Prosodic Level Research Articles

De l'ambiguïté des segments averbaux en français parlé

Verbless segments can play many different roles. These roles range from working as a sentential constituent to being a predicate. Yet, their analysis in spoken French is not always straightforward. In this article, we shall study two cases when the syntactic interpretation of verbless segments in spoken French is ambiguous: (i) verbless segments which can be given different statuses. (ii) complements which can be postponed, verbless constituents which partake of the three following categories: integrated elements, detached elements and autonomous elements. We shall assess the efficiency of the various means used to disambiguate the interpretation of these segments, i.e. the use of the linguistic context, of the prosodic and semantic levels, of word order, etc.

Boundary- and prominence-related lengthening and their interaction.

The durational effects of prosodic boundaries and prosodic prominence are well known, but their interaction is less well-understood. Recent studies in English [Turk and Shattuck-Hufnagel (2007); Byrd and Riggs (2008)] indicate that the two effects might be interdependent. Two acoustic experiments are presented, examining the role of boundary- and prominence-related lengthening and their interaction in Greek. The first experiment explores the effects of boundary-adjacent lengthening (conditions: no boundary, intermediate phrase, intonational phrase) and prominence (on the first, second, and third syllable away from the boundary). The second experiment examines these same effects, but post-boundary. Data from eight speakers were collected. The results from five speakers are presented. Results show that pre-boundary, both prominence and boundary have an effect on segment duration. Two prosodic levels (the level of word and one higher level) are distinguished. Furthermore, for two speakers, there is an interaction effect, such that three prosodic levels are distinguished when the prominent syllable is boundary-adjacent. The post-boundary condition shows less systematic results, with different speakers exhibiting different patterns. These results suggest that prosodic effects are speaker dependent, more stable pre-boundary, and that prominence and boundaries interact, increasing prosodic boundary marking. Implications for models of prosodic structure are discussed.

Portuguese converted deverbal nouns: Constraints on their bases

The aim of this paper is to analyse the constraints that govern the formation of Portuguese converted deverbal nouns (‘CDNs’). To this end, we have analysed a total of 1323 converted deverbal nouns. The first type of constraints investigated are those found at the prosodic level: these deal with the configurational formatting of the minimal root. We will focus on the relationship between these constraints and the semantic and phonological identification of the lexical item in perceptual terms. The second level of constraints may be characterised as morphophonological-pragmatic. Many verbs contain morphological units that block the formation of CDNs because their [+Latinate] character has a certain stylistic effect in terms of pragmatic usage. The third level of constraints pertains to lexical-conceptual structure combined with argument structure.We will show that (i) the constraints on the bases of CDNs are not exclusively founded on morphological incompatibilities between derivational affixes and other morphological units with which they join. Thus, a derivational mechanism that does not involve affixation, such as conversion, is also ruled by restrictions on the bases. (ii) Structural constraints – based on argument structures and on prosodic, morphological and lexical-semantic structures – are interwoven with processing and pragmatic conditions. (iii) Word formation is organised in structures that have interfaces with each other.

Children with specific language impairment: The role of prosodic processes in explaining difficulties in processing syntactic information

Children with specific language impairment (SLI) experience great difficulties in language comprehension and/or production whereby the majority of these children have particular problems in acquiring syntactic rules. In the speech stream boundaries of major syntactic constituents are reliably marked by prosodic cues. Therefore, prosodic information provides an important cue for discovering the syntactic structure of a language [Jusczyk, P.W., 2002. How infants adapt speech-processing capacities to native language structure. Curr. Dir. Psychol. Sci. 11, 15–18.]. Following this, the question is, whether children with SLI differ in the processing of syntactic information from normally developing children and to what extent this is related to the processing of the inherent prosodic information. Children heard either correct sentences or sentences with a word category violation (syntactic level) and a joined prosodic incongruity (prosodic level) while event-related brain potentials (ERPs) were recorded. Judging the sentence's correctness, control children performed better than children with SLI for all types of sentences. With respect to the ERPs, control children showed a bilateral early starting anterior negativity sustaining into a late anterior negativity and a P600 in posterior regions in response to incorrect sentences. Children with SLI showed a comparable P600 but unlike the control children there was only a late, clearly left lateralized anterior negativity. The complete absence of a right anterior negativity in children with SLI suggests that they may not access prosodic information in the same way normal children do. The differences in prosodic processing may in turn hamper the development of syntactic processing skills as indicated by the absence of the syntax-related early left anterior negativity.

