Lisa Lavoie | Emerson College (original) (raw)
Papers by Lisa Lavoie
The term lenition is often used to describe sound changes, both historical and synchronic,<br&... more The term lenition is often used to describe sound changes, both historical and synchronic,<br> but the range and limits of lenition have not been well-defined. This report presents the results emerging from a data base of consonant strength alternations from over 165 languages. The data base represents a significant advance in the study of consonant strength by providing specific examples in sufficient quantity to compare types and frequency of alternations across languages. These generalizations demonstrate that common notions of lenition, rooted in historical change, over-regularize the phenomenon. The logically possible types of lenition and fortition are not evenly distributed; some are overwhelmingly common and others practically non-existent. The environments of alternations are often overlooked but crucial in determining if an alternation is lenition or fortition. The data base shows that consonant strength behavior exhibits asymmetries when compared to standard assumpt...
Yindjibarndi, a Ngarluma language spoken in northwestern Australia, displays an intriguing patter... more Yindjibarndi, a Ngarluma language spoken in northwestern Australia, displays an intriguing pattern of stress in which all long vowels may be pronounced as a sequence of two short vowels. Previous analyses (Wordick 1982, Kager 1993) have treated this as vowel breaking, but I show that the language's morphology, historical and synchronic phonology, and metrical structure support an analysis in which identical adjacent short vowels coalesce into a single long vowel, but only where they can serve as the head of a foot. I construct an Optimality Theoretic account of these facts, drawing on numerous well-established constraint types including correspondence constraints (McCarthy & Prince 1995), positional faithfulness constraints (Beckman 1997), prosodic constraints (McCarthy & Prince 1993), and a version of the Weight to Stress Principle (Prince 1990). Additionally, I motivate a sociolinguistic constraint, *RUDE, which forces consonant lenition. In addition to accounting for the inte...
We present an approach for the manual labeling of speech at the articulatory feature level, and a... more We present an approach for the manual labeling of speech at the articulatory feature level, and a new set of labeled conversational speech collected using this approach. A detailed transcription, including overlapping or reduced gestures, is useful for studying the great pronunciation variability in conversational speech. It also facilitates the testing of feature classifiers, such as those used in articulatory approaches to automatic speech recognition. We describe an effort to transcribe a small set of utterances drawn from the Switchboard database using eight articulatory tiers. Two transcribers have labeled these utterances in a multi-pass strategy, allowing for correction of errors. We describe the data collection methods and analyze the data to determine how quickly and reliably this type of transcription can be done. Finally, we demonstrate one use of the new data set by testing a set of multilayer perceptron feature classifiers against both the manual labels and forced align...
Consonants exhibit more variation in their phonetic realization than is typically acknowledged, b... more Consonants exhibit more variation in their phonetic realization than is typically acknowledged, but that variation is linguistically constrained. Acoustic analysis of both read and spontaneous speech reveals that consonants are not necessarily realized with the manner of articulation they would have in careful citation form. Although the variation is wider than one would imagine, it is limited by the phoneme inventory. The phoneme inventory of the language restricts the range of variation to protect the system of phonemic contrast. That is, consonants may stray phonetically into unfilled areas of the language's sound space. Listeners are seldom consciously aware of the consonant variation, and perceive the consonants phonemically as in their citation forms. A better understanding of surface phonetic consonant variation can help make predictions in theoretical domains and advances in applied domains. 1 Consonant variation The consonant variation of interest here goes beyond well...
� Motivations: A detailed transcription, including overlapping/reduced gestures • is useful for s... more � Motivations: A detailed transcription, including overlapping/reduced gestures • is useful for studying pronunciation variation • facilitates the testing of automatic articulatory feature classifiers � Alternatives: • Narrow phonetic transcription (e.g.
We investigate monosyllabic words with rimes consisting of a diphthong or non-low tense vowel fol... more We investigate monosyllabic words with rimes consisting of a diphthong or non-low tense vowel followed by a liquid, such as file, foul, foil, feel, fool, fail; fire, flour and foyer, which we term sesquisyllables. Evidence from phonological distribution, speaker intuition, metrical properties, variant pronunciations, and an acoustic study converges on the interpretation that these are trimoraic monosyllables. Comparison of durations for V, Vd, Vl, and Vld rimes for low vowels and diphthongs revealed systematic duration differences attributable to proposed mora count. The CV and CVd cases, both argued to be bimoraic, are closely parallel in duration. However there is a systematic difference for the CVl and CVld cases, argued to be bimoraic for low vowels and trimoraic for diphthongs. We account for these results by integrating the assignment of duration to moras and segments.
