A Bayesian-Network Approach to Lexical Disambiguation

Lexical ambiguity can be syntactic if it involves more than one grammatical category for a single word, or semantic if more than one meaning can be associated with a word. In this article we discuss the application of a Bayesian-network model in the resolution of lexical ambiguities of both types. The network we propose comprises a parsing subnetwork, which can be constructed automatically for any context-free grammar, and a subnetwork for semantic analysis, which, in the spirit of Fillmore's (1968) case grammars, seeks to fulfill the required cases of all candidates for verb of the sentence. Solving for the highest joint probability of the variables conditioned upon the evidences to the network yields the most likely candidate with its meaning, along with its cases and respective meanings. This is achieved by fixing the values of all evidence nodes concurrently, and then performing a stochastic simulation in which the remaining nodes are updated probabilistically with a high degree of parallelism. The process of disambiguation is directed neither by the syntax nor the semantics, but rather by the interrelation between the two subnetworks. The use of a Bayesian-network model allows us to express this interrelation between the two subnetworks and among their constituents in a rather direct and rigorous way that, in connection with the convergence properties of the stochastic simulation, reveals a very robust model.     

Race-Based Parsing and Syntactic Disambiguation

We present a processing model that integrates same important psychological claims about the human sentence-parsing mechanism: namely, that processing is influenced by limitations an working memory and by various syntactic preferences. The model uses time-constraint information to resolve conflicting preferences in a psychologically plausible way. The starting paint far this proposal is the Sausage Machine model (Fodor & Frazier, 1980: Frazier & Fodor, 1978). From there, we attempt to overcome the original model's dependence an ad hoc aspects of its grammar, and its omission of verb-frame preferences. We also add mechanisms far lexical disambiguation and semantic processing in parallel with syntactic processing.     

Toward a Unified Model of Lexical Acquisition and Lexical Access

Much effort has gone into constructing models of how children segment speech and thereby discover the words of their language. Much effort has also gone into constructing models of how adults access their mental lexicons and thereby segment speech into words. In this paper, I explore the possibility of a model that could account for both word discovery by children and on-line segmentation by adults. In particular, I discuss extensions to the distributional regularity (DR) model of Brent and Cartwright (1996) that could yield an account of on-line segmentation as well as word discovery.     

Effects and Limitations of Prosodic and Semantic Biases on Syntactic Disambiguation

This paper examined the effects of prosody on the syntactic ambiguity resolution of Japanese sentences, especially with reference to the interaction with semantic bias. Syntactically ambiguous sentences with different types of semantic bias were constructed. The degree of bias in each sentence was evaluated through visual presentation experiments. Three types of sentences were selected based on the results of visual presentation experiments, were recorded with prosody maximally favoring each possible interpretation of the sentences, and were used as the stimuli for the auditory presentation experiments. The results showed that prosodic cues can influence the interpretation of a sentence even when the sentence is strongly semantically biased. The results also showed a limitation to prosodic cues. The prosodic biases alone were not sufficient to fully determine the interpretation of the sentences even when the sentences were neutrally biased semantically.     

Single-Stage Prediction Models Do Not Explain the Magnitude of Syntactic Disambiguation Difficulty

The disambiguation of a syntactically ambiguous sentence in favor of a less preferred parse can lead to slower reading at the disambiguation point. This phenomenon, referred to as a garden-path effect, has motivated models in which readers initially maintain only a subset of the possible parses of the sentence, and subsequently require time-consuming reanalysis to reconstruct a discarded parse. A more recent proposal argues that the garden-path effect can be reduced to surprisal arising in a fully parallel parser: words consistent with the initially dispreferred but ultimately correct parse are simply less predictable than those consistent with the incorrect parse. Since predictability has pervasive effects in reading far beyond garden-path sentences, this account, which dispenses with reanalysis mechanisms, is more parsimonious. Crucially, it predicts a linear effect of surprisal: the garden-path effect is expected to be proportional to the difference in word surprisal between the ultimately correct and ultimately incorrect interpretations. To test this prediction, we used recurrent neural network language models to estimate word-by-word surprisal for three temporarily ambiguous constructions. We then estimated the slowdown attributed to each bit of surprisal from human self-paced reading times, and used that quantity to predict syntactic disambiguation difficulty. Surprisal successfully predicted the existence of garden-path effects, but drastically underpredicted their magnitude, and failed to predict their relative severity across constructions. We conclude that a full explanation of syntactic disambiguation difficulty may require recovery mechanisms beyond predictability.     

