The topic of this dissertation is the nature of linguistic competence, the capacity to understand and produce sentences of natural language. I defend the empiricist account of linguistic competence embedded in the connectionist cognitive science. This strand of cognitive science has been opposed to the traditional symbolic cognitive science, coupled with transformational-generative grammar, which was committed to nativism due to the view that human cognition, including language capacity, should be construed in terms of symbolic representations and hardwired rules. Similarly, linguistic (...) competence in this framework was regarded as being innate, rule-governed, domain-specific, and fundamentally different from performance, i.e., idiosyncrasies and factors governing linguistic behavior. I analyze state-of-the-art connectionist, deep learning models of natural language processing, most notably large language models, to see what they can tell us about linguistic competence. Deep learning is a statistical technique for the classification of patterns through which artificial intelligence researchers train artificial neural networks containing multiple layers that crunch a gargantuan amount of textual and/or visual data. I argue that these models suggest that linguistic competence should be construed as stochastic, pattern-based, and stemming from domain-general mechanisms. Moreover, I distinguish syntactic from semantic competence, and I show for each the ramifications of the endorsement of a connectionist research program as opposed to the traditional symbolic cognitive science and transformational-generative grammar. I provide a unifying front, consisting of usage-based theories, a construction grammar approach, and an embodied approach to cognition to show that the more multimodal and diverse models are in terms of architectural features and training data, the stronger the case is for the connectionist linguistic competence. I also propose to discard the competence vs. performance distinction as theoretically inferior so that a novel and integrative account of linguistic competence originating in connectionism and empiricism that I propose and defend in the dissertation could be put forward in scientific and philosophical literature. (shrink)

Our understanding of dual-process models of cognition may benefit from a consideration of language processing, as language comprehension involves fast and slow processes analogous to those used for reasoning. More specifically, De Neys's criticisms of the exclusivity assumption and the fast-to-slow switch mechanism are consistent with findings from the literature on the construction and revision of linguistic interpretations.

Subject relative clauses (SRCs) are typically processed more easily than object relative clauses (ORCs), but this difference is diminished by an inanimate head‐noun in semantically non‐reversible ORCs (“The book that the boy is reading”). In two eye‐tracking experiments, we investigated the influence of animacy on online processing of semantically reversible SRCs and ORCs using lexically inanimate items that were perceptually animate due to motion (e.g., “Where is the tractor that the cow is chasing”). In Experiment 1, 48 children (aged 4;5–6;4) (...) and 32 adults listened to sentences that varied in the lexical animacy of the NP1 head‐noun (Animate/Inanimate) and relative clause (RC) type (SRC/ORC) with an animate NP2 while viewing two images depicting opposite actions. As expected, inanimate head‐nouns facilitated the correct interpretation of ORCs in children; however, online data revealed children were more likely to anticipate an SRC as the RC unfolded when an inanimate head‐noun was used, suggesting processing was sensitive to perceptual animacy. In Experiment 2, we repeated our design with inanimate (rather than animate) NP2s (e.g., “where is the tractor that the car is following”) to investigate whether our online findings were due to increased visual surprisal at an inanimate as agent, or to similarity‐based interference. We again found greater anticipation for an SRC in the inanimate condition, supporting our surprisal hypothesis. Across the experiments, offline measures show that lexical animacy influenced children's interpretation of ORCs, whereas online measures reveal that as RCs unfolded, children were sensitive to the perceptual animacy of lexically inanimate NPs, which was not reflected in the offline data. Overall measures of syntactic comprehension, inhibitory control, and verbal short‐term memory and working memory were not predictive of children's accuracy in RC interpretation, with the exception of a positive correlation with a standardized measure of syntactic comprehension in Experiment 1. (shrink)

We investigate the order in which speakers produce the proper names of couples they know personally in English and Japanese, two languages with markedly different constituent word orders. Results demonstrate that speakers of both languages tend to produce the name of the person they feel closer to before the name of the other member of the couple (N = 180). In this way, speakers’ unique personal histories give rise to a remarkably systematic linguistic generalization in both English and Japanese. Insofar (...) as closeness serves as an index of cognitive accessibility, the current work demonstrates that systematicity emerges from a domain-general property of memory. (shrink)

Stereotypes shape inferences in philosophical thought, political discourse, and everyday life. These inferences are routinely made when thinkers engage in language comprehension or production: We make them whenever we hear, read, or formulate stories, reports, philosophical case-descriptions, or premises of arguments – on virtually any topic. These inferences are largely automatic: largely unconscious, non-intentional, and effortless. Accordingly, they shape our thought in ways we can properly understand only by complementing traditional forms of philosophical analysis with experimental methods from psycholinguistics. This (...) paper seeks, first, to bring out the wider philosophical relevance of stereotypical inference, well beyond familiar topics like gender and race. Second, we wish to provide philosophers with a toolkit to experimentally study these ubiquitous inferences and what intuitions they may generate. This paper explains what stereotypes are, and why they matter to current and traditional concerns in philosophy – experimental, analytic, and applied. It then assembles a psycholinguistic toolkit and demonstrates through two studies how potentially questionnaire-based measures can be combined with process measures to garner evidence for specific stereotypical inferences and study when they ‘go through’ and influence our thinking. (shrink)

