The time course of semantic and syntactic processing in Chinese sentence comprehension: Evidence from eye movements

In the present study, we examined the time course of semantic and syntactic processing when Chinese is read. Readers’ eye movements were monitored, and the relation between a single-character critical word and the sentence context was manipulated such that three kinds of sentences were developed: (1) congruent, (2) those with a semantic violation, and (3) those with both a semantic and a syntactic violation. The eye movement data showed that the first-pass reading times were significantly longer for the target region in the two violation conditions than in the congruent condition. Moreover, the semantic + syntactic violation caused more severe disruption than did the pure semantic violation, as reflected by longer first-pass reading times for the target region and by longer go-past times for the target region and posttarget region in the former than in the latter condition. These results suggest that the effects of, at least, a semantic violation can be detected immediately by Chinese readers and that the processing of syntactic and semantic information is distinct in both first-pass and second-pass reading.     

Lexical decision and aphasia: Evidence for semantic processing

Wernicke's and Broca's-Conduction aphasics and a Global aphasic were presented with a lexical-decision task in which English words and pronounceable nonwords were preceded by semantically related, unrelated, or nonword primes. The patients were also given a simple semantic-judgment task using the word pairs from the lexical-decision task. Wernicke's aphasics performed similar to normals and Broca's-Conduction aphasics showing significantly shorter latencies in making real-word identifications when preceded by a semantically related word. In addition, both superordinate and coordinate associates showed semantic-priming effects. Performance on the semantic-judgment task showed significantly more impairment in the aphasic group than in the normal controls. These results suggest that aphasics with even severe language impairments retain stored semantic information that may be automatically activated, yet is inaccessible to conscious semantic decision during metalinguistic tasks.     

The interplay between semantic and referential aspects of anaphoric noun phrase resolution: Evidence from ERPs

In this event-related brain potential (ERP) study, we examined how semantic and referential aspects of anaphoric noun phrase resolution interact during discourse comprehension. We used a full factorial design that crossed referential ambiguity with semantic incoherence. Ambiguous anaphors elicited a sustained negative shift (Nref effect), and incoherent anaphors elicited an N400 effect. Simultaneously ambiguous and incoherent anaphors elicited an ERP pattern resembling that of the incoherent anaphors. These results suggest that semantic incoherence can preclude readers from engaging in anaphoric inferencing. Furthermore, approximately half of our participants unexpectedly showed common late positive effects to the three types of problematic anaphors. We relate the latter finding to recent accounts of what the P600 might reflect, and to the role of individual differences therein.     

The processing of negative sentences in fluent aphasics: semantic and pragmatic aspects.

This paper examines the processing of negative sentences with respect to their truth value in fluent aphasic subjects. Three groups of 25 subjects each (men and women) were studied: the first group consisted of fluent aphasics, the second consisted of right temporal lobe-damaged subjects, and the third of normal subjects. An experiment of sentence verification was carried out. A quantitative study of reaction time (RT) and a qualitative study of the types of answers were carried out. The results show that the fluent aphasics were impaired in their logical-semantic processing of negative sentences and showed a predilection for the affirmative in answering strategies. The right temporal lobe-damaged subjects showed differences only in respect to the normal subjects in the RT, but not in answering strategies. The results suggest that in Wernicke's Aphasics there is a tendency to substitute a type of contextual-pragmatic processing for the logical-semantic processing.     

Making sense of sentences in schizophrenia: electrophysiological evidence for abnormal interactions between semantic and syntactic processing

Event-related potentials to critical verbs were measured as patients with schizophrenia and healthy controls read sentences word by word. Relative to their preceding context, critical verbs were (a) congruous, (b) incongruous and semantically unrelated to individual preceding words (pragmatic-semantic violations), (c) incongruous but semantically related to individual preceding words (animacy-semantic violations), or (d) syntactically anomalous. The N400 was modulated normally in patients, suggesting that semantic integration between individual words within sentences was normal in schizophrenia. The amplitude of the P600 to both syntactic and animacy-semantic violations was reduced in patients relative to controls. The authors suggest that, in schizophrenia, an abnormality in combining semantic and syntactic information online to build up propositional meaning leaves sentence processing to be primarily driven by semantic relationships between individual words.     

