Publications

Abstract

The author contests several claims of La Heij, Starreveld, and Kuipers (this issue) concerning the modelling of spoken word planning. The claims are about the relevance of error findings, the interaction between semantic and phonological factors, the explanation of word-word findings, the semantic relatedness paradox, and production rules.

Abstract

Although humans are the ultimate “natural language generators”, the area of psycholinguistic modeling has been somewhat underrepresented in recent approaches to Natural Language Generation in computer science. To draw attention to the area and illustrate its potential relevance to Natural Language Generation, I provide an overview of recent work on psycholinguistic modeling of language production together with some key empirical findings, state-of-the-art experimental techniques, and their historical roots. The techniques include analyses of speech-error corpora, chronometric analyses, eyetracking, and neuroimaging.
The overview is built around the issue of cognitive control in natural language generation, concentrating on the production of single words, which is an essential ingredient of the generation of larger utterances. Most of the work exploited the fact that human speakers are good but not perfect at resisting temptation, which has provided some critical clues about the nature of the underlying system.

Abstract

Four experiments examined crossmodal versions of the Stroop task in order (1) to look for Stroop asymmetries in color naming, spoken-word naming, and written-word naming and to evaluate the time course of these asymmetries, and (2) to compare these findings to current models of the Stroop effect. Participants named color patches while ignoring spoken color words presented with an onset varying from 300 msec before to 300 msec after the onset of the color (Experiment 1), or they named the spoken words and ignored the colors (Experiment 2). A secondary visual detection task assured that the participants looked at the colors in both tasks. Spoken color words yielded Stroop effects in color naming, but colors did not yield an effect in spoken-word naming at any stimulus onset asynchrony. This asymmetry in effects was obtained with equivalent color- and spoken-word-naming latencies. Written color words yielded a Stroop effect in naming spoken words (Experiment 3), and spoken color words yielded an effect in naming written words (Experiment 4). These results were interpreted as most consistent with an architectural account of the color-word Stroop asymmetry, in contrast with discriminability and pathway strength accounts.

Abstract

In a recent series of papers, Caramazza and Miozzo [Caramazza, A., 1997. How many levels of processing are there in lexical access? Cognitive Neuropsychology 14, 177-208; Caramazza, A., Miozzo, M., 1997. The relation between syntactic and phonological knowledge in lexical access: evidence from the 'tip-of-the-tongue' phenomenon. Cognition 64, 309-343; Miozzo, M., Caramazza, A., 1997. On knowing the auxiliary of a verb that cannot be named: evidence for the independence of grammatical and phonological aspects of lexical knowledge. Journal of Cognitive Neuropsychology 9, 160-166] argued against the lemma/lexeme distinction made in many models of lexical access in speaking, including our network model [Roelofs, A., 1992. A spreading-activation theory of lemma retrieval in speaking. Cognition 42, 107-142; Levelt, W.J.M., Roelofs, A., Meyer, A.S., 1998. A theory of lexical access in speech production. Behavioral and Brain Sciences, (in press)]. Their case was based on the observations that grammatical class deficits of brain-damaged patients and semantic errors may be restricted to either spoken or written forms and that the grammatical gender of a word and information about its form can be independently available in tip-of-the-tongue stales (TOTs). In this paper, we argue that though our model is about speaking, not taking position on writing, extensions to writing are possible that are compatible with the evidence from aphasia and speech errors. Furthermore, our model does not predict a dependency between gender and form retrieval in TOTs. Finally, we argue that Caramazza and Miozzo have not accounted for important parts of the evidence motivating the lemma/lexeme distinction, such as word frequency effects in homophone production, the strict ordering of gender and pho neme access in LRP data, and the chronometric and speech error evidence for the production of complex morphology.

Abstract

B. Rapp and M. Goldrick (2000) claimed that the lexical and mixed error biases in picture naming by
aphasic and nonaphasic speakers argue against models that assume a feedforward-only relationship
between lexical items and their sounds in spoken word production. The author contests this claim by
showing that a feedforward-only model like WEAVER ++ (W. J. M. Levelt, A. Roelofs, & A. S. Meyer,
1999b) exhibits the error biases in word planning and self-monitoring. Furthermore, it is argued that
extant feedback accounts of the error biases and relevant chronometric effects are incompatible.
WEAVER ++ simulations with self-monitoring revealed that this model accounts for the chronometric
data, the error biases, and the influence of the impairment locus in aphasic speakers.

Abstract

Simple name-retrieval models of spoken word planning (Bloem & La Heij, 2003; Starreveld & La Heij, 1996) maintain (1) that there are two levels in word planning, a conceptual and a lexical phonological level, and (2) that planning a word in both object naming and oral reading involves the selection of a lexical phonological representation. Here, the name retrieval models are compared to more complex models with respect to their ability to account for relevant data. It appears that the name retrieval models cannot easily account for several relevant findings, including some speech error biases, types of morpheme errors, and context effects on the latencies of responding to pictures and words. New analyses of the latency distributions in previous studies also pose a challenge. More complex models account for all these findings. It is concluded that the name retrieval models are too simple and that the greater complexity of the other models is warranted

Abstract

WEAVER++ has no backward links in its form-production network and yet is able to explain the lexical
and mixed error biases and the mixed distractor latency effect. This refutes the claim of B. Rapp and M.
Goldrick (2000) that these findings specifically support production-internal feedback. Whether their restricted interaction account model can also provide a unified account of the error biases and latency effect remains to be shown.

