Review Article on Brazil et al.

Discourse Intonation and Language Teaching

This appeared in 1986 in Vol. 16 at pp 54-62 of the  Journal of the International Phonetic Association.
The following version contains a few minor emendations etc.

D. BRAZIL, M. COULTHARD and C. JOHNS, Discourse Intonation and Language Teaching. (Pp. xi + 205. Paperback, £4.90. ISBN 0 582 55366 O. Longman, London, 1980.) Accompanied by a cassette containing two transcribed recordings and excercise material.

The first chapter of this book is entitled 'Tone'. The authors gloss the term as MAJOR PITCH MOVEMENT but do not define it any further. They go on to remark that 'there is always and only one tone in each tone unit' (p.13). They recognise five tones and deal first with the two they find most fundamental, the FALLING and the FALLING-RISING. They report finding that 'these two tones are by far the most frequent found in our data' (p.13) though the two types are not predominant in their transcripts (pp. 145-197). Nor are they so reported by Crystal (1969:225); see also Overall (1985:17). It is suggested that these two tones 'embody the basic meaning distinction carried by tone' (p.13). After various illustrative examples, the authors summarise the chapter by saying that they have offered: '[an] interactional explanation of the significance of tone [in which] tone choice [is seen as] not dependent on linguistic features of the message but rather on the speaker's assessment of the relationship between the message and the audience .... [making] moment by moment decisions to refer to sections of his message as common ground or to proclaim them as an addition to it' (p.18). This conviction leads them to adopt a tone indication system which is semantics-based, unlike most others which are pitch-orientation based. Their intonation notation has the following features:

(i) initial and final unit boundaries are signalled by pairs of slashes. There is no mention of why two obliques are used, and they make no differential use of a single slash. A triple set is used later (in unexplained alternation with three-line-height vertical bars, eg at p.65) to signal the boundary of what they call a 'pitch sequence'. It will be remembered that the 'double slash' is also the TONE GROUP BOUNDARY in Halliday's (1970) representations, though the rationale there is that a single oblique is his foot boundary marker;

(ii) the tone choice is indicated by p or r immediately after the initial boundary sign. Later in the book three other signs are used: p+ for a rise-fall tone, r+ for a rising tone and o for a tone referred to variously as neutral, level or zero. The letters p and r stand for PROCLAIMING and REFERRING;

(iii) 'the tonic syllable .... is underlined and capitalised' (p.16). We see later at p.39 that capitalisation without underlining is used to identify pre-nuclear accents, here termed 'prominent' syllables. This use of capitals, which follows a convention introduced by Crystal, entails their withdrawal from their normal orthographical applications.

Chapter 2, 'Key', deals with the pitch elevation character of the (beginning of) the unit. Three significant pitch elevation contrasts are recognised; high, mid and low. Most notably in their usage they say that MID-KEY 'is not defined as the norm for the speaker'. Choices are made with reference to the pitch of the preceding unit: 'there are no absolute values for high, mid and low key, even for a particular speaker' (p.24). The drawings on p.25 representing their high, mid and low proclaiming and referring tones resemble the Kingdon (1958) emphatic, high and low falls and fall-rises (they are, unlike similar idealised pitch-meter traces used later, not given identification numbers). The authors suggest that the speaker's choice of key as high, mid or low corresponds exactly and only to his wish to mark the matter of the unit semantically as respectively CONTRASTIVE, EQUATIVE or ADDITIVE (see especially pp.42 and 64). The suggestion that when 'a speaker begins in low key he indicates ... equivalence ... between what he says' and something the previous speaker has said is acceptable in respect of many situations but it surely must often also not be the case. For example in reply to It's 'just `right the comment It's too `big may employ low key merely from disanimation, eg reflecting disappointment. Cf Cruttenden 1986:117. Their labels show advances on the excessively narrow and context-bound ones of many previous writers in the field of intonation semantics but their claims for them tend to be very sweeping. The approach recalls Jassem (1952:70) where the remarks 'falling nuclear tunes have a proclamatory value' and 'rising nuclear tunes have evocative value' occur. Jassem (1952) was referred to in Brazil (1975) but in the present work no debt to its ideas is mentioned .

