This thesis concerns the nature of the gestures performed by five Swedish children. The children are followed from 18 to 30 months of age: an age range which is characterized by a rapid succession of developmental changes in children's abilities to communicate by means of both spoken language and gesture. There are few studies of gesture in children of these ages, making it essential to ask a number of basic questions: What sort of gestural actions do the children perform? How does the use of gesture change over time, from 18 to 30 months of age? How are the gestures performed in coordination with speech? The answers provided to these questions are both quantitative and qualitative in kind. Several transitions in the use of gesture are identified, relating to developmental changes in the organization of speech — highlighting the symbiotic relationship between gesture and speech in the communicative ecology. Considerable attention is paid to the even more basic question of what sort of actions qualify for the label "gesture". Instead of treating gestural qualities as a matter of a binary distinction between actions counting as gesture and those that do not, a multi-level approach is advocated. This approach allows for descriptions of gestures in terms of several different levels of complexity. Furthermore, a distinction is made between levels of communicative explicitness on the one hand, and levels of semiotic complexity on the other. This distinction allows for the recognition that some gestural actions are semiotically complex, without being explicitly communicative, and vice versa: that some gestural actions are explicitly communicative, without being semiotically complex. The latter is particularly consequential for this thesis, since a large number of communicative gestural actions reside in the borderland between practical action and expressive gesture. Hence, the gestures analyzed include not only the prototypical "empty-handed" gestures, but also gestures that involve handling of physical objects. Overall, the role of conventionality in children's gestures is underscored. The approach is (a) cognitive in the sense that it pays attention to the knowledge and bodily skills involved in the performance of the gestures, (b) social and interactive in the sense that it views gestures as visible and accountable parts of mutually organized social activities, and (c) semiotic in the sense that the analysis tries to explicate how signification is brought about, in contrast to treating the meanings of gestures as transparently given, the way participants themselves often do when engaged in social interaction. (Less)

/pdf/children-s-gestures-from-18-to-30-months-496hgnxfcb.pdf

Children's Gestures from 18 to 30 Months

Speaker age is an important paralinguistic feature in speech which has to be considered in the study of phonetic variation. Knowledge about this feature may be used to improve speech technology applications, e.g. automatic speech recognition and speech synthesis. The present thesis describes six studies of several phonetic aspects of age-related variation in speech.

As the speech production mechanism changes from young adulthood to old age, speech is affected in numerous ways. Human perception of speaker age is based on cues such as pitch, speech rate and voice quality, and is fairly accurate. However, it is still unclear which cues are the most important ones. The first study included in this thesis investigated the role of F0 and speech rate (word duration) in age perception. It was found that while these cues may be less important than spectral ones (e.g. formant frequencies), they still correlate with chronological as well as perceived age.

In the second study, two stimulus types of various lengths were compared. Results indicated that while longer stimulus duration (regardless of speech type) seems to improve the age estimation of females, spontaneous speech (regardless of duration) appears to contain more important cues for perception of male speaker age.

In the next two studies, several automatic estimators of speaker age were built, none of which reached the same accuracy as humans. Important features in machine perception of age were also investigated. It was found that prosodic features seem to be more important in the estimation of female age, while spectral features (e.g. F2 ) appear to be more important for male age.

Although several acoustic correlates of speaker age are known, their relative importance has not yet been established. The next study analysed 161 features, automatically extracted from segments in six words produced by 527 speakers. Normalised means were used to ensure that the features could be compared directly. The most important acoustic correlates of speaker age were identified to be speech rate (segment duration) and intensity range. However, F0 and some spectral measures (e.g. F1 and F2 ) may also, if used in combination with other features, be important correlates of age.

Synthetic speech may sound more natural if speaker age is included as a parameter. The final study developed a research tool which used data- driven formant synthesis and age-weighted linear interpolation to simulate an age between the ages of any two of four female differently aged reference speakers. Evaluation of the tool showed that speaker age may in fact be simulated using formant synthesis. The tool will be used in further studies of analysis by synthesis of speaker age.

