Ambiguity in sentence processing

As listeners and readers, we rarely notice the ambiguities that pervade our everyday language. When we hear the proverb `Time flies like an arrow' we might ponder its meaning, but not the fact that there are almost 100 grammatically permissible interpretations of this short sentence. On occasion, however, we do notice sentential ambiguity: headlines, such as `Two Sisters Reunited After 18 Years in Checkout Counter', are amusing because they so consistently lead to the unintended interpretation (presumably, the sisters did not spend 18 years at the checkout). It is this consistent preference for one interpretation-and one grammatical structure-rather than another that has fuelled research into sentence processing for more than 20 years. Until relatively recently, the dominant belief had been that these preferences arise from general principles that underlie our use of grammar, with certain grammatical constructions being preferred over others. There has now accrued, however, a considerable body of evidence demonstrating that these preferences are not absolute, but can change in particular circumstances. With this evidence have come new theories of sentence processing, some of which, at first glance, radically question the standard notions of linguistic representation, grammar and understanding.     

Predicted errors in children's early sentence comprehension

Children use syntax to interpret sentences and learn verbs; this is syntactic bootstrapping. The structure-mapping account of early syntactic bootstrapping proposes that a partial representation of sentence structure, the set of nouns occurring with the verb, guides initial interpretation and provides an abstract format for new learning. This account predicts early successes, but also telltale errors: Toddlers should be unable to tell transitive sentences from other sentences containing two nouns. In testing this prediction, we capitalized on evidence that 21-month-olds use what they have learned about noun order in English sentences to understand new transitive verbs. In two experiments, 21-month-olds applied this noun-order knowledge to two-noun intransitive sentences, mistakenly assigning different interpretations to "The boy and the girl are gorping!" and "The girl and the boy are gorping!". This suggests that toddlers exploit partial representations of sentence structure to guide sentence interpretation; these sparse representations are useful, but error-prone.     

Hierarchial chunking in sentence processing

In order to evaluate a left-to-right hierarchical chunking model of sentence perception, Johnson’s Hierarchical Clustering Scheme (HCS) technique was applied to data obtained from sentence intelligibility tests. One hundred and twenty Ss listened to sentences disturbed by white noise. After each presentation they wrote down what they had heard. For each sentence, a table of conditional probabilities p(j/i) was computed, where p(j/i) is the probability that word j had been correctly identified. given correct identification of word i. This was done for all i’s and j’s from the sentence. HCS analysis of the off-diagonal submatrices for which words i precede words j (“forward conditional probabilities ”) yielded satisfactory results. Apparently there is a latent hierarchical structure to these data. The large chunks that appear from these analyses do generally correspond to major syntactic constituents. Minor constituents, however, are very often not reflected in the chunking pattern.     

Phonological typicality and sentence processing

In studies of language, it is widely accepted that the form of a word is independent of its meaning and syntactic category. Thus, the relationship between phonological form and grammatical class would not be expected to affect reading time. However, Farmer et al. have now shown that the phonological typicality of a noun or verb influences how rapidly it is read. This finding has implications for both sentence processing and the interpretation of fixation patterns in reading.     

Prototypicality in sentence production

Three cued-recall experiments examined the effect of category typicality on the ordering of words in sentence production. Past research has found that typical items tend to be mentioned before atypical items in a phrase—a pattern usually associated with lexical variables (like word frequency), and yet typicality is a conceptual variable. Experiment 1 revealed that an appropriate conceptual framework was necessary to yield the typicality effect. Experiment 2 tested ad hoc categories that do not have prior representations in long-term memory and yielded no typicality effect. Experiment 3 used carefully matched sentences in which two category members appeared in the same or in different phrases. Typicality affected word order only when the two words appeared in the same phrase. These results are consistent with an account in which typicality has its origin in conceptual structure, which leads to differences in lexical accessibility in appropriate contexts.     

Meaning specificity in sentence processing.

Two experiments were conducted to test the principle of encoding specificity in sentence recognition when the target sentence was embedded in a meaningful discourse. The results indicate that change in meaning context resulted in significantly less recognition of the target sentences while change in sentence form and the target-test interval had no significant effects. The findings were discussed within the framework of a semantic interpretation of the encoding specificity principle.     

Closed-class immanence in sentence production

The closed-class hypothesis asserts that function words play a privileged role in syntactic processes. In language production, the claim is that such words are intrinsic to, identified with, or immanent in phrasal skeletons. Two experiments tested this hypothesis with a syntactic priming procedure. In both, subjects tended to produce utterances in the same syntactic forms as priming sentences, with the structures of the self-generated sentences varying as a function of differences in the structures of the primes. Changes in the closed-class elements of the priming sentences had no effect on this tendency over and above the impact of the structural changes. These results suggest that free-standing closed-class morphemes are not inherent components of the structural frames of English sentences.     