Abstract

The trigger for shifting gaze between stimuli requiring vocal and manual responses was examined. Participants were presented with picture–word stimuli and left- or right-pointing arrows. They vocally named the picture (Experiment 1), read the word (Experiment 2), or categorized the word (Experiment 3) and shifted their gaze to the arrow to manually indicate its direction. The experiments showed that the temporal coordination of vocal responding and gaze shifting depends on the vocal task and, to a lesser extent, on the type of relationship between picture and word. There was a close temporal link between gaze shifting and manual responding, suggesting that the gaze shifts indexed shifts of attention between the vocal and manual tasks. Computer simulations showed that a simple extension of WEAVER++ [Roelofs, A. (1992). A spreading-activation theory of lemma retrieval in speaking. Cognition, 42, 107–142.; Roelofs, A. (2003). Goal-referenced selection of verbal action: modeling attentional control in the Stroop task. Psychological Review, 110, 88–125.] with assumptions about attentional control in the coordination of vocal responding, gaze shifting, and manual responding quantitatively accounts for the key findings.

Abstract

In 4 chronometric experiments, influences of spoken word planning on speech recognition were examined. Participants were shown pictures while hearing a tone or a spoken word presented shortly after picture onset. When a spoken word was presented, participants indicated whether it contained a prespecified phoneme. When the tone was presented, they indicated whether the picture name contained the phoneme (Experiment 1) or they named the picture (Experiment 2). Phoneme monitoring latencies for the spoken words were shorter when the picture name contained the prespecified phoneme compared with when it did not. Priming of phoneme monitoring was also obtained when the phoneme was part of spoken nonwords (Experiment 3). However, no priming of phoneme monitoring was obtained when the pictures required no response in the experiment, regardless of monitoring latency (Experiment 4). These results provide evidence that an internal phonological pathway runs from spoken word planning to speech recognition and that active phonological encoding is a precondition for engaging the pathway. (PsycINFO Database Record (c) 2007 APA, all rights reserved)

Abstract

The authors discuss the issue of how visual orienting, selective stimulus processing, and vocal response planning are related in Stroop-like tasks. The evidence suggests that visual orienting is dependent on both visual processing and verbal response planning. They also discuss the issue of selective perceptual processing in Stroop-like tasks. The evidence suggests that space-based and object-based attention lead to a Trojan horse effect in the classic Stroop task, which can be moderated by increasing the spatial distance between colour and word and by making colour and word part of different objects. Reducing the presentation duration of the colour-word stimulus or the duration of either the colour or word dimension reduces Stroop interference. This paradoxical finding was correctly simulated by the WEAVER++ model. Finally, the authors discuss evidence on the neural correlates of executive attention, in particular, the ACC. The evidence suggests that the ACC plays a role in regulation itself rather than only signalling the need for regulation.

Abstract

According to most models of speech production, the planning of spoken words involves the independent retrieval of segments and metrical frames followed by segment-to-frame association. In some models, the metrical frame includes a specification of the number and ordering of consonants and vowels, but in the word-form encoding by activation and verification (WEAVER) model (A. Roelofs, 1997), the frame specifies only the stress pattern across syllables. In 6 implicit priming experiments, on each trial, participants produced 1 word out of a small set as quickly as possible. In homogeneous sets, the response words shared word-initial segments, whereas in heterogeneous sets, they did not. Priming effects from shared segments depended on all response words having the same number of syllables and stress pattern, but not on their having the same number of consonants and vowels. No priming occurred when the response words had only the same metrical frame but shared no segments. Computer simulations demonstrated that WEAVER accounts for the findings.

Abstract

This article reports 7 experiments investigating whether utterances are planned in a parallel or rightward incremental fashion during language production. The experiments examined the role of linear order, length, frequency, and repetition in producing Dutch verb–particle combinations. On each trial, participants produced 1 utterance out of a set of 3 as quickly as possible. The responses shared part of their form or not. For particle-initial infinitives, facilitation was obtained when the responses shared the particle but not when they shared the verb. For verb-initial imperatives, however, facilitation was obtained for the verbs but not for the particles. The facilitation increased with length, decreased with frequency, and was independent of repetition. A simple rightward incremental model accounts quantitatively for the results.

Abstract

The ability to explain the occurrence of errors in children’s speech is an essential component of successful theories of language acquisition. The present study tested some generativist and constructivist predictions about error on the questions produced by ten English-learning children between 2 and 5 years of age. The analyses demonstrated that, as predicted by some generativist theories [e.g. Santelmann, L., Berk, S., Austin, J., Somashekar, S. & Lust. B. (2002). Continuity and development in the acquisition of inversion in yes/no questions: dissociating movement and inflection, Journal of Child Language, 29, 813–842], questions with auxiliary DO attracted higher error rates than those with modal auxiliaries. However, in wh-questions, questions with modals and DO attracted equally high error rates, and these findings could not be explained in terms of problems forming questions with why or negated auxiliaries. It was concluded that the data might be better explained in terms of a constructivist account that suggests that entrenched item-based constructions may be protected from error in children’s speech, and that errors occur when children resort to other operations to produce questions [e.g. Dąbrowska, E. (2000). From formula to schema: the acquisition of English questions. Cognitive Liguistics, 11, 83–102; Rowland, C. F. & Pine, J. M. (2000). Subject-auxiliary inversion errors and wh-question acquisition: What children do know? Journal of Child Language, 27, 157–181; Tomasello, M. (2003). Constructing a language: A usage-based theory of language acquisition. Cambridge, MA: Harvard University Press]. However, further work on constructivist theory development is required to allow researchers to make predictions about the nature of these operations.

