Lexical Segmentation Research Articles

Classifying coherent versus nonsense speech perception from EEG using linguistic speech features

When a person listens to natural speech, the relation between features of the speech signal and the corresponding evoked electroencephalogram (EEG) is indicative of neural processing of the speech signal. Using linguistic representations of speech, we investigate the differences in neural processing between speech in a native and foreign language that is not understood. We conducted experiments using three stimuli: a comprehensible language, an incomprehensible language, and randomly shuffled words from a comprehensible language, while recording the EEG signal of native Dutch-speaking participants. We modeled the neural tracking of linguistic features of the speech signals using a deep-learning model in a match-mismatch task that relates EEG signals to speech, while accounting for lexical segmentation features reflecting acoustic processing. The deep learning model effectively classifies coherent versus nonsense languages. We also observed significant differences in tracking patterns between comprehensible and incomprehensible speech stimuli within the same language. It demonstrates the potential of deep learning frameworks in measuring speech understanding objectively.

Research on Product Conceptual Design Scheme Configurations from a Designer–User Conflict Perspective

In the context of rapid product iteration, design conflicts arise from discrepancies in designers’ understanding of user needs, influenced by subjective preferences, behavioural stances, and other factors. This paper proposes a product conceptual design approach based on the design conflict perspective. First, user comments and design documents are collected. Natural language processing (NLP) methods, including cleaning, filtering, lexical segmentation, feature clustering, and sentiment analysis, are employed to identify design themes. The intuitionistic fuzzy sets (IFSs) and term frequency–inverse document frequency (TF-IDF) algorithms are then utilized to obtain evaluation matrices for the products from both users and designers. Subsequently, design conflicts between users and designers are calculated, and an optimal configuration for product conceptual design is determined through regression analysis and planning methods. Finally, the proposed method is validated using a mobile phone as a product example, and suggestions for product improvement are presented. The results indicate that considering design conflicts as a factor in product design and synthesizing designer and user product concepts enhance the accuracy and reliability of product conceptual design generation. The findings of this study offer new insights into the conceptual design configuration for product iteration.

The Interplay of Syllable Structure and Consonant Sonority in L2 Speech Segmentation

The present study investigated whether L1 and L2 Spanish speakers show sensitivity to matching/mismatching syllable structure and consonant sonority in lexical segmentation in Spanish. A total of 81 English–Spanish learners and 72 Spanish–English learners completed a fragment-monitoring task. They listened to lists of Spanish words as they saw a CV or CVC syllable (e.g., “pa” or “pal”) and pressed a button when the word began with the syllable shown on the screen. The task manipulated syllable structure (CV or CVC) and consonant sonority (fricative, nasal, or liquid) of target syllables and carrier words. Target syllables either matched or did not match the structure of the first syllable in target carrier words (e.g., “pa—pa.lo.ma”; “pa—pal.me.ra”). The results showed that consonant sonority modulated sensitivity to syllable structure in both groups of participants. Spanish–English learners responded faster to matching syllable structure in words that had a fricative or a nasal as the second consonant, and English–Spanish learners responded faster only with a fricative consonant. Higher L2 Spanish proficiency correlated with faster target-syllable identification, but sensitivity to matching/mismatching structure did not vary as a function of proficiency. The study highlights the influence of phonetic factors in the development of L2 lexical segmentation routines.

Adequacy, fluency and cognitive processes: evidence from translating English news articles into Arabic

ABSTRACT This article reports on an empirical study on cognitive processes [i.e. (dis) fluency, temporal management of the translation process, and revision] and their relationship with fluency and adequacy. Based on Jakobsen and Schou’s (1999) computational model of human translation, the author utilizes keylogging data retrieved from a Translog-II-based task to investigate the cognitive processes behind Arab translator trainees’ translations of an English news article into Arabic and unfold whether these processes predict fluency and adequacy (N = 32). Results demonstrate trainees’ segmentation of a news article into frequent short lexical segments manifested in the mean translation units (M = 141.34, SD = 44.37), reliance on online revision over end revision, and temporal delay in the drafting stage. Multiple-line regression analysis shows that translation quality is predicted by revision time, orientation time, and translation speed in addition to a positive association between translation units and translation adequacy. According to the findings, translation trainees should be equipped with ST comprehension, and TT production and revision skills.

