to copy or not is made. In addition,
these children have to make a choice between alternative responses (to copy or
not to copy) each time the adult performs an action. Children who mistakenly
imitate a ‘do this’ action probably intend to make a choice when they start the
game, but quickly forget and end up simply executing a single action sequence
(copying every move).
Motor cortex
Premotor
cortex
Prefrontal
cortex
BOX 2
Phineas Gage
In 1848 Phineas Gage, a railway worker in the USA, was using an iron rod to tamp down
gunpowder for blasting when a spark caused an explosion and the rod was propelled
straight up through one eye socket and both left and right frontal lobes (see Figure 2).
Initial interest was focused on the remarkable fact that he had survived this injury at all,
and the wider implications of the case were not appreciated until a follow-up report 20
years later. In this report, Harlow drew attention to the fact that whilst Gage had
indeed recovered from the accident with his speech, learning and memory intact, he
5 EXECUTIVE FUNCTIONS IN CHILDHOOD: DEVELOPMENT AND DISORDER 213
was left with such a profound impairment of social behaviour and personality that
according to friends and co-workers ‘he was no longer Gage’. Changeux quotes from
Harlow’s account of Gage:
He is fitful, irreverent, indulging at times in the grossest profanity ... impatient
of restraint or advice when it conflicts with his desires ... capricious and
vacillating, devising many plans of future operations, which are no sooner
arranged than they are abandoned in favor of others.
Cognitive flexibility
Cognitive flexibility refers to the aspect of executive function that enables people
to think and behave appropriately according to the changing needs of a complex
environment, and in line with their plans and goals. One of the hallmarks of
mature human behaviour is its flexibility and responsiveness to the constantly
changing features and demands of the environment, particularly of the social
214 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
RESEARCH SUMMARY 1
t
Activity 2 The Tower of London task
Allow about This activity gives you practical experience of doing the Tower of London task and illustrates how it
5 minutes measures planning.
Look at Figure 3. Moving one ball at a time, what is the minimum number of moves that is
needed to change the pattern of balls in the ‘initial position’ into the pattern shown as ‘goal
position no. 1’? Starting from the initial position in each case, work out the minimum number
of moves needed for goal positions no. 2 and no. 3.
Comment
Turn to the end of the chapter to find out the answers. The real task is somewhat easier to do
because it involves real pegs and balls that can be moved around (or ones that can be moved
around on a computer screen). Using static pictures, as you have been, places an additional
load on ‘working memory’, because for every move that you make you must actively hold in
your memory the position to which each ball has been moved.
s
As you can see from Activity 2, working memory is related to planning in this
view of executive function. Working memory refers to that aspect of remembering
that involves holding things actively in mind in support of ongoing plans of
action. For example, in the previous example of conversing with someone,
working memory would be used to maintain a representation of what the other
person is saying, or has just said, while planning a response to them. Working
memory is in itself a huge area of psychological research, which is beyond the
scope of this chapter. It is important to recognize, however, that it can be
regarded as a component of executive function.
t
Activity 3 The richness of the sensory world
Allow about This activity encourages you to reflect on the complexity of your everyday environment.
5 minutes
Consider your sensory world. Look around the environment you are in. Try to observe
everything you can see. Then sit back, shut your eyes and listen to everything that you can hear.
After that, focus on what is touching you. Attend to each part of your body in turn. What can
you feel? It takes a few minutes to get the most out of this activity, so do not hurry it.
Comment
The richness and complexity of the assaults on your senses are considerable. However, it is
hard to experience this complexity to its full extent because the mature human mind is so
expert at filtering out irrelevance, and building stimuli into simpler, more meaningful (and
organized) patterns. If this ‘filtering’ did not take place, and you gave equal weight to all
incoming sensory information, it would be impossible to behave in anything but a chaotic
manner.
s
Your current plan of action probably revolves around reading this section of the
chapter. In order to do this successfully you have to be able to ignore most of the
218 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
sensory world around you. You have to be able to prioritize the meaningful
stimuli – the print that makes up words on the page – and inhibit responses to
irrelevant stimuli, in order to enact your plan of reading the current paragraph,
and to achieve the goal of finishing it. What if you were unable to do this? What if
you could not give the words on the page any greater priority in cognition than
the grain of the wood on the table on which your book is resting? What if you
were unable to inhibit responses to the sound of the computer fan as it whirrs in
the background, or the feel of your feet on the carpet?
If you were unable to inhibit responses to stimuli that do not relate to the task
that you have planned to do, then it would probably be impossible to complete it
and achieve your goal. You would be drawn from one stimulus to another, in a
haphazard fashion, and it would be impossible to undertake any coherently
organized action. This is a rather extreme way of conveying the point, but young
children and people with executive function related disorders of inhibitory
control do have difficulty in prioritizing their response to task-related stimuli, and
do have difficulty in inhibiting responses to what are referred to as ‘prepotent’
stimuli.
A prepotent stimulus is a stimulus that draws a person’s attention towards it,
and which seems to cause the person to behave in a particular way (the prepotent
response). Prepotency is a very important feature of effective everyday
functioning. It is to be hoped, for example, that a red traffic light would draw a
driver’s attention towards it, and cause the driver to behave in a certain way. The
sight, smell and feel of the mother’s breast are the most likely prepotent stimuli
for the young breast-feeding infant.
In the course of typical development it is possible to observe infants and young
children being distracted by inappropriate prepotent stimuli. By ‘inappropriate’
we mean stimuli that are nothing to do with the child’s current plan of action. For
example, one might observe an 8-month-old infant catch sight of a toy on the
other side of the room and begin crawling towards it. It is clear to an observer that
they are enacting a plan to get the toy, but halfway across the room the infant
notices a scrap of paper on the floor. This seems to ‘capture’ their behaviour and
their attention. They pick it up, sit down and inspect it. The original plan is now
lost and they have been catapulted onto another stream of behaviour, which
might involve another plan, which might itself get interrupted by another
prepotent stimulus, and so on and so forth. This executive function analysis of a
familiar scene offers one explanation of why infant behaviour sometimes appears
somewhat haphazard and disorganized to an adult onlooker – according to this
view it is because executive functions are as yet undeveloped.
One aspect of child development that psychologists have become interested in,
then, is the way in which children develop an ability to inhibit responses to
stimuli that are nothing to do with their current plan of action. Put another way,
this amounts to an ability to prioritize responses to task-relevant (as opposed to
task-irrelevant) stimuli. When children begin to be able to do this, their behaviour
becomes less haphazard, and progressively more strategic and organized.
5 EXECUTIVE FUNCTIONS IN CHILDHOOD: DEVELOPMENT AND DISORDER 219
t
Activity 4 The Stroop task
Allow about This activity allows you to do the Stroop task and gives you a way of experiencing prepotency directly.
15 minutes
Get three coloured pens – perhaps a blue one, a red one and a black one. Write down a list
of fifteen words as follows (using the names of the colours that you have chosen). Note that
each word is written in the wrong colour.
BLUE (write in red ink)
RED (write in black ink)
BLACK (write in blue ink)
RED (write in blue ink)
RED (continue writing each word in the ‘wrong’ colour, sometimes writing ‘RED’ in
blue, sometimes in black)
BLUE
BLACK
BLUE
RED
BLACK
BLACK
BLUE
RED
BLUE
BLACK
Now, out loud, and working as accurately and as a quickly as you can, call out the colour of
the ink in which each word is written. What happens?
Comment
You should find that after a few words you start to get confused, wanting to call out the word
that you are reading, rather than the colour of the ink in which it is written. The meaning
associated with the word is acting as a powerful prepotent stimulus. You have to inhibit
everything you have learned about words and their meaning in order to call out the colour of
the ink. You could take two measures of performance from this, each of which would give
some information about inhibitory control: speed of completion, and number of errors made.
Why do you think that this task is not suitable for children of, say, 3 years of age?
s
The Stroop task tends to be used with older children and adults because of the
demands it makes on literacy skills, which in themselves are not a component of
220 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
BOX 3
The Handgame
The Handgame is a task that has the same basic structure as the Stroop task from
Activity 4. The child must inhibit a prepotent response in order to execute a rule-
guided action. Specifically, the child is first asked to imitate two hand actions (making a
fist and pointing a finger). Then, in the conflict condition, children must make the
opposite responses (making a fist when the experimenter points their finger and vice
versa). This involves:
. inhibiting the prepotent response to imitate; and
. performing an action guided by the rule ‘do the opposite of what the experi-
menter is doing’.
The measure of executive function that this task provides is the number of errors in the
conflict condition. The task is based on work by Luria and has been used with pre-
school children. Other variants of this task include Luria’s Knock/Tap game (in which
the child must knock when the experimenter taps the table and vice versa), the
Opposite Worlds task (in which school-aged children are asked to say ‘one’ when they
see a ‘2’ and to say ‘two’ when they see a ‘1’), and the Day/Night Stroop task (Gerstadt
et al., 1994; see Figure 4) in which children are instructed to say the word ‘day’ when
shown a line drawing of the moon and stars, and ‘night’ when shown a line drawing of
the sun. You should be able to see how these tasks relate conceptually to the Stroop
task, and how they are better suited to minds that are not yet at the stage of having
overlearned (automated) the ability to read words.
RESEARCH SUMMARY 2
BOX 4
Target card
Subject tries …
Shape Wrong
Colour Wrong
Number Right!
n
As executive function develops, so children’s abilities to learn new skills improve,
and they are increasingly able to behave in a planned, strategic and organized
manner. They are able to stay ‘on-task’ longer. They are able, when necessary, to
override habitual responses to prepotent stimuli. They become more skilled and
flexible in ‘orchestrating’ elements of their thinking and behaviour, and they are
able to engage in increasingly sophisticated planning and decision making.
Inhibitory control is only one component of this developmental trajectory, but it is
of fundamental importance. One way to assess this importance is to look at the
implications for child development of a failure to develop typical levels of
inhibitory control. This is the topic of Section 4.
5 EXECUTIVE FUNCTIONS IN CHILDHOOD: DEVELOPMENT AND DISORDER 225
4 Executive dysfunction
This section looks at executive function from a different perspective. What might
the consequences be if the development of executive function were impaired?
This is an important question to ask for two reasons. The first of these is to
understand the experiences and needs of those children and adults who are
directly affected. The second reason is that investigating the dysfunction of
cognitive processes is one way of increasing psychological knowledge about how
the processes work, and of finding out about what contribution they make in
typical development. We will focus on problems in the development of inhibitory
control, but you should bear in mind that executive dysfunction can also take the
form of cognitive inflexibility, planning difficulties, and problems associated with
working memory.
t
Activity 5 Implications of under-developed inhibitory control
Allow about This activity encourages you to consider the developmental consequences of poor inhibitory control.
10 minutes
What would you predict to be the implications for a child of a failure to develop typical levels
of inhibitory control? Note down possible consequences of under-developed inhibitory
control with respect to cognitive, social and emotional development.
Comment
If you were to compare a group of children with poor inhibitory control with a group of
typically developing children of the same ages, you might expect the former group to be, on
average:
. more distractible;
. less able to undertake tasks and learn skills that require sustained levels of concentration;
. more socially awkward (perhaps saying things at inappropriate moments; ‘blurting’ out);
226 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
5 Conclusion
The development of executive function is a process by which children come to be
able to act effectively in a complex, and not entirely predictable world. It is about
how children learn to organize – or ‘orchestrate’ – the cognitive and behavioural
skills that they are developing, in order to be able to respond flexibly to the
demands of the world around them, in line with their own plans and goals. As
you have seen, executive function is particularly strongly associated with the early
stages of learning new skills, and is involved in the effortful production of
controlled (as opposed to automatic, or over learned) actions. As such, executive
function depends on high-level cognitive processes. You will read about another
aspect of high-level cognition in Chapter 6, which is about ‘theory of mind’.
Answers to Activity 2
Goal position 1: two moves.
You would move the grey ball from peg 2 to peg 3 and the blue ball from peg 1 to
peg 2.
Goal position 2: four moves.
You would move the grey ball from peg 2 to peg 3; the blue ball from peg 1 to peg 2;
the black ball from peg 1 to peg 2 and the grey ball from peg 3 to peg 1.
Goal position 3: five moves.
You would move the blue ball from peg 1 to peg 2; the black ball from peg 1 to peg 3;
the blue ball from peg 2 to peg 1; the grey ball from peg 2 to peg 1 and the black ball
from peg 3 to peg 1.
References
Baker, L. and Brown, A. L. (1984) ‘Metacognitive skills and reading’, in Pearson, P.
D. (ed.) Handbook of Reading Research, pp. 353–94, White Plains, NY, Longman.
Barkley, R. A. (1997) ‘Behavioral inhibition, sustained attention and executive
functions: constructing a unified theory of ADHD’, Psychological Bulletin, vol. 121,
pp. 65–94.
Brophy, M., Taylor, E. and Hughes, C. (2002) ‘To go or not to go: inhibitory
control in ‘hard-to-manage’ children’, Infant and Child Development, Special Issue
on Executive Functions and Development, vol. 11, pp. 125–40.
Brown, T. E. (1999) ‘Does ADHD diagnosis require impulsivity-hyperactivity?: a
response to Gordon & Barkley’, ADHD Report, vol. 7, pp. 1–7.
5 EXECUTIVE FUNCTIONS IN CHILDHOOD: DEVELOPMENT AND DISORDER 229
Contents
1 Introduction 233
1.1 The importance of understanding minds 233
1.2 Understanding minds and egocentrism 234
6 Conclusion 256
References 256
6 UNDERSTANDING MINDS 233
Learning o u t c o m e s
After you have studied this chapter you should be able to:
1 understand and define what is meant by the term ‘theory of mind’;
2 understand the rationale underlying the tests used to establish whether
young children have developed the ability to reflect on another person’s
mind;
3 recognize what a ‘false belief’ is and why it is important;
4 understand the difference between first-order and second-order theory of
mind;
5 appreciate that the development of a theory of mind allows other skills to
develop;
6 identify the relationship between skills which develop early in infancy and
a subsequent appreciation of minds;
7 understand the relationship between children’s cognitive development and
their social environment;
8 understand the relationship between children’s communication and the
development of their theory of mind.
1 Introduction
1.1 The importance of understanding minds
You are gazing out of the window and see ten people standing in a line on the
pavement by the side of a road. Most of them are facing in the same direction,
looking towards the oncoming traffic. Beside the first person in the line is a pole
with a sign at the top. What are they doing there?
You probably recognize, instantly, that this is a queue at a bus stop. In doing so,
you identify that the people standing in the queue intend to catch a bus, believe
that a bus will arrive and hope that it will arrive on time. In making that
assessment you are demonstrating your understanding that other people – here
the ones in the queue – have ‘minds’, and that what goes on in their minds
dictates their behaviour. Without such an assumption you would only see these
people standing, pointlessly, in a line. There would be no apparent reason for
their behaviour because you would not appreciate their knowledge, desires,
hopes and intentions; this, and all other human behaviour, would appear similarly
unsystematic and meaningless.
As adults we appreciate that other people have minds like our own and that
these contain both information and desires, intentions, hopes and beliefs
(sometimes referred to as ‘mental states’) which inform and determine how they
behave. It also follows that we appreciate that other people have the same
234 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Psychologists’ interest in the development of theory of mind skills began, not with
human beings, but with research on chimpanzees. In a study which has since
become a classic (Premack and Woodruff, 1978), an adult chimpanzee called
Sarah was shown a series of videotaped scenes in which a human actor struggled
with problems of different kinds. Some of the problems appeared simple, such as
trying to get hold of bananas that were out of reach. Others were more complex,
such as trying to escape from a locked cage. For each scene Sarah was presented
with two photographs that showed different actions, only one of which was an
effective solution to the problem. So, for the scene in which the actor was locked
in a cage, Sarah was presented with a photograph of a key and a photograph of a
solution to a different problem, for example, a stick to get the out-of-reach
bananas. On the overwhelming majority of occasions (21 out of the 24) Sarah
selected the ‘correct’ solution. Premack and Woodruff concluded that the
chimpanzee ‘understood’ the actor’s purpose and therefore ‘understood’ the
contents of the actor’s mind.
t
Activity 1 Understanding Sarah’s mind
Allow about This activity will help you identify possible explanations for Sarah’s behaviour.
10 minutes
From what you have just read of the experiment involving Sarah, do you think the results
indicated that the chimpanzee understood the actor’s purpose and the contents of the
actor’s mind? What other interpretation might there be?
Comment
The philosopher Daniel Dennett (1978) has provided an important line of criticism of
Premack and Woodruff’s conclusions. Dennett suggested that their evidence did not show
that chimpanzees have an understanding of other people’s intentions, desires or knowledge.
Sarah could have shown that level of success on the task without reflecting on the actor’s
mental states at all. Instead, she could have solved the problem by drawing on her knowledge
of the associations between objects in the real world, what is sometimes referred to as the
external contingencies of objects. So, for example, on seeing the actor in a cage, rather than
impute the actor’s mental state, i.e. her desire and intention to escape from the cage, Sarah
could see the cage and match this to its common partner, a key, as shown in one of the
solution photographs.
s
236 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
0% 57% 92%
Extensive research on the development of theory of mind has been carried out
since Wimmer and Perner’s study. Improvements have been made to the tasks,
such as reducing the linguistic demands (for example, using simpler test
questions) which has resulted in enhanced performance particularly for younger
children. However, regardless of these modifications, the age of 4 years still
remains as the point at which the critical change in children’s understanding of
false beliefs appears to occur. Before 4 years, children rely on the current state of
reality to answer questions in the unexpected transfer task. From about 4 years
on, children begin to hold representations of people’s mental states, reflecting on
the belief of the story character and recognizing that that belief is false.
6 UNDERSTANDING MINDS 237
BOX 1
5
Figure 1
The unexpected
transfer task
(adapted from Frith,
1989, p. 160).
n
238 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Much of the research on the development of theory of mind has been based on
samples of children in Europe and North America. How do children in contrasting
cultures fare on the unexpected transfer task? Avis and Harris (1991) investigated
this with children of the Baka tribe, who live in the rainforests of south-east
Cameroon. The Baka are non-literate pygmies who have a hunter-gatherer
lifestyle with different values from those which predominate in Western societies.
