Search

This brief commentary addresses the question of what reading researchers and psycholinguists should make of the 'sentence superiority effect' (SSE) first reported by Snell and Grainger (2017), and explored in additional papers by the same group (Massol & Grainger, 2021; Massol, Mirault, & Grainger, 2021; Wen, Snell, & Grainger, 2019). The main conclusion is that Snell and Grainger's results from a 2AFC version of their task, reported in a footnote, tell us that the SSE is most likely due to a simple guessing strategy.

Published Jun 2022 Although research in sentence comprehension has suggested that processing long-distance dependencies involves maintenance between the elements that form the dependency, studies on maintenance of long-distance subject–verb (SV) dependencies are scarce. The few relevant studies have delivered mixed results using self-paced reading or phoneme-monitoring tasks. In the current study, we used eye tracking during reading to test whether maintaining a long-distance SV dependency results in a processing cost on an intervening adverbial clause. In Experiment 1, we studied this question in Spanish and found that both go-past reading times and regressions out of an adverbial clause to the previous regions were significantly increased when the clause interrupts a SV dependency compared to when the same clause doesn’t interrupt this dependency. We then replicated these findings in English (Experiment 2), observing significantly increased go-past reading times on a clause interrupting a SV dependency. The current study provides the first eye-tracking data showing a maintenance cost in the processing of SV dependencies cross-linguistically. Sentence comprehension models should account for the maintenance cost generated by SV dependency processing, and future research should focus on the nature of the maintained representation. This research was partially funded by Ministerio de Economía y Competitividad, Agencia Estatal de Investigación & Fondo Europeo de Desarrollo Regional Grants PSI2015-65694-P, and RTI2018-096311-B-I00 to Nicola Molinaro, and RYC-2017–22015 and FFI2016-76432-P_LAMPT to Simona Mancini; by Eusko Jaurlaritza Grants PI_2016_1_0014 to Nicola Molinaro, PRE_2018_2_0074 and EP_2018_1_0042 to Bojana Ristic; and by Agencia Estatal de Investigación’s Severo Ochoa excellence program Grant SEV2015– 0490 to the Basque Center on Cognition, Brain and Language.

We introduce the statistical concept of 'compensatory selection', which arises when selecting a subset of applicants based on multiple predictors, such as when standardized test scores are used in combination with other predictors required in a school application (e.g., previous grades, references letters, and personal statements). Post-hoc analyses often fail to find a positive correlation between test scores and subsequent success, and this failure is sometimes taken as evidence against the predictive validity of the standardized test. The present analysis reveals that the failure to find a negative correlation indicates that the standardized test is in fact a valid predictor of success. This is due to compensation between predictors during selection: Some students are admitted despite a low test score because their application is exceptional in other respects, while other students are admitted primarily based on a high test score despite weakness in the rest of their application. This compensatory selection process introduces a negative correlation between test scores and other predictors among those admitted. If test scores are valid predictors of success, this negative correlation between the predictors counteracts the positive correlation between test scores and success that would have been observed if all applicants were admitted. If test scores are not predictive of success, but were nevertheless used in a compensatory selection process, there would be a spurious negative correlation between test scores and success (i.e., an admitted student with a weak application except for a high test score would be unlikely to succeed). The selection effect that is described here is fundamentally different from the well-known 'restricted range' problem and can powerfully alter results even in situations that accept most applicants.

Previous research has demonstrated effects of both orthographic neighborhood size and neighbor frequency in word recognition in Chinese. A large neighborhood—where neighborhood size is defined by the number of words that differ from a target word by a single character—appears to facilitate word recognition, while the presence of a higher-frequency neighbor has an inhibitory effect. The present study investigated modulation of these effects by a word’s predictability in context. In two eye-movement experiments, the predictability of a target word in each sentence was manipulated. Target words differed in their neighborhood size (Experiment 1) and in whether they had a higher-frequency neighbor (Experiment 2). The study replicated the previously observed effects of neighborhood size and neighbor frequency when the target word was unpredictable, but in both experiments neighborhood effects were absent when the target was predictable. These results suggest that when a word is preactivated by context, the activation of its neighbors may be diminished to such an extent that these neighbors do not effectively compete for selection.

Distinctions related to person and animacy have long been known to impact both the grammar and incremental processing in a way that can be described through "prominence" scales. We put the generalizability of these scales to the test by examining the processing effects of a typologically uncommon distinction known as obviation, which is found in Ojibwe, an Indigenous language of North America. Obviation contrasts the single most discourse-salient animate third person (proximate) with other non-salient third persons (obviative). Using a visual world paradigm, we show that obviation influences parsing and interpretation commitments under incremental ambiguity: Proximate nouns are assumed to be the agent of an action, while obviative nouns do not lead to strong incremental commitments. This result parallels previous findings in other languages with distinctions related to animacy and person, supporting a theory where the effect of prominence information in processing is the result of a common set of constraints derived from the alignment of scales related to person, syntactic position, and thematic role.

Agreement attraction (e.g., ∗The key to the cabinets are rusty) is not attributable to the linear proximity between the local noun and verb (Franck, Vigliocco, & Nicol, 2002). However, agreement with a disjoined subject (e.g., The horses or the clock is red) is specifically sensitive to the number of the nearer noun (Haskell & MacDonald, 2005). The present study highlights other differences between the influence on agreement of a local noun in the classic attraction configuration and the nearer noun in a coordinate subject. Experiments using a two-alternative forced-choice paradigm and eyetracking during reading show, first, that a singular second conjunct tends to elicit a singular verb; this influence of a singular noun contrasts with the lack of effect from a singular attractor. Second, in comprehension a singular second conjunct both facilitates processing of an ungrammatical singular verb and inhibits processing of a grammatical plural verb. This symmetrical effect contrasts with the lack of an agreement attraction effect in comprehension of grammatical sentences. It is proposed that variable agreement with coordinate subjects should be given distinct theoretical treatment, relating these phenomena to the cross-linguistic phenomenon of closest conjunct agreement.

