Refine
Year of publication
Document Type
- Article (78)
- Postprint (16)
- Review (3)
- Monograph/Edited Volume (2)
- Part of Periodical (2)
- Other (1)
Keywords
- German (10)
- eye-tracking (9)
- interference (8)
- locality (8)
- Eye movements (7)
- Bayesian data analysis (6)
- Reading (6)
- individual differences (6)
- self-paced reading (6)
- sentence processing (6)
Institute
In eye-movement control during reading, advanced process-oriented models have been developed to reproduce behavioral data. So far, model complexity and large numbers of model parameters prevented rigorous statistical inference and modeling of interindividual differences. Here we propose a Bayesian approach to both problems for one representative computational model of sentence reading (SWIFT; Engbert et al., Psychological Review, 112, 2005, pp. 777-813). We used experimental data from 36 subjects who read the text in a normal and one of four manipulated text layouts (e.g., mirrored and scrambled letters). The SWIFT model was fitted to subjects and experimental conditions individually to investigate between- subject variability. Based on posterior distributions of model parameters, fixation probabilities and durations are reliably recovered from simulated data and reproduced for withheld empirical data, at both the experimental condition and subject levels. A subsequent statistical analysis of model parameters across reading conditions generates model-driven explanations for observable effects between conditions.
Individuals with agrammatic Broca's aphasia experience difficulty when processing reversible non-canonical sentences. Different accounts have been proposed to explain this phenomenon. The Trace Deletion account (Grodzinsky, 1995, 2000, 2006) attributes this deficit to an impairment in syntactic representations, whereas others (e.g., Caplan, Waters, Dede, Michaud, & Reddy, 2007; Haarmann, Just, & Carpenter, 1997) propose that the underlying structural representations are unimpaired, but sentence comprehension is affected by processing deficits, such as slow lexical activation, reduction in memory resources, slowed processing and/or intermittent deficiency, among others. We test the claims of two processing accounts, slowed processing and intermittent deficiency, and two versions of the Trace Deletion Hypothesis (TDH), in a computational framework for sentence processing (Lewis & Vasishth, 2005) implemented in ACT-R (Anderson, Byrne, Douglass, Lebiere, & Qin, 2004). The assumption of slowed processing is operationalized as slow procedural memory, so that each processing action is performed slower than normal, and intermittent deficiency as extra noise in the procedural memory, so that the parsing steps are more noisy than normal. We operationalize the TDH as an absence of trace information in the parse tree. To test the predictions of the models implementing these theories, we use the data from a German sentence—picture matching study reported in Hanne, Sekerina, Vasishth, Burchert, and De Bleser (2011). The data consist of offline (sentence-picture matching accuracies and response times) and online (eye fixation proportions) measures. From among the models considered, the model assuming that both slowed processing and intermittent deficiency are present emerges as the best model of sentence processing difficulty in aphasia. The modeling of individual differences suggests that, if we assume that patients have both slowed processing and intermittent deficiency, they have them in differing degrees.
We present a computational evaluation of three hypotheses about sources of deficit in sentence comprehension in aphasia: slowed processing, intermittent deficiency, and resource reduction. The ACT-R based Lewis and Vasishth (2005) model is used to implement these three proposals. Slowed processing is implemented as slowed execution time of parse steps; intermittent deficiency as increased random noise in activation of elements in memory; and resource reduction as reduced spreading activation. As data, we considered subject vs. object relative sentences, presented in a self-paced listening modality to 56 individuals with aphasia (IWA) and 46 matched controls. The participants heard the sentences and carried out a picture verification task to decide on an interpretation of the sentence. These response accuracies are used to identify the best parameters (for each participant) that correspond to the three hypotheses mentioned above. We show that controls have more tightly clustered (less variable) parameter values than IWA; specifically, compared to controls, among IWA there are more individuals with slow parsing times, high noise, and low spreading activation. We find that (a) individual IWA show differential amounts of deficit along the three dimensions of slowed processing, intermittent deficiency, and resource reduction, (b) overall, there is evidence for all three sources of deficit playing a role, and (c) IWA have a more variable range of parameter values than controls. An important implication is that it may be meaningless to talk about sources of deficit with respect to an abstract verage IWA; the focus should be on the individual's differential degrees of deficit along different dimensions, and on understanding the causes of variability in deficit between participants.