The researchers acted out a version of the unexpected transfer task, adapted to be
more appropriate to the Baka children. In this version a member of the group
cooks some mangoes and leaves them in a cooking pot. While the cook is absent
the mangoes are transferred to a second, closed pot. The children were then
asked where the cook would look for his mangoes when he returned. The 5 and
6 year olds judged that the cook would look for the mangoes in the cooking pot.
Younger children made reality-based errors, saying that the cook would look for
them in the closed pot. Avis and Harris’s results are very similar to those found by
Wimmer and Perner and other researchers, and they suggest that the
development of an understanding of false beliefs is not specific to one culture and
that important changes occur at a similar age for different children around the
world.
BOX 2
1 2 3
4 5
Gopnik and Astington (1988) found that children older than about 4 years were able to
appreciate that Sooty would hold a false belief about the contents of the tube – that it
contained sweets. However, younger children who are reality-biased, said that Sooty
would say there were pencils in the tube. Ingeniously, the researchers then asked the
children about their own, previous false belief, ‘When you first saw this tube and it was
all shut up like this, what did you think was inside?’ (Frame 5). Again, children of 4 years
or older were able to reflect on their own mental states, and even though they now
knew the tube contained pencils, answered that they had previously believed that it
contained sweets. By contrast, the younger children stated that originally, before the
240 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
tube had been opened, they had believed that it contained pencils. These findings
suggest that as well as being unable to reflect on other people’s false beliefs, young
children remain reality-biased and are unable to reflect on their own, previous false
beliefs.
As is often the case with research of this sort, questions arise as to whether the results
might be affected by the experimental procedures or other factors. So, further studies
using the Smarties task have concluded that the young children’s responses are not
attributable to difficulties in understanding the specific moment of time to which the
question refers. This was established by asking, ‘What did you think was in the box
before I took the top off?’ (Lewis and Osborne, 1990). Nor were the results due to the
children’s embarrassment at admitting their original ignorance. Wimmer and Hartl
(1991) introduced children to a silly puppet, who always made mistakes. Adding this
element to the design of the experiment enabled the researchers to work out whether
children’s incorrect answers were caused by their embarrassment at admitting their
previous ignorance, rather than lack of theory of mind skills. When the silly puppet was
introduced, if children were simply embarrassed about getting the question wrong,
they would fail questions about their own previous false belief by stating that they
always knew the contents of the tube. However, they would be expected to pass
questions regarding the false belief of the silly puppet, and to state that the puppet
previously believed that the tube contained Smarties since they would not be expected
to feel embarrassed about the silly puppet’s false belief. However, young children
without theory of mind skills were just as likely to get the false belief question wrong,
irrespective of whether they were asked about their own previous false belief, or the
false belief of the puppet.
n
The overall conclusion appears to confirm that before the age of about 4 years
children rely on reality to predict their own and other people’s behaviour.
However, from about 4 years onwards, they begin to appreciate that a
psychological level of information (the beliefs that people hold about the world)
dictates people’s behaviour and that sometimes these beliefs are wrong.
6 UNDERSTANDING MINDS 241
Second-order Thus far, this chapter has been about first-order theory of mind, that is the
theory of mind ability to appreciate and reflect on the contents of another person’s mind. Second-
The ability to
attribute beliefs
order theory of mind is the ability to attribute beliefs about beliefs, or beliefs about
about beliefs or intentions.
beliefs about Figure 3 illustrates the difference between first-order and second-order theory
intentions.
of mind skills. The boy on the left is simply thinking about taking the girl’s apple;
he is not using theory of mind skills. The girl, by contrast, is using first-order
theory of mind skills as she is thinking about the boy on the left’s desire and
intention (mental states) to steal her apple. Unfortunately for her, the boy on the
right is using second-order theory of mind skills. He realizes that he can steal the
girl’s apple, for he believes (correctly, as it happens) that she is preoccupied with
thinking about the other boy’s intention to steal her apple.
Second-order theory of mind skills are believed to develop between the ages of
6 and 8 years.
A lie or a joke?
Do children need to have attained second-order theory of mind skills in order to
be able to distinguish lies from jokes? This question was answered by Sullivan
et al. (1995) in a study in which 48 children, whose ages ranged from 5 to 9 years,
were told four brief stories. Two of the stories assessed the children’s second-
order ignorance (for example, does John know that Mary knows X?) and second-
order belief understanding (for example, what does John think Mary thinks?). The
6 UNDERSTANDING MINDS 243
other two stories assessed the children’s ability to discriminate between a lie and a
joke. In one story a boy who did not clean his room lies to his mother by saying, ‘I
did a really good job cleaning my room’. In a second story a boy does not finish
eating his dinner and jokingly says to his mother, while they both sit at the table, ‘I
did a really good job finishing my peas’. At the end of each story the participants
were asked whether the story character was lying or joking. The results showed
that children typically were able to distinguish a lie from a joke only after they
could attribute second-order ignorance, but before they could attribute second-
order false belief. Therefore, the skill that appears to be crucial when
distinguishing a joke from a lie is the ability to appreciate and reflect on what
different people know or do not know, rather than being able to reflect on the
knowledge that people have about other people’s minds.
Such conclusions seem to make sense, since jokes differ from lies in terms of
the knowledge of the listener and the speaker. With jokes, the speaker knows that
the listener also knows the truth. In the example, both the boy and his mother
could see the peas remaining on his plate when he said, ‘I did a really good job
finishing my peas’. However, with lies the speaker and listener have different
knowledge and the speaker is aware of this difference; the boy who did not clean
his room knows that his mother does not know he has not cleaned his room when
he says, ‘I did a really good job cleaning my room’. Therefore, understanding lies
and jokes requires an understanding of another person’s knowledge or ignorance
of facts rather than an understanding of someone’s beliefs about beliefs.
t
Activity 2 Cause and effect
Allow about This activity will help you to appreciate some of the difficulties in establishing cause and effect.
5 minutes
Would you say that Astington and Jenkins’ conclusion was justified?
Comment
One problem with research that shows a correlation between two variables is that it is
difficult, if not impossible, to identify with any certainty whether an increase in one variable
causes an increase in the other. In this case, do enhanced theory of mind skills lead to an
improvement in social interactions, or are children who are more socially skilled more likely to
develop a theory of mind before those who are less socially skilled?
s
In a subsequent study (Jenkins and Astington, 2000) the researchers sought to
shed light on the direction of causality by using a different research design. They
carried out a longitudinal study that tested 20 children aged 3 and 4 on three
occasions in the course of approximately 7 months. On each occasion the
children were presented with a series of false belief understanding tasks, and they
were video-recorded during play with a friend. The play sessions generated
measures of the amount of pretend play, joint proposals, and explicit role
assignment (where the child assigns a pretend role to himself or herself, or to
another child, by suggesting, for example, ‘Let’s be firegirls now’). It was found
that the children’s performance on the theory of mind understanding tests taken
during the first recording session predicted joint proposals and role assignment
during play in the second session. However, there was no evidence that social
behaviours predicted children’s theory of mind. Therefore, it appears that the
development of theory of mind skills brings with it a change in the quality of
children’s interactions.
3.3 Bullying
The previous example of the relationship between theory of mind skills and
children’s interactions during play is positive in every respect. It would be
reassuring to think that the ability to appreciate the feelings and thoughts of
others would make children more altruistic and sensitive to those around them.
However, this is not always the case. The development of a theory of mind may
also bring with it an improved capacity for children to bully others and also to lie
and deceive. To be sure, children may lie, deceive or bully before they
understand mental states, but once they have developed a theory of mind they
become more sophisticated in the strategies that they can use. By being able to
appreciate another person’s mind, bullies are better able to identify their victim’s
weaknesses and vulnerabilities and use this understanding to refine their bullying
strategies.
By contrast with the popular stereotype of an ‘oafish’ bully lacking in social
skills and understanding, bullies may be manipulative experts in social situations,
organizing gangs and using subtle, indirect methods to bully. One study
6 UNDERSTANDING MINDS 245
t
Activity 3 Analysing young children’s talk
Allow about This activity asks you to analyse short extracts of children’s talk to identify patterns in their use of
10 minutes desire-based and belief-based terms.
Look at the following six examples from four of the children in the study by Wellman and
Bartsch and identify which contain desire-based terms and which belief-based terms. Note
the ages of the children at the time of the recording: does any pattern emerge?
Comment
Extracts 1 and 3 contain desire-based terms – ‘You don’t want a spoon’ and ‘I want a band-
aid’. Extracts 2, 4, 5 and 6 contain belief-based terms – ‘so de mailman can know where I are’, ‘I
know he could rip it’, ‘I thought I would’ and ‘You don’t know where the pieces go’. The two
examples of desire-based terms come from Adam and Ross when they were younger by
several months than in the other extracts.
s
This hint of an age trend across these examples was borne out by the fuller
analysis of the data. Wellman and Bartsch found that genuine reference to the
subjective mental state of desire occurred around 2 years of age. At this age, not
only do children refer to their own desires but they are also able to refer to other
people’s desires, as Adam does in Extract 1. This use is distinct from some early
uses of the word ‘want’ – ‘I want this’, ‘I don’t want to’– that may not truly reflect a
child’s understanding of mind, but rather may simply be a tool they have learned
to use in order to obtain desired objects or not to have to undertake undesired
activities. These children may have no psychological understanding of the
psychological state ‘desire’, but may simply be using ‘want’ because they have
built up a conditioned response which achieves desired goals. The use of
‘genuine reference’ at the beginning of this paragraph distinguishes those ‘learned
responses’ from uses which indicate a true understanding of the mental state of
desire.
The Sally/Anne and Smarties tasks provide ways of establishing whether
children really do understand the psychological consequences of mental states
rather than simply rote learning mental state terms. Just as with analyses of
everyday language, studies employing these tasks show that children pass tests
for an understanding of desires before they pass tests for false belief
understanding. Research summary 1 provides an example of one such study by
Repacholi and Gopnik (1997).
248 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
RESEARCH SUMMARY 1
n
As can be seen from these examples, both naturalistic observations and
experimental paradigms have shown that there is a critical shift at about 18
months as children begin to appreciate that different people can have different
desires. It is likely that this understanding is an important stepping stone to
understanding false beliefs.
RESEARCH SUMMARY 2
Understanding intentions
The example of the bus queue, which began this chapter, illustrated that in order to
understand the intentions of another person it is necessary to impute desires and
goal states to their actions – the people are standing, waiting, because they are
hoping that a bus will arrive. In order to understand that people have minds and
that their minds dictate their behaviour, it is necessary to appreciate that behaviours
are purposeful. If behaviours are not seen as purposeful then there would be no
need to try to explain them because they would appear to be unsystematic and
therefore meaningless. When do children show evidence of beginning to
understand the intentions behind people’s actions? Research summary 3 describes
a piece of research by Meltzoff (1995) which addresses this question.
250 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
RESEARCH SUMMARY 3
(a)
(b)
Figure 4 (a) The human experimenter attempts to pull the dumbbells apart but his fingers slip off the
end. (b) The mechanical ‘arms’ and ‘fingers’ located in front of the human experimenter fail in a similar
fashion to pull the dumbbells apart (from Meltzoff, 1995, p. 844).
n
6 UNDERSTANDING MINDS 251
t
Activity 4 How do we learn to think about others?
Allow about This activity will help you think about experiences that might encourage an understanding of other
10 minutes people’s minds.
Think of some experiences within a young child’s day-to-day environment that might
encourage the development of an understanding of other people’s minds. What might assist
them to ‘stand in another person’s shoes’? And how, in turn, might these different factors
affect children’s cognitive development?
Note down your thoughts and review them as you read the rest of this section.
s
252 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
t
Activity 5 T h e s i g n i fi c a n c e o f s i b l i n g s
Allow about This activity will give you further experience of analysing and interpreting data.
5 minutes
Study the scores shown in Table 2. What conclusions can you draw from these results?
Comment
Where children have no siblings there appears to be a relationship between their level of
language ability and their competence at theory of mind tests – those with lower language
ability do less well. However, when children have two siblings there is a clear indication that
this difference is reduced – the presence of siblings appears to compensate for slower
language development in developing false belief understanding.
s
If siblings make a difference in this way it would seem plausible that their age
might be of significance. When Ruffman et al. (1998) examined the results of 444
English and Japanese children whose ages ranged from 3 years and 1 month to 6
6 UNDERSTANDING MINDS 253
years and 11 months, they found that the number of theory of mind tests that they
passed increased with the number of older siblings a child had. However, there
was no connection between children’s success on the tests and either the number
of younger siblings or the gender of the siblings (whether older or younger). How
might the older siblings facilitate a child’s theory of mind development? Ruffman
et al. point to activities such as pretend play, deception, teasing and talk about
feelings, all of which would feature less in interactions with younger siblings.
5.2 Communication
The data in Table 3 indicate that the children in the late-signing group showed
less well-developed theory of mind skills than either the native-signers (p <0.001)
or a sample of 4-year-old hearing children (p < 0.001). It would appear that access
to early conversation – through whatever medium – is an important factor in the
development of a theory of mind.
5.3 Gender
There is much research to show that the gender of a child can affect the types of
interaction they have with their environment. For example, mothers tend to talk
about emotions more to their 2-year-old daughters than to their 2-year-old sons
(Dunn et al., 1987). Older siblings tend to mention feeling states more frequently to
girls than boys (Brown et al., 1996). Girls also tend to use words which signify
emotional states earlier and more frequently than boys (Cervantes and Callanan,
1998). This suggests that girls might succeed on tests of false belief understanding
6 UNDERSTANDING MINDS 255
before boys. Evidence from nearly 1,500 children, aged between 2 years and 4
months and 6 years and 3 months, indeed showed a slight advantage for girls on
false belief understanding tests (Charman et al., 2002). However, this was only so
for children younger than 4 years and 8 months, not for children older than that,
and in any event the researchers concluded that if there is an age-specific
advantage for girls in the acquisition of false belief understanding it is only a very
weak effect.
6 Conclusion
At the beginning of this chapter you were asked to imagine what the world would
be like if we were not able to appreciate the minds of other people. It should now
be clear that a theory of mind is an exceptionally important ability that allows us
to perform complex behaviours and draws upon some fundamentally important
skills. Before the age of about 4 years, children do not fully appreciate other
people’s mental states and so they are unable to use this rich source of
psychological information to predict a person’s behaviour. Young children
believe that people behave according to how the world really is, rather than how
each person believes the world to be. However, research using tests of false belief
understanding has shown that, at about 4 years, children begin to appreciate the
mental states of others. This change in cognitive skills appears not to be restricted
to Western cultures.
Having a new realm of information upon which to draw allows children to
develop many other skills, including improved interactions with others and the
appreciation of intentions in lies, jokes, and ironic or sarcastic comments. The
development of a theory of mind also provides children with the tools to bully
and lie more effectively.
The development of theory of mind skills draws upon many cognitive abilities,
including joint attention and the understanding of desires and intentions. Being
able to share one’s perceptual focus with another person and being able to
appreciate the desires and intentions of others by watching their actions appear to
underpin the development of an understanding of false beliefs. Furthermore, the
environment within which a child lives can also affect theory of mind
development. The interactions that children have with their siblings, friends,
parents and other adults affect how they see other people. From about 4 years of
age a child sees other individuals not just as people, but as people who have
minds of their own.
References
Astington, J. and Jenkins, J. (1995) ‘Theory of mind development and social
understanding’, Cognition and Emotion, vol. 9, pp. 151–65.
Avis, M. and Harris, P. (1991) ‘Belief-desire reasoning among Baka children:
evidence for a universal conception of mind’, Child Development, vol. 62,
pp. 460–7.
Brown, J., Donelan-McCall, N. and Dunn, J. (1996) ‘Why talk about mental states?
The significance of children’s conversations with friends, siblings and mothers’,
Child Development, vol. 67, pp. 836–49.
Cervantes, C. and Callanan, M. (1998) ‘Labels and explanations in mother–child
emotion talk: age and gender differentiation’, Developmental Psychology, vol. 34,
pp. 88–98.
6 UNDERSTANDING MINDS 257
Charman, T., Baron-Cohen, S., Swettenham, J., et al. (2000) ‘Testing joint
attention, imitation, and play as infancy precursors to language and theory of
mind’, Cognitive Development, vol. 15, pp. 481–98.
Charman, T., Ruffman, T. and Clements, W. (2002) ‘Is there a gender difference in
false belief development?’, Social Development, vol. 11, pp. 1–10.
Dennett, D. (1978) ‘Beliefs about beliefs’, Behavioral and Brain Sciences, vol. 1,
pp. 568–70.
Donaldson, M. (1978) Children’s Minds, Glasgow, Fontana.
Dunn, J., Bretherton, I. and Munn, P. (1987) ‘Conversations about feeling states
between mothers and their young children’, Developmental Psychology, vol. 23,
pp. 132–9.
Frith, U. (1989) Autism: explaining the enigma, Oxford, Blackwell.
Gopnik, A. and Astington, J. (1988) ‘Children’s understanding of representational
change and its relation to the understanding of false belief and the appearance-
reality distinction’, Child Development, vol. 59, pp. 26–37.
Jenkins, J. and Astington, J. (1996) ‘Cognitive factors and family structure
associated with theory of mind development in young children’, Developmental
Psychology, vol. 32, pp. 70–8.
Jenkins, J. and Astington, J. (2000) ‘Theory of mind and social behavior: causal
models tested in a longitudinal study’, Merrill-Palmer-Quarterly, vol. 46,
pp. 203–20.
Lewis, C., Freeman, N., Kyriadidou, C., Maridaki-Kassotaki, K. and Berridge, D.
(1996) ‘Social influences on false belief access’, Child Development, vol. 67,
pp. 2930–47.
Lewis, C. and Osborn, A. (1990) ‘Three-year-olds’ problems with false belief:
conceptual deficit or linguistic artifact?’, Child Development, vol. 61, pp. 1514–19.