Verbal working memory and on-line syntactic processing: Evidence from self-paced listening

Two experiments investigated the relationship between performance on standard tests of verbal working-memory and the on-line construction of syntactic form. In Experiment 1, working-memory was measured in 100 college students on a version of the Daneman and Carpenter (1980) reading-span task, and online syntactic processing was assessed using a self-paced listening task with four sentence types. In Experiment 2, working-memory was measured in 48 college students on two versions of the reading-span task and two other tests of verbal working-memory, and on-line syntactic processing was assessed using the self-paced listening task with an additional sentence type. In both experiments, there was no relationship between working-memory capacity and the increase in processing time seen for the on-line construction of syntactic form for either syntactically more complex or syntactically simpler sentences. The results indicate that the capacity of the working-memory system that is measured by standard working-memory tests does not determine the efficiency of on-line syntactic processing. They are consistent with the view that the working-memory system used for parsing is at least partially separate from that measured by traditional measures of working-memory capacity.     

On differentiating syntactic and semantic features in emerging grammars: Evidence from empty form usage

Recent emphasis on underlying semantic relations in the child's acquisition of grammar has left ignored those cases where syntactic operations can be observed relatively independent of semantic relations. Such operations may reveal optional processes in the child's transition from single-word utterances to grammatical usage that may be related to specific linguistic rather than general cognitive abilities.     

Integration of syntactic and semantic information in predictive processing: cross-linguistic evidence from German and English

Two visual-world eyetracking experiments were conducted to investigate whether, how, and when syntactic and semantic constraints are integrated and used to predict properties of subsequent input. Experiment 1 contrasted auditory German constructions such as, The hare-nominative eats ... (the cabbage-acc) versus The hare-accusative eats ... (the fox-nom), presented with a picture containing a hare, fox, cabbage, and distractor. We found that the probabilities of the eye movements to the cabbage and fox before the onset of NP2 were modulated by the case-marking of NP1, indicating that the case-marking (syntactic) information and verbs' semantic constraints are integrated rapidly enough to predict the most plausible NP2 in the scene. Using English versions of the same stimuli in active/passive voice (Experiment 2), we replicated the same effect, but at a slightly earlier position in the sentence. We discuss the discrepancies in the two Germanic languages in terms of the ease of integrating information across, or within, constituents.     

The role of verb argument structure in sentence processing: Distinguishing between syntactic and semantic effects

This paper describes an ongoing research program designed to investigate how syntactic and semantic aspects of lexical information become available to the sentence processing system. The two experiments described here distinguished between syntactic and semantic representations by using cross-modal naming and lexical decision in a new way. The relationship between the main verb and the probe word was varied such that the probe word met either the syntactic criteria to be an argument, the semantic criteria, neither, or both the syntactic and semantic criteria. Lexical decision times were sensitive to both syntactic and semantic congruity, while naming times were sensitive only to syntactic congruity. The two tasks were then used to investigate syntactic and semantic representations when verb argument structure was ambiguous. Subcategorized structures were constructed without regard for biasing context, but the contextually inappropriate thematic frame was ruled out while the inappropriate syntactic frame was still available.     

Assessment of associative relations in alzheimer's disease: Evidence for preservation of semantic memory

Abstract

Semantic priming was assessed in Alzheimer's disease (AD), elderly control, and young control subjects with three lexical decision, continuous priming experiments. the stimuli were intracategory, associated pairs in Experiments 1 and 2 and varied types of associated pairs in Experiment 3. AD priming effects were equal to those of elderly control subjects; elderly and young control priming effects were also equal. We interpret this as evidence for the relative preservation of the semantic memory network in early AD, as in normal aging, at least to the extent necessary for normal access of concept nodes and normal automatic spreading activation between concept nodes. In a final study (Experiment 4) knowledge of the associations between the targets and related primes used in Experiments 1 and 2 was explicitly assessed; 20 out of 22 AD subjects showed perfect or close to perfect performance.     