We present evidence that preschool children oftentimes understand disjunctive sentences as if they were conjunctive. The result holds for matrix disjunctions as well as disjunctions embedded under every. At the same time, there is evidence in the literature that children understand or as inclusive disjunction in downward-entailing contexts. We propose to explain this seemingly conflicting pattern of results by assuming that the child knows the inclusive disjunction semantics of or, and that the conjunctive inference is a scalar implicature. We make (...) two assumptions about implicature computation in the child: that children access only a proper subset of the adult alternatives, and that children possess the adult capacity to strengthen sentences with implicatures. As a consequence, children are expected to sometimes not compute any implicatures at all, but in other cases they are expected to compute an implicature that is different from the adult implicature. We argue that the child’s conjunctive strengthening of disjunctive sentences realizes the latter possibility: the adult infers that the conjunction is false but the child infers that the conjunction is true. This behaviour is predicted when our assumptions about child development are coupled with the assumption that a covert exhaustive operator is responsible for strengthening in both the child and the adult. Specifically, children’s conjunctive strengthening is predicted to follow from the same mechanism used by adults to compute conjunctive free choice implicatures in response to disjunctive permission sentences. We furthermore argue that this parallel between the child and the adult extends to disambiguation preferences. In particular, we present evidence that children prefer to strengthen disjunctions to conjunctions, in matrix and embedded positions ; this result mirrors previous findings that adults prefer to compute free choice, at the root and under every. We propose a disambiguation strategy that explains the preference for conjunctive strengthening – by both the child and the adult – even though there is no general preference for exhaustification. Specifically, we propose that the preference for a conjunctive strengthening follows from a pragmatic preference for a complete answer to the Question Under Discussion. (shrink)

There is now considerable evidence that human sentence processing is expectation based: As people read a sentence, they use their statistical experience with their language to generate predictions about upcoming syntactic structure. This study examines how sentence processing is affected by readers' uncertainty about those expectations. In a self-paced reading study, we use lexical subcategorization distributions to factorially manipulate both the strength of expectations and the uncertainty about them. We compare two types of uncertainty: uncertainty about the verb's complement, reflecting (...) the next prediction step; and uncertainty about the full sentence, reflecting an unbounded number of prediction steps. We find that uncertainty about the full structure, but not about the next step, was a significant predictor of processing difficulty: Greater reduction in uncertainty was correlated with increased reading times. We additionally replicated previously observed effects of expectation violation, orthogonal to the effect of uncertainty. This suggests that both surprisal and uncertainty affect human RTs. We discuss the consequences for theories of sentence comprehension. (shrink)

A critical issue in perception is the manner in which top‐down expectancies guide lower level perceptual processes. In speech, a common paradigm is to construct continua ranging between two phonetic endpoints and to determine how higher level lexical context influences the perceived boundary. We applied this approach to music, presenting participants with major/minor triad continua after brief musical contexts. Two experiments yielded results that differed from classic results in speech perception. In speech, context generally expands the category of the expected (...) stimuli. We found the opposite in music: The major/minor boundary shifted toward the expected category, contracting it. Together, these experiments support the hypothesis that musical expectancy can feed back to affect lower‐level perceptual processes. However, it may do so in a way that differs fundamentally from what has been seen in other domains. (shrink)

Language is a collaborative act: To communicate successfully, speakers must generate utterances that are not only semantically valid but also sensitive to the knowledge state of the listener. Such sensitivity could reflect the use of an “embedded listener model,” where speakers choose utterances on the basis of an internal model of the listener's conceptual and linguistic knowledge. In this study, we ask whether parents’ spatial descriptions incorporate an embedded listener model that reflects their children's understanding of spatial relations and spatial (...) terms. Adults described the positions of targets in spatial arrays to their children or to the adult experimenter. Arrays were designed so that targets could not be identified unless spatial relationships within the array were encoded and described. Parents of 3–4-year-old children encoded relationships in ways that were well-matched to their children's level of spatial language. These encodings differed from those of the same relationships in speech to the adult experimenter. In contrast, parents of individuals with severe spatial impairments did not show clear evidence of sensitivity to their children's level of spatial language. The results provide evidence for an embedded listener model in the domain of spatial language and indicate conditions under which the ability to model listener knowledge may be more challenging. (shrink)