Meltzoff, A. N. (1995) ‘Understanding the intentions of others: re-enactment of
intended acts by 18-month-old children’, Developmental Psychology, vol. 31,
pp. 838–50.
Piaget, J. (1932) The Moral Judgement of the Child, New York, Harcourt Brace.
Piaget, J. and Inhelder, B. (1968) The Psychology of the Child, New York, Basic
Books.
Premack, D. and Woodruff, G. (1978) ‘Does the chimpanzee have a theory of
mind?’, Behavioral and Brain Sciences, vol. 1, pp. 515–26.
Repacholi, B. M. and Gopnik, A. (1997) ‘Early reasoning about desires: evidence
from 14- and 18-month-olds’, Developmental Psychology, vol. 33, pp. 12–21.
Ruffman, T., Perner, J., Naito, M., Parkin, L. and Clements, W. (1998) ‘Older (but
not younger) siblings facilitate false belief understanding’, Developmental
Psychology, vol. 34, pp. 161–74.
Ruffman, T., Perner, J. and Parkin, L. (1999) ‘How parenting style affects false
belief understanding’, Social Development, vol. 8, pp. 395–411.
258 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Sullivan, K., Winner, E. and Hopfield, N. (1995) ‘How children tell a lie from a
joke: the role of second-order mental state attributions’, British Journal of
Developmental Psychology, vol. 13, pp. 191–204.
Sutton, J., Smith, P. and Swettenham, J. (1999) ‘Social cognition and bullying:
social inadequacy or skilled manipulation?’, British Journal of Developmental
Psychology, vol. 17, pp. 435–50.
Wellman, H. and Bartsch, K. (1994) ‘Before belief: children’s early psychological
theory’, in Lewis, C. and Mitchell, P. (eds) Children’s Early Understanding of
Mind: origins and development, pp. 331–54, Hove, Lawrence Erlbaum Associates.
Whiten, A. (1991) Natural Theories of Mind: evolution, development and
simulation of everyday mindreading, Oxford, Blackwell.
Wimmer, H. and Hartl, M. (1991) ‘Against the Cartesian view on mind: young
children’s difficulty with own false beliefs’, British Journal of Developmental
Psychology, vol. 9, pp. 125–38.
Wimmer, H. and Perner, J. (1983) ‘Beliefs about beliefs: representation and
constraining function of wrong beliefs in young children’s understanding of
deception’, Cognition, vol. 13, pp. 103–28.
Woolfe, T., Want, S. C. and Siegal, M. (2002) ‘Signposts to development: theory of
mind in deaf children’, Child Development, vol. 73, pp. 768–78.
Chapter 7
Mathematical and scientific
thinking
Terezinha Nunes and Peter Bryant
Contents
1 Introduction 261
1.1 Cognition and language in context 261
1.2 What is the nature of mathematical and
scientific knowledge? 262
5 Conclusion 297
References 298
7 MATHEMATICAL AND SCIENTIFIC THINKING 261
Learning o u t c o m e s
After you have studied this chapter you should be able to:
1 understand the difference between generative and reproductive
knowledge;
2 distinguish between the ways in which children think in mathematics and
science;
3 understand the importance of situations, and not just of computations, in
problem solving in mathematics;
4 assess the impact of cultural settings on mathematical and scientific
thinking;
5 discuss possible connections between the development of intelligence and
learning science and mathematics;
6 compare different theories about how children progress in their
mathematical and scientific thinking and the implications of this analysis for
education.
1 Introduction
1.1 Cognition and language in context
So far in this book you have seen that as children gain an increased understanding
of their linguistic and physical environments they attempt to organize and
systematize this knowledge in a number of important ways. They develop explicit
categories that capture not just key characteristics of their environment, but also
the similarities and differences between aspects of it. They apply verbal labels to
them and to their own internal mental states. They learn how to use language to
communicate their ideas and desires, and through this communication they
acquire knowledge about how others think about and understand the world. This
in turn pushes their own understanding of their environment to new levels of
sophistication. They generate hypotheses about how language is formed, how the
world ‘works’, and how other people think.
At each stage children are moving towards the ability to represent their world in
symbolic ways. Words in language are used to represent real objects, events and
feelings, and enable the discussion of things that can only be imagined. Through
such discussions children begin to deal with unknown factors and hypothetical
situations. This is especially true when they begin to acquire mathematical and
scientific understanding of their environment.
This chapter will introduce you to the nature of mathematical and scientific
understanding. We will introduce the idea that mathematical and scientific
knowledge is not simply a ‘collection of facts’ but a ‘way of thinking’ and we will
present you with an overview of the psychological research that has attempted to
uncover how children come to acquire this way of thinking. What you will
discover is that children often rely on their everyday understanding of the world
to generate solutions to mathematical and scientific problems.
262 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
t
Activity 1 Learning to count in Japanese
Allow about This activity will help you to appreciate the difference between reproductive and generative learning.
5 minutes
Study the numbers below, which are in Japanese but written phonetically in English.
1 ichi 10 ju 20 ni ju
2 ni 11 ju ichi 21 ni ju ichi
3 san 12 ju ni 22 ni ju ni
4 shi 13 ju san
5 go 14 ju shi
6 roku
7 sichi
8 hachi
9 ku
How do you say 15 in Japanese? How do you say 43? How do you say 67? Do you think you
could count to 99? How did you get your answers?
Comment
Your performance in Activity 1 can show you that you know more facts about Japanese
numbers than you were taught. Learning mathematics means understanding a system that
goes beyond the examples you learn from. For this reason mathematics learning is said to be
‘generative’, that is, learning the system allows you to generate new facts that you were never
taught about. If mathematical and scientific knowledge were simply reproductive (the
repetition of what the person was taught), you could not have answered the questions in
Activity 1.
s
7 MATHEMATICAL AND SCIENTIFIC THINKING 263
t
Activity 2 Generating numbers
Allow about This activity encourages you to consider the need to look at learning processes (how children come to
5 minutes understand something) rather than just at the outcomes (what they are able to do).
Look at Figure 1. It contains numbers written by four children aged between 5 and 6 years of
age. They had not been taught how to write multi-digit numbers in school. However, they
see numbers around them and do their own analysis of how numbers should be written. Can
you figure out how each one generates the writing of numbers? Which children seem to
think similarly about writing numbers? What similarities and differences are there?
(9)
(5)
(25) (9)
(45)
(5)
(100) (25)
(40)
(123)
(100)
(109)
(103)
(1000)
(109)
(2000) (1000)
(1002)
(2009)
(1009)
(2568)
(1598)
one hundred
eight
two hundred
one hundred
and nine
five
one thousand
twenty-five
one thousand
and five
forty
two thousand
and fifty-three
Comment
In order to understand children’s learning of mathematics, it is necessary to analyse their
productions – the way they count, write numbers, solve problems and so on – and try to
figure out how they think. It is not enough to try to find out what number facts they know or
do not know. Consider first Alice’s production. Alice seems to have the idea that, for each
number word, you put down a digit. She does not know which digits she should put down for
some words – for example, twenty, hundred and thousand – so she uses one line for twenty
and two number-like shapes for hundred and thousand, respectively. Although not correct,
her writing of numbers is not random and can be understood if you crack her system. Luke
and Susanna seem to use similar systems. It is quite likely that they have learned how to write
two digit numbers so they get these right (if you disregard Luke’s inversion of the digit 2). For
three digit numbers, they have created a system: they write each number in sequence, as if
they wrote the words in sequence. Megan succeeds in keeping three digit numbers within the
conventional writing but finds it hard to do so with four digit numbers. For these three
children, number learning is not a simple sequence to be memorized, with the larger ones
being learned later: they succeed with 2,000, yet do not write smaller numbers correctly. They
have a system, but it happens not to be the one that adults use.
s
It may seem that this notion of generative knowledge applies more readily to
mathematical than to scientific knowledge. It could be argued that there are many
‘scientific facts’ about the world that children can be taught in school and they will
learn them without difficulty – and this is to some extent true. However, scientific
information is often considered to be something that is transmitted in school as
‘mere fact’, when this information is not only that. Often what people call
‘scientific fact’ is important because it reflects a way of thinking. Try Activity 3
now and think about facts and ways of thinking in science.
t
Activity 3 Facts and ideas
Allow about This activity will help you to consider the distinction between understanding a fact, and showing
10 minutes awareness of the underlying principles that explain that knowledge.
It has been known now for a long time that the world is round and not flat, as it was thought
in the past. This is a fact. But is this a ‘mere fact’ or is it a way of thinking? This is what
Nussbaum, Novack and their colleagues at the University of Cornell decided to figure out
(Nussbaum and Novack, 1976; Nussbaum, 1985).
Look at the picture of the world in Figure 2. It illustrates the idea that there are two girls, one
who lives at the North Pole and one who lives at the South Pole. They have two bottles each,
one with a cork in and one that is open. Their bottles are half filled with juice. Can you draw
in the juice in their bottles?
7 MATHEMATICAL AND SCIENTIFIC THINKING 265
Look at the second picture of the world in Figure 3. It shows three boys, one living at the
North Pole, one living on the Equator, and one living at the South Pole. They each have a ball
in their hand. What will happen to the ball if they drop it? Can you draw in the path the ball
will follow?
Figure 2 What will happen to the juice in Figure 3 What will happen to the ball if it is
the bottles? dropped?
Comment
As with mathematical knowledge, scientific knowledge is not the accumulation of facts but is
defined by ‘ways of thinking’. Most children in primary school will know and tell you (if asked)
that the world is round. Yet, their thinking may be governed by the idea that the world is flat,
with the sky above and the ground below. This is what their drawings of people on earth and
of rain clouds suggest (see Figure 4). Their answers to the questions in Activity 3 (Figure 5)
also indicate a way of thinking that is consistent with a flat world conception.
(a) (b)
Lucas Frances
Lucas (6 years) and Frances (8 years) were asked (following Nussbaum and colleagues) to
imagine that they were astronauts and were looking at the world from a distance. What would
the world look like? The answer to this question typically is ‘round’ or ‘like a ball’ or something
of the sort – and both Lucas and Frances answered in this way. They were then asked to draw
the world as it would be seen from the spacecraft. After this was accomplished, they were
asked to put in some people, some clouds and some rain in the picture. The children were
also asked to indicate the North and the South Pole as a preparation for the next activity.
Both drew a round world and both decided to put in some countries. Lucas put in a rather
good outline of South America and England plus what he thought was not a good outline of
Australia, but he wanted to have it in anyway (see Figure 4a). Frances put in some countries
but was not sure what they were supposed to look like (see Figure 4b). As you can see, apart
from these similarities, their drawings of people on earth, clouds and rain differ in significant
ways. For Lucas, each country had its own cloud ‘on top’ and its rain falling from top to
bottom. His round world contains people all standing in one direction with clouds above and
not all around. Even though 6-year-old children have been told that the world is round and
draw it round when asked to, they place people, clouds and rain in a way that is more fitting
with a belief in a flat world. They have learned the fact but not the way of thinking that goes
with it. For Frances, the conception of the round world prevails in this drawing; she also put in
some countries, each one with a rain cloud but the clouds are positioned all around the world.
INTERVIEWER TO FRANCES: ... Can you draw this for me?
FRANCES: Yes. [see (4b)] (Later, as she draws the South Pole) ... I used to think that people
would fall off the earth. I didn’t know why they wouldn’t fall off when they were
here.
INTERVIEWER: Do they fall off?
FRANCES: No (laughing).
INTERVIEWER: Why not?
FRANCES: My father told me. It’s because of gravity. (She draws the rain falling on the South
Pole and remarks that ‘Isn’t it funny to think of rain falling this way?’)
s
The drawings by the children are consistent with the solutions they give to the
problems posed in Activity 3 – presented in Figure 5. Lucas’s solution is typical for
children of his age. The drink in the uncorked bottles owned by the little boys at
the South Pole and at the Equator falls out of the bottles. The balls also fall the
same way as the liquid. Frances thought that the liquid would stay in the bottles
but was quite doubtful when giving her answer about the little girl’s liquid at the
South Pole. When asked about the ball, she commented it would go ‘right down
into space’. Frances seemed to alternate between a round and a flat world way of
thinking.
7 MATHEMATICAL AND SCIENTIFIC THINKING 267
So, what can be seen from this example is that knowledge of mathematics and
science is about knowing how to think appropriately about a problem, rather
than repeating facts that have been learned by rote. Moreover the example shows
that although children often know particular facts, they can be seen to apply their
knowledge in a way that betrays a lack of understanding about the underlying
‘laws’ that account for those facts.
We will now consider children’s development in understanding these laws in
relation to specific aspects of mathematical and scientific reasoning. We will turn
to mathematical reasoning first, and consider the question of whether children are
always susceptible to these kinds of errors, or whether mathematical
understanding is mediated by certain contextual factors.
Concrete
operations
coins would have increased the quantity. Through their own interactions with
The stage at
which children objects, by changing displays in several ways and checking the results of such
can think logically changes, they would later on come to understand the invariance of number.
about objects and The discovery of the invariance of number in spite of spatial displacements that
events in their
environment, and make sets look larger or smaller (around the age of 6 or 7) was considered so
represent them important by Piaget that it was taken as a mark of a new stage in children’s
symbolically. intellectual development: the stage of concrete operations.
However, they are
still unable to
This was not the end point of development, though, for at this stage children’s
generate mathematical understanding was, according to Piaget, restricted to one-variable
hypotheses problems for which addition and subtraction are sufficient. If 7-year-old children
systematically
have to solve a problem that requires establishing a proportional relationship
combining
possible different between two different variables, they are not able to cope. An example of
situations and proportional relations is given in Activity 4 below.
outcomes.
t
Activity 4 Proportional relationships
Allow about This activity will encourage you to test your understanding of the nature of proportional reasoning.
5 minutes
Here is an eel:
Comment
The information in the first two examples should have indicated to you that there is an
underlying ratio of eel length to amount of food of 1 cm: 2 g, and this should enable you to
work out that Ernie needs 10 g of food each day. In a study by Inhelder and Piaget (1958)
children were told that eels have to be fed amounts of food that vary with their size. They
270 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
were asked to look at some values in both variables – length of the eel and amount of food it
needed – and then say how much food an eel of a certain length should receive. Children at
age 7 normally suggested higher values for longer eels but did not derive the values
systematically. In contrast, older children (about 11–13 years of age) tried to establish a
relationship between the two variables and to use this relationship to deduce how much food
the longer eel would need.
s
Piaget and his colleagues (Inhelder and Piaget, 1958; Piaget et al., 1968)
considered the achievement of proportional reasoning as another landmark in the
development of intelligence. Proportional reasoning involves the recognition that
the relationship between two variables remains the same (invariant) although the
values in both variables are changed. This ability to establish an invariant relation
among relations is a higher order operation that, according to Piaget, indicates the
Formal achievement of a new level of thinking, that of formal operations.
operations
The stage at
which children 2.2 Limitations in Piaget’s ideas
are able to reason Although Piaget’s contribution to current approaches to the study of children’s
in an abstract way mathematical and scientific thinking is broadly recognized, some aspects of his
without reference
to concrete work have been criticized.
experience. They The most serious criticisms have been directed at his idea that children’s
can tackle intellectual structures determine how they think, above and beyond any other
problems in a
systematic and
influences. Much of the recent research on mathematical reasoning seems to
scientific manner suggest that both children and adults show different levels of success when
and are able to solving problems that involve the same intellectual structures but differ in other
generate
respects, such as the content of the problem, the particular mathematical
hypotheses about
the world based representation they are using, or the social situation in which they are engaged.
on their With respect to the content of the problem, for example, it is easy to imagine
accumulated that someone may grasp that the relationship between quantity purchased and
knowledge.
amount of money that has to be paid is proportional. In other words, the more
sweets you buy, the more you pay, and the amounts vary in a fixed ratio – for
example, 5 p per sweet. Yet, this same person may not realize that, if you enlarge a
rectangle, the ratio between length and width must be kept constant for the figure
to look similar – if the length is twice the width in the small rectangle, it must also
be twice the width in the larger one, otherwise the figures look different.
With respect to the type of mathematical representation, Nunes (1993) has
shown that pupils (of approximately 12–13 years) who were able to solve
problems with negative numbers orally made significantly more errors if they were
asked to write the information down before solving the problem. Although the
content of the problem and the social situation were the same, the written
representation had characteristics that confused the pupils when they were solving
the problems.
Lave (1988) showed that adults in California performed very differently when
solving the same type of proportion problems in different social situations. They
were much more successful in the supermarket than in a written test. Thus it now
seems clear that while intellectual structures may influence performance in
7 MATHEMATICAL AND SCIENTIFIC THINKING 271
problem solving, they do not directly determine what a person will actually do.
Other aspects of the problem situation need consideration.
Gérard Vergnaud, a French developmental psychologist who studied with
Piaget, synthesized these ideas in his theory, which is known as the theory of
conceptual fields. According to Vergnaud (1985), in order to analyse mathematical
concepts, the invariant properties of the concept must be considered, as Piaget
proposed, and also the situations that give meaning to the concept and the symbols
used in its representation. He further proposed that in mathematics one should
think not of isolated concepts, but of conceptual fields where the different concepts
are connected to the same core invariants, situations and symbolic representations.
For example, he distinguished between the conceptual field of additive reasoning
and the field of multiplicative reasoning. Although there are connections between
these two domains of reasoning, Vergnaud’s view was that the differences between
them are important enough to require that children’s understanding in these two
domains must be analysed separately. We will use additive reasoning as an
example to illustrate our discussion of mathematical development.
t
Activity 5 H o w d i f fi c u l t a r e t h e s e p r o b l e m s ?
Allow about This activity will encourage you to reflect a little on the nature of addition problems, and in particular
5 minutes what makes some addition tasks difficult for young children.
Read the arithmetic problems presented below. They have different levels of difficulty for
young children. Try to figure out which one is easier, which is more difficult and why.
Problem 1
John had some marbles. He played with a friend and won four marbles. Now he has nine.
