Workshop on Prosodic Development
Barcelona | April 16th, 2010
Workshop on Prosodic Development
Barcelona | April 16th, 2010
The Workshop on Prosodic Development is hosted by the Universitat Pompeu Fabra on April 16th, 2010.
The main goal of this one-day workshop is to discuss different aspects of the children’s prosodic development in different languages. The project members will present the results of two recent projects related to “Rhythmic and Intonational Development in Catalan, English, and Spanish”, financed both by the Generalitat de Catalunya (Batista i Roca projects) and by the BritishAcademy. Invited speakers will present their work on a variety of aspects related to phonological development.
Discussions will address specific questions regarding the influence of language-specific distributional and frequency properties on language development and also the influence of general production and perception constraints. One of the goals will be to try to bridge the gap between perception and production studies in prosodic development.
We are grateful to the UPF's Departament de Traducció i Ciències del Llenguatge for their financial support of this workshop.
program| index
page09:00–09:30 / WELCOME – RECEPTION
09:30–10:15 / On the acquisition of word stress and Intonation:
Evidence from early perception
Paula Fikkert & Aoju Chen
Radboud University & Max Planck Institute / 7
10:15–11:00 / Stress pattern preference in Spanish-learning infants:
The importance of syllable weight
Ferran Pons & Laura Bosch
Universitat de Barcelona / 9
11:00–11:30 / COFFEE BREAK
11:30–12:15 / Prosodic structure in early child speech:
Evidence from intonation, tempo, and coda production
SóniaFrota
University of Lisbon / 11
12:15–13:00 / How do German-Spanish bilingual children ask questions in their two languages?
ConxitaLleo & MartinRakow
HamburgUniversity / 13
13:00–14:30 / LUNCH
14:30–15:15 / Variation in intonational marking of topic and focus in children
Aoju Chen
Max Planck Institute / 15
PRESENTATION OF PROJECT RESULTS
15:15–15:45 / Speech rhythm as durational marking of prosodic heads and edges.
Evidence from Catalan, English, and Spanish
Pilar Prieto & Maria del Mar Vanrell
ICREA-Universitat Pompeu Fabra & Universitat Autònoma de Barcelona-Universitat Pompeu Fabra / 17
15:45–16:15 / Phonological factors in rhythmic development: A cross-linguistic study
BrechtjePost
CambridgeUniversity / 19
16:15–16:45 / Early acquisition of word-level prosody
Lluïsa Astruc
The Open University - CambridgeUniversity / 21
16:45–17:15 / Early intonational development in Catalan and Spanish
Maria del Mar Vanrell & Pilar Prieto
Universitat Autònoma de Barcelona-Universitat Pompeu Fabra & ICREA-Universitat Pompeu Fabra / 23
17:15–17:45 / Prosodic manipulation in child-directed speech: a cross-linguistic study
ElinorPayne
University of Oxford / 25
17:45–18:15 / COFFEE BREAK
18:15–19:00 / Rhythm cues and language discrimination in infancy: a review
Laura Bosch
Universitat de Barcelona / 27
On the acquisition of word-stress and Intonation: Evidence from early perception
Paula Fikkert1 Aoju Chen2
1RadboudUniversityNijmegen, 2Max Plank Institute
Background and aims. Dutch children’s early disyllabic words are typically trochaic. Their first two-word utterances are declaratives with H*L pitch accent. Acoustically, both patterns are very similar. However, before their first birthday infants can segmentize both trochaic and iambic words (Jusczyk et al., 1999) and are sensitive to intonation. This paper aims to disentangle children’s developing perceptual knowledge of word-stress and intonation.
Methods. We tested fifteen 14- and sixteen 24-month-old infants in a mispronunciation paradigm. They saw pairs of pictures with a distractor word and an iambic or trochaic target word (ballon ‘balloon’ or konijn ‘rabbit’; or schommel ‘seesaw’ en varken ‘pig’). The target words were produced with correct or incorrect word-stress in a carrier sentence that required a question intonation (L*H H%) (Zie je het konijn? ‘Do you see the rabbit?’) or in a carrier sentence that required a statement intonation (H*L). The intonation was either appropriate for the carrier sentence, or inappropriate.