Strategies for interactive listening in modern foreign language learning

In two-way interactive listening, listeners are expected to use interactional skills or strategies to understand meaning, recognize interlocutors’ intentions, make responses, and establish common ground. However, strategy use can be affected by learner differences and affective factors. The present study investigated the effects of group difference (students majoring in five modern foreign languages) and an affective factor (self-perceived communication confidence) on Chinese native speakers’ strategies for interactive listening. A questionnaire survey was conducted to explore the listening strategies used by 445 college students, and semi-structured interviews were conducted with 20 students to uncover listening difficulties in their language majors. A two-way between-groups MANOVA test revealed statistically significant main effects for language majors and communication confidence on overall and on certain types of strategy for interactive listening. Spanish language majors were found to use strategies more frequently than English, French, German, or Japanese major students. Students in all five language majors with higher self-perceived communication confidence used inference-making, phonological and nonverbal strategies more frequently than low-confidence students. Certain listening difficulties were found to be language specific. In general, vocabulary size as well as lexical segmentation were reported as main listening difficulties, followed by morphological and syntactic variation, and fast speech rate.

How much does prosody help word segmentation? A simulation study on infant-directed speech

Infants come to learn several hundreds of word forms by two years of age, and it is possible this involves carving these forms out from continuous speech. It has been proposed that the task is facilitated by the presence of prosodic boundaries. We revisit this claim by running computational models of word segmentation, with and without prosodic information, on a corpus of infant-directed speech. We use five cognitively-based algorithms, which vary in whether they employ a sub-lexical or a lexical segmentation strategy and whether they are simple heuristics or embody an ideal learner. Results show that providing expert-annotated prosodic breaks does not uniformly help all segmentation models. The sub-lexical algorithms, which perform more poorly, benefit most, while the lexical ones show a very small gain. Moreover, when prosodic information is derived automatically from the acoustic cues infants are known to be sensitive to, errors in the detection of the boundaries lead to smaller positive effects, and even negative ones for some algorithms. This shows that even though infants could potentially use prosodic breaks, it does not necessarily follow that they should incorporate prosody into their segmentation strategies, when confronted with realistic signals.

Exploring the relationships between L2 vocabulary knowledge, lexical segmentation, and L2 listening comprehension

The capacity to perceive and meaningfully process foreign or second language (L2) words from the aural modality is a fundamentally important aspect of successful L2 listening. Despite this, the relationships between L2 listening and learners’ capacity to process aural input at the lexical level has received relatively little research focus. This study explores the relationships between measures of aural vocabulary, lexical segmentation and two measures of L2 listening comprehension (i.e., TOEIC & Eiken Pre-2) among a cohort of 130 tertiary level English as a foreign language (EFL) Japanese learners. Multiple regression modelling indicated that in combination, aural knowledge of vocabulary at the first 1,000-word level and lexical segmentation ability could predict 34% and 38% of total variance observed in TOEIC listening and Eiken Pre-2 listening scores respectively. The findings are used to provide some preliminary recommendations for building the capacity of EFL learners to process aural input at the lexical level.

On Trainee Interpreters’ English Listening Comprehension Barriers: The Case of Local University Undergraduate Translation Majors in China

Listening is the first step of interpretation, and it determines whether an interpreter can successfully grasp information received in the source utterance and fulfill the interpreting task. Being an important part of interpreting process, it is also the bottleneck of many trainee interpreters. For the majority of Chinese students majoring in English or translation, it is a problem needing to be addressed urgently. The present study aims at finding the trainees’ barriers for listening comprehension in English to Chinese interpretation. A questionnaire study was conducted on 164 trainee interpreters. The results and the following informal interviews indicate that some factors exert negative effects on the trainees’ listening and information receiving, including linguistic barriers and non-linguistic ones. The linguistic barriers revealed from the data are perception of sounds, lexical segmentation, sentence structure recognition etc. The Non-linguistic barriers include lack of background knowledge, logic organization of the information, speaker’s accent, unfamiliar vocabularies, unbalanced effort coordination and psychological stress, etc. Based on the theoretical and pedagogic analysis of students’ problems, some teaching and training strategies are suggested to cope with the problems.