How many did he have before the game?
Problem 2
Mary had nine sweets. She gave four to her sister. How many does she have now?
Problem 3
Paul had nine buttons in his pocket. His pocket had a hole and some fell out. Now he has four
buttons. How many buttons did he lose?
272 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
100
Comment
Look at Figure 7 and check whether your predictions 80
were correct.
Percentage correct
As you can see from Figure 7, the problem involving a 60
relation between addition and subtraction (1) was the
most difficult to solve. This shows the importance of
understanding the relation between addition and 40
0
Figure 7 Percentage of first graders who solved problems 1 2 3
1 to 3 correctly (data from Riley et al., 1983). Problem type
s
In view of the strong connection between addition and subtraction, it is now
considered more appropriate to discuss the development of additive reasoning as
a whole rather than the development of each of its parts (addition and
subtraction) in isolation. Much research has been carried out comparing the
difficulties that children face in solving addition and subtraction problems
(Carpenter and Moser, 1982; Riley et al., 1983). This research shows that children
may know how to solve a particular numerical computation – that is, they may
know, for example, that 9 – 4 = 5, but may still not be able to solve problems that
require just that computation. The understanding of addition and subtraction does
not depend only on knowledge of number facts but also on the children’s ability
to analyse the situations (Brown, 1981; Vergnaud, 1982).
The simplest additive reasoning situations are related to questions in which
elements are added to or taken away from groups or ‘sets’ (referred to as
transformations). A related and similarly easy type of problem has to do with
joining or separating two sets and asking about the results of this union/
separation, for example ‘In a family there are three girls and two boys. How many
children are there altogether?’ Problems that involve the comparison of two sets,
like ‘Mary has five books; Tom has three books; how many more books does
Mary have than Tom?’ are rather more difficult.
The analysis of the situation described in the problem is not sufficient to
characterize its level of difficulty. The information that is unknown is also
important. For example, a transformation problem is very easy when the
information to be calculated is the result of the transformation (e.g. in Problem 2
of Activity 5, the calculation is ‘9 – 4 = ?’). In this case, the action in the story and
the operation to solve the problem are directly related: a transformation that
increases the number will be solved by addition, and one that decreases the
number will be solved by subtraction.
In contrast, when the initial situation is unknown and must be calculated on the
basis of information about the transformation and its end result (for example, in
7 MATHEMATICAL AND SCIENTIFIC THINKING 273
they relied mostly on written procedures, which proved much less efficient: their
rate of correct responses was 74 per cent to word problems and 37 per cent to
computation exercises.
The comparison between their successful performance in the streets and their
rather high rate of failure when solving computation exercises suggests that
symbolic systems are not merely accessories to people’s reasoning. They mediate
complex reasoning. In other words they are part of the reasoning process and as
such change the nature of the activity by their use (Luria, 1979). In a later study,
Nunes et al. (1993) analysed the influence of symbolic systems not only on the
rates of correct and incorrect responses but also on the size of the errors made by
children when using either oral or written arithmetic. Three different error bands
were defined for this comparison: errors that fell within 10 per cent of the value of
the correct answer (for example, between 18 and 22 if the correct response to the
problem was 20); errors that were larger than 10 per cent but did not differ by
more than 20 per cent from the correct answer; and errors that differed from the
correct answer by more than 20 per cent. The percentages of correct responses
and errors in each error band for oral and written addition and subtraction are
presented in Figure 8. The figure clearly shows that written arithmetic led to larger
errors, a tendency that was supported by a statistical analysis of the association
between type of strategy used in solving the problem (oral versus written) and
error band.
/00)
<#'=)>#*$".;#")
6#7&&"%)>#*$".;#")
10)
!"#$"%&'(")*+,"#-".)
20)
30)
40)
0)
5*##"$&) 67&87%) 67&87%) 9*#")&8'%) 5*##"$&) 67&87%) 67&87%) 9*#")&8'%)
/0:) 40:) 40:)*;&) /0:) 40:) 40:)*;&)
?'@) AB>")*C)#",>*%,",) ?+@) AB>")*C)#",>*%,",)
Figure 8 The percentage of correct responses and errors for oral and written addition and
subtraction (data from Nunes et al., 1993, p. 47). (a) Percentage of correct responses and of errors
within each error band in addition problems. (b) Percentage of correct responses and of errors
within each error band in subtraction problems.
children seemed to keep the meaning of the problem in mind. For example, in
the division problem below, the boy clearly keeps in mind the fact that he is
trying to figure out how many marbles each of the five children will get:
The references to marbles and children are clear throughout the problem solving
procedure. In contrast, in written arithmetic, references to the problem and even
to the relative values of digits are set aside. For example, when solving in written
arithmetic the same computation, 75 divided by 5, the digit ‘7’ is spoken of as
‘seven’, not seventy, which would take into account its relative value. This loss of
meaning in written arithmetic is probably one of the reasons for children’s
acceptance of responses to computations that would seem, under other
circumstances, unacceptable, as illustrated by this example:
Such examples illustrate the relative difficulty of written arithmetic for these
children. This is further illustrated by a study conducted by Nunes (1993)
described in Research summary 1.
7 MATHEMATICAL AND SCIENTIFIC THINKING 277
RESEARCH SUMMARY 1
n
278 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
t
Activity 6 Comparing oral and written arithmetic
Allow about This activity will encourage you to reflect on the key differences between oral and written arithmetic
5 minutes for these children, and thereby identify what kinds of factors might increase the ease or the difficulty
of learning mathematics in school.
Consider the examples of oral and written arithmetic above and the account given in
Research summary 1. Why do you think that the children find oral arithmetic easier than
written arithmetic?
Comment
One reason why oral arithmetic practices appear to be more successful for children than
written forms of mathematical problem solving may be because this form of reasoning appears
to retain the ‘human sense’ of the problem. The problem retains its everyday, practical
meaning and is thought about with reference to quantities. However, by presenting the
problem on paper, its human sense is lost. Moreover, the process of translating the elements
of the problem into correct mathematical notation also presents the children with an
additional step in which errors and confusions can be introduced.
s
equal pretend cakes (cardboard rectangles) into halves, cutting one along the
height (resulting in two smaller rectangles) and the other along the diagonal
(resulting in two triangles). They ensured that the children realized that the two
parts of each cake were of the same size. They then asked the children: if one
child ate one of the rectangular parts and another child ate one of the triangular
parts, would they be eating the same amount of cake?
If the children understood the idea of half as the result of an exact division into
two, they should conclude that halves of two equal wholes indicate the same
amount irrespective of their appearance. However, children appeared to judge
the parts on their appearance and did not rely on the underlying logic of ‘dividing
the whole’. In a recent study (Nunes et al., 2002), we confirmed Piaget’s results
using a paper-and-pencil task: 45 per cent of the Year 4 children (aged about 8
years) and 20 per cent of the Year 5 children (aged about 9 years) did not
recognize that the two different looking halves were equivalent amounts of cake.
For concrete representations to be successful, there must be visual similarity
across each of the ‘portions’ of the whole.
Representing fractions through conventional notation does not appear to make
the task any easier. Children have a great deal of difficulty realizing that 1⁄8 is a
smaller number than 1⁄6. Ordering fractions in this way is much more difficult for
them than ordering whole numbers. By the age of 8 or 9, children have no
difficulty at all in ordering numbers up to 100 or 1,000 by their size or indicating
the place of a whole number in a number line. In contrast, Kerslake (1986) found
that only one of fifteen youngsters in the age range 12–14 could correctly place 2⁄3
on a number line. Kerslake’s study was carried out with a small sample, but the
results are in line with other research. Mack (1993) reports that the vast majority of
her sample of 12 year olds in the United States indicated that 1⁄6 was a smaller
fraction than 1⁄8 because 6 is smaller than 8. However, Mack reports that this was a
difficulty observed only when the youngsters were presented with the problems
symbolically. If asked to imagine that a pizza was divided into six pieces and an
identical pizza was divided into eight pieces, and then to indicate which pieces
would be larger, those obtained from the first or the second pizza, the youngsters
had no difficulty in giving the correct answer.
Our recent investigations bring support to these results. We presented 142
youngsters in the age range 8 to 10 years with two comparable problems about
fractions. In one problem, the youngsters were told that two boys would share
fairly one pie and three girls would share fairly an identical pie. Most of the
youngsters correctly indicated that each boy would receive more pie than each
girl, with rates of correct response varying from 89–96 per cent between the 8 and
the 10 year olds. When the youngsters were asked to compare the fractions 1⁄2 and
1
⁄3, the rate of correct responses dropped to 11 per cent, 31 per cent and 59 per
cent, respectively, for the 8, 9 and 10 year olds. These results suggest that the way
in which number problems are represented is another powerful influence on the
solutions that children will arrive at.
280 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
t
Activity 7 Assumptions about the physical world
Allow about This activity illustrates that it may be possible to show an implicit appreciation of principles relating to
5 minutes physics, without being able to demonstrate an explicit awareness of them.
Find a room or other open space and place a box or bin in the centre. Make a paper ball or
find a similar object. Now move quickly past the box and drop the ball into it as you pass by.
Can you do it?
Now, imagine you are running across a room and want to drop (not throw) the ball into the
box. Where should you be when you drop the ball in order for it to go into the box? Look at
Figure 9 and choose which picture correctly represents where you should be when you drop
the ball.
Figure 9 Where
should you drop the
ball if you want it to
fall into the box? (a) (b) (c)
Comment
You should find that the physical task is quite easy. However, it is likely that you have a wrong
idea of where you needed to be to get the ball in the box. The correct answer is to release
the ball before you get to the box (picture a). Although you may have found the question
difficult to answer because it demands an explicit awareness of the physical forces acting on
the ball, intuitively you knew when to release the ball when you were required to do so for
real.
s
The task you tried in Activity 7 was studied by McClosky et al. (1980), who asked
college students in physics and other students to say where they should drop the
ball from if they were running across a room and wanted to make sure it landed
in the box. The students were asked to answer by choosing the correct picture
from a set of three similar to the ones in Figure 9.
Physics students were significantly more accurate in answering this question
although, as McClosky and his colleagues suggested, they were unlikely to be
282 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
more successful in actually dropping the ball into the basket. Whereas 73 per cent
of the physics students chose the picture in which the stick figure dropped the
ball before reaching the target, only 13 per cent of the non-physics students made
this correct choice. Among the physics students, the answers were divided
between the correct choice and the picture where the ball was dropped directly
above the box (27 per cent of the physics students chose this alternative). In
contrast, 80 per cent of the non-physics students chose the picture of the stick
figure dropping the ball directly over the target and there were also choices in
which the ball was dropped after the figure had already run past the box (7 per
cent). These results seem to indicate that physics students were thinking about the
action differently from the non-physics students with respect to force and motion.
The physics students took into account the fact that the ball was already in motion
when carried across the room by the running stick figure and thought of its path
as it was dropped as the result of this initial force and direction plus the effect of
gravity, whereas the non-physics students disregarded the motion of the ball
while it was in the stick figure’s hand, and thought only about gravity.
This illustrates a key point about understanding scientific concepts: while
people can operate successfully within the physical world, they often do so
without a genuine understanding of the physical principles that underpin it.
Scientific learning is often a challenge to everyday understanding of cause and
effect.
everyday life with a world that is continuous, in which objects are solid and
undivided. Yet, to understand many of the changes they observe in the world, it is
necessary to develop a way of thinking that describes solid objects as bundles of
‘particles’, ‘molecules’, or ‘atoms’ – that is, discontinuous elements that are
somehow kept together.
Piaget and Inhelder (1974) were pioneers in the investigation of children’s
understanding of the particulate nature of matter. They set out a pattern of
investigation by pointing out that it is when children have to understand change
that they come to ‘invent’ an atomic theory about the world. Piaget and Inhelder
asked children to explain what happened to sugar when it was put into water and
then stirred. Whereas the younger children seemed to believe that the sugar
somehow disappeared, those at the ages of 11–13 were aware of the fact that if
the taste of sugar remained in the water, then the sugar itself must still be present
in some form. This permanence of a property of sugar – its taste – contradicted
the apparent disappearance of sugar from the viewpoint of the older children. In
order to eliminate the contradiction between the disappearance of sugar and the
preservation of one of its properties – the sweet taste – the older children
‘invented’ an atomic theory about physical quantities.
Many studies have refined Piaget’s original ideas about children’s
understanding of the particulate nature of matter. Driver et al. (1985) and Hesse
and Anderson (1992), for example, studied other chemical transformations such
as combustion (what happens when wood is burned?) and oxidation (rusting and
cleaning bits of iron) and also investigated pupils’ conceptions when they
observed other changes of state, such as evaporation. An example of the type of
responses observed by Driver et al. (1985) is presented in Figure 10, where one
child is shown attempting to illustrate what happens to particles in each of the
three states of matter: solid, liquid and gas. The figure illustrates the difficulty of
moving from a world of
observables to a conceptual
world of non-observed entities.
It also illustrates children’s own
conceptions of changes of state
rather than taught solutions.
A third type of difficulty lies
in the everyday world view of
things as stable, with properties
that are part of the very matter
of which they are made. As
discussed earlier, Piaget and his
colleagues hypothesized that
children develop a theory about
the particulate nature of matter
in their attempt to explain the
sweetness of the water after the Figure 10 A drawing of an 11 year old’s
disappearance of sugar: if a representation of the three states of matter: solid, liquid
property of sugar, sweetness, and gas (Driver et al., 1985, p. 145).
284 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
still exists in water, then the sugar must still be there. A conception of the world
that only deals in terms of matter and its properties leaves a range of phenomena,
such as light, heat and electrical current, unexplained. Chi and her colleagues
(Chi and Slotta, 1993; Chi, 1997) pointed out that not all scientific ideas are based
on the notion of matter: many important scientific ideas are based on the notion
of processes. These researchers suggest that the main obstacle to children’s
understanding of these scientific ideas is that they need to switch categories, from
matter to processes.
Examples of this kind of difficulty are easy to find (Reiner et al., 2000). For
example, children may have a natural tendency to treat heat and temperature as
substances with the essential property of hotness. Certainly young children find it
hard to think of heat and cold as existing independently of objects and thus treat
temperature as a property of objects. Similarly, Clement (1982) has provided
evidence that young children treat forces as properties of objects. In the case of
inanimate objects, children see forces as acquired properties that cause
movement; and think that objects stop when this acquired property dissipates in
the environment (Vosniadou et al., 2001).
A fourth obstacle to the development of scientific thinking is the fact that many
scientific concepts represent intensive rather than extensive quantities. Extensive
quantities, such as height and weight, are measured through the simple
application of a unit of measurement that is repeated until the number of units
completely describes the quantity. For example, when length is measured in
centimetres, the value that is obtained by this measurement is equivalent to the
number of times 1 cm – that is the unit – can be fitted onto the object measured.
In contrast, intensive quantities, which include density, speed and force, are
measured through a ratio between two other measures. So, density is measured as
a ratio of mass to volume. Similarly, speed involves a relation between distance
and time, and force involves a relation between mass and acceleration.
An everyday intensive quantity is taste. For example, how sweet a glass of
lemon juice will taste depends on how much sugar you add to the lemon juice
and how much lemon juice there is in the glass. There is a direct relation between
the amount of sugar and the sweetness: the more sugar you add, the sweeter the
lemon juice will taste. There is also an inverse relation between the amount of
lemon juice and sweetness: the more lemon juice there is in the glass, the less
sweet it will taste. The results of the investigation reported in Research summary 2
illustrate the difficulty of inverse relations for children.
7 MATHEMATICAL AND SCIENTIFIC THINKING 285
RESEARCH SUMMARY 2
Intensive quantities
Nunes et al. (2002) conducted an investigation of children’s understanding of inverse
relations. The children (142 in total) were shown pictures of glasses of lemon juice, as
those presented in Figure 11.
The children were asked first whether they thought that the lemon juice would taste
the same in the two glasses. Those children who thought that the taste would be
different were then asked to say which would taste sweeter. The percentages of
correct responses in Table 1 refer to the total number of children by age for both
questions.
Age in years
7 8 9
Will the lemon juice in the two glasses taste the 44% 70% 60%
same or different?
If the child said different: which one will taste 44% 40% 51%
sweeter?
Two results should be pointed out from the data in Table 1. First, there were no
significant differences between the age groups; so the small improvement in percentages
with age could be due to chance. Thus the problem is difficult even for children at age 9.
Second, quite a few 8 year olds realized that the taste would be different in the two
glasses but made the mistake of thinking that the more juice, the sweeter it would taste.
Rather than thinking of an inverse relation, they assumed that the amount of juice and
the amount of sweetness would be directly related. In fact, when children are asked to
respond to situations where all that matters is the variable that is directly related to the
intensive quantity, their performance is perfect. When we asked them to say which
lemon juice would taste sweeter when the two glasses had the same amount of juice and
different amounts of sugar, the rate of correct responses was close to 100 per cent.
These results were replicated in an earlier study by Desli (unpublished).
n
286 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
The intensive quantities that occur in science are usually more complex than
those encountered in everyday life because, as we showed earlier, they require
distinctions that people do not make in everyday life. It is likely that science
teaching involving intensive quantities could profit from the use of analogies with
simpler intensive quantities used in everyday life, in the same way that
mathematics teaching seems to profit from the use of everyday situations as
models for teaching about mathematical concepts and operations. This is a
reasonable hypothesis, but further research is needed to test it.
A second formal difference is that scientists strive for consistency and are (or
should be) willing to reject their explanations if they are inconsistent with
available evidence or with other valid theories. In everyday life people may not
be so ready to use consistency as a criterion for validating their knowledge. For
example, Hatano (1993) has observed that young children can respond correctly
to some questions about heat, but give different answers to other problems which
are inconsistent with their earlier responses. If asked what they would do in order
7 MATHEMATICAL AND SCIENTIFIC THINKING 287
to cool the water in a bath, even young children will say that they would add cold
water to it and thereby treat the resulting temperature in the bath as being
somewhere between that of the hot and the cold water. However, when asked
what is the temperature of the water in a container into which one litre of water at
the temperature of 60 degrees and one litre of water at the temperature of 20
degrees were poured, the most frequent answer given by young pupils is ‘80
degrees’. Whereas in the first problem about the temperature of the water in the
bath the resulting temperature is intermediate between hot and cold, in the
science classroom problem the resulting temperature is viewed as the sum of the
temperature in the two volumes of water. On the basis of a different set of
examples, di Sessa (1993) also argued that children’s conceptions lack the
coherence of scientific theories. However, there is disagreement on this issue.