International audience; Existing evidence suggests that the parser avoids positing a movement dependency if the grammar does not require doing so. By investigating the processing of two syntactic ambiguities that have not been the subject of processing studies before, we provide more conclusive evidence for this parsing bias in two Romance languages: French and Italian. In two acceptability-judgment experiments and two self-paced-reading studies, we found that sentences that involved a filler–gap dependency (indirect questions in Italian and free relatives in French) were dispreferred compared to sentences involving the same lexical material but no filler–gap dependency (declarative complement clauses in both languages). Crucially, the filler–gap dependency was not dispreferred when there was no available competitor. We conclude by discussing the relevance of these results for syntactic theory, in particular for the questionable status of Merge over Move as a grammatical principle.

The time a reader's eyes spend on a word is influenced by visual (e.g., contrast) as well as lexical (e.g., word frequency) and contextual (e.g., predictability) factors. Well-known visual word recognition models predict that visual and higher-level manipulations may have interactive effects on early eye movement measures, because of cascaded processing between levels. Previous eye movement studies provide conflicting evidence as to whether they do, possibly because of inconsistent manipulations or limited statistical power. In the present study, 2 highly powered experiments used sentences in which a target word's frequency and predictability were factorially manipulated. Experiment 1 also manipulated visual contrast, and Experiment 2 also manipulated font difficulty. Robust main effects of all manipulations were evident in both experiments. In Experiment 1, interactions between the effect of contrast and the effects of frequency and predictability were numerically small and statistically unreliable in both early (word skipping, first fixation duration) and later (gaze duration, go-past time) measures. In Experiment 2, frequency and predictability did demonstrate convincing interactions with font difficulty, but only in the later measures, possibly implicating a checking mechanism. We conclude that although the predicted interactions in early eye movement measures may exist, they are sufficiently weak that they are difficult to detect even in large eye movement experiments. (PsycInfo Database Record (c) 2020 APA, all rights reserved).

We introduce the statistical concept of ’compensatory selection’, which arises when selecting a subset of applicants based on multiple predictors, such as when standardized test scores are used in combination with other predictors required in a school application (e.g., previous grades, references letters, and personal statements). Post-hoc analyses often fail to find a positive correlation between test scores and subsequent success, and this failure is sometimes taken as evidence against the predictive validity of the standardized test. The present analysis reveals that the failure to find a negative correlation indicates that the standardized test is in fact a valid predictor of success. This is due to compensation between predictors during selection: Some students are admitted despite a low test score because their application is exceptional in other respects, while other students are admitted primarily based on a high test score despite weakness in the rest of their application. This compensatory selection process introduces a negative correlation between test scores and other predictors among those admitted (a ’collider bias’ or ’Berkson’s paradox’ effect). If test scores are valid predictors of success, this negative correlation between the predictors counteracts the positive correlation between test scores and success that would have been observed if all applicants were admitted. If test scores are not predictive of success, but were nevertheless used in a compensatory selection process, there would be a spurious negative correlation between test scores and success (i.e., an admitted student with a weak application except for a high test score would be unlikely to succeed). The selection effect that is described here is fundamentally different from the well-known ’restricted range’ problem and can powerfully alter results even in situations that accept most applicants.

In English, when two nouns in a disjunctive subject differ in number (e.g., the dogs or the cat), the verb tends to agree with the number of the nearer noun. This is exceptional, as a noun's linear proximity to the verb does not generally play a role in agreement. In the present study, we investigate a further puzzle about agreement with disjunction, namely, the existence of a pattern in which two singular disjuncts trigger plural agreement (e.g., The lawyer or the accountant are…). Two eyetracking studies in English show that plural agreement with a disjunction of singulars does not reliably disrupt readers' eye movements, in contrast to the immediate disruptive effect of other agreement violations. Three off-line rating studies in English show that plural agreement results in only a small decrement in acceptability, compared to other agreement violations, and that in some structural configurations there is no decrement at all. On the whole, the data do not support the hypothesis that plural agreement is licensed only when or has an inclusive reading; even when it has an exclusive reading, there is only a small penalty for plural agreement. Finally, we explored this issue in Italian, which has a richer system of inflectional morphology. Italian speakers showed a plural preference in a completion experiment, and singular and plural agreement did not differ in acceptability in a rating experiment. We conclude that agreement with disjunction is a grammatical lacuna or gap, in the sense that speakers' grammar simply does not prescribe a verb number following a disjunctive subject.

Previous research ( Mirault, Snell, & Grainger, 2018 ) has demonstrated that subjects sometimes incorrectly judge an ungrammatical sentence as grammatical when it is created by the transposition of two words in a grammatical sentence (e.g., The white was cat big). Here we present two eye-tracking experiments designed to assess the prevalence of this phenomenon in a more natural reading task, and to explore theoretical explanations. Readers failed to notice transpositions at about the same rate as in Mirault et al. (2018) . Failure to notice the transposition was more common when both words were short, and when readers' eyes skipped, rather than directly fixated, one of the two words. The status of the transposed words as open- or closed-class did not have a reliable effect. The transposed words caused disruption in the eye movement record only on trials when participants ultimately judged the sentence to be ungrammatical, not when they judged the sentence to be grammatical. We argue that the results are not entirely consistent with the account offered by Mirault et al. (2018) , which attributes failure to notice transpositions to parallel processing of adjacent words, or with a late, post-perceptual rational inference account ( Gibson, Bergen, & Piantadosi, 2013 ). We propose that word recognition is serial, but post-lexical integration of each word into its context may not be perfectly incremental.

Forced choice recognition is usually assumed to involve a relative judgment process in which each test alternative is matched to memory and the one with the highest memory strength is selected. We monitored eye movements during a forced-choice recognition test to determine if absolute judgments also play a role; that is, do participants ever select an item because its memory strength exceeds an absolute criterion without comparing it to the other item? The results strongly supported a role for absolute judgments. Participants sometimes selected a response without looking at one of the test alternatives, and they were most likely to do this when they looked at the target first and correctly selected it as the studied item. Participants were also more accurate when they looked at the target first than when they looked at the lure first, which would be expected if they sometimes failed to consider the actual target word because they made an incorrect absolute judgment that the lure was studied. Finally, response times were faster when the word selected as the studied item was considered first than when it was considered second; that is, correct responses were faster when the target was viewed first and errors were faster when the lure was viewed first. This would be expected if participants sometimes make absolute judgments that the first item was studied, thus eliminating the additional time needed to consider the second item.

