Showing papers on "Utterance published in 2001"

Spoken word production: A theory of lexical access

Abstract: A core operation in speech production is the preparation of words from a semantic base. The theory of lexical access reviewed in this article covers a sequence of processing stages beginning with the speaker’s focusing on a target concept and ending with the initiation of articulation. The initial stages of preparation are concerned with lexical selection, which is zooming in on the appropriate lexical item in the mental lexicon. The following stages concern form encoding, i.e., retrieving a word’s morphemic phonological codes, syllabifying the word, and accessing the corresponding articulatory gestures. The theory is based on chronometric measurements of spoken word production, obtained, for instance, in picture-naming tasks. The theory is largely computationally implemented. It provides a handle on the analysis of multiword utterance production as well as a guide to the analysis and design of neuroimaging studies of spoken utterance production.

First Steps toward a Usage-Based Theory of Language Acquisition

Abstract: Usage-based models of language focus on the speci®c communicative events in which people learn and use language. In these models, the psycholinguistic units with which individuals operate are determined not by theoretical ®at but by observation of actual language use in actual communicative events. This data-based approach make these models especially congenial for the analysis of children's language, since children do not learn and use the same units as adults. In this paper I employ a usage-based model of language to argue for ®ve fundamental facts about child language acquisition: (1) the primary psycholinguistic unit of child language acquisition is the utterance, which has as its foundation the expression and understanding of communicative intentions; (2) early in their language development children are attempting to reproduce not adult words but whole adult utterances; (3) children's earliest utterances are almost totally concrete in the sense that they are instantiations of item-based schemas or constructions; (4) abstractions result from children generalizing across the type variation they observe at particular ``slots'' in otherwise recurrent tokens of the same utterance; and (5) children create novel utterances for themselves via usage-based syntactic operations in which they begin with an utterance-level schema and then modify that schema for the exigencies of the particular communicative situation (usage event) at hand.

Pragmatic Markers and Sociolinguistic Variation: A Relevance-Theoretic Approach to the Language of Adolescents

Abstract: This book combines theoretical work in linguistic pragmatics and sociolinguistics with empirical work based on a corpus of London adolescent conversation. It makes a general contribution to the study of pragmatic markers, as it proposes an analytical model that involves notions such as subjectivity, interactional and textual capacity, and the distinction between contextual alignment/divergence. These notions are defined according to how information contained in an utterance interacts with the cognitive environment of the hearer. Moreover, the model captures the diachronic development of markers from lexical items via processes of grammaticalisation, arguing that markerhood may be viewed as a gradient phenomenon. The empirical work concerns the use of like as a marker, as well as a characteristic use of two originally interrogative forms, innit and is it, which are used as attitudinal markers throughout the inflectional paradigm, despite the fact that they contain a third person singular neuter pronoun. The author provides an in-depth analysis of these features in terms of pragmatic functions, diachronic development and sociolinguistic variation, thus adding support to the hypothesis that adolescents play an important role in language variation and change.

Reference and Reflexivity

Abstract: In this volume, author John Perry develops a "reflexive-referential" account of indexicals, demonstratives and proper names. On these issues the philosophy of language in the 20th century was shaped by two competing traditions, descriptivist and referentialist. Referentialist tradition is portrayed as holding that indexicals contribute content that involves individuals without identifying conditions on them. Descriptivist tradition is portrayed as holding that referential content does not explain all of the identifying conditions conveyed by names and indexicals. This text reveals a coherent and structured family of contents - from reflexive contents that place conditions on their actual utterance to the fully incremental contents that place conditions only on the objects of reference - reconciling the insights of both traditions.

Method of comparing utterances for security control

Abstract: A security method compares a present verbal utterance with a previously recorded verbal utterance by comparing time-frequency domain representations of the utterances, with multiple repeat utterances forming a basis for determining a variation in repetitious performance by an individual, and similar differences between enrollment and challenge utterances forming a basis for a similar analysis of variance between enrollment and challenge utterances. In one embodiment a set of enrollment data is searched by each challenge until either a match is made, indicating an action, possibly dependent upon the specific match, or no match is made indicating an abort. In one application an individual is accepted or rejected as an imposter, in another applicaton, a selected action is accepted as corresponding to a verbal command.

