Participant Reference in Genesis 37

INTRODUCTION

For a proper understanding of a narrative it is crucial to know who are the "actors" on the scene, and what their roles and mutual relationships are. 1 The way in which participants are introduced and the referring expressions used for them throughout the story have both linguistic effects (affecting, e.g., the cohesion and structure of a text) and literary consequences (having an impact on, e.g., the characterization of the participants).

This article investigates the various ways in which participants are introduced and referred to in Genesis 37. It investigates how a linguistic approach to participant reference can help addressing exegetical questions and how it interacts with other historical-critical as well as synchronic, literary approaches. Its starting point is the model for the description of participant reference developed by S E Runge in his contribution to a previous issue of JNSL (2006) and in his PhD dissertation (2007) and tests its applicability to the Genesis chapter under discussion. 1 In the autumn of 2011 I gave an MA seminar about participant reference and participant identification in Biblical Hebrew. I am indebted to the participants of this seminar for their useful feedback, which sharpened my ideas on this topic. I am also indebted to Steven E Runge, who was willing to contribute to the discussion board in the electronic learning environment of this seminar. The investigations have been supported by the Netherlands Organisation for Scientific Research (NWO).

THE INTRODUCTION OF NEW PARTICIPANTS

Patterns of participant introduction in Biblical Hebrew

When participants enter the scene for the first time (or after a considerable period of absence), they are introduced in a certain way. Various patterns of participant introduction can be discerned. A systematic analysis of these patterns can be found in the works of Runge (2006:89-90;2007:91-94) and L J de Regt (1999:32-33). Runge distinguishes two patterns. In the first the new participant's referring expression and anchoring relation (on which see below, the end of this section and further Section 3) are introduced in a Topic-Comment articulation. Compare the activation of Abraham and Lot in Runge's second pattern is a construction with two clauses. The first clause predicates the existence of the participant and establishes a relation. The second clause assigns the referring expression. Although the formulation is somewhat different, this is similar to the pattern that De Regt describes as "a clause referring to the participant in terms of class and membership", for which he gives the following example: and [Judah] turned in to) a certain Adullamite, whose name was Hirah".

If a new participant is introduced only by name, and the reader does not have prior knowledge about this participant, it is still "hanging in the air". Therefore, new participants are often anchored. The anchoring relation "guides the reader in how to ground the participant to the discourse context" (Runge 2006:89). Runge (2007:92) 3. Interpersonal relations, e.g. Gen 12:5 ‫ת-לוֹט‬ ‫אֶ‬ ‫וְ‬ ‫יו‬ ִ ‫ח‬ ‫ן-אָ‬ ‫בֶּ‬ "Lot, his nephew".

The introduction of new participants in Genesis 37

When we turn to Genesis 37, we see that the patterns that Runge and De Regt describe are not used for the introduction of Joseph (37:2), Reuben (37:21) and Judah (37:26). Nor are these participants anchored to other participants in the way described by Runge: In Gen 37:2 the story starts with Joseph, without an anchoring expression such as "Jacob's son", which would link Joseph to the immediately preceding superscription "These are the Toledot of Jacob". Likewise, the information that Judah and Reuben belong to Joseph's brothers is understood. It is not indicated by anchoring expressions. Hence we can say that Joseph, Judah and Reuben are not typically introduced as new participants. For this untypical introduction of a new participant by name alone, two explanations are possible (based on Revell 1996):

1. They belong to the great figures in the past that were so well known that further identification was not needed. 4 In this respect it is worth noting that apparently Joseph, Judah and Reuben did not need an explicit introduction, whereas Bilhah 3 Note that here the participant is anchored both by a genealogical relation ("the Adullamite") and by a interpersonal relation ("his friend"); cf. the third item in this list. 4

