Blog Archives
A person is not a number: Discourse involvement in subject-verb agreement computation
Agreement is a very important mechanism for language processing. Mainstream psycholinguistic research on subject-verb agreement processing has emphasized the purely formal and encapsulated nature of this phenomenon, positing an equivalent access to person and number features. However, person and number are intrinsically different, because person conveys extra-syntactic information concerning the participants in the speech act. To test the person-number dissociation hypothesis we investigated the neural correlates of subject-verb agreement in Spanish, using person and number violations. While number agreement violations produced a left-anterior negativity followed by a P600 with a posterior distribution, the negativity elicited by person anomalies had a centro-posterior maximum and was followed by a P600 effect that was frontally distributed in the early phase and posteriorly distributed in the late phase. These data reveal that the parser is differentially sensitive to the two features and that it deals with the two anomalies by adopting different strategies, due to the different levels of analysis affected by the person and number violations.
from Brain Research
Cue-dependent interference in comprehension
The role of interference as a primary determinant of forgetting in memory has long been accepted, however its role as a contributor to poor comprehension is just beginning to be understood. The current paper reports two studies, in which speed-accuracy tradeoff and eye-tracking methodologies were used with the same materials to provide converging evidence for the role of syntactic and semantic cues as mediators of both proactive (PI) and retroactive interference (RI) during comprehension. Consistent with previous work (e.g., Van Dyke & Lewis, 2003), we found that syntactic constraints at the retrieval site are among the cues that drive retrieval in comprehension, and that these constraints effectively limit interference from potential distractors with semantic/pragmatic properties in common with the target constituent. The data are discussed in terms of a cue-overload account, in which interference both arises from and is mediated through a direct-access retrieval mechanism that utilizes a linear, weighted cue-combinatoric scheme.
from the Journal of Memory and Language
Shedding light on words and sentences: Near-infrared spectroscopy in language research
Investigating the neuronal network underlying language processing may contribute to a better understanding of how the brain masters this complex cognitive function with surprising ease and how language is acquired at a fast pace in infancy. Modern neuroimaging methods permit to visualize the evolvement and the function of the language network. The present paper focuses on a specific methodology, functional near-infrared spectroscopy (fNIRS), providing an overview over studies on auditory language processing and acquisition. The methodology detects oxygenation changes elicited by functional activation of the cerebral cortex. The main advantages for research on auditory language processing and its development during infancy are an undemanding application, the lack of instrumental noise, and its potential to simultaneously register electrophysiological responses. Also it constitutes an innovative approach for studying developmental issues in infants and children. The review will focus on studies on word and sentence processing including research in infants and adults.
from Brain and Language
Unification of sentence processing via ear and eye: An fMRI study
We present new evidence based on fMRI for the existence and neural architecture of an abstract supramodal language system that can integrate linguistic inputs arising from different modalities such that speech and print each activate a common code. Working with sentence material, our aim was to find out where the putative supramodal system is located and how it responds to comprehension challenges. To probe these questions we examined BOLD activity in experienced readers while they performed a semantic categorization task with matched written or spoken sentences that were either well-formed or contained anomalies of syntactic form or pragmatic content. On whole-brain scans, both anomalies increased net activity over non-anomalous baseline sentences, chiefly at left frontal and temporal regions of heteromodal cortex. The anomaly-sensitive sites correspond approximately to those that previous studies () have found to be sensitive to other differences in sentence complexity (object relative minus subject relative). Regions of interest (ROIs) were defined by peak response to anomaly averaging over modality conditions. Each anomaly-sensitive ROI showed the same pattern of response across sentence types in each modality. Voxel-by-voxel exploration over the whole brain based on a cosine similarity measure of common function confirmed the specificity of supramodal zones.
from Cortex
Direct-access retrieval during sentence comprehension: Evidence from Sluicing
Language comprehension requires recovering meaning from linguistic form, even when the mapping between the two is indirect. A canonical example is ellipsis, the omission of information that is subsequently understood without being overtly pronounced. Comprehension of ellipsis requires retrieval of an antecedent from memory, without prior prediction, a property which enables the study of retrieval in situ ([Martin and McElree, 2008] and [Martin and McElree, 2009]). Sluicing, or inflectional-phrase ellipsis, in the presence of a conjunction, presents a test case where a competing antecedent position is syntactically licensed, in contrast with most cases of nonadjacent dependency, including verb–phrase ellipsis. We present speed–accuracy tradeoff and eye-movement data inconsistent with the hypothesis that retrieval is accomplished via a syntactically guided search, a particular variant of search not examined in past research. The observed timecourse profiles are consistent with the hypothesis that antecedents are retrieved via a cue-dependent direct-access mechanism susceptible to general memory variables.