Epub 2018 Apr 12 A word’s predictability, as measured by its cloze probability, has a robust influence on the time a reader’s eyes spend on the word, with more predictable words receiving shorter fixations. However, several previous studies using the boundary paradigm have found no apparent effect of predictability on early reading time measures when the reader does not have valid parafoveal preview of the target word. The present study directly assesses this pattern in two experiments, demonstrating evidence for a null effect of predictability on first fixation and gaze duration with invalid preview, supported by Bayes factor analyses. While the effect of context independent word frequency is shown to survive with invalid preview, consistent with previous studies, the effect of predictability is eliminated with both unrelated word previews and random letter string previews. These results suggest that a word’s predictability influences early stages of orthographic processing, and does so only when perceptual evidence is equivocal, as is the case when the word is initially viewed in parafoveal vision. Word frequency may influence not only early orthographic processing, but also later processing stages. (PsycINFO Database Record (c) 2018 APA, all rights reserved) This work was supported by a grant from the National Science Foundation (BCS 1732008) to the Adrian Staub

The authors of this introductory chapter express their gratitude for the many contributions Lyn Frazier has made to the field of psycholinguistics and to her students, colleagues, and friends. Her introduction of garden-path theory gave new life to the study of sentence comprehension and shaped research on the topic for many years. Throughout her career, she has provided stimulating, often controversial, analyses of how ellipses are processed and of the roles semantics and prosody play in understanding language. Her lively curiosity has led her to explore many other topics in psycholinguistics, including effects of discourse structure and of not-at-issue content, among others. The chapter concludes with an appreciation of the impact she has had as a mentor, colleague, and collaborator, as well as a few remembrances of Lyn’s particular style as a scientist.

This study assessed the reliability of individual differences among fluent adult readers in the effects of four variables - word frequency, predictability, visual contrast, and font difficulty – on eye fixation duration measures, word skipping probability, and regression probability. Split-half reliability was computed in a reanalysis of data from two large, previously published experiments (Staub, 2020) by correlating simple effects in two halves of each experiment (e.g., Hedge, Powell, & Sumner, 2018) and by estimating, in the context of mixed-effects models, a correlation parameter between by-subject slopes for each half (Rouder & Haaf, 2019). The reliability of the effects was generally low, though the second of these methods revealed a few notable exceptions. First, the effects of visual contrast were quite reliable, as expected based on presumed individual differences in contrast sensitivity. Second, the frequency effect on gaze duration was also reliable, but only when raw (as opposed to log) gaze duration was used as the dependent measure. The effect of predictability demonstrated poor reliability for all dependent measures. Model comparison confirmed that model fit was improved by inclusion of by-subject slopes for those effects that showed substantial reliability. These results have implications for the feasibility of studies on individual differences in eye movements in reading, as only experimental effects that demonstrate substantial reliability are good candidates to be explored in individual difference studies.

Two experiments used eyetracking during reading to examine the processing of the matrix verb following object and subject relative clauses. The experiments show that the processing of the matrix verb following an object relative is indeed slowed compared to the processing of the same verb following a subject relative. However, this difficulty is entirely eliminated if additional material intervenes between the object gap and the matrix verb. An explanation in terms of spillover processing is ruled out, suggesting that it is the gap-matrix verb sequence that is itself responsible for the difficulty. We consider two accounts of this difficulty, one emphasizing the potential difficulty of rapidly switching between the sentential subject's thematic or syntactic role in the embedded clause and its role in the matrix clause, and one emphasizing the potential difficulty of performing two demanding memory retrievals in rapid succession. The present experiments also closely replicate the previous findings from eyetracking that the noun phrase and the verb within an object relative are both loci of processing difficulty, but that the former induces substantially greater difficulty.

Listeners tend to categorize an ambiguous speech sound so that it forms a word with its context (Ganong, 1980). This effect could reflect feedback from the lexicon to phonemic activation (McClelland & Elman, 1986), or the operation of a task-specific phonemic decision system (Norris, McQueen, & Cutler, 2000). Because the former account involves feedback between lexical and phonemic levels, it predicts that the lexicon's influence on phonemic decisions should be delayed and should gradually increase in strength. Previous response time experiments have not delivered a clear verdict as to whether this is the case, however. In 2 experiments, listeners' eye movements were tracked as they categorized phonemes using visually displayed response options. Lexically relevant information in the signal, the timing of which was confirmed by separate gating experiments, immediately increased eye movements toward the lexically supported response. This effect on eye movements then diminished over the course of the trial rather than continuing to increase. These results challenge the lexical feedback account. The present work also introduces a novel method for analyzing data from 'visual-world' type tasks, designed to assess when an experimental manipulation influences the probability of an eye movement toward the target. (PsycINFO Database Record

We tested whether failure to notice repetitions of function words during reading (e.g., Amanda jumped off the the swing and landed on her feet.) is due to the eyes' tendency to skip one of the instances of the word. Eye movements were recorded during reading of sentences with repetitions of the word the or repetitions of a noun, after which readers were asked whether an error was present. A repeated the was detected on 46% of trials overall. On trials on which both instances of the were fixated, detection was still only 66%. A repeated noun was detected on 90% of trials, with no significant effect of eye movement patterns. Detecting an omitted the also proved difficult, with eye movement patterns having only a small effect. Readers frequently overlook function word errors even when their eye movements provide maximal opportunity for noticing such errors, but they notice content word repetitions regardless of eye movement patterns. We propose that readers overlook function word errors because they attribute the apparent error to noise in the eye movement control system.