Cf. Revell (1996:70; Section 4.8.1): "In a small proportion of the cases, a character is designated, where first mentioned, by name alone. This is easily understood in the case of the great figures of the past: Abraham, Isaac, Jacob/Israel, Aaron, Moses. Their names were so well known in the community that further identification was superfluous". See also Heimerdinger (1999:133-134 These two options affect our reading of Genesis 37 within the whole of Genesis. Should Genesis 37 (or rather 37-50) be read as a story on its own, or only as part of the larger Genesis account, including the birth stories of Jacob's children? And if the participant references presuppose knowledge of previous elements, does this also apply to narrative elements? Then Shechem, mentioned in Gen 37:12, recalls Genesis 34, which may even explain the brothers' quick departure from there; Reuben's name is spoiled in Gen 35:22 and the reference to Hebron in 37:14 is just a continuation of 35:27. In such a reading, one has to come to terms with the fact that Rachel, who died in Gen 35:16-20, implicitly appears on the scene again by the reference to Joseph's mother in 37:9-10. 6 An argument against the second option is the presence of various phenomena that suggest that Gen 37:1/2 marks an absolute start. Especially the Toledot formula in 37:2, the final in a series that provides a major structural framework to the book of Genesis, sets Chapters 37-50 apart as Jacob's family history. 7 The Subjectqatal pattern in the clause immediately following the Toledot formula corroborates its structuring role. One could challenge our use of the Toledot argument because other Toledot sections, too, contain references to participants that have been introduced before: Noah's Toledot formula is given in Gen 6:9, but he has already been mentioned in Gen 5:29-32 and 6:8 (cf. Pirson 1999:110), and Terah's Toledot formula is given in Gen 11:27, but he has already been mentioned in 11:24-26. In these two examples, however, most of the 5

Cf. Revell (1996:70 information given before the Toledot formula (about Noah's and Terah's sons respectively), is repeated after it and one can, so to speak, start reading from the Toledot formula onwards, without missing essential information. The case of Genesis 37 is different, because here some essential information is missing regarding the family relations involved. Thus in this case one can only start reading from the Toledot formula onwards if some information about the heroes Joseph, Jacob, Judah and Reuben can be presupposed. Taking these observations into account, I prefer the first of the two options mentioned above, because it accounts for the fact that Bilhah and Zilpah are anchored while the other participants are not, and because it does justice to those phenomena that mark Gen 37:2 as the start of a completely new section. 8 A participant that had not played any role in the preceding chapters of Genesis is the anonymous man in Gen 37:15a ‫הוּ‬ ‫אֵ‬ ‫צָ‬ ‫מְ‬ ‫יִּ‬ ‫וַ‬ ‫ישׁ‬ ִ ‫א‬ "And a man found him".

Unnamed participants are not very common in biblical narrative (De Regt 1999:4), and there is no scholarly consensus about the man's role in the narrative. Longacre (1989:152) argues that the anonymous man "is made thematic for a brief span", because he is mentioned explicitly three times in Gen 37:15-17, whereas Joseph is only referred to by suffixes and affixes: "It seems that by these devices the newly introduced (albeit anonymous) man is made thematic for a brief span". However, according to Runge (2007:71-72), the idea that "the man" is thematic is misguided; he is "a face without a name, simply a prop to move the plot forward". He argues that "thematicity is NOT directly proportional to the use of lexical NPs, and that the lexical references to this participant are either 8 Another phenomenon that suggests that some knowledge on the readers' side is presupposed concerns the identification of "their father" with "Jacob" and "Israel". Especially the shifts between "Jacob" and "Israel" presuppose the knowledge that they refer to the same participant. Nowhere in this chapterand nowhere else in the Bible -do we find an explicit identification of the type "Jacob, that is Israel" (compare semantically required (37:15a, 15c), or are intended to signal the beginning of a new development unit (37:17a)". 9 At the end of the chapter, another new participant is introduced, without any anchoring relation to the participants already mentioned:

‫ח‬ ‫בָּ‬ ‫טַּ‬ ‫הַ‬ "Meanwhile the Midianites had sold him in Egypt to Potiphar, an officer of Pharaoh, the captain of the guard".

Potiphar is anchored to another participant who has not yet become active, Pharaoh. Here the anchoring is titular rather than relational (cf. Section 2.1, end). 10 3. ANCHORING

The anchoring of Joseph's brothers

To give the participants a place in the story, they are often anchored. Different types of anchoring have been mentioned at the end of Section 2.1. In Section 2.2 we discussed some examples where the anchoring expressions are strikingly missing in the introduction of new participants in Genesis 37. Anchoring is used, however, throughout the story, for Joseph's brothers. They are anchored to Joseph ("his brothers"; also once anchored to Judah [Gen 37:26] and once to Reuben [Gen 37:30]) or Jacob ("his sons"). The switching reference expressions (switching between e.g. anchoring to Joseph and anchoring to Jacob) reflect various perspectives (cf. below, Section 3.2). However, if we take a closer look at the various reference expressions used for Joseph's family, it appears to be not only a group that is viewed from different perspectives, but also a rather diffuse group, because they are referred to with various designations, that do not exactly overlap:

