For new authors:
free, easy and fast
For registered authors
17 Pages, Grade: 1.0
3.Centering Theory in contrast to the demonstrative description’s capability of ensuring
unexpected referential continuity
3.1.The resolution of demonstrative descriptions in correlation to referential continuity
3.2.1.The fundamentals of Centering Theory
3.2.2.An analysis of this -NPs
3.3.Is there a contradiction between Centering Theory and the anadeictic value of demonstratives to ensure referential continuity by “breaking” it?
Comparing the apparent contrasting statements of Centering Theory and Fossard, Garnham and Cowles (2012) in regards to demonstrative descriptions, I particularly examined reference resolution as an influencing factor of text coherence. According to the results of Fossard et al. (2012), the reference resolution of demonstrative descriptions was only facilitated when the predicative information referred to the subordinate character in a gender-ambiguous condition. As demonstrative descriptions have an exceptional discourse function due to their anadeictic dimension, they are capable of ensuring unexpected referential continuity by marking a discontinuity with the previous context. In regards to Centering Theory, however, a discourse is more coherent if utterances preserve the same topic and keep it as the highest-focused entity without diminishing its relative ranking. This continue transition state is considered most preferable. While demonstrative descriptions serve in an unexpected and unique way, their exceptionality seems to be beyond the Centering Theory’s stated preferences for text coherence and hence not against it.
What provides the cohesion of a text? Since readers or listeners have certain assumptions of what makes a text coherent, the writer or speaker has to convey those assumptions. In order to get a deeper insight into those ongoing mental processes, a variety of linguistic theories have been developed: e.g., theories about how pronouns or discourse connectives are processed or how text coherence can be characterized at all. One of the most significant factors that contribute to cohesion is continuity. Thus, salience-based approaches of reference resolution were conducted to further analyse anaphoric expressions and their referential continuity. Among them was the study of Fossard et al. (2012). This study particularly investigated the discourse function of demonstrative descriptions as they create referential continuity by shifting the reader’s attention to an unexpected referent. Moreover, a theory about Centering was formulated to state preferences for text coherence based on the manner of how discourse entities are introduced and discussed. This theory also focused on salience and reference resolution. In this context, it claimed that a discourse is more coherent if its utterances preserve the same topic and keep it as the highest-focused entity. For this would lead to a continue transition state which is most preferable.
In this paper I will investigate whether there is a contradiction between Centering Theory and the claim of Fossard et al. (2012) that a demonstrative description can ensure unexpected referential continuity by marking a discontinuity with the previous context. In order to contrast these two approaches, I will focus on the reference resolution of demonstrative descriptions first and outline the study of Fossard et al. (2012) with regard to demonstrative descriptions in particular. Then, I will present the essentials of Centering Theory that are relevant for the present paper. After introducing its fundamentals, I will proceed with an analysis of demonstrative expressions based on the first pilot study of Poesio (2008). Finally, I will discuss the two approaches and argue that they are not, in fact, mutually exclusive.
In the study of Fossard et al. (2012), the authors concluded that the demonstrative description preferentially orients processing towards the less salient referent in a discourse context when used anaphorically. The following section shall thereby provide a deeper insight into reference resolution of particularly demonstrative descriptions. It shall, moreover, present the study’s essential experimental results of whether the demonstrative description marks referential continuity within the previous discourse context.
