Syntactic Prediction Research Articles

Syntactic predictions and asyntactic comprehension in aphasia: Evidence from scope relations

People with aphasia (PWA) often fail to understand syntactically complex sentences. This phenomenon has been described as asyntactic comprehension and has been explored in various studies cross-linguistically in the past decades. However, until now there has been no consensus among researchers as to the nature of sentence comprehension failures in aphasia. Impaired representations accounts ascribe comprehension deficits to loss of syntactic knowledge, whereas processing/resource reduction accounts assume that PWA are unable to use syntactic knowledge in comprehension due to resource limitation resulting from the brain damage. The aim of this paper is to use independently motivated psycholinguistic models of sentence processing to test a variant of the processing/resource reduction accounts that we dub the Complexity Threshold Hypothesis. According to this hypothesis, PWA are capable of building well-formed syntactic representations, but, because their resources for language processing are limited, their syntactic parser fails when processing complexity exceeds a certain threshold. The source of complexity investigated in the experiments reported in this paper is syntactic prediction. We conducted two experiments involving comprehension of sentences with different types of syntactic dependencies, namely dependencies that do not require syntactic prediction (i.e. unpredictable dependencies in sentences that require Quantifier Raising) and dependencies whose resolution requires syntactic predictions at an early stage of processing based on syntactic cues (i.e. predictable dependencies in movement-derived sentences). In line with the predictions of the Complexity Threshold Hypothesis, the results show that the agrammatic patients that participated in this study had no difficulties comprehending sentences with the former type of dependencies, whereas their comprehension of sentences with the latter type of dependencies was impaired.

Language structure in the brain: A fixation-related fMRI study of syntactic surprisal in reading

How is syntactic analysis implemented by the human brain during language comprehension? The current study combined methods from computational linguistics, eyetracking, and fMRI to address this question. Subjects read passages of text presented as paragraphs while their eye movements were recorded in an MRI scanner. We parsed the text using a probabilistic context-free grammar to isolate syntactic difficulty. Syntactic difficulty was quantified as syntactic surprisal, which is related to the expectedness of a given word's syntactic category given its preceding context. We compared words with high and low syntactic surprisal values that were equated for length, frequency, and lexical surprisal, and used fixation-related (FIRE) fMRI to measure neural activity associated with syntactic surprisal for each fixated word. We observed greater neural activity for high than low syntactic surprisal in two predicted cortical regions previously identified with syntax: left inferior frontal gyrus (IFG) and less robustly, left anterior superior temporal lobe (ATL). These results support the hypothesis that left IFG and ATL play a central role in syntactic analysis during language comprehension. More generally, the results suggest a broader cortical network associated with syntactic prediction that includes increased activity in bilateral IFG and insula, as well as fusiform and right lingual gyri.

Online Syntactic Prediction in Agrammatic Aphasia: Evidence from ERPs

Online Syntactic Prediction in Agrammatic Aphasia: Evidence from ERPs

Structural prediction in aphasia

Structural prediction in aphasia

Multi-voxel pattern analysis of noun and verb differences in ventral temporal cortex

Recent evidence suggests a probabilistic relationship exists between the phonological/orthographic form of a word and its lexical-syntactic category (specifically nouns vs. verbs) such that syntactic prediction may elicit form-based estimates in sensory cortex. We tested this hypothesis by conducting multi-voxel pattern analysis (MVPA) of fMRI data from early visual cortex (EVC), left ventral temporal (VT) cortex, and a subregion of the latter – the left mid fusiform gyrus (mid FG), sometimes called the “visual word form area.” Crucially, we examined only those volumes sampled when subjects were predicting, but not viewing, nouns and verbs. This allowed us to investigate prediction effects in visual areas without any bottom-up orthographic input. We found that voxels in VT and mid FG, but not in EVC, were able to classify noun-predictive trials vs. verb-predictive trials in sentence contexts, suggesting that sentence-level predictions are sufficient to generate word form-based estimates in visual areas.

