Does Text Genre Variability Play A Part

Published Date: 02 Nov 2017

Abstract: This study was an attempt to investigate the effect of text genre variability on the identification of referents and the choice of various appropriate anaphoric expressions. A total number of 35 homogenous graduate and undergraduate female EFL learners were opted for the study. The data was mainly collected via email messages. Two English cloze tests including an expository text and a narrative one were sent for all the participants through an email message. The participants were provided ample time to study the texts carefully, identify the ellipted referents and choose appropriate anaphoric expressions. The results of qualitative analysis of the data indicated that: a) in some environments identification is difficult regardless of short referential distance; b) there may be NP slots in which the majority of participants identify the referent differently from the writer; c) the presence of grammatical constraint dictates full NP selection conformity; d) recency of prior mention strongly determines the selectional performance; e) NP slot may contain referents whose mentions are purely optional; f) a long referential distance, between five to 20 clauses intervening between the two subsequent mentions of a given referent, may not impact on the difficulty of identification of referent; g) there is no identical pattern of reference in all cases of NP selection. In all, concerning the identification of referents, the results revealed a significant difference between expository and narrative text genres (p=.003 0.05) though the difference between means were not high. With regards to the choice of anaphoric expressions, the results suggested significant difference between narrative and expository texts in terms of choice of NP of (p=.003 0.05) and PRN (p=.000 0.05) .

Keywords: Text genre variability, expository text; narrative Text; identification of referents; choice of anaphoric expressions;

1. Introduction

The topic anaphora refers to the linguistic referential device used, the extra linguistic entity referred to, or the relationship that can be assumed between coreferring pronouns and antecedents. This linguistic issue has been subject to a variety of different terminologies, cf. referential choice (Clancy, 1980), continuity marking device (Givon, 1983), the syntax of reference (Tolmin, 1987), anaphora (Fox, 1987), accessibility markers (Ariel, 1990), and so forth (For Further information, see Eslami Rasekh, 1997). Anaphora resolution (AR) which most commonly appears as pronoun resolution is the problem of resolving references to earlier or later items in the discourse. In other words, it is the process of finding that previous item (Mitkov, 2002).

According to Huang (2000, p. 302), the problem of anaphoric distribution in discourse boils down to this:

"on the one hand, from the perspective of anaphoric production, what contributes to the speaker's choice of an appropriate anaphoric form; and on the other, from the vantage point of anaphoric resolution, what enables the addressee to identify the intended referent of that form at a given point in discourse (cf., e.g., Sacks and Schegloff, 1979; Fornel, 1987; Huang, 1989, 1994; Gundel et al., 1993)?"

The present study was to examine the effect of text genre variability on the identification and the choice of various appropriate anaphoric expressions. The task students involved in was comprehension of two pieces of English expository and narrative texts and production of appropriate anaphoric expressions, the former being a receptive performance and the latter a productive one. In other words, this study is to answer the following questions:

Does text genre variability (narrative vs. expository) have any effect on the identification of the appropriate anaphoric expressions?

Does text genre variability (narrative vs. expository) have any effect on the choice of different anaphoric expressions?

2. Theoretical Framework

This section reviews three ongoing research approaches in the study reference: the context (givenness) approach, the distant approach, and the focus approach. The three approaches are taken chronologically, beginning with the context theories, which were initiated earlier than the other two. Distant approaches evolved from givenness studies. Focus theories should be considered as the most recent and advanced of all.

2.1. The Context Approach to Discourse Anaphora

The context approach begins with theories of givenness that were initiated earlier than other approaches to the study of anaphora resolution. A referential concept is taken to be active after introduction, and is believed to remain active within an episode; it loses the status when the episode is closed down and a new discourse begins (Eslami Rasekh, 1997).

2.1.1. Definiteness

Definiteness can fulfill "the maxim of antecedent" (clark &Halivand, 1977) which is related to coherence and meets the demands of a topical referent. The maxim of antecedent also meets a basic condition of coherence in discourse. The condition for definiticity of a referent is set by Allen as: "speaker uses a definite NP where he judges that hearer can, at least conceptually, identify the NP’s denotatum from background knowledge or from information evoked by or inferred from context" (Allan 1986b, p.125).

