Faculty of Arts

University of Ljubljana








A paper for an MA Seminar in Phonology




October 1998



Traditionally, according to Roach, intonation has been described as having two main functions: grammatical, and emotional or attitudinal, i.e. following the work of Halliday, and O’Connor and Arnold, respectively. In this paper I will compare the work of O’Connor and Arnold with that of David Brazil, who advocated a communicative approach and insisted upon the contextual significance of intonational meaning. In the EFL environment it is customary to teach a general model of intonation. Following the comparison of the two advocated models, I will propose which, if either or both, of the two models of intonation should be applied in the teaching of intonation to EFL students in Slovenia.


Many of the best-known descriptions of English intonation in the past have regarded the attitudinal function as primary and central, i.e. O’Connor and Arnold, for example. The main part O’Connor and Arnold’s work was dedicated to the attitudinal meaning of intonation, i.e. attitude which can be detected in a speaker’s voice: whether the speaker is angry, or civil, or grumpy, or enthusiastic. They ascribed different meanings to various patterns; for example, something is ‘categorical, weighty, judicial or considered’, or it can be ‘impressed, awed, complacent, self-satisfied, challenging, censorious, disclaiming responsibility’.

According to Tench (1996: 112), they thus have an ‘intonational lexicon’ of twenty patterns, each of which has a range of meanings. For all the patterns, however, different ranges of meaning can be ascribed to statements, wh-questions, yes/no questions, commands and interjections. He concludes that the range of different meanings seems so wide that it may actually be too wide to posses any general sense.

According to Roach (1996: 45), the claim that “…we use intonation to express our attitudes is fundamentally wrong”. He then adds that work by phoneticians on emotions and attitudes in speech has tended to have a rather simplistic view of the subject, and it has become perhaps rather commonplace among phonology teachers to quote some of the “...more speculative and unscientific statements which were made by O’Connor and Arnold” (Roach, ibid.).

As Tench (1996) points out, many of the supposedly attitudinal functions of intonation, which have been suggested by earlier writers, are in fact better viewed as interactional. We therefore need studies to show how people use what we usually think of as emotional or attitudinal expression in order to achieve some interactional result, for instance ‘sounding reluctant to agree’, ‘sounding friendly’, ‘sounding doubtful’ (Roach: 48).

On the other hand, even Roach seems to believe that there is a place for attitude after all. “Emotions”, he says, “typically make themselves evident in spite of the speaker’s attempts to control them…while attitudes are usually deliberately displayed” (Roach: ibid.) which takes into account Couper-Kuhlen’s proposition that emotions should be regarded as related to the speaker’s inner state, while attitude relates to the speaker’s observable behaviour.

In the past, the attitudinal function of intonation was the main feature of intonation model taught to students of English as a foreign language. The students were even taught that using inappropriate or “wrong” intonation pattern in a certain social environment might result in offense being taken by a person spoken to. Such a view, for obvious reasons, probably caused a lot of anxiety in learners of English. According to Couper-Kuhlen (1996), emotions, just like facial expressions or other body language, make themselves evident in spite of a speaker ‘s attempts to control them. On the other hand, attitudes are usually deliberately displayed. Given that they are intentionally shown to influence or even manipulate people, claims Roach (1996: 48), they should have no place in the teaching of intonation.

Other descriptions of English intonation, including the work by David Brazil, have made the organisation of information and discourse central, but have nevertheless acknowledged the additional component of attitude.


Discourse intonation (DI) is a theory of intonation which relates ‘stress’, ‘tone’, and ‘pitch height’ to categories of meaning. Discourse intonation, as proposed by David Brazil, attempts to make the simplest possible description, and it attempts to adopt the language user’s, not the linguist’s perspective: contextual factors are of paramount importance, and the speaker’s perceptions are central. Brazil (1985: 238) claims that there is a “…need for stating the communicative value of intonation in terms of the projected contextual implications of the tone unit: only if we regard intonation as a ‘situation-creating’ device, … can we give proper recognition to its ability to carry independent meanings”. He also says that prominent syllables, like tones, are distributed on the basis of what context of interaction the speaker chooses to project.

DI therefore does not aim to provide a link to categories of grammar, neither does it attempt to establish links between attitude and intonation, but it offers a way of accounting for them outside its own systems. Cauldwell and Hewings (1996: 51) claim that meanings like surprise, irony, sarcasm, grumpiness are therefore features of particular contexts and are not attributable to any one choice, such as a high-falling tone.

Discourse intonation takes the stance that most speech is divided into tone units which have either one or two prominences. Tone units may or may not be separated by a pause. What is important is that there is only one tone in each tone unit: every tone must be in a separate tone unit. DI accounts for four systems: prominence, tone, key and termination, and the maximum number of choices on any one syllable is three. In the next few sections the central issues of discourse intonation, namely that of tone, key and termination and referring and telling, will be briefly addressed.

3.1 Tone

Cadwell and Hewings made a table which illustrates that the last prominence in each tone unit is the location of one of five tones recognised by discourse intonation: the fall, the rise, the fall-rise, the rise-fall and the (mid) level. See table 1 on the following page.









Telling (dominant)


ì î








î ì


Referring (dominant)







mid level


According to Brazil, we choose one of the five or a certain combination. A particular communicative value is associated with each of the tones in a way that holds true for all occurrences of that tone. Another aspect of intonation dealt with by Brazil is the introduction of key and termination, two notions which will be further explained in the following section.

3.2 Key and Termination

KEY is the choice associated with the onset syllable, or the first prominent syllable, while TERMINATION is the choice associated with the tonic syllable, or the last prominent syllable in an intonation group.

Key and termination choices combine freely with all tone choices.

On prominences there is a choice of pitch height. Table 2 on the following page illustrates that pitch height on prominences can be either high, mid or low.












pitch opening








invites adjudication

pitch sequence closure


The effect of termination decides an aspect of the context of interaction that the speaker projects. By specifying how many possibilities of choice there are in an intonation unit, where they are made, and what relations of interdependence exist among them, predictions about its meaning can be made. Brazil says (1985: 112) that “…verbal communication is a linear and cumulative process which has as an intended outcome an increase in the area of speaker/hearer convergence”. Some stretches of discourse constitute a step forward in this process, while others do not.




Discourse intonation makes a basic distinction between telling, i.e. the proclaiming tone, which is associated with the ‘fall’ and ‘rise-fall’, and referring, associated with the ‘rise’ and ‘fall-rise’. The significance of the PROCLAIMING TONE, is that, by producing it, the speaker offers to further the process by changing the hearer’s world. It is not yet present in the common ground, the ground shared by a speaker and a hearer. On the other hand, the REFERRING TONE is already present in the common ground. Referring tone units can also be characterised negatively as those in which the speaker recognises that they are saying nothing which will constitute a step further, i.e. such a feature of discourse will in no way alter the state of speaker/hearer