Chapter 3 on 'The tone unit', though it says it deals with its NATURE, STRUCTURE and FUNCTION, is mainly devoted to attempts to convey what the authors wish to be understood by their term PROMINENCE. They explain it in semantic terms, eg: 'Making any word prominent ... constitutes a meaningful choice' and say that it represents the speaker's assessment of the 'relative information load carried by particular elements in his discourse.' Despite their claim of difference from previous concepts it is difficult to see how their term differs notably from the FULL STRESS of Kingdon 1958 or ACCENTED as employed in O'Connor & Arnold 1961,1973. Another suggested novelty of analysis is that 'we see all intonation meaning as carried by the tonic segment' (p.42). This seems to be either very obvious or an unexplicit rejection of what Kingdon called 'special preheads'. If the latter it is unfortunate to omit further discussion of what it implies.

In the next section they suggest that 'there is a consistent relationship between pitch variation and the distribution of prominence' (p.43) and purport to supply acoustic evidence of the truth of their assertions by offering for comparison a set of idealised versions of pitch-meter fundamental frequency traces. When the corresponding recordings were subjected to analysis in our phonetics laboratory at Leeds, the traces obtained failed to match the idealisations offered in a variety of ways. Notably the non-'prominent' go of example (25) on p.44 did not give the level trace they represent but the same kind of arc-shape as was found with their items (23) and (24) where it is 'prominent'. For further details see Overall (1985). The kind of pictorial representation used would be perfectly acceptable as pedagogically helpful simplification of the presentation of matters that were universally accepted as factual, but as an attempt to establish a controversial point it seems remarkably ill-judged. At any rate their claim to have established a 'general basis for distinguishing prominent syllables from tonic syllables' is quite unjustified and their transcriptions show large numbers of cases where their choice between showing a tone as onset or tonic appears to be quite arbitrary. See Overall (1985: Chapter 5).

The section headed 'Key' seems again to be appealing pointlessly to instrumental analysis, this time to confirm the surely obvious suggestion that whereas fall and fall-rise tones are to be classified as high, mid or low according to the values of their initial pitches, the pitch classification of rise-falls and rises may be based on their peaks. The final section is headed 'tone unit boundaries', which they declare, despite their emphatic notation, to be 'not of great importance'. Their suggestion that 'all intonational meaning is carried by the tonic segment' (p.45) will not stand up, since the peak of a rise-fall will often occur in the 'enclitic' segment. Boundaries are shown in numbers of cases where the rhythmic integration of the words on either side of the double slash is maximal, for example on the several occasions which assign the weak article the to a separate unit from its following substantive.

Chapter 4, 'More on tone', begins with the observations that choice of tone can 'carry the social meanings of convergence/divergence, or solidarity/separateness' (p.51) and that some items like // r ACtually // 'serve to insinuate intimacy or solidarity' (p.51), while // p SUREly // may be used 'in emphasising the likely lack of agreement on a point' (p.52). These seem to be interesting suggestions.

The main part of the chapter is concerned with introducing 'the rise, symbol r+ and the rise-fall p+ ... accounted for as variants of the referring and proclaiming tones' (p.52). They contrast the r tone whose common ground references are said to be 'vividly present background' (p.53) with r+ which is described as referring to matter which, while 'present in the area of convergence, has need of reactivation' (p.53). They also say that the DOMINANT role relationship, meaning that of 'the person who has greater freedom in making linguistic choices' (or is so claiming) is expressed by the use of r+ tone (p.53). Thus r and r+ are distinguished 'most generally ... by reference to their social implications' (p.55). The explanation they offer of the rise-fall p+ 'option' rejects the usual intonation-manuals label of 'surprise' as relevant only 'in certain circumstances' (p.56). They prefer to say that the p+ tone signals 'that the speaker is simultaneously adding information to the common ground but also to his own store of knowledge' or that his information is 'doubly new', or that he is suggesting: 'that alters my world view' (p.56). These rather extravagant expressions fit ill with many possible applications of p+, eg // p+ that's NOthing new // or // p+ it's BORing // or // p+ I've HEARD them all //. Once again no details are given of the physical characteristics of the two tones. This is particularly unfortunate as regards p+ since only 16 illustrations of it are offered in the book of which only six contain any audible rise element (see Overall 1985:45).

