This book constitutes a retrospective of Michael Halliday's model of intonation, which is set within the functional systemic approach inspired by Firth and long associated with Halliday. The ways in which language ‘means’, according to Halliday & Greaves (henceforth H&G), are threefold: firstly it construes human experience, secondly it enacts human relationships, and finally it enables the construction of discourse. Throughout this book they endeavour to account for the role of intonation (pitch and rhythm) in making these meanings.

The book is divided into three parts. Part I (chapters 1–4) focuses on the analysis of sound and the structure of intonation, part II (chapters 5–7) focuses on how intonation contributes to meaning, and part III, a single chapter, consists of a guide to Halliday's system. In addition, each part is followed by worked analyses of spoken texts, applying the model described in the book. There is an accompanying CD that contains a pdf file of the entire book with embedded sounds, so that the reader can hear all the examples and read the book on screen. There are also separate sound files and video clips which can be used alongside the book.

The first chapter provides a broad overview of the role of speech in language, and begins by considering different aspects of speech sounds. The authors offer a brief view of speech from the acoustic and biological perspective (hearing and production), and also from an engineering perspective, including the recording of speech, and applications in speech technology and audiology. They then consider the role of speech in the expression of meaning as it is studied by philosophers, dialectologists, sociologists, psychologists, clinicians and therapists. Such an overview, however brief, is of value in itself in setting out the broader context for the study of speech. This is followed by an eclectic overview of related theories and approaches: the IPOFootnote ¹ approach to modelling English intonation (’t Hart et al. Reference ’t Hart, Collier and Cohen1990), the ToBI system of transcription, metrical and autosegmental phonology and Optimality Theory. One might have wished for a more detailed discussion of how these differ from a systemic functional approach, especially since generative models have come to dominate the study of intonational phonology in recent years.

Chapter 2 offers a concise introduction to the instrumental analysis of speech, covering waveforms, spectrograms and fundamental frequency (pitch) contours. The aim is to prepare readers to understand the illustrations that later in the book accompany extended data analyses. Some of the detail is perhaps unnecessary, although the chapter is a useful introduction in its own right. It is a pity that the alignment of text with the waveform or spectrogram is not always clear, and that some of the pitch contours are overlaid on sections of spectrogram that are too blurred to be useful.

Chapter 3 begins with a discussion of the mapping between information units and tone units. It then outlines the system of tonal contours proposed by Halliday, five simple tones and two complex, i.e. seven in all. These can be preceded by ‘pretonic’ patterns which depend on the choice of tonic. The authors then exemplify the way in which various meanings can be realized through tone choice, tone placement (‘tonicity’) and tone units (‘tonality’). Choice of tone is shown to contribute to ‘logical meaning’, i.e. by signalling interclausal relationships such as coordination and subordination, and also to interpersonal meaning in the way they are combined with different kinds of clauses – declarative, interrogative and imperative. For each clause type H&G suggest a default, neutral, choice of tone, while other combinations are marked and give rise to a range of attitudinal nuances. To illustrate this they give examples of declaratives with different tones and describe the attitudes expressed; for example, a declarative with a low rise (tone 3) may sound non-committal, while the same declarative with a full rise may sound challenging. These glosses are for the most part intuitively acceptable, but ascribing such meanings to the combination of clause type and tone may well mislead readers into thinking that these meanings will arise consistently. The authors point out, much later in the book, that ‘the exact meaning of any one instance will be determined by its environment’ (p. 165), but the illustrations tend to suggest a more direct link between tone choice and attitudinal meaning, which is unfortunate. It is also unfortunate that the sound files illustrating ‘non-committal’ and ‘challenging’ declaratives (‘I like it’) display other differences, including the pitch of the initial syllable, one high and one low (HL*H vs LL*H), with additional loudness and rougher voice quality for the ‘challenging’ version. It does indeed sound more ‘challenging’ but not necessarily because of the choice of tone. The subsequent account of tone placement and information focus gives a good overview of the basic effect of signalling given and new information, and there is finally a brief discussion of the mapping between tone units and information units. H&G claim that ‘other things being equal’ an information unit will be coextensive with a grammatical clause. Since in general other things are not equal, this may be a little misleading to intonation novices.