Results. The results for iambic targets in question sentences show that both word stress and intonation play a role. Word stress is less important for the 24-month-olds, but intonation remains important for word recognition. The results of the other conditions are currently being analysed.
Conclusion. Children develop knowledge of intonation and word-stress independently, and show knowledge of both at 24 months.
Stress pattern preference in Spanish-learning infants: The importance of syllable weight
Ferran Pons & Laura Bosch
Universitat de Barcelona
Infants’ sensitivity toward prosodic information of the speech input is already present early in life (i.e., Mehler et al., 1988). Attunement of early sensitivities to the specific prosodic characteristics of the input can be identified by emergent attention biases observed in infants’ prosodic preferences. Studies with English- and German-learning infants observed a bias toward trochaic words (the most frequent lexical stress pattern in these languages), but a trochaic preference could no be found in French-learning infants (Jusczyk, Cutler, & Redanz, 1993; Höhle, Bijeljac-Babic, Herold, Weissenborn, & Nazzi, 2009). An explanation for this bias is thus attributed to the language infants are exposed to. Here we present results from studies with Spanish-learning infants that challenge the proposal of a general or universal trochaic bias and add more evidence to the importance of the speech input characteristics infants are exposed to. Spanish differs from English in its general rhythmic pattern, predominant lexical types, lexical stress assignment rules, and the distribution of trochaic and iambic words in the lexicon. Spanish also differs from French in its lexical stress properties, but these two languages share the syllable as the basic rhythmic unit.
We first explored 9-month-old Spanish-learning infants’ lexical stress preferential patterns and no clear trochaic preference could be found (Pons & Bosch, 2007). To further analyze stress preference patterns we focused on syllable weight, a factor that plays a role in stress assignment in Spanish and that can be related to the emergence of lexical stress preferences (Pons & Bosch, 2010). Our results revealed that in Spanish syllable weight turned out to be a crucial aspect for a lexical stress preference to be observed in infants (either trochaic or iambic). These new results broaden our current knowledge on the factors that play a role in the emergence of a preference bias for the native language pattern of word stress.
Our results would suggest that Spanish-learning infants’ preference patterns are not based on general rhythmic properties of the most frequent words in their language (trochaic disyllables), but on more specific knowledge that involves linking stress information to syllable structure
References
Höhle, B., Bijeljac-Babic, R., Herold, B., Weissenborn, J., & Nazzi, T. (2009). The development of language specific prosodic preferences during the first year of life: Evidence from German and French. Infant Behavior and Development, 32, 262–274.
Jusczyk, P. W., Cutler, A., & Redanz, N. (1993). Preference for the predominant stress pattern of English words. Child Development, 64, 675–687.
Mehler, J., Jusczyk, P., Lambertz, G., Halsted, N., Bertoncini, J., & Amiel-Tison, C. (1988). A precursor of language acquisition in young infants. Cognition, 29, 143–178.
Pons, F. & Bosch, L. (2007). The perception of stress patterns by Spanish and Catalan infants. In: P. Prieto, J. Mascaró, & M. J. Solé (Eds.). Segmental and prosodic issues in (Romance) Linguistics.Current Issues in Linguistic Theory. John Benjamins, Amsterdam (pp. 199-218)
Pons, F. & Bosch, L. (to appear in 2010). Stress pattern preference in Spanish-learning infants: the role of syllable weight, Infancy, 15(3).
Prosodic structure in early child speech:
Evidence from intonation, tempo and coda production
Sónia Frota
University of Lisbon
Recent studies on prosodic development in European Portuguese (EP) have independently gathered production data on intonational development, word stress, and duration patterns at segmental, syllabic and phrasal levels, in early child speech (Frota & Vigário 2008a,b, Frota & Matos 2009, Matos 2010, Frota et al. in progress). Intonational patterns were analysed, namely the inventory of pitch accents and boundary tones and their realizations, together with the patterns of pitch accent distribution in one-word and multiword phrases. Word stress was considered taking into account pitch accent type and the presence/absence of a pitch accent in the relevant syllable, as well as the syllable position in the utterance or intonational phrase (IP). The analysis of duration patterns included the duration of intervocalic stop consonants, the duration of syllables as a function of word stress, position in the prosodic word (PW) and position in the IP, and the duration of syllables and PWs as a function of the number of syllables within a given domain, namely the PW and the IP. The empirical database consists of the speech of one EP child, from 1;00 to 2;04. In the present talk, these data will be described and discussed in an integrated way with the goal of answering to the following question: what does intonation and tempo tell us about the development of prosodic structure? In particular, we will put into test the path of development in three stages suggested in Frota & Vigário (2008b). A related question also to be addressed is whether the child uses the acquired prosodic structure to constrain the development of other aspects of phonology. Coda development data of the same child will be shown to provide evidence in this direction (Jordão 2009, Jordão & Frota 2009).