from the Journal of Memory and Language
Objects, events and “to be” verbs in Spanish – An ERP study of the syntax–semantics interface
In Spanish, objects and events at subject position constrain the selection of different forms of the auxiliary verb “to be”: locative predicates about objects require “estar en”, while those relating to events require “ser en”, both translatable as “to be in”. Subjective ratings showed that while the “object + ser + en” is considered as incorrect, the “event + estar + en” combination is also perceived as unacceptable but to a lesser degree. In an ERP study, we evaluated the impact of a purely semantic distinction (object versus events) on the subsequent processing of these auxiliary verbs followed by locatives in Spanish. For the “ser en” predicate, the P600 component was larger when the subject was an object than when it was an event. This P600 effect is consistent with an online repair of the defining predicate when it does not fit with the adequate semantic properties of the subject. On the other hand, for the “estar en” predicate, event subjects when compared to object subjects showed more positive ongoing amplitudes between 280 and 380 ms after the presentation of the “en” preposition, followed by a longer positive wave starting around 400 ms and lasting until 700 ms after the presentation of the following determiner, with central and frontal scalp distributions respectively. Thus, the different subject-predicate combinations, depending on the semantic features of the subjects, triggered syntactic reparatory processes at a structural level. These findings are consistent with an incremental interpretation of sentence meaning based on the interaction between syntactic and semantic information.
from Brain and Language
Register affects language comprehension: ERP evidence from article omission in newspaper headlines
Language processing involving syntax-discourse interface operations has been claimed to be particularly resource-consuming. In production, this additional complexity is claimed to be the source of article omission in the speech of young children and certain language-impaired speakers. In comprehension, article omission in some “special registers” (e.g., newspaper headlines) has been attributed to the trade-off between spending more processing resources and increasing processing speed. We investigated the comprehension of noun phrases (NPs) with and without articles (e.g., (a) policeman arrests (a) monk) when readers were or were not aware of reading headlines by recording electrophysiological responses. The presence of an N400-effect suggests first that comprehension of article-less NPs exerts processing demands and indicates that article omission does not result in costs from morphosyntactic processing (no LAN), but from linking processes at the syntax-discourse interface supported by discourse-semantic memory. Differences between the instruction modalities suggest that awareness of the special register plays a role. The electrophysiological data thus demonstrate that register has an effect on the discourse-semantic integration of NPs and provide evidence for a certain degree of top-down processing.
from the Journal of Neurolinguistics
The need for quantitative methods in syntax and semantics research
The prevalent method in syntax and semantics research involves obtaining a judgement of the acceptability of a sentence/meaning pair, typically by just the author of the paper, sometimes with feedback from colleagues. This methodology does not allow proper testing of scientific hypotheses because of (a) the small number of experimental participants (typically one); (b) the small number of experimental stimuli (typically one); (c) cognitive biases on the part of the researcher and participants; and (d) the effect of the preceding context (e.g., other constructions the researcher may have been recently considering). In the current paper we respond to some arguments that have been given in support of continuing to use the traditional nonquantitative method in syntax/semantics research. One recent defence of the traditional method comes from Phillips (2009), who argues that no harm has come from the nonquantitative approach in syntax research thus far. Phillips argues that there are no cases in the literature where an incorrect intuitive judgement has become the basis for a widely accepted generalisation or an important theoretical claim. He therefore concludes that there is no reason to adopt more rigorous data collection standards. We challenge Philips’ conclusion by presenting three cases from the literature where a faulty intuition has led to incorrect generalisations and mistaken theorising, plausibly due to cognitive biases on the part of the researchers. Furthermore, we present additional arguments for rigorous data collection standards. For example, allowing lax data collection standards has the undesirable effect that the results and claims will often be ignored by researchers with stronger methodological standards. Finally, we observe that behavioural experiments are easier to conduct in English than ever before, with the advent of Amazon.com’s Mechanical Turk, a marketplace interface that can be used for collecting behavioural data over the internet.