The development of relative clauses in spontaneous child speech

Abstract: This study examines the development of relative clauses in the speech of four English-speaking children between 1;9 and 5;2 years of age. It is shown that the earliest relative clauses occur in presentational constructions that express a single proposition in two ®nite clauses. Starting from such simple sentences, children gradually learn the use of more complex constructions in which the relative clause modi®es the noun of a fulledged main clause. Five factors are considered that might contribute to the development of relative clauses in spontaneous child speech: (1) the ambient language, (2) the formulaic character of the main clause, (3) the information structure of the whole utterance, (4) the communicative function of presentational relatives, and (5) the limited processing capacity of young children.

Automatically improving a voice recognition system

Abstract: Systems and methods for automatically improving a voice recognition system are provided. In one embodiment, the systems and methods retrieve voice recognition information produced by a voice recognition system in response to recognizing a user utterance. The voice recognition information comprises a recognized voice command associated with the user utterance and a reference to an audio file that includes the user utterance. The audio file is played and it is determined if the recognized voice command matches the user utterance included in the audio file. The user utterance is then transcribed to create a transcribed utterance, if the recognized voice command does not match the user utterance. The transcribed utterance is then recorded in association with the recognized voice command to monitor recognition accuracy.

System and method for determining utterance context in a multi-context speech application

Abstract: A speech center coordinates speech services for a number of speech-enabled applications performing on a computer system. The speech center includes a conversation manager that manages conversations between a user and the speech-enabled applications. The conversation manager includes a context manager that maintains a context list of speech-enabled applications that the user has accessed. If the user speaks an utterance, the context manager determines which speech-enabled application should receive a translation of the utterance from the context list, such as by determining the most recently accessed application.

The role of pronouns in young children's acquisition of the English transitive construction.

Abstract: Two studies investigating the linguistic representations underlying English-speaking 2 1/2-year-olds' production of transitive utterances are reported. The first study was a training study in which half the children heard utterances with full nouns as agent and patient, and half the children heard utterances with both pronouns (i.e., He's [verb]-ing it) and also full nouns. In subsequent testing, only children who had been trained with pronouns and nouns were able to produce a transitive utterance creatively with a nonce verb. The second study reported an analogous set of findings, but in comprehension. Together, the results of these 2 studies suggest that English-speaking children build many of their early linguistic constructions around certain specific lexical or morphological items and patterns, perhaps especially around particular pronoun configurations.

Multi-stage large vocabulary speech recognition system and method

Abstract: Multiple processing stages are provided with different vocabulary databases to improve processing time, efficiency, and accuracy in speech recognition. The entire vocabulary is divided into smaller vocabulary subsets, which are associated with particular keywords. A small vocabulary subset is generated or retrieved based on certain information, such as a calling party's locality. A user is prompted to provide input information, such as the locality in which a business whose phone number is requested is located, in the form of a spoken utterance to the system. If the utterance matches one of the entries in the initial small vocabulary subset, then the utterance is considered to be recognizable. If the utterance is not recognizable when compared to the initial small vocabulary subset, then the utterance is stored for later use. The user is then prompted for a keyword related to another subset of words in which his initial utterance may be found. A vocabulary subset associated with the received keyword is generated or retrieved. The initial stored utterance is then retrieved and compared to the newly loaded vocabulary subset. If the utterance matches one of the entries in the newly loaded vocabulary subset, then the utterance is recognizable. Otherwise, it is determined that the initial utterance was unrecognizable, and the user is prompted to repeat the initial utterance.

