Hannah Rohde

A probabilistic reconciliation of coherence-driven and centering-driven theories of pronoun interpretation

Abstract Two classic theories of pronoun interpretation have each sought to specify the relationship between pronoun use and discourse coherence, but make seemingly irreconcilable claims. According to Hobbs (1979, 1990), pronoun interpretation is not governed by an independent mechanism, but instead comes about as a by-product of utilizing world knowledge during the inferential establishment of discourse coherence relations. Factors pertaining to the grammatical form and information structure of utterances do not come into play. According to Centering Theory (Grosz, Joshi & Weinstein 1995; inter alia), on the other hand, pronoun interpretation is predominantly determined by information structural relationships within and between utterances (e.g., topic transitions) and the grammatical roles occupied by potential referents. Factors pertaining to world knowledge and the establishment of informational coherence relations do not come into play. In this paper, we describe a series of psycholinguistic experiments that ultimately suggest a reconciliation of these diverse approaches. These experiments reveal a definitive role for coherence relationships of the Hobbsian sort, demonstrating that pronoun interpretation is affected by (i) probabilistic expectations that comprehenders have about what coherence relationships will ensue, and (ii) their expectations about what entities will be mentioned next which, crucially, are conditioned on those coherence relationships. However, these experiments also reveal a role played by the topichood status of potential referents. These data are reconciled by a probabilistic model that combines the comprehenders prior coherence-driven expectations about what entities will be referred to next and Centering-driven likelihoods that govern the speakers choice of referential form. The approach therefore situates pronoun interpretation within a larger body of work in psycholinguistics, according to which language interpretation results when top-down predictions about the ensuing message meet bottomup linguistic evidence.

Grammatical and information-structural influences on pronoun production

A standard assumption in psycholinguistic research on pronoun interpretation is that production and interpretation are guided by the same set of contextual factors. A line of recent research has suggested otherwise, however, arguing instead that pronoun production is insensitive to a class of semantically driven contextual biases that have been shown to influence pronoun interpretation. The work reported in this paper addresses three fundamental questions that have been left unresolved by this research. First, research demonstrating the insensitivity of production to semantic biases has relied on referentially unambiguous settings in which the comprehenders ability to resolve the pronoun is not actually at stake. Experiment 1, a story continuation study, demonstrates that pronoun production is also insensitive to semantic biases in settings in which a pronoun would be referentially ambiguous. Second, previous research has not distinguished between accounts in which production biases are driven by grammatical properties of intended referents (e.g., subject position) or by information-structural factors (specifically, topichood) that are inherently pragmatic in nature. Experiment 2 examines this question with a story continuation study that manipulates the likelihood of potential referents being the topic while keeping grammatical role constant. A significant effect of the manipulation on rate of pronominalisation supports the claim that pronoun production is influenced by the likelihood that the referent is the current topic. Lastly, the predictions of Kehler et al.s Bayesian analysis of the relationship between production and interpretation have never been quantitatively examined. The results of both experiments are shown to support the analysis over two competing models.

Where's Wally: the influence of visual salience on referring expression generation

Referring expression generation (REG) presents the converse problem to visual search: given a scene and a specified target, how does one generate a description which would allow somebody else to quickly and accurately locate the target?Previous work in psycholinguistics and natural language processing has failed to find an important and integrated role for vision in this task. That previous work, which relies largely on simple scenes, tends to treat vision as a pre-process for extracting feature categories that are relevant to disambiguation. However, the visual search literature suggests that some descriptions are better than others at enabling listeners to search efficiently within complex stimuli. This paper presents a study testing whether participants are sensitive to visual features that allow them to compose such “good” descriptions. Our results show that visual properties (salience, clutter, area, and distance) influence REG for targets embedded in images from the Wheres Wally? books. Referring expressions for large targets are shorter than those for smaller targets, and expressions about targets in highly cluttered scenes use more words. We also find that participants are more likely to mention non-target landmarks that are large, salient, and in close proximity to the target. These findings identify a key role for visual salience in language production decisions and highlight the importance of scene complexity for REG.

Form and function: Optional complementizers reduce causal inferences

Many factors are known to influence the inference of the discourse coherence relationship between two sentences. Here, we examine the relationship between two conjoined embedded clauses in sentences like The professor noted that the student teacher did not look confident and (that) the students were poorly behaved. In two studies, we find that the presence of that before the second embedded clause in such sentences reduces the possibility of a forward causal relationship between the clauses, i.e., the inference that the student teacher’s confidence was what affected student behavior. Three further studies tested the possibility of a backward causal relationship between clauses in the same structure, and found that the complementizer’s presence aids that relationship, especially in a forced-choice paradigm. The empirical finding that a complementizer, a linguistic element associated primarily with structure rather than event-level semantics, can affect discourse coherence is novel and illustrates an interdependence between syntactic parsing and discourse parsing.

Giving Good Directions : Order of Mention Reflects Visual Salience

In complex stimuli, there are many different possible ways to refer to a specified target. Previous studies have shown that when people are faced with such a task, the content of their referring expression reflects visual properties such as size, salience, and clutter. Here, we extend these findings and present evidence that (i) the influence of visual perception on sentence construction goes beyond content selection and in part determines the order in which different objects are mentioned and (ii) order of mention influences comprehension. Study 1 (a corpus study of reference productions) shows that when a speaker uses a relational description to mention a salient object, that object is treated as being in the common ground and is more likely to be mentioned first. Study 2 (a visual search study) asks participants to listen to referring expressions and find the specified target; in keeping with the above result, we find that search for easy-to-find targets is faster when the target is mentioned first, while search for harder-to-find targets is facilitated by mentioning the target later, after a landmark in a relational description. Our findings show that seemingly low-level and disparate mental “modules” like perception and sentence planning interact at a high level and in task-dependent ways.