GRAMMAR OF TIME IN UDMURT SOCREALISTIC POETRY: GAI SABITOV’s CREATIVITY

This article discusses the individual structural-semantic, grammatical and stylistic peculiarities of the poetic texts of Gui Sabitov, who is one of the brightest representatives of the Udmurt socialist realism. Udmurt socrealistic idiostyles as a national-specific variant of this artistic method has never been the object of interpretation. The relevance of the study is dictated by the insufficiency of the linguistic and poetic approach in considering the phenomenon of development of Udmurt literature, the lack of scientific works focused both on the grammar of poetry and on the temporality category of the poetic whole, its artistic contexts. The article analyzes the multi-level relationship of the grammatical sphere and the artistic semantics of G. Sabitov’s texts, shows the interconversion of form and text content, linguistic elements, the ideological plan of expression and the individual author’s worldview. The novelty is the selection of the main grammatical and lexical segments of the lyrical translation of temporality, the identification of scenarios of individual author semantization of the present tense, the study of Gui Sabitov’s poetic corpus from the point of view of linguistic realities that are projected onto the Udmurt Soviet literature in general.

Objective Intelligibility Assessment by Automated Segmental and Suprasegmental Listening Error Analysis.

Purpose Subjective speech intelligibility assessment is often preferred over more objective approaches that rely on transcript scoring. This is, in part, because of the intensive manual labor associated with extracting objective metrics from transcribed speech. In this study, we propose an automated approach for scoring transcripts that provides a holistic and objective representation of intelligibility degradation stemming from both segmental and suprasegmental contributions, and that corresponds with human perception. Method Phrases produced by 73 speakers with dysarthria were orthographically transcribed by 819 listeners via Mechanical Turk, resulting in 63,840 phrase transcriptions. A protocol was developed to filter the transcripts, which were then automatically analyzed using novel algorithms developed for measuring phoneme and lexical segmentation errors. The results were compared with manual labels on a randomly selected sample set of 40 transcribed phrases to assess validity. A linear regression analysis was conducted to examine how well the automated metrics predict a perceptual rating of severity and word accuracy. Results On the sample set, the automated metrics achieved 0.90 correlation coefficients with manual labels on measuring phoneme errors, and 100% accuracy on identifying and coding lexical segmentation errors. Linear regression models found that the estimated metrics could predict a significant portion of the variance in perceptual severity and word accuracy. Conclusions The results show the promising development of an objective speech intelligibility assessment that identifies intelligibility degradation on multiple levels of analysis.

Rapid Transformation from Auditory to Linguistic Representations of Continuous Speech

During speech perception, a central task of the auditory cortex is to analyze complex acoustic patterns to allow detection of the words that encode a linguistic message [1]. It is generally thought that this process includes at least one intermediate, phonetic, level of representations [2-6], localized bilaterally in the superior temporal lobe [7-9]. Phonetic representations reflect a transition from acoustic to linguistic information, classifying acoustic patterns into linguistically meaningful units, which can serve as input to mechanisms that access abstract word representations [10, 11]. While recent research has identified neural signals arising from successful recognition of individual words in continuous speech [12-15], no explicit neurophysiological signal has been found demonstrating the transition from acoustic and/or phonetic to symbolic, lexical representations. Here, we report a response reflecting the incremental integration of phonetic information for word identification, dominantly localized to the left temporal lobe. The short response latency, approximately 114ms relative to phoneme onset, suggests that phonetic information is used for lexical processing as soon as it becomes available. Responses also tracked word boundaries, confirming previous reports of immediate lexical segmentation [16, 17]. These new results were further investigated using a cocktail-party paradigm [18, 19] in which participants listened to a mix oftwo talkers, attending to one and ignoring the other. Analysis indicates neural lexical processing of only the attended, but not the unattended, speech stream. Thus, while responses to acoustic features reflect attention through selective amplification of attended speech, responses consistent with a lexical processing model reveal categorically selective processing.