Early research on definiteness by philosophers (Frege 1892; Russell 1905; Strawson 1950) focused on definite descriptions, singular noun phrases equipped with the definite article. According to Laury (2001), while the views of these scholars differed in significant ways, a central insight in their work was that the use of definite descriptions in general presupposed a unique referent.

Abbott (2001) believes that what definite NPs all have in common is that they conventionally implicate a unique referent. This minimal semantic characterization, abstracts away from contexts including addressees and their assumed abilities. This idea of having a minimal conventional meaning, and deriving particular usage phenomena pragmatically, is suggested by the analysis of "pragmatic definites" in Löebner (1987, P.319).

2.1.2. Discourse Givenness

Chafe (1976, 1987) defines given information as "knowledge which the speaker assumes to be in the consciousness of the addressee" (1976, p.30). In Chafe’s initial binary framework of given and new, given information is previously activated, whereas new information is activated only by the current segment of text. Chafe then introduces a distinction between new, given, and a third category, ‘quasi-given’ (1976, p. 34). This third category is related to the inferential availability of information, and has been a central concept in modern approaches.

Prince (1981) identifies three different sources of givenness. First, Predictability/Recoverability is based on the speaker’s assumption "that the hearer can predict or could have predicted that a particular linguistic item will or would occur in a particular position within a sentence" (1981). Second, Saliency is based on the speaker’s assumption "that the hearer has or could appropriately have some particular thing/entity/… in his/her consciousness at the time of hearing the utterance." Third, Shared Knowledge is based on the speaker’s assumption "that the hearer ‘knows,’ assumes, or can infer a particular thing (but is not necessarily thinking about it" (Hempelmann et al., 2005).

2.2. The Distant Approach to Discourse Anaphora

Sentential givenness allows only for a binary division between given and new, but a diverse set of referring expressions are used to mark various degrees of accessibility. The importance of a topical referent has the effect of forcing it to occur before the comment; the continuity of the topical entity, however, pushes it to the end of the sentence. The principle predicts "a more predictable topic follows the comment, a less predictable topic precedes the comment" (Givon, 1989, p.225).

2.2.1. Topic Continuity (TC)

This approach is best represented by Givon (1983, 1985, &1990). The main premise of this model is that anaphoric encoding in discourse is essentially determined by topic continuity. Continuity of topic in discourse is measured primarily by factors such as linear distance (the number of clause/sentence between the two mentions of a referent), referential interference (the number of interfering referents), and thematic information (maintenance or change of the protagonist). Roughly, what the model predicts is this: the shorter the linear distance, the fewer the competing referents and the more stable the thematic status of the protagonist, the more continuous a topic, the more likely that it will be encoded in terms of a reduced anaphoric expression. Thus, on this account, different types of anaphoric expression may just be hierarchically reordered as topic-coding devices (e.g. Givon, 1983); (For further information, see Huang, 2000). The major criticism against TC is that referential distance does not account for unusual referential choices across episode boundaries (cf. Eslami Rasekh, 1997). The claims made by topic continuity pertain to narrative genre; other genres may reveal different structural characteristics (Allan, 1987b).

2.2.2. Accessibility Theory (AT)

Many theories of reference (e.g., Ariel, 1990; Chafe, 1994; Givón, 1983; Prince, 1981) assume that speakers choose referring expressions (e.g., pronouns, definite noun phrases) depending on how accessible they believe the referent is in the addressee's mental representation. When the referent is assumed to be highly accessible for their addressee, speakers produce less explicit referring expressions such as pronouns. In contrast, when the referent is assumed to be less accessible for the addressee, they tend to use more explicit referring expressions such as definite noun phrases or proper names to facilitate comprehension. That is, the choice of referring expressions is driven by the speaker's assumptions about the addressee's current focus of attention (Fukumura & van Gompel, 2009).

Ariel (1990, 2001), develops an Accessibility Theory in which high accessibility markers consist of less linguistic material, and signal the default choice of continued activation. By contrast, low accessibility markers consist of much linguistic material, and signal termination of activation of the current (topical) referent, and the (re)introduction of a different referent. Consistent with Ariel (1990), Arnold & Tanenhaus (2007) hypothesize that reference to highly accessible or focused entities, that is, those entities that could be referred to with a pronoun, will be produced less prominently than entities that are not in focus and could not naturally be referred to with a pronoun.