Some researchers suggest that children’s theories are coherent even though they
are less powerful than adults’ theories (Posner et al., 1982).
correlation relied on those items which related to reasoning that Inhelder and
Piaget expected to develop only at the level of formal operations, whereas those
that can be solved at the concrete operational level were not related to success in
science.
t
Activity 8 Understanding correlations
Allow about This activity will help you to understand the importance of reflecting on the nature of the evidence
5 minutes that theoretical claims are based on.
What problems can you see in interpreting Piburn’s results?
Comment
Piburn’s results cannot be used on their own to infer that propositional reasoning ability
contributed to the students’ attainment in science. This is because their grades and their
propositional reasoning ability were obtained at the same point in time. While there may be a
relationship it is not possible to identify the direction of that relationship: it could equally be
the case that the students’ science grades contributed to their propositional reasoning ability.
To determine the direction of any relationships a longitudinal study is required, in which the
science learners are tested for their propositional reasoning before they start the science
course.
s
The second prediction from Inhelder and Piaget’s theory is about the
generalization of formal operational reasoning across contexts. The formal
operational stage was conceived by Piaget (Piaget and Garcia, 1971) as a stage in
which the operations of thought (such as propositional reasoning) become
abstract and therefore independent of their content. This means, according to
Piaget, that thinking can now function in a formal way, which makes it possible
for children to learn about any scientific domain without the development of new
operations of reasoning. This prediction about the generalization of the structures
of thought has met a growing degree of scepticism.
Specifically with respect to propositional reasoning, much work has been
carried out that indicates its susceptibility to influences from the content of the
proposition. For example, Cheng and Holyoak (1985) and Girotto et al. (1988)
looked at young people’s reasoning about several types of if–then propositions.
They have suggested that people reason about these propositions not in a formal,
content-free way, as originally suggested by Piaget and his colleagues, but rather
they rely on pragmatic schemas that involve both formal and content specific
elements. For example, when the if–then proposition is a prohibition people are
likely to make inferences that differ from those made when the proposition is a
promise. Thus the content of a proposition has a clear influence on the inferences
that are made even if the propositions have the same form. Activity 9 presents an
example of the contrast between prohibitions and promises.
290 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
t
Activity 9 Does the content of a proposition affect reasoning?
Allow about This activity will help you to see how the content of a proposition can affect the way that people
5 minutes reason about it, as they often rely on their everyday experiences to guide them to an appropriate
conclusion.
Here are two types of if–then propositions: promises and prohibition/permissions.
1 If you mow the lawn, I will give you £5. You didn’t mow the lawn. Do I give you £5?
2 If you drive a lorry, you must not drive through the city centre. You drive a car. Can you
drive through the city centre?
In (1) what do you conclude?
In (2) what do you conclude?
Are there differences? Why?
Comment
In the first proposition, you probably concluded that you were not given £5 because you did
not mow the lawn. Although this conclusion seems ‘logical’ based on your everyday
experiences of promises like this one, in fact it is not the logical conclusion. Just because you
did not do the task, it does not follow (in strict logical terms) that you will not be given £5. In
fact, the proposition does not specify what will happen in this circumstance, only what will
happen when you do mow the lawn. Similarly, in the second, prohibitive statement, you cannot
‘logically’ conclude that you can drive through the city centre. The proposition simply states
that you cannot drive through the city centre if you are driving a lorry. Although your common
sense tells you that you can imply that you are permitted because you are not explicitly
prohibited (‘they would have mentioned it if I was not!’), in fact you cannot come to that
conclusion on the basis of the proposition alone. From these examples you can see that when
reasoning about promises or prohibitions, people rely on their cultural knowledge to draw
their conclusions, rather than on strict logical deduction.
s
formulae, the terminology, and yet still maintain their previous conceptions. This
difficulty has been illustrated many times, for example, when instructed students
are interviewed about heat and temperature or forces, as discussed earlier. It is
often identified by teachers as a difficulty in applying the concepts learned in the
classroom; students may be able to repeat a formula but fail to use the concept
represented by the formula when they explain observed events.
Piaget suggested an interesting hypothesis relating to the process of cognitive
change. Cognitive change was expected to result from the pupils’ own intellectual
activity. When confronted with a result that challenges their thinking – that is when
faced with conflict – pupils realize that they need to think again about their own
ways of solving problems, regardless of whether the problem is one in mathematics
or in science. Conflict was hypothesized to bring about disequilibrium and would
result in the setting off of equilibration processes that would ultimately produce
cognitive change. For this reason, according to Piaget and his colleagues, in order
for pupils to progress in their thinking they need to be actively engaged in solving
problems that will challenge their current mode of reasoning.
However, Piaget also pointed out that young children do not always discard
their ideas in the face of contradictory evidence. They may actually discard the
evidence and keep their theory. Activity 10 describes a science lesson where
children aged about 6 were asked to observe whether objects floated or sank in
water and later in the afternoon were interviewed about what they had learned in
their science lesson.
t
A c t i v i t y 1 0 D o h e a v y t h i n g s s i n k a n d l i g h t o n e s fl o a t ?
Allow about This activity will encourage you to reflect on the problematic nature of supporting conceptual change
10 minutes in children.
Read the transcript below and note how Billy integrates the evidence that he collects (his
observation of the floating carton) with his existing ‘theory’ that heavy things sink.
The first of the interactions was recorded during a science lesson in which children were
asked to make predictions about whether some things would float or sink when put in a
basin full of water. The second is an excerpt of an interview carried out with one of the pupils
after the lesson.
TEACHER: You have to tell us what you’ve chosen (to place in the water).
BILLY: I’ve got a milk.
T: You’ve chosen a carton of a milk.
B: And it’s a bit heavy.
T: Yeah!
B: And you can drink out of it and eh ...
T: What’s it made from?
LOUISA: It’s a carton.
T: It’s a carton. What’s a carton made from? ... Do you know what cartons are made from?
Louisa?
292 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Comment
What do you think of Billy’s conclusion after the lesson: did he discard his explanation for why
some things float and others sink? What does this imply about children’s reaction to conflicting
evidence? It would seem that Billy’s theory that ‘heavy things sink’ remained intact, despite the
‘heavy’ milk carton floating. To account for this, he simply re-categorized the carton as ‘light’ in
his explanation. In this way, Billy’s theory remained unchanged – but his view of the evidence
was altered.
s
A new concept was developed by Vygotsky in order to address the social nature
of cognitive change: the concept of the zone of proximal development. The zone
of proximal development is the difference between what a child can accomplish
in solving a problem working independently and what can be accomplished by
the same child with adult help.
out by the children. She thus structured the activity so that the children could mix
the chemicals, compare results, and discuss what they had seen without being
unduly pressured by the planning of the mixing of chemicals. The planning of the
experiment was ‘off-loaded’ by the teacher’s assignment of chemicals (one to
each child) so that they could (as a pair) effectively generate a full table of mixes
and use the results to come to conclusions. This example illustrates how pupils
were allowed to use some aspects of a system of knowledge without its full
appreciation. Thus although the design of the experiment may have been beyond
their level of cognitive development, they were still able to participate in a
research activity, recording observations, comparing results, and trying to come to
conclusions that would be relevant to the learning of chemical mixtures.
t
Activity 11 H o w d o c h i l d r e n fi g u r e o u t w h a t s c i e n c e i s ?
Allow about This activity will show you that children and adults have different ideas about what science is about,
10 minutes and encourage you to consider how these differences might be resolved.
Read the transcript below and consider the difference between a teacher’s and some
students’ views of what science is. See if you can think of a way of introducing these students
to the teacher’s idea of what science is.
CAIN: What’s science?
TEACHER: Oh, go on ... you said ‘what’s science?’
CAIN: What’s science?
TEACHER: What’s science? Does anybody have any ideas what science is? ... Marcia.
MARCIA: It’s when ... you sit down and do your work.
TEACHER: Yeah! What sort of work is science? Daniel.
DANIEL: When you stay in your place.
TEACHER: When you stay in your place. Sarah.
SARAH: Number work.
TEACHER: Number work, is it? Something to do with numbers? Billy.
BILLY: Like when you work with the sand.
TEACHER: Like when you work with the sand. What sort of ... What sort of sand work is
science work, because sometimes I get you to work with the sand for number,
don’t I?
BILLY: Em ...
TEACHER: What sort of things do I ask you to do of ... say ... if it was science and I asked
you to go and work with the sand? What sort of things would you be doing?
CAIN: Measuring work.
TEACHER: Em ...
CAIN: I know.
TEACHER: Sh ... we have hands, don’t we? (Shushing Cain, who started the whole thing.) Paul.
PAUL: Measuring how much sand.
TEACHER: That’s more a number job, isn’t it? Excuse me, Sarah. ... Jodie.
JODIE: Like when you put heavy things.
296 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Comment
Working in the zone of proximal development means developing some common ground for
understanding science as children gain experience under the teacher’s guidance. The transcript
is an illustration of just how difficult it is for children to gain an understanding of what science is.
They participate in concrete activities that are not direct illustrations of what science is. The
teacher’s view of science is about asking questions and testing ideas in a systematic way. The
children only understand science as a label that describes activities that they do in school. The
teacher could orientate the children to her viewpoint by setting up a session such that the
children have to generate some ‘I wonder’ questions for themselves which the teacher could
guide them through answering.
s
5 Conclusion
Mathematical and scientific conceptual development does not seem to be simply
a question of learning facts or procedures. Both involve the achievement of new
ways of thinking and reasoning in generative ways. But it is not enough to
analyse the structures of reasoning, as Piaget originally proposed, in order to
understand children’s progress in mathematics and science. Children do not
interact only with the physical environment and do not discover only by
themselves the properties of actions and of objects. Societies have developed
systems of representation and knowledge to which children are exposed both in
school and out of school. Through their participation in social situations, where
people use these knowledge systems, the children come to use this knowledge,
even if they do not (at least initially) grasp it in the same way that adults do.
Whilst exploring their ideas with peers or participating with adults in situations
(such as experiments) which they themselves could not create, children become
engaged in mathematical and scientific thinking. In this way they may be able to
restructure their ways of reasoning and make progress.
However, theorists currently acknowledge that many reasoning problems make
specific intellectual demands irrespective of any general level of reasoning ability.
Fractions, for example, pose specific difficulties to children and adults alike,
which no general theory of mathematical development deals with adequately.
The distinction between properties and processes, like heat and light, is another
case in point. It causes extraordinary and quite particular problems for children
and many adults and, as for fractions, the distinction needs specific research and
probably specific educational methods as well.
Current theories about mathematical and scientific development recognize that
there are general ways in which children’s intellects change as they grow older,
but they also take into account the specific problems that arise when children
have to add particular techniques or forms of knowledge to their repertoire. This
new breed of theories also recognizes that children’s ways of thinking about the
world may change in particular ways as a result of finally mastering these new
and quite difficult forms of knowledge (Vygotsky, 1978).
So while psychologists can recognize that children do move towards a general
ability to represent their environment in increasingly abstract and organized ways
by inferring rules based on their experiences of the world, often these initial
hypotheses are problematic exactly because they are limited by the experiences
that the children may have.
298 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Further reading
Nunes, T., Schliemann, A. D. and Carraher, D. W. (1993) Street Mathematics and
School Mathematics, Cambridge, Cambridge University Press.
Nunes, T. and Bryant, P. (1996) Children Doing Mathematics, Oxford, Blackwell.
Stavy, R. and Tirosh, D. (2000) How Students (Mis-)Understand Science and
Mathematics. Intuitive Rules, New York, Teachers College Press.
References
Behr, M., Lesh, R. Post, T. R. and Silver, E. A. (1983) ‘Rational number concepts’, in
Lesh, R. and Landau, M. (eds) Acquisition of Mathematical Concepts and
Processes, pp. 91–126, New York, Academic Press.
Brown, M. (1981) ‘Number operations’, in Hart, K. M. (ed.) Children’s
Understanding of Mathematics: 11–16, pp. 23–46, London, John Murray.
Bryant, P., Christie, C. and Rendu, A. (1999) ‘Children’s understanding of the
relation between addition and subtraction: inversion, identity and
decomposition’, Journal of Experimental Child Psychology, vol. 74, pp. 194–212.
Carpenter, T. P. and Moser, J. M. (1982) ‘The development of addition and
subtraction problem-solving skills’, in Carpenter, T. P., Moser, J. M. and Romberg,
T. A. (eds) Addition and Subtraction: a cognitive perspective, pp. 9–24, Hillsdale,
NJ, Lawrence Erlbaum Associates.
Carraher, T. N., Carraher, D. W. and Schliemann, A. D. (1985) ‘Mathematics in the
streets and in schools’, British Journal of Developmental Psychology, vol. 3,
pp. 21–9.
Cheng, P. W. and Holyoak, K. J. (1985) ‘Pragmatic reasoning schemas’, Cognitive
Psychology, vol. 18, pp. 293–328.
Chi, M. T. H. (1997) ‘Creativity: shifting across ontological categories flexibly’, in
Ward, T. B., Smith, S. M. and Vaid, J. (eds) Creative Thought: an investigation of
conceptual structures and processes, pp. 209–34, Washington, DC, American
Psychological Association.
Chi, M. T. H. and Slotta, J. D. (1993) ‘The ontological coherence of intuitive
physics’, Cognition and Instruction, vol. 10, pp. 249–60.
Clement, J. (1982) ‘Students’ preconceptions in introductory mechanics’,
American Journal of Physics, vol. 50, pp. 66–71.
Desli, D. (1999, unpublished) ‘Children’s understanding of intensive quantities’,
PhD thesis, Institute of Education, University of London.
di Sessa, A. (1993) ‘Toward an epistemology of physics’, Cognition and
Instruction, vol. 10, pp. 105–225.
7 MATHEMATICAL AND SCIENTIFIC THINKING 299
Doise, W. and Mugny, G. (1984) The Social Development of the Intellect, Oxford,
Pergamon.
Driver, R., Guesne, E. and Tiberghien, A. (1985) ‘Children’s ideas and the learning
of science’, in Driver, R., Guesne, E. and Tiberghien, A. (eds) Children’s Ideas in
Science, pp. 145–69, Milton Keynes, Open University Press.
Erickson, G. and Tiberghien, A. (1985) ‘Heat and temperature’, in Driver, R.,
Guesne, E. and Tiberghien, A. (eds) Children’s Ideas in Science, pp. pp. 52–84,
Milton Keynes, Open University Press.
Gelman, R. and Meck, B. (1992) ‘Early principles aid initial but not later
conceptions of number’, in Bideaud, J., Meljac, C. and Fishcher, J.-P. (eds)
Pathways to Number: children’s developing numerical abilities, pp. 171–89,
Hillsdale, NJ, Lawrence Erlbaum Associates.
Ginsburg, H. P., Posner, J. K. and Russel, R. L. (1981) ‘The development of mental
addition as a function of schooling and culture’, Journal of Cross-cultural
Psychology, vol. 12, pp. 163–79.
Girotto, V., Light, P. H. and Colbourn, C. J. (1988) ‘Pragmatic schemas and
conditional reasoning in children’, Quarterly Journal of Experimental Psychology,
vol. 40, pp. 469–82.
Groen, G. J. and Resnick, L. B. (1977) ‘Can pre-school children invent addition
algorithms?’, Journal of Educational Psychology, vol. 69, pp. 645–52.
Hatano, G. (1993) ‘Children’s concepts of heat and temperature’, paper presented
at SRCD Biennial Meeting, New Orleans, March 1993.
Hesse, J. J. and Anderson, C. W. (1992) ‘Students’ conceptions of chemical
change’, Journal of Research in Science Teaching, vol. 29, pp. 277–99.
Howe, C., Tolmie, A. and Anderson, A. (1991) ‘Information technology and group
work in physics’, Journal of Computer Assisted Learning, vol. 7, pp. 133–43.
Howe, C., Tolmie, A. and Rodgers, C. (1992) ‘The acquisition of conceptual
knowledge in science by primary school children: group interaction and the
understanding of motion down an incline’, British Journal of Developmental
Psychology, vol. 10, pp. 113–30.
Inhelder, B. and Piaget, J. (1958) The Growth of Logical Thinking from Childhood
to Adolescence, London, Routledge.
Kerslake, D. (1986) Fractions: Children’s Strategies and Errors: a report of the
strategies and errors in secondary mathematics project, Windsor, NFER-Nelson.
Kieren, T. (1988) ‘Personal knowledge of rational numbers: its intuitive and
formal development’, in Hiebert, J. and Behr, M. (eds) Number Concepts and
Operations in the Middle-grades, pp. 53–92, Reston, VA, National Council of
Teachers of Mathematics.
Lave, J. (1988) Cognition in Practice: mind, mathematics and culture in everyday
life, Cambridge, Cambridge University Press.
Luria, A. (1979) Curso de Psicologia Geral, Rio de Janeiro, Civilização Brasileira.
300 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Reiner, M., Slota, J. D., Chi, M. and Resnick, L. B. (2000) ‘Naive physics reasoning:
a commitment to substance-based components’, Cognition and Instruction,
vol. 18, pp. 1–34.
Riley, M. S., Greeno, J. G. and Heller, J. I. (1983) ‘Development of children’s
problem-solving ability in arithmetic’, in Ginsburg, H. P. (ed.) The Development of
Mathematical Thinking, pp. 153–200, New York, Academic Press.