In initial productions (until 1;04), disyllabic targets, if uttered with two syllables, tend to be produced with one pitch accent per syllable (usually the falling accent), and level stress predominates. The intervocalic stop consonant in these productions is realized with a duration longer than its counterpart in productions showing just one pitch accent and similar to the duration of the stop interval between two target words (Demuth & McCullough 2009 also use interval duration as evidence for PW status). PW duration, in this period, and unlike in later production, is not correlated with number of syllables per PW. Position in the PW or the IP is irrelevant for syllable duration. These findings strongly suggest that the syllable is being treated as a PW and a prosodic phrase. After 1;04, targets are uttered with one accent and patterns of stress shift predominate over level stress, showing an interaction with pitch accent type (which may be explained along the lines of de Lacy 2002, 2007, Yip 2007). These facts indicate that syllables coincide no longer with PWs. However, the relation between number of pitch accents an number of PWs is very close, and the position of the syllable in the IP is still irrelevant for syllable duration, suggesting that each PW is treated as a prosodic phrase. At 1;09 a temporal reorganization occurs at phrase level, as shown by the U-shaped correlation pattern between syllable duration and number of syllables in the IP. After this reorganization, syllables that are prosodically final become longer, or in other words, final lengthening emerges. Further, the close relation between number of pitch accents and number of PWs is broken, with many more PWs being unaccented (matching the pitch accent distribution of the language). These facts indicate that PWs and prosodic phrases are now treated differently. This same difference is shown by coda development data: codas start to be treated at IP prominent and IP final positions, and not at PW prominent or final positions.
Taken together, these findings strongly suggest a path of development in the acquisition of prosodic structure along the lines proposed in Frota & Vigário (2008b): an initial stage where there is a matching between syllable, PW, and phrase; a second stage where PW and phrase match; and a third stage where the three basic prosodic domains are fully independent (syllable≈PW≈phrase> syllable≠PW≈phrase> syllable≠PW≠phrase
How do German-Spanish bilingual children ask questions in their two languages?
Conxita Lleó & Martin Rakow
ResearchCenter on Multilingualism, University of Hamburg
From various interrogative sentences described in the literature, both in Spanish and German, we have focused on the most common ones in early child production, namely those seeking information. In both languages, information-seeking yes/no questions show a falling-rising contour and wh-questions have a rising-falling contour. Yes/no interrogative utterances in the target languages, German and Spanish, exhibit four comparable intonation points: a first peak, a falling F0, an L and a final H%. However, there are differences regarding the alignment and scaling of H and L tones across the two languages. Yes/no questions by German and Spanish monolinguals and German-Spanish bilinguals aged 2;0 and 3;0 are analyzed, with the aim of finding out whether children at these ages have already acquired the intonation contours of these questions. Results show that monolingual children already produce both the F0 contours as well as the alignment and scaling of the four crucial points before 3;0, whereas bilinguals correctly produce the contours, but some of them tend to not differentiate alignment and/or scaling in the two languages. Information-seeking wh-questions are also characterized by four comparable points in both languages: a first peak, a falling F0, an L and a final L%, which again show differences of alignment and scaling across the two languages. Results based on the production of information-seeking wh-questions by the same children are comparable: whereas monolingual children produce not only the intonation points but also alignment and scaling target-like in the two languages, bilinguals tend to produce indistinguishable alignment and scaling in their two languages. Some theoretical consequences of these results are discussed, with the aim of trying to sort out those constructs that because of their added semantic import must be characterized as phonemic, from the phonetic ones, without added semantic import, which may lead to later acquisition, especially in the bilingual condition.