Structural syntactic prediction measured with ELAN: Evidence from ERPs

The current study used event-related potentials (ERPs) to investigate how and when argument structure information is used during the processing of sentences with a filler-gap dependency. We hypothesize that one specific property – animacy (living vs. non-living) – is used by the parser during the building of the syntactic structure. Participants heard sentences that were rated off-line as having an expected noun (Who did the Lion King chase the caravan with?) or an unexpected noun (Who did Lion King chase the animal with?). This prediction is based on the animacy properties relation between the wh-word and the noun in the object position. ERPs from the noun in the unexpected condition (animal) elicited a typical Early Left Anterior Negativity (ELAN)/P600 complex compared to the noun in the expected condition (caravan). Firstly, these results demonstrate that the ELAN reflects not only grammatical category violation but also animacy property expectations in filler-gap dependency. Secondly, our data suggests that the language comprehension system is able to make detailed predictions about aspects of the upcoming words to build up the syntactic structure.

Active prediction of syntactic information during sentence processing

We describe an eye-tracking experiment that tested the effect of syntactic predictability on skipping rates during reading. We found that plural noun phrases were skipped more often than singular noun phrases, in syntactic contexts which induced a high expectation for a plural. We interpret this effect as evidence that the plural noun phrase has been predicted ahead of time. The results indicate that the examination of skipping rates might be a useful tool for the investigation of syntactic prediction effects.

Before the N400: Effects of lexical–semantic violations in visual cortex

There exists an increasing body of research demonstrating that language processing is aided by context-based predictions. Recent findings suggest that the brain generates estimates about the likely physical appearance of upcoming words based on syntactic predictions: words that do not physically look like the expected syntactic category show increased amplitudes in the visual M100 component, the first salient MEG response to visual stimulation. This research asks whether violations of predictions based on lexical–semantic information might similarly generate early visual effects. In a picture–noun matching task, we found early visual effects for words that did not accurately describe the preceding pictures. These results demonstrate that, just like syntactic predictions, lexical–semantic predictions can affect early visual processing around ∼100 ms, suggesting that the M100 response is not exclusively tuned to recognizing visual features relevant to syntactic category analysis. Rather, the brain might generate predictions about upcoming visual input whenever it can. However, visual effects of lexical–semantic violations only occurred when a single lexical item could be predicted. We argue that this may be due to the fact that in natural language processing, there is typically no straightforward mapping between lexical–semantic fields (e.g., flowers) and visual or auditory forms (e.g., tulip, rose, magnolia). For syntactic categories, in contrast, certain form features do reliably correlate with category membership. This difference may, in part, explain why certain syntactic effects typically occur much earlier than lexical–semantic effects.