The effect of word frequency on eye movement behaviour during reading has been reported in many experimental studies. However, the vast majority of these studies compared only two levels of word frequency (high and low). Here we assess whether the effect of log word frequency on eye movement measures is linear, in an experiment in which a critical target word in each sentence was at one of three approximately equally spaced log frequency levels. Separate analyses treated log frequency as a categorical or a continuous predictor. Both analyses showed only a linear effect of log frequency on the likelihood of skipping a word, and on first fixation duration. Ex-Gaussian analyses of first fixation duration showed similar effects on distributional parameters in comparing high- and medium-frequency words, and medium- and low-frequency words. Analyses of gaze duration and the probability of a refixation suggested a nonlinear pattern, with a larger effect at the lower end of the log frequency scale. However, the nonlinear effects were small, and Bayes Factor analyses favoured the simpler linear models for all measures. The possible roles of lexical and post-lexical factors in producing nonlinear effects of log word frequency during sentence reading are discussed.

International audience; Three eye movement experiments investigated the processing of the syntactic ambiguity in strings such as the information that the health department provided, where the that-clause can be either a relative clause (RC) or the start of a nominal complement clause (CC; the information that the health department provided a cure). The experiments tested the prediction that comprehenders should avoid the RC analysis because it involves an unforced filler-gap dependency. Readers showed difficulty upon disambiguation toward the RC analysis, and showed facilitated processing of the ambiguous material itself when the CC analysis was available; both patterns suggest rapid initial adoption of the CC analysis in preference to the RC analysis. The strength of the bias of a specific head noun (e.g., information) to appear with a CC did not modulate these effects, nor were these effects reliably modulated by the tendency of an ambiguous string to be completed off-line as a CC or an RC. These results add to the evidence that structural principles guide the processing of filler-gap dependencies

Several studies have examined effects of explicit task demands on eye movements in reading. However, there is relatively little prior research investigating the influence of implicit processing demands. In this study, processing demands were manipulated by means of a between-subject manipulation of comprehension question difficulty. Consistent with previous results from Wotschack and Kliegl, the question difficulty manipulation influenced the probability of regressing from late in sentences and re-reading earlier regions; readers who expected difficult comprehension questions were more likely to re-read. However, this manipulation had no reliable influence on eye movements during first-pass reading of earlier sentence regions. Moreover, for the subset of sentences that contained a plausibility manipulation, the disruption induced by implausibility was not modulated by the question manipulation. We interpret these results as suggesting that comprehension demands influence reading behavior primarily by modulating a criterion for comprehension that readers apply after completing first-pass processing. Refereed/Peer-reviewed

The E-Z Reader 10 model of eye movements in reading (Reichle, Warren, & McConnell, 2009) posits that the process of word identification strictly precedes the process of integration of a word into its syntactic and semantic context. The present study reports a single large-scale ( N = 112) eyetracking experiment in which the frequency and plausibility of a target word in each sentence were factorially manipulated. The results were consistent with E-Z Reader’s central predictions: frequency but not plausibility influenced the probability that the word was skipped over by the eyes rather than directly fixated, and the two variables had additive, not interactive, effects on all reading time measures. Evidence in favor of null effects and null interactions was obtained by computing Bayes factors, using the default priors and sampling methods for ANOVA models implemented by Rouder, Morey, Speckman, and Province (2012). The results suggest that though a word’s plausibility may have a measurable influence as early as the first fixation duration on the target word, in fact plausibility may be influencing only a post-lexical processing stage, rather than lexical identification itself.

In research on the role of lexical predictability in language comprehension, predictability is generally defined as the probability that a word is provided as a sentence continuation in the cloze task (Taylor, 1953), in which subjects are asked to guess the next word of a sentence. The present experiments investigate the process by which subjects generate a cloze response, by measuring the latency to initiate a response in a version of the task in which subjects produce a spoken continuation to a visually presented sentence fragment. Higher probability responses were produced faster than lower probability responses. The latency to produce a response was also influenced by item constraint: A response at a given level of probability was issued faster when the context was more constraining, i.e., a single response was elicited with high probability. We show that these patterns are naturally produced by an activation-based race model in which potential responses independently race towards a response threshold. Implications for the interpretation of cloze probability as a measure of lexical predictability are discussed.

Two eye movement while reading experiments address the issue of how reading of an unpredictable word is influenced by the presence of a more predictable alternative. The experiments replicate the robust effects of predictability on the probability of skipping and on early and late reading time measures. However, in both experiments, an unpredictable but plausible word was read no more slowly when another word was highly predictable (i.e., in a constraining context) than when no word was highly predictable (i.e., in a neutral context). In fact, an unpredictable word that was semantically related to the predictable alternative demonstrated facilitation in the constraining context, in relatively late eye movement measures. These results, which are consistent with Luke and Christianson’s (2016) corpus study, provide the first evidence from a controlled experimental design for the absence of a prediction error cost, and for facilitation of an unpredictable but semantically related word, during normal reading. The findings support a model of lexical predictability effects in which there is broad pre-activation of potential continuations, rather than discrete predictions of specific lexical items. Importantly, pre-activation of likely continuations does not result in processing difficulty when some other word is actually encountered.

Low-frequency (LF) words have higher hit rates and lower false alarm rates than high-frequency (HF) words in recognition memory, a phenomenon termed the mirror effect . Visual word recognition latencies are longer for LF words. We examined the relationship between eye fixation durations during study and later recognition memory for individual words to test whether (1) increased fixation time on a word is associated with better memory, and (2) increased fixation times on LF words can account for their hit rate advantage. In Experiments 1 and 2, words of various frequencies were presented in lists in an intentional study design. In Experiment 3, HF and LF critical words were presented in matched sentence frames in an incidental study design. In all cases, the standard frequency effect on eye movements emerged, with longer reading times for lower frequency words. At test, studied words and new words from each frequency class were presented. The hit rate portion of the mirror effect was evident in all experiments. The time spent fixating a word did predict memory performance in the intentional encoding experiments, but critically, the frequency effect on hit rates was independent of this effect. Time spent fixating a word during incidental encoding did not predict later memory performance. These results suggest that the hit rate advantage for LF words is not due to the additional time spent on these words at encoding, which is consistent with retrieval-stage models of the mirror effect.

