Lexical Bias Research Articles

Avoiding attachment ambiguities: The role of constituent ordering*1

Three experiments investigated whether speakers use constituent ordering as a mechanism for avoiding ambiguities. In utterances like “Jane showed the letter to Mary to her mother,” alternate orders would avoid the temporary PP-attachment ambiguity (“Jane showed her mother the letter to Mary,” or “Jane showed to her mother the letter to Mary”). A preference judgment experiment confirmed that comprehenders prefer the latter orders for dative utterances when the former order would have contained an ambiguity. Nevertheless, speakers in two on-line production experiments showed no evidence of an ambiguity avoidance strategy. In fact, they were slightly more likely to use the former order when it was ambiguous than when it was not. Speakers’ failure to disambiguate with ordering cannot be explained by the use of other ambiguity mechanisms, like prosody. A prosodic analysis of the responses in Experiment 3 showed that while speakers generally produced prosodic patterns that were consistent with the syntactic structure, these patterns would not strongly disambiguate the PP-attachment ambiguity. We suggest that speakers do not consistently disambiguate local PP-attachment ambiguities of this type, and in particular do not use constituent ordering for this purpose. Instead, constituent ordering is driven by factors like syntactic weight and lexical bias, which may be internal to the production system.

The developing constraints on parsing decisions: The role of lexical-biases and referential scenes in child and adult sentence processing

Two striking contrasts currently exist in the sentence processing literature. First, whereas adult readers rely heavily on lexical information in the generation of syntactic alternatives, adult listeners in world-situated eye-gaze studies appear to allow referential evidence to override strong countervailing lexical biases ( Tanenhaus, Spivey-Knowlton, Eberhard, & Sedivy, 1995). Second, in contrast to adults, children in similar listening studies fail to use this referential information and appear to rely exclusively on verb biases or perhaps syntactically based parsing principles ( Trueswell, Sekerina, Hill, & Logrip, 1999). We explore these contrasts by fully crossing verb bias and referential manipulations in a study using the eye-gaze listening technique with adults (Experiment 1) and five-year-olds (Experiment 2). Results indicate that adults combine lexical and referential information to determine syntactic choice. Children rely exclusively on verb bias in their ultimate interpretation. However, their eye movements reveal an emerging sensitivity to referential constraints. The observed changes in information use over ontogenetic time best support a constraint-based lexicalist account of parsing development, which posits that highly reliable cues to structure, like lexical biases, will emerge earlier during development and more robustly than less reliable cues.

Avoiding attachment ambiguities: The role of constituent ordering

Three experiments investigated whether speakers use constituent ordering as a mechanism for avoiding ambiguities. In utterances like “Jane showed the letter to Mary to her mother,” alternate orders would avoid the temporary PP-attachment ambiguity (“Jane showed her mother the letter to Mary,” or “Jane showed to her mother the letter to Mary”). A preference judgment experiment confirmed that comprehenders prefer the latter orders for dative utterances when the former order would have contained an ambiguity. Nevertheless, speakers in two on-line production experiments showed no evidence of an ambiguity avoidance strategy. In fact, they were slightly more likely to use the former order when it was ambiguous than when it was not. Speakers’ failure to disambiguate with ordering cannot be explained by the use of other ambiguity mechanisms, like prosody. A prosodic analysis of the responses in Experiment 3 showed that while speakers generally produced prosodic patterns that were consistent with the syntactic structure, these patterns would not strongly disambiguate the PP-attachment ambiguity. We suggest that speakers do not consistently disambiguate local PP-attachment ambiguities of this type, and in particular do not use constituent ordering for this purpose. Instead, constituent ordering is driven by factors like syntactic weight and lexical bias, which may be internal to the production system.