Abstract

Many current generativist theorists suggest that young children possess the grammatical principles of inversion required for question formation but make errors because they find it difficult to learn language-specific rules about how inversion applies. The present study analyzed longitudinal spontaneous sampled data from twelve 2–3-year-old English speaking children and the intensive diary data of 1 child (age 2;7 [years;months] to 2;11) in order to test some of these theories. The results indicated significantly different rates of error use across different auxiliaries. In particular, error rates differed across 2 forms of the same auxiliary subtype (e.g., auxiliary is vs. are), and auxiliary DO and modal auxiliaries attracted significantly higher rates of errors of inversion than other auxiliaries. The authors concluded that current generativist theories might have problems explaining the patterning of errors seen in children's questions, which might be more consistent with a constructivist account of development. However, constructivists need to devise more precise predictions in order to fully explain the acquisition of questions.

Abstract

Various accounts of metaphor interpretation propose that it involves constructing an ad hoc concept on the basis of the concept encoded by the metaphor vehicle (i.e. the expression used for conveying the metaphor). This paper discusses some of the differences between these theories and investigates their main empirical prediction: that metaphor interpretation involves enhancing properties of the metaphor vehicle that are relevant for interpretation, while suppressing those that are irrelevant. This hypothesis was tested in a cross-modal lexical priming study adapted from early studies on lexical ambiguity. The different patterns of suppression of irrelevant meanings observed in disambiguation studies and in the experiment on metaphor reported here are discussed in terms of differences between meaning selection and meaning construction.

Abstract

In this paper, I will give an overview of some well-studied multimodal signals that humans produce while they communicate with other humans, and discuss the implications of those studies for HCI. I will first discuss a conceptual framework that allows us to distinguish between functional and sensory modalities. This distinction is important, as there are multiple functional modalities using the same sensory modality (e.g., facial expression and eye-gaze in the visual modality). A second theoretically important issue is redundancy. Some signals appear to be redundant with a signal in another modality, whereas others give new information or even appear to give conflicting information (see e.g., the work of Susan Goldin-Meadows on speech accompanying gestures). I will argue that multimodal signals are never truly redundant. First, many gestures that appear at first sight to express the same meaning as the accompanying speech generally provide extra (analog) information about manner, path, etc. Second, the simple fact that the same information is expressed in more than one modality is itself a communicative signal. Armed with this conceptual background, I will then proceed to give an overview of some multimodalsignals that have been investigated in human-human research, and the level of understanding we have of the meaning of those signals. The latter issue is especially important for potential implementations of these signals in artificial agents. First, I will discuss pointing gestures. I will address the issue of the timing of pointing gestures relative to the speech it is supposed to support, the mutual dependency between pointing gestures and speech, and discuss the existence of alternative ways of pointing from other cultures. The most frequent form of pointing that does not involve the index finger is a cultural practice called lip-pointing which employs two visual functional modalities, mouth-shape and eye-gaze, simultaneously for pointing. Next, I will address the issue of eye-gaze. A classical study by Kendon (1967) claims that there is a systematic relationship between eye-gaze (at the interlocutor) and turn-taking states. Research at our institute has shown that this relationship is weaker than has often been assumed. If the dialogue setting contains a visible object that is relevant to the dialogue (e.g., a map), the rate of eye-gaze-at-other drops dramatically and its relationship to turn taking disappears completely. The implications for machine generated eye-gaze are discussed. Finally, I will explore a theoretical debate regarding spontaneous gestures. It has often been claimed that the class of gestures that is called iconic by McNeill (1992) are a “window into the mind”. That is, they are claimed to give the researcher (or even the interlocutor) a direct view into the speaker’s thought, without being obscured by the complex transformation that take place when transforming a thought into a verbal utterance. I will argue that this is an illusion. Gestures can be shown to be specifically designed such that the listener can be expected to interpret them. Although the transformations carried out to express a thought in gesture are indeed (partly) different from the corresponding transformations for speech, they are a) complex, and b) severely understudied. This obviously has consequences both for the gesture research agenda, and for the generation of iconic gestures by machines.

Abstract

In this paper, I compare three different assumptions about the relationship between speech, thought and gesture. These assumptions have profound consequences for theories about the representations and processing involved in gesture and speech production. I associate these assumptions with three simplified processing architectures. In the Window Architecture, gesture provides us with a 'window into the mind'. In the Language Architecture, properties of language have an influence on gesture. In the Postcard Architecture, gesture and speech are planned by a single process to become one multimodal message. The popular Window Architecture is based on the assumption that gestures come, as it were, straight out of the mind. I argue that during the creation of overt imagistic gestures, many processes, especially those related to (a) recipient design, and (b) effects of language structure, cause an observable gesture to be very different from the original thought that it expresses. The Language Architecture and the Postcard Architecture differ from the Window Architecture in that they both incorporate a central component which plans gesture and speech together, however they differ from each other in the way they align gesture and speech. The Postcard Architecture assumes that the process creating a multimodal message involving both gesture and speech has access to the concepts that are available in speech, while the Language Architecture relies on interprocess communication to resolve potential conflicts between the content of gesture and speech.

Abstract

In this paper, I will argue that although the study of multimodal interaction offers exciting new prospects for Human Computer Interaction and human-human communication research, language is the primary form of communication, even in multimodal systems. I will support this claim with theoretical and empirical arguments, mainly drawn from human-human communication research, and will discuss the implications for multimodal communication research and Human-Computer Interaction.

Abstract

Listeners’ signals of recipiency, such as “Mm-hm” or “uh-huh” in English, are the most elementary or minimal “conversational turns” possible. Minimal, because apart from acknowledging recipiency and inviting the speaker to continue with his/her next turn, they do not add any new information to the discourse of the conversation. The goal of this project is to gather cross cultural information on listeners’ feedback behaviour during conversation. Listeners in a conversation usually provide short signals that indicate to the speaker that they are still “with the speaker”. These signals could be verbal (like for instance “mm hm” in English or “hm hm” in Dutch) or nonverbal (visual), like nodding. Often, these signals are produced in overlap with the speaker’s vocalisation. If listeners do not produce these signals, speakers often invite them explicitly (e.g. “are you still there?” in a telephone conversation). Our goal is to investigate what kind of signals are used by listeners of different languages to signal “recipiency” to the speaker.