Is text-to-speech synthesis ready for use in computer-assisted language learning?

Text-to-speech (TTS) synthesis, the generation of speech from text input, offers another means of providing spoken language input to learners in Computer-Assisted Language Learning (CALL) environments. Indeed, many potential benefits (ease of creation and editing of speech models, generation of speech models and feedback on demand, etc.) and uses (talking dictionaries, talking texts, dictation, pronunciation training, dialogue partner, etc.) of TTS synthesis in CALL have been put forward. Yet, the use of TTS synthesis in CALL is not widely accepted and only a few applications have found their way onto the market. One potential reason for this is that TTS synthesis has not been adequately evaluated for this purpose. Previous evaluations of TTS synthesis for use in CALL, have only addressed the comprehensibility of TTS synthesis. Yet, CALL places demands on the comprehensibility, naturalness, accuracy, register and expressiveness of the output of TTS synthesis. In this paper, the aforementioned aspects of the quality of the output of four state-of-the-art French TTS synthesis systems are evaluated with respect to their use in the three different roles that TTS synthesis systems may assume within CALL applications, namely: (1) reading machine, (2) pronunciation model and (3) conversational partner [Handley, Z., Hamel, M.-J., 2005. Establishing a methodology for benchmarking speech synthesis for computer-assisted language learning (CALL). Language Learning and Technology Journal 9(3), 99–119. Retrieved from: http://llt.msu.edu/vol9num3/handley/default.html.]. The results of this evaluation suggest that the best TTS synthesis systems are ready for use in applications in which they ‘add value’ to CALL, i.e. exploit the unique capacity of TTS synthesis to generate speech models on demand. An example of such an application is a dialogue partner. In order to fully meet the requirements of CALL, further attention needs to be paid to accuracy and naturalness, in particular at the prosodic level, and expressiveness.

Licensing by Information Structure

n/a

Sentence recognition in native- and foreign-language multi-talker background noise

Studies of speech perception in various types of background noise have shown that noise with linguistic content affects listeners differently than nonlinguistic noise [e.g., Simpson, S. A., and Cooke, M. (2005). "Consonant identification in N-talker babble is a nonmonotonic function of N," J. Acoust. Soc. Am. 118, 2775-2778; Sperry, J. L., Wiley, T. L., and Chial, M. R. (1997). "Word recognition performance in various background competitors," J. Am. Acad. Audiol. 8, 71-80] but few studies of multi-talker babble have employed background babble in languages other than the target speech language. To determine whether the adverse effect of background speech is due to the linguistic content or to the acoustic characteristics of the speech masker, this study assessed speech-in-noise recognition when the language of the background noise was either the same or different from the language of the target speech. Replicating previous findings, results showed poorer English sentence recognition by native English listeners in six-talker babble than in two-talker babble, regardless of the language of the babble. In addition, our results showed that in two-talker babble, native English listeners were more adversely affected by English babble than by Mandarin Chinese babble. These findings demonstrate informational masking on sentence-in-noise recognition in the form of "linguistic interference." Whether this interference is at the lexical, sublexical, and/or prosodic levels of linguistic structure and whether it is modulated by the phonetic similarity between the target and noise languages remains to be determined.