2.3. The Focus Approach to Discourse Anaphora

The choice of referring expressions appears to be very much a function of episodic boundaries rather than the effect of distance. Referential distance fails to account for the significant role of episode boundaries, ignoring a large number of unusual referential choices. Proponents of the recency model concede that there are a great many potential counter examples but suggest no systematic explanation for them (Tomlin, 1987).

2.3.1. Centering

Centering (Grosz et al., 1983; Grosz et al., 1995) is intended to model discourse coherence, inference by conversational participants, and anaphora resolution (Beaver, 2003). The basic idea underlying this approach is that anaphoric encoding in discourse is largely determined by cognitive processes such as activation and attention. Activation and attention are two intimately related, though distinct cognitive phenomena. Roughly, in the domain of reference, activation of a referent in one's current short-term memory at moment is a result of focusing one's attention on that referent at a previous moment (e.g. Chafe, 1994; Kibrik, 1996).

In other words, the central empirical claim of the centering model is that full NPs are predicted to be used when the targeted referent is currently not addressee-activated, whereas reduced anaphoric expressions such as pronouns and zero anaphors are predicted to be selected when such a referent is estimated to be currently both speaker- and addressee-activated (e.g. Tomlin, 1987; Tomlin and Pu, 1991; Gundel et al., 1993; Kibrik, 1996).

2.3.2. Hierarchical Discourse Structure

Under this approach, it is assumed that the most important factor that influences anaphoric selection is the hierarchical structure of discourse. From this assumption, mentions (initial or non-initial) at the beginning or peak of a new discourse structural unit tend to be done by a full NP, whereas subsequent mentions within the same discourse structural unit tend to be achieved by a reduced anaphoric expression(Huang, 2000).

The most representative of the hierarchy model is Fox's (1987) study of discourse anaphora in both English written texts and conversational discourse. The key concept underlying Fox's analysis is that anaphoric encoding and discourse organization are closely correlated. She assumes two modes of description: what she calls the 'context-determines-use' mode and the 'use-accomplishes-context' mode. In the first, the hierarchical structure of discourse is claimed to determine the choice of a particular anaphoric form; in the second, the use of a particular anaphoric form is said to establish a particular structural pattern of discourse. In other words, on Fox's view, anaphoric distribution is largely determined by, and itself determines, the hierarchical structure of discourse.

3. Research methodology

3.1. Participants

A total number of thirty-five advanced Iranian graduate and undergraduate students of English, studying or studied English at the University of Isfahan, Faculty of Foreign Languages, and Department of English participated in this study. Their age ranged between 21 and 28. The sample consisted of mostly senior students. In order to control the effect of gender on the result of the study, just female students were selected.

3.2. Instrumentation

Two rational cloze tests, each with 35 NP slots, including one narrative text and one expository text were used in this study. 35 NP slots were selected because research indicated that the length of a test improves the reliability up to a certain point. Up to 30-35 items, the reliability increases sharply. Beyond 35 items, however, the increase of reliability is so smooth that it can be ignored (For further information, see Farhady, Jafarpur, & Birjandi, 2007).

3.2. 1. Expository Cloze Test

A piece of expository text with 35 empty slots pertaining to topical participants was given to the students to study carefully and identify the ellipted referents of the empty NP slots (see Appendix 1). Only the first mention of the topical participants of the expository text was given; all the subsequent references were indicated by empty NP slots left for the participants to spell out.

The readability of the text selected for this study came out to be 28, meaning very difficult. Among the five instructors, who taught reading courses both at higher academic levels, three considered it difficult and appropriate for advanced students; one consider it difficult, however, suitable for upper intermediate students; and one evaluated it fairy difficult and appropriate for advanced students. In order to make sure about the level of difficulty of the selected text, in a pilot study, the text was given to both advanced and upper intermediate students. The indexes of item facility and item difficulty of each slot were calculated; furthermore, participants’ own evaluations of the level of difficulty of the text were checked out. Results revealed that the text is mostly appropriate for advanced students rather than upper intermediate ones. In all, based on the Flesch-Kincaid readability formula, five instructors’ ideas, and pilot study, the text was selected for advanced students.

