This book is a collection of short papers dedicated to Professor Wiktor Jassem, an eminent researcher in the fields of rhythm, intonation, and the prosody–segment relationship, among others. Unlike in some other dedicated volumes and Festschrifts, the personality, contribution, and influence of Professor Jassem form a strong unifying line throughout the entire 19 contributions.
The book consists of editors’ introduction, four core thematically-oriented sections – each including individual contributions – and the final section providing Jassem's biography and bibliography. The introduction presents a concise preview of the individual contributions, highlights their main novel aspects, and identifies the relevance of Jassem's work to the issues discussed in the papers. The summaries provide sufficient information on the individual contributions for those who wish to read only a subset of the papers.
The first section, entitled ‘Methodology’, includes papers by the guest editors, Dafydd Gibbon, Daniel Hirst and Nick Campbell, and concludes with a paper by Grażyna Demenko. If the reader expects well-rounded, experiment-based papers, they will probably meet with disappointment. But if the reader expects sharp identification of issues, ambitious proposals and a solid description of methodological approaches to these issues, he or she will be fully gratified.
Daniel Hirst takes on the formidable task of identifying universal prosodic elements as a tool for describing both intrinsic differences and environmental effects in a language- and theory-neutral way. He starts with introducing Jassem's Narrow Rhythmic Unit and uses it for setting out, concisely, the differences between Jassem's and Halliday's models. He then proposes the representational framework that can be used for testing the universality of prosodic constituents.
The abstract of Dafydd Gibbon's paper, with four rather substantial aims, leads one to expect a full research paper. Instead, Gibbon provides an intuitive method for understanding rhythm and its basic properties, with explication within a formal computational model informed by Jassem's work and with avenues for further research using the model. One very favourable feature is that both lexical and post-lexical tone patterning can be stated in the same formal language; and all this within a very short text (nine pages).
Nick Campbell draws attention to cognitive representations and how they interact with the prosodic aspects of speech, for example suggesting conversational rhythmic units based on human–robot spoken interactions. He also continues to advocate using large corpora of real conversational speech for researching prosody.
Grażyna Demenko's paper deals with intonation in speech technology and provides a short overview of the issues and some methodological approaches in this applied domain. Given that this is an overview paper, the reader might wish to have better referencing of the claims, explanation of the terms (e.g. micro-prosody), and broader synthesis of disjoint issues, but given space limitations for these short contributions, the paper provides a useful starting point for researchers wishing to explore prosody in speech technology.
Minor editorial glitches in this first section (absent in the other sections), such as referring to a non-existent Figure 7 instead of Figure 6 (p. 31), referring to footnote 2 instead of 3 (p. 38), and some other ones, do not present major distractions for the reader.
The second core section, ‘Rhythm – duration and timing’, contains four papers (Hill, Gutt, Klessa, and Malisz & Wagner). David Hill's is a personal paper reviewing the development of the first synthesizers and recollecting a six-month collaboration between Hill and Jassem. The results of their research give a quantifiably better account of speech rhythm with the Jassem's model than with the traditional stress feet model of Abercrombie (Reference Abercrombie1967) and Halliday (Reference Halliday1985).
The paper by Ulrike Gutt provides a critical and compact overview of the recent advances in the study of speech rhythm with respect to the usefulness of traditional divisions and experimental metrics commonly used for studying speech rhythm. The novel aspect of the paper is rhythm in second language (L2; non-native) speech and the options the rhythm metrics provide for gauging L2 proficiency. Gutt concludes with a suggestion that rhythm metrics do not in fact measure rhythm and that rhythm should be considered a phonological phenomenon that is not linked to measurable aspects of speech in any direct way. While I largely agree with Gutt, I would postpone the abandoning of rhythm metrics until we reach a similar conclusion for not only durationally-based metrics but also for more sophisticated ones that measure the alternation of prominent and non-prominent syllables using a combination of duration, f0 and intensity.
Katarzyna Klessa's paper is a small study of the duration of Polish segments with an overview of past work on this topic. The pilot findings based on a subset of a large corpus of read and semi-spontaneous speech suggest differential effects of phrasal stress on the duration of vowels and consonants in two dialectal groups of Polish (East and West).
Finally, Zofia Malisz & Petra Wagner take on the formidable challenge of examining the effects of intensity, pitch movement, duration and spectral tilt on signalling word stress and sentence prominence in Polish. This is a very balanced, data-driven paper examining the hypothesis that all factors mentioned above participate in cuing prominence in a corpus of conversational speech. Intensity, pitch movement and duration were found to robustly signal major accentual prominence. But neither these three prosodic dimensions nor the hypothesized spectral tilt were found to mark lexical stress in a consistent way. The authors suggest that fixed penultimate word stress in Polish is a highly influential expectation perceived by native speakers but not consistently signalled in the acoustic signal. This is in line with Gutt's conclusions and, to my mind, provides a useful hypothesis for studies of other fixed-stressed languages such as Czech, Slovak or Hungarian.
The third core section, ‘Melody – intonation and pitch’, contains five papers (Esling & Moisik, Rakowski, Connel, Karpinski, and Wagner). John Esling & Scott R. Moisik's paper provides much needed descriptions of physiological factors and strategies in producing pitch variation, especially the dynamic characteristics of laryngeal gestures and their temporal coordination. The paper presents the novel technique of Simultaneous Laryngoscopy and Laryngeal Ultrasound (SLLUS) enabling such research. An interesting finding concerns the role larynx height plays in producing pitch variation despite its high-mass slow-moving dynamics.