System and method for relating syntax and semantics for a conversational speech application

Abstract: A conversation manager processes spoken utterances from a user of a computer. The conversation manager includes a semantics analysis module and a syntax manager. A domain model that is used in processing the spoken utterances includes an ontology (i.e., world view for the relevant domain of the spoken utterances), lexicon, and syntax definitions. The syntax manager combines the ontology, lexicon, and syntax definitions to generate a grammatic specification. The semantics module uses the grammatic specification and the domain model to develop a set of frames (i.e., internal representation of the spoken utterance). The semantics module then develops a set of propositions from the set of frames. The conversation manager then uses the set of propositions in further processing to provide a reply to the spoken utterance.

System, method and computer program product for self-verifying file content in a speech recognition framework

Abstract: A system, method and computer program product are provided for verifying utterances in a database. Initially, a plurality of utterances are stored in a database each in a separate file. Next, a transcription associated with each of the utterances is identified. The utterance in each file is subsequently recognized utilizing speech recognition. The utterances in the database are then verified utilizing an output of the speech recognition step and the transcription associated with the utterance in each file.

Influencing a voice recognition matching operation with user barge-in time

Abstract: A system, method and computer program product for automated voice recognition using the barge-in time of a user's utterance is provided. A list of items is audibly provided to a user. The user selects one item from the list by providing an utterance representing the user's selection. The utterance time (i.e., the barge-in time) is calculated from the utterance of the user. Then, the utterance is compared to each item in the list so as to generate a confidence score for each item in the list, with the confidence score being based on the similarity between the item and the utterance. The confidence scores are also based on the barge-in time. One of the items from the list is selected based on the confidence scores of the items.

User barge-in enablement in large vocabulary speech recognition systems

Abstract: An interactive voice response unit which provides beneficial operation by including means to handle unconstrained input such as natural speech and to allow barge-in includes a prompter, a recognizer of speech signals, a meaningful phrase detector and classifier, and a turn-taking module, all under control of a dialog manager. In the course of listening to user input while outputting a voiced message, the voice response unit processes the received signal and ascertains whether it is receiving an utterance that is intended to interrupt the prompt, or merely noise or an utterance that is not meant to be used by the arrangement. The unit is sensitive to the speed and context of the speech provided by the user and is thus able to distinguish between a situation where a speaker is merely pausing and a situation where a speaker is done speaking.

Speech recognition using word-in-phrase command

Abstract: A method (1400, 1435) is described that corrects incorrect text associated with recognition errors in computer-implemented speech recognition. The method includes the step of performing speech recognition on an utterance to produce a recognition result (1405) for the utterance. The command includes a word and a phrase (1500). The method includes determining if a word closely corresponds to a portion of the phrase (1505). A speech recognition result is produced if the word closely corresponds to a portion of the phrase (1520, 1525).

polysp: a polysystemic, phonetically-rich approach to speech understanding

Abstract: We outline an approach to speech understanding, Polysp (for POLYsystemic SPeech understanding) that combines a richly-structured, polysystemic linguistic model derived from Firthian prosodic analysis and declarative phonology, with psychological and neuropsychological approaches to the organization of sensory experience into knowledge. We propose that the type of approach exemplified by Polysp promises a fruitful way of conceptualising how meaning is understood from spoken utterances, partly by ascribing an important role to all kinds of systematic fine phonetic detail available in the physical speech signal and by rejecting assumptions that the physical signal is analysed as early as possible into abstract linguistic units. Polysp provides a framework by which episodic multimodal sensory experience of speech can be simultaneously processed into different types of linguistic and non-linguistic knowledge at a variety of levels of abstraction, with the emphasis always on understanding meaning in order to interact with another person rather than on building a complete description of a given utterance at successive, obligatory stages of formal linguistic analysis. We discuss phonetic data consistent with these views.