Lexical and Prosodic Effects on Syntactic Ambiguity Resolution in Aphasia

The purpose of this study was to determine whether and when individuals with aphasia and healthy controls use lexical and prosodic information during on-line sentence comprehension. Individuals with aphasia and controls (n = 12 per group) participated in a self-paced listening experiment. The stimuli were early closure sentences, such as "While the parents watched(,) the child sang a song." Both lexical and prosodic cues were manipulated. The cues were biased toward the subject- or object- of the ambiguous noun phrase (the child). Thus, there were two congruous conditions (in which both lexical cues and prosodic cues were consistent) and two incongruous conditions (in which lexical and prosodic cues conflicted). The results showed that the people with aphasia had longer listening times for the ambiguous noun phrase (the child) when the cues were conflicting, rather than consistent. The controls showed effects earlier in the sentence, at the subordinate verb (watched or danced). Both groups showed evidence of reanalysis at the main verb (sang). These effects demonstrate that the aphasic group was sensitive to the lexical and prosodic cues, but used them on a delayed time course relative to the control group.     

Syllabic Effects in Italian Lexical Access

Two cross-modal priming experiments tested whether lexical access is constrained by syllabic structure in Italian. Results extend the available Italian data on the processing of stressed syllables showing that syllabic information restricts the set of candidates to those structurally consistent with the intended word (Experiment 1). Lexical access, however, takes place as soon as possible and it is not delayed till the incoming input corresponds to the first syllable of the word. And, the initial activated set includes candidates whose syllabic structure does not match the intended word (Experiment 2). The present data challenge the early hypothesis that in Romance languages syllables are the units for lexical access during spoken word recognition. The implications of the results for our understanding of the role of syllabic information in language processing are discussed.     

双语词汇通达的认知控制

基于熟练性的认知控制研究认为,双语词汇通达的非选择性和选择性是由于认知控制能力的发展结果。基于语境的认知控制研究认为,语境对认知控制的影响可能发生在词汇通达之后,也可能发生在通达之前,由此造成非选择性和选择性通达。考虑到认知控制能在词汇通达之前就可以利用语境来限制非目标语言的激活,这种结果既与双语交互激活模型关于词汇激活的非选择性假设不一致,也对认知控制能力的发展主要依赖熟练性的看法提出疑问,新的研究也因此被期待     

方言和外语学习对词汇通达能力的影响

使用图片命名任务和TOT(舌尖现象)任务考察方言和外语学习对词汇通达能力的影响。实验结果发现,在图片命名任务和TOT任务中,单语者的表现均好于粤语-普通话双言者和粤语-普通话-英语三语者,粤语-普通话双言者和粤语-普通话-英语三语者的表现没有显著差异。该结果说明,与外语学习一样,方言也会影响个体的词汇通达能力,但方言和外语学习对词汇通达能力的影响依赖于他们的使用频率。     

The Use of Lexical Prosody for Lexical Access of the Japanese Language

It has been suggested that English listeners do not use lexical prosody for lexical access in word recognition. The present study was designed to examine whether this argument could be generalized to the Japanese language. Two experiments using a cross-modal priming task were conducted. The participants made a lexical decision regarding a visual target following an auditory prime that was either the prosodically congruent or incongruent homophone of the target. In experiment 1, the primes were presented as complete words, and in experiment 2, they were presented as word fragments. In both experiments, the priming effects were observed only in the congruent condition. These results suggest that the prime activated only the representations of prosodically congruent words. We therefore concluded that Japanese listeners use lexical prosody for lexical access of their language and that the role of lexical prosody is determined language-specifically.     