Prosodic levelling during language shift: Okinawan approximations of Japanese pitch‐accent1

This article demonstrates how the language death process potentially contributes to the emergence of a new dialect. A sociolinguistic survey of the prosodic phonology of Japanese spoken in Yaeyama, Okinawa shows that the obsolescence process of Okinawan has resulted in a new dialect of Japanese. First, Okinawan is undergoing processes typical of language death, including increased variability and the elimination of phonological contrasts at the prosodic level. These processes are also occurring in the incoming language, Japanese. Second, these processes have influenced the acquisition of the prosody of Standard Japanese, resulting in a new prosodic system. Third, a positive attitude towards the local variant of Japanese is prevalent among the younger members of the community, particularly the working‐class males. These results suggest that a local prosodic system is emerging, resulting in a new dialect of Japanese.

Analysis of a case of the foreign accent syndrome in terms of the framework of gestural phonology

Brain damage which results in aphasia or apraxia of speech frequently causes deviant articulation; and in a limited number of cases the patients’ speech takes on characteristics normally associated with a dialect that is not their own, or it resembles the performance of a non-native speaker of the language. This condition is referred to as the foreign accent syndrome (FAS). FAS has been classified as an apraxic condition, a motor speech disorder in which the capacity to program the positioning of speech musculature and the sequencing of muscle movements for the volitional production of speech, has been disturbed. The present study is an analysis of the speech of a Norwegian FAS patient who after having suffered a stroke, developed what sounded like an English accent when speaking Norwegian. Her speech exhibited deviant features both on the segmental and on the prosodic level: deviant vowel and consonant articulations, deviant rhythmical patterns and deviant pitch patterns. This complex pattern of deviant features can be accounted for in terms of dysfunction of specific components in a gestural phonological model. The deviations can all be seen as the result of abnormal scaling or abnormal phasing of gestures.

Mapping of prosodic structure onto words and phrases in children's and adults' speech production

Direct recordings of lip and jaw movement were obtained to investigate how children and adults differentiate prosodic structure both within iambic and trochaic words and across word boundaries. It was hypothesised that children, especially those with less language skills, would be particularly likely to produce strong-weak (or trochaic) sequences. This experiment examined children's (both normally developing and specifically language impaired, aged 4–6 years) and adults’ productions of prosodic sequences that were controlled for syntactic and phonetic content, but that differed in rhythmic alternations, both within words (i. e., iambs vs. trochees) and across words (i. e., optimal strong-weak vs. stress gap or stress clash). To avoid final position effects, target words were in utterance internal position. Analyses included kinematic measures of prosodic structure (i. e., amplitude and duration of movements in weak vs. strong syllables; ratios of weak-to-strong movements). Consistent with prior work, weak syllables in iambs were produced differentially from those in trochees; though these effects were now observed in utterance position. However, in noun-noun sequences within a phrase, even when the rhythm rule was violated, prosodic foot structure was not reorganised as a function of rhythmic properties of the utterance. Overall, children with the least mature language skills demonstrated poor control of temporal and amplitude aspects of movement, especially in weak syllables. Findings are related to the organisation of speech production units across prosodic levels.

The feature [stiff] interacts with intonation to affect vocal-fold vibration characteristics

As part of a larger study of the effect of prosody on segmental cues, previous work has shown that in a high pitch environment, F0 is significantly increased relative to a baseline following voiceless obstruents, but F0 closely follows the baseline following voiced obstruents. When a syllable carries a low or no pitch accent, F0 is increased only slightly following all obstruents. It is suggested that this difference occurs because demands of the segmental and prosodic levels of speech production conflict. In particular, results support a theory that the primary feature of voiced or voiceless obstruents is [.-stiff] or [+stiff] vocal folds, respectively. Enhancing features such as [-.spread] or [+spread] vocal folds are secondary. In a high pitch environment, the feature [-.stiff] conflicts with the need to raise pitch. Because of enhancing gestures, the prosodic level can override the segmental. Likewise, in a low pitch environment, the feature [+stiff] conflicts, so the vocal folds will not be stiffened for voiceless obstruents. If so, one might expect that stop consonants in low pitch environments will have longer or stronger prevoicing than in high pitch environments, and preliminary data show that speakers do show this tendency. [Work supported by NIH Grant DC04331.]