The present study investigated the relationship between the location and skew of an individual reader's fixation duration distribution. The ex-Gaussian distribution was fit to eye fixation data from 153 subjects in five experiments, four previously presented and one new. The τ parameter was entirely uncorrelated with the μ and σ parameters; by contrast, there was a modest positive correlation between these parameters for lexical decision and speeded pronunciation response times. The conclusion that, for fixation durations, the degree of skew is uncorrelated with the location of the distribution's central tendency was also confirmed nonparametrically, by examining vincentile plots for subgroups of subjects. Finally, the stability of distributional parameters for a given subject was demonstrated to be relatively high. Taken together with previous findings of selective influence on the μ parameter of the fixation duration distribution, the present results suggest that in reading, the location and the skew of the fixation duration distribution may reflect functionally distinct processes. The authors speculate that the skew parameter may specifically reflect the frequency of processing disruption.

The conditions under which speakers make syntactic errors, and the manner in which listeners respond to them, provide insight into how complex computational problems related to the rules of language are solved. One of the important syntactic rules of English is that a subject and its corresponding verb must agree in number. However, the presence of a number-bearing element between the subject and verb results in frequent production errors and has also been shown to complicate comprehension. When asked to press a button in response to anomalies in several narrated short stories, participants in the current study were better able to detect subject–verb agreement violations when there were no intervening words. In a separate event-related potential (ERP) experiment in which participants listened to the same stories for comprehension, simple subject–verb agreement violations elicited the predicted anterior negativity and later posterior positivity (P600). In contrast, when a singular noun phrase appeared betw...

We examine whether judgments of posterior probabilities in Bayesian reasoning problems are affected by reasoners' beliefs about corresponding real-world probabilities. In an internet-based task, participants were asked to determine the probability that a hypothesis is true (posterior probability, e.g., a person has a disease, given a positive medical test) based on relevant probabilities (e.g., that any person has the disease and the true and false positive rates of the test). We varied whether the correct posterior probability was close to, or far from, independent intuitive estimates of the corresponding 'real-world' probability. Responses were substantially closer to the correct posterior when this value was close to the intuitive estimate. A model in which the response is a weighted sum of the intuitive estimate and an additive combination of the probabilities provides an excellent account of the results.

This chapter discusses the current state of the art with regard to the timecourse of sentence processing in the brain. It outlines the challenges associated with studying timecourse information at the sentence level from a neurobiological perspective and describes competing theoretical and empirical perspectives in this domain. In addition to drawing on findings from neurophysiological methods (electroencephalography [EEG]; magnetoencephalography [MEG]), insights from eye movement measures during natural reading are also taken into account. The chapter concludes that while we are currently unable to make absolute claims about the timecourse of sentence processing from a neurobiological perspective, current evidence supports a cascaded architecture that dynamically combines top-down (including predictive) and bottom-up information sources.

The present study replicated the well-known demonstration by Altmann and Kamide (1999) that listeners make linguistically guided anticipatory eye movements, but used photographs of scenes rather than clip-art arrays as the visual stimuli. When listeners heard a verb for which a particular object in a visual scene was the likely theme, they made earlier looks to this object (e.g., looks to a cake upon hearing The boy will eat …) than when they heard a control verb (The boy will move …). New data analyses assessed whether these anticipatory effects are due to a linguistic effect on the targeting of saccades (i.e., the where parameter of eye movement control), the duration of fixations (i.e., the when parameter), or both. Participants made fewer fixations before reaching the target object when the verb was selectionally restricting (e.g., will eat). However, verb type had no effect on the duration of individual eye fixations. These results suggest an important constraint on the linkage between spoken languag...

In three experimental tasks Stephen and Mirman (2010) measured gaze steps, the distance in pixels between gaze positions on successive samples from an eyetracker. They argued that the distribution of gaze steps is best fit by the lognormal distribution, and based on this analysis they concluded that interactive cognitive processes underlie eye movement control in these tasks. The present comment argues that the gaze step distribution is predictable based on the fact that the eyes alternate between a fixation state in which gaze is steady and a saccade state in which gaze position changes rapidly. By fitting a simple mixture model to Stephen and Mirman’s gaze step data we reveal a fixation distribution and a saccade distribution. This mixture model captures the shape of the gaze step distribution in detail, unlike the lognormal model, and provides a better quantitative fit to the data. We conclude that the gaze step distribution does not directly suggest processing interaction, and we emphasize some important limits on the utility of fitting theoretical distributions to data.

Participants' eye movements were recorded as they read single sentences presented normally, presented entirely in faint text, or presented normally except for a single faint word. Fixations were longer when the entire sentence was faint than when the sentence was presented normally. In addition, fixations were much longer on a single faint word embedded in normal text, compared to when the entire sentence was faint. The primary aim of the study was to examine the influence of stimulus quality on the distribution of fixation durations. Ex-Gaussian fitting revealed that stimulus quality affected the mean of the Normal component, but in contrast to results from single-word tasks (Plourde & Besner, 1997), stimulus quality did not affect the exponential component, regardless of whether one or all words were faint. The results also contrast with the finding (Staub, White, Drieghe, Hollway, & Rayner, 2010) that the word frequency effect on fixation durations is an effect on both of the critical distributional parameters. These findings are argued to have implications for the interpretation of the role of stimulus quality in word recognition, and for models of eye movement control in reading.

An important question in research on eye movements in reading is whether word frequency and word predictability have additive or interactive effects on fixation durations. A fair number of studies have reported only additive effects of the frequency and predictability of a target word on reading times on that word, failing to show significant interactions. Recently, however, Hand, Miellet, O'Donnell, and Sereno (see record 2010-19099-001) reported interactive effects in a study that included the distance of the prior fixation from the target word (launch site). They reported that when the saccade into the target word was launched from very near to the word (within 3 characters), the predictability effect was larger for low frequency words, but when the saccade was launched from a medium distance (4-6 characters from the word) the predictability effect was larger for high frequency words. Hand et al. argued for the importance of including launch site in analyses of target word fixation durations. Here we describe several problems with Hand et al.'s use of analyses of variance in which launch site is divided into distinct ordinal levels. We describe a more appropriate way to analyze such data-linear mixed-effect models-and we use this method to show that launch site does not modulate the interaction between frequency and predictability in two other data sets.