Left inferior frontal lesions impair syntactic predictions at the sensory level

Event Abstract Back to Event Left inferior frontal lesions impair syntactic predictions at the sensory level Anna S. Hasting1*, Maria Jakuszeit1 and Sonja A. Kotz1 1 MPI for Human Cognitive and Brain Sciences, Germany Recent event-related EEG and MEG studies have shown that the detection of syntactic errors in spoken and written language coincides with sensory processes in the 100 ms range both temporally (e.g. Hasting & Kotz, 2008) and spatially (e.g. Herrmann et al., 2009). It has further been shown that sensory cortices are indeed sensitive to certain syntactic cues (Dikker et al., 2009). The present study aims to clarify whether such early sensory syntactic processing is based on syntactic rule representations within the sensory cortices, or whether it reflects rapid mismatch responses that are enabled by top-down syntactic predictions at the sensory level. The left inferior frontal cortex has frequently been found to serve syntactic functions (e.g. Friederici & Kotz, 2003) and is therefore a prime candidate region for the generation of syntactic predictions. We tested 10 patients with left inferior frontal lesions but intact temporal cortices in a passive auditory event-related potential (ERP) paradigm that had reliably elicited early grammaticality effects in the past (Hasting & Kotz, 2008). Neither word category violations nor subject-verb agreement violations elicited early grammaticality effects in these patients, whereas a group of 10 age-matched controls showed reliable negativities in response to both violation types (150 - 250 ms and 100 - 300 ms, respectively). These results show that intact auditory cortices and surrounding temporal regions do not suffice to generate early syntax-specific ERP effects. They rather suggest that early sensory aspects of syntactic analysis critically depend on top-down predictions generated by the left-inferior frontal cortex. Keywords: EEG, Language Conference: XI International Conference on Cognitive Neuroscience (ICON XI), Palma, Mallorca, Spain, 25 Sep - 29 Sep, 2011. Presentation Type: Poster Presentation Topic: Poster Sessions: Neural Bases of Language Citation: Hasting AS, Jakuszeit M and Kotz SA (2011). Left inferior frontal lesions impair syntactic predictions at the sensory level. Conference Abstract: XI International Conference on Cognitive Neuroscience (ICON XI). doi: 10.3389/conf.fnhum.2011.207.00510 Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters. The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated. Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed. For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions. Received: 22 Nov 2011; Published Online: 28 Nov 2011. * Correspondence: Dr. Anna S Hasting, MPI for Human Cognitive and Brain Sciences, Leipzig, Germany, hasting@cbs.mpg.de Login Required This action requires you to be registered with Frontiers and logged in. To register or login click here. Abstract Info Abstract The Authors in Frontiers Anna S Hasting Maria Jakuszeit Sonja A Kotz Google Anna S Hasting Maria Jakuszeit Sonja A Kotz Google Scholar Anna S Hasting Maria Jakuszeit Sonja A Kotz PubMed Anna S Hasting Maria Jakuszeit Sonja A Kotz Related Article in Frontiers Google Scholar PubMed Abstract Close Back to top Javascript is disabled. Please enable Javascript in your browser settings in order to see all the content on this page.

Syntactic Prediction in Sentence Reading: Evidence from Eye Movements

The current study was conducted to examine syntactic prediction and structural representation of Chinese coordinate structure during the process of Chinese sentence comprehension. We recorded the eye movements of each participant by using SensoMotoric Instruments (SMI) iView Hi-Speed system while they were seated in front of a monitor (resolution of 1024 × 768 pixel). Sixty native speakers of Chinese were recruited, and were paid for their participation. All of them were native speakers of Mandarin Chinese who had normal or corrected-to-normal vision. The experiment materials were composed by 24 critical sentences and other 60 filler sentences. A 2 × 2 factorial design was adopted, with syntactic prediction and temporal ambiguity as within-participant variables. The results from four reading time measures, including the first fixation duration, first pass reading time, go-past time as well as the percent regressions, showed that syntactic prediction could facilitate the processing of Chinese coordinate sentence with local ambiguity. For both types of ambiguous sentences, the occurrence of the word "either" could greatly reduced the time of first fixation duration, first pass reading, go-past, and percent regressions on the critical region and the spillover region, although there was no significant influence of syntactic prediction on the NP region. As revealed by the statistic analysis, there was a significant interaction between these two experimental variables. Moreover, the manipulation of temporal ambiguity was also significant when syntactic prediction was unavailable but not when syntactic processing could be successfully predicted. In contrast with the results of previous English studies, the present study demonstrates that syntactic prediction can facilitate Chinese sentence comprehension and help people deal with ambiguous syntactic constituents. Therefore, in our study, when readers met the sentences with syntactic prediction, they would adopt the sentence-coordination analysis. However, if the sentences were not predicated, they would prefer the NP-coordination analysis and treated the second NP as part of the direct object rather than as the subject of a clause.