Polysp: a polysystemic, phonetically-ric h approach to speech understanding

Abstract: We outline an approach to speech understanding, Polysp (for POLYsystemic SPeech understanding) that combines a richly-structured, polysystemic linguistic model derived from Firthian prosodic analysis and declarative phonology, with psychological and neuropsychological approaches to the organization of sensory experience into knowledge. We propose that the type of approach exemplified by Polysp promises a fruitful way of conceptualisi ng how meaning is understood from spoken utterances, partly by ascribing an important role to all kinds of systematic fine phonetic detail available in the physical speech signal and by rejecting assumptions that the physical signal is analysed as early as possible into abstract linguistic units. Polysp provides a framework by which episodic multimodal sensory experience of speech can be simultaneousl y processed into different types of linguistic and non-linguistic knowledge at a variety of levels of abstraction, with the emphasis always on understanding meaning in order to interact with another person rather than on building a complete description of a given utterance at successive, obligatory stages of formal linguistic analysis. We discuss phonetic data consistent with these views. This paper explores the contribution of phonetic knowledge to how we understand words, and some implications for what makes a plausible model of spoken word understanding. We show that certain types of fine phonetic detail systematically reflect not just the phonemic content but the wider phonological and grammatical structure of the message; that while some systematic differences in phonetic fine detail are relatively localised in the speech signal, others stretch over several syllables; and that both types can make speech easier to understand. We make the case that one consequence of neglecting fine phonetic detail in models of spoken word recognition and understanding is that other processes and stages of analysis may be given inappropriate emphasis, and that this has happened in models which adopt the convenient fiction that the phoneme is the basic input unit to the lexicon. In consequence, no current phonetic or psycholinguistic theory accounts satisfactorily for how normal connected speech is understood.

Selective speaker adaptation for an in-vehicle speech recognition system

Abstract: Disclosed herein is a method of improving the recognition accuracy of an in-vehicle speech recognition system. The method of the present invention selectively adapts the system's speech engine to a speaker's voice characteristics using an N-best matching technique. In this method, the speech recognition system receives and processes a spoken utterance relating to a car command and having particular speaker-dependent speech characteristics so as to select a set of N-best voice commands matching the spoken utterance. Upon receiving a training mode input from the speaker, the system outputs the N-best command set to the speaker who selects the correct car command. The system then adapts the speech engine to recognize a spoken utterance having the received speech characteristics as the user-selected car command.

Computer-implemented multi-scanning language method and system

Abstract: A computer-implemented method and system for speech recognition of a user speech input. The user speech input which contains utterances from a user is received. A first language model recognizes at least a portion of the utterances from the user speech input. The first language model has utterance terms that form a general category. A second language model is selected based upon the identified utterances from use of the first language model. The second language model contains utterance terms that are a subset category of the general category of utterance terms in the first language model. Subset utterances are recognized with the selected second language model from the user speech input.

Secondary topic as a relation in information structure

Abstract: Information structure may be understood as the pragmatic structuring of a proposition in terms of the speaker’s assumptions concerning the addressee’s state of mind at the time of the utterance. The commonly assumed binominal partition of the information structure into topic‐focus, theme‐rheme, topic‐ comment, focus‐open proposition, or focus‐presupposition is sometimes insuYcient. The paper addresses the notion of secondary topic, which, although sometimes used in research on information structure, has not been studied closely. The main focus of the paper is Ostyak (Uralic), where the secondary topic is systematically expressed by object agreement. The results of the paper indicate that multiple topic constructions suggested in the literature do not necessarily involve topicalization as (multiple) adjunction but, rather, argue for the possibility of two clause-internal argument topics. The paper suggests some independent criteria for identifying secondary topic and shows how it interacts with other information-structure relations, as well as with syntax and semantics. This further raises the question of cross-linguistic variation in the grammatical realization of the secondary topic function within a clause.

Root infinitives in Dutch early child language: an effect of input?

Abstract: Children who acquire Dutch as their first language show a strong preference for using infinitival verb forms during the early stages of grammatical development. This exemplifies the 'root infinitive' (RI) phenomenon, which has played a significant role in recent discussions on the development of syntax. Most accounts proposed thus far invoke an immaturity of the child's grammatical competence. We explore the possibility that the early predominance of infinitival forms is related to patterns in the language input. We analysed a corpus of utterances addressed by two Dutch-speaking mothers to their two- to three-year old sons. Root infinitive utterances amounted to 10%, and auxiliary-plus infinitive main verb constructions, which in terms of word order are maximally similar to RIs, constituted 30% of all verb-containing utterances. These figures render an account in terms of exposure to utterance structures unlikely. There is a moderate (but significant) correlation between frequency of occurrence of individual verb forms in the input and age of acquisition. However, infinitive verb forms are often acquired earlier than their input frequency would predict, and this may be related to an overall increased salience, due to their systematic appearance in sentence-final position and their relatively high conceptual transparency as compared to finite verbs.