Cf. Runge (2007:87 note 83; criticism of Heimerdinger): "On the other hand, consider the introduction of an unanchored generic referent as topic i[n] Gen 37:15 ‫ישׁ‬ ִ ‫א‬ ‫הוּ‬ ‫אֵ‬ ‫צָ‬ ‫מְ‬ ‫יִּ‬ ‫.וַ‬ The apparently acceptable introduction of 'a man' without numerical specification as the topic in this verse would seem to indicate that much less information is needed to bootstrap a participant than Heimerdinger allows". 10

Cf. Runge (2007:92 note 86). Perhaps we can regard Potiphar's introduction ("Potiphar, an officer of Pharaoh, the captain of the guard") as a case of a long designation of new participant, reflecting a negative assessment as described by De Regt (1999:5). 11

For more details see Van Peursen (forthcoming b).

1. "His brothers, the sons of Bilhah and the sons of Zilpah, his father's wives" (Gen 37:2); apparently excluding the two brothers mentioned by name later on in the chapter, Reuben and Judah, who are sons of Jacob's wife Leah.

2. "All his [Jacob's] sons" (Gen 37:3), apparently including Leah's sons. 12 3. "All his [Joseph's] brothers" (Gen 37:4): the same group as "all his sons", but with a different anchor.

4. "All his [Jacob's] sons and daughters" (Gen 37:35).

5. "His father and his brothers" (Gen 37:10).

6. "I and your mother and brothers" (Gen 37:10) in Jacob's response to "the sun, the moon and eleven stars" in Gen 37:9. The number eleven is only valid if Joseph's younger brother, Benjamin is included. 13 The various designations of Joseph's family show that in concrete texts participant reference may be a more complex phenomenon than in a theoretical framework. 14 Runge (2007:90) accounts for extensions of a subject by his category S1+, which he describes as "participant was the subject of the immediately preceding clause, and at least one other subject participant is added in the present clause to create a compound subject". However, to deal with the complexity we find in Genesis 37 we should at also introduce a context S1-(where at least one subject participant of the immediately preceding clause is omitted) as well as the categories N1+ and N1-for similar variations in non-subject contexts. 15 Without these 12 "All" may be used to mark the difference with the restricted reference in Gen 37:2; on Benjamin, a son of Jacob's wife Rachel, see the discussion below. 13

The inclusion of both Joseph's mother and his younger brother Benjamin is problematic because Gen 35:16-20 tells about Rachel's death at Benjamin's birth. Moreover, the reference to Benjamin combines badly with the story in which Joseph is Jacob's only favourite son. See Van Peursen (forthcoming b). 14 This was highlighted by Che Napoleon in his contribution to the MA-seminar mentioned in note 1. 15

In Runge's description the non-subject positions are designated by N; the numbers refer to the their contexts, e.g. INT = "introduction of a brand new participant"; S1 = "participant was the subject of the immediately preceding clause"; S2 = "participant was the addressee of a speech reported in preceding clause"; N1 = "participant was in the same non-subject role in the preceding clause". emendations of Runge's model, it is impossible to deal with, for example, the shifts from "his brothers" (non-subject context) to "his brothers and his father" (both in non-subject contexts) and again to "his father" in Gen 37:10.

The anchor as the central participant

The fact that Joseph's brothers are only referred to by anchoring expressions highlights their secondary role in the story. The counterpart of their secondary role as reflected in the way in which they are always anchored to someone else, is Joseph's role as the main anchor for the other participants; this reflects his status as central character (cf. Revell 1996:176). However, even though Joseph is the central participant, he is mainly described through the eyes of others: his father who loved him, his brothers who hated him, and the anonymous man who found him (cf. Berlin 1983:48).

In the second half of the story Joseph apparently loses control and is also grammatically driven into object roles rather than subject roles. Thus in Gen 37:15 "and a certain man found him … and asked him" -rather than "and he found a man" etc. -Joseph becomes the object of "find" and "ask" (cf. Fokkelman 1996:159). Joseph's subsequent arrival at his brothers is also told from the brothers' perspective, rather than from Joseph's. Compare Berlin's (1983:48) comment on Gen 37:18 "and when they saw him":

The reader travels with Joseph on his search for his brothers, but the perspective changes suddenly when he approaches their location. The point of view is that of the brothers -"They saw him from afar and before he could reach them they conspired to put him to death" (v. 18). It is from the brothers' point of view that we see this episode. They hate him, they plot, collectively and individually, they look up and see a passing caravan, they (presumably) sell Joseph, and then bring his blood-stained garment to their father.