Chapter 4 is a further discussion of the role intonation plays in generating meaning – in H&G's view, intonation plays a part in all but the expression of experiential meaning. This is mainly a chapter that sets out the authors’ fundamental position on the nature of language, which is that language is ‘not an inventory of structures but. . . a semiotic system’ (p. 61). The emphasis on this point serves to show how the authors differ fundamentally in their view of language from the many prosodic phonologists working in a generative paradigm. They vehemently oppose the autosegmental (AM) pitch level system, partly because they believe the contours themselves generate meaning, not the component parts, and partly because they feel it treats intonation as a secondary resource rather than being directly implicated in the making of meaning. The first is understandable, especially coming from a Firthian perspective. I am not sure of the justification for the second, but H&G's convincing argument for the centrality, even primacy, of prosody deserves more elaboration, including evidence from child language, than it is given here.

A unique feature of this book is the extent to which worked examples of discourse are presented with the original video and/or sounds, so that the reader can see how the model can be applied. Some of these examples constitute appendices to parts I and II, and others are incorporated into part III. While such worked examples constitute some of the richness of the book, they are also to some extent a hostage to fortune. In the appendix to part I, a short extract from a doctor–patient interview is analysed. The reader has access to both the sound files and the original video clip, and can thus follow the analysis closely. It is unfortunate that the first line (‘What happens when you take penicillin?’) is analysed with a rising tonic, when the pitch trace in the illustration (fig. A.1.1b) shows a very shallow fall, which is also what we hear.Footnote ² The interpretation (‘mild, interpersonal’), while subjectively accurate, is to be ascribed not to a rising tone but to the very narrow range – the fall ends in the middle of the speaker's range. A fall to low would indeed have been more assertive.

While intonation specialists are by no means always in agreement over what they hear, it is a pity that a book aimed at the novice provides such easily contestable examples. This is of course the nature of ‘real’ data, but it is in my view more helpful to point out indeterminate cases as such, rather than leave the readers to think the indeterminacy is the result of their own inadequacy. It would also have been useful to point out how much the attitudinal interpretation is due to the speaker's voice quality, the overall height of the utterance, and the narrowness of range, none of which play a part in Halliday's system. The same observation applies to the second clip about smoking:

The doctor's intonation in the first question is the default for a wh-question (fall) and spoken in a ‘neutral’ manner, but the patient's response, while using the same default falling tone, is spoken very rapidly and very breathily. The choice of tone may be neutral, as the authors state, but other prosodic features are definitely not. What the speaker meant is another matter, and is probably not recoverable. The doctor's response (nope, great) is interpreted as follows:

In most cases, the attitudinal labels used by the authors to describe meanings conveyed by this data are intuitively right, but the step from ‘marked’ tone choice to interpretation is often an unexplained step. The step from the textual meaning of tone 3 (coordinating) or the interpersonal meaning (non-committal) to ‘less judgemental’ than tone 1 and ‘less patronising’ than tone 5 is not explained. As H&G point out: ‘(s)uch prosodic features are not random and they are not paralinguistic’ (p. 94). In my view they are pragmatic.

Chapter 5 deals in greater depth with the way in which intonation contributes to ‘textual’ meaning, i.e. signalling interclausal relationships. The discussion of what constitutes a tone group is confusing – on the one hand it is suggested that a tone unit normally maps onto a clause and yet the examples quoted appear to mark boundaries before accented syllables, even when they are word-internal, as in this example:

H&G note explicitly that ‘every tone unit must begin at a foot boundary, i.e. with a salient syllable’ (p. 101). If this is indeed what is meant, it reveals a very different analysis from what is widely practised in intonation research, in both the British tradition and the AM tradition. It is well known that boundaries are not easy to identify, but most researchers agree on the presence of ‘final lengthening’ as a cue, and also that a sequence of unstressed syllables may be distributed across two domains, the boundary being marked by a change in pitch level. Unstressed syllables at the beginning of a tone unit are seemingly disallowed by the model described here. In the account of the interaction between tone groups and accent placement, H&G rightly point out the powerful meaning potential of their combined forces. They suggest an evolutionary explanation for the packaging of discourse into quanta of discourse, i.e. ‘one quantum of experience. . . equals one quantum of exchange’ (p. 108) While somewhat at odds with tone units that could begin in the middle of a word, this kind of reflection, offered almost as an aside, would have been worthy of further development.