Abstract

Psycholinguistic theories propose different models of inflectional processing of regular and irregular verbs: dual mechanism models assume separate modules with lexical frequency sensitivity for irregular verbs. In contradistinction, connectionist models propose a unified process in a single module.We conducted a PET study using a 2 x 2 design with verb regularity and frequency.We found significantly shorter voice onset times for regular verbs and high frequency verbs irrespective of regularity. The PET data showed activations in inferior frontal gyrus (BA 45), nucleus lentiformis, thalamus, and superior medial cerebellum for both regular and irregular verbs but no dissociation for verb regularity.Our results support common processing components for regular and irregular verb inflection.

Abstract

The research presented in this dissertation examined the influence of prosodically-conditioned detail on the recognition of spoken words. The main finding is that subphonemic information in the speech signal that is conditioned by constituent-level prosodic structure can affect lexical processing systematically. It was shown that such information, as indicated by and estimated from the lengthening of speech sounds in the vicinity of prosodic boundaries, can help listeners to distinguish onset-embedded words (e.g. 'ham') from longer words that have this word embedded at their onset (e.g. 'hamster'). Furthermore, it was shown that variation in the realization of a spoken word that is associated with its position in the prosodic structure of an utterance can effect lexical processing. The pattern of competitor activation associated with the recognition of a monosyllabic spoken word in utterance-final position, where the realization of the word is strongly affected by the utterance boundary, is different from that associated with the recognition of the same word in utterance-medial position, where the realization of the word is less strongly affected by the following prosodic-word boundary. Taken together, the findings attest to the extraordinary sensitivity of the spoken-word recogntion system by demonstrating the relevance for lexical processing of very fine-grained phonetic detail conditioned by prosodic structure.

Abstract

Eye movements were monitored as participants followed spoken instructions to manipulate one of four objects pictured on a computer screen. Target words occurred in utterance-medial (e.g., Put the cap next to the square) or utterance-final position (e.g., Now click on the cap). Displays consisted of the target picture (e.g., a cap), a monosyllabic competitor picture (e.g., a cat), a polysyllabic competitor picture (e.g., a captain) and a distractor (e.g., a beaker). The relative proportion of fixations to the two types of competitor pictures changed as a function of the position of the target word in the utterance, demonstrating that lexical competition is modulated by prosodically conditioned phonetic variation.

Abstract

The structure of relative clauses has attracted considerable attention in recent years, and a number of authors have carried out analyses of the syntax of relativization. In our investigation of syntactic structure and change in New Guinea Tok Pisin, we find that the basic processes involved in relativization have much broader discourse functions, and that relativization is only a special instance of the application of general ‘bracketing’ devices used in the organization of information. Syntactic structure, in this case, can be understood as a component of, and derivative from, discourse structure.

Abstract

Most of the study of basic emotion expressions has focused on facial expressions and little work has been done to specifically investigate happiness, the only positive of the basic emotions (Ekman & Friesen, 1971). However, a theoretical suggestion has been made that happiness could be broken down into discrete positive emotions, which each fulfil the criteria of basic emotions, and that these would be expressed vocally (Ekman, 1992). To empirically test this hypothesis, 20 participants categorised 80 paralinguistic sounds using the labels achievement, amusement, contentment, pleasure and relief. The results suggest that achievement, amusement and relief are perceived as distinct categories, which subjects accurately identify. In contrast, the categories of contentment and pleasure were systematically confused with other responses, although performance was still well above chance levels. These findings are initial evidence that the positive emotions engage distinct vocal expressions and may be considered to be distinct emotion categories.

Abstract

Several theorists have proposed that distinctions are needed between different positive emotional states, and that these discriminations may be particularly useful in the domain of vocal signals (Ekman, 1992b, Cognition and Emotion, 6, 169–200; Scherer, 1986, Psychological Bulletin, 99, 143–165). We report an investigation into the hypothesis that positive basic emotions have distinct vocal expressions (Ekman, 1992b, Cognition and Emotion, 6, 169–200). Non-verbal vocalisations are used that map onto five putative positive emotions: Achievement/Triumph, Amusement, Contentment, Sensual Pleasure, and Relief. Data from categorisation and rating tasks indicate that each vocal expression is accurately categorised and consistently rated as expressing the intended emotion. This pattern is replicated across two language groups. These data, we conclude, provide evidence for the existence of robustly recognisable expressions of distinct positive emotions.

Abstract

A series of experiment tested Ekman’s (1992) hypothesis that there are a set of positive basic emotions that are expressed using vocal para-linguistic sounds, e.g. laughter and cheers. The proposed categories investigated were amusement, contentment, pleasure, relief and triumph. Behavioural testing using a forced-choice task indicated that participants were able to reliably recognize vocal expressions of the proposed emotions. A cross-cultural study in the preliterate Himba culture in Namibia confirmed that these categories are also recognized across cultures. A recognition test of acoustically manipulated emotional vocalizations established that the recognition of different emotions utilizes different vocal cues, and that these in turn differ from the cues used when comprehending speech. In a study using fMRI we found that relative to a signal correlated noise baseline, the paralinguistic expressions of emotion activated bilateral superior temporal gyri and sulci, lateral and anterior to primary auditory cortex, which is consistent with the processing of non linguistic vocal cues in the auditory ‘what’ pathway. Notably amusement was associated with greater activation extending into both temporal poles and amygdale and insular cortex. Overall, these results support the claim that ‘happiness’ can be fractionated into amusement, pleasure, relief and triumph.