The text selected for this study was a piece of news titled Holder defends terror trials in civilian courts revealed by Associated Press, June 16, 2011. The text was written by journalist Nedra Pickler. The news story was about Attorney General Eric Holder who were fought over the prosecution of terrorism suspects in civilian court and urged by the top-ranking Senate Republican Mitch McConnell to send two Iraqis Waad Ramadan Alwan and Mohanad Shareef Hammadi to Guantanamo Bay rather than try them in Kentucky.

3.2. 2. Narrative Cloze Test

A piece of narrative text with 35 empty slots pertaining to topical participants was given to the students to study carefully and identify the ellipted referents of the empty NP slots (see Appendix 2). Only the first mention of the topical participants of the expository text was given; all the subsequent references were indicated by empty NP slots left for the participants to spell out.

The story selected for this study was taken from the book The Chamber by John Grisham, retold by Sue Harmes Series Editors: Andy Hopkins and Jocelyn Potter. The book has been evaluated by Penguin Readers, designed in association with Longman, appropriate for advanced students. The Chamber consisted of fourteen chapters. The story selected for this study was opted from the first chapter, A Delicate Exercise. The story was about the Mississippi Klan leader Dogan who accompany with Sam Cayhall and Rollie Wedge planned to assassinate the Jewish lawyer Martin B. Kramer.

3.3. Data Collection Procedures

The data of this study was mainly collected via email messages. Ten days before the main study, a pilot study using participants with a background similar to the participants of the main study was conducted in an effort to verify the appropriateness of all the tests designed for the study. Based on the result of the pilot study, internal consistencies (Cranach’s alphas) were calculated for both expository and narrative text genres; though cloze tests are not potentially reliable due to their interrelatedness of their items, internal reliability estimates were respectable, ranging from .68 for the thirty five slots on the expository text genre to .73 for the thirty five slots on the narrative text genre. Next, two English text genres were sent for all the participants through email message. The participants were provided ample time to study the texts carefully, identify the ellipted referents and choose appropriate anaphoric expressions.

3.4. Data Analysis & Scoring

Appropriate descriptive statistical procedures such as frequencies, percentages and means were used to interpret the results of the study quantitatively. In addition, a paired t-test was run to determine the significant difference between students’ performance in both identification of referents and choice of referring expressions within the expository text genre vs. Narrative one. Concerning scoring, In order to score the cloze tests, the acceptable scoring method was used. That is, any response which is acceptable in the context obtained a score and any incorrect response was not given any.

4. Findings

4.1. Students’ Performance in Identification of the Referents

Following are the results of the analysis of expository text concerning identification of referents. The analysis of data will be presented quantitatively slot by slot.

In order to ascertain whether there is a significant difference in identification of referents between the two text genres, the paired sample t-test was performed. The results suggested significant difference between narrative and expository texts in terms of identification of referents (p=.003 0.05), though the difference between means were not high; students could identify 81.06% of referents in narrative text genres and 63.59% of referents in expository one (see tables 4.1.3, 4.1.4).

4.2. Students’ Performance in the Choice of the Referring Expressions

In analysis of the choice of the referring expressions, the percent of the use of full NP in comparison to the use of pronoun investigated quantitatively. The term full NP refers to any definite referent marked by a non-pronominal expression, whether or not the expression in question is a first name, a last name, a noun without any attachment, or a noun with a determiner, a title, a description, etc. In analysis of the choice anaphoric expressions, first the original writer’s choice of expressions will be presented, then participants’ choices will be discussed.

In order to ascertain whether there is a significant difference in the choice of referring expressions between the two text genres, the paired sample t-test was performed. The results suggested significant difference between narrative and expository texts in terms of choice of NP of (p=.003 0.05) and PRN (p=.000 0.05) (see tables 4.2.8, 4.2.9, 4.2.10, & 4.2.11).

