Petra Hendriks

Optimality theoretic semantics

The aim of this article is to elucidate the processes that characterize natural language interpretation. The basic hypothesis is that natural language interpretation can be characterized as an optimization problem. This innovative view on interpretation is shown to account for the crucial role of contextual information while avoiding certain well-known problems associated withcompositionality. This will become particularly clear in the context of incomplete expressions. Our approach takes as a point of departure total freedom ofinterpretation in combination with the parallel application of soft constraints on possible interpretations. These constraints can be contextual, intonational or syntactic in nature. The integration of pragmatic andsyntactic/semantic information in a system of ranked constraints is proposed to correctly derive the optimal interpretations in cases of nominal anaphorization, determiner quantification and elliptical comparatives.

When production precedes comprehension: An optimization approach to the acquisition of pronouns

Data from child language comprehension show that children make errors in interpreting pronouns as late as age 6;6 yet correctly comprehend reflexives from the age of 3;0. On the other hand, data from child language production show that children correctly produce both pronouns and reflexives from the age of 2 or 3. Current explanations of this asymmetric delay in comprehension either have rejected the comprehension data outright or have argued that the problems are pragmatic or caused by processing limitations. In contrast, our account, formulated in the framework of Optimality Theory, handles the comprehension data as well as the production data by arguing that children acquire the ability to take into account the alternatives available to their conversational partner relatively late. It is this type of bidirectional optimization, we argue, that is necessary for correctly interpreting pronouns.

Cognitive architectures and language acquisition: A case study in pronoun comprehension

In this paper we discuss a computational cognitive model of childrens poor performance on pronoun interpretation (the so-called Delay of Principle B Effect, or DPBE). This cognitive model is based on a theoretical account that attributes the DPBE to childrens inability as hearers to also take into account the speakers perspective. The cognitive model predicts that child hearers are unable to do so because their speed of linguistic processing is too limited to perform this second step in interpretation. We tested this hypothesis empirically in a psycholinguistic study, in which we slowed down the speech rate to give children more time for interpretation, and in a computational simulation study. The results of the two studies confirm the predictions of our model. Moreover, these studies show that embedding a theory of linguistic competence in a cognitive architecture allows for the generation of detailed and testable predictions with respect to linguistic performance.

Referential choice across the lifespan: why children and elderly adults produce ambiguous pronouns.

In this study, children, young adults and elderly adults were tested in production and comprehension tasks assessing referential choice. Our aims were (1) to determine whether speakers egocentrically base their referential choice on the preceding linguistic discourse or also take into account the perspective of a hypothetical listener and (2) whether the possible impact of perspective taking on referential choice changes with increasing age, with its associated changes in cognitive capacity. In the production task, participants described picture-based stories featuring two characters of the same gender, making it necessary to use unambiguous forms; in the comprehension task, participants interpreted potentially ambiguous pronouns at the end of similar orally presented stories. Young adults (aged 18–35) were highly sensitive to the informational needs of hypothetical conversational partners in their production and comprehension of referring expressions. In contrast, children (aged 4–7) did not take into account possible conversational partners and tended to use pronouns for all given referents, leading to the production of ambiguous pronouns that are unrecoverable for a listener. This was mirrored in the outcome of the comprehension task, where children were insensitive to the shift of discourse topic marked by the speaker. The elderly adults (aged 69–87) behaved differently from both young adults and children. They showed a clear sensitivity to the other persons perspective in both production and comprehension, but appeared to lack the necessary cognitive capacities to keep track of the prominence of discourse referents, producing more potentially ambiguous pronouns than young adults, though fewer than children. In conclusion then, referential choice seems to depend on perspective taking in language, which develops with increasing linguistic experience and cognitive capacity, but also on the ability to keep track of the prominence of discourse referents, which is gradually lost with older age.

Coherence relations, ellipsis, and contrastive topics

It has been observed (Kehler 1996, 2000, 2002) that ellipsis resolution processes interact with the inference processes underlying the establishment of coherence relations in discourse. For example, gapping only co-occurs with the coherence relation of Resemblance. In this paper I show that the reason why certain ellipsis processes only co-occur with certain types of coherence relations does not lie in the (im)possibility to reconstruct the missing material. Rather, ellipsis processes differ in their relation to the topic of the sentence. The way in which different coherence relations construct their topic (i.e. as a contrastive topic or as a non-contrastive topic) restricts the types of ellipsis they can occur with. This conclusion is supported by observed differences between gapping and subject deletion in Dutch SGF-constructions.