/pdf/perception-analysis-and-synthesis-of-speaker-age-3rirpi17h4.pdf

Perception, Analysis and Synthesis of Speaker Age

This thesis aims at describing the developmental patterns of text production in four text types: spoken and written narrative texts, and spoken and written expository texts (n=316), produced by fou ...

/pdf/developmental-aspects-of-text-production-in-writing-and-4bvsnkf3zk.pdf

Developmental aspects of text production in writing and speech

One of the most important functions of prosody is to divide the flow of speech into chunks. The chunking, or prosodic phrasing, of speech plays an important role in both the production and perception of speech. This study represents a move away from the laboratory speech examined in previous, related studies on prosodic phrasing in Swedish, since a spontaneous, Southern Swedish speech material is investigated. The study is, however, not primarily intended as a study of the Southern Swedish dialect; rather Southern Swedish is used as a convenient object on which to test various hypotheses about the phrasing function of prosody in spontaneous speech. The study comprises both analyses of production data and perception experiments, and both the phonetics and phonology of prosodic phrasing is dealt with. First, the distribution of prosodic phrase boundaries in spontaneous speech is examined by considering it as a reflection of optimality theoretic constraints that restrain the production and perception of speech. Secondly, the phonetic realization of prosodic phrase boundaries is investigated in a study on articulation rate changes within the prosodic phrase. Evidence of phrase-final lengthening, a reduction of the articulation rate in the final part of the prosodic phrase, is found. The tonal means used to signal coherence within the prosodic phrase is subsequently investigated. An attempt is made to test the two Lund intonation models’ capacities for describing spontaneous speech. The two approaches have different implications for the amount of preplanning needed, which makes them particularly interesting to compare by testing spontaneous data. The results indicate that no or little preplanning is needed to produce tonally coherent phrases. No evidence is found to suggest e.g. that speakers accommodate for the length of the upcoming phrase by starting longer phrases with a higher F0 than short phrases. An explanation is sought for variation in F0 starting points found in the data despite F0’s insensitivity to phrase length. It is concluded that F0 is used to signal coherence even across prosodic phrase boundaries. It is furthermore found that tonal coherence signals are used to override strong boundary signals in spontaneous speech, thereby making initially unplanned additions possible. Finally, the perception of boundary strength is examined in two perception experiments. Listeners are found to agree well in their perceptual judgments of boundary strength, and it is shown that the main correlate to perceived boundary strength in spontaneous speech is pause length. The useful distinction between weak, prosodic phrase boundaries and strong, prosodic utterance boundaries in descriptions of read speech is found to be inappropriate for descriptions of spontaneous speech. It fails to capture the conflicting local and global signals of boundary strength and coherence that arise when strong boundary signals are overriden by coherence signals. The possibility to use conflicting signals in this way is seen as an important asset to the speaker as it makes changes in the speech plan possible, and it is regarded to be a characteristic of prosodic phrasing in spontaneous speech.

/pdf/prosodic-phrasing-in-spontaneous-swedish-3249cwy1b3.pdf

Prosodic Phrasing in Spontaneous Swedish

Une etude aerodynamique, fibroscopique, acoustique et perceptive a ete menee sur les voyelles nasales du francais. Les resultats aerodynamiques montrent que la propagation du debit d'air nasal est plutot progressive que regressive, cette tendance s'amenuise en parole "spontanee". Les resultats fibroscopiques confirment que les mouvements velaires commencent avant le debut de la voyelle. Ils peuvent se terminer avant la voyelle nasale, notamment en parole "spontanee". Les donnees aerodynamiques couplees aux donnees fibroscopiques suggerent qu'il n'existe pas une ouverture velaire precise pour laquelle l'air va passer dans le conduit nasal. Les donnees acoustiques montrent qu'il existe une correlation entre le debit d'air nasal et le signal acoustique pour les voyelles [2A] et [2E], la correlation entre le signal acoustique et les mouvements velaires n'existe pas. Les resultats des tests de perception suggerent que le debit d'air nasal n'est pas l'indice principal pour discriminer la voyelle nasale.