Effects of Lexical Prosody and Word Familiarity on Lexical Access of Spoken Japanese Words

Lexical prosody (e.g., stress and pitch accent) has been shown to constrain lexical activation of spoken words in various languages. In the present study, whether or not the constraint of lexical prosody is affected by word familiarity in lexical access of Japanese words was examined using a cross-modal priming task. The stimuli were pairs of prosodically different homophones (minimal accent pairs). When the targets were more familiar members of minimal accent pairs, the responses were facilitated by prior presentations of primes that were prosodically different homophones of the targets, suggesting that lexical prosody did not constrain lexical activation. In contrast, when less familiar members of minimal accent pairs were used as the targets, the prosodically different homophones did not facilitate the responses to the targets. These results suggest that the constraint of lexical prosody is not so strong but is affected by the factor of word relative familiarity.     

言语产生中的词汇通达理论

首先简单介绍了言语产生中词汇产生研究方法,然后重点介绍了言语产生中的2类重要的词汇通达理论：交互激活理论和独立阶段理论,围绕词汇通达中的3个重要问题：词汇通达的阶段、词汇通达的时间进程、词汇通达阶段是分离独立的还是交互作用的展开论述,阐述了2类理论的相同点和相异点。     

Reading Words and Wirds: Phonology and Lexical Access

Naming latency for printed words is inversely related to their frequency. Four experiments were run to test whether the naming of non-words that are homophones of words (pseudohomophones) is similarly influenced by the frequency of those words. McCann and Besner (1987) failed to find such a frequency effect for pseudohomophones when they were presented in a list of non-words. The present studies show that list structure is critical: A frequency effect occurs for pseudohomophones in a list only of homophones and in a list containing words. The list structure effect was found for three different stimulus lists and suggests that lexical access is strategic. If none of the items in a list has a lexical entry, then pronunciation may be the product of a non-lexical process. If all items have a lexical entry that may be accessed orthographically or phonologically, then pronunciation will be the product of a lexical process.     

双语言语产生中的词汇提取机制

文章主要介绍了双语言语产生中词汇提取机制的两种理论：非特定语言提取的理论和特定语言提取的理论。重点围绕着双语言语产生中词汇提取的一个关键问题,即非目标语言的词汇是否对目标语言的词汇提取形成竞争或者干扰详细地论述了二者的理论分歧,非特定语言提取的理论认为非目标语言的词汇会对目标语言的词汇提取形成竞争,而特定语言提取的理论认为不会形成竞争。文章最后介绍了双语言语产生中词汇提取机制的最新研究趋势,表明了非特定语言提取的理论和特定语言提取的理论逐步走向融和     

Semantic Degradation and Lexical Access in Age-Related Naming Failures

This study investigated the impaired lexical access and semantic degradation hypotheses as two potential explanations of naming failures in normal aging. Naming responses on the Boston Naming Test (BNT) and Action Naming Test (ANT) were analyzed across three test sessions for 39 adults from three age groups (50s, 60s, and 70s). Failures to name before and after cues were classified as either impaired access if failures occurred at an earlier test session followed by successful naming at a later test session or semantic degradation if naming was successful at an earlier test session followed by failures at a later test session. The results indicated that on both the BNT and ANT all age groups produced more naming failures attributed to impaired access than to semantic degradation. However, for object naming, the failures showed significantly more semantic degradation for people in their 70s compared to the younger age groups. By contrast, for action naming, semantic degradation was negligible, possibly masked by a ceiling effect, and the only age-difference result that approached significance indicated that adults in their 70s produced more naming failures attributed to impaired access than adults in their 50s.     