Evaluating Contemporary Models of Figurative Language Understanding

Abstract: The 7 articles in this collection describe new views of figurative language understanding. These articles differ in the types of figurative language they examine and address different linguistic levels (i.e., word vs. utterance) in the models proposed. Yet the articles make important claims about the use of contextual information in inferring figurative meaning and similarities and differences in processing figurative and literal language. These models are evaluated; places where the models conflict are noted; and ways these new ideas are improvements over more general, traditional theories of figurative language understanding are suggested. Several issues and topics are also raised that are not addressed in these articles but should be of concern to figurative language scholars.

System, method and computer program product for dynamic billing using tags in a speech recognition framework

Abstract: A system, method and computer program product are disclosed for providing dynamic billing in a speech recognition framework in accordance with an embodiment of the present invention. An utterance from a user is received via a speech recognition portal. The utterance is representative of a request for a service. The request for the service associated with the utterance is recognized utilizing a speech recognition process. Subsequently, an event for executing the requested service is issued utilizing a tag associated with an extensible markup language. The requested service is executed utilizing the tag. The tag is also utilized to generate a bill for the execution of the requested service.

Method and system for semantic speech recognition

Abstract: The present invention discloses a computer-implemented method to understand queries or commands spoken by users when they use natural language utterances similar to those that people use spontaneously to communicate. More precisely, the invention discloses a method that identifies user queries or commands from the general information involved in spoken utterances directly by the speech recognition system, and not by a post-process as is conventionally used. In a phase of preparation of the system, a vocabulary of items representing data and semantic identifiers is created as well as a syntax module having valid combinations of items. When the system is in use, a user utterance is first discretized into a plurality of basic speech units which are compared to the items in the vocabulary and a combination of items is selected according to the evaluation from the syntax module in order to generate the most likely sequence of items representative of the user utterance. Finally the semantic identifiers and the data extracted from the user utterance are used to call the appropriate function that process the user request.

Covariation between spatial language and cognition, and its implications for language learning

Abstract: Word learning: the scope of the problem Much of this volume is concerned with the problem of how children learn the meanings of words, or more exactly morphemes of various kinds. The answers may depend on many factors of course: different kinds of morphemes may be, indeed must be, learnt in different ways, the ways themselves may be opened up by conceptual development within the child over time, and so on. However, we can (not without some danger) abstract away from these divergent factors and ask general questions about the scope of the “induction” or “mapping” problem. Now in the chapters that immediately precede, those by Bowerman & Choi, de Leon, and Brown, an issue is raised that has significant bearing on the dimensions of the “mapping problem.”What these chapters show is that, in a number of spatial domains, the kinds of categories that need to be associated with the meanings of words can vary rather drastically across languages, and moreover that, at least by the beginning of systematic speech, there is no evidence of a uniform initial state of the learning machine, i.e. little evidence that children are presuming certain kinds of natural categories, later discarding them in favor of the local idiosyncrasies. Nor does this picture change when one starts to plumb comprehension before the age of complex utterance production. In short, the semantic categories look almost as variable as the phonological strings onto which they must be mapped.