Abstract

The difficulty of the task of segmenting a speech signal into its words is immediately clear when listening to a foreign language; it is much harder to segment the signal into its words, since the words of the language are unknown. Infants are faced with the same task when learning their first language. This study provides a better understanding of the task that infants face while learning their native language. We employed an automatic algorithm on the task of speech segmentation without prior knowledge of the labels of the phonemes. An analysis of the boundaries erroneously placed inside a phoneme showed that the algorithm consistently placed additional boundaries in phonemes in which acoustic changes occur. These acoustic changes may be as great as the transition from the closure to the burst of a plosive or as subtle as the formant transitions in low or back vowels. Moreover, we found that glottal vibration may attenuate the relevance of acoustic changes within obstruents. An interesting question for further research is how infants learn to overcome the natural tendency to segment these ‘dynamic’ phonemes.

Abstract

Articulatory feature (AF) modelling of speech has received a considerable amount of attention in automatic speech recognition research. Although termed ‘articulatory’, previous definitions make certain assumptions that are invalid, for instance, that articulators ‘hop’ from one fixed position to the next. In this paper, we studied two methods, based on support vector classification (SVC) and regression (SVR), in which the articulation continuum is modelled without being restricted to using discrete AF value classes. A comparison with a baseline system trained on quantised values of the articulation continuum showed that both SVC and SVR outperform the baseline for two of the three investigated AFs, with improvements up to 5.6% absolute.

Abstract

This paper addresses the problem of recognizing a vocabulary of over 50,000 city names in a telephone access spoken dialogue system. We adopt a two-stage framework in which only major cities are represented in the first stage lexicon. We rely on an unknown word model encoded as a phone loop to detect OOV city names (referred to as ‘rare city’ names). We use SpeM, a tool that can extract words and word-initial cohorts from phone graphs from a large fallback lexicon, to provide an N-best list of promising city name hypotheses on the basis of the phone graph corresponding to the OOV. This N-best list is then inserted into the second stage lexicon for a subsequent recognition pass. Experiments were conducted on a set of spontaneous telephone-quality utterances; each containing one rare city name. It appeared that SpeM was able to include nearly 75% of the correct city names in an N-best hypothesis list of 3000 city names. With the names found by SpeM to extend the lexicon of the second stage recognizer, a word accuracy of 77.3% could be obtained. The best one-stage system yielded a word accuracy of 72.6%. The absolute number of correctly recognized rare city names almost doubled, from 62 for the best one-stage system to 102 for the best two-stage system. However, even the best two-stage system recognized only about one-third of the rare city names retrieved by SpeM. The paper discusses ways for improving the overall performance in the context of an application.

Abstract

This paper addresses the issue of large-vocabulary recognition in a specific word class. We propose a two-pass strategy in which only major cities are explicitly represented in the first stage lexicon. An unknown word model encoded as a phone loop is used to detect OOV city names (referred to as rare city names). After which SpeM, a tool that can extract words and word-initial cohorts from phone graphs on the basis of a large fallback lexicon, provides an N-best list of promising city names on the basis of the phone sequences generated in the first stage. This N-best list is then inserted into the second stage lexicon for a subsequent recognition pass. Experiments were conducted on a set of spontaneous telephone-quality utterances each containing one rare city name. We tested the size of the N-best list and three types of language models (LMs). The experiments showed that SpeM was able to include nearly 85% of the correct city names into an N-best list of 3000 city names when a unigram LM, which also boosted the unigram scores of a city name in a given state, was used.

Abstract

Humans are able to recognise a word before its acoustic realisation is complete. This in contrast to conventional automatic speech recognition (ASR) systems, which compute the likelihood of a number of hypothesised word sequences, and identify the words that were recognised on the basis of a trace back of the hypothesis with the highest eventual score, in order to maximise efficiency and performance. In the present paper, we present an ASR system, SpeM, based on principles known from the field of human word recognition that is able to model the human capability of ‘early recognition’ by computing word activation scores (based on negative log likelihood scores) during the speech recognition process. Experiments on 1463 polysyllabic words in 885 utterances showed that 64.0% (936) of these polysyllabic words were recognised correctly at the end of the utterance. For 81.1% of the 936 correctly recognised polysyllabic words the local word activation allowed us to identify the word before its last phone was available, and 64.1% of those words were already identified one phone after their lexical uniqueness point. We investigated two types of predictors for deciding whether a word is considered as recognised before the end of its acoustic realisation. The first type is related to the absolute and relative values of the word activation, which trade false acceptances for false rejections. The second type of predictor is related to the number of phones of the word that have already been processed and the number of phones that remain until the end of the word. The results showed that SpeM’s performance increases if the amount of acoustic evidence in support of a word increases and the risk of future mismatches decreases.

Abstract

In this paper, we investigate the ability of SpeM, our recognition system based on the combination of an automatic phone recogniser and a wordsearch module, to determine as early as possible during the word recognition process whether a word is likely to be recognised correctly (this we refer to as ‘on-line’ early word recognition). We present two measures that can be used to predict whether a word is correctly recognised: the Bayesian word activation and the amount of available (acoustic) information for a word. SpeM was tested on 1,463 polysyllabic words in 885 continuous speech utterances. The investigated predictors indicated that a word activation that is 1) high (but not too high) and 2) based on more phones is more reliable to predict the correctness of a word than a similarly high value based on a small number of phones or a lower value of the word activation.