In 3 experiments, the author examined how readers' eye movements are influenced by joint manipulations of a word's frequency and the syntactic fit of the word in its context. In the critical conditions of the first 2 experiments, a high- or low-frequency verb was used to disambiguate a garden-path sentence, while in the last experiment, a high- or low-frequency verb constituted a phrase structure violation. The frequency manipulation always influenced the early eye movement measures of first-fixation duration and gaze duration. The context manipulation had a delayed effect in Experiment 1, influencing only the probability of a regressive eye movement from later in the sentence. However, the context manipulation influenced the same early eye movement measures as the frequency effect in Experiments 2 and 3, though there was no statistical interaction between the effects of these variables. The context manipulation also influenced the probability of a regressive eye movement from the verb, though the frequency manipulation did not. These results are shown to confirm predictions emerging from the serial, staged architecture for lexical and integrative processing of the E-Z Reader 10 model of eye movement control in reading (Reichle, Warren, & McConnell, 2009). It is argued, more generally, that the results provide an important constraint on how the relationship between visual word recognition and syntactic attachment is treated in processing models.

A word's predictability in context has a well-established effect on fixation durations in reading. To investigate how this effect is manifested in distributional terms, an experiment was carried out in which subjects read each of 50 target words twice, once in a high-predictability context and once in a low-predictability context. The ex-Gaussian distribution was fit to each subject's first-fixation durations and single-fixation durations. For both measures, the μ parameter increased when a word was unpredictable, while the τ parameter was not significantly affected, indicating that a predictability manipulation shifts the distribution of fixation durations but does not affect the degree of skew. Vincentile plots showed that the mean ex-Gaussian parameters described the typical distribution shapes extremely well. These results suggest that the predictability and frequency effects are functionally distinct, since a frequency manipulation has been shown to influence both μ and τ. The results may also be seen as consistent with the finding from single-word recognition paradigms that semantic priming affects only μ.

Speakers are known to make subject–verb agreement errors both when a number-mismatching noun intervenes between the head of the subject phrase and the verb (e.g., ∗The key to the cabinets are on the table) and in configurations in which there is a number-mismatching noun that does not intervene (e.g., ∗The cabinets that the key open are on the second floor). Using a two-choice response time (RT) paradigm, Staub (2009) found that correct agreement decisions were also slowed in both cases. The present article reports a new experiment designed to explore whether these two RT effects are qualitatively similar or different. Fitting of the ex-Gaussian distribution ( Ratcliff, 1979 ) to individual subjects’ RT data, in each condition, demonstrated that the effect of an intervening number attractor on correct RT is due to both a shifting of the distribution to the right and to increased skewing, while the effect of a non-intervening attractor is almost entirely a skewing effect. A non-parametric vincentizing procedure supported these conclusions. These findings are taken to support the view that these two types of number attraction involve distinct processing mechanisms.

Using a word-by-word self-paced reading paradigm, T. A. Farmer, M. H. Christiansen, and P. Monaghan (2006) reported faster reading times for words that are phonologically typical for their syntactic category (i.e., noun or verb) than for words that are phonologically atypical. This result has been taken to suggest that language users are sensitive to subtle relationships between sound and syntactic function and that they make rapid use of this information in comprehension. The present article reports attempts to replicate this result using both eyetracking during normal reading (Experiment 1) and word-by-word self-paced reading (Experiment 2). No hint of a phonological typicality effect emerged on any reading-time measure in Experiment 1, nor did Experiment 2 replicate Farmer et al.'s finding from self-paced reading. Indeed, the differences between condition means were not consistently in the predicted direction, as phonologically atypical verbs were read more quickly than phonologically typical verbs, on most measures. Implications for research on visual word recognition are discussed.

A central issue in sentence-processing research is whether the parser entertains multiple analyses of syntactically ambiguous input in parallel, and whether these analyses compete for selection. In this article, we review theoretical positions for and against such competitive parallelism. We then review empirical evidence, primarily drawing on reading time studies, bearing on the prediction made by parallel competitive models that some cost ought to be associated with processing syntactically ambiguous material. We argue that this prediction is not confirmed by the data, and we discuss recent claims that the models in question do not actually make this prediction. We also emphasize the contrast with lexical ambiguity, where there is clearly a processing cost associated with competition between alternate meanings. Finally, we review a different kind of recent evidence suggesting that two syntactic analyses may indeed coexist under specific circumstances.

Two eye movement experiments examined effects on syntactic reanalysis when the correct analysis was briefly entertained at an earlier point in the sentence. In Experiment 1, participants read sentences containing a noun phrase coordination/clausal coordination ambiguity, while in Experiment 2 they read sentences containing a subordinate clause object/main clause subject ambiguity. The critical conditions were designed to induce readers to construct the ultimately correct analysis just prior to being garden-pathed by the incorrect analysis. In both experiments, the earliest measures of the garden path effect were not modulated by this manipulation. However, there was significantly less regressive re-reading of the sentence in those conditions in which the correct analysis was likely to have been constructed, then abandoned, at an earlier point. These results suggest that a syntactic analysis that is abandoned in the course of processing a sentence is not lost altogether, and can be re-activated or retrieved from memory. Implications for models of initial syntactic analysis and reanalysis are discussed.

Several previous studies (Adams, Clifton, & Mitchell, 1998; Mitchell, 1987; van Gompel & Pickering, 2001) have explored the question of whether the parser initially analyzes a noun phrase that follows an intransitive verb as the verb's direct object. Three eyetracking experiments examined this issue in more detail. Experiment 1 strongly replicated the finding (van Gompel & Pickering, 2001) that readers experience difficulty on this noun phrase in normal reading, and found that this difficulty occurs even with a class of intransitive verbs for which a direct object is categorically prohibited. Experiment 2, however, demonstrated that this effect is not due to syntactic misanalysis, but is instead due to disruption that occurs when a comma is absent at a subordinate clause/main clause boundary. Exploring a different construction, Experiment 3 replicated the finding (Pickering & Traxler, 2003; Traxler & Pickering, 1996) that when a noun phrase “filler” is an implausible direct object for an optionally transitive relative clause verb, processing difficulty results; however, there was no evidence for such difficulty when the relative clause verb was strictly intransitive. Taken together, the three experiments undermine the support for the claim that the parser initially ignores a verb's subcategorization restrictions.