Vergnaud, G. (1982) ‘A classification of cognitive tasks and operations of thought
involved in addition and subtraction problems’, in Carpenter, T. P., Moser, J. M.
and Romberg, T. A. (eds) Addition and Subtraction: a cognitive perspective,
Hillsdale, pp. 39–59, NJ, Lawrence Erlbaum Associates.
Vergnaud, G. (1985) ‘Concepts et schemes dans une théorie opératoire de la
représentation’, Psychologie Française, vol. 30, pp. 245–52.
Vosniadou, S., Ioannides, C., Dimitrakopoulou, A. and Papademetriou, E. (2001)
‘Designing learning environments to promote conceptual change in science’,
Learning and Instruction, vol. 11, pp. 381–419.
Vygotsky, L. S. (1978) Mind in Society: the development of higher psychological
processes, Cambridge, MA, Harvard University Press.
Chapter 8
A socio-cognitive perspective on
learning and cognitive development
Anne-Nelly Perret-Clermont, Felice Carugati and John Oates
Contents
1 Introduction 305
8 Conclusion 327
References 328
8 A SOCIO-COGNITIVE PERSPECTIVE ON LEARNING AND COGNITIVE DEVELOPMENT 305
Learning o u t c o m e s
After you have studied this chapter you should be able to:
1 describe Piaget’s views on the significance of authority, peer collaboration
and individual activity for cognitive development;
2 describe Vygotsky’s position on the role of social interactions in cognitive
development;
3 describe Mead’s theory of symbolic interactionism;
4 explain how these theorists’ socio-historical contexts may have influenced
their thinking;
5 define the concept of ‘socio-cognitive conflict’ and describe its role in
fostering cognitive growth;
6 explain why intersubjectivity is significant in understanding teaching;
7 define ‘didactic contract’ and explain its significance;
8 discuss how Doise’s four levels of analysis offer ways into the study of
learning and development.
1 Introduction
In this chapter, we start by reviewing three theoretical positions on cognitive
development that have been prominent in research on learning and have had
major influences on educational practice. We then explore the importance of
recognizing the ways in which children’s learning is embedded in and formed by
social relationships and practices, and the implications that these have for theory.
This leads to a new way of thinking about the processes that are active when
learning takes place. It also gives a broader perspective on the factors that may
need to be considered when seeking to improve the effectiveness of learning
situations, when trying to understand how, why and when children learn.
The ideas covered in this final chapter also offer a new set of perspectives from
which to view the ideas and findings presented earlier in the book. In particular,
this chapter broadens the frame within which cognitive development can be
viewed. It suggests, in a number of ways, that development does not just ‘happen’
within the individual child, but that it occurs in a context of social customs,
understandings and relationships that play an important role in either facilitating
or hindering the individual’s progress towards more advanced modes of thinking.
Further, we argue that this context is an integral part of learning, which defines
the meanings that children and teachers give to their joint activities. Learning is
embedded in particular social practices, with specific expectations and values,
especially, but not only, in school situations.
Our aim in this chapter is to offer a critical analysis of the basic assumptions and
orientations of three main theories in developmental psychology, those of Jean
Piaget, Lev Vygotsky and George Herbert Mead. We locate their ideas within the
306 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
dominant cultural ideologies of Europe, Soviet Russia and the United States and
consider some of their implications for the ways in which education and learning
are conceived and structured. We then bring in new ways of thinking about
learning and teaching situations, illustrated and supported with research
evidence, to build a richer understanding of the social and psychological
processes involved. A particular focus throughout the chapter is on the mutual
understandings that participants in such situations hold about the nature of the
social relationships supporting teaching and learning interactions; what we
describe as the ‘didactic contract’.
behaviourist view, arguing instead for the importance of the child’s own activity
in the construction of knowledge. Educational policy was, to some extent,
inspired by the Piagetian position. One important development was that
‘discovery learning’ became a fashionable approach, in which children were
offered richly resourced learning environments, and the role of the teacher was
reworked as a ‘facilitator’ of children’s own constructive endeavours, rather than
purely being a ‘transmitter of knowledge’.
However, despite the ‘progressive’ label attached to such approaches, and the
marshalling of Piagetian ideas as a counter to the behaviourist theory of learning,
the individualistic constructivist perspective throws up some serious issues that are
also open to critical analysis. Perhaps the most damaging of these is the potential
that Piaget’s theories offer for ethnocentric interpretations of delays in children’s
development in different socio-economic and cultural circumstances. For example,
it is often the case that children in socially disadvantaged groups, or from countries
where there is much poverty and hardship, are found to be delayed in their
development relative to their more advantaged peers with better standards of
living. Given such findings, the view that development is essentially endogenous
(i.e. springing from within the child) inevitably leads to the suggestion that such
differences are due to inherent deficits in these cultural groups. Highlighting this
problem raises the issue of how the relation between culture and the development
of thinking can be reconciled with the Piagetian tradition.
Piaget recognized this issue, but he gave little direct attention to the problem,
and neither have those who have built on his theory in orthodox ways. However,
as we will show during the course of this chapter, the question is closely aligned
with a related issue that Piaget did consider to be of great importance; the relation
between the development of the individual and what we will call in this chapter
‘authority’. By this we mean the influence of those who can exert power and
direction over learners in various ways, by their control of resources,
communication, time and routine, by virtue of their greater age or greater
knowledge, or by holding positions of respect, such as teacher, lecturer or school
head. Perret-Clermont (1996) has suggested that the socio-historical context of the
first half of the twentieth century and Piaget’s own personal biography were
sources of the ethical beliefs he came to adopt. These in turn had a major
influence on the role he attached to ‘authority’ in his theory and the image he had
of rational thought and what it could achieve.
Piaget lived in Europe through the blind violence of the First and Second World
Wars, which he blamed on the prevailing religious and cultural traditions and their
autocratic use of authority over people. Perhaps in reaction to this experience, he
rejected the notion of ‘authority’ and placed his faith in rationality and the
autonomy of the individual in non-hierarchical social relationships with peers. In
line with this way of thinking, he focused his efforts on identifying the conditions
that could lead a child towards developing autonomous, critical and rational
thought. As a crucial foundation for this, he placed importance on a capacity to
reflect on one’s own actions and to apply logical reasoning to one’s own behaviour
as well as to problems in the world. Piaget doubted whether the external,
authoritative influences of education and other social institutions could ever do
308 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
more than just load children with cultural constraints. This ‘baggage’ he saw as only
giving pre-formed answers to questions that children had neither formulated for
themselves, nor attempted to answer in the course of individually developing their
own ways of thinking. At the heart of this problem, Piaget believed, lies the
asymmetry of status between adult and child, which he saw as constraining
children from developing the capacity for autonomous, reflective thought.
This line of thinking is often seen as demoting culture to just a collection of
beliefs, social rules and constraints rather than seeing it as providing a legacy of
meanings, understandings, narratives and social structures transmitted from one
generation to the next. This creates a major problem for seeing how formal
teaching can play any kind of role in fostering cognitive development, especially
when taken in combination with Piaget’s stance that authority inhibits the
development of independent thought and reflection. Indeed, Piaget saw formal
teaching as an obstacle and a hindrance to cognitive development; for him,
interaction between peers was much more important. Piaget’s positive view of
peer interaction as a fundamental facilitator of cognitive development is an
important point, but one that is rarely noted by commentators, who have tended
to portray social processes as insignificant in Piagetian theory. And when Piaget
himself stressed the importance of peer interactions, he seems to have been
unaware that these commonly occur in social settings organized by adults.
Under the banner of ‘active learning’, some teachers have taken on these three
strands of Piaget’s thinking (i.e. negative towards authority and positive towards
peer collaboration and individual activity) and have concluded that, apart from
providing a rich environment, they need to leave children to learn entirely on
their own. In contrast to the educational ideology that children learn best by
imitating ‘correct’ models, teachers who wish to be ‘Piagetian’ seem to feel that
they have to stay ‘backstage’. Although teachers’ direct experience is likely to tell
them that an educational environment needs to be carefully structured if it is
going to engage children’s interest, the received Piagetian view is that
organization must come from the children. From this perspective, the teacher
must stay in the background, so that their authority and knowledge do not hold
back the child. This creates a tension, for example when children might justifiably
ask ‘If the teacher knows the right answer, why not give it to us rather than
expecting us to discover it for ourselves?’.
This approach can be contrasted with more traditional teaching situations,
where teachers are likely to seem omnipresent. In such contexts, teachers often
start sentences which students are meant to complete (e.g. ‘The name of the
largest lake in Switzerland is ...?’), and students are then rewarded if they give the
‘correct’ response. This is a method which is thought to encourage students’
participation, but in fact the ultimate aim is for the child to give a response that
matches that of the adult. Yet a common theme in both traditional and progressive
approaches is that the final aim is for the child to achieve the adult’s answer. Is
this expectation a type of cultural ethnocentrism, or simply a lack of awareness of
cultural constructions?
8 A SOCIO-COGNITIVE PERSPECTIVE ON LEARNING AND COGNITIVE DEVELOPMENT 309
individual development was located within, even subservient to, broader socio-
cultural development. Second, he also emphasized the role of technology,
through his continued reference to tools (in a broad sense, he was referring to
what we can call cultural ‘tools of thinking’ as well as physical artefacts). Third, he
saw mental development as springing from the use of tools embedded in social
activities. His position was broadly aligned with the Marxist view that the
development of thought is the outcome of both material conditions and people’s
activities and struggles.
However, despite these broad differences between the two theorists, there is a
parallel with Piaget’s reaction to the context in which he was located. Vygotsky also
stressed the individual, constructive aspects of development. He did not see the
child as simply a passive recipient of culture’s transmission of tools, but also as a
creative builder of their own thinking. This was not consonant with the prevailing
ideology of Soviet Russia, and Vygotsky’s work was not widely published for many
years, so it is for this reason that Vygotsky’s work is often seen as following Piaget,
although in fact the two theorists were to some extent independently developing
their ideas at much the same time. Another point of agreement between the two
lies in their recognition that biological maturation also plays a role, albeit a
secondary one, in making certain developments possible on the cognitive plane.
On balance, Piaget saw the control of development as being essentially inside
the child, while Vygotsky saw the child’s development as being the internalization
of outside influences or social co-ordinations. In his theory, Piaget justified his
position by seeing the logical progression of development as arising from
children’s reflections on their own actions, culminating in the achievement of
formal operational thought, that is, fully abstract and hypothetico-deductive
reasoning. For Vygotsky, however, this achievement was the result of a social,
rather than an individual process.
appropriate ways by ‘significant others’’. For the child, seeing a hammer used in
specific ways by another person and hence in a social context, imbues the object
with social significance. Through such repeated experiences whereby objects,
actions and language come to acquire social meanings, the child also begins to
form simultaneously a representation of the self and of what Mead called the
‘generalized other’. This means that the child becomes able to see themselves and
their actions from the perspective of another person, and as a result, begins to
internalize meanings and values. According to Mead, it is this process of
internalization that allows children’s thinking to advance from the level of
immediate experience to a level of self-reflection. Mead believed that this is an
important part of children forming a conscious concept of self, and he believed
that the play of young children, in which role-playing other people and their
activities tends to be common, was a primary vehicle for this development. This
stands in contrast to Piaget’s view of the role of other people. According to Piaget,
play is at first solitary, and then later (as two or more children play together)
becomes ‘parallel play’, where the players merely co-ordinate their behaviour, but
do not truly co-operate. Only in Piaget’s final stage of development do they
properly take account of other people’s real or potential perspectives.
The reason why these different interpretations of the role of play are important
is that play can be seen as an excellent example of a form of social activity that is
not bound by external authority structures, nor by imposed power relations.
Although at times one child or another inevitably ‘sets the agenda’ for where the
play is going, in general, play is an activity that children engage in for its own
sake, not for its purpose in relation to some defined outcome. The significance of
such reciprocal peer relations for development will be explored in more detail as
this chapter progresses.
the classic Piagetian conservation of liquid task, a child is first asked to confirm
that the amounts of liquid in two identical containers are, indeed, the same. Then
the liquid from one container is poured into another container with different
proportions and the child is asked whether the two amounts of liquid are the
same. Children find it much easier to solve this problem correctly if the task is
presented as a ‘fair shares’ reward of a drink to each child for their participation in
the experiment. Here, the task of judging whether the two containers hold the
same amount is supported by the notions of fairness and equity. This particular
social marking is therefore based on the concept of ‘distributed justice’. But, of
course, social marking is itself subject to meaning construction and how it
operates in a particular situation will depend on how the children make sense of
the social rules that the adult intends to appeal to. For example, evoking the
concept of distributed justice is more effective if children are placed in a co-
operative situation than if they are expected to compete with each other. Research
that has made use of rules to do with the ‘right’ to have equal quantities of objects
or drink have confirmed this phenomenon.
In this section, then, we have shown two broad patterns of findings regarding
the social context of experimental conditions. The first contextual effect that we
have found is that the cognitive demands that a task makes of a child, and the
child’s chances of reaching a solution, do not simply reside in the task itself. They
are made harder or easier by the social context of the child, the task and other
people. A degree of social conflict around possible routes to solving the task can
make the finding of a correct solution easier. Once such a joint solution has been
found, through the resolution of the socio-cognitive conflict, it can then be
available for the child to apply to new tasks with similar demands when working
independently. In this way the child’s cognitive development has been fostered.
The second important contextual effect that we have highlighted concerns the
‘match’ of the social context to the rules that have to be applied to the task in
hand. For example, as we illustrated in respect of the Piagetian conservation task,
a match between task and social context also helps children to come to a
successful solution. In the favourable circumstances outlined above, then,
children can perform at higher levels than those predicted by theoretical positions
that only offer individualistic interpretations of cognitive development, such as
those of Piaget.
there is a trick!’. Other children started the test, but then turned to the researcher,
asking ‘Now, should I tell her that there is a trick or not?’. Further, Grossen
observed that children who were non-conservers (i.e. they believed that moving a
quantity of liquid from one container to another of different proportions changed
the amount of liquid) tended to produce correspondingly non-conserving
responses from the children that they were testing. Conserving children, on the
other hand, tended to elicit conserving responses from their partners.
Leoni (1990) has also shown how younger children are more successful at solving
problems when the tester is a psychologist who is making a game out of a
problem-solving situation than if the tester is a teacher. In contrast, children who
have had more experience of school are more likely to succeed if the task is
presented as a ‘school’ task and not as a game.
conversations, as a result, are structured in very different ways from most other
situations in everyday life.
The implicit assumption that guides most classroom activity is that the teacher
poses the questions and knows the right answers. On this basis, a teacher
evaluates the student’s responses and then uses this information to judge the
student’s level of competence. Where a group of children show that they
understand and accept this basic, underlying social rule, the flow of
communication in their classroom is then well regulated on the basis of this
consensus. This is generally seen as best fitted to meeting the mutual expectations
for good teaching. Edwards and Mercer (1987) and Mercer (1995) have shown
how, in classes of 8–10-year-old students, classroom activity and the teacher’s
language use build shared understandings about these basic social rules which
students then use to attribute meaning to their interactions. The teacher uses
language to point up and mark these implicit understandings, so that the way talk
is organized and used in the classroom by the teacher develops a shared
vocabulary which fixes the respective roles of teacher and student. This sets up
the particular perspective in which the teacher is expected to lead the student
along a path to the acquisition of new concepts and the reorganization and
extension of knowledge.
based not just on a reflective analysis of the answer but also on the child’s
perception that they need to give a different, more ‘correct’ response.
A specific type of classroom dialogue that has attracted the interest of
researchers is where groups of children seem to be able to develop new
understandings through ‘externalized reasoning’ where they all participate in
more of a true debate. In this sort of situation, the teacher’s interventions take on
a different meaning. Their role can be to get closer to understanding the students’
modes of thought, rather than encouraging the student to better understand the
teacher’s thought processes. Rather than guide them along a preconceived path,
the teacher may facilitate the students’ discourse and cognitive processes to
support the development of new understandings (Mercer, 1995).
These insights have led to a major shift in perspective that now puts at centre
stage the relationship between student–teacher debate and the development of
thinking (Pontecorvo, et al., 1991; Pontecorvo, 1993; DeGroot and Schwarz, 2003;
Schwarz et al., 2003). Seeing the classroom situation as a context in which
characteristic types of conversational events take place, where the development
of modes of thought are either facilitated or hindered, points up the reciprocal
relation between the social milieu and the knowledge of the child (Barth, 1994).
The shared cultural practices of school life, the objects that they involve and the
communication patterns within them all contribute to a ‘referential framework’
(Resnick, 1991) for the teaching and learning process.
rarely use arithmetic approaches even when these have been studied and learned
in the school setting.
As we saw in Chapter 7, research on this theme was carried out in Recife in
Northern Brazil by Nunes et al. (1993) who studied the mathematical abilities of
street children. At ages of 6 or 7, the researchers found that these children are
already surviving independently by re-selling oranges, or doing small jobs, and
some of them quickly become highly accurate in working out monetary
transactions and giving the correct change. Yet these same children systematically
fail in mathematics in their first year of schooling and tend to be rapidly rejected
by the educational system. Why is it that they are unable to transfer the
knowledge acquired from their street life to school tasks? On the streets, children
learn highly complex algorithms to deal efficiently with the buying and selling
situations that they encounter, situations that demand accurate calculation. In the
school situation, the teacher does not know that the students already possess
these practical techniques and when they spontaneously use them in class the
teacher tends to treat their strategies as clumsy or incorrect. This then contradicts
the children’s experiences and does not allow them to reflect on the strengths and
limitations of these strategies (in large part oral) that they use in their commercial
transactions.
The question thrown up by this research is the nature and extent to which there
is a link between the context in which one acquires knowledge and the context in
which it is used or transferred, since the desired transfer of knowledge and skills
often fails to occur. There can be many reasons for this; perhaps the teacher fails
to understand the significance of this aspect, or fails to clarify it, or perhaps the
students themselves fail to grasp the relations between the knowledge they have
gained in one situation and that required in another. It may also be the case, at
times, that knowledge is context-bound without any of the participants being
aware of this. Thus we have to consider further the three-way relations in the
triangle of teacher, student and knowledge.