Abstract

Although researchers studying human speech recognition (HSR) and automatic speech recognition (ASR) share a common interest in how information processing systems (human or machine) recognize spoken language, there is little communication between the two disciplines. We suggest that this lack of communication follows largely from the fact that research in these related fields has focused on the mechanics of how speech can be recognized. In Marr's (1982) terms, emphasis has been on the algorithmic and implementational levels rather than on the computational level. In this article, we provide a computational-level analysis of the task of speech recognition, which reveals the close parallels between research concerned with HSR and ASR. We illustrate this relation by presenting a new computational model of human spoken-word recognition, built using techniques from the field of ASR that, in contrast to current existing models of HSR, recognizes words from real speech input.

Abstract

RU Radboud Universiteit Nijmegen, 16 september 2005

Abstract

In this paper, we illustrate the close parallels between the research fields of human speech recognition (HSR) and automatic speech recognition (ASR) using a computational model of human word recognition, SpeM, which was built using techniques from ASR. We show that ASR has proven to be useful for improving models of HSR by relieving them of some of their shortcomings. However, in order to build an integrated computational model of all aspects of HSR, a lot of issues remain to be resolved. In this process, ASR algorithms and techniques definitely can play an important role.

Abstract

The fields of human speech recognition (HSR) and automatic speech recognition (ASR) both investigate parts of the speech recognition process and have word recognition as their central issue. Although the research fields appear closely related, their aims and research methods are quite different. Despite these differences there is, however, lately a growing interest in possible cross-fertilisation. Researchers from both ASR and HSR are realising the potential benefit of looking at the research field on the other side of the ‘gap’. In this paper, we provide an overview of past and present efforts to link human and automatic speech recognition research and present an overview of the literature describing the performance difference between machines and human listeners. The focus of the paper is on the mutual benefits to be derived from establishing closer collaborations and knowledge interchange between ASR and HSR. The paper ends with an argument for more and closer collaborations between researchers of ASR and HSR to further improve research in both fields.

Abstract

The ultimate goal of our research is to develop a computational model of human speech recognition that is able to capture the effects of fine-grained acoustic variation on speech recognition behaviour. As part of this work we are investigating automatic feature classifiers that are able to create reliable and accurate transcriptions of the articulatory behaviour encoded in the acoustic speech signal. In the experiments reported here, we analysed the classification results from support vector machines (SVMs) and multilayer perceptrons (MLPs). MLPs have been widely and successfully used for the task of multi-value articulatory feature classification, while (to the best of our knowledge) SVMs have not. This paper compares the performance of the two classifiers and analyses the results in order to better understand the articulatory representations. It was found that the SVMs outperformed the MLPs for five out of the seven articulatory feature classes we investigated while using only 8.8–44.2% of the training material used for training the MLPs. The structure in the misclassifications of the SVMs and MLPs suggested that there might be a mismatch between the characteristics of the classification systems and the characteristics of the description of the AF values themselves. The analyses showed that some of the misclassified features are inherently confusable given the acoustic space. We concluded that in order to come to a feature set that can be used for a reliable and accurate automatic description of the speech signal; it could be beneficial to move away from quantised representations.

Abstract

Good teachers know their students, and exploit this knowledge to adapt or optimise their instruction. Traditional teachers know their students because they interact with them face-to-face in classroom or one-to-one tutoring sessions. In these settings, they can build student models, i.e., by exploiting the multi-faceted nature of human-human communication. In distance-learning contexts, teacher and student have to cope with the lack of such direct interaction, and this must have detrimental effects for both teacher and student. In a past study we have analysed teacher requirements for tracking student actions in computer-mediated settings. Given the results of this study, we have devised and implemented a tool that allows teachers to keep track of their learners'interaction in e-learning systems. We present the tool's functionality and user interfaces, and an evaluation of its usability.

Abstract

This study investigates whether or not the representation of lexical stress information can be primed during speech production. In four experiments, we attempted to prime the stress position of bisyllabic target nouns (picture names) having initial and final stress with auditory prime words having either the same or different stress as the target (e.g., WORtel–MOtor vs. koSTUUM–MOtor; capital letters indicate stressed syllables in prime–target pairs). Furthermore, half of the prime words were semantically related, the other half unrelated. Overall, picture names were not produced faster when the prime word had the same stress as the target than when the prime had different stress, i.e., there was no stress-priming effect in any experiment. This result would not be expected if stress were stored in the lexicon. However, targets with initial stress were responded to faster than final-stress targets. The reason for this effect was neither the quality of the pictures nor frequency of occurrence or voice-key characteristics. We hypothesize here that this stress effect is a genuine encoding effect, i.e., words with stress on the second syllable take longer to be encoded because their stress pattern is irregular with respect to the lexical distribution of bisyllabic stress patterns, even though it can be regular with respect to metrical stress rules in Dutch. The results of the experiments are discussed in the framework of models of phonological encoding.

Abstract

Any complete theory of speaking must take the dialogical function of language use into account. Pickering & Garrod (P&G) make some progress on this point. However, we question whether their interactive alignment model is the optimal approach. In this commentary, we specifically criticize (1) their notion of alignment being implemented through priming, and (2) their claim that self-monitoring can occur at all levels of linguistic representation.

Abstract

This study investigates whether or not masked form priming effects in the naming task depend on the number of shared segments between prime and target. Dutch participants named bisyllabic words, which were preceded by visual masked primes. When primes shared the initial segment(s) with the target, naming latencies were shorter than in a control condition (string of percent signs). Onset complexity (singleton vs. complex word onset) did not modulate this priming effect in Dutch. Furthermore, significant priming due to shared final segments was only found when the prime did not contain a mismatching onset, suggesting an interfering role of initial non-target segments. It is concluded that (a) degree of overlap (segmental match vs. mismatch), and (b) position of overlap (initial vs. final) influence the magnitude of the form priming effect in the naming task. A modification of the segmental overlap hypothesis (Schiller, 1998) is proposed to account for the data.