The effect of noun phrase type on working memory saturation during sentence comprehension

Double centre-embedded structures such as “the rat the cat the boy chased ate was brown” seem ungrammatical to many human subjects. Using an offline complexity judgement task, Gibson and Thomas (1999) demonstrated that people found such sentences no more difficult to understand when the second verb phrase (VP) was omitted, relative to a condition where all the required VPs were present. According to the Syntactic Prediction Locality Theory (SPLT; Gibson, 1998), this syntactic illusion is determined by the high working memory cost associated with the integration of the second VP. This cost could be reduced by replacing the third noun phrase (the boy) by a pronoun, making the reader more sensitive to the omission of the second VP. This hypothesis was tested in two experiments using French sentences. Both experiments confirmed the syntactic illusion when the second VP was not a pronoun. The second experiment measured the reading times of the VPs and showed that the pronoun induced a longer reading time of the final VP when the second VP was omitted. The overall results indicate a condition under which human subjects could process the most complex part of a sentence with more than one embedded relative clause. The overall results are consistent with most of the hypotheses derived from the SPLT although offline complexity judgements could not be the most sensitive measure to test some of these hypotheses.

Sensitivity to syntax in visual cortex

One of the most intriguing findings on language comprehension is that violations of syntactic predictions can affect event-related potentials as early as 120 ms, in the same time-window as early sensory processing. This effect, the so-called early left-anterior negativity (ELAN), has been argued to reflect word category access and initial syntactic structure building ( Friederici, 2002). In two experiments, we used magnetoencephalography to investigate whether (a) rapid word category identification relies on overt category-marking closed-class morphemes and (b) whether violations of word category predictions affect modality-specific sensory responses. Participants read sentences containing violations of word category predictions. Unexpected items varied in whether or not their word category was marked by an overt function morpheme. In Experiment 1, the amplitude of the visual evoked M100 component was increased for unexpected items, but only when word category was overtly marked by a function morpheme. Dipole modeling localized the generator of this effect to the occipital cortex. Experiment 2 replicated the main results of Experiment 1 and eliminated two non-morphology-related explanations of the M100 contrast we observed between targets containing overt category-marking and targets that lacked such morphology. Our results show that during reading, syntactically relevant cues in the input can affect activity in occipital regions at around 125 ms, a finding that may shed new light on the remarkable rapidity of language processing.

Textos padronizados em português (BR) para medida da velocidade de leitura: comparação com quatro idiomas europeus

To develop standardized texts in Brazilian Portuguese to assess reading speed and compare performances among four European languages. 10 texts were designed by a linguistic expert at the level of a sixth grade reading material (reading ages 10-12 years) and were matched for length and syntactic complexity, according to the syntactic prediction locality theory of Gibson. Normally sighted native speaking volunteers aged 18-35 years read each text aloud at random. The reading speed was on average 1100 +/- 167 (standard deviation) characters per minute. Only small differences were found between the measured reading speeds of the 10 texts, and these differences were not statistically significant in groups of at least 6 texts. The mean reading speed in age-matched volunteers in Finnish was: 1263 +/- 142; French: 1214 +/- 152; English: 1234 +/- 147; and German: 1126 +/- 105. The authors developed a set of standardized, homogeneous, and comparable texts in Brazilian Portuguese. These texts will be a valuable tool to measure reading speed for repeated measurements and in international studies in the field of reading and low vision research.

Locality, Cyclicity, and Resumption: At the Interface between the Grammar and the Human Sentence Processor