We explore the interaction between oculomotor control and language comprehension on the sentence level using two well-tested computational accounts of parsing difficulty. Previous work (Boston, Hale, Vasishth, & Kliegl, 2011) has shown that surprisal (Hale, 2001; Levy, 2008) and cue-based memory retrieval (Lewis & Vasishth, 2005) are significant and complementary predictors of reading time in an eyetracking corpus. It remains an open question how the sentence processor interacts with oculomotor control. Using a simple linking hypothesis proposed in Reichle, Warren, and McConnell (2009), we integrated both measures with the eye movement model EMMA (Salvucci, 2001) inside the cognitive architecture ACT-R (Anderson et al., 2004). We built a reading model that could initiate short Time Out regressions (Mitchell, Shen, Green, & Hodgson, 2008) that compensate for slow postlexical processing. This simple interaction enabled the model to predict the re-reading of words based on parsing difficulty. The model was evaluated in different configurations on the prediction of frequency effects on the Potsdam Sentence Corpus. The extension of EMMA with postlexical processing improved its predictions and reproduced re-reading rates and durations with a reasonable fit to the data. This demonstration, based on simple and independently motivated assumptions, serves as a foundational step toward a precise investigation of the interaction between high-level language processing and eye movement control.
Traxler, Pickering, and Clifton (1998) found that ambiguous sentences are read faster than their unambiguous counterparts. This so-called ambiguity advantage has presented a major challenge to classical theories of human sentence comprehension (parsing) because its most prominent explanation, in the form of the unrestricted race model (URM), assumes that parsing is non-deterministic. Recently, Swets, Desmet, Clifton, and Ferreira (2008) have challenged the URM. They argue that readers strategically underspecify the representation of ambiguous sentences to save time, unless disambiguation is required by task demands. When disambiguation is required, however, readers assign sentences full structure—and Swets et al. provide experimental evidence to this end. On the basis of their findings, they argue against the URM and in favor of a model of task-dependent sentence comprehension. We show through simulations that the Swets et al. data do not constitute evidence for task-dependent parsing because they can be explained by the URM. However, we provide decisive evidence from a German self-paced reading study consistent with Swets et al.'s general claim about task-dependent parsing. Specifically, we show that under certain conditions, ambiguous sentences can be read more slowly than their unambiguous counterparts, suggesting that the parser may create several parses, when required. Finally, we present the first quantitative model of task-driven disambiguation that subsumes the URM, and we show that it can explain both Swets et al.'s results and our findings.
Among theories of human language comprehension, cue-based memory retrieval has proven to be a useful framework for understanding when and how processing difficulty arises in the resolution of long-distance dependencies. Most previous work in this area has assumed that very general retrieval cues like [+subject] or [+singular] do the work of identifying (and sometimes misidentifying) a retrieval target in order to establish a dependency between words. However, recent work suggests that general, handpicked retrieval cues like these may not be enough to explain illusions of plausibility (Cunnings & Sturt, 2018), which can arise in sentences like The letter next to the porcelain plate shattered. Capturing such retrieval interference effects requires lexically specific features and retrieval cues, but handpicking the features is hard to do in a principled way and greatly increases modeler degrees of freedom. To remedy this, we use well-established word embedding methods for creating distributed lexical feature representations that encode information relevant for retrieval using distributed retrieval cue vectors. We show that the similarity between the feature and cue vectors (a measure of plausibility) predicts total reading times in Cunnings and Sturt's eye-tracking data. The features can easily be plugged into existing parsing models (including cue-based retrieval and self-organized parsing), putting very different models on more equal footing and facilitating future quantitative comparisons.