Chapter 5, 'Key and termination', observes that the speaker can 'choose pitch level meaningfully twice in a single tonic segment' (p.60). Just as they use the term KEY for the pitch elevation value of the first 'prominent' word in the tone unit so they assign to the 'pitch of the tonic syllable' the expression TERMINATION. They also set up a term for a 'phonological unit larger than the tone unit' (p.61) suggesting that:'[the] downward drift of pitch is ... exploited as an organising mechanism producing a phonological unit of indefinite length ... the PITCH SEQUENCE, which begins immediately following a tone unit with LOW TERMINATION and includes all succeeding tone units until the next one with low termination' (p.61). It is difficult to share their enthusiasm for identifying unity in such amorphous groupings but the reference to 'sequence-ending implications of low termination' makes good sense. The examples offered in this chapter make a reasonable case for the usefulness of classifying high, mid and low items as respectively CONTRASTIVE, ADDITIVE and EQUATIVE at least in many circumstances if not so universally as the authors claim.

Chapter 6, 'Intonation and discourse structure', turns to 'discussion of the function of intonation choices at points of speaker change' (p.73) with reference to the model of discourse structure of Sinclair and Coulthard (1975), a description developed to analyse TEACHER-PUPIL interaction. Most of the chapter is concerned with matters of pitch elevation ('key' and 'termination'). Though it occupies the authors of this book so extensively, and though it is put forward as something which will be useful to EFL teachers to study, it is surely one of the features of intonation least likely to be language-specific. The observations on PITCH CONCORD between MOVES are very interesting (MOVES are defined as contributions from at least two participants falling into one of three major classes viz: opening, answering and follow-up). 'High termination, ... at the end of an utterance, anticipates high-key opening in the next utterance' (p.76). Mid expects mid, low leaves the choice open. There are only half a dozen examples in this chapter but at least two of them cause misgivings. A high-key response Time to go to a mid-termination remark it's three o'clock is described as 'asking' and 'in some sense questioning' (p.77) but as performed it doesn't sound in the least necessarily interrogative. Another item is 'a typical high key, proclaimed, moodless initiation', as a doctor 'seeks more information' viz: Tight pain. In this context the choice of referring tone, it is claimed, 'would be positively worrying to the patient' (p.77). One should rather say that, although it could be so, with appropriate tempo and voice-quality characteristics — two features of speech totally neglected by these authors — it need not at all be so.

The chapter on 'Reading intonation' begins with some thoughtful comments on the relation that the reader assumes with his text when reading aloud: "he can either 'perform' it as if he himself were speaking or he can stand outside it saying “this is what the text says” "(p.83). The middle section of the chapter, entitled 'Orientation', readdresses itself to the 'fundamental theoretical distinction' between orientation towards the hearer which they call DIRECT ORIENTATION and towards the language which they call OBLIQUE ORIENTATION. The former we are told involves choosing between 'referring and proclaiming tones' but the latter 'between proclaiming and neutral tones' (p.89). Now for the first time (p.88) we meet their fifth and final tone type, the neutral or level tone, symbolised by o (and referred to less happily as ZERO TONE in the tapescript). Of a seven-line passage whose dozen marked p/o possibilities are said to be read with zero at all choices, only one is read as described! The final section of the chapter, headed 'Pitch and pitch sequences', turns from tone selection to PITCH-LEVEL PHENOMENA. The examples given contain gross reading inaccuracies (copious in No. 28); No. 27 is referred to as 'the special case of a word which might be expected to be non-prominent being made tonic ... to achieve sequence closure'. One can only say it does so at the sacrifice of well-formedness. The chapter ends with another example performed very unrealistically. This and various other recorded items could be misleading or at least puzzling to foreign teachers who expect to receive illustrations of acceptable usages.