Fossard et al. (2012) stated that there were two complemental discourse-referring management procedures, which are called anaphora and deixis. These operate on the mental model of the discourse and allow the interlocutors to coordinate their attention in a discourse. This is mainly achieved by their contrasting functions. Anaphora prototypically serves to maintain attention where a referent is already established. In doing so, it signals referential continuity. Deixis, on the other hand, shifts the attention to a new referent. According to the level of the referent’s accessibility in the addressee’s mental model, anaphoric forms are also characteristically used to refer to the highly-focused referent in the discourse representation; whereas, demonstrative expressions are preferentially interpreted as referring to the less salient referent, which is not in focus, but “activated”. In a nutshell, anaphora orients processing towards the highly-focused referent and thereby keeps the interlocutors’ attention, which is perceived as referential continuity. Deixis, by contrast, refers to the less salient referent and thus shifts the attention to a new referent. In contrast to the traditional view, the present study proposes that anaphoric and deictic expressions are not mutually exclusive indexical categories, but can be ordered by their relative degrees of deicticity and anaphoricity. In this connection, Fossard et al. (2012) refer to Cornish’s (2007) scale, which is enclosed below (see Figure 1). The middle of this scale shows various overlapping expressions between the two poles of pure deixis (1st and 2nd personal pronouns) and pure anaphora (3rd person reflexive pronouns). These are called “anadeixis”. The specific characteristic of anadeictic expressions, among which demonstratives (pronouns or noun phrases (NPs)) are the best example, is that their use comprises a partly anaphoric and a partly deictic reference. The demonstrative that, for instance, can serve three functions, of which only the first is relevant for the present study. In terms of the anaphoric function, that refers to an identifiable entity, established via NP (e.g., “Peter dreaded Suzie’s furies. That woman was unpredictable”). In regards to the deictic function, there is no identifiable entity displayed (e.g., “Look at that girl!”); whereas concerning the discourse-deictic function, the referent is created from the context since there is no independent discourse entity (e.g., “Peter pushed Suzie. That behaviour shocked her”). These three distinct possible functions exemplify the deictic and anaphoric dimensions of the demonstrative that as an anadeictic expression. Consequently, through their profoundly deictic dimension, they are capable of referring to a less salient referent. And through their anaphoric value, they assume a reference frame in which the intended referent is already introduced. This anadeictic value enables demonstrative expressions to fulfill a unique function in discourse construction when used anaphorically.
In the present study, Fossard et al. (2012) conducted two on-line studies to examine their hypothesis that, due to its deictic value, the demonstrative NP is a good indexical tool to access less salient referents which are, however, unexpected to ensure referential continuity even though they are already “known”. To gain a deeper insight into anaphor processing, they focused on the influencing factors of processing salience and conceptual gender agreement while contrasting the demonstrative NP with the third-person anaphoric pronoun—a prototypically anaphoric expression. In the course of their study, the authors suggested an adapted version of what they called “Sanford and Garrod’s model of anaphor processing” (Fossard et al. 2012: 1400). In this, they proposed that the “bonding process”, which is an automatic matching process between the anaphor and its antecedent, is guided by gender agreement and anaphor type. The latter either directs processing towards the main character (anaphor pronoun) or the subordinate character (demonstrative pronoun). After this first step follows the “reference resolution”, which aims to identify the referent of an anaphor to integrate it into the interpretation process. If there are two potential antecedents of different gender in a context, the pronoun is disengaged from the focused noun when its gender does not agree. This causes a higher activation of the anaphor’s referent than of the non-referent.
Figure 1. Cornish’s scale of anaphoricity and deicticity (Cornish, 2007: Fig. 1, p. 149).
Abbildung in dieser Leseprobe nicht enthalten
For their study, they also manipulated two factors in the target sentence: the form of the anaphor (anaphoric third-person pronoun he / she vs. demonstrative description that man / that woman) and the character type, which implies the saliency of the referent (highly-focused main character vs. less salient subordinate character).
Before the two on-line experiments, however, the authors conducted a norming study and a sentence-completion task—both without gender cue. Concerning the results of the norming study, which are relevant for the present paper, the authors stated that the participants did not consider a demonstrative description which referred to the Subordinate Character (the less salient character) an unusual referential choice. On the contrary, they thought demonstrative descriptions were more acceptable than pronouns when referring to the Subordinate Character and more appropriate when in reference to the Subordinate Character than the Main Character. In regards to the sentence- completion task, in which participants had to continue passages themselves, the authors affirmed their assumption that the third-person anaphoric pronoun was almost always chosen to maintain reference to the Main Character. The demonstrative description, by contrast, was preferred to indicate reference to the Subordinate Character, so Fossard et al. (2012) concluded that both anaphors are used very complementarily since they work in opposite ways.