Context effects and coding stations in sentence processing

Blakar, R. M. Context effects and coding stations in sentence processing. Scand. J. Psychol., 1973, 14, 103–105.-Using a proof-reading technique, with a sentence split up in such a way that the 16 subjects participating had to correct each typing error at different coding stations within the sentence, it was found: (a) Extending the context perceived represented an active process, where every new part included in the context influenced the element processed (at the moment); even total re-structuring of the interpretation of an element was evident. (b) Some of the coding points artificially introduced by the technique represented adequate coding stations, while others did not, either because they resulted in uncertainty as regards the missing "letter" or gave misleading interpretations. The critical difference between adequate and inadequate coding stations would appear to lie in the varying access given at each point to relevant context.     

Planning units and syntax in sentence production

The study was conducted to determine, first, whether it is the deep or the surface clause that is more important as a speech planning unit, and second, whether syntactic decisions are made during sentence production. Subjects, while talking, heard tones to which they had to respond by pressing a button; reaction times to these tones were taken as an index of processing load during production. It was found that there were increased RTs at the ends compared with the beginnings of deep structure clauses. No difference was found between RTs at the beginnings and ends of surface clauses not corresponding to a deep clause. The results were interpreted as showing that deep clauses are major planning units and that some planning for clauses occurs at the end of the preceding clause. Differences were found between RTs during clauses of different syntactic structure. These results were interpreted as indicating that syntax influences production and were discussed in relation to previous studies of pausing and speech disruption.     

Verb processing during sentence comprehension in aphasia

This study examines verb processing during on-line sentence comprehension in aphasia. We describe two experiments that explore whether a group of Broca's aphasics, who were agrammatic in comprehension as well as speech, a group of fluent aphasics, and a group of normal controls are sensitive to the argument structure arrangements of verbs. Subjects had to perform a complex secondary task both in the immediate vicinity of the verb and also at a point well past the verb while listening to sentences for meaning. Reaction times to this secondary task show that both normal controls and agrammatic Broca's aphasic subjects activate multiple argument structure possibilities for a verb in the vicinity of the verb, yet at a point downstream from the verb such effects disappear. These data suggest that the problems agrammatic subjects show with verbs in sentence comprehension, and the general lexical access deficit also recently claimed to be part of the agrammatics' problem, may not extend to the real-time processing of verbs and their arguments. Fluent aphasic subjects, on the other hand, do not show sensitivity to the argument structure properties of verbs, suggesting that these patients may have a semantic-like sentence processing deficit.     

Attentional resource and processing speed limitations during sentence processing in Parkinson's disease

Several studies have suggested that patients with Parkinson's disease (PD) have sentence comprehension difficulty in part because of their limited executive resources. However, these assessments confound the executive resources contributing to sentence comprehension with the resources needed for task performance. In the present study, we used a word detection technique that minimizes task demands in order to evaluate attentional and processing speed resources during the comprehension of simple sentences without subordinate clauses and sentences containing subject-relative and object-relative center-embedded subordinate clauses. We found that PD patients have poor sensitivity to phonetic errors embedded in unbound grammatical morphemes, regardless of the clausal structure of the sentence, suggesting difficulty attending to grammatical morphemes. We also found that PD patients are significantly slowed in their sensitivity to phonetic errors in content words embedded in object-relative center-embedded sentences. Slowed sensitivity to content words in object-relative sentences was correlated with timed executive measures of planning. On a traditional measure of comprehension, these PD patients were impaired for sentences containing object-relative center-embedded clauses compared to sentences with subject-relative center-embedded clauses, and comprehension of object-relative sentences was correlated with executive measures. Our findings are consistent with the claim that limited executive resources for strategic attention and processing speed contribute to the sentence comprehension difficulties of PD patients.     

Language integration in bilingual sentence production

To what extent are processes used in sentence production integrated between the different languages of a bilingual and to what extent are they kept separate? We consider three models that differ in their assumptions about the degree of integration: De Bot's [De Bot, K. (1992). A bilingual production model: Levelt's Speaking model adapted. Applied Linguistics, 13, 1-24] bilingual blueprint of the speaker, Ullman's [Ullman, M. T. (2001). The neural basis of lexicon and grammar in first and second language: The declarative/procedural model. Bilingualism: Language and Cognition, 4, 105-122] declarative/procedural model of bilingualism, and Hartsuiker et al.'s [Hartsuiker, R. J., Pickering, M. J., & Veltkamp, E. (2004). Is syntax separate or shared between languages? Cross-linguistic syntactic priming in Spanish/English bilinguals. Psychological Science, 15, 409-414] integrated model. A review of the evidence from bilingual sentence production studies shows that Hartsuiker et al.'s predictions are supported, but argues against the other two models. We discuss some repercussions for bilingual language use.     