Syntactic frame and verb bias in aphasia: Plausibility judgments of undergoer-subject sentences

This study investigates three factors that have been argued to define “canonical form” in sentence comprehension: Syntactic structure, semantic role, and frequency of usage. We first examine the claim that sentences containing unaccusative verbs present difficulties analogous to those of passive sentences. Using a plausibility judgment task, we show that a mixed group of aphasics performed significantly better on unaccusatives than on passives. We then turn to the observation that passives are generally harder than actives for aphasics. We show that this effect is modulated by lexical bias, i.e., the likelihood that a verb appears in a given syntactic structure: Passives of passive-bias verbs were significantly easier than passives of active-bias verbs. More generally, sentences whose structure matches the lexical bias of the main verb are significantly easier than sentences in which structure and lexical bias do not match. These findings suggest that “canonical form” reflects frequency and lexical biases.

The influence of sublexical and lexical representations on the processing of spoken words in English

Previous research suggests that sublexical and lexical representations are involved in spoken word recognition. The current experiment examined when sublexical and lexical representations are used in the processing of real words in English. The same set of words varying in phonotactic probability/neighbourhood density was presented in three different versions of a same‐different matching task: (1) mostly real words as filler items, (2) an equal number of words and nonsense words as filler items and (3) mostly nonsense words as filler items. The results showed that lexical representations were used in version 1 of the same‐different matching task to process the words, whereas sublexical representations were used in version 3 of the same‐different matching task to process the words. Finally, in version 2 of the same‐different matching task individual variation was observed in the form of distinct sublexical and lexical biases. Implications for the processing of spoken words are discussed.

What Should I Do? Behavior Regulation by Language and Paralanguage in Early Childhood

This article explores the functional significance of affective messages for behavior in early childhood. Previous research indicates that children's affective judgments are influenced more by what is said than by how it is said. Of particular interest is the extent to which this tendency toward literal interpretation has real consequences for behavior. The effect of consistent and conflicting affective messages on child behavior was assessed in a social-referencing procedure. What was said had a stronger effect than facial and vocal paralanguage on children's exploration of novel objects. This suggests that the lexical bias evident in children's interpretations reflects a genuine developmental transition in the primary cues on which attributions are based, and these cues have direct consequences for behavior regulation.

Lexical effects on compensation for coarticulation: the ghost of Christmash past

AbstractThe question of when and how bottom‐up input is integrated with top‐down knowledge has been debated extensively within cognition and perception, and particularly within language processing. A long running debate about the architecture of the spoken‐word recognition system has centered on the locus of lexical effects on phonemic processing: does lexical knowledge influence phoneme perception through feedback, or post‐perceptually in a purely feedforward system? Elman and McClelland (1988) reported that lexically restored ambiguous phonemes influenced the perception of the following phoneme, supporting models with feedback from lexical to phonemic representations. Subsequently, several authors have argued that these results can be fully accounted for by diphone transitional probabilities in a feedforward system (Cairns et al., 1995; Pitt & McQueen, 1998). We report results strongly favoring the original lexical feedback explanation: lexical effects were present even when transitional probability biases were opposite to those of lexical biases.

Lexical effects on compensation for coarticulation: the ghost of Christmash past

The question of when and how bottom-up input is integrated with top-down knowledge has been debated extensively within cognition and perception, and particularly within language processing. A long running debate about the architecture of the spoken-word recognition system has centered on the locus of lexical effects on phonemic processing: does lexical knowledge influence phoneme perception through feedback, or post-perceptually in a purely feedforward system? Elman and McClelland (1988) reported that lexically restored ambiguous phonemes influenced the perception of the following phoneme, supporting models with feedback from lexical to phonemic representations. Subsequently, several authors have argued that these results can be fully accounted for by diphone transitional probabilities in a feedforward system ( Cairns et al., 1995; Pitt & McQueen, 1998). We report results strongly favoring the original lexical feedback explanation: lexical effects were present even when transitional probability biases were opposite to those of lexical biases.