Deep Learning of Inflection and the Cell-Filling Problem

Machine learning offers two basic strategies for morphology induction: lexical segmentation and surface word relation. The first approach assumes that words can be segmented into morphemes. Inferring a novel inflected form requires identification of morphemic constituents and a strategy for their recombination. The second approach dispenses with segmentation: lexical representations form part of a network of associatively related inflected forms. Production of a novel form consists in filling in one empty node in the network. Here, we present the results of a task of word inflection by a recurrent LSTM network that learns to fill in paradigm cells of incomplete verb paradigms. Although the task does not require morpheme segmentation, we show that accuracy in carrying out the inflection task is a function of the model’s sensitivity to paradigm distribution and morphological structure.

The role of character positional frequency on Chinese word learning during natural reading.

Readers’ eye movements were recorded to examine the role of character positional frequency on Chinese lexical acquisition during reading and its possible modulation by word spacing. In Experiment 1, three types of pseudowords were constructed based on each character’s positional frequency, providing congruent, incongruent, and no positional word segmentation information. Each pseudoword was embedded into two sets of sentences, for the learning and the test phases. In the learning phase, half the participants read sentences in word-spaced format, and half in unspaced format. In the test phase, all participants read sentences in unspaced format. The results showed an inhibitory effect of character positional frequency upon the efficiency of word learning when processing incongruent pseudowords both in the learning and test phase, and also showed facilitatory effect of word spacing in the learning phase, but not at test. Most importantly, these two characteristics exerted independent influences on word segmentation. In Experiment 2, three analogous types of pseudowords were created whilst controlling for orthographic neighborhood size. The results of the two experiments were consistent, except that the effect of character positional frequency was absent in the test phase in Experiment 2. We argue that the positional frequency of a word’s constituent characters may influence the character-to-word assignment in a process that likely incorporates both lexical segmentation and identification.

Lexical Segmentation in Artificial Word Learning: The Effects of Converging Sublexical Cues.

This study examines how French listeners segment and learn new words of artificial languages varying in the presence of different combinations of sublexical segmentation cues. The first experiment investigated the contribution of three different types of sublexical cues (acoustic-phonetic, phonological and prosodic cues) to word learning. The second experiment explored how participants specifically exploited sublexical prosodic cues. Whereas complementary cues signaling word-initial and word-final boundaries had synergistic effects on word learning in the first experiment, the two manipulated prosodic cues redundantly signaling word-final boundaries in the second experiment were rank-ordered with final pitch variations being more weighted than final lengthening. These results are discussed in light of the notions of cue type, cue position and cue efficiency.

Minding the gaps: literacy enhances lexical segmentation in children learning to read.

Can emergent literacy impact the size of the linguistic units children attend to? We examined children's ability to segment multiword sequences before and after they learned to read, in order to disentangle the effect of literacy and age on segmentation. We found that early readers were better at segmenting multiword units (after controlling for age, cognitive, and linguistic variables), and that improvement in literacy skills between the two sessions predicted improvement in segmentation abilities. Together, these findings suggest that literacy acquisition, rather than age, enhanced segmentation. We discuss implications for models of language learning.

The role of stress and word size in Spanish speech segmentation.

In English, the predominance of stressed syllables as word onsets aids lexical segmentation in degraded listening conditions. Yet it is unlikely that these findings would readily transfer to languages with differing rhythmic structure. In the current study, the authors seek to examine whether listeners exploit both common word size (syllable number) and stress cues to aid lexical segmentation in Spanish. Forty-seven Spanish-speaking listeners transcribed two-word Spanish phrases in noise. As predicted by the statistical probabilities of Spanish, error analysis revealed that listeners preferred two- and three-syllable words with penultimate stress in their attempts to parse the degraded speech signal. These findings provide insight into the importance of stress in tandem with word size in the segmentation of Spanish words and suggest testable hypotheses for cross-linguistic studies that examine the effects of degraded acoustic cues on lexical segmentation.