Abstract

To investigate the role of the syllable in Dutch speech production, five experiments were carried out to examine the effect of visually masked syllable primes on the naming latencies for written words and pictures. Targets had clear syllable boundaries and began with a CV syllable (e.g., ka.no) or a CVC syllable (e.g., kak.tus), or had ambiguous syllable boundaries and began with a CV[C] syllable (e.g., ka[pp]er). In the syllable match condition, bisyllabic Dutch nouns or verbs were preceded by primes that were identical to the target’s first syllable. In the syllable mismatch condition, the prime was either shorter or longer than the target’s first syllable. A neutral condition was also included. None of the experiments showed a syllable priming effect. Instead, all related primes facilitated the naming of the targets. It is concluded that the syllable does not play a role in the process of phonological encoding in Dutch. Because the amount of facilitation increased with increasing overlap between prime and target, the priming effect is accounted for by a segmental overlap hypothesis.

Abstract

The study endeavors a detailed and systematic classification of linguistic simultaneity expressions. Further, it aims at a well described survey of how simultaneity is expressed by native speakers in their own language. On the basis of real production data the book answers the questions of how native speakers express temporal simultaneity in general, and how learners at different levels of proficiency deal with this situation under experimental test conditions. Furthermore, the results of this study shed new light on our understanding of aspect in general, and on its acquisition by adult learners.

Abstract

Neuronal groups can interact with each other even if they are widely separated. One group might modulate its firing rate or its internal oscillatory synchronization to influence another group. We propose that coherence between two neuronal groups is a mechanism of efficient interaction, because it renders mutual input optimally timed and thereby maximally effective. Modulations of subjects' readiness to respond in a simple reaction-time task were closely correlated with the strength of gamma-band (40 to 70 hertz) coherence between motor cortex and spinal cord neurons. This coherence may contribute to an effective corticospinal interaction and shortened reaction times.

Abstract

Gender assignment relates to a native speaker's knowledge of the structure of the gender system of his/her language, allowing the speaker to select the appropriate gender for each noun. Whereas categorical assignment rules and exceptional gender assignment are well investigated, assignment regularities, i.e., tendencies in the gender distribution identified within the vocabulary of a language, are still controversial. The present study is an empirical contribution trying to shed light on the gender assignment system native German speakers have at their disposal. Participants presented with a category (e.g., predator) and a pair of gender-marked pseudowords (e.g., der Trelle vs. die Stisse) preferentially selected the pseudo-word preceded by the gender-marked determiner ‘‘associated’’ with the category (e.g., masculine). This finding suggests that semantic regularities might be part of the gender assignment system of native speakers.

Abstract

Previous studies have indicated that vocal and facial expressions of the ‘basic’ emotions share aspects of processing. Thus amygdala damage compromises the perception of fear and anger from the face and from the voice. In the current study we tested the hypothesis that there exist positive basic emotions, expressed mainly in the voice (Ekman, 1992). Vocal stimuli were produced to express the specific positive emotions of amusement, achievement, pleasure, contentment and relief.

Abstract

We have used the genome-wide marker genotypes from Genetic Analysis Workshop 15 Problem 2 to explore joint evidence for genetic linkage to rheumatoid arthritis across several samples. The data consisted of four high-density genome scans on samples selected for rheumatoid arthritis. We cleaned the data, removed intermarker linkage disequilibrium, and assembled the samples onto a common genetic map using genome sequence positions as a reference for map interpolation. The individual studies were combined first at the genotype level (mega-analysis) prior to a multipoint linkage analysis on the combined sample, and second using the genome scan meta-analysis method after linkage analysis of each sample. The two approaches were compared, and give strong support to the HLA locus on chromosome 6 as a susceptibility locus. Other regions of interest include loci on chromosomes 11, 2, and 12.

Abstract

Miraña, an endangered Witotoan language spoken in the Colombian Amazon region, has an inventory of over 60 noun class markers, most of which denote the shape of nominal referents. Class markers in this language are ubiquitous in their uses for derivational purposes in nouns and for agreement marking in virtually all other nominal expressions, such as pronouns, numerals, demonstratives, and relative clauses, as well as in verbs. This study provides a comprehensive analysis of this system by giving equal attention to its morphosyntactic, semantic, and discourse-pragmatic properties. The particular properties of this system raise issues in a number of ongoing theoretical discussions, in particular the typology of systems of nominal classification and the typology of reference tracking.

Abstract

Based on the conversation analysts’ insights into the various forms of third person reference in English, this paper first presents the inventory of forms Kilivila, the Austronesian language of the Trobriand Islanders of Papua New Guinea, offers its speakers for making such references. To illustrate such references to third persons in talk-in-interaction in Kilivila, a case study on gossiping is presented in the second part of the paper. This case study shows that ambiguous anaphoric references to two first mentioned third persons turn out to not only exceed and even violate the frame of a clearly defined situational-intentional variety of Kilivila that is constituted by the genre “gossip”, but also that these references are extremely dangerous for speakers in the Trobriand Islanders’ society. I illustrate how this culturally dangerous situation escalates and how other participants of the group of gossiping men try to “repair” this violation of the frame of a culturally defined and metalinguistically labelled “way of speaking”. The paper ends with some general remarks on how the understanding of forms of person reference in a language is dependent on the culture specific context in which they are produced.