Phonological Weakness in English, 2009
2007 IEEE International Conference on Acoustics, Speech and Signal Processing - ICASSP '07, 2007
Journal of the International Phonetic Association, 2002
The Journal of the Acoustical Society of America, 2000
Journal of the International Phonetic Association, 2002
The Journal of the …, Jan 1, 2006
ABSTRACT This study investigates the manual labeling of speech, and in particular conversational ... more ABSTRACT This study investigates the manual labeling of speech, and in particular conversational speech, at the articulatory feature level. A detailed transcription, including subtleties such as overlapping or reduced gestures, is useful for studying the great pronunciation variability in conversational speech. This type of labeling also facilitates the testing of automatic feature classifiers, such as those used in articulatory approaches to automatic speech recognition. For this study, approximately 100 utterances drawn from the switchboarddatabase have been transcribed using eight articulatory tiers rather than the traditional single phonetic tier. The tiers include: place and degree for up to two constrictions, nasality, glottal state, lip rounding, and vowel quality. Two transcribers have labeled this set of utterances in a multipass strategy, allowing for correction of errors. Preliminary analysis shows a high degree of intertranscriber agreement. Further analysis of the data is being performed to address a number of questions, such as: How quickly and reliably can this type of transcription be done? What are its advantages and disadvantages relative to purely phone‐based transcription? What characteristics of the utterances correspond to high or low transcriber agreement? What can be learned from the data regarding articulatory phenomena such as reduction and asynchrony?
cf.linguistlist.org
Abstract: Consonant weakening, or lenition, is comprehensively examined in this dissertation whic... more Abstract: Consonant weakening, or lenition, is comprehensively examined in this dissertation which consists of acoustic and articulatory phonetic studies of consonant strength, as well as a cross-linguistic survey of phonological patterns of lenition and ...
The term lenition is often used to describe sound changes, both historical and synchronic,<br&... more The term lenition is often used to describe sound changes, both historical and synchronic,<br> but the range and limits of lenition have not been well-defined. This report presents the results emerging from a data base of consonant strength alternations from over 165 languages. The data base represents a significant advance in the study of consonant strength by providing specific examples in sufficient quantity to compare types and frequency of alternations across languages. These generalizations demonstrate that common notions of lenition, rooted in historical change, over-regularize the phenomenon. The logically possible types of lenition and fortition are not evenly distributed; some are overwhelmingly common and others practically non-existent. The environments of alternations are often overlooked but crucial in determining if an alternation is lenition or fortition. The data base shows that consonant strength behavior exhibits asymmetries when compared to standard assumpt...
Yindjibarndi, a Ngarluma language spoken in northwestern Australia, displays an intriguing patter... more Yindjibarndi, a Ngarluma language spoken in northwestern Australia, displays an intriguing pattern of stress in which all long vowels may be pronounced as a sequence of two short vowels. Previous analyses (Wordick 1982, Kager 1993) have treated this as vowel breaking, but I show that the language's morphology, historical and synchronic phonology, and metrical structure support an analysis in which identical adjacent short vowels coalesce into a single long vowel, but only where they can serve as the head of a foot. I construct an Optimality Theoretic account of these facts, drawing on numerous well-established constraint types including correspondence constraints (McCarthy & Prince 1995), positional faithfulness constraints (Beckman 1997), prosodic constraints (McCarthy & Prince 1993), and a version of the Weight to Stress Principle (Prince 1990). Additionally, I motivate a sociolinguistic constraint, *RUDE, which forces consonant lenition. In addition to accounting for the inte...