Chapter 8, 'A comparison with two other descriptions of intonation', contains the remark that for them the 'physical features of intonation have always been regarded as a secondary area of investigation'. Not an understatement as we have seen: Cruttenden (1986:123) is obliged to speculate about the nature of rises in Brazil 1975. They limit their comparisons with O'Connor and Arnold's Intonation of Colloquial English (1973), hereafter ICE, to only three of its 10 tone 'groups'. First the rise-fall type on which they quote, not entirely accurately, various ICE semantic labellings, fairly justifiably suggesting that they are rather too local to the contexts they occur in to satisfactorily express the semantic value of the tone. But one doubts that their single expression 'acknowledging an incursion into the speaker's world' is more helpful than the ICE plethora of terms. The rest of their attention to ICE is devoted to the two patterns called by its authors the LOW DROP and HIGH DROP, with the difficulty that, translating its two-height distinctions into their three, they cannot be sure (they say of the 'low drop') 'whether we are dealing with cases of (mid key +) low termination or (high key +) mid termination' (p.100). They find that in the ICE recording the second speaker in Did you lock the back door? - Of course produces a mid-termination of course 'which expects a further response'. It's not certain that everyone would accept that this expectation is a necessity. Incidentally they, without comment, convert the question from a simple rise nucleus (their r+ ) to a fall-rise. There are no fall-rise questions in ICE at all. At p.101 they give the example (9) // A: p that's (sic for ICE What) a VERy nice HOUSE // B: p YES // p ISn't it // and say 'the second speaker elects to approve the verdict with the contrastive high key “yes before agreeing with it in mid key'. The teacher unaware of the 'meaning-changing significance' of such 'different key selections' is, they say, 'likely to increase rather than reduce the student's confusion'. May not confusion result from their describing the answer 'yes' as 'contrastive' in such a context? Their comment (pp.102-103) that when ICE's authors attribute 'seriousness' or 'urgency' to their examples, they seem to be describing 'not the effects of the intonation so much as the situation which the lexis ... evokes' is justified though of a type often made before. But when they suggest (p.102) that certain kinds of question with high, mid and low termination expect hearer-responses of the type you-are-right, I-agree and nil respectively they seem to be offering a new insight. Even so they are probably too insistent on pitch alone as the divider and they don't refer to the very likely wide ambiguous overlaps. Similarly they make an interesting suggestion with their contention at p.103 that high termination on eg Can I help you can represent an unresolved issue while mid presents it as already resolved. There are some puzzling references to heard values (eg at p.100 'the second speaker in fact produces ...') in reference to items apparently not included in the ICE recordings. Unlike the pages discussing ICE, the 8-page discussion of Halliday (1970) in this chapter contains helpful page references to the text discussed. In it they 'try to demonstrate that our method of representing intonation as a simultaneous but quite separate set of options ... enables us to say more and say it more economically' than Halliday. They quote Halliday's glosses for his Tone 4 (their r tone) 'expressing reservation, contrast or personal opinion' and point out that they are compatible with the 'common-ground implications' of their r tone. Adding a follow-up example of their own they note that r+ tone is more likely in: (26)

A: //r i DON'T like the new OFfice block //

B:// r+ it's VERy efFICient //

where both dominance and the sense of invoking some quality which will surely be shared are present in B's reply. Unfortunately, none of the examples in this chapter are included on the cassette. Whether it truly sounds dominant must perhaps depend largely on prosodic features they don't deal with, but in any case it makes a good demonstration of the awkward ambiguities of their notation. With its 'key' mid, as here, its pretonic prominent word very can either have two upper pitches (either the same or stepping down) or low pitch on its first and mid pitch on its second syllable. The dominance will be expressed in the former case haughtily and in the latter patronisingly. This highlights two weaknesses in their theory: (i) that (p.10) the ONSET SYLLABLE determines the key (it can't for polysyllabic onset words if the reference-point diagram for r+ is to be taken seriously) and (ii) that pre-nuclear pitch MOVEMENT is not communicatively significant (p.9).

About ten Halliday examples are discussed re-illustrating points already made earlier rather than breaking new ground. In several places they seem not to have indicated in their translations certain of Halliday's 'salient' syllables as 'prominent', notably at examples 27, 28, 40, 41, 58, 59 and 60. (Actually 40 and 41 contain obvious mistakes of theirs. But their transcriptions of 27, 28, 58 and 59 correspond to Halliday's own reading better than his does, though they compound his errors in 33 and 34.)