Variation in intonational marking of topic and focus in children
Aoju Chen
Max Planck Institute
It has long been recognised that children acquire different aspects of language (e.g. Mean Length of Utterance, vocabulary size, vocabulary comprehension, syntactic complexity) in different rates. Further, children can also differ in the development stages or route of development (Wells 1986). In this talk, I will consider variations in intonational realisation of topic and focus in Dutch 4- to 5-year-olds. Specifically, I will discuss the differences between a group of 4- to 5-year-olds with a ‘playful’ manner of speaking (characterised with a wide pitch span, more pitch dynamics, intra-speakers variations in voice quality) and a group of 4- to 5-year-olds with a more consistent and everyday speaking style. In the light of these data, I will present hypotheses on differences in both rate and route of development regarding intonational realisation of topic and focus and suggest methods addressing these hypotheses.
References
Wells, G. (1986). Variation in child language. In P. Fletcher & M. Garman (eds.), Language acquisition: studies in first language development (2nd edition). Cambridge: CambridgeUniversity Press.
Speech rhythm as durational marking of prosodic heads and edges.
Evidence from Catalan, English, and Spanish
Pilar Prieto1,2 Maria del Mar Vanrell3,2
1 ICREA, 2Universitat Pompeu Fabra, 3 Universitat Autònoma de Barcelona
One of the unsolved issues in the phonetic sciences is the quest for reliable acoustic correlates of perceived differences in linguistic rhythm present in the speech signal and which allow human (and certain animals) to distinguish languages according to rhythmic classes. One of the leading views on this issue is that the rhythm percept reflects language-specific phonological properties, which in turn are signaled by the acoustic/phonetic properties of speech. Various metrics based on variability in the duration of consonantal and vocalic intervals, and the comparative proportions of vocalic and consonantal intervals have partially succeeded in relating the durational properties of the speech signal with traditional rhythm types (Ramus et al. 1999, Grabe & Low 2002, Low, Grabe & Nolan 2000, Dellwo 2004, 2006, among others; for a review, see White & Mattys 2007a). These measures are known to be partially dependent on the syllabic structure types that are present in the language.
The goal of this study is to examine the extent to which the rhythmic differences perceived between three languages that are reported to belong to different rhythmic classes (English: ‘stress-timed’, Spanish: ‘syllable-timed’, Catalan: ‘intermediate’) are attributable to effects of syllable structure. Data from a total of 24 speakers reading 720 utterances from these languages clearly show that differences in the rhythmic metrics - in particular, the vocalic variability measures, nPVI-V, ΔV, and VarcoV - emerge even when syllable structure is controlled for in the experimental materials, at least between English on the one hand and Spanish/Catalan on the other. This strongly suggests that important differences in durational processes exist in these languages, and thus that the rhythmic percept is not solely dependent on the presence of these two phonological properties in a given language. The consonantal interval variability metrics ΔC and VarcoC are not able to discriminate between languages when syllable structure is controlled for, and would therefore appear purely to capture phonotactic differences, and any rhythmic effects resulting from these. Similarly, the interval measure %V is shown to be a much weaker indicator of rhythmic difference than vocalic variability, once syllable structure is factored out.
Further analyses of the data indicate that the rhythmic class distinctions under consideration finely correlate with differences in the way languages instantiate two prosodic timing processes, namely durational marking of prosodic heads, and prosodic boundary lengthening. These results are compatible with the findings and conclusions of a study by Fant, Kruckenberg & Nord 1991 on Swedish, English and French, and of a recent study of two varieties of Italian (White, Payne and Mattys, 2009), which advance the hypothesis that multiple factors contribute to the perception of rhythmic differences, including prosodic timing. Our analysis finds robust crosslinguistic support in favour of a prosody-based hypothesis of linguistic rhythm.
Phonological factors in rhythmic development: A cross-linguistic study
Brechtje Post
CambridgeUniversity
The characteristic rhythm of a language – traditionally referred to in terms of stress- and syllable-timing – has been claimed to emerge from various phonological properties, especially vowel reduction and syllable complexity (e.g. Bertinetto 1981, Dasher and Bolinger 1982, Roach 1982, Dauer 1983; cf. Prieto et al. submitted). If so, learning to produce an appropriate rhythmic pattern as a child depends not only on sufficient motor control, but also on the acquisition of those properties. This implies that, while rhythmic development may start early (e.g. Nazzi et al 1998), it potentially encompasses the entire period of phonological development until approximately age 9 (Ruscello 2003).