Mètre et rythmes du vers classique: Corneille et Racine by Valérie Beaudouin (review)

MLRy 99.2, 2004 489 The whole enterprise is characterized by the thoroughness of its scholarship. The texts are meticulously edited with close reproduction ofseventeenth-century spellings and punctuation. Unfamiliar words and obscure references are explained in footnotes, always well judged in their scope. A detailed account ofthe royal tours ofthe provinces between the years 1615 and 1660, a chronological classification of all the entries inventoried , and a comprehensive bibliography complete this well-illustrated volume. Notwithstanding the decision only to reproduce extracts, this is a useful book which offersvaluable insights into political ritual in the Ancien Regime, and should provide the stimulus for furtherresearch in the field. Goldsmiths, University of London Marie-Claude Canova-Green Metre et rythmes du vers classique: Corneille et Racine. By Valerie Beaudouin. (LettresNumeriques, ii) Paris: Champion. 2002. 620pp. ?80. ISBN 2-74530509 -3. As Valerie Beaudouin observes in her introduction, previous formal studies ofthe clas? sical alexandrine have been limited to corpora of no more than 1,000 lines. With this, the published version of her doctoral thesis (2000), supervised by Jacques Roubaud, Beaudouin studies the complete dramatic works of Corneille and Racine, a corpus of 77,000 lines. An empirical study of such unprecedented magnitude is possible thanks to the metrometre,a computer program which is as much Beaudouin's focus as le vers itself. Following a useful overview of differentrhythmic theories and analytical approaches to the alexandrine (Chapter 1), Chapters 4-6 deal at considerable length with the 'mecanisation des regles', the painstaking task of creating the metrometre itself. Although this concentration on technical matters teaches the informed reader little new about fundamental versificatory rules, the results explored elsewhere give many fresh insights into the alexandrine's essential rhythmic characteristics. Beaudouin effectivelyproves that the alexandrine represents the concatenation of two hexasyllabic metrical segments, almost identical on many vocalic, syntactic, and prosodic levels. Chapter 7 provides extensive evidence ofthe similarities between each hemistich, clearly presented in numerous graphs charting accurately and exhaustively the most subtle rhythmic details. For example, Beaudouin demonstrates that certain vowel combinations are more 'rhythmic' than others, with noyaux vocaliques of sev? eral graphemes often closing the hemistich. Giving the dominant accentual patterns 001001 and 010101 as the two most regular structures, she also highlights a tendency to make the second hemistich more regular than the first. This is particularly in? teresting when Racine and Corneille are compared with nineteenth-century poets, confirming many of our suspicions with empirical evidence, such as Baudelaire's fondness for a 3/3 structure, and Rimbaud's sustained accentuation of syllables 1,5, 7, and 11. Chapter 3 argues that there can be no absolute definition of rhyme quality, suggesting instead a contextual, relative nomenclature according to the frequency of a rhyme unit (rimeme) in the corpus. Thus, for Beaudouin, the rime suffisante'desir/ soupir' is merely pauvre, and although such a system raises questions as to its practical implementation, or its relevance in terms of aural rhyme harmony, the 170-page ap? pendix catalogues all rimemessystematically in order to allow useful future research. For readers with only a passing interest in versificatoryminutiae, Chapter 8 demon? strates the value of the metrometrebeyond the prosodic. Beaudouin investigates the relationship between verse rhythm and semantic content, and highlights two princi? pal semantic fields: love and death. Within these, eight subcategories are charted to provide a precise overview of each play's thematic texture, and several close read? ings explore the relationship between the rhythm and content of individual lines. 49? Reviews Furthermore, the 'love' field is characterized by a greater frequency of irregular lines, with the 'death' field, more regular, representing the calm anticipated afterthe turbulent passions of tragedy. Occasionally, Beaudouin's thematic interpretations are unconvincing, with only perfunctory comment on some surprising examples of trans? gressive versification, and a rather simplistic view ofthe characteristics of tragedy and comedy. On the whole, however, this clearly presented work paves the way forsimilar statistical analyses, allowing us both to chart the history of alexandrine rhythm and to compare in exhaustive detail the works of other poets and playwrights on formal and thematic levels. University of Edinburgh David Evans Le Comte d'Essex. By Thomas Corneille. Ed. by Wendy Gibson...

