Trends in Phonetics and Phonology

Introduction

Genesis of this volume

This volume was inspired by the 9^th edition of the Phonetik & Phonologie conference, held in Zürich in October 2013. ‘Phonetik & Phonologie’ (P&P) conferences first came to life in 2004 in Potsdam, with Caroline Féry and Hubert Truckenbrodt being the two initiators of the P&P conferences. Along with the P&P conference came the P&P mailing list, which is frequently used to disseminate information on phonetics and phonology conferences in Europe. The idea of P&P was to provide an informal context for an exchange of research in phonetics and phonology that is currently taking place in German-speaking Europe. Traditionally, no registration fees are charged. Following its inception in Potsdam, the 2^nd P&P took place in 2005 in Tübingen, followed by 2006 in Stuttgart, 2007 in Nijmegen, 2009 in Köln, 2010 in Frankfurt, 2011 in Osnabrück, and 2012 in Jena. In October 2013, the Phonetics Laboratory of the University of Zurich (UZH) hosted the 9^th ‘Phonetik & Phonologie’ conference. This was the first time the conference was held in Switzerland.

The idea to host the 9^th edition of P&P in Zurich came when Adrian Leemann and Marie-José Kolly attended the 8^th P&P in Jena. The first call for papers for the 9^th edition of the P&P conference went out on 14 January 2013, the second call was released on 23 April 2013. Deadline for abstract submission was early June and the conference was held on October 11 and 12 in the main building of the University of Zurich. We received 68 submissions for the P&P9 conference. Keeping in line with the tradition of previous P&P conferences, selection criteria were not highly competitive, given that P&P is an informal platform for researchers in German-speaking Europe to present their research. We divided the 68 contributions into 22 talks and 46 posters. The talks were grouped according to the following 7 main categories: articulation, sociophonetics, perception, prosody, vowels, varieties, and varia. A comprehensive abstract book was put together, which is available for download (http://www.pholab.uzh.ch/pundp9/programm/a_BookofAbstractsPP9.pdf).

During the planning of the conference the idea came about to work on an edited volume with a selection of P&P9 contributions. The timeline we adhered to was paper submission on January 15, 2014, with the review process ← 9 | 10 → taking slightly more than a month. Each paper was reviewed by at least two external reviewers in blind peer reviewing. Each editor was assigned 6 or 7 papers and made a decision on the manuscript based on his/her assessment of the two external reviews and his/her judgment of the paper. The papers were sent back to the authors with the respective revision requirements (major or minor revisions, reject). Each contributor sent back the revised manuscript along with a point-by-point response of how s/he addressed the comments by the referees. The point-by-point responses were then again checked by the four editors, before copy editing and typesetting the manuscript so that it could be sent to the Swiss National Science Foundation.

Contents of the current volume

The current volume includes state of the art research on phonetics and phonology in various languages (8 different languages) and from interdisciplinary contributors. Most contributors worked on German (Augursky et al., Brandstätter et al., Cunha et al., Insam & Schuppler, Kalmanovitch, Maurer et al., Mixdorff et al., Schauffler et al., Tomaschek et al., Trouvain, Weirich & Simpson, and Zimmerer & Trouvain). A number of papers further analyzed regional varieties of German (Hove et al., Kolly & Leemann, Otto). Bressmann, Nakamura, Ordin et al., and Schützler worked on English; Dellwo & Schmid on Italian, Dimos et al. on French, Ritzau on Danish, Voeltzel on Icelandic and Faroese. One paper examines Brazilian sign language. Thematically, we structured the edited volume into the following eight sections: segmentals, suprasegmentals/prosody, articulation in spoken and sign language, perception, phonology, crowdsourcing phonetic data, second language speech, and arts (with inevitable overlap between these areas).