The final chapter 'Intonation and language teaching', twice as long as any of the others, has as its main theme that 'the omission of instruction on intonation, and indeed other aspects of the sound system of English, is a striking characteristic of general course books and ESP materials'. Its second section on 'Accounts of intonational meaning in teaching materials' begins with a brief recapitulation (with the same misquotation of quietly for greatly) of the first point made about ICE in the previous chapter. The main contention is that 'There are no arguments for teaching intonation in terms of attitude because the rules for use are too obscure, too amorphous, and too easily refutable.' A few examples are given, repeats from earlier in the book, which illustrate well the point that 'certain grammatical features of sentences' act as clues to the discourse setting (p.121). As before with ICE, now with Armstrong and Ward, they speak of them as if they had made very sweeping statements and say 'It is not difficult to refute these generalisations', ignoring the various disclaimers made by the authors in question. A section on 'Approaches to devising materials' repeats their claim that compensation must be made for 'the marked dominance of the teacher in all spoken classroom interaction' but they may well underrate the student's ability to recognise this fact and allow for it. Most of the rest of the final chapter is devoted to 'An intonation syllabus'. The considerable number of contrastive analyses of English intonation with other languages that are available are ignored. At 'Exercise types' some of the items seem rather unsuitable. For example: (p.139)

// r THEY'RE not // r MINE // would surely be better with a fall in the first unit.

And in: (p.141 )

A // r excuse ME //

B // r YES //

A // p have you seen my BRIEF case anywhere //

surely // r exCUSE me // was intended and surely also r would be better pedagogy for the last remark.

In a brief section on 'Language testing' there is yet another too sweeping remark: 'there is no such thing as a questioning intonation' when so many languages use a high-rising pitch on a non-speech sound like [m] for interrogation. The book closes with two continuous transcripts stretching over fifty pages displaying the cumbersomeness and unreadability of their notation. The amount of information could, Kingdon-style, have been given in half a dozen pages. The longer transcript is a 13-minute near-monologue by a primary-school teacher. The two dozen brief interactions from his pupils are almost all inaudible; at least only just over two dozen words of them are transcribed, and so far from interacting the speaker repeatedly either verbally brushes aside or totally ignores the pupils' interjections. The shorter transcript is a 4-minute monologue addressed by the MP Denis Healey to an interviewer. Both are strange choices for a book which places so much emphasis on the interactive aspect of intonation.

This stimulating work is without doubt one of the most notable recent contributions to intonation semantics. However, as Overall (1985) has shown, detailing over 700 non-trivial errors of notation or performance (including over 13% of all tone ascriptions), it badly needs re-writing since its inadequacies of execution come very near to marring the whole enterprise. At the same time if various of its claims to have solved long-standing problems were tempered to something more realistic, its real merits might be more easily recognised. Also more realistic would be a new title with no reference to language teaching. Not a single EFL real-life intonation problem gets a mention anywhere in the book.


ARMSTRONG, L. E. & WARD, I. C. (1926). A Handbook of English Intonation. Cambridge: Heffer.

BRAZIL, D. C. (1975). Discourse Intonation. Discourse Analysis Monographs 1, Birmingham, English Language Research: University of Birmingham.

CRUTTENDEN, A. (1986). Intonation. London: Cambridge University Press.

CRYSTAL, D. (1969). Prosodic Systems and Intonation in English. London: Cambridge University Press.

HALLIDAY, M.A.K. (1970). A Course in Spoken English: Intonation. London: Oxford University Press.

JASSEM, W. (1952). Intonation of Conversational English. Wrocław, Poland: Wrocław Society of Sciences & Letters.

KINGDON, R. (1958). The Groundwork of English Intonation. London: Longman.

O'CONNOR, J. D. & ARNOLD, G. F. (1961). Intonation of Colloquial English. (Second Edition 1973.) London: Longman.

OVERALL, C.S. (1985). An Examination of Brazil et al. Discourse Intonation & Language Teaching. Unpublished dissertation, University of Leeds.

SINCLAIR, J. M. & COULTHARD, R. M. (1975). Towards an Analysis of Discourse. London: Oxford University Press.