Semantic memory and the granularity of semantic relations: Evidence from speed-accuracy decomposition

In the present study we examined whether semantic relations are atomistic unitary associations, or are complex concepts consisting of a number of relational elements. The complexity of the ownership relation was assessed by combining a relation verification task (“Many people own [cars/ comets]”) with the speed-accuracy decomposition procedure (Meyer, Irwin, Osman, & Kounios, 1988). The latter permits one to determine whether subjects achieve their final state of response accuracy in a single, discrete all-or-none transition, or whether the relevant processes yield partial information representing intermediate states of knowledge. The rationale was that the retrieval of a unitary relational link from a classical associative network should be an all-or-none affair. In contrast, a set of relational elements need not be processed as a unitary bundle, thereby allowing partial response-information states. In two experiments, we found evidence of such partial information (i.e., sensitivity in units ofd′), lending support to the notion that relations are complex. Furthermore, the results suggest that the accumulation of guessing sensitivity was linear over time, weighing against alternate theoretical interpretations.     

SCEGRAM: An image database for semantic and syntactic inconsistencies in scenes

Our visual environment is not random, but follows compositional rules according to what objects are usually found where. Despite the growing interest in how such semantic and syntactic rules – a scene grammar – enable effective attentional guidance and object perception, no common image database containing highly-controlled object-scene modifications has been publically available. Such a database is essential in minimizing the risk that low-level features drive high-level effects of interest, which is being discussed as possible source of controversial study results. To generate the first database of this kind – SCEGRAM – we took photographs of 62 real-world indoor scenes in six consistency conditions that contain semantic and syntactic (both mild and extreme) violations as well as their combinations. Importantly, always two scenes were paired, so that an object was semantically consistent in one scene (e.g., ketchup in kitchen) and inconsistent in the other (e.g., ketchup in bathroom). Low-level salience did not differ between object-scene conditions and was generally moderate. Additionally, SCEGRAM contains consistency ratings for every object-scene condition, as well as object-absent scenes and object-only images. Finally, a cross-validation using eye-movements replicated previous results of longer dwell times for both semantic and syntactic inconsistencies compared to consistent controls. In sum, the SCEGRAM image database is the first to contain well-controlled semantic and syntactic object-scene inconsistencies that can be used in a broad range of cognitive paradigms (e.g., verbal and pictorial priming, change detection, object identification, etc.) including paradigms addressing developmental aspects of scene grammar. SCEGRAM can be retrieved for research purposes from http://www.scenegrammarlab.com/research/scegram-database/.     

Attention and semantic processing during speech: an fMRI study

This fMRI study was conducted to investigate whether language semantics is processed even when attention is not explicitly directed to word meanings. In the “unattended” condition, the subjects performed a visual detection task while hearing semantically related and unrelated word pairs. In the “phoneme” condition, the subjects made phoneme judgements between prime and target words, and in the “word” condition, they indicated whether the words in each word pair were semantically related or unrelated. In the word condition, stronger activity was obtained for unrelated than related words in the superior temporal gyrus/sulcus (STG/STS), lingual/fusiform gyri and cuneus, whereas in the phoneme condition no such effect was observed. However, during the unattended condition, stronger activity was found in the right STG. Our results suggest that both semantic judgement of word pairs and their passive hearing activate the same neural networks but this activation is more restricted in the passive hearing.     

Neural responses to morphological, syntactic, and semantic properties of single words: an fMRI study

Dissociations in the recognition of specific classes of words have been documented in brain-injured populations. These include deficits in the recognition and production of morphologically complex words as well as impairments specific to particular syntactic classes such as verbs. However, functional imaging evidence for distinctions among the neural systems underlying these dissociations has been inconclusive. We explored the neural systems involved in processing different word classes in a functional Magnetic Resonance Imaging study, contrasting four groups of words co-varying morphological complexity (simple, monomorphemic words vs complex derived or inflected words) and syntactic class (verbs vs nouns/adjectives). Subtraction of word from letter string processing showed activation in left frontal and temporal lobe regions consistent with prior studies of visual word processing. No differences were observed for morphologically complex and simple words, despite adequate power to detect stimulus specific effects. A region of posterior left middle temporal gyrus showed significantly increased activation for verbs. Post hoc analyses showed that this elevated activation could also be related to semantic properties of the stimulus items (verbs have stronger action associations than nouns, and action association is correlated with activation). Results suggest that semantic as well as syntactic factors should be considered when assessing the neural systems involved in single word comprehension.     