Segmentals

Trouvain presents a typology of the non-phonemic use of clicks in German. Different functions of clicks are discussed, including clicks as interjections, as coarticulatory by-products or as discourse markers. A corpus analysis shows that the frequency of clicks occurring in German dialogues is highly speaker-idiosyncratic and thus potentially useful in the field of forensic phonetics. Schützler compares different approaches to transforming acoustic vowel data. ← 10 | 11 → The author analyzes commonly used transformation routines and discusses the complementary functions of psychoacoustic transformations. Finally, he proposes multi-dimensional scaling as a method to assess multi-dimensional results, which are common when different approaches to vowel transformations, are applied. Friedrichs, Maurer, Suter, and Dellwo’s contribution illustrates the practical difficulties in the acoustic study of vowels, illustrating the practical difficulties that may arise with the use of the software Praat (2014). The paper mainly addresses a number of issues related to formant analysis such as the inspection of spectrograms and the setting of parameters for algorithms based on linear predictive coding. Finally, the authors raise an intriguing theoretical problem concerning formant estimation at higher fundamental frequency which is also addressed in the contribution by Maurer et al. Cunha, Harrington, Moosmüller, and Brandstätter’s contribution investigates the contrast between tense and lax high vowels in two varieties, i.e. Standard Austrian German and Standard German. The results show that in Standard German the difference between /i y u/ and /ɪ ʏ ʊ/ is realized both in terms of duration and of vowel height, whereas Standard Austrian German mainly relies on the quantity feature and displays an approximation of the two categories as far as vowel height is concerned. Brandstätter, Moosmüller, and Kaseß investigate quality contrasts in high vowel pairs of Standard Austrian German. As opposed to Standard German vowels, where /u, ʊ/, /y, ʏ/ and /i, ɪ/ are distinguished by quality as well as by quantity, the quality contrast in Standard Austrian German high vowels /u, ʊ/ and /y, ʏ/ tends to be neutralized. The authors quantify the degree of neutralization with formant measurements in read pseudowords of 18 Standard Austrian and 7 Standard German speakers. The paper ‘The role of long term acquaintances in speech accommodation’ by Kalmanovitch examines segmental accommodation between speakers that have known each other for a long time. Kalmanovitch’s case study shows phonetic convergence between interlocutors with regard to global phonetic features: for example, long-term acquaintances share a tendency towards voiced realizations of /z/ in Standard German regardless of their regional provenance.

Suprasegmentals

Weirich & Simpson examine the impact of accent realization and speaker’s sex on vowel length in German. They find tense vowels to be more strongly affected by the different accent conditions tested. Moreover, they report ← 11 | 12 → between-sex differences in vowel quantity realizations in the unaccented accent condition. They argue that males have a stronger degree of undershoot than females. In the contribution ‘Segmental effects on prosody: Modelling German argument structure’, Schauffler, Schweitzer, Schweitzer, and Augurzky examine if context-independent prosodic features such as tendency of tonal alternation affect the choice of pitch accent types. The study reports that the distance between two successive accents affects the probability that the two accents differ in type. Based on these analyses they conclude that prosody-inherent factors need to be taken into account when phonological and information structural categories are associated with each other. Augurzky, Riester, and Tomaschek investigate the effect of consonantal segmental characteristics on the placement of prosodic breaks. They find that both the number of prosodic breaks as well as their perceptual salience are increased: they argue that the presence of consonant clashes increases the number of prosodic boundaries in speech production. Mixdorff, Hayashi, and Ushiyama study the perception of word boundaries and accented syllables in German by native German listeners as opposed to non-natives. The authors report that even non-German speakers detect a large number of word boundaries and accented syllables. Overall, non-native German speakers showed improvement in the task the more experienced they were with German.

Articulation in spoken and sign language

Bressmann presents a novel method for the reconstruction of pseudo-3D tongue surfaces. The proposed method represents a fast, non-invasive alternative to other imaging techniques for obtaining data on tongue shapes. Bressmann used a custom-made head stabilizer with a custom-made ultrasound transducer stabilizer to obtain several coronal ultrasound images of the tongue. This allowed him to measure tongue height, anteriority and concavity at different view angles. The efficiency of the method and the stability of the results are demonstrated with data from three English speakers. Barbosa, Temoteo, and Nogueria deal with the description of the position of arm and forearm in the dominant hand’s location for Brazilian Sign Language signs. The authors use goniometric measures to analyze 32 signs, including the following locations: the angle of flexion of the arm, angle of vertical abduction, the angle of horizontal abduction, lateral rotation of the arm, and the flexion of the elbow and rotation of the wrist. ← 12 | 13 →

Perception

Tomaschek, Truckenbrodt, and Hertrich study the discrimination sensitivities and identification patterns of vowel quality and duration in German /u/ and /o/. Results reveal vowel quantity to be the principal cue for vowel discrimination, while spectral discrimination shows no sensitivity to tense/lax contrasts. They argue that vowel quality nevertheless supports contrast marking in that it increases the perceptual distance between short and long vowels. In ‘Vertical variation in East Thuringian’, Otto reports on an acoustic and perceptual study of the variability between Standard German and the variety of German spoken in East Central Germany. She finds that a reduced horizontal dimension in the vocalic space in East Thuringian compared to Standard German is the strongest cue to perceptual difference between the two varieties. The paper ‘Perceptual magnets in different neighbourhoods’ by Duran combines an exemplar-theoretic model with the model of the perceptual magnet effect. A perceptual magnet is the apparent warping of the perceptual phonetic space due to which two distinct speech sounds are perceived as more similar if they are close to a prototypical representation in a given category.