Student Teacher
well as a student. We have shown so far in this chapter how the development of
the individual’s knowledge cannot be easily isolated from the other two points,
the teacher and the object of knowledge itself, since all of these are situated
within an institutional and cultural context. We have also tried to show how much
there is to be gained in enriching understandings of cognitive development by
considering all three points together. Generally, trainee teachers see knowledge
as being their focus, so, for example, in mathematics they may concentrate on
how to design a series of lessons, how to demonstrate a formula or how to
choose examples to be worked through. We would argue that this is a narrow
view of teaching and that to study this aspect unquestioningly, without also
considering the other points of the triangle and their relations, is too limited and
reductionist.
As far as the teacher is concerned, theoretical and research interest has mainly
tended to focus on the socio-emotional aspects of the teacher–student relationship
rather than on the ways in which the teacher represents the child’s points of view.
It has also tended to see teaching in relation to the evaluation of student abilities. It
is only very recently that studies have been conducted on how teachers think
about intelligence, learning and development, and their mental models of students
(for an exposition of the dynamics of causal attributions see Monteil, 1989; with
respect to the educational implications of social representations, see Mugny and
Carugati, 1985; Selleri et al., 1994; Carugati and Selleri,1996). However, little is
known about the reasons why teachers find it difficult to decentre from their own
points of view.
The study of this teacher–student–knowledge triangle in teaching situations is
becoming increasingly important. How do teacher-trainers, teachers themselves
and psychologists studying ‘the triangle’ see it functioning? The implicit didactic
contract seems to be that the teacher sets tasks that correspond with what he or
she wishes the students to know. Because students come to understand what they
are expected to know, they then expect to be given exercises that will lead to
success in the task. One can talk of a genuine ‘micro-culture’, with slogans such as
‘You have to do this, because it is how you learn properly’; ‘If you do it in this
way, you get a good mark’; ‘If you do this, it means that you have understood’. In
other words, students’ responses and the roles that they take on are not
spontaneous constructions, but are the product of particular teaching practices.
So how can these issues around the meanings of learning situations and the
corresponding expected behaviour(s) be addressed? A crucial theme here is the
modes of constructing shared understandings between teacher and students that
support or do not support students in discovering the knowledge that the teacher
wishes to transmit. In what circumstances do children genuinely learn what one is
seeking to teach them? The school system and the life of the classroom are usually
sufficiently based on well-oiled daily routines to allow students to respond
correctly without having properly understood. Students do not often see the aim of
schoolwork as being to gain understanding, but rather to pass tests successfully.
For students, essential achievements are to gain a satisfactory mark for homework,
to meet the teacher’s expectations and to show that they know what they ought to
know. To understand is an optional bonus! This is often what happens if, when a
8 A SOCIO-COGNITIVE PERSPECTIVE ON LEARNING AND COGNITIVE DEVELOPMENT 325
teacher says ‘try to understand’, the student hears this as ‘try to understand how
you have to answer, so that I believe that you have understood’! Some students
even gain good qualifications without ever having properly experienced the
satisfaction of having truly understood something. We believe that to understand is
an autonomous act, internal and free, which does not depend on hierarchical
structures, nor can it be prescribed or demanded. So there is a risk that this level of
deep comprehension, if it has not been experienced, may end up playing no part
in the students’ field of experience, or even in their expectations.
Summary of Section 7
. Doise has proposed that the social nature of cognition can be analysed at
four different levels. These four levels focus on different objects of
concern.
. Level 1 is the level of the individual engaged in constructing their modes
of thought and their knowledge, trying to make sense of their experience.
. Level 2 is the level of actual relationships, interaction and discourse, for
example, student–teacher or peer–peer expectations. It focuses on, for
instance, the construction of intersubjectivity and mutual expectations.
. Level 3 refers to roles and statuses that frame and constrain the sorts of
behaviour that are considered ‘proper’ for teaching and learning in
different contexts.
. Level 4 is the level of national policies and ideologies, systems of social
representations, that define situations (such as schools) in specific ways.
These may often be dictated by professional training and traditions, and
by government policies that define situations in specific ways. But these
are always reinterpreted by the participants, for example by teachers and
students in specific schools and situations.
. These levels are not independent: one level can affect or transform
another.
8 Conclusion
In this chapter we have set out to show how communication contracts and
didactic contracts, both of which are largely implicit, govern the transmission of
understandings and the conditions of learning. Learning lies at the heart of
specific ways of relating which are closely interlinked with the rest of the daily life
of the class. Learning is also linked more generally with the ways in which the
school operates so as to organize understandings and the dynamics of teacher–
student interaction. We have also explained the importance of the ways in which
teachers and students construct their own meanings for these practices and
traditions. The transmission of experience and knowledge from one generation to
the next is seen as something that can be appropriated and interpreted. It can
become both a personal and also a collective acquisition, and is thus not a
328 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Acknowledgement
This chapter is an updated and modified version of the chapter published in
Italian (Carugati, F., and Perret-Clermont, A. -N. (1999) ‘La prospettiva psico-
sociale: intersoggettività e contratto didattico’, in Pontecorvo, C. (ed.) Manuale di
Psicologia dell’educazione, pp. 41–66, Bologna: Il Mulino.) and an abridged
version of the original paper is in press in French.
References
Azmitia, M. (1996) ‘Peer interactive minds: developmental, theoretical, and
methodological issues’, in Baltes P. B. and Staudiner, U. M. (eds) Interactive
Minds: life-span perspectives on the social foundations of cognition, pp. 133–61,
Cambridge, Cambridge University Press.
Barth, B. M. (1994) Le Savoir en Construction: former à une pédagogie de la
compréhension, Paris, Retz.
Baucal, A., Muller, N., Perret-Clermont, A.-N. and Marro, P. (2002) ‘Nice designed
experiment goes to the local community’, paper presented at the Fifth Congress of
the International Society for Cultural Research and Activity Theory, Amsterdam,
June 2002.
8 A SOCIO-COGNITIVE PERSPECTIVE ON LEARNING AND COGNITIVE DEVELOPMENT 329
Howe, C., Tolmie, A., Greer, K. and Mackenzie, M. (1995) ‘Peer collaboration and
conceptual growth in physics: task influence on children’s understanding of
heating and cooling’, Cognition and Instruction, vol. 13, pp. 483–503.
Hundeide, K. (1985) ‘The tacit background of children’s judgements’, in Wertsch,
J. V. (ed.) Culture, Communication and Cognition: Vygotskian perspectives,
pp. 306–22, Cambridge, Cambridge University Press.
Hundeide, K. (1988) ‘Metacontracts for situational definitions and for presentation
of cognitive skills’, Quarterly Newsletter of the Laboratory of Comparative Human
Cognition, vol. 10, pp. 85–91.
Hundeide, K. (1992) ‘The message structure of some Piagetian experiments’, in
Wold, A. H. (ed.) The Dialogical Alternative: towards a theory of language and
mind, pp. 139–56, Oslo, Scandinavian University Press.
Labov, W. (1972) ‘The study of language in its social context’ in Giglioli, P. P. (ed.)
Language and Social Context, pp. 283–307, Harmondsworth, Penguin Education.
Lévy, M. and Grossen, M. (1991) ‘Contrat experimental et acte de questionnement:
deux illustrations empiriques de l’articulation entre processus et activité cognitive
de l’enfant dans une situation de test piagetienne’, Bulletin de Psychologie, vol. 44,
pp. 229–38.
Light, P. and Perret-Clermont, A.-N. (1989) ‘Social context effects in learning and
testing’, in Sloboda, J. A. (ed.) Cognition and Social Worlds, pp. 99–112, Oxford,
(Oxford Science Publications), Oxford University Press.
Littleton, K. and Light, P. (eds) (1999) Learning with Computers: analyzing
productive interaction, London, Routledge.
Mead, G. H. (1934) Mind, Self and Society, Chicago, Chicago University Press.
Mercer, N. (1995) The Guided Construction of Knowledge: talk amongst teachers
and learner, Clevedon, Multilingual Matters.
Monteil, J. M. (1989) Eduquer et Former, Grenoble, Presses Universitaires de
Grenoble.
Mugny, G. and Carugati, F. (1985) l’Intelligence au Pluriel, Cousset, Delval.
Nicolet, M. (1995) Dynamiques Relationnelles et Processus Cognitifs, Lausanne,
Delachaux et Niestlé.
Nunes, T., Schliemann, A. D. and Carraher, D. W. (1993) Street Mathematics and
School Mathematics, Cambridge, Cambridge University Press.
Perret, J. F. and Perret-Clermont, A.-N. (2001) Apprendre un Metier dans un
Contexte de Mutations Technologiques, Fribourg, Editions Universitaires Fribourg.
Perret-Clermont, A.-N. (1979) La Construction de l’Intelligence dans l’Interaction
Sociale, Berne, Peter Lang.
Perret-Clermont, A.-N. (1996) La Construction de l’Intelligence dans l’Interaction
Sociale (4th edn), Peter Lang, Berne.
8 A SOCIO-COGNITIVE PERSPECTIVE ON LEARNING AND COGNITIVE DEVELOPMENT 331
Acknowledgements
Grateful acknowledgement is made to the following sources for permission to
reproduce material within this book. Every effort has been made to contact
copyright holders. If any have been inadvertently overlooked the publishers will
be pleased to make the necessary arrangements at the first opportunity.
Chapter 1
Table
Table 1: Younger, B. and Gotlieb, S. (1988) ‘Development of categorization skills:
changes in the nature or structure of infant form categories?’, Developmental
Psychology, (24), pp. 611–19, The American Psychological Association, Inc.
Figures
Figure 1: Harold Taylor/Oxford Scientific Films; Figure 3: Younger, B. and
Gotlieb, S. (1988) ‘Development of categorization skills: changes in the nature or
structure of infant form categories?, Developmental Psychology, (24), pp. 611–19,
The American Psychological Association, Inc.
Chapter 2
Text
Reading A: Christophe, A. and Morton, J. (1998) ‘Is Dutch native English?
Linguistic analysis by 2-month-olds’, Developmental Science, vol. 1 (2), pp. 215–
19, Blackwell Publishers Ltd; Reading B: Harris, M. and Chasin, J. (1999)
‘Developments in early lexical comprehension: a comparison of parental report
and controlled testing’, Journal of Child Language, 26, pp. 453–60, Cambridge
University Press.
Figures
Figure 2: Johnson, E. K. and Jusczyk, P. W. (2001) ‘Word segmentation by 8-
month-olds: when speech cues count more than statistics’, Journal of Memory
and Language, vol. 44, p. 555, Academic Press; Figure 3: Johnson, E. K. and
Jusczyk, P. W. (2001) ‘Word Segmentation by 8-month-olds: when speech cues
count more than statistics’, Journal of Memory and Language, vol. 44, p. 557,
Academic Press; Figure 5 and Figure 9: Fenson, L., Dale, P., Reznick, S., Bates, E.,
Thal, D. and Pethick, S. J. (1994) ‘Variability in early communicative
development’, Monographs of the Society for Research in Child Development,
vol. 59 (5), p. 42, The Society for Research in Child Development, Inc., University
of Michigan; Figure 6: Elman, J., Bates, E., Johnson, M., Karmiloff-Smith, A., Parisi,
D. and Plunkett, K. (1996) ‘Rethinking Innateness: a connectionist perspective on
development’, MIT Press; Figure 7: Jones, S. et al. (1992) ‘Human speech and
language’, The Cambridge Encyclopedia of Human Evolution, Cambridge
University Press; Figure 8 and Figure 10: Fenson, L., Dale, P., Reznick, S., Bates,
334 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Chapter 3
Text
Reading A: Mehler, J. and Dupoux, E. (1994) ‘The organ of language’, What
Infants Know: the new cognitive science of early development, pp. 146–50,
Blackwell Publishers Ltd; Reading B: Karmiloff-Smith, A. (1992) Beyond
Modularity: a developmental perspective on cognitive science, MIT Press.
Figures
Figure 4: Reprinted by permission of the publisher from The Postnatal
Development of the Human Cerebral Cortex vols I-VIII by Jesse LeRoy Conel,
Cambridge, MA.: Harvard University Press, Copyright #, 1939, 1975 by the
President and Fellows of Harvard College; Figure 5: Photo courtesy of Leslie
Tucker, Centre for Brain and Cognitive Development, Birkbeck College; Figure 6:
Scott Camazine/Science Photo Library; Figures 7 and 8: Neville, H. J. et al. (1998)
‘Cerebral organisation for language in deaf and hearing subjects: biological
constraints and effects of experience’, Proceedings of The National Academy of
Science, U.S.A., vol. 95, # Copyright 2002, The National Academy of Sciences,
U.S.A.; Figure 9: Reprinted from Brain and Language, vol. 61, p. 351, Reilly, J. S.
et al., ‘Narrative discourse in children with early focal brain injury’, Copyright
(1998), with permission from Elsevier.
Chapter 4
Text
Reading A: Aitchison, J. (1998, 4th edn), ‘A blueprint in the brain’, The Articulate
Mammal: an introduction to psycholinguistics, Routledge/Taylor & Francis.
Chapter 6
Figures
Figure 1: Frith, U. (1989) ‘Figure 3: The false belief paradigm’, Autism: explaining
the Enigma, Blackwell Publishers Ltd; Figure 3: Whiten, A. (1991) Natural
Theories of Mind: evolution, development and simulation of everyday
mindreading, Blackwell Publishers Ltd; Figure 4: Meltzoff, A. N. (1995)
‘Understanding the intentions of others: re-enactment of intended acts by 18-
month-old children, Developmental Psychology, vol. 31, (5) Copyright # 1995 by
the American Psychological Association. Reprinted with permission.
ACKNOWLEDGEMENTS 335
Chapter 7
Text
The observations in Activities 10 and 11 were collected by S. Allen in a small-scale
study for an Open University course and are reprinted with permission.
Figure
Figure 10: Driver, R. (1985) ‘Beyond Appearances: the conservation of matter
under physical and chemical transformations’, in Driver, R., Guesne, E. and
Tiberghien, A. (eds) Children’s Ideas in Science, Open University Press.
Cover photographs
ª Getty Images (Photodisc).