Phonological and semantic priming: Evidence for task-independent effects

The questions asked in the present experiments concern the generality of semantic and phonological priming effects: Do these effects arise automatically regardless of target task, or are these effects restricted to target tasks that specifically require the retrieval of the primed information? In Experiment 1, subjects produced faster color matching times on targets preceded by a masked rhyming prime than on targets preceded by an orthographic control or an unrelated prime. This result suggests that automatic priming effects on the basis of phonological similarity can be obtained even when the target task does not make use of phonological information. This claim was reinforced in Experiment 2 in which a rhyme priming effect and a semantic priming effect were found in a semantic categorization task. In Experiment 3, the target task was phonological (rhyme detection), and, again, both phonological and semantic priming effects were observed. Finally, in Experiments 4 and 5, in a replication and an extension of Experiment 1, phonological and semantic priming effects were found in a color matching task, a task involving neither phonological nor semantic processing. These results are most straightforwardly interpreted by assuming that both semantic and phonological priming effects are, at least in part, due to automatic activation of memorial representations.     

A comparison of sentence- and discourse-level semantic processing: an ERP study

Eighteen subjects listened to sentences ending with semantically congruent or incongruent words. Each congruent sentence was embedded at the end of a short story so that the final word was semantically acceptable at the sentence level but incongruent in terms of discourse context. The same stories were also presented with congruent endings. Auditory ERPs elicited by the final words in these sentences and stories were compared. Both sentence- and discourse-level semantic integration were associated with N400 and Late Positive Component (LPC) effects in addition to a new component, P550. Local and global semantic processing, although evoking the same components, were characterized by differential effects on ERP amplitudes according to the amount of text integrated and its congruence. These results indicate similar cognitive processes of context build up, underlying sentence- and discourse-semantic processing.     

Negative priming between pictures and words in a selective attention task: Evidence for semantic processing of ignored stimuli

This study examined the processing of ignored pictures and words when attention was directed to a different picture or word. Previous work by Tipper (1985) demonstrated that the priming effect of an ignored picture on a subsequent categorically related picture is inhibitory. This effect was termednegative priming. Tipper concluded that ignored pictures achieved abstract categorical levels of internal representation, and that these representations were inhibited during selection of a simultaneously presented picture. This conclusion, however, was premature. Observation of the figures used by Tipper suggests that objects within a category have greater structural similarity than do objects in different categories. The negative priming effect could therefore be at a structural level of representation. The present study examined priming across symbolic domains (pictures and words) where there was no structural relationship between objects. Negative priming was again observed and was equivalent to the negative priming observed within symbolic domain. These data suggest that ignored drawings and words do achieve abstract categorical levels of representation, and that the mechanisms underlying negative priming operate at, or beyond, this level.     

Sentence processing in an artificial language: Learning and using combinatorial constraints

A study combining artificial grammar and sentence comprehension methods investigated the learning and online use of probabilistic, nonadjacent combinatorial constraints. Participants learned a small artificial language describing cartoon monsters acting on objects. Self-paced reading of sentences in the artificial language revealed comprehenders’ sensitivity to nonadjacent combinatorial constraints, without explicit awareness of the probabilities embedded in the language. These results show that even newly-learned constraints have an identifiable effect on online sentence processing. The rapidity of learning in this paradigm relative to others has implications for theories of implicit learning and its role in language acquisition.     

Auditory language comprehension: an event-related fMRI study on the processing of syntactic and lexical information

The functional specificity of different brain areas recruited in auditory language processing was investigated by means of event-related functional magnetic resonance imaging (fMRI) while subjects listened to speech input varying in the presence or absence of semantic and syntactic information. There were two sentence conditions containing syntactic structure, i.e., normal speech (consisting of function and content words), syntactic speech (consisting of function words and pseudowords), and two word-list conditions, i.e., real words and pseudowords. The processing of auditory language, in general, correlates with significant activation in the primary auditory cortices and in adjacent compartments of the superior temporal gyrus bilaterally. Processing of normal speech appeared to have a special status, as no frontal activation was observed in this case but was seen in the three other conditions. This difference may point toward a certain automaticity of the linguistic processes used during normal speech comprehension. When considering the three other conditions, we found that these were correlated with activation in both left and right frontal cortices. An increase of activation in the planum polare bilaterally and in the deep portion of the left frontal operculum was found exclusively when syntactic processes were in focus. Thus, the present data may be taken to suggest an involvement of the left frontal and bilateral temporal cortex when processing syntactic information during comprehension.