A large section of the chapter is devoted to showing that there is a pattern of typical, ‘unmarked, relationships between the semantic categories of (major) speech function and the grammatical categories of mood’ (p. 123). A number of examples are given, with accompanying sound clips, illustrating default associations, for example of a falling tonic with a declarative, and marked examples together with an interpretative gloss. These glosses are reminiscent of early pedagogical accounts of English intonation, e.g. O'Connor & Arnold (Reference O'Connor and Arnold1961), and are equally problematic. The problem lies in trying to specify those meanings generated by marked association of tone and mood. As soon as you give one isolated example, e.g. ‘the combination of a tone 5 with an imperative clause means something like “have you thought of doing that?” or “don't expect me to do it for you’”, it is immediately possible to find others which in their context mean something different: ‘Leave the /\room please’Footnote ⁴ is unlikely to be glossed with either of those meanings. In other words, the only claim we can make is that the combination is marked, and what the speaker means has to be inferred. H&G agree that there is a ‘constant meaning associated with each of the different tones, deriving . . . from the basic opposition between rising and falling’ (p. 123). They also agree that this meaning is very general (something like complete, incomplete or, as they also say, unconditional assertion, conditional assertion, p. 128) so to get closer to the meaning we look at the combination of tone and grammatical environment. This is indeed a useful intermediate step, but the specific meaning is highly context-dependent and giving examples out of context is, I think, misleading.

The final section of chapter 5, explaining how tone choice reflects how the speaker perceives the relationship between successive grammatical units, is a useful one. H&G show how intonation can signal paratactic and hypotactic relationships, and then deal with semantic relationships between clauses such as expansion, elaboration and extension. Elaboration, for example, is indicated by tone concord. Such relationships are also signalled by the overall pitch height and range of a clause (e.g. Wichmann Reference Wichmann2000) but this is clearly outside Halliday's system.

Chapter 6 is a detailed analysis of a short spoken text, a fragment of a radio talk show. The illustrations, using waveforms or spectrograms and superimposed pitch contours, are not as clear as they might be, and could have been more helpful if the alignment of segments with pitch contour had been identifiable. The analysis is broadly helpful, but the interpretation still relies to some extent on context, including paralinguistic cues, such as a ‘slight hint of irony’ and the tone is ‘warm’ (p. 152). This is admittedly part of an attempt to set the conversation in its context, including the genre, the power relations between speakers, and the function (e.g. partly phatic) of the exchange. Other interpretations rely on more elusive criteria, as in: ‘he sounds friendly, with a jovial voice quality, and the steadily descending pitch of his Pretonic suggests that he is confident and relaxed’ (p. 154).

In the final chapter of part II, chapter 7, the authors discuss ‘secondary’ tones, such as falls starting high, mid or low, classed respectively as ‘strong’, ‘neutral’ and ‘mild’, when associated with declaratives. They also describe different ‘pretonic patterns’ – they suggest three variants: ‘steady’, ‘bouncing’ and ‘listing’, again very reminiscent of holistic descriptions in the early pedagogical literature. The different effects of pretonic patterns together with different falls are interpreted attitudinally – ‘drawing attention, unexpected’, ‘serious’ and ‘unmarked, lacking in overtone of absurdity or unexpectedness’. Such interpretations are subjective, and rely on the overall context, as the authors themselves state, but there is still room for a closer look at why certain patterns give rise to certain meanings in certain contexts. This makes it all the more tantalizing that a short section on phonaesthesia observes the connection between high fall and ‘strong’ but doesn't make explicit why this is iconic. Given the current interest in iconicity and intonation, as for example in work by Gussenhoven (Reference Gussenhoven2004), an extended discussion from such experienced linguists would have been most welcome.

The appendix to part II again offers worked analyses of spoken data, and is followed by a short part III, consisting of an overview of the system – intended as a checklist for the analysis of spoken discourse – and further worked examples.

According to the blurb, this book was written ‘for scholars interested in language but not necessarily linguists or phoneticians’. It presents a comprehensive overview of a model of intonation firmly integrated within Functional Systemic Linguistics. It is a book that makes only passing reference to other models and is thus largely self-referential. Much of what the authors say, however, has wider validity than just within its own framework. It relates intonation to a functional view of language in general, thus ensuring that readers understand not only the minutiae of intonational analysis but also its place within a broader theory of language. Some of the discussions touched upon deserved more extended treatment, especially in relation to the evolution of language. The discussion of iconicity is also tantalizingly brief, but there was no doubt a difficult balance to be sought between focusing on the detail of the analysis, including worked examples, and setting them within a broader theoretical framework. This book is idiosyncratically organized, and an index would have been helpful; the bibliography is limited, but carefully chosen, and reflects the particular focus of the book. The material provided on the accompanying CD is of very great value, especially since non-specialists are mostly obliged to take intonational analysis on trust. With this book, it is possible to follow, and in some cases, as I have shown above, dispute the authors’ analysis, but on balance it is far better that the data should generate discussion and controversy than that it should be obscured.