In order to find out more about the time course of demonstrative resolution, they conducted two self-paced reading tasks. An example of the material used in both experiments is given in Table 1. In these, participants were shown three segmented sentences, among which the third represents the target sentence. In regards to the latter, two reading time measures were captured: for the “anaphoric segment” (the first part of the sentence, which contained the anaphor) and for the “predicative segment” (the second part of the sentence). Moreover, gender as another influencing factor for processing was taken into account for both self-paced reading tasks. For, gender agreement as a strong morphosyntactic/semantic cue allows the reader or listener to unambiguously decide which discourse entity could be the antecedent. Only the first of the two experiments, however, provided texts in the “gender cue” version, while the other used material in which the gender of the two characters was identical (either feminine or masculine). This manipulation was considered necessary to examine the correlation between the anaphoric/deictic properties of the demonstrative description and the morphosyntactic/semantic cue to resolution. The first sentence of each piece of experimental material referred to a highly-focused character (the Main Character) in an explicit setting (e.g., a restaurant). Through the following sentence it became apparent that the already presented character did not only serve as the first sentence topic, but also as the discourse topic. Thus it became the main protagonist of the specified situation since it occurred in the subject position, was introduced by a proper name and was then mentioned twice again by a third-person subject pronoun in the second sentence. While maintaining this character as the discourse topic, a second and less salient character was described by its role in the setting (e.g., “the waitress” in a restaurant scene). This subordinate character appeared in direct object position in the subordinate clause of the second sentence and hence played a tangential role in the presented discourse because of its deep embedded position in the sentence structure. In avoidance of the effect of the linear order—that the high-focused character appears first in the discourse and the less salient last—, the Main Character was repeated after the Subordinate Character, either via a possessive or nonsubject third-person pronoun. The target sentence as the last of the discourse was then constructed to refer to either the Main or the Subordinate Character, either via the anaphoric third-person pronoun he / she or via the demonstrative description that man / that woman —the verbs changed in each case. So, there were four alternative versions of the third sentence which varied in two essential factors. The first was character-type, which was implied semantically by the predication in the predicative segment, facilitating processing either towards the highly-focused character or the less salient one. The second factor referred to the form of the anaphor, which served as the grammatical subject of the anaphoric segment.
The key results concerning the demonstrative description in the first experiment were that in the anaphoric segment there was no essential effect on character type, but on anaphor type. The latter is, however, very likely due to the different character length between the pronoun and demonstrative. In the predicative segment, reading time was measured faster when it oriented toward the main character and when the anaphoric pronoun was used in the anaphoric segment than when the demonstrative description was used. This, however, might also be because of spillover processing from the anaphor. Concerning reference resolution of the demonstrative description, the predicative information for the less salient character was not integrated much faster with a demonstrative description than with a pronoun. This was striking for it revealed that in contrast to the pronoun, the convergence of saliency and gender cues was not sufficient for rapid integration of the preferred referent of the demonstrative description. A reason for this might be the “preliminary activation” of the Main Character in the comprehender’s mental representation. For, the highly- focused character serves as the primary target of inferential processes. Therefore, it is so strong and preliminarily activated that the way that reference resolution is delayed can affect the interpretation of the demonstrative description. Fossard et al. (2012) suggested this in their present study. Furthermore, they proposed that two incompatible factors were influencing the demonstrative description: firstly, it inherently preferred the subordinate character as referent, which also agreed in gender in the first case (see Example 1a)—while, secondly, making the main character easier accessible. As a consequence, the demonstrative description had to counteract the main character’s activation by restraining it or by raising the subordinate character’s activation in order to resolve the less salient and gender matching character.
illustration not visible in this excerpt
Note: The double slashes (//) mark the presentation of the text shown on the computer screen in the self-paced reading tasks (Experiments 1 and 2) (Fossard et al., 2012: Table 1, p. 1390).
1a. In fact, that woman simply recommended the dish of the day.
In the other case the gender of the main character matched with the demonstrative description (see Example 1b).This also delayed the integration of the predicative information since this gender agreement conflicted with the preferred referent of the demonstrative description.
1b. In fact, that man simply ordered the dish of the day.
These two cases, however, exemplify how time-consuming this processing is, which consequently made it hard to detect a significant effect for demonstrative descriptions in the first experiment of the present study. This also seems to contradict the results obtained in the sentence-completion task, which particularly stressed the definite preference of the demonstrative description to orient processing towards the subordinate character. Yet, the latter used experimental material with no gender cues, which was not the case until Experiment 2.
In contrast to Experiment 1, the second experiment used texts with no gender cues. This affected the processing of the demonstrative description as follows: when the predicative information was used to refer to the subordinate character, participants read the predicative segment faster with a
Research Paper (postgraduate), 7 Pages
Master's Thesis, 77 Pages
Bachelor Thesis, 48 Pages
Seminar Paper, 7 Pages
Diploma Thesis, 181 Pages
Scientific Essay, 13 Pages
Research Paper (undergraduate), 22 Pages
Research Paper (undergraduate), 17 Pages
GRIN Publishing, located in Munich, Germany, has specialized since its foundation in 1998 in the publication of academic ebooks and books. The publishing website GRIN.com offer students, graduates and university professors the ideal platform for the presentation of scientific papers, such as research projects, theses, dissertations, and academic essays to a wide audience.
Free Publication of your term paper, essay, interpretation, bachelor's thesis, master's thesis, dissertation or textbook - upload now!