Spoken word access processes: An introduction

Abstract: We introduce the papers in this special issue by summarising the current major issues in spoken word recognition. We argue that a full understanding of the process of lexical access during speech comprehension will depend on resolving several key representational issues: what is the form of the representations used for lexical access; how is phonological information coded in the mental lexicon; and how is the morphological and semantic information about each word stored? We then discuss a number of distinct access processes: competition between lexical hypotheses; the computation of goodness-of-et between the signal and stored lexical knowledge; segmentation of continuous speech; whether the lexicon ineuences prelexical processing through feedback; and the relationship of form-based processing to the processes responsible for deriving an interpretation of a complete utterance. We conclude that further progress may well be made by swapping ideas among the different sub-domains of the discipline. The proportion of the world’s population that is kept awake at night worrying about spoken word access processes is, undoubtedly, vanishingly small. After all, hardly anyone has even heard the phrase Spoken word access processes. Such things are hardly the stuff mobile-phone conversations are made on. And yet in another way, as we psycholinguists know, spoken word access processes are what those conversations depend on. An English listener chosen at random might not understand this phrase as a whole, but this would not be because of a failure to recognise the words themselves. Spoken word recognition is remarkably robust and seemingly effortless. Chances are high that our native speaker of English would be able to recognise our four-word phrase with no difeculty, even if it were spoken over a mobile phone in a noisy station by someone whose voice he or she had never heard before, irrespective of how fast that talker spoke, and

Essays in speech act theory

Abstract: 1. Chapter 1. Introduction (by Vanderveken, Daniel) 2. Part I. General Theory 3. Chapter 2. Universal Grammar and Speech Act Theory (by Vanderveken, Daniel) 4. Chapter 3. Verbal Mood and Sentence Moods in the Tradition of Universal Grammar (by Leclerc, Andre) 5. Chapter 4. How Performatives Work (by Searle, John R.) 6. Chapter 5. Possible Directions of Fit between Mind, Language and the World (by Sousa Melo, Candida J. de) 7. Part II. Discourse and Interlocution 8. Chapter 6. Speech Acts and the logic of mutual understanding (by Trognon, Alain) 9. Chapter 7. Utterance acts and speech acts (by Davis, Steven) 10. Chapter 8. An Ascription-Based Theory of Illocutionary Acts (by Yamada, Tomoyuki) 11. Chapter 9. An approach for modelling and simulating conversations (by Moulin, Bernard) 12. Part III. Speech Acts in Linguistics 13. Chapter 10. Illocutionary Morphology and Speech Acts (by Kubo, Susumu) 14. Chapter 11. Speech-Act Constructions, Illocutionary Forces, and Conventionality (by Yamanashi, Masa-aki) 15. Chapter 12. Speech act theory and the analysis of conversation (by Moeschler, Jacques) 16. Chapter 13. Speech Acts and Relevance Theory (by Dominicy, Marc) 17. Notes 18. References 19. Notes on Contributors 20. Subject Index 21. Name Index

Natural language and dialogue generation processing

Abstract: A system and method for interacting with objects, via a computer using utterances, speech processing and natural language processing. A Dialogue Generation Log provides a conversational memory by relating earlier generated utterances with possible new utterances. The Dialogue Generation Log encompasses a memory structure relating the prior spoken utterances with objects, and conversational contexts. The speech processor searches the Dialogue Generation Log for a matching phrase for the proposed utterance, and determines whether the utterance should be repeated. The determination may be made on a case-by-case basis, depending upon the context of the conversational utterance. Custom user information may be stored globally using a global user registry for remote access. The system reduces the repetitive nature of computer generated conversation by not voicing utterances that have been generated recently, and by learning user preferences from already voiced conversations.

You're not from 'round here, are you?: naive Bayes detection of non-native utterance text

Abstract: Native and non-native use of language differs, depending on the proficiency of the speaker, in clear and quantifiable ways. It has been shown that customizing the acoustic and language models of a natural language understanding system can significantly improve handling of non-native input; in order to make such a switch, however, the nativeness status of the user must be known. In this paper, we show that naive Bayes classification can be used to identify non-native utterances of English. The advantage of our method is that it relies on text, not on acoustic features, and can be used when the acoustic source is not available. We demonstrate that both read and spontaneous utterances can be classified with high accuracy, and that classification of errorful speech recognizer hypotheses is more accurate than classification of perfect transcriptions. We also characterize part-of-speech sequences that play a role in detecting non-native speech.