One of the most basic functions of human language is to convey who did what to whom. In the world's languages, the order of these three constituents (subject [S], verb [V], and object [O]) is uneven, with SOV and SVO being most common. Recent experiments using experimentally elicited pantomime provide a possible explanation of the prevalence of SOV, but extant explanations for the prevalence of SVO could benefit from further empirical support. Here, we test whether SVO might emerge because (a) (...) SOV is not well suited for describing reversible events (a woman pushing a boy) and (b) pressures to be efficient and mention subjects before objects conspire to rule out many other alternatives. We tested this by asking participants to describe reversible and non‐reversible events in pantomime, and we instructed some participants to be consistent in the form of their gestures and to teach them to the experimenter. These manipulations led to the emergence of SVO in speakers of both English (SVO) and Turkish (SOV). (shrink)

Theories of natural language syntax often characterize grammatical knowledge as a form of abstract computation. This paper argues that such a characterization is correct, and that fundamental properties of grammar can and should be understood in terms of restrictions on the complexity of possible grammatical computation, when defined in terms of generative capacity. More specifically, the paper demonstrates that the computational restrictiveness imposed by Tree Adjoining Grammar provides important insights into the nature of human grammatical knowledge.

In this article we use proportional hazards models to examine how low-level processes affect the probability of making a saccade over time, through the period of fixation, during reading. We apply the Cox proportional hazards model to investigate how launch distance (relative to word beginning), fixation location (relative to word center), and word frequency affect the hazard of a saccadic response. This model requires that covariates have a constant impact on the hazard over time, the assumption of proportional hazards. We (...) show that this assumption is not supported. The impact of the covariates changes with the time passed since fixation onset. To account for the non-proportional hazards we fit step functions of time, resulting in a model with time-varying effects on the hazard. We evaluate the ability to predict the timing of saccades on held-out fixation data. The model with time-varying effects performs better in predicting the timing of saccades for fixations as short as 100 ms and as long as 500 ms, when compared both to a baseline model without covariates and a model which assumes constant covariate effects. This result suggests that the time-varying effects model better recovers the time course of low-level processes that influence the decision to move the eyes. (shrink)

In this paper, I argue for a modified version of what Devitt calls the Representational Thesis. According to RT, syntactic rules or principles are psychologically real, in the sense that they are represented in the mind/brain of every linguistically competent speaker/hearer. I present a range of behavioral and neurophysiological evidence for the claim that the human sentence processing mechanism constructs mental representations of the syntactic properties of linguistic stimuli. I then survey a range of psychologically plausible computational models of comprehension (...) and show that they are all committed to RT. I go on to sketch a framework for thinking about the nature of the representations involved in sentence processing. My claim is that these are best characterized not as propositional attitudes but, rather, as subpersonal states whose representational properties are determined by their functional role. Finally, I distinguish between explicit and implicit representations and argue that the latter can be drawn on as data by the algorithms that constitute our sentence processing routines. I conclude that skepticism concerning the psychological reality of grammars cannot be sustained. (shrink)

Memory limitations and probabilistic expectations are two key factors that have been posited to play a role in the incremental processing of natural language. Relative clauses (RCs) have long served as a key proving ground for such theories of language processing. Across three self-paced reading experiments, we test the online comprehension of Hungarian subject- and object-extracted RCs (SRCs and ORCs, respectively). We capitalize on the syntactic properties of Hungarian that allow for a variety of word orders within RCs, which helps (...) us to delineate the processing costs associated with memory demand and violated expectations. Results showed a processing cost at the RC verb for structures that have longer verb-argument distances, despite those structures being more frequent in the corpus. These findings thus support theories that attribute processing difficulty to memory limitations, rather than theories that attribute difficulty to less expected structures. (shrink)

Much previous work has suggested that word order preferences across languages can be explained by the dependency distance minimization constraint (Ferrer‐i Cancho, 2008, 2015; Hawkins, 1994). Consistent with this claim, corpus studies have shown that the average distance between a head (e.g., verb) and its dependent (e.g., noun) tends to be short cross‐linguistically (Ferrer‐i Cancho, 2014; Futrell, Mahowald, & Gibson, 2015; Liu, Xu, & Liang, 2017). This implies that on average languages avoid inefficient or complex structures for simpler structures. But (...) a number of studies in psycholinguistics (Konieczny, 2000; Levy & Keller, 2013; Vasishth, Suckow, Lewis, & Kern, 2010) show that the comprehension system can adapt to the typological properties of a language, for example, verb‐final order, leading to more complex structures, for example, having longer linear distance between a head and its dependent. In this paper, we conduct a corpus study for a group of 38 languages, which were either Subject–Verb–Object (SVO) or Subject–Object–Verb (SOV), in order to investigate the role of word order typology in determining syntactic complexity. We present results aggregated across all dependency types, as well as for specific verbal (objects, indirect objects, and adjuncts) and nonverbal (nominal, adjectival, and adverbial) dependencies. The results suggest that dependency distance in a language is determined by the default word order of a language, and crucially, the direction of a dependency (whether the head precedes the dependent or follows it; e.g., whether the noun precedes the verb or follows it). Particularly we show that in SOV languages (e.g., Hindi, Korean) as well as SVO languages (e.g., English, Spanish), longer linear distance (measured as number of words) between head and dependent arises in structures when they mirror the default word order of the language. In addition to showing results on linear distance, we also investigate the influence of word order typology on hierarchical distance (HD; measured as number of heads between head and dependent). The results for HD are similar to that of linear distance. At the same time, in comparison to linear distance, the influence of adaptability on HD seems less strong. In particular, the results show that most languages tend to avoid greater structural depth. Together, these results show evidence for “limited adaptability” to the default word order preferences in a language. Our results support a large body of work in the processing literature that highlights the importance of linguistic exposure and its interaction with working memory constraints in determining sentence complexity. Our results also point to the possible role of other factors such as the morphological richness of a language and a multifactor account of sentence complexity remains a promising area for future investigation. (shrink)