Two Very reliable influences on eye fixation durations in reading are word frequency, as measured by corpus counts, and word predictability, as measured by cloze norming. Several studies have reported strictly additive effects of these 2 variables. Predictability also reliably influences the amplitude of the N400 component in event-related potential studies, However, previous research suggests that while :frequency affects the N400 in single-word tasks, it may have little or no effect on the N400 when a word is presented with a preceding sentence context. The present study assessed this apparent dissociation between the results from the 2 methods using a coregistration paradigm in which the frequency and predictability of a target word were manipulated while readers' eye movements and electroencephalograms were simultaneously recorded. We replicated the pattern of significant, and additive; effects of the 2 manipulations on eye fixation durations. We also replicated the predictability effect on the N400, time-locked to the onset of the reader's :first fixation on the target word. However there was no indication of a frequency effect in the electroencephalogram record. We suggest that this pattern has implications both for the interpretation of the N400 and for the interpretation of frequency and predictability effects in language comprehension. Refereed/Peer-reviewed

It is well known that people tend to perform poorly when asked to determine a posterior probability on the basis of a base rate, true positive rate, and false positive rate. The present experiments assessed the extent to which individual participants nevertheless adopt consistent strategies in these Bayesian reasoning problems, and investigated the nature of these strategies. In two experiments, one laboratory-based and one internet-based, each participant completed 36 problems with factorially manipulated probabilities. Many participants applied consistent strategies involving use of only one of the three probabilities provided in the problem, or additive combination of two of the probabilities. There was, however, substantial variability across participants in which probabilities were taken into account. In the laboratory experiment, participants’ eye movements were tracked as they read the problems. There was evidence of a relationship between information use and attention to a source of information. Participants’ self-assessments of their performance, however, revealed little confidence that the strategies they applied were actually correct. These results suggest that the hypothesis of base rate neglect actually underestimates people’s difficulty with Bayesian reasoning, but also suggest that participants are aware of their ignorance.

Readers’ eye movements were monitored as they read sentences in which two noun phrases or two independent clauses were connected by the word or (NP-coordination and S-coordination, respectively). The word either could be present or absent earlier in the sentence. When either was present, the material immediately following or was read more quickly, across both sentence types. In addition, there was evidence that readers misanalyzed the S-coordination structure as an NP-coordination structure only when either was absent. The authors interpret the results as indicating that the word either enabled readers to predict the arrival of a coordination structure; this predictive activation facilitated processing of this structure when it ultimately arrived, and in the case of S-coordination sentences, enabled readers to avoid the incorrect NP-coordination analysis. The authors argue that these results support parsing theories according to which the parser can build predictable syntactic structure before encountering the corresponding lexical input.

In their reply to our article (Staub, Grant, Clifton, & Rayner, 2009), Farmer, Monaghan, Misyak, and Christiansen (2011) argue that prediction is important in sentence processing, that the phonological typicality of nouns and verbs can affect judgments about sentences, and, most centrally, that the phonological typicality of a word with a predicted part of speech has an immediate and substantial effect on on-line reading. The present debate is entirely about the last of these claims.1 The initial report by Farmer, Christiansen, and Monaghan (2006) suggested that phonological atypicality slowed self-paced reading of nouns and verbs by nearly 50 ms. Because we considered an effect of this magnitude to be surprising, we (Staub et al., 2009) conducted and reported two experiments, one using eyetracking and one using self-paced reading, attempting to replicate the effect that Farmer et al. (2006) reported. In neither study did any hint of this effect appear. Farmer et al. (2011) now argue that we did not obtain a phonological typicality effect in our experiments because we intermixed the critical noun and verb stimuli, in contrast to Farmer et al. (2006), who conducted separate experiments with nouns and with verbs. In our intermixed design, they suggest, “it is likely that subjects implicitly learn to recognize the structure shared between the N and V items….and that when such a word order is used, the main verb can be followed by either an N- or V- structure.” They then present data from a new self-paced reading experiment with an intermixed design. As in our experiments, they do not find evidence of a phonological typicality effect overall. However, they argue that there is evidence of an effect at the very beginning of the experiment, which decreases over time, consistent with their proposal that structure-based expectations for a noun or verb diminish over time when the two sentence types are intermixed. We have two objections to this account. The first is simply that there is very little support for Farmer et al.’ s main empirical claim, i.e., that there is a phonological typicality effect at the beginning of their experiment with an intermixed design. The second is that, even if the data were stronger, their hypothesis about the effect of an intermixed experimental design is puzzling. We take these points in order.

OBJECTIVES The purpose of this study was to examine eye-movement patterns in older and younger adults to identify differences in how they respond to both to-be-attended and to-be-ignored speech. DESIGN The study described in this article used an eye-tracking paradigm to provide insight into the factors underlying competing speech understanding in older (n = 23) and younger (n = 22) listeners. Participants attended to a sentence presented in one ear and were instructed to click on a visually displayed word that was heard in that ear while their eye movements were monitored. A foil word also was shown on the screen. Either no sound, steady state noise, or competing speech was presented to the other ear. RESULTS Comparisons between younger and older listeners on all three types of indicators measured in this study (percent correct, response time, and eye movement patterns) demonstrated that older adults were more greatly affected by competing speech than were younger adults. Differences between the groups could not be attributed to the presence of hearing loss in the older participants, as performance for all subjects was at ceiling in quiet and none of the performance metrics was significantly associated with degree of hearing loss. CONCLUSIONS Results of this study support the idea that age-related changes other than lack of audibility or susceptibility to energetic masking negatively affect the ability to understand speech in the presence of a competing message.