Experiments on Auditory-Visual Perception of Sentences by Users of Unilateral, Bimodal, and Bilateral Cochlear Implants.

Five experiments probed auditory-visual (AV) understanding of sentences by users of cochlear implants (CIs). Sentence material was presented in auditory (A), visual (V), and AV test conditions to listeners with normal hearing and CI users. (a) Most CI users report that most of the time, they have access to both A and V information when listening to speech. (b) CI users did not achieve better scores on a task of speechreading than did listeners with normal hearing. (c) Sentences that are easy to speechread provided 12 percentage points more gain to speech understanding than did sentences that were difficult. (d) Ease of speechreading for sentences is related to phrase familiarity. (e) Users of bimodal CIs benefit from low-frequency acoustic hearing even when V cues are available, and a second CI adds to the benefit of a single CI when V cues are available. (f) V information facilitates lexical segmentation by improving the recognition of the number of syllables produced and the relative strength of these syllables. Our data are consistent with the view that V information improves CI users' ability to identify syllables in the acoustic stream and to recognize their relative juxtaposed strengths. Enhanced syllable resolution allows better identification of word onsets, which, when combined with place-of-articulation information from visible consonants, improves lexical access.

Using Predictability for Lexical Segmentation.

This study investigates a strategy based on predictability of consecutive sub-lexical units in learning to segment a continuous speech stream into lexical units using computational modeling and simulations. Lexical segmentation is one of the early challenges during language acquisition, and it has been studied extensively through psycholinguistic experiments as well as computational methods. However, despite strong empirical evidence, the explicit use of predictability of basic sub-lexical units in models of segmentation is underexplored. This paper presents an incremental computational model of lexical segmentation for exploring the usefulness of predictability for lexical segmentation. We show that the predictability cue is a strong cue for segmentation. Contrary to earlier reports in the literature, the strategy yields state-of-the-art segmentation performance with an incremental computational model that uses only this particular cue in a cognitively plausible setting. The paper also reports an in-depth analysis of the model, investigating the conditions affecting the usefulness of the strategy.

Reading between the words: The effect of literacy on second language lexical segmentation

ABSTRACTThere is evidence that the ability to segment an utterance into words improves with literacy, yet previous research makes it difficult to disentangle the effect of literacy from that of age or cognitive abilities. We tested the hypothesis that literacy increases lexical segmentation in a second language in a unique sample of adult illiterates learning to read in their second language, controlling for cognitive abilities and using a task that taps language processing rather than only metalinguistic awareness. Participants’ segmentation was correlated with first language reading at the beginning of an intensive literacy course for illiterate adults. At the end of the course, those learning to read for the first time benefited more in terms of their segmentation abilities. We discuss implications for models of second language learning.

Cognitive–Perceptual Examination of Remediation Approaches to Hypokinetic Dysarthria

To determine how increased vocal loudness and reduced speech rate affect listeners' cognitive-perceptual processing of hypokinetic dysarthric speech associated with Parkinson's disease. Fifty-one healthy listener participants completed a speech perception experiment. Listeners repeated phrases produced by 5 individuals with dysarthria across habitual, loud, and slow speaking modes. Listeners were allocated to habitual ( n = 17), loud ( n = 17), or slow ( n = 17) experimental conditions. Transcripts derived from the phrase repetition task were coded for overall accuracy (i.e., intelligibility), and perceptual error analyses examined how these conditions affected listeners' phonemic mapping (i.e., syllable resemblance) and lexical segmentation (i.e., lexical boundary error analysis). Both speech conditions provided obvious perceptual benefits to listeners. Overall, transcript accuracy was highest in the slow condition. In the loud condition, however, improvement was evidenced across the experiment. An error analysis suggested that listeners in the loud condition prioritized acoustic-phonetic cues in their attempts to resolve the degraded signal, whereas those in the slow condition appeared to preferentially weight lexical stress cues. Increased loudness and reduced rate exhibited differential effects on listeners' perceptual processing of dysarthric speech. The current study highlights the insights that may be gained from a cognitive-perceptual approach.