Prosodic and melodic processing in adults and children. Behavioral and electrophysiologic approaches.

The results of a series of experiments aimed at directly comparing the prosodic level of processing in language with the melodic level of processing in music are reported. The first series of experiments was conducted on adults, musicians and nonmusicians, and the second one on 7- to 9-year-old musician and nonmusician children. However, as this last study is still in progress, only preliminary results will be presented. The theoretic framework within which these experiments are taking place is described. The first problem concerns the specificity of the perceptive and cognitive computations necessary to perceive and understand language. We argue that comparing language with music can provide interesting insights into this complex issue. The second problem is linked to the relationship between different types of learning. Does early musical training influence the way in which musicians process some aspects of language as prosody? These two problems are considered and the results of the experiments are described.

Prosodic domain-initial effects on the acoustic structure of vowels

In the process of language change, vowels tend to shift in ‘‘chains,’’ leading to reorganizations of entire vowel systems over time. A long research tradition has described such patterns, but little is understood about what factors motivate such shifts. Drawing data from changes in progress in American English dialects, the broad hypothesis is tested that changes in vowel systems are related to prosodic organization and stress patterns. Changes in vowels under greater prosodic prominence correlate directly with, and likely underlie, historical patterns of shift. This study examines acoustic characteristics of vowels at initial edges of prosodic domains [Fougeron and Keating, J. Acoust. Soc. Am. 101, 3728–3740 (1997)]. The investigation is restricted to three distinct prosodic levels: utterance (sentence-initial), phonological phrase (strong branch of a foot), and syllable (weak branch of a foot). The predicted changes in vowels /e/ and /ε/ in two American English dialects (from Ohio and Wisconsin) are examined along a set of acoustic parameters: duration, formant frequencies (including dynamic changes over time), and fundamental frequency (F0). In addition to traditional methodology which elicits list-like intonation, a design is adapted to examine prosodic patterns in more typical sentence intonations. [Work partially supported by NIDCD R03 DC005560-01.]

Bootstrapping in the acquisition of word stress in Brazilian Portuguese

This paper deals with the acquisition of word stress in Brazilian Portuguese. In the course of acquisition, children use several strategies to mark stress prominence before the adult algorithm of primary stress is used productively. I argue that the prominences found in children´s early utterances do not reflect word stress but prominences of a higher prosodic level. In other words, children use the stress information available in higher prosodic domains as a cue for the acquisition of the algorithm of primary stress.

Planning of syllables in children with developmental apraxia of speech

The aim of the present study was to investigate whether children with developmental apraxia of speech (DAS) show a deficit in planning syllables in speech production. Six children with DAS and six normally speaking (NS) children produced high- and low-frequency of occurrence syllable utterances, in which the syllable structure was systematically manipulated in an otherwise unchanging phoneme sequence. Anticipatory coarticulation, using second formant trajectories, and durational structure were analysed. The results showed stronger coarticulation in the children with DAS when compared to the normally speaking children, but in contrast to our expectations, in neither group was a systematic effect of syllable structure on the second format trajectory found. Effects of syllable structure did emerge for durational structure in that durational adjustments were found in the segments of the second syllable. These adjustments were less systematic in children with DAS when compared to normally speaking children. Furthermore, at the prosodic level, normally speaking children showed metrical contrasts that were not realized by the children with DAS. The latter results are interpreted as evidence for a problem in the planning of syllables in speech production of children with DAS, in particular concerning prosodic aspects, which is discussed in relation to the automation of speech production.