Etude aérodynamique, fibroscopique, acoustique et perceptive des voyelles nasales du français

Nasal speech sounds occur in most of the languages of the world. Two goals concerning nasalisation in speech were pursued in the studies presented in this volume. One deals with the development and testing of a method for recording nasal vibration with an accelerometer microphone attached to the outer part of the nose, and analysing the signal with the help of a wide band spectrogram. The other aim was the investigation of speech production strategies which are used to characterise the mora nasal in Osaka Japanese and to distinguish it from the non-moraic nasal.

The use of a lightweight accelerometer microphone and the interpretation of its signal with the help of spectral analysis was considered a practical method for gathering information about nasality in speech in a field work situation. Fibreoptic recordings of the velopharyngeal port confirmed that formant structure in a wide band spectrogram of the accelerometer signal differs according to the presence or absence of oral-nasal coupling.

There are different speech production strategies for distinguishing the mora nasal from the non-moraic nasal in Osaka Japanese. These strategies include variation in duration, but also a variation in vowel quality, based on other aspects than velar opening. Anticipatory velar lowering procedures did not prove to account for a distinction between the two nasal categories. Classical analyses, claiming that the mora nasal in intervocalic position has to be a nasal vowel can be rejected for Osaka Japanese, since a variant with oral closure occurs. Its use, however, depends on the speaker’s age and on the quality of the post-nasal vowel.

Nasals and Nasalisation in Speech Production: With Special Emphasis on Methodology and Osaka Japanese

It has been suggested that speakers with an L1 with lexical tones may have an advantage when it comes to perceptually discriminating between different tones in another tone language (Kaan, Wayland, ...

/pdf/appropriate-tone-accent-production-in-l2-swedish-by-l1-1htp1z5qj7.pdf

Appropriate Tone Accent Production in L2-Swedish by L1-Speakers of Somali?

A comparison between second language learners of Swedish, some of whom have an L1 with lexical tones, while the others do not, and their ability to produce adequate tonal word accents in Swedish is presented. The L1s in question are Farsi, Somali, Vietnamese and Thai. There is no clear evidence that it is easier to learn the correct Swedish word accent patterns and their correct application for speakers whose L1 is a tonal language of any sort. Instead, it does seem to be an advantage to be an L1-speaker of a particular type of tonal language.

Tendencies of Swedish word accent production by L2-learners with tonal and non-tonal L1

In this contribution a project is presented which aims to expand and update the course material for the education of future teachers in Swedish as a foreign language. Due to new patterns of global migration in the last decades, the existing material needs to be complemented, which entails further analysis and description. As global migration affects Sweden in a similar way to many other countries, this contribution aims to trigger the discussion about how teaching of pronunciation for this group of student teachers is addressed elsewhere.

New Foreign Accents in Swedish

The number of immigrants with Albanian L1 has increased in Sweden during the last two decades. This particular group is also present in the SFI-classroom. This contribution aims to present typical pronunciation features in L2-Swedish produced by Albanian L1-speakers, based on recordings of two Albanian speakers living in Sweden.

/pdf/observed-pronunciation-features-in-swedish-l2-produced-by-l1-1yc9jvy9wp.pdf

Mechtild Tronnier

Papers

Nasals and Nasalisation in Speech Production: With Special Emphasis on Methodology and Osaka Japanese

Appropriate Tone Accent Production in L2-Swedish by L1-Speakers of Somali?

Tendencies of Swedish word accent production by L2-learners with tonal and non-tonal L1

New Foreign Accents in Swedish

Observed pronunciation features in Swedish L2 produced by L1-speakers of Albanian