Jacob is anchored to Joseph as well, in Gen 37:10, 11, 22 (unless the suffix pronoun refers to Reuben), 35. In some other cases, however, Jacob is the anchor. Anchoring to Jacob occurs in Gen 37:3 ‫יו‬ ‫נָ‬ ‫ל-בָּ‬ ‫כָּ‬ "all his sons" and To account for the subtleties described here, we can distinguish between "central", "main" and "dominant" participant. 16 Jacob is the main participant, because it is his story, his Toledot and as long as he lives, the stories dealing with his family are his stories. 17 Joseph is the central participant, in literary terms the "hero", to whom the other participants are anchored. In the first part of the chapter Joseph is also the dominant participant, but that is a role that he rapidly loses in the second half of the story.

VARIOUS TYPES OF PARTICIPANT REFERENCE

Once a participant is introduced, there are several means to refer to it: a proper noun, a common noun, a pronoun, verbal inflection or a combination of them (e.g. proper noun + common noun as in "David, the king"). Several scholars, including F I Andersen (1994), De Regt (1991-1992) and Runge (2006;, make a distinction between unmarked or default reference and marked reference, or, in other words, between minimal encoding (the minimal encoding that is needed for the unequivocal identification of the participants and the avoidance of ambiguity) and overencoding (any encoding that is more than the minimal encoding). In the case of overencoding, Runge discerns on the one hand a processing function, and, on the other hand, pragmatic functions. Where the encoding is less than the minimal encoding, we speak of underencoding. There is not a one-to-one relationship between these types of encoding and the various types of referring expressions. What in one context is the minimal encoding, may in another context be a marked encoding. In one context a proper noun is required to avoid ambiguity, and hence the "minimal encoding", in others it supplies "superfluous" information and hence is a case of overencoding. In this section we will discuss examples of the various types of encoding in Genesis 37. 16 For this distinction I am indebted to Che Napoleon's contributions to the MAseminar mentioned in note 1. 17

Cf. above, Section 2.2, and Van Peursen (forthcoming a).

Minimal encoding

As indicated above, the definition of the default or minimal encoding depends on the context. It may be inflection, a pronoun, a common noun or a proper noun.

4.1.1 Inflection.

The most common usage of this type of minimal encoding is when there is subject continuity, e.g.

"But when his brothers saw… they hated him."

In the case of a dialogue an explicit reference to the subject is not needed, even if the subject changes. This is evident, for example, in the dialogue between Israel and Joseph in Other examples can be found in the introduction of direct speech in Gen 36:6 ("and [he, i.e. Joseph] said to them [i.e. his brothers]") and in the dialogue between Joseph and the anonymous man in Gen 37:15-17.

A pronoun.

In some contexts the use of a pronoun is marked, as in

A common noun or a proper noun.

When a new participant is introduced, a common noun or a proper noun is the minimal encoding. Examples with both a proper noun ("Joseph", "Judah") and a common noun 20 ("his brothers") occur in: Here the explicit reference to Joseph's father is needed for disambiguation, although the fact that this proper noun is used rather than "their father" as in preceding clause, has a processing function (cf. below, Section 4.2). If a participant is already active, a proper noun is still the minimal encoding if there is a change of role, e.g., the object of the preceding clause becomes the subject (Runge 2007:102-103), as in:

Gen 37:4-5 ‫אוּ‬ ‫נְ‬ ְ ‫שׂ‬ ‫ יִּ‬ ֽ ‫וַ‬ ‫ֹתוֹ‬ ‫א‬ ‫ֹא‬ ‫ל‬ ‫וְ‬ ‫לוּ‬ ‫יָכְ‬ ‫רוֹ‬ ‫בְּ‬ ַ ‫דּ‬ ‫ם‬ ֽ ָ ‫שׁ‬ ‫לְ‬ ‫ם‬ ֲ ‫ח‬ ‫יַּ‬ ‫וַ‬ ‫ף‬ ‫יוֹסֵ‬ ‫לוֹם‬ ֲ ‫ח‬ "they hated him (i.e. Joseph) and could not speak peaceably to him. Now Joseph (already active, but now moved to Subject role) had a dream".

After this discussion of various patterns of minimal encoding, we will turn to cases where more than the unmarked minimal encoding is used. In Runge's model, all these cases are marked. He distinguishes two functions of these marked constructions: a processing and a pragmatic one. 20 Cf. also "the child" in Gen 37:30, quoted above. 21

Hence the reference to Judah functions to resolve ambiguity rather than to mark the start of a dialogue; cf. De Regt (1999:19); pace Longacre (1989:162).