This guide accompanies the following article: Charles Clifton Jr and Adrian Staub, ‘Parallelism and Competition in Syntactic Ambiguity Resolution’, Language and Linguistics Compass 2 (2008): 234–250, doi: 10.1111/j.1749-818x.2008.00055.x Introduction Two metaphors have dominated cognitive psychology throughout its history: ‘activation’ and ‘computation’. Activation (itself metaphorically based on neural firing rate) assumes that representations (mental symbols or patterns of non-symbolic ‘nodes’) exist at varying degrees of activation, and high activation of a representation amounts to something like perception or recall. Computation assumes that representations are instead constructed from more elementary components, and that a representation does not exist prior to its construction. We examine the differential implications of these metaphors in the domain of sentence comprehension. Most theories that claim the representation of a sentence is something that is activated by input which proposes that multiple representations are at least temporarily activated, and in order for one representation to be selected, it must de-activate the others in a time-consuming process of competition. Theories that claim that the representation of a sentence is constructed, in contrast, have to posit rules for how the input guides construction, but by and large, these theories do not claim that alternative possible representations compete with each other. We review evidence indicating that time-consuming competition does exist in the process of recognizing individual words, but propose that nearly all existing evidence denies competition in the case of sentence comprehension. Annotated References Clifton, C., Jr., A. Staub, and K. Rayner. 2007. Eye movements in reading words and sentences. Eye movement research: insights into mind and brain, ed. by. R. van Gompel, M. Fisher, W. Murray and R. L. Hill, 341–71. New York: Elsevier. doi:10.1016/B978-008044980-7/50017-3 This chapter contains an extensive review of experiments on eye movements made while reading sentences, examining a number of questions in addition to the one addressed here, namely, whether the eyes slow down while reading a syntactically ambiguous phrase. Duffy, S., G. Kambe, and K. Rayner. 2001. The effect of prior disambiguating context on the comprehension of ambiguous words: evidence from eye movements. On the consequences of meaning selection: perspectives on resolving lexical ambiguity, ed. by D. S. Gorfein, 27–43. Washington, DC: American Psychological Association. doi: 10.1037/10459-002 An accessible review of eyetracking research supporting the existence of competition between alternative meanings of ambiguous words. Elman, J. L., M. Hare, and K. McRae. 2004. Cues, constraints, and competition in sentence processing. Beyond nature-nurture: essays in honor of Elizabeth Bates, ed. by M. Tomasello and D. Slobin, 111–138. Mahwah, NJ: Lawrence Erlbaum Associates. This chapter presents an implemented constraint-based model of sentence comprehension, making a clear prediction that time-consuming competition exists during the reading and the resolution of a syntactic ambiguity, and presenting data that indicate that reading is slowed only during the resolution. Frazier, L. 1987. Sentence processing: a tutorial review. Attention and performance XII, ed. by M. Coltheart, 559–86. Hillsdale, NJ: Lawrence Erlbaum Associates. This is probably the most-cited presentation of the ‘garden-path’ model discussed in the paper. It predates the full development of constraint-based competition models. Frazier, L. (1995). Constraint satisfaction as a theory of sentence processing. Journal of Psycholinguistic Research 24.437–68. doi: 10.1007/BF02143161 This article presents a variety of criticisms of constraint-based models of sentence processing in addition to the current claim that competition in the region of a syntactic ambiguity is not observed. Green, M. J., and D. C. Mitchell. 2006. Absence of real evidence against competition during syntactic ambiguity resolution. Journal of Memory and Language 55.1–17. doi: 10.1016/j.jml.2006.03.003 Green and Mitchell present an interesting claim that local ambiguity does not necessarily result in competition. The claim seems to be correct, but we argue that it does not plausibly apply to most instances of sentence comprehension. MacDonald, M. C., N. J. Pearlmutter, and M. S. Seidenberg. 1994. The lexical nature of syntactic ambiguity resolution. Psychological Review 101.676–703. doi: 10.1037/0033-295X.101.4.676 This is an important article, advocating the claim that sentence comprehension is much the same as word recognition, and that both are characterized by competition between multiple possible representations. van Gompel, R. P. G., M. J. Pickering, and M. J. Traxler, M. 2001. Reanalysis in sentence processing: evidence against current constraint-based and two-stage models. Journal of Memory and Language 45.225–258. doi: 10.1006/jmla.2001.2773 One of a series of studies indicating that syntactic ambiguity can speed reading rather than slowing it. Focus Questions 1. Assuming the authors’ perspective that there is parallel activation of, and competition between, multiple meanings of a word, but not between multiple syntactic analyses, why might this be the case? Are there considerations of efficiency or resource constraints that would give rise to this difference? 2. In what other areas of cognition is there evidence of competition for selection between activated representations, or between response options? 3. The authors suggest that the argument by Green and Mitchell (2006) is implausible because it assumes pre-activation of all possible sentence continuations. Do you agree that this is implausible? Why or why not? 4. In discussing Levy’s (2008) proposals, the authors raise the issue of whether processing behavior at the point of syntactic disambiguation is bimodal or unimodal. Why is this important? How could you tell if there is bimodality? 5. Can you think of predictions made by an account of syntactic processing that assumes parallel activation of multiple alternatives, other than the ones discussed in the article? 6. One possibility that is alluded to briefly in the article is that the reading time advantage obtained by van Gompel and colleagues for globally ambiguous sentences may be because of a failure to fully resolve the ambiguity. Do you know of any specific evidence suggestive of this? Can you think of critical experiments that might address this issue? Topical Outline Syntactic parsing A. Early Models: Heuristics, Delay B. The Garden Path (GP) Model • Focuses on ambiguity resolution • Proposes structural simplicity as primary principle • Assumes a separate reanalysis stage when first-pass parsing fails C. Constraint-Based Models • Simultaneous (optimal) use of many information types • Parallel activation of multiple analyses • Reanalysis as re-weighting or re-ranking D. The State of the Art • Problems for the GP model ○ Evidence for rapid use of non-syntactic information ○ Cases in which parsing preferences do not conform to simplicity metrics • Problems for constraint-based models ○ Failure to show reversals of GP-predicted preferences ○ Failure to show evidence of competition during ambiguity (PRESENT REVIEW GOES HERE) • Emergence of new perspectives ○ Frequency-based accounts ○ Importance of structural prediction.