Verb effects during sentence processing.

We describe a study that explores how the properties of a verb's lexical entry contribute to on-line sentence processing. Schmauder (1991) has recently suggested that the verb-processing effects shown previously by Shapiro and his colleagues (Shapiro & Levine, 1990; Shapiro, Zurif, & Grimshaw, 1987; 1989) are not generalizable and may have been artifacts precipitated by experimental design factors. In this paper we logically analyze such a possibility and then present experiments that systematically investigate the design differences between the Shapiro et al. studies cited above and the Schmauder study. These analyses and experiments provide further evidence that verb effects during sentence processing are real, are to be expected given the architectures of recent parsing models, and are replicable using a dual task that is modified in reasonable ways.     

Measuring lexical access during sentence processing

The results from “on-line” investigations of sentence comprehension are often difficult to interpret since it is not always apparent what component processes are reflected in the response measure. The results of two experiments reported here indicate that response latencies from phoneme-triggered lexical decision (PTLD) reflect the time needed for lexical access during sentence processing. Listeners were presented with sentences and were asked to make a word/nonword judgment for items beginning with a particular word-initial target phoneme. Speed of lexical access was manipulated by varying the semantic predictability of the target-bearing word. WORD judgments were faster for words that were preceded by semantically related verbs than were WORD judgments for words that were preceded by neutral verbs. The present results are consistent with other studies showing semantic facilitation of lexical access during the processing of fluent speech. It is argued that the phoneme-triggered lexical-decision task is a more suitable measure of lexical access during sentence processing than phoneme monitoring (Foss, 1969) or word monitoring (Marslen-Wilson & Tyler, 19751. In addition, it is pointed out that the phoneme-triggered lexical-decision task lends itself to modifications which should enable investigators to study various aspects of on-line sentence processing.     

Large-scale neural network for sentence processing

Our model of sentence comprehension includes at least grammatical processes important for structure-building, and executive resources such as working memory that support these grammatical processes. We hypothesized that a core network of brain regions supports grammatical processes, and that additional brain regions are activated depending on the working memory demands associated with processing a particular grammatical feature. We used functional magnetic resonance imaging (fMRI) to test this hypothesis by comparing cortical activation patterns during coherence judgments of sentences with three different syntactic features. We found activation of the ventral portion of left inferior frontal cortex during judgments of violations of each grammatical feature. Increased recruitment of the dorsal portion of left inferior frontal cortex was seen during judgments of violations of specific grammatical features that appear to involve a more prominent working memory component. Left posterolateral temporal cortex and anterior cingulate were also implicated in judging some of the grammatical features. Our observations are consistent with a large-scale neural network for sentence processing that includes a core set of regions for detecting and repairing several different kinds of grammatical features, and additional regions that appear to participate depending on the working memory demands associated with processing a particular grammatical feature.     

The development of complex sentence processing strategies

Children aged 3, 4, and 5 years and adults heard sentences with clauses connected by after, and, or before, saw a picture, and indicated whether or not the picture matched one of the events of the sentence. Response times were taken as a measure of immediate accessibility to the meaning of the clause that the picture was about. Temporal organization of sentence meanings was dominant in 3-year-olds and adults, but not in 4- or 5-year-olds. The 3-year-olds and especially the adults processed and-sentences as implicitly temporal. The results for 4- and 5-year-olds are interpreted as indicating experimentation with alternate strategies for organizing sentences based on the structural/presuppositional properties of clauses.     

Event-related potential indices of ambiguous sentence processing

An event-related potential (ERP) probe was used to examine various models of ambiguous sentence processing. ERPs to light flashes were recorded during and immediately after auditorily presented ambiguous and unambiguous target sentences. Each target sentence was preceded by either a relevant or a neutral context sentence. Principal component analyses of the ERPs indicated that although certain components varied as a function of ambiguity, none of the components varied as a function of preceding context. These findings provided some support for a postdecision model of ambiguity processing which suggests that both meanings of an ambiguity are always processed, even when prior disambiguating context is available.     

Effects of clausal structure and word frequency in sentence processing

A phoneme-monitoring task was employed to test the effects of clausal structure and lexical ambiguity on sentence processing. Results supported the hypothesis that the clause serves as a psychologically real unit of sentence processing, with the semantic interpretation of each clause being assigned at the clause boundary. The frequency of the ambiguous or control word preceding the critical item in the phoneme-monitoring task was also found to affect the results obtained, with higher frequency words leading to longer mean reaction times.This research is based on a thesis submitted by the first author as partial fulfillment of the M.A. degree at The University of Iowa. Portions of this paper were presented at the meeting of the Psychonomic Society, San Antonio, Texas, 1978.