336 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
Name index
Aitchison, J. 172, 201–2 Callanan, M. 254 Emiliani, F. 326
Anderson, C. W. 283 Camaioni, L. 107 Erickson, G. 282
Arterberry, M. E. 42 Carpenter, T. P. 272
Asher, S. R. 227 Carraher, T. N. 274 Fantz, R. L. 28
Astington, J. 239, 243, 244, 252 Carugati, F. 312, 313, 324 Fearing, D. D. 47
Avis, M. 238 Casey, B. J. 131, 222 Fenson, L. 78, 79, 83, 84, 92, 93,
Azmitia, M. 312 Cavanagh, J. C. 214 106, 107, 109, 110, 111
Cervantes, C. 254 Fifer, W. 67
Baer, R. A. 223 Changeux, J. -P. 124, 132, 213 Fodor, J. A. 14, 127–8, 129, 135, 156,
Baker, L. 214 Charman, T. 249, 255 157, 158, 159, 207
Baldwin, D. A. 77 Chasin, J. 78, 106–11 Fox, N. A. 146
Barkley, R. A. 226, 227 Chelune, G. J. 223 French, R. M. 51
Barlow-Brown, F. 107 Cheng, P. W. 289 Frith, U. 237
Barrett, M. 107 Chevallard, Y. 317 Furrow, D. 196
Barrett, M. D. 85, 88 Chi, M. T. H. 284 Fuster, J. M. 119, 145
Barth, B. M. 322 Choi, S. 89
Bartsch, K. 246, 247 Chomsky, N. 11, 14, 64, 135, 137, Gage, P. 212–13
Bates, E. 85, 93, 106, 107, 110, 111 143, 144, 160, 165, 172, 192, 194, Garcia, R. 289
Baucal, A. 316 197, 201–3 Gelman, R. 160, 278
Bavelier, D. 140 Christophe, A. 67, 69, 98–104 Gerstadt, C. L. 220, 221
Behl-Chadha, G. 36 Clement, J. 284 Gilly, M. 312, 326
Behr, M. 278 Cowan, M. W. 122 Ginsburg, H.P. 274
Bell, M. A. 146 Crago, M. 189 Girotto, V. 289
Bell, N. 316 Csibra, G. 147, 148 Goldfield, B. A. 86, 111
Bellugi, U. 189 Cutler, A. 102 Goldfield, E. C. 133
Benedict, H. 107, 110 Goldman-Rakic, P. S. 119, 145, 146
Benigni, L. 107 Dale, P. 106 Goldstein, K. 212
Berg, E. A. 223 DeCasper, A. J. 66, 67, 120 Gopnik, A. 54, 56, 89, 239, 247, 248
Bertoncini, J. 98 DeGroot, R. 322 Gopnik, M. 189
Bickerton, D. 136 Dehaene-Lambertz, G. 99 Gottlieb, S. 30, 31, 32, 33
Bishop, D. V. M. 190 Deloache, J. S. 48 Graham, S. 47
Bomba, P. C. 30, 53 Demetras, M. J. 193 Grant, D. A. 223
Bornstein, M. H. 42 Dennett, D. 235, 236 Groen, G. J. 273
Borton, R. W. 124 Desli, D. 285 Grossen, M. 315–16, 319
Bosch, L. 104 di Sessa, A. 282, 287 Gunzi, S. 110
Braine, M. D. S. 195, 197 Diamond, A. 146, 221
Bretherton, I. 107, 110 Doise, W. 19, 292, 305, 312, 313, Hanlon, C. 181, 193
Brooks, S. 107 325, 327 Harlow, J. M. 212–13
Brophy, M. 227 Donaldson, M. 234, 318 Harris, M. 74, 75, 77–8, 78, 84, 85,
Brousseau, G. 317 Donaldson, M. 17–18 86, 87, 88, 89, 106–11
Brown, A. L. 214 Dore, J. 79, 107 Harris, P. 238
Brown, J. 254 Drewe, E. A. 221 Hart, B. 89
Brown, M. 272 Driver, R. 283 Hartl, M. 240
Brown, R. 170, 174, 175, 179, 181, Dunn, J. 215, 254 Hatano, G. 286
184, 193 Dupoux, E. 67, 100, 129, 154–7 Hebb, D. 133
Brown, T. E. 227 Hesketh, S. 99
Bruner, J. S. 63, 73–4, 76, 87, 150, Edwards, D. 321 Hesse, J.J. 283
293, 325 Eimas, P. D. 34, 35, 40, 43–4, 48, 52, Holyoak, K. J. 289
Bryant, P. 273 53 Howe, C. 292–3, 312
Burgess, P. 213 Elbers, E. 316 Hughes, C. 215, 221, 227
Butterworth, G. 41, 77 Elman, J. 129, 133, 196 Hundeide, K. 316, 318
NAME INDEX 337
Tardif, T. 89
Tatham, M. 68
Tees, R. C. 81
Thal, D. 106, 142
Thatcher, R. W. 147
Thomas, M. S. C. 190
Tiberghien, A. 282
Tomasello, M. 107, 194, 196
Trauner, D. 148
Tucker, L. A. 148
Turkewitz, G. 124
Tzourio, N. 131
Waxman, S. R. 43
Wellman, H. 246, 247
Welsh, M. C. 223
Werker, J. F. 81
Whiten, A. 242
Williams, D. 127
Wimmer, H. 238, 240
Wood, D. 19
Woodruff, G. 235, 236
Woolfe, T. 254
Wyndhamn, J. 318
Yeeles, C. 107
Younger, B. A. 30, 31, 32, 33, 47
Zittoun, T. 328
SUBJECT INDEX 339
Subject index
A-not-B task 146, 211, 222 attention shifting, and the brain structure and function 149
abstract concepts, infant prefrontal cortex 148 and changes in vocal tract
categorization of 38, 48, 49 auditory cues, and categorization anatomy 83
action 39 and cognitive modules 11
controlled actions and authority, and Piagetian theories genetic and environmental
automatic actions 211, 228 307–8 influences in 14
slips of 210–11 autism 127 innateness and epigenesis
action words, and early and domain specificity 160 116–18
comprehension vocabulary 108, auxiliary verbs and language development 11,
109, 111 defining the term 167 14, 135–45
active learning 308 and the formation of questions physiological features of 118–19
active sentences 167 138 and plasticity 118, 123, 143, 158
additive reasoning 268, 271–4 and linguistic development 175 postnatal 119, 120, 122–3, 143
adult–child intersubjectivity 315–17 subject-auxiliary inversion 192 prenatal 119, 120–1, 130, 143
children testing children 315–16 and selectionism 124–5, 133
psychologists testing children babbling 81, 92–3 and self-organization 10, 132–4
316 babies see also cerebral cortex;
relational context 315 development of the vocal tract prefrontal cortex
adults, theory of mind and 82–3 brain damage
interaction with 253 learning to say words 81–3 and cognitive modularity 154–6
amalgams 174 postnatal brain development and domain specificity 160
animals 122–3 and executive function 211–13
categorical forms of mental and the postnatal and language development
structure in 26 environment 119, 120 140–2, 143
infant categorization of 50–1 selectionism 124 and the prefrontal cortex 146,
and early language and speech recognition 65–73 148
development 54 in different languages 67, brain imaging methods 130–2
and humans 51–3 98–104 and inhibitory control 222
and image schemas 44 Baka tribe (Cameroon), and the Broca’s aphasia 154
levels of categorization 35, unexpected transfer task 238 bullying, and theory of mind 244–5,
36–7, 47, 48 basic/intermediate categories, and 256
and memory 34–5 infant categorization 35, 47, 50
non-human animals 50–1 behaviour, organization of 207 canonical babbling 81, 93
and object examination 47 behaviourism cat categories
and the sequential touching and constructivism 306–7 and infant categorization 51
procedure 41–2 and domain-general differentiating cats from dogs
see also cat categories; dog development 159 39–40
categories; horses belief-based terms in children’s and the familiarization/
antisocial behaviour, talk, and mental state terms 246–8 novelty-preference method
communication skills and bilingual babies (Spanish/Catalan) 29, 30
executive function 215–16 104 of humans 52, 53
aphasia, and cognitive modularity bottom-up processing, and infant and levels of categorization
154–6 categorization of animals 50, 51 36
arithmetic, oral versus written boys and memory 34–5
274–5 and early word production 83, categorization
ASL (American Sign Language), 84 and language 23
neuroscientific studies of deaf and theory of mind and re-cognition 25–6
participants and 139–40 development 254–5 see also infant categorization
attention deficit hyperactivity and vocabulary content 93 cause and effect, and theory of
disorder (ADHD) children, and brain and cognitive development mind 244
inhibitory control 12, 226–7 10–11, 115–61 cerebral cortex 116, 118, 119
340 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
hemispheres 138, 139 and theory of mind and infant categorization 39–42,
and cognitive modularity development 253–4 57
154–6 communication contracts 317, 327 dynamic movement cues 42
modules 125–35 commutativity of addition 274 and early language
and plasticity 123 compounding, and morphology development 54
and self-organization 133 168 sequential touching
cerebral palsy 126 computational modelling 185 procedure 41–2, 47, 50
child directed speech (motherese) concept formation 9 morphological 172
196 conceptual categorization 27–8 syntactic 170–1, 172
children testing children 315–16 conceptual change, and to reference, and first word
chimpanzees mathematical and scientific learning 77–8
and theory of mind skills 235 thinking 290–6 to word boundaries 68–72
vocal tract of 82, 83 conceptual fields, and cultural practices
classroom psychology, and mathematical thinking 271 and oral versus written
cognitive development 320–1 conceptual primitives 44 arithmetic 274–8
clinical psychology, executive concrete operations stage of of school life 322
function in 211–13 development 269 cultural tools 17
cognition in everyday life 8–9 conflict and Vygotsky’s social
cognitive development 305–28 cognitive 19 constructivism 309, 310
and constructivism 14–16, 159, and conceptual change as a and the zone of proximal
255, 306–9, 310 social process 292–3 development (ZPD) 294
and knowledge out of the socio-cognitive 312, 313, 314 culture
classroom 323–5 connectionist models (neural and Piagetian theories 307–8
and mathematical and scientific networks) 149, 183, 184 and social constructivism 309
thinking 288–96 connectionist computational
and the psychology of school modelling 149 Darwinian selection, and brain
life 320–5 and grammatical development development 124
and social constructivism 17–19, 196, 198 Day/Night Stroop task 220, 221
309–10 of infant categorization 44, 48 deaf children, and theory of mind
and social interactions 312–14 and self-organization in the 254
and social practices 305 brain 134 deaf people
socio-cognitive perspective on of vocabulary learning 80 and sign language
13, 325–7 conservation, and mathematical and the innateness of
and symbolic interactionism reasoning 268–9 language 136
310–11 conservation of liquid task 314, 316 neuroscientific studies of
see also brain and cognitive consistency, and scientific thinking 139–40
development 286–7 dendrites
cognitive flexibility, and executive constructivism in human infants 122, 123
function 213–14, 214–15, 221, 223 Piaget’s theory of 14–16, 159, and prenatal brain development
cognitive modularity 116, 127–32, 255, 306–9, 310 120–1, 122
154–7 see also social constructivism desire-based terms in children’s
and Fodor’s model of the mind context-bound words talk, and mental state terms 246–8
128, 129, 158–9 and early comprehension developmental cognitive
cognitive psychology, executive vocabulary 108, 109, 110 neuroscience
function in 209–11 and early word meaning 85, 86 brain imaging methods 130–2
cognitive skills, and theory of mind contextually flexible words 85, 86 and language development
development 248–50, 256 controlled actions, and automatic 138–42, 143
cognitive structure 24–7 actions 211 developmental triangle 323–4
collaborative learning 293 counting strategies 273–4 didactic contracts 306, 317–19, 327
commission errors, and inhibitory creoles, and the development of children’s interpretation of
control 221, 222 language 136 adults’ questions 318–19
communication cues the student’s view of ‘being a
and executive function and defining the term 68 student’ 317–18
communication skills 214–16 and first words 173
SUBJECT INDEX 341
and the role of experience development 254–5 if-then propositions, and scientific
87–8 and vocabulary content 93 understanding 289–90
types of words 85–7 global/superordinate categories image schemas, and infant
recognizing speech 65–73 and infant categorization 35, 37, categorization 44
understanding and producing 41–2, 47, 50 immigrant children, and Pinker on
words 63–5 single-process model of 43, the innateness of language 136–7
see also speech 48 individuals, and socio-cognitive
fMRI (functional magnetic Go/NoGo tasks 221, 222, 227 research 326
resonance imaging) 131, 139, 222 grammar 11, 165–203 infant categorization 9–10, 23–57
of ADHD children 227 defining grammatical terms 167 and cues 39–42, 57
focal lesions, and language grammatical development dual-process model of 44–6
development 140–2 173–81 and early language
formal operations stage of diary studies of 195–6 development 54–6
development 270, 310 early combinations 174 formation of 30–3
and scientific understanding empiricist theories of 165, of human and non-human
288–90 187, 194–6, 197–8 animals 50–3
formulaic speech 174 and first words 173 levels of representation 35–7,
fractions, and mathematical nativist theories of 65, 187, 46–8, 50
thinking 278–9, 297 193–4, 197 measuring 28–30
French language to full blown grammar perceptual and conceptual
and infant speech recognition 179–81 categorization 27–8
67, 100–2, 103 and utterance length 174–9 single-process model of 43–4,
and phonology 168 word endings 182–90 45–6, 48
furniture, infant categorization of, grammatical errors and brain- of spatial relations 24, 37–8,
levels of categorization 36, 37, 47, damaged children 141–2 48–9
48 and the nature of spoken storing categories in memory
language 166–73 33–5
gaze direction, and first word learning morpho-syntax inflectional morphology 168, 182
learning 77 171–2 single and dual route theories of
gender differences phonology 166, 167–8 182–90, 197
in early word production 83, 84 and Pinker on the innateness of information encapsulation, and
and theory of mind language 136–7 Fodor’s modularity of mind 128
development 254–5 plural forms 169, 182, 184, 188 inhibition of return, and the
and vocabulary content 93 see also morphology; nouns; prefrontal cortex 148
see also boys; girls syntax; UG (Universal inhibitory control 12, 213, 217–24
generative learning, mathematics as Grammar); verbs and ADHD children 226–7
262–4 grammatical morphemes 168 and brain activation 222
genetic determination, and brain in child development 222–5
development 117, 126 habituation/dishabitation sucking implications of under-
genetic influences on brain method, testing infants’ developed 225–6
development perception of languages 98–104 measuring the development of
and cognitive neuroscience 142 the Handgame 220, 221 219–22
and epigenesis 116–18 hearing loss 126 and prepotent stimuli 218, 221
innate differences versus heat and temperature, and and the richness of the sensory
modularization 127–32 children’s understanding of world 217
and the shared environment science 282, 284, 286–7, 290 innate modularity versus
126–7 Hebb rule 133 modularization 127–32
see also nativism hiding from policemen task 18 innateness
German language, plural inflection horses, infant categorization of 43, and epigenesis 116–18
errors 188–9 52, 53 Pinker on the innateness of
girls humans, infant categorization of 50, language 135–8
and early word production 83, 51–3 input systems (modules), and
84 hybrid stimuli, and infant Fodor’s modularity of mind 128
and theory of mind categorization 40
SUBJECT INDEX 343
intensive quantities, and children’s and theory of mind matter, children’s understanding of
understanding of science 284–6 development 283–4
intentions, understanding 249–50 language ability 252–3, 255 Mean Length of Utterance (MLU)
inter-modal preferential looking language use in the family 174–6
task 177, 178–9 253–4 means–end tests, and infant
interference effects, learning about see also first words; grammar; categorization 55
word endings 184 speech; vocabulary memory
intermediate categories see basic/ language delay 89–90, 91, 92 infant storage of categories in
intermediate categories language experience, and the 33–5
internalization, and symbolic development of word meaning learning about word endings
interactionism 311 87–8 182–90
interpersonal interaction, and learning by appropriation 294–5 planning and working memory,
socio-cognitive research 326 left-hemisphere damage (LHD), and executive function 213,
intersubjectivity see adult–child and language development 141–2 214–16
intersubjectivity lemon juice puzzle 285 and re-cognition 26
intransitive verbs 178 lies, and theory of mind 241–2, 244, working memory 145
inversion problems in addition and 256 mental growth, social nature of 10
subtraction 273 local interactions, and self- meta-representation, and theory of
irony, and second-order theory of organization in the brain 134 mind 238
mind 241 modules (functional units of the
Italian language, and infant speech MacArthur Communicative brain) 125–35
recognition 99 Development Inventories innate modularity versus
and early word comprehension modularization 127–32,
Japanese language 78, 79, 84, 91, 106 158–61
and infant speech recognition and early word production 84, and the shared environment
67, 98–102, 103 91 126–7
learning to count 262 and vocabulary content and money
joint attention 248–9 development 93 and integers 278
joint proposals 243 Mandarin Chinese, and the and oral arithmetic 274
joint social goals 249 development of word meaning 89 monolingual babies 104
jokes, and second-order theory of mandatory processing, and Fodor’s morphemes, grammatical 168
mind 241, 243 modularity of mind 128 morphology 166, 168–70
mathematical thinking 12–13, learning about word endings
Knock/Tap game 220, 221 261–80 182–90
knowledge, transferring out of the and cognitive development learning morpho-syntax 171–2
classroom 322–3 288–96 nativist theories of 65
Korean language, and the conceptual change in and over-regularization errors
development of word meaning 89 individuals 290–2 169, 170, 184, 185–6
and fractions 278–9, 297 U-shaped development in 169,
language and generative learning 262–4 182, 185
and brain development 11, 14, and knowledge out of the motherese (child directed speech)
135–45 classroom 322–3 196
and categorization 23 and mathematical representation mothers
and cognitive modularity 154–7 270 children’s first words and
and infant categorization 26, 27, oral versus written arithmetic language experience 87–8
28, 57 274–5 and language delay 89–90
early language development Piaget’s research into and theory of mind
54–6 mathematical reasoning 268–70 development 253–4, 255
single-process model of 43 problem solving and social motor cortex, delayed formation of
and spatial relations 38 situations 270–1 myelin sheaths 122
and mathematical and scientific and scientific reasoning 280 motor disabilities, and the
understanding 261 understanding addition 268, environment 127
social constructivism and 271–4 multiplicative reasoning,
language development 17 see also scientific thinking conceptual field of 271
344 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN
procedural knowledge, and infant and cognitive development cognitive skills 248–50, 256
categorization 44 288–96 SLI (specific language impairment)
proportional reasoning 269–70 conceptual change in children, double dissociation and
propositional reasoning 288–90 individuals 290–2 word inflections 189
prosodic speech cues, and word conceptual change as a social slips of action 210–11
boundaries 72 process 292–6 Smarties task, and theory of mind
prosody, and infants’ perception of formal operations stage 238–40, 245, 247
languages 67, 98, 102 288–90 snowflakes, self-organization and
prototype abstraction 33–4, 38, 53 different ideas about science patterns in 133, 134
prototypes, infant categorization of 295–6 social constructivism 17–19, 309–10
30–3, 38 facts and ideas in 264–7 social context, of early word
psychologists testing children 316 and intensive quantities 284–6 learning 73–8
and mathematical development social factors, in theory of mind
quantitative enrichment 43 280 development 251–5, 256
question formation parsimony and consistency social interactions, and theory of
and Pinker on the innateness of 286–7 mind 243–4
language 137–8 physics and the physical world social marking 313–14
and syntax 192, 196 281–2 social process, conceptual change
questions, and the didactic contract second-order theory of mind 241–3 as a 292–6
318–19 selectionism 124–5 social representations, and socio-
and the Hebb rule 133 cognitive research 326–7
reading self-organization, and brain social situations, and mathematical
executive control demands of development 10, 132–4 problem-solving 270–1
214 sensory disabilities, and the social support of learning and
executive function and reading environment 126–7 development 19
skill 223 sentences socio-cognitive conflict 312, 313,
receptive language, and active and passive 167 314
grammatical development 177 and the nature of spoken socio-cognitive perspective 13,
reduplicated babbling 81 language 166 325–7
referential style, and vocabulary object of a sentence 167, 178 sorting behaviour, developmental
content 93 subject of a sentence 167, 170 changes in 54, 56
representational re-description, and syntax 170–1 Soviet Russia, and Vygotsky’s social
infant categorization 44–5 sequential touching procedure, and constructivism 309–10
right-hemisphere damage (RHD), infant categorization 41–2, 47, 50 Spanish/Catalan bilingual babies
and language development 141–2 sibling interaction, and theory of 104
role assignment, and theory of mind development 252–3, 255, spatial relations, infant
mind 244 256 categorization of 24, 37–8, 48–9
roles, and socio-cognitive research sign language specific/subordinate categories,
326 neuroscientific studies of deaf and infant categorization 35, 47,
Russian language, and infant participants and 139–40 50
speech recognition 67, 99 and Pinker on the innateness of speech
language 136–7 babbling 81, 92–3
Sally/Anne task, and theory of and theory of mind and the development of the
mind 236, 237, 245, 247 development 254 vocal tract 82–3
sarcasm, and second-order theory ‘Simon says’ games, and executive and early word comprehension
of mind 241 function 208–9, 210, 211, 221 63, 64, 73
scaffolding 19 single-process model, of infant and early word production
school life, cognitive development categorization 43–4, 45–6, 48 83–4, 94
and the psychology of 320–5 single route theory of inflection use and grammatical development
scientific thinking 12–13, 264–8, 183–9, 197 173–81
280–97 skill acquisition linking sounds to meaning 73–8
children’s understanding of and executive function 222, 223 recognizing 65–73
matter 283–4 and the prefrontal cortex 147–9 cues to word boundaries
theory of mind and associated 68–72
346 COGNITIVE AND LANGUAGE DEVELOPMENT IN CHILDREN