The next two papers explore the relationship between speech melody and music. Andrzej Rakowski's paper provides a comprehensive and refreshing survey of issues surrounding absolute pitch – the ability to identify the notational label on a chromatic scale for a tone – and its various sub-categories (pseudo-absolute, quasi-absolute, active/passive absolute pitch). The results of a test designed to screen those demonstrating full absolute pitch showed the effect of memory on absolute pitch categorization: decreasing test interval from four seconds to one second caused significant deterioration in performance for those possessing quasi and pseudo absolute pitch, supporting thus the author's view that those types are based on non-auditory memory.
Bruce Connel's paper presents an analysis of a drinking song in the African tone language Mambila comparing the tunes stemming from the succession of lexical tones (speech) and the song's melody. This limited but interesting study concludes that both lexical tones and the melody of the song contribute to overall pitch realization and that the only robust difference between the two is the increased pitch range of the song's melody compared to speech.
The paper by Maciej Karpiński is a focused study of phrase-final rises (PFRs) in task-oriented dialogues with a theory-neutral operationalization of this phenomenon, and a bottom–up approach trying to link the acoustic data with the pragmatic and discourse functions of PFRs. Interesting findings include the effect of the task role (information giver vs. follower), and turn management (overlapped vs. non-overlapped phrase and turn-final vs. turn-medial positions).
The final paper, by Agnieszka Wagner, presents an approach to the much needed automatic annotation of prosody. It describes a reasonably small set of features extractable from the acoustic signal and lexical labelling and four models for detecting (i) the presence and type of prominence, and (ii) the presence and type of boundary comparing the performance of four algorithms for each (decision trees, radial basis function, multi-layer perceptron, and discriminant function analysis). The feasibility of this approach is underscored by the performance of the model, comparable with other state-of-the-art models and human annotators.
The final section, entitled ‘Harmony: Segments in context’, contains five papers (Kohler, Gonet, Balas, Shockey & Bond, and Rosenhouse). It starts with a very entertaining and controversial paper by Klaus J. Kohler re-visiting the issue of incomplete neutralization of voicing. I mean entertaining in a positive sense and not just with reference to the Epilogue of the paper, which summarizes the development of neutralization with ingenious melodic expressions. Kohler advocates the Firthian syntagmatic approach to phonology over the structuralist, generative, and laboratory approaches, which treat the voicing contrast purely paradigmatically and fail, in his opinion, to show the incompleteness of neutralization in real communicative data. It is worth mentioning that context-induced voicing of word-final voiceless plosives that Kohler uses for supporting the syntagmatic approach to phonology are not limited to German but are also known for Kraków/Poznań Polish and Slovak as voice sandhi. Certainly, Kohler's call for improved understanding and ultimately bridging of the segment–prosody divide is very much warranted. Yet, in my view, the critique of the laboratory phonology approach especially is too harsh, since experimentally testing the predictions of phonological theories experimentally goes a long way towards improving our understanding of the cognitive system underlying human speech.
The paper by Wiktor Gonet takes the issue of word-final voicing in English and its repercussions for teaching English pronunciation to learners who are native speakers of Polish. In a corpus of English extracted from TV shows, the measurements of length of voicing during obstruent closure suggest a minimal effect of the preceding context and whether the obstruent is a stop or fricative but a significant effect of the following context such that following voiced segments facilitate lengthening of voicing in the target word-final obstruents.
Anna Balas’ paper examines the production of glottal stops in English vowel-initial words with the goal of facilitating language instruction to non-native (Polish) speakers of English. The author describes a rather surprising finding that glottal stops, hypothesized to be a delimitative boundary signal, were more frequent in Polish-accented English before unaccented function words than accented lexical words. This might be influenced by the vowel identity, for which the study did not control.
The following paper, by Linda Shockey & Zinny Bond, describes data from a gating experiment showing how phonetic reduction strategies in native and non-native speech might provide a great deal of insight into phonological (cognitive) patterning in speech. This paper thus echoes in some way the approach by Kohler in stressing the importance of communication in casual speech and syntagmatic relations, which is represented in the current paper by the ‘perceptual window’.
The final paper, by Judith Rosenhouse, explores the phonetic and prosodic aspects of first language (L1) interference in L2 on the basis of the features in the speech of Hungarian-born speakers of Hebrew in Israel. It is a descriptive study informing the areas of language instruction, bilingualism, and code-switching; it also points to the potential such comparative studies might provide for improving language instruction programs for immigrants.
The persistent feeling I had when reading the book was that the restricted space for each contribution affected the papers in that while they present very useful targeted overviews of topics, the space devoted to methodology and results was rather limited. I commonly had questions about the methodological approach (e.g. regarding balanced order of the three conditions in Rakowski's study, pitch categorization in A. Wagner's study, VIC-interval normalization and statistical tests in Gonet’s, and several others). Yet, the novel experimental results, despite being pilot or suggestive at times, provide inspiration for future research not only in L1 but especially in L2. Overall, this volume thus presents a very useful set of short, targeted, empirically-based, and thought-provoking studies useful for a wide readership of students and researchers in the area of speech prosody.