Abstract

One of the central aims of the "Cognitive Anthropology Research Group" (since 1998 the "Department of Language and Cognition of the MPI for Psycholinguistics") is to research the relationship between language, culture and cognition and the conceptualization of space in various languages and cultures. Ever since its foundation in 1991 the group has been developing methods to elicit cross-culturally and cross-linguistically comparable data for this research project. After a brief summary of the central considerations that served as guidelines for the developing of these elicitation devices, this paper first presents a broad selection of the "space games" developed and used for data elicitation in the groups' various fieldsites so far. The paper then discusses the advantages and shortcomings of these data elicitation devices. Finally, it is argued that methodologists developing such devices find themselves in a position somewhere between Scylla and Charybdis - at least, if they take the requirement seriously that the elicited data should be comparable not only cross-culturally but also cross-linguistically.

Abstract

This article discusses how the Trobriand Islanders speak about body and mind. It addresses the following questions: do the linguistic datafit into theories about lexical universals of body-part terminology? Can we make inferences about the Trobrianders' conceptualization of psychological and physical states on the basis of these data? If a Trobriand Islander sees these idioms as external manifestations of inner states, then can we interpret them as a kind of ethnopsychological theory about the body and its role for emotions, knowledge, thought, memory, and so on? Can these idioms be understood as representation of Trobriand ethnopsychological theory?

Abstract

When we communicate, we communicate in a certain context, and this context shapes our utterances. Natural languages are context-bound and deixis 'concerns the ways in which languages encode or grammaticalise features of the context of utterance or speech event, and thus also concerns ways in which the interpretation of utterances depends on the analysis of that context of utterance' (Stephen Levinson). The systems of deixis and demonstratives in the Oceanic languages represented in the contributions to this volume illustrate the fascinating complexity of spatial reference in these languages. Some of the studies presented here highlight social aspects of deictic reference illustrating de Leon's point that 'reference is a collaborative task' . It is hoped that this anthology will contribute to a better understanding of this area and provoke further studies in this extremely interesting, though still rather underdeveloped, research area.

Abstract

Arbeitet man als muttersprachlicher Sprecher des Deutschen mit Corpora gesprochener oder geschriebener deutscher Sprache, dann reflektiert man in aller Regel nur selten über die Vielzahl von kulturspezifischen Informationen, die in solchen Texten kodifiziert sind – vor allem, wenn es sich bei diesen Daten um Texte aus der Gegenwart handelt. In den meisten Fällen hat man nämlich keinerlei Probleme mit dem in den Daten präsupponierten und als allgemein bekannt erachteten Hintergrundswissen. Betrachtet man dagegen Daten in Corpora, die andere – vor allem nicht-indoeuropäische – Sprachen dokumentieren, dann wird einem schnell bewußt, wieviel an kulturspezifischem Wissen nötig ist, um diese Daten adäquat zu verstehen. In meinem Vortrag illustriere ich diese Beobachtung an einem Beispiel aus meinem Corpus des Kilivila, der austronesischen Sprache der Trobriand-Insulaner von Papua-Neuguinea. Anhand eines kurzen Auschnitts einer insgesamt etwa 26 Minuten dauernden Dokumentation, worüber und wie sechs Trobriander miteinander tratschen und klatschen, zeige ich, was ein Hörer oder Leser eines solchen kurzen Daten-Ausschnitts wissen muß, um nicht nur dem Gespräch überhaupt folgen zu können, sondern auch um zu verstehen, was dabei abläuft und wieso ein auf den ersten Blick absolut alltägliches Gespräch plötzlich für einen Trobriander ungeheuer an Brisanz und Bedeutung gewinnt. Vor dem Hintergrund dieses Beispiels weise ich dann zum Schluß meines Beitrags darauf hin, wie unbedingt nötig und erforderlich es ist, in allen Corpora bei der Erschließung und Kommentierung von Datenmaterialien durch sogenannte Metadaten solche kulturspezifischen Informationen explizit zu machen.

Abstract

One of the many results of the "Space" research project conducted at the MPI for Psycholinguistics is that there are three "Frames of spatial Reference" (FoRs), the relative, the intrinsic and the absolute FoR. Cross-linguistic research showed that speakers who prefer one FoR in verbal spatial references rely on a comparable coding system for memorizing spatial configurations and for making inferences with respect to these spatial configurations in non-verbal problem solving. Moreover, research results also revealed that in some languages these verbal FoRs also influence gestural behavior. These results document the close interrelationship between language, culture and cognition in the domain "Space". The proper description of these interrelationships in the spatial domain requires language and culture specific ontologies.

Abstract

This handbook chapter summarizes some of the problems of nominal classification in language, presents and illustrates the various systems or techniques of nominal classification, and points out why nominal classification is one of the most interesting topics in Cognitive Linguistics.

Abstract

Human ethologists have shown that humans are both attracted to others and at the same time fear them. They refer to this kind of fear with the technical term ‘social fear’ and claim that “it is alleviated with personal acquaintance but remains a principle characteristic of interpersonal behaviour. As a result, we maintain various degrees of greater distance between ourselves and others depending on the amount of confidence we have in the other” (Eibl-Eibesfeldt 1989: 335). The goal of this task is to conduct exploratory, heuristic research to establish a new subproject that – based on a corpus of video data – will investigate various forms of human spatial behaviour cross-culturally.

Abstract

A new sign language has been created by deaf Nicaraguans over the past 25 years, providing an opportunity to observe the inception of universal hallmarks of language. We found that in their initial creation of the language, children analyzed complex events into basic elements and sequenced these elements into hierarchically structured expressions according to principles not observed in gestures accompanying speech in the surrounding language. Successive cohorts of learners extended this procedure, transforming Nicaraguan signing from its early gestural form into a linguistic system. We propose that this early segmentation and recombination reflect mechanisms with which children learn, and thereby perpetuate, language. Thus, children naturally possess learning abilities capable of giving language its fundamental structure.