Overencoding with processing function

The processing function relates to the interaction between participant reference and text structure. Overencoding is often used to mark the beginning or the end of a textual unit. Its task is "overcoming disruptions in the flow of information by providing more encoding" (Runge 2007:11). An example is Gen 37:3 ‫ל‬ ‫אֵ‬ ָ ‫ר‬ ְ ‫שׂ‬ ‫יִ‬ ‫וְ‬ ‫ב‬ ‫הַ‬ ‫אָ‬ ‫ף‬ ‫ת-יוֹסֵ‬ ‫אֶ‬ "Now Israel loved Joseph", with both participants explicit. In this case not only the explicit references mark the start of a new unit, but also the use of "Israel" rather than, e.g., "his father", which would link this verse closer with the preceding.

22

In direct speech we find:

Gen 37:33 ‫ת‬ ‫ֹנֶ‬ ‫ת‬ ‫כְּ‬ ‫י‬ ‫נִ‬ ‫בְּ‬ … ‫ֹף‬ ‫ר‬ ‫טָ‬ ‫ף‬ ַ ‫ֹר‬ ‫ט‬ ‫ף‬ ֽ ֵ ‫יוֹס‬ "My son's robe … Joseph is without doubt torn to pieces".

In this case the overencoding at the end of Jacob's direct speech gives the conclusion its own place and emphasis.

More than once there is a discrepancy between a segmentation of the text based on the processing function of participant reference and a division of the text into units in a literary approach. For example, one of the major representatives of the latter approach, J P Fokkelman (1996), distinguishes in Gen 37:2-11 three paragraphs (vv. 2-4, 5-6, 9-11), each of them closing with a remark about the brothers' hatred. This is followed by 37:12-17, which Fokkelman regards as an intermediate section. What follows after that is 37:18-33, a literary unit hold together by an elegant ABCDXD‫׳‬C‫׳‬B‫׳‬A‫׳‬ structure, in which the corresponding elements are characterized by recurring words or expressions and which places Judah's discourse in vv. 26-27 in the centre (the X in the structure). Thus A (37:18-20) and A‫׳‬ (37:31-33) both have "a ferocious animal devoured him" (in A in the brother's speech; in A‫׳‬ in Jacob's words) and a reference to Joseph's brothers; B (vv. 21-22) and B‫׳‬ (v. 30) both have Reuben's speech; C (vv. 23-24) and C‫׳‬ (v. 29) both have an empty cistern (in C: no water; in C‫:׳‬ no Joseph) and garments that are torn; and both D (v. 25) and D‫׳‬ (v. 28) deal with the tradesmen going to Egypt (see also Fokkelman 2000:79-82). Longacre (1989:144), too, considers Gen 37:12-17 as a distinct unit, bracketed by the proper noun "Joseph" in vv. 13 and 17.

The overspecification of Joseph may also be related to his change of role; cf. above, under "minimal encoding".

This literary structure, however, does not concur with a division of the text based on participant reference. In Gen 37:3, quoted above, the proper nouns "Israel" and "Joseph" seem to have a processing function, but in Fokkelman's analysis they occur in the middle of a paragraph. Gen 37:9 is the start of a new unit in Fokkelman's text division, but the use of "he" rather than an explicit reference expression suggests continuity. On the basis of participant reference, it is more appropriate to assume that Gen 37:17d rather than 37:18a marks the start of a new paragraph, because it has the explicit references "Joseph" and "his brothers" and because it contains a new location: "Dothan". 23 Perhaps we can also establish such a discrepancy between the two approaches in Gen 37:12/13. In Fokkelman's division Gen 37:12 is the start of a new unit, but the participant reference is less decisive. One could argue that the place indication "Shechem" in Gen 37:12 suggests indeed the start of a new unit, 24 but the referring expressions used suggest that 37:13 (with the proper nouns "Israel" and "Joseph") marks a new start rather than 37:12 (with the anchored expressions "his brothers" and "their father's flock").

Finally it should be noted that within the ABCDXD‫׳‬C‫׳‬B‫׳‬A‫׳‬ structure, in v. 29, the introduction of a new participant (Reuben) and the different time frame (at least implicitly: some time has past) suggest the start of a new section. Although in this case it goes perhaps too far to oppose the literary and the linguistic text division because also in a literary approach a break in events is suggested, because suddenly the question rises: Where had Reuben been while Joseph was sold? (cf. Berlin 1983:121).