We present a detailed process theory of the moment-by-moment working-memory retrievals and associated control structure that subserve sentence comprehension. The theory is derived from the application of independently motivated principles of memory and cognitive skill to the specialized task of sentence parsing. The resulting theory construes sentence processing as a series of skilled associative memory retrievals modulated by similarity-based interference and fluctuating activation. The cognitive principles are formalized in computational form in the Adaptive Control of Thought- Rational (ACT-R) architecture, and our process model is realized in ACT-R. We present the results of 6 sets of simulations: 5 simulation sets provide quantitative accounts of the effects of length and structural interference on both unambiguous and garden-path structures. A final simulation set provides a graded taxonomy of double center embeddings ranging from relatively easy to extremely difficult. The explanation of center-embedding difficulty is a novel one that derives from the model's complete reliance on discriminating retrieval cues in the absence of an explicit representation of serial order information. All fits were obtained with only 1 free scaling parameter fixed across the simulations; all other parameters were ACT-R defaults. The modeling results support the hypothesis that fluctuating activation and similarity-based interference are the key factors shaping working memory in sentence processing. We contrast the theory and empirical predictions with several related accounts of sentence-processing complexity
Several studies (e.g., Wicha et al., 2003b; DeLong et al., 2005) have shown that readers use information from the sentential context to predict nouns (or some of their features), and that predictability effects can be inferred from the EEG signal in determiners or adjectives appearing before the predicted noun. While these findings provide evidence for the pre-activation proposal, recent replication attempts together with inconsistencies in the results from the literature cast doubt on the robustness of this phenomenon. Our study presents the first attempt to use the effect of gender on predictability in German to study the pre-activation hypothesis, capitalizing on the fact that all German nouns have a gender and that their preceding determiners can show an unambiguous gender marking when the noun phrase has accusative case. Despite having a relatively large sample size (of 120 subjects), both our preregistered and exploratory analyses failed to yield conclusive evidence for or against an effect of pre-activation. The sign of the effect is, however, in the expected direction: the more unexpected the gender of the determiner, the larger the negativity. The recent, inconclusive replication attempts by Nieuwland et al. (2018) and others also show effects with signs in the expected direction. We conducted a Bayesian random-ef-fects meta-analysis using our data and the publicly available data from these recent replication attempts. Our meta-analysis shows a relatively clear but very small effect that is consistent with the pre-activation account and demonstrates a very important advantage of the Bayesian data analysis methodology: we can incrementally accumulate evidence to obtain increasingly precise estimates of the effect of interest.
Argument-head distance and processing complexity: Explaining both locality and antilocality effects
(2006)
Although proximity between arguments and verbs (locality) is a relatively robust determinant of sentence-processing difficulty (Hawkins 1998, 2001, Gibson 2000), increasing argument-verb distance can also facilitate processing (Konieczny 2000). We present two self-paced reading (SPR) experiments involving Hindi that provide further evidence of antilocality, and a third SPR experiment which suggests that similarity-based interference can attenuate this distance-based facilitation. A unified explanation of interference, locality, and antilocality effects is proposed via an independently motivated theory of activation decay and retrieval interference (Anderson et al. 2004).*
Linear mixed-effects models have increasingly replaced mixed-model analyses of variance for statistical inference in factorial psycholinguistic experiments. Although LMMs have many advantages over ANOVA, like ANOVAs, setting them up for data analysis also requires some care. One simple option, when numerically possible, is to fit the full variance covariance structure of random effects (the maximal model; Barr, Levy, Scheepers & Tily, 2013), presumably to keep Type I error down to the nominal a in the presence of random effects. Although it is true that fitting a model with only random intercepts may lead to higher Type I error, fitting a maximal model also has a cost: it can lead to a significant loss of power. We demonstrate this with simulations and suggest that for typical psychological and psycholinguistic data, higher power is achieved without inflating Type I error rate if a model selection criterion is used to select a random effect structure that is supported by the data. (C) 2017 The Authors. Published by Elsevier Inc.
This tutorial analyzes voice onset time (VOT) data from Dongbei (Northeastern) Mandarin Chinese and North American English to demonstrate how Bayesian linear mixed models can be fit using the programming language Stan via the R package brms. Through this case study, we demonstrate some of the advantages of the Bayesian framework: researchers can (i) flexibly define the underlying process that they believe to have generated the data; (ii) obtain direct information regarding the uncertainty about the parameter that relates the data to the theoretical question being studied; and (iii) incorporate prior knowledge into the analysis. Getting started with Bayesian modeling can be challenging, especially when one is trying to model one’s own (often unique) data. It is difficult to see how one can apply general principles described in textbooks to one’s own specific research problem. We address this barrier to using Bayesian methods by providing three detailed examples, with source code to allow easy reproducibility. The examples presented are intended to give the reader a flavor of the process of model-fitting; suggestions for further study are also provided. All data and code are available from: https://osf.io/g4zpv.
With the arrival of the R packages nlme and lme4, linear mixed models (LMMs) have come to be widely used in experimentally-driven areas like psychology, linguistics, and cognitive science. This tutorial provides a practical introduction to fitting LMMs in a Bayesian framework using the probabilistic programming language Stan. We choose Stan (rather than WinBUGS or JAGS) because it provides an elegant and scalable framework for fitting models in most of the standard applications of LMMs. We ease the reader into fitting increasingly complex LMMs, using a two-condition repeated measures self-paced reading study.