P(ragmatic)-trees. Coherence, intentionality and cognitive content in discourse and non-verbal action

Abstract The starting point of this study is the adoption of the simple device of a tree structure as a working model for the analysis of pragmatic cohesion relations between successive actions of both discourse and non-verbal behavior. It is claimed that (i) cohesion relations, as represented by the nodes of such a P(ragmatic)-tree, are in the case of discourse classifiable into a limited number of recursively applicable types, and that (ii) this pragmatic structure, representing the intentionality of behavior, constitutes a level that underlies the more superficial syntactic and prosodic levels of discourse (‘pragmatics first’). The aims of this paper are: (1) to address a number of technical issues concerning the application of the tree device to spontaneous speech, concentrating on traditional preoccupations of the interpretative approach to discourse analysis, such as thematization, intonation and the scope of discourse markers and topics; (2) to investigate the implications of the ‘pragmatics first’ claim and of assuming a fundamental analogy between verbal and non-verbal actions for such theoretical issues as the relations between pragmatic structure and (semantic) content, and the dynamics of cognition in general and discourse in particular.

Prosodic Structure and Perception of Korean Domain-initial Coronal Stops

Abstract The perception of Korean's three-way voiceless stop contrast depends on the combination of a number of acoustic properties including closure duration, f 0, VOT and spectral slope. Production studies indicate that these same properties systematically vary according to a stop's position in the hierarchical prosodic structure of an utterance. This study investigated interactions between prosodic structure and the recognition of tense, lax and aspirated stops by using stimuli that contained stops produced at one position in the prosodic hierarchy cross-spliced into utterances at a different prosodic level. The results of the experiment showed that recognition of tense and aspirated stops remained high regardless of the degree of mismatch between the underlying prosodic position of a segment and the domain into which it was inserted. Misrecognition rates were higher for cross-spliced lax stops, and underlying Utterance and Intonation Phrase initial lax stops spliced into the syllable level were perceived as aspirated by a majority of listeners. These findings are interpreted in terms of Cho & Keating's (2001) study of domain-initial strengthening and the acoustic overlap that is observed between stop manner and prosodic position. These results are taken as preliminary evidence that could support Cho & Keating's (2001) suggestion that prosodically conditioned acoustic differences aid listeners in discerning intended phrasings.

Articulatory properties of initial segments in several prosodic constituents in French

This study reports on how what is usually called “segmental” articulation may be affected by prosodic structure. Articulatory properties of segments of various types are compared in initial position of four prosodic constituents in French: an Intonational Phrase, an Accentual Phrase, a Word and a Syllable. Modification of linguopalatal contact, nasal flow, acoustic measures of glottal articulation, as well as durational properties of segments are examined depending on these prosodic positions. Results show that the articulation of initial segments varies according to the prosodic level of the constituent. Initial segments in higher prosodic domains tend to have more linguopalatal contact, somewhat less nasal flow for nasals, and vowels tend to be more frequently glottalized than in initial position in lower domains. The magnitude and the realization of the articulatory modification observed in these positions is found to depend on speakers, articulators, and segment types. However, the regular trend observed suggests that these articulatory properties can reflect the prosodic encoding of constituents of various levels within utterances.

Discriminating between constructivist and nativist positions: fillers as evidence of generalization

In her note, Ann Peters distinguishes between two types of fillers: (1) premorphological, whose presence in the utterance is motivated by purely phonological considerations (the child attempts to produce an adult utterance at the prosodic level, or approximates the prevalent rhythmic pattern in the language); and (2) protomorphological, which function as place-holders for grammatical morphemes and eventually differentiate into various grammatical morphemes. I would like to suggest that there may also be a third type which sometimes appears when the child begins to generalize over a set of related words. Just as protomorphological fillers can provide valuable information about how children acquire function words, these ‘generalized’ fillers may offer insights into how children form categories of function words. I will illustrate this phenomenon with data from the Naomi corpus in the CHILDES database (Sachs, 1983).