Overencoding with pragmatic function

4.3.1 Thematic highlighting Runge (2007:151) distinguishes two subcategories within the third category, that of the pragmatic function: thematic and cataphoric. The thematic highlighting "assigns added prominence to the information that is crucial to understanding the interpretative point of the story". A certain type of overencoding is the use of supplementary reference expressions (e.g., "Isaac, his father", where "Isaac" would suffice; Runge 2007:157-161). Of this phenomenon, which is very common in some other stories in Genesis, 25

Underencoding

Underencoding is the phenomenon that less encoding is given than what we would expect in the default or minimal encoding. Underencoding may lead to ambiguity. This happens in 37:28, where the subject of a finite verb is underencoded: Underencoding does not always lead to ambiguity. In the case of a predicative participle, for example, the minimal encoding of the Subject in Biblical Hebrew is a personal pronoun. This means that the following example is a case of underencoding, even though there is no confusion about the identity of the subject:

CONCLUSIONS

The starting point of our investigation was Runge's model for the analysis of participant reference, which we applied to Genesis 37. Runge's model worked well, which by itself is an important conclusion, because of the large differences in participant reference between Runge's sample chapter (Genesis 27; cf. above, note 25) and the chapter under discussion (Genesis 37), and shows the broader applicability of this model.

28

Cf. Runge's (2007:72) criticism of Longacre's view that the repetition of the NP "Reuben" indicates that he is a thematic participant. What Runge expresses within his linguistic framework, was previously observed by Berlin in her literary approach. See Berlin (1983:118) on v. 22, where she argues that the repetition of "he said" may indicate a pause or a change of tone, rather than a later addition in the story. 29

For details see Van Peursen (forthcoming b).

However, from our analysis of Genesis 37 also some suggestions for modifications emerge, partly because this chapter contained some phenomena that receive little attention in Runge's model. We can summarize them as follows:

1. In the light of our analysis of Genesis 37 and the examples collected by Revell in his monograph, the introduction by name alone should be treated as an alternative way of participant introduction, rather than as exceptional.

2. A treatment of underencoding only in terms of exceptions is unsatisfying. In Genesis 37 we saw only two examples, but an analysis over a larger corpus would be worthwhile, both to discover the mechanisms that are at work and to see whether we can develop strategies to identify ambiguous participants.

3. The model needs elaboration to handle "diffuse participants" more appropriately, for example by adding "S1-", "N1+" and "N1-" to Runge's categories (above, Section 3.1).

Regarding the various patterns for anchoring new participants, we should remark that although Joseph, Reuben and Judah are not anchored in the typical way described by Runge and others, the family relations are made clear by explicit references such as Gen 37:2 "Joseph … with his brothers", and Gen 37:26 "Judah said to his brothers". It is worth considering applying a somewhat broader definition on anchoring, and not to restrict it to some established patterns such as the apposition construction of the type "Judah, their brother". Regarding Genesis 37 we could make the following observations:

1. The participants are not introduced in a way that is typical for the introduction of new participants. I have argued that this implies that Joseph, Judah and Reuben belong to the great heroes of the past and hence did not need to be anchored; unlike Bilhah and Zilphah, Jacob's wives, who are indeed anchored.

2. The variation in participant reference expressions give the various characters their place in the story;

a. Joseph is the central participant, but he is not dominant.

b. The secondary role of the brothers is underlined by the way in which they are anchored.

c. "The man" is a prop (cf. Runge, pace Longacre); he remains a face without a name.

3. The division of the text into "development units" on the basis of the processing function of participant reference does not run parallel with the division of the text on literary grounds. Thus there is a discrepancy between the linguistic text-syntactic approaches of Longacre, De Regt and Runge and the literary approach of Fokkelman and others. This raises the question as to the priority in the analytical procedure of the linguistic analysis and the literary analysis. Elsewhere we have argued for the priority of the linguistic analysis, expressing our agreement with the following quotation from Eep Talstra (1996:12):

[On the basis of] the assumption that a linguistic analysis referring to language as a system comes prior to a stylistic analysis referring to the phenomena that mark the structure of a specific textual composition, it is my view that observations on the level of grammar and lexicon should have priority over observations in terms of semantics or stylistics.

This does not deny that the literary analysis, especially when it leads to different results, poses challenges to the linguistic analysis. An analysis of the complex interaction between linguistic hierarchy, literary structure and text composition deserves further study. This is, however, beyond the scope of the present study.