Information Structure Prediction for Visual-world Referring Expressions

We investigate the order of mention for objects in relational descriptions in visual scenes. Existing work in the visual domain focuses on content selection for text generation and relies primarily on templates to generate surface realizations from underlying content choices. In contrast, we seek to clarify the influence of visual perception on the linguistic form (as opposed to the content) of descriptions, modeling the variation in and constraints on the surface orderings in a description. We find previously-unknown effects of the visual characteristics of objects; specifically, when a relational description involves a visually salient object, that object is more likely to be mentioned first. We conduct a detailed analysis of these patterns using logistic regression, and also train and evaluate a classifier. Our methods yield significant improvement in classification accuracy over a naive baseline.

Evoking Context with Contrastive Stress: Effects on Pragmatic Enrichment

Although it is widely acknowledged that context influences a variety of pragmatic phenomena, it is not clear how best to articulate this notion of context and thereby explain the nature of its influence. In this paper, we target contextual alternatives that are evoked via focus placement and test how the same contextual manipulation can influence three different phenomena that involve pragmatic enrichment: scalar implicature, presupposition, and coreference. We argue that focus placement influences these three phenomena indirectly by providing the listener with information about the likely question under discussion (QUD) that a particular utterance answers (Roberts, 1996/2012). In three listening experiments, we find that the predicted interpretations are indeed made more available when focus placement is added to the final element (to the scalar adjective, to an entity embedded under the negated presupposition trigger, and to the predicate of a pronoun). These findings bring together several distinct strands of work on the effect of focus placement on interpretation all in the domain of pragmatic enrichment. Together they advance our empirical understanding of the relation between focus placement and QUD and highlight commonalities between implicature, presupposition, and coreference.

In-situ and ex-situ wh-question constructions in Moro

Abstract This paper addresses the formation of wh-questions in Thetogovela Moro, a Kordofanian language spoken in the Nuba Mountains of Sudan. Moro has both in-situ and ex-situ wh-questions, but exhibits a subject/non-subject asymmetry: while non-subjects may employ either construction, subjects must appear in the ex-situ form. Ex-situ wh-questions are analyzed as wh-clefts, and they share several properties with clefts and relative clauses. The fronted element is marked with a cleft particle and for noun phrases, a demonstrative that we analyze as a relative pronoun is used. Verbal tone patterns are those that are found in dependent clauses rather than main clauses. Subject questions, clefts and relative clauses are marked with a verbal prefix é-, while non-subject questions, clefts and relative clauses are marked with a verbal prefix ə́-. We analyze these prefixes as dependent clause markers and provide evidence of additional dependent clause uses in the language. Finally, non-subject wh-questions bear an optional particle nə́- on the subject and/or verb. We offer several arguments that this is best analyzed as a complementizer

Recovering discourse relations: Varying influence of discourse adverbials

Discourse relations are a bridge between sentence-level semantics and discourselevel semantics. They can be signalled explicitly with discourse connectives or conveyed implicitly, to be inferred by a comprehender. The same discourse units can be related in more than one way, signalled by multiple connectives. But multiple connectives aren’t necessary: Multiple relations can be conveyed even when only one connective is explicit. This paper describes the initial phase in a larger experimental study aimed at answering two questions: (1) Given an explicit discourse adverbial, what discourse relation(s) do naive subjects take to be operative, and (2) Can this be predicted on the basis of the explicit adverbial alone, or does it depend instead on other factors?

Aspects of a theory of pronoun interpretation

On Processing Models and Their Complexity Several commentators ask how our proposal fits within a larger theory of language interpretation and processing, so let us start there. Following Rohde, Levy & Kehler (2011) and others, we envision our results situated in a strongly incremental theory, one in which both sentence-level and discourse-level decisions are made utilizing a broad variety of information sources on a moment-by-moment basis. This comes with two important ramifications. First, it predicts a particular time course of processing, whereby linguistic expressions influence context at the time they are encountered, and hence many only affect the interpretation of a subsequent pronoun indirectly. For instance, in interpreting a passage like Amanda fired Brittany and she immediately filed a lawsuit, the comprehender will have a particular set of biases toward coherence and next mention after the first clause (presumably favoring Explanation and Brittany respectively, since fired is an object-biased IC verb), which then get updated when and is encountered, and then again when the pronoun she is encountered. Crucially, the updates proceed along opposite paths in these two cases. On the one hand, and updates P(CR) to increase the likelihood of the coherence relations with which it is most compatible (Occasion, Result, Parallel) and to decrease it for others (e.g., Explanation will drop to near zero), which cascades to update P(referent). On the other hand, she updates P(referent) to become P(referent | pronoun) (by including the production probability P(pronoun | referent)) which, in pulling the distribution toward the subject referent, cascades to update P(CR) to increase the likelihood of subject-biased relations per the results of Rohde & Kehler (2008). The second ramification is that interpretation does not stop there; these probabilities will continue to be updated as subsequent linguistic material is encountered through the end of the clause. This answers Arnold’s and de Hoop’s question about how an ultimate