@article {490, title = {Prosodic expression of sarcasm in L2 English after explicit training}, year = {2021}, publisher = {Nederlandse Vereniging voor Fonetische Wetenschappen}, address = {online}, abstract = {Sarcasm occurs frequently in everyday interactions and is typically expressed via prosody. However, speakers usually experience difficulty with expressing sarcasm prosodically in a second language. Recent research by Smorenburg and co-authors has shown that short prosodic training can already improve perceived sarcasm in Dutch learners of English (L2E learners). It remains to be investigated what exactly have changed in the prosodic production after training. In this study, we analysed the production of sarcastic prosody by L2E learners (N =12) before and after the training in Smorenburg et al.{\textquoteright}s study. The production data were elicited via a simulated telephone conversation task in which the participants were prompted to give a sarcastic response to a friend{\textquoteright}s remark. Mean pitch, maximum, minimum pitch, duration were extracted from the key word of each response (e.g. {\textquoteleft}healthy{\textquoteright} in {\textquoteleft}She is a healthy lady) using ProsodyPro and were subsequently analysed using linear mixed-effect modelling. We have found that prosodic improvement depended on utterance type and gender. The training helped L2E learners to lengthen duration in declaratives and tag-questions but not in wh-exclamatives and lower mean pitch and minimum pitch across utterance types to sound more sarcastic, similar to native speakers of English. Male L2E learners lengthened duration to a larger extent than female L2E learners, similar to female native speakers of English; female L2 learners did not lower mean pitch more than male L2 learners, different from female native speakers of English. We will discuss the nature of the prosodic improvement: L1 transfer and/or L2 learning. }, author = {Zandee, Femke and Jansen, Nelleke and Laura Smorenburg and Aoju Chen} } @article {475, title = {Automatic Analysis of Speech Prosody in Dutch}, year = {2020}, publisher = {Nederlandse Vereniging voor Fonetische Wetenschappen}, address = {online}, abstract = {In this talk we present the first publicly available tool for automatic analysis of speech prosody (AASP) in Dutch. Incorporating the state-of-the-art analytical frameworks, AASP enables users to analyze prosody from two different theoretical perspectives. Structurally, AASP analyzes prosody in terms of prosodic events within the auto-segmental metrical framework, hypothesizing prosodic labels in accordance with Transcription of Dutch Intonation (ToDI). Holistically, by means of the Functional Principal Component Analysis (FPCA) AASP generates mathematical functions that capture changes in the shape of a pitch contour. Regarding ToDI, AASP performs four tasks including pitch accent detection, pitch accent classification, prosodic boundary detection, and prosodic boundary tone classification. Using SVM, AASP performs with accuracy comparable to similar tools for other languages for pitch accent detection, prosodic boundary detection, and prosodic boundary tone classification. Notably, we have found that by combining functional features extracted from FPCA with conventional acoustic features, AASP can attain a higher accuracy for pitch accent classification (76.87\%) than AuToBI for English using conventional acoustic features (71.6\%). Regarding FPCA, AASP outputs the weights of principal components that capture core variations in the shape of pitch contours in a .csv file, which can be directly used for further statistical analysis. Published as a Docker container, AASP can be set up on various operating systems in only two steps. Moreover, the tool is accessed through a graphic user interface, making it accessible to users with limited programming skills. It has also the potential to be adapted for prosodic analysis in other languages. }, author = {Hu, Na and Janssen, Berit and Hanssen, Judith and Carlos Gussenhoven and Aoju Chen} } @article {455, title = {Prosody differs between objective and subjective causal relations in English}, year = {2019}, publisher = {Nederlandse Vereniging voor Fonetische Wetenschappen}, address = {Amsterdam, The Netherlands}, abstract = {Causality can be established either objectively or subjectively. The causality expressed in example (1) is objective, because both the consequence and the cause are real events; while the causality in example (2) is subjective, because {\textquotedblleft}Heidi is talented{\textquotedblright} is one{\textquoteright}s opinion. In languages like Dutch, these two types of causality are expressed by specialized connectives (omdat/want). However, in English, they are both expressed by because. Hence, the types of causality are left unspecified on the lexical level (if other lexical cues are not present). This leads to the question of whether these two types of causality are distinguished by non-lexical cues, e.g. prosody. We investigated this question in both forward and backward causals using a dialogue task. Taking the Bayesian approach, we examined a wide range of prosodic features, including not only static measures such as pitch and duration, but also dynamic measures on the shape of pitch contours extracted by Functional Principal Component Analysis. The results showed that in comparison with objective causals, subjective causals were produced with higher F0 maximum, lower F0 minimum, longer duration, and also with distinctive contour shapes. These results indicate a trade-off between lexical and prosodic cues. (1a) Heidi is thrilled because she won the first prize at the art festival. [backward] (1b) Heidi won the first prize at the art festival so she is thrilled. [forward] (2a) Heidi is talented because she won the first prize at the art festival. [backward] (2b) Heidi won the first prize at the art festival so she is talented. [forward] }, author = {Hu, Na and Aoju Chen and Hugo Quen{\'e} and Ted Sanders} } @article {347, title = {Intonational realisation of topic and focus in child Dutch}, year = {2007}, publisher = {Nederlandse Vereniging voor Fonetische Wetenschappen}, address = {Nijmegen, The Netherlands}, abstract = {