Recent research has shown that brain potentials time-locked to fixations in natural reading can be similar to brain potentials recorded during rapid serial visual presentation (RSVP). We attempted two replications of Hagoort, Hald, Bastiaansen, and Petersson [Hagoort, P., Hald, L., Bastiaansen, M., & Petersson, K. M. Integration of word meaning and world knowledge in language comprehension. Science, 304, 438-441, 2004] to determine whether this correspondence also holds for oscillatory brain responses. Hagoort et al. reported an N400 effect and synchronization in the theta and gamma range following world knowledge violations. Our first experiment (n = 32) used RSVP and replicated both the N400 effect in the ERPs and the power increase in the theta range in the time-frequency domain. In the second experiment (n = 49), participants read the same materials freely while their eye movements and their EEG were monitored. First fixation durations, gaze durations, and regression rates were increased, and the ERP showed an N400 effect. An analysis of time-frequency representations showed synchronization in the delta range (1-3 Hz) and desynchronization in the upper alpha range (11-13 Hz) but no theta or gamma effects. The results suggest that oscillatory EEG changes elicited by world knowledge violations are different in natural reading and RSVP. This may reflect differences in how representations are constructed and retrieved from memory in the two presentation modes.
Background: In addition to the canonical subject-verb-object (SVO) word order, German also allows for non-canonical order (OVS), and the case-marking system supports thematic role interpretation. Previous eye-tracking studies (Kamide et al., 2003; Knoeferle, 2007) have shown that unambiguous case information in non-canonical sentences is processed incrementally. For individuals with agrammatic aphasia, comprehension of non-canonical sentences is at chance level (Burchert et al., 2003). The trace deletion hypothesis (Grodzinsky 1995, 2000) claims that this is due to structural impairments in syntactic representations, which force the individual with aphasia (IWA) to apply a guessing strategy. However, recent studies investigating online sentence processing in aphasia (Caplan et al., 2007; Dickey et al., 2007) found that divergences exist in IWAs' sentence-processing routines depending on whether they comprehended non-canonical sentences correctly or not, pointing rather to a processing deficit explanation. Aims: The aim of the current study was to investigate agrammatic IWAs' online and offline sentence comprehension simultaneously in order to reveal what online sentence-processing strategies they rely on and how these differ from controls' processing routines. We further asked whether IWAs' offline chance performance for non-canonical sentences does indeed result from guessing. Methods Procedures: We used the visual-world paradigm and measured eye movements (as an index of online sentence processing) of controls (N = 8) and individuals with aphasia (N = 7) during a sentence-picture matching task. Additional offline measures were accuracy and reaction times. Outcomes Results: While the offline accuracy results corresponded to the pattern predicted by the TDH, IWAs' eye movements revealed systematic differences depending on the response accuracy. Conclusions: These findings constitute evidence against attributing IWAs' chance performance for non-canonical structures to mere guessing. Instead, our results support processing deficit explanations and characterise the agrammatic parser as deterministic and inefficient: it is slowed down, affected by intermittent deficiencies in performing syntactic operations, and fails to compute reanalysis even when one is detected.
Sentence comprehension requires that the comprehender work out who did what to whom. This process has been characterized as retrieval from memory. This review summarizes the quantitative predictions and empirical coverage of the two existing computational models of retrieval and shows how the predictive performance of these two competing models can be tested against a benchmark data-set. We also show how computational modeling can help us better understand sources of variability in both unimpaired and impaired sentence comprehension.
An English double-embedded relative clause from which the middle verb is omitted can often be processed more easily than its grammatical counterpart, a phenomenon known as the grammaticality illusion. This effect has been found to be reversed in German, suggesting that the illusion is language specific rather than a consequence of universal working memory constraints. We present results from three self-paced reading experiments which show that Dutch native speakers also do not show the grammaticality illusion in Dutch, whereas both German and Dutch native speakers do show the illusion when reading English sentences. These findings provide evidence against working memory constraints as an explanation for the observed effect in English. We propose an alternative account based on the statistical patterns of the languages involved. In support of this alternative, a single recurrent neural network model that is trained on both Dutch and English sentences is shown to predict the cross-linguistic difference in the grammaticality effect.