Representational complexity and memory retrieval in language comprehension
Mental representations formed from words or phrases may vary considerably in their feature-based complexity. Modern theories of retrieval in sentence comprehension do not indicate how this variation and the role of encoding processes should influence memory performance. Here, memory retrieval in language comprehension is shown to be influenced by a target’s representational complexity in terms of syntactic and semantic features. Three self-paced reading experiments provide evidence that reading times at retrieval sites (but not earlier) decrease when more complex phrases occur as filler phrases in filler-gap dependencies. The data also show that complexity-based effects are not dependent on string length, syntactic differences, or the amount of processing the stimuli elicit. Activation boosting and reduced similarity-based interference are implicated as likely sources of these complexity-based effects.
Binding in agrammatic aphasia: Processing to comprehension
These data suggest that comprehension difficulties with binding constructions seen in agrammatic aphasic patients are not due to a deficit in automatic syntactic processing or delayed processing. Rather, they point to a possible deficit in lexical integration.
from Aphasiology
Grammatical weight and relative clause extraposition in English
In relative clause extraposition (RCE) in English, a noun is modified by a non-adjacent RC, resulting in a discontinuous dependency, as in: Three people arrived here yesterday who were from Chicago. Although discourse focus is known to influence the choice of RCE over truth-conditionally equivalent sentences with canonical structure (Rochemont and Culicover, English focus constructions and the theory of grammar, Cambridge University Press, 1990; Takami, A functional constraint on Extraposition from NP, John Benjamins, 1999), Hawkins (Efficiency and complexity in grammars, Oxford University Press, 2004) and Wasow (Postverbal behavior, CSLI Publications, 2002) have proposed in addition that RCE should be preferred when the relative clause is long (or ‘heavy’) relative to the VP because such structures are processed more efficiently in comprehension and production. The current study tested this hypothesis based on Hawkins’ (Efficiency and complexity in grammars, Oxford University Press, 2004) domain minimization principles. In an acceptability judgment task, canonical sentences were rated significantly higher than extraposition sentences when the RC was light, but this difference disappeared when the RC was heavy. In a self-paced reading task, extraposition sentences were read significantly faster than canonical sentences when the RC was heavy, but there was no difference when the RC was light. In an analysis of RCE in the ICE-GB corpus, extraposed RCs were significantly longer than the VP on average, whereas canonical RCs were significantly shorter, and the proportion of sentences with extraposition decreased as the ratio of VP-to-RC length increased. These findings support Hawkins’ (Efficiency and complexity in grammars, Oxford University Press, 2004) domain minimization principles and help explain why a discontinuous dependency is allowed and sometimes preferred even in a language with relatively fixed word order.
Syntactic priming in comprehension: Parallelism effects with and without coordination
Although previous research has shown a processing facilitation for conjoined phrases that share the same structure, it is currently not clear whether this parallelism advantage is specific to particular syntactic environments such as coordination, or whether it is an example of more general effect in sentence comprehension. Here, we report three eye-tracking experiments that test for parallelism effects both in coordinated noun phrases and in subordinate clauses. The first experiment replicated previous findings, showing that the second conjunct of a coordinated noun phrase was read more quickly when it had the same structure as the first conjunct, compared with when it did not. Experiment 2 examined parallelism effects in noun phrases that were not linked by coordination. Again, a reading time advantage was found when the second noun phrase had the same structure as the first. Experiment 3 compared parallelism effects in coordinated and non-coordinated syntactic environments. The parallelism effect was replicated for both environments, and was statistically equivalent whether or not coordination was involved. This demonstrated that parallelism effects can be found outside the environment of coordination, suggesting a general syntactic priming mechanism as the underlying explanation.
from the Journal of Memory and Language
Plasticity of grammatical recursion in German learners of Dutch
Previous studies have examined cross-serial and embedded complement clauses in West Germanic in order to distinguish between different types of working memory models of human sentence processing, as well as different formal language models. Here, adult plasticity in the use of these constructions is investigated by examining the response of German-speaking learners of Dutch using magnetoencephalography (MEG). In three experimental sessions spanning their initial acquisition of Dutch, participants performed a sentence-scene matching task with Dutch sentences including two different verb constituent orders (Dutch verb order, German verb order), and in addition rated similar constructions in a separate rating task. The average planar gradient of the evoked field to the initial verb within the cluster revealed a larger evoked response for the German order relative to the Dutch order between 0.2 to 0.4 s over frontal sensors after 2 weeks, but not initially. The rating data showed that constructions consistent with Dutch grammar, but inconsistent with the German grammar were initially rated as unacceptable, but this preference reversed after 3 months. The behavioural and electrophysiological results suggest that cortical responses to verb order preferences in complement clauses can change within 3 months after the onset of adult language learning, implying that this aspect of grammatical processing remains plastic into adulthood.
COMD News 