3.4. Right-dislocation

Right-dislocation involves the same process as left-dislocation but with the movement rightwards of material, that is to the end of the sentence. Like clefting and existential, right-dislocation introduces a pronoun that in a sense announces what is coming up in a discourse. Unlike those other three constructions, though, here the pronoun is referential (and isn’t just an empty place-holder). This can be an effective way of building up expectations:

There he lay, a vast red golden dragon, fast asleep.

Topic 4. Discourse: Organization and Support

1. Discourse Processing

1.1. Methods of Studying Discourse Processing (Research Paradigms)

1.1.1. On-Line Paradigms

1.1.2. Off-Line Methods

1.2. Theories of Discourse Processing

1.2.1. Construction-Integration Model

1.2.2. Structure-Building Framework

1.2.3. Event-Indexing Model

1.2.4. Memory-Based Approach

1.3. Theoretical and Empirical Aspects of Discourse Processing

1.3.1. Integrating Sentences into a Coherent Discourse

1.3.2. Generating Inferences during Discourse Processing

1.3.3. Determining Reference in Discourse Processing

2. Elements Supporting Discourse: Discourse Markers

2.1. The Conversational Approach

2.2. The Grammatico-Syntactic Approach

2.3. The Discourse-Cognitive Approach

In our daily lives, we are often asked to comprehend language in units larger than the single sentence. For example, when we have a conversation, the meaning of a particular utterance may be based on our understanding of previous utterances (e.g., the meaning of ‘him’ in the sentence ‘‘Ask him where the briefcase is’’ depends on our understanding of the referent’s previous mention). When we read or listen to stories, the narrative is built up across many sentences, paragraphs, and typically chapters.

The preceding examples of language use are examples of discourse, and discourse processing is the processing involved in building meaning across several sentences or utterances.

1. Methods of Studying Discourse Processing

Over the past several decades, discourse-processing research has focused on the comprehension mechanisms that operate while readers comprehend a text or listeners understand a conversation – as it unfolds. The methods used in this pursuit have been a mixture of both on-line and off-line measures. On-line measures are often timing-based measures designed to assess the moment-by-moment operation of the language comprehension system. Off-line measures are typically based on responses such as paraphrases and answers to comprehension questions; such measures are not necessarily intended to directly tap the moment-by-moment operation of the comprehension system.

1. On-Line Paradigms

There are two major methodological paradigms that are based principally on on-line measures: the inconsistency detection paradigm and the probe response paradigm. These paradigms use reading times, lexical decision times, and naming times (among other measures) as primary dependent measures.

a. Inconsistency Detection The inconsistency detection paradigm was introduced by O’Brien and Albrecht (1992). Participants are asked to read a passage such as the following:

Jane woke up on Thursday morning in a panic. Her rent check was due the next day, but there was no money in her bank account. When she walked into the kitchen, she remembered that she hadn’t yet deposited the large refund check that she had gotten from the IRS. After getting dressed, she grabbed the check and drove to the beach.

The narrative begins with a sentence or two designed to introduce a fact about (or a goal for) the main character. In this case, Jane needs to get some money into her bank account. At the end of the narrative, there is a critical sentence (or clause) that conflicts with information presented earlier in the passage. In the example narrative, the bold clause at the end of the passage is the critical one: if Jane needed to get the check to the bank, readers should find it strange that she took the check to the beach. The comprehension difficulty slows the readers down, such that they read the critical sentence more slowly than a version of the same sentence that is consistent with the information presented at the beginning of the narrative (e.g., ‘‘After getting dressed, she grabbed the check and drove to the bank’’).

The inconsistency detection paradigm has been used to explore several issues in discourse processing. For example, the paradigm provides a means of assessing what information readers remember as they work their way through a narrative. If their reading times slow down for a sentence that is inconsistent with information provided earlier in the narrative, their slowed reading times suggest that they remember the previously presented information. The inconsistency detection paradigm has also been used to investigate how readers draw inferences during discourse processing. Sometimes the inconsistent information is inconsistent only with implicit rather than explicit information.

b. Probe Response

The second major paradigm that has been used in studies of discourse processing is the probe response task. Participants are asked to listen to or read passages such as the example presented above, and at some point during the passage the participants are presented with a probe word (i.e., a word that may or may not be related to the theme of the narrative). For example, at the end of the passage presented above, participants might see the probe word ‘bank.’ In one version of the probe response methodology, participants are asked to make a lexical decision on the probe word. If the concept named by the probe word (bank) is currently active in the comprehender’s representation, they should respond to the probe word faster than they would respond to a probe word that is unrelated to the narrative (e.g., ‘bird’ is a word that is unrelated to the ‘Jane’ passage). If the concept named by the probe word is not currently active, it should not be responded to faster than an unrelated probe word.

The probe response methodology has been used to study a wide range of topics in discourse processing. Variants of this task have been employed to assess when comprehenders draw inferences while processing narratives, the nature of the inferences drawn, the way that certain elements of a discourse remain active in memory (or not) as more of the discourse is processed, how comprehenders keep track of the characteristics and goals of the protagonists in a discourse narrative, and many other issues.

2. Off-Line Methods

Several kinds of off-line measures have been employed to understand the comprehension of discourse. Paraphrase tasks, in which participants are asked to paraphrase what they have heard or read, have been used to study what comprehenders remember about various aspects of a discourse (e.g., how much of the surface form do they remember? How many of the details do they remember?) and what inferences they have drawn. Another off-line measure that has been used is the question-answering task.

Participants’ answers to comprehension questions are used to assess the representations that comprehenders generated for a discourse. Think-aloud paradigms, in which participants are asked to think aloud about their developing understanding of a discourse, have been used to explore the comprehension strategies employed as comprehenders work their way through a discourse. Whereas there is some question about the extent to which these off-line measures provide a true reflection of the processing mechanisms that operate during, rather than after, comprehension, measures such as those listed above provide valuable information about the representations that are constructed from a discourse.