The Influence of Native-language Phonology on Lexical Access: Exemplar-Based Versus Abstract Lexical Entries

This study used medium-term auditory repetition priming to investigate word-recognition processes. Highly fluent Catalan-Spanish bilinguals whose first language was either Catalan or Spanish were tested in a lexical decision task involving Catalan words and nonwords. Spanish-dominant individuals, but not Catalan-dominant individuals, exhibited repetition priming for minimal pairs differing in only one feature that is nondistinctive in Spanish (e.g., /net@/ vs. /nEt@/), thereby indicating that they processed these words as homophones. This finding provides direct evidence both that word recognition uses a language-specific phonological representation and that lexical entries are stored in the mental lexicon as abstract forms.     

Picture Naming and Lexical Access in Italian Children and Adults

Italian-speaking adults and 5- to 6-year-old children were compared in a timed picture-naming task, measuring latency as well as name agreement, in relation to frequency, age of acquisition (AoA), picture complexity, word length, word complexity, semantic category, phonological characteristics, and grammatical gender. Although children were slower and less accurate than adults, correlations of dependent variables with lexical predictors were similar in both groups. However, word complexity had effects on adults that were not seen in children, and grammatical gender had effects on children that were not seen in adults. Adult ratings of AoA had strong effects on both groups, but an objective measure of AoA only affected children. Adult and child reaction times were also differentially affected by semantic category (especially animal names). Results have implications for developmental and cross-linguistic studies of lexical access, in and out of context.     

The Effect of Inhibition of Return on Lexical Access

Abstract—Inhibition of return refers to a bias against returning attention to a location or object that has been recently attended. Recent research has shown that inhibition of return can be found not only in simple detection tasks, but also in tasks requiring relatively simple discrimination judgments. The present experiments examined whether inhibition of return occurs in tasks that require complex discriminations such as lexical decision and categorization tasks. Not only was inhibition of return found in both experiments, but a greater amount of inhibition was found for low-frequency words than for high-frequency words. These findings suggest that inhibition of return not only inhibits returning attention to previously attended locations, but can also affect the processing that is required for lexical access.     

ERP Evidence for the Activation of Syntactic Structure During Comprehension of Lexical Idiom

The present study used event-related potentials to investigate whether the syntactic structure was activated in the comprehension of lexical idioms, and if so, whether it varied as a function of familiarity and semantic transparency. Participants were asked to passively read the “1+2” structural Chinese lexical idioms with each being presented following 3–5 contextual “1+2” (congruent-structure condition) or “2+1” structural Chinese phrases (incongruent-structure condition). The N400 ERP responses showed more positivity in congruent-structure condition relative to incongruent-structure condition in idioms with high familiarity and high semantic transparency, but less positivity in congruent-structure condition in idioms with high familiarity but low semantic transparency, idioms with low familiarity but high semantic transparency, and idioms with low familiarity and low semantic transparency. Our results suggest that syntactic structure, as the unnecessarity of lexical idiomatic words, was nevertheless activated, independent of familiarity and semantic transparency.     

On the Nature of Semantic Constraints on Lexical Access

We present two eye-tracking experiments that investigate lexical frequency and semantic context constraints in spoken-word recognition in German. In both experiments, the pivotal words were pairs of nouns overlapping at onset but varying in lexical frequency. In Experiment 1, German listeners showed an expected frequency bias towards high-frequency competitors (e.g., Blume, 'flower') when instructed to click on low-frequency targets (e.g., Bluse, 'blouse'). In Experiment 2, semantically constraining context increased the availability of appropriate low-frequency target words prior to word onset, but did not influence the availability of semantically inappropriate high-frequency competitors at the same time. Immediately after target word onset, however, the activation of high-frequency competitors was reduced in semantically constraining sentences, but still exceeded that of unrelated distractor words significantly. The results suggest that (1) semantic context acts to downgrade activation of inappropriate competitors rather than to exclude them from competition, and (2) semantic context influences spoken-word recognition, over and above anticipation of upcoming referents.