5. Discussion

According to Brewer’s (1980), narrative texts describe events that occur through time and are ‘related through a causal or thematic chain’ (p. 223). They are more likely to convey life experiences, person-oriented dialogue, and familiar language in the oral tradition than are expository texts (Bruner, 1986; Rubin, 1995; Tonjes, Wolpow, & Zintz, 1999). Although narrative texts may contain unfamiliar information, most readers have, from first-hand experience, well-developed schemas about the settings, actions, and events described by them (Cot´e et al., 1998; Nelson, 1996). Accordingly, most readers possess the knowledge necessary to understand narrative texts. Furthermore, narrative texts are assumed to use high-frequency word. Many narrative texts also follow a simple structure—a sequence of causally related events. As a result, readers usually understand narrative text structure (e.g., Williams et al., 2004). Thus, working memory requirements are reduced and comprehension is facilitated (Cain, 1996; Cot´e et al., 1998). Therefore, identification of referents is easier and participants tend to use more pronominal expressions than nominal ones.

Expository texts, on the other hand, describe the structure and processes involved in a system or event. In contrast to narrative texts, the purpose of expository texts is to inform the reader of new information (McNamara, Graesser, & Louwerse, in press). Thus, they tend to include less familiar concepts and words, and require more inferences on the part of the reader. Expository texts generally demand more knowledge that fewer people possess and tend to place increased processing demands on the reader due to their greater structural complexity, greater information density, and greater knowledge demands (McNamara, Graesser, & Louwerse, in press). In terms of structure, expository texts often contain abstract and logical relations, which can be difficult to interpret (Stein & Trabasso, 1981); (For further information, see Best, Floyd, &McNamara, 2008, pp.139-140). According to Fox (1987), English expository texts are structured into rhetorical units within which propositions carry core or ancillary meaning. The rhetorical units are analyzed based on one’s judgment about how meaning is organized into core and adjuncts based on semantic criteria. Accordingly, identification of referents in an expository text contains more cognitive cost than a narrative one; this cognitive demand leads readers to choose nominal expressions to supply the missing referents than pronominal ones (see Eslami Rasekh, 1997).

6. Conclusion

The present study was aimed at examining the effect of text genre variability on the identification and the choice of various appropriate anaphoric expressions. Two English cloze tests including an expository text and a narrative one were sent for all the participants through email message. The participants were provided ample time to study the texts carefully, identify the ellipted referents and choose appropriate anaphoric expressions.

The results of qualitative analysis of both narrative and expository text genres indicated that: a) in some environments identification is difficult regardless of short referential distance (cf. expository slots 7, 9, 31, 32 & 33; narrative slots 34 & 35); b) there may be NP slots in which the majority of subjects identify the referent differently from the writer (cf. expository slots 31 & 32; narrative slots 20 & 21); c) the presence of grammatical constraint dictates full NP selection conformity (cf. expository slots 12, 14, & 18; narrative slot 10); d) recency of prior mention strongly determines the selectional performance (cf. expository slots 34 & 35; narrative slot 3,4, 5, 6, & 35); e) NP slot may contain referents whose mentions are purely optional (cf. expository slots 7 & 8); f) a long referential distance, between five to 20 clauses intervening between the two subsequent mentions of a given referent, may not impact on the difficulty of identification of referent (cf. expository slots 25, 26 & 27; narrative slot 31) ; g) there is no identical pattern of reference in all cases of NP selection.

In all, concerning the identification of referents, the results revealed a significant difference between English expository and English narrative text genres (p=.003 0.05), though the difference between means were not high. With regards to the choice of anaphoric expressions, the results suggested significant difference between narrative and expository texts in terms of choice of NP of (p=.003 0.05) and PRN (p=.000 0.05) .

This study like other studies might have many inherent limitations. The major drawback of the study is the small size of the sample with 35 female advanced EFL learners, which may not have been a true representative of the larger population of Iranian EFL learners. Furthermore, the data was obtained from only one expository and one narrative text genres; thus not allowing for a more exhaustive generalization about the effect of text genre variability in identification of referents and choice of referring expressions. Given the fact that the study examined only advanced students, the results may not be true about other levels of language proficiency. Therefore, more studies can be conducted to make out whether the findings of the study are true with larger number of expository and narrative texts, across different levels of language proficiency including larger sample of the participants from different educational settings.