We present an approach for the manual labeling of speech at the articulatory feature level, and a... more We present an approach for the manual labeling of speech at the articulatory feature level, and a new set of labeled conversational speech collected using this approach. A detailed transcription, including overlapping or reduced gestures, is useful for studying the great pronunciation variability in conversational speech. It also facilitates the testing of feature classifiers, such as those used in articulatory approaches to automatic speech recognition. We describe an effort to transcribe a small set of utterances drawn from the Switchboard database using eight articulatory tiers. Two transcribers have labeled these utterances in a multi-pass strategy, allowing for correction of errors. We describe the data collection methods and analyze the data to determine how quickly and reliably this type of transcription can be done. Finally, we demonstrate one use of the new data set by testing a set of multilayer perceptron feature classifiers against both the manual labels and forced align...
Consonants exhibit more variation in their phonetic realization than is typically acknowledged, b... more Consonants exhibit more variation in their phonetic realization than is typically acknowledged, but that variation is linguistically constrained. Acoustic analysis of both read and spontaneous speech reveals that consonants are not necessarily realized with the manner of articulation they would have in careful citation form. Although the variation is wider than one would imagine, it is limited by the phoneme inventory. The phoneme inventory of the language restricts the range of variation to protect the system of phonemic contrast. That is, consonants may stray phonetically into unfilled areas of the language's sound space. Listeners are seldom consciously aware of the consonant variation, and perceive the consonants phonemically as in their citation forms. A better understanding of surface phonetic consonant variation can help make predictions in theoretical domains and advances in applied domains. 1 Consonant variation The consonant variation of interest here goes beyond well...
� Motivations: A detailed transcription, including overlapping/reduced gestures • is useful for s... more � Motivations: A detailed transcription, including overlapping/reduced gestures • is useful for studying pronunciation variation • facilitates the testing of automatic articulatory feature classifiers � Alternatives: • Narrow phonetic transcription (e.g.
We investigate monosyllabic words with rimes consisting of a diphthong or non-low tense vowel fol... more We investigate monosyllabic words with rimes consisting of a diphthong or non-low tense vowel followed by a liquid, such as file, foul, foil, feel, fool, fail; fire, flour and foyer, which we term sesquisyllables. Evidence from phonological distribution, speaker intuition, metrical properties, variant pronunciations, and an acoustic study converges on the interpretation that these are trimoraic monosyllables. Comparison of durations for V, Vd, Vl, and Vld rimes for low vowels and diphthongs revealed systematic duration differences attributable to proposed mora count. The CV and CVd cases, both argued to be bimoraic, are closely parallel in duration. However there is a systematic difference for the CVl and CVld cases, argued to be bimoraic for low vowels and trimoraic for diphthongs. We account for these results by integrating the assignment of duration to moras and segments.
Phonological Weakness in English, 2009
2007 IEEE International Conference on Acoustics, Speech and Signal Processing - ICASSP '07, 2007
Journal of the International Phonetic Association, 2002
The Journal of the Acoustical Society of America, 2000
Journal of the International Phonetic Association, 2002
The Journal of the …, Jan 1, 2006
ABSTRACT This study investigates the manual labeling of speech, and in particular conversational ... more ABSTRACT This study investigates the manual labeling of speech, and in particular conversational speech, at the articulatory feature level. A detailed transcription, including subtleties such as overlapping or reduced gestures, is useful for studying the great pronunciation variability in conversational speech. This type of labeling also facilitates the testing of automatic feature classifiers, such as those used in articulatory approaches to automatic speech recognition. For this study, approximately 100 utterances drawn from the switchboarddatabase have been transcribed using eight articulatory tiers rather than the traditional single phonetic tier. The tiers include: place and degree for up to two constrictions, nasality, glottal state, lip rounding, and vowel quality. Two transcribers have labeled this set of utterances in a multipass strategy, allowing for correction of errors. Preliminary analysis shows a high degree of intertranscriber agreement. Further analysis of the data is being performed to address a number of questions, such as: How quickly and reliably can this type of transcription be done? What are its advantages and disadvantages relative to purely phone‐based transcription? What characteristics of the utterances correspond to high or low transcriber agreement? What can be learned from the data regarding articulatory phenomena such as reduction and asynchrony?
cf.linguistlist.org
Abstract: Consonant weakening, or lenition, is comprehensively examined in this dissertation whic... more Abstract: Consonant weakening, or lenition, is comprehensively examined in this dissertation which consists of acoustic and articulatory phonetic studies of consonant strength, as well as a cross-linguistic survey of phonological patterns of lenition and ...