Lexical biases in aphasic sentence comprehension: An experimental and corpus linguistic study

Background: This study investigates the role of lexical information in normal and aphasic sentence comprehension. Effects of verb biases in normal comprehension have been well documented in previous studies (e.g., Spivey-Knowlton & Sedivy, 1995; Trueswell, Tanenhaus, & Kello, 1993), but their role in aphasic language processing has largely been ignored (with the exceptions of Menn et al., 1998, and Russo, Peach, & Shapiro, 1998). Aims: The aim of the study is to test the lexical bias hypothesis, i.e., the hypothesis that sentence comprehension is influenced by lexical biases in aphasic listeners, as well as in normals. Method & Procedures: Using a sentence plausibility judgement task, we probe for sensitivity to verb transitivity bias, i.e., the likelihood, as estimated from corpus counts, that a verb will be transitive, rather than intransitive. Five normal controls and eighteen participants with aphasia (six with Broca's aphasia, four with Wernicke's aphasia, two with conduction aphasia, and six with anomic aphasia) are included in the study. Based on the lexical bias hypothesis, we predicted that participants would make more errors in sentences with a mismatch of verb bias and syntactic structure, such as a transitive sentence containing a verb with intransitive bias. Outcomes & Results: Both the group of normal controls and the mixed group of aphasic patients make significantly more errors on sentences in which there is a mismatch between verb bias and syntactic structure, as predicted by the lexical bias hypothesis. Specifically, patients with fluent aphasia types, particularly anomic aphasia, show a sensitivity to verb bias, contrary to earlier findings. Conclusions: These results are consistent with the view that lexical factors, not purely syntactic ones, are to blame for many previously observed patterns in aphasic comprehension. The results are further consistent with the view that many aphasic errors differ not qualitatively but quantitatively from normal comprehension errors.

The acquisition of compound vs. phrasal stress: the role of prosodic constituents.

This paper investigates the acquisition of compound vs. phrasal stress (hót dog vs. hot dóg) in English. This has previously been shown to be acquired quite late, in contrast to recent research showing that infants both perceive and prefer rhythmic patterns in their own language. Subjects (40 children in four groups the averages ages of which are 5;4, 7;2, 9;3 and 11;6 and 10 adults) were shown pairs of pictures representing a compound word and the corresponding phrase. They heard a prerecorded tape with the names of the items, and were asked to indicate which one they heard. In addition to 9 real compounds and corresponding phrases, 9 novel compounds were presented (rédcup = invented type of flower vs. red cúp). A gradual increase in overall correct scores until age twelve was found along with a significant effect of real vs. novel compounds (p < 0.001), and an overwhelming tendency for the younger children to prefer compounds regardless of stress. We conclude that the results are due to the slow development of the ability to use prosodic information to override a strong lexical bias.

Understanding the Constraints on Syntactic Generation: Lexical Bias and Discourse Congruency Effects on Eye Movements

We measured eye movements as people read short stories. The target sentences contained noun/verb homographs (e.g., duck) and were preceded by a biasing context sentence. The homograph in the target sentence was always disambiguated by a case-marked pronoun, e.g., She saw his/him duck. Lexical bias effects (reflecting the relative frequency of the noun and verb forms) were found in the initial fixations on the homograph. In contrast, discourse congruency effects were first observed several words downstream in the probability of a regressive eye movement. Strong discourse congruency effects were also observed in the second pass reading times. We concluded that the lexical bias effects reflect processing difficulty during the initial generation of syntactic structure, while the discourse congruency effects reflect later anomaly detection. Thus, the data challenge syntactic processing models in which all relevant and available constraints are brought to bear uniformly and simultaneously.

The source of a lexical bias in the Verbal Transformation Effect

Comparisons of responses to words and pseudowords Žgure prominently in our efforts to understand how spoken words are recognised. This was apparent in many of the presentations at the SWAP meeting (e.g., Fowler & Brancazio, 2000; Frauenfelder & Content, 2000). Differences in responding to words and pseudowords are assumed to demonstrate the inuence of lexical memory in the task at hand, from which we infer the structural and functional characteristics of the processing system. Each task we use in this enterprise is often best suited to address a subset of theoretical issues. For example, phoneme detection and identiŽcation are used most often to examine issues about prelexical processing, such as whether there is lexical feedback and what the structure of phonetic categories is. Lexical decision and other word-based tasks tend to be used to explore the properties of lexical memory. So when a lexical effect is identiŽed in an unfamiliar task, it is important to determine what we can learn about word processing from it. This has been the goal of our work into the origin of word-pseudoword differences in the Verbal Transformation Effect (VTE).