Clark highlights a neglected issue in research on language use: the process by which speakers and addressees anchor utterances with respect to individual entities in their common ground. In his review, he identifies the challenges linked to investigations of anchoring, but also displays the pitfalls of evading it.

The empirical study of language is a young field in contemporary linguistics. This being the case, and following a natural development process, the field is currently at a stage where different research methods and experimental approaches are being put into question in terms of their validity. Without pretending to provide an answer with respect to the best way to conduct linguistics related experimental research, in this article we aim at examining the process that researchers follow in the design and implementation (...) of experimental linguistics research with a goal to validate specific theoretical linguistic analyses. First, we discuss the general challenges that experimental work faces in finding a compromise between addressing theoretically relevant questions and being able to implement these questions in a specific controlled experimental paradigm. We discuss the Granularity Mismatch Problem (Poeppel and Embick, 2005 ) which addresses the challenges that research that is trying to bridge the representations and computations of language and their psycholinguistic/neurolinguistic evidence faces, and the basic assumptions that interdisciplinary research needs to consider due to the different conceptual granularity of the objects under study. To illustrate the practical implications of the points addressed, we compare two approaches to perform linguistic experimental research by reviewing a number of our own studies strongly grounded on theoretically informed questions. First, we show how linguistic phenomena similar at a conceptual level can be tested within the same language using measurement of event-related potentials (ERP) by discussing results from two ERP experiments on the processing of long-distance backward dependencies that involve coreference and negative polarity items respectively in Dutch. Second, we examine how the same linguistic phenomenon can be tested in different languages using reading time measures by discussing the outcome of four self-paced reading experiments on the processing of in-situ wh -questions in Mandarin Chinese and French. Finally, we review the implications that our findings have for the specific theoretical linguistics questions that we originally aimed to address. We conclude with an overview of the general insights that can be gained from the role of structural hierarchy and grammatical constraints in processing and the existing limitations on the generalization of results. (shrink)

In Jackendoff's Parallel Architecture, the well-formed expressions of a language are licensed by correspondences between phonology, syntax, and conceptual structure. I show how this architecture can be used to make sense of the existence of parasitic gap constructions. A parasitic gap is one that is rendered acceptable because of the presence of another gap in the same sentence. Compare *a person whoi everyone who talks to ti likes Chris, which shows an illicit extraction from a relative clause, and a person (...) whoi everyone who talks to pgi likes ti, which shows a parasitic gap in the relative clause. Languages differ in terms of the range of configurations in which they allow parasitic gaps; these configurations appear to form a hierarchy. These observations raise some fundamental questions: Why do parasitic gaps exist at all? Why are different syntactic configurations possible for P-gaps and why just these configurations? Why is there a parasitic gap hierarchy? The answers that I propose make crucial use of constructional overgeneralization of across-the-board extraction in coordinate constructions, formulated straightforwardly in the framework of Jackendoff's Parallel Architecture. (shrink)

In the generative tradition, the language faculty has been shrinking—perhaps to include only the mechanism of recursion. This paper argues that even this view of the language faculty is too expansive. We first argue that a language faculty is difficult to reconcile with evolutionary considerations. We then focus on recursion as a detailed case study, arguing that our ability to process recursive structure does not rely on recursion as a property of the grammar, but instead emerges gradually by piggybacking on (...) domain-general sequence learning abilities. Evidence from genetics, comparative work on non-human primates, and cognitive neuroscience suggests that humans have evolved complex sequence learning skills, which were subsequently pressed into service to accommodate language. Constraints on sequence learning therefore have played an important role in shaping the cultural evolution of linguistic structure, including our limited abilities for processing recursive structure. Finally, we re-evaluate some of the key considerations that have often been taken to require the postulation of a language faculty. (shrink)