We present an experimental investigation of the role of resumptive pronouns. We investigate object extraction in WH-questions for a range of syntactic configurations (nonislands, weak islands, strong islands) and for multiple levels of embedding (single, double, and triple). In order to establish the crosslinguistic properties of resumption, parallel experiments were conducted in three languages, viz. English, Greek, and German. Three main experimental results are reported. First, resumption does not remedy island violations: resumptive pronominals are at most as acceptable as gaps, but not more acceptable. This result disconfirms claims in the literature that resumptives can ‘save’ island violations. Second, embedding reduces acceptability even in extraction out of nonislands and declaratives, structures standardly assumed to be fully grammatical. Third, nonislands and weak islands pattern together and contrast with strong islands in terms of the effect of resumption and embedding. Our experimental findings show a remarkable consistency across the three languages we investigate; crosslinguistic variation appears confined to quantitative differences in crosslinguistically identical principles. We argue that these experimental results can be explained by the interaction of grammatical principles with resource limitations of the human parser. In particular, extraction from nonislands and weak islands imposes increased demands on the computational resources of the parser. We extend Gibson’s (1998) syntactic prediction locality theory in order to formalize this intuition and account for the processing complexity of A-bar dependencies.

영한 기계번역에서 구문 분석 정확성 향상을 위한 구문 범주 예측

영한 기계번역에서 구문 분석 정확성 향상을 위한 구문 범주 예측

Syntactic prediction in language comprehension: Evidence from either...or.

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.

The syntactic prediction with token automata: application to HandiAS system

This paper presents a finite-state machine to compute the probability of a word appearance when one knows the left syntactic context. We memorize the token number of words in a dictionary and the token number of syntactic categories on finite-state automata. We compute a word probability with these numbers. If we have not predicted the awaited word, we take into account the first letter for a new prediction, and so on. This system has been implemented on a prototype software for disabled communication aid, called HandiAS and it is a part of the Research project CNHL of the LI, the Computer Laboratory of the Tours University.

The Complexity of Nested Structures in Japanese

This article presents two questionnaire experiments that investigate the processing complexity of a variety of nested constructions in Japanese. The first experiment demonstrated that embedded structures containing a direct object NP in the most embedded clause were more complex than comparable nested structures that lacked an object NP in the most embedded clause. The second experiment demonstrated that a construction consisting of a relative clause embedded within a sentential complement is less complex than the reverse embedding, consisting of sentential complement embedded within a relative clause. These results are discussed in terms of the syntactic prediction locality theory (Gibson 1998).

Linguistic complexity: locality of syntactic dependencies

This paper proposes a new theory of the relationship between the sentence processing mechanism and the available computational resources. This theory – the Syntactic Prediction Locality Theory (SPLT) – has two components: an integration cost component and a component for the memory cost associated with keeping track of obligatory syntactic requirements. Memory cost is hypothesized to be quantified in terms of the number of syntactic categories that are necessary to complete the current input string as a grammatical sentence. Furthermore, in accordance with results from the working memory literature both memory cost and integration cost are hypothesized to be heavily influenced by locality (1) the longer a predicted category must be kept in memory before the prediction is satisfied, the greater is the cost for maintaining that prediction; and (2) the greater the distance between an incoming word and the most local head or dependent to which it attaches, the greater the integration cost. The SPLT is shown to explain a wide range of processing complexity phenomena not previously accounted for under a single theory, including (1) the lower complexity of subject-extracted relative clauses compared to object-extracted relative clauses, (2) numerous processing overload effects across languages, including the unacceptability of multiply center-embedded structures, (3) the lower complexity of cross-serial dependencies relative to center-embedded dependencies, (4) heaviness effects, such that sentences are easier to understand when larger phrases are placed later and (5) numerous ambiguity effects, such as those which have been argued to be evidence for the Active Filler Hypothesis.

Parsing Complex Sentences with Structured Connectionist Networks

A modular, recurrent connectionist network is taught to incrementally parse complex sentences. From input presented one word at a time, the network learns to do semantic role assignment, noun phrase attachment, and clause structure recognition, for sentences with both active and passive constructions and center-embedded clauses. The network makes syntactic and semantic predictions at every step. Previous predictions are revised as expectations are confirmed or violated with the arrival of new information. The network induces its own "grammar rules" for dynamically transforming an input sequence of words into a syntactic/semantic interpretation. The network generalizes well and is tolerant of ill-formed inputs.