Phonology

In ‘Conditioning factors in word-final coronal stop deletion in British English’, Nakamura explores articulatory and acoustic characteristics of t/d deletion in word-final consonant clusters in British English. Nakamura finds that preceding and following phonological contexts are significant in promoting the occurrence of t/d deletion while morphology seems to play a subordinate role. The paper ‘Dissimilation in Western Nordic’ by Voeltzel examines dissimilation phenomena in Icelandic and Faroese in which sonorant clusters turn into /d + sonorants/. She finds that morphemic boundaries play an important role in triggering this dissimilation process, whereas phrase boundaries seem to obstruct it. Voeltzel’s account shows an interaction at the interface of morphology and phonology. ← 13 | 14 →

Crowdsourcing phonetic data

Kolly & Leemann present the functionalities of the smartphone application Dialäkt Äpp which has obtained a considerable interest on the part of the Swiss public (more than 61 000 downloads, June 2014). This application predicts the local Swiss German dialect of users on the basis of their indicated variants of 16 words; moreover, users may record their own pronunciation of these words, which are then uploaded on a server. For the first time, this method of crowdsourcing acoustic linguistic data allows for large-scale quantitative analyses of regional pronunciation variants. Hove, Leemann, Kolly, Dellwo, Goldman, Almajai, and Wanitsch’s contribution presents a follow-up research to Dialäkt Äpp which aims at gaining new insights through the collection of additional acoustic data from Swiss German dialects. The development of a new smartphone application Voice Äpp takes the crowdsourcing method one step further, using automatic speech recognition based on data retrieved through crowdsourcing (Dialäkt Äpp, cf. Kolly & Leemann this volume) in order to determine the dialect of the user. Additionally, Voice Äpp enhances the interactive possibilities of the application, in that it offers to users a number of information about their own voices, e.g. by calculating speech rate or F0 range.

Second language speech

Zimmerer and Trouvain investigate an interesting phenomenon in the pronunciation of German as an L2 by French learners, i.e. the realization of the phoneme /h/ which is lacking in their L1. Nevertheless, an analysis of read speech produced by 7 native speakers of French shows that /h/ is only rarely deleted in their German speech, while different realizations of this consonant can be found – not only fricatives, but also glottal stops. Insam & Schuppler present a case study on the effects of pronunciation training on non-native speech. Five native English learners of German received perceptual training in this L2. Acoustic analyses of their vowel productions before and after the training as well as perceptual ratings of their foreign accent show that the phonetic training was successful. Furthermore, the data suggests that acoustic and perceived characteristics of L2 vowels are related. Ordin, Polyanskaya, and Wagner’s research is concerned with the acquisition of English as a second language by German and French learners. Here, the focus is not on the production of a particular segment, but rather on the overall rhythmic ← 14 | 15 → patterns of the languages involved. Initial German learners are shown to cope better than French learners with the durational variability of English vocalic and consonantal intervals; moreover, both German and French learners exhibit a greater (and thus more target-like) durational variability when the overall proficiency of their English increases. Timing patterns in a situation of language contact constitutes the topic of another contribution to this volume. Dellwo and Schmid investigate the speech of 5 Italian-German bilinguals in both of their languages, partially applying the same rhythm metrics as Ordin et al. (this volume). An interesting result of this study comes from the fact that the timing patterns in the two languages do not only confirm some well-known characteristics from the literature but there are also some speaker-specific characteristics which appear in both languages of the bilinguals, in particular with regard to the amount and the variability of voicing patterns and vowel durations. Ritzau investigates the influence of orthographic input on the pronunciation of twelve German learners’ Danish speech. Danish has a far more opaque writing system than German, and the pronunciation of assimilations at word boundaries is expected to be particularly difficult for learners. Ritzau’s results show that learners’ speech is more target-like in spontaneous speech – her data suggest a stronger influence of the writing system in read speech. These findings are discussed in the context of the language classroom: At which stage should second language learners receive orthographic input?