Dynamical models make specific assumptions about cognitive processes that generate human behavior. In data assimilation, these models are tested against timeordered data. Recent progress on Bayesian data assimilation demonstrates that this approach combines the strengths of statistical modeling of individual differences with the those of dynamical cognitive models.
Dependency Resolution Difficulty Increases with Distance in Persian Separable Complex Predicates
(2016)
Delaying the appearance of a verb in a noun-verb dependency tends to increase processing difficulty at the verb; one explanation for this locality effect is decay and/or interference of the noun in working memory. Surprisal, an expectation-based account, predicts that delaying the appearance of a verb either renders it no more predictable or more predictable, leading respectively to a prediction of no effect of distance or a facilitation. Recently, Husain et al. (2014) suggested that when the exact identity of the upcoming verb is predictable (strong predictability), increasing argument-verb distance leads to facilitation effects, which is consistent with surprisal; but when the exact identity of the upcoming verb is not predictable (weak predictability), locality effects are seen. We investigated Husain et al.'s proposal using Persian complex predicates (CPs), which consist of a non-verbal element—a noun in the current study—and a verb. In CPs, once the noun has been read, the exact identity of the verb is highly predictable (strong predictability); this was confirmed using a sentence completion study. In two self-paced reading (SPR) and two eye-tracking (ET) experiments, we delayed the appearance of the verb by interposing a relative clause (Experiments 1 and 3) or a long PP (Experiments 2 and 4). We also included a simple Noun-Verb predicate configuration with the same distance manipulation; here, the exact identity of the verb was not predictable (weak predictability). Thus, the design crossed Predictability Strength and Distance. We found that, consistent with surprisal, the verb in the strong predictability conditions was read faster than in the weak predictability conditions. Furthermore, greater verb-argument distance led to slower reading times; strong predictability did not neutralize or attenuate the locality effects. As regards the effect of distance on dependency resolution difficulty, these four experiments present evidence in favor of working memory accounts of argument-verb dependency resolution, and against the surprisal-based expectation account of Levy (2008). However, another expectation-based measure, entropy, which was computed using the offline sentence completion data, predicts reading times in Experiment 1 but not in the other experiments. Because participants tend to produce more ungrammatical continuations in the long-distance condition in Experiment 1, we suggest that forgetting due to memory overload leads to greater entropy at the verb.
Dependency Resolution Difficulty Increases with Distance in Persian Separable Complex Predicates
(2016)
Delaying the appearance of a verb in a noun-verb dependency tends to increase processing difficulty at the verb; one explanation for this locality effect is decay and/or interference of the noun in working memory. Surprisal, an expectation-based account, predicts that delaying the appearance of a verb either renders it no more predictable or more predictable, leading respectively to a prediction of no effect of distance or a facilitation. Recently, Husain et al. (2014) suggested that when the exact identity of the upcoming verb is predictable (strong predictability), increasing argument-verb distance leads to facilitation effects, which is consistent with surprisal; but when the exact identity of the upcoming verb is not predictable (weak predictability), locality effects are seen. We investigated Husain et al.'s proposal using Persian complex predicates (CPs), which consist of a non-verbal element—a noun in the current study—and a verb. In CPs, once the noun has been read, the exact identity of the verb is highly predictable (strong predictability); this was confirmed using a sentence completion study. In two self-paced reading (SPR) and two eye-tracking (ET) experiments, we delayed the appearance of the verb by interposing a relative clause (Experiments 1 and 3) or a long PP (Experiments 2 and 4). We also included a simple Noun-Verb predicate configuration with the same distance manipulation; here, the exact identity of the verb was not predictable (weak predictability). Thus, the design crossed Predictability Strength and Distance. We found that, consistent with surprisal, the verb in the strong predictability conditions was read faster than in the weak predictability conditions. Furthermore, greater verb-argument distance led to slower reading times; strong predictability did not neutralize or attenuate the locality effects. As regards the effect of distance on dependency resolution difficulty, these four experiments present evidence in favor of working memory accounts of argument-verb dependency resolution, and against the surprisal-based expectation account of Levy (2008). However, another expectation-based measure, entropy, which was computed using the offline sentence completion data, predicts reading times in Experiment 1 but not in the other experiments. Because participants tend to produce more ungrammatical continuations in the long-distance condition in Experiment 1, we suggest that forgetting due to memory overload leads to greater entropy at the verb.