Lexical biases are useful

I present two criticisms: (1) Merge is a model of performance in speech perception experiments but not an ecologically valid model integrating both word recognition and speech perception, and (2) Merge's implementation of the variability of lexical effects must be made more precise or the model is indistinguishable from its alternatives.

Recency and lexical preferences in Spanish.

One experiment provided evidence in support of Gibson, Pearlmutter, Canseco-Gonzalez, and Hickok's (1996) claim that a recency preference applies to Spanish relative clause attachments, contrary to the claim made by Cuetos and Mitchell (1988). Spanish speakers read stimuli involving either two or three potential attachment sites in which the same lexical content of the two-site conditions appeared in a different structural configuration in the three-site conditions. High attachment was easier than low attachment when only two sites were present, but low attachment was preferred over high attachment, which was in turn preferred over middle attachment, when three sites were present. The experiment replicated earlier results and showed that (1) attachment preferences are determined in part by a preference to attach recently/low, and (2) lexical biases are insufficient to explain attachment preferences.

Top-down influences on phonetic processing as a function of aging

A phonetic identification experiment was conducted in order to assess the extent to which lexical information influenced low-level speech perception processes in three different age groups of subjects (20–29 years, 60–69 years, and 70–79 years). Subjects were required to label syllable-initial, bilabial stop consonants which varied in voice onset time (VOT) as either /b/ or /p/. Following Ganong [J.E.P.: H.P.P. 6, 110–125 (1980)], a pair of VOT continua with opposing lexical biases was created: one continuum ranged from the word /buk/ to the nonword /puk/ while the other continuum ranged from the nonword /bu t/ to the word /put/. VOT values for the syllable-initial /b/–/p/ phoneme were the same at corresponding steps in the two continua. Results indicated a significant influence of lexical status on phonetic identifications for all age groups. Furthermore, a significant difference in the location of the phonetic crossover boundary was observed, with older subjects showing longer VOT values than young subjects. While findings suggest that top-down processing is still a prevalent part of phonetic perception in elderly individuals, it appears that aging may bring changes in the values of the acoustic parameters used to define phonetic categories. [Work supported by McGill University.]

Transitional probability, not lexical knowledge, influences compensation

The perceptual system compensates for fricative–stop coarticulation. Ambiguous stops between /t/ and /k/ tend to be heard as /k/ after /s/-final words and as /t/ after /S/-final words, even when the word–final fricatives are replaced with an ambiguous phoneme [J. L. Elman and J. L. McClelland, J. Mem. Lang. 27, 143–165 (1988)]. Biases in the transitional probabilities (TPs) of /s/ and /S/ in the words in that study mean that the effect following ambiguous fricatives could be due either to lexical involvement in prelexical compensation for coarticulation or to a sensitivity to TP at the prelexical level. In three experiments, listeners categorized nonword–final fricatives with TP biases, word–final fricatives where TPs were controlled, and word–initial stops following both fricative contexts. Categorization of ambiguous fricatives and stops was influenced by the TPs of the fricatives in the nonwords. But although there was a lexical bias in the categorization of the word–final ambiguous fricatives, there was no lexical influence on the categorization of the stops. These results suggest that the lexicon does not influence prelexical processing, and that the prelexical level is sensitive to TP. [Work supported by NIDCD, the Human Frontiers Science Project, and the Max Planck Society.]

Resolving Syntactic Category Ambiguities in Discourse Context: Probabilistic and Discourse Constraints

Three experiments investigated the resolution of syntactic category ambiguities in a self-paced reading paradigm, using moving window presentation. Sentences containing noun/verb homographs likeduck(e.g.,She saw her duck and chickens/stumble near the barn.) were embedded in short stories. The first experiment reports the collection of several types of norms to measure the degree to which each homograph was biased toward use as a noun or use as a verb, and the strength of the discourse bias for each item. The goal was to determine how probabilistic lexical bias and discourse bias influence the generation and selection of syntactic structure in the ambiguous region, and this was investigated by varying the length of the ambiguous region in Experiments 2–4. The data suggest that only probabilistic lexical bias influenced the generation of structure, while both lexical bias and discourse bias guided selection of the appropriate structure.