Languages express the topic-focus distinction in different ways (e.g. word order, particles, intonation). In this study we focus on intonation and in particular look at type of pitch accent and phrasing. There has been relatively little discussion on the use of intonation to express the topic-focus distinction in child language. Prior work is mostly concerned with the use of accentuation in expressing contrast. In this study, we examined how Dutch children use pitch accent types (including deaccentuation) and phrasing to mark topic and focus in different sentence positions and how they differ from adults. The topic and focus under investigation were non-contrastive and realised as full NPs (e.g. what did the boy draw? [The boy]topic drew [a castle]focus).

A picture-matching game was used to elicit topic-focus structures as answers to WH-questions. Two variables were controlled for in the answer sentences: PRAGMATIC CONDITION (topic, focus), SENTENCE POSITION (initial, final). Data were collected from monolingual Dutch children (aged 4-5 years, 7-8 years and 10-11 years) as well as adults. The intonation patterns were transcribed following ToDI notation.

Our analysis has revealed five major findings:

  1. Children of all age groups and adults employ a similar set of pitch accent types. These accent types (e.g. L*H, H*L, H*, !H*L) form the core of the inventory of pitch accents in Dutch. This finding thus shows that children as young as 4 have adult-like inventory of accent types.
  2. Like adults, children of all age groups deaccent topic more frequently than focus independent of sentence position. This result indicates children{\textquoteright}s early sensitivity to the accentuation-focus and deaccentuation-topic associations, as suggested in prior work.
  3. Children acquire H*L as the typical {\textquoteleft}focus accent{\textquoteright} at the age of 7 or 8. 4- to 5-year-olds exhibit a weak preference for H*L over other accent types in sentence-initial position and no preference for H*L in sentence-final position. Possibly, frequent use of H*L in sentence-initial topic in adult Dutch has made it difficult for young children to associate H*L primarily with focus.
  4. Children also acquire deaccentuation as the typical {\textquoteleft}topic intonation{\textquoteright} at the age of 7 or 8. 4- to 5- year-olds realise topic similarly frequently with deaccentuation, L*H and H*L in sentence-final position.
  5. Only 4- to 5-year-olds appear to use phrasing to realise topic, which forms its own intonational phrase (IP). Older children and adults utter the topic-focus structure mostly as one IP.
}, author = {Aoju Chen} } @article {270, title = {Universality and Language-dependence of Intonational Meaning}, year = {2001}, publisher = {Nederlandse Vereniging voor Fonetische Wetenschappen}, address = {Utrecht, The Netherlands}, abstract = {

Two contradictory views in previous studies have been identified on the nature of intonational meaning across languages (Ladd 1981): The Strong Universalist Hypothesis and The Nuclear Tone Hypothesis. The two views differ, among other things, in their claim on whether intonational meaning is innately specified, i.e. in accordance with the natural physiological states of speaker, and hence universal. Because of the obvious drawbacks of both views, in the present project, we have adoped the conception recently proposed by Gussenhoven. That is, intonational meaning is universal and language-specific at the same time; universal and language-specific meanings of intonation are derived from two different language components, namely, the phonetic implementation module and intonatonal lexicon, i.e., the set of morphemes invested with intonational meaning. Our aim in this project is to provide empirical evidence for the universality and the language-dependence of intonational meaning by studying the use of three biologically determined codes, the Frequency Code (Ohala 1983, 1984, 1994), the Effort Code and the Production Code (Gussenhoven forthcoming) in human vocal communication.

In the present talk, I will give an overview of the project design and then consider findings from two cross-language perception experiments. The two experiments are concerned with the use of the Frequency Code, which is based on the fact that a larger larynx predicts low-pitched voice. The informational interpretation of the Frequency Code is that high pitch signals questions and low pitch signals statements. From an affective point of view, high pitch signals politeness, friendliness, uncertainty and submissiveness while low pitch signals self-confidence, assertiveness and dominance. Findings from the two experiments support the universality of intonational meanings derived from the Frequency Code and shed light on a new type of language-dependence of intonational meaning.

}, author = {Aoju Chen} }