Arts

Dimos, Dick, and Dellwo provide an overview article of the phonetics of composed theatre. This avant-garde type of theatre is characterized by a mixture of sounds and visual elements, all of which contribute to the artistic result. Their paper gives a first overview of how prosodic features are mannered and highly variable in this as of yet under-researched domain of speech production. Maurer, Suter, Friedrichs, and Dellwo investigate the acoustic characteristics of voice in music and straight theatre. Stage voices do indeed exhibit significant differences with respect to ‘normal speech’ both in the F0 and the spectral domain. Notions such as ‘formant tuning’ or ‘singer’s formant’ are discussed and illustrated with a number of vowel spectra. The empirical evidence for this research comes from a new database that is currently being built in a project on the acoustics of stage voices. One of the aims of the project is ← 15 | 16 → to highlight fundamental problems of acoustic vowel analysis, in particular regarding the spectral properties of vowels at higher fundamental frequencies.

About the Zürich Phonetics laboratory

Before the P&P9 conference, the Phonetics Laboratory hosted a number of conferences and workshops in Zurich. In 2009, the Laboratory hosted the 5^th AISV conference (Associazione Italiana de Scienze della Voce, http://www.pholab.uzh.ch/labor/konferenzen/aisv2009_de.html). In 2012 the Laboratory hosted the first workshop on Research on Prosody in Switzerland (SWIP I, http://www.pholab.uzh.ch/labor/konferenzen/prosody-workshop-270412_de.html). The aim for this workshop was to bring together researchers in speech prosody working in Switzerland. Delegates came from Neuchâtel, Geneva, Fribourg, Bern, and Zurich. SWIP II was subsequently held in Neuchâtel in 2013, and SWIP III will take place in Geneva in 2014.

The Phonetics Laboratory at the University of Zurich was established in 1935. Eugen Dieth, Professor of English, dialectologist, and phonetician, is the founding father of the Laboratory. The Laboratory became an official University Department in 1943. Two of the most significant works of Dieth are the Schwyzertütschi Dialäktschrift (1938), an attempt at a standardization of Swiss German dialect spelling, and the Vademekum der Phonetik (1950), a phonetics primer that was written based on his introductory courses on phonetics Dieth taught at the University of Zurich. The first full time assistant at the Phonetics Laboratory was Rudolf Brunner. Together with his teacher Eugen Dieth, Brunner was dedicated to the documentation and research of Swiss German dialects. Following Brunner, Sonja Spörri-Bütler took over the assistant position until 1995. Spörri-Bütler, too, worked on Swiss German dialects, for example in ‘A voice onset time analysis of initial Swiss German stops’ (1981), published with Daly H. Engstrom in Folia Phoniatrica 33. Stephan Schmid followed Sonja Spörri-Bütler in 1995; his research interests mainly deal with Swiss German and Italian dialects, phonological typology, and second language speech. Volker Dellwo now occupies the first Assistenzprofessur in Phonetics at UZH which was established in 2010; he acquired funding for a number of research projects in different fields such as speech rhythm, forensic phonetics, acoustic variability of vowels, theatrical speech and the so-called ‘cocktail party effect’. Adrian Leemann and Marie-José Kolly joined the Laboratory in late 2011. Both of them, Adrian Leemann as postdoctoral researcher and ← 16 | 17 → Marie-José Kolly as PhD student, work in the project on ‘Speaker identification based on temporal information’. In 2012 more doctoral students joined the Laboratory; Daniel Friedrichs in a project on vowel production and perception at high fundamental frequencies, Kostis Dimos in a project on the phonetics of theatrical speech and Lei He in a project on speaker identification using temporal information. The projects are possible thanks to the funding of the Swiss National Science Foundation (SNSF) and the Gebert-Rüf Stiftung (http://www.pholab.uzh.ch/leute/dellwo/researchGrants.html). More information about the Phonetics Lab can also be found in Dellwo & Studer (2011).

Finally, we express our gratitude to the sponsors of P&P9 in Zürich and the sponsors of this edited volume (ZUNIV, Hochschulstiftung UZH, and The Swiss National Science Foundation).

Zürich, June 2015

Adrian Leemann