Phonological Errors in Aphasic Naming: Comprehension, Monitoring and Lexicality

This paper investigates the production of phonological errors in aphasic naming, examining the relationship between these errors and deficits in comprehension. The predictions of Dell and O'Seaghda's (1991) computational model of speech production were tested by lesioning. The set of lesioned models demonstrated a strong correlation between proportion of phonologically related errors in naming and comprehension accuracy. In contrast, no correlation between proportion of phonological errors in naming and comprehension accuracy was found for a group of fifteen aphasics. This paper also examines monitoring behaviours, such as the presence of self-corrections and interrupted responses, again finding no relationship with auditory comprehension. There was also no evidence for a lexical bias in the phonological errors for these aphasic subjects. Phonologically related errors were argued to be words only by chance; this was supported by the fact that they occurred more frequently on short words than long words, and that the proportion of real word errors was not significantly different to that observed in a pseudocorpus of errors. We conclude that a production-based monitor would be compatible with these results but that any comprehension-based monitor is not reliably employed by some or all of the aphasic subjects examined here. Additionally, these data are incompatible with models of language processing where speech input and output share the same processing components.

Effects of lexical status on native and non-native English-speaking adults’ vowel perception

Monolingual, English-speaking adults and native Chinese (Mandarin) speakers who learned English as a second language heard stimuli from two ‘‘native,’’ synthetic continua, in which the vowels ranged from English /i/ to /i/ in the context /b–b/ or /b–p/. Thus the end points of the first continuum constituted an English word and a nonword (‘‘bib’’ vs *‘‘beeb’’); the reverse held for the second continuum (*‘‘bip’’ vs ‘‘beep’’). These same subjects also heard stimuli from two ‘‘foreign’’ continua, where the vowels ranged from English /i/ to a foreign (non-English) vowel /Y/ in the contexts described above. Thus the end points of the first continuum corresponded to a word and a nonword (‘‘bib’’ vs *‘‘bYb’’); both end points of the second continuum corresponded to nonwords (*‘‘bip’’ vs *‘‘bYp’’). After training on end points, subjects’ identifications of the nine stimuli of a given continuum were examined to assess whether: the Chinese speakers, like native English speakers, exhibit a ‘‘lexical bias’’ effect for English vowels (from the native continua); vowel categories not bounded by another native/English vowel (as in the foreign continua) expand outward or become better defined with increased exposure to English and/or lexical status. [Work supported by NIH.]

Effects of lexical status on children’s and adults’ perception of native and non-native vowels

Monolingual, English-speaking 5-year-olds, 9-year-olds, and adults heard stimuli from two ‘‘native,’’ synthetic continua, in which the vowels ranged from English /i/ to /i/ in the context /b—b/ or /b—p/. Thus the endpoints of the first continuum constituted a word and a nonword (‘‘bib’’ vs *‘‘beeb’’); the reverse held for the second continuum (*‘‘bip’’ vs ‘‘beep’’). Other subjects heard stimuli from two ‘‘foreign’’ continua, where the vowels ranged from English /i/ to a foreign vowel /y/ in the contexts described above. Thus the endpoints of the first continuum corresponded to a word and a nonword (‘‘bib’’ vs *‘‘bYb’’); both endpoints of the second continuum corresponded to nonwords (*‘‘bip’’ vs *‘‘bYp’’). After training on endpoints, subjects’ identifications of the nine stimuli of a given continuum were examined to assess whether: children, like adults, exhibit a ‘‘lexical bias’’ effect for familiar vowels (from the ‘‘native’’ continua); vowel categories not bounded by another native vowel (as in the ‘‘foreign’’ continua) expand outward or become better defined with increasing age and/or lexical status.