Age Differences in Adults' Use of Referring Expressions

The aim of this article is to investigate whether choosing the appropriate referring expression requires taking into account the hearer’s perspective, as is predicted under some versions of bidirectional Optimality Theory but is unexpected under other versions. We did this by comparing the results of 25 young and 25 elderly adults on an elicitation task based on eight different picture stories, and a comprehension task based on eight similar written stories. With respect to the elicitation task, we found that elderly adults produce pronouns significantly more often than young adults when referring to the old topic in the presence of a new topic. With respect to the comprehension task, no significant differences were found between elderly and young adults. These results support the hypothesis that speakers optimize bidirectionally and take into account hearers when selecting a referring expression. If the use of a pronoun will lead to an unintended interpretation by the hearer, the speaker will use an unambiguous definite noun phrase instead. Because elderly adults are more limited in their processing capacities, as is indicated by their smaller working memory capacity, as speakers they will not always be able to reason about the hearer’s choices. As a result, they frequently produce non-recoverable pronouns.

Processing the noun phrase versus sentence coordination ambiguity: Thematic information does not completely eliminate processing difficulty

When faced with the noun phrase (NP) versus sentence (S) coordination ambiguity as in, for example, The thief shot the jeweller and the cop …, readers prefer the reading with NP-coordination (e.g., “The thief shot the jeweller and the cop yesterday”) over one with two conjoined sentences (e.g., “The thief shot the jeweller and the cop panicked”). A corpus study is presented showing that NP-coordinations are produced far more often than S-coordinations, which in frequency-based accounts of parsing might be taken to explain the NP-coordination preference. In addition, we describe an eye-tracking experiment investigating S-coordinated sentences such as Jasper sanded the board and the carpenter laughed, where the poor thematic fit between carpenter and sanded argues against NP-coordination. Our results indicate that information regarding poor thematic fit was used rapidly, but not without leaving some residual processing difficulty. This is compatible with claims that thematic information can reduce but not completely eliminate garden-path effects.

Who Is He? Children with ASD and ADHD Take the Listener into Account in Their Production of Ambiguous Pronouns

During conversation, speakers constantly make choices about how specific they wish to be in their use of referring expressions. In the present study we investigate whether speakers take the listener into account or whether they base their referential choices solely on their own representation of the discourse. We do this by examining the cognitive mechanisms that underlie the choice of referring expression at different discourse moments. Furthermore, we provide insights into how children with Autism Spectrum Disorder (ASD) and Attention Deficit Hyperactivity Disorder (ADHD) use referring expressions and whether their use differs from that of typically developing (TD) children. Children between 6 and 12 years old (ASD: n=46; ADHD: n=37; TD: n=38) were tested on their production of referring expressions and on Theory of Mind, response inhibition and working memory. We found support for the view that speakers take the listener into account when choosing a referring expression: Theory of Mind was related to referential choice only at those moments when speakers could not solely base their choice on their own discourse representation to be understood. Working memory appeared to be involved in keeping track of the different referents in the discourse. Furthermore, we found that TD children as well as children with ASD and children with ADHD took the listener into account in their choice of referring expression. In addition, children with ADHD were less specific than TD children in contexts with more than one referent. The previously observed problems with referential choice in children with ASD may lie in difficulties in keeping track of longer and more complex discourses, rather than in problems with taking into account the listener.

Asymmetries between Language Production and Comprehension

This book asserts that language is a signaling system rather than a code, based in part on such research as the finding that 5-year-old English and Dutch children use pronouns correctly in their own utterances, but often fail to interpret these forms correctly when used by someone else. Emphasizing the unique and sometimes competing demands of listener and speaker, the author examines resulting asymmetries between production and comprehension. The text offers examples of the interpretation of word order and pronouns by listeners, and word order freezing and referential choice by speakers. It is explored why the usual symmetry breaks down in children but also sometimes in adults. Gathering contemporary insights from theoretical linguistic research, psycholinguistic studies and computational modeling, Asymmetries between Language Production and Comprehension presents a unified explanation of this phenomenon.

Why Don't You See What I Mean? Prospects and Limitations of Current Automatic Sign Recognition Research

This article presents an overview of current automatic sign recognition research. A review of recent studies, as well as on our own research, has identified several problem areas that hamper successful sign recognition by a computer. Some of these problems are shared with automatic speech recognition, whereas others seem to be unique to automatic sign recognition. These latter difficulties include context dependency, determination of the basic units of modeling, the ability to distinguish between signs and gestures, movement epenthesis, and repetition within signs. As a possible solution to these problems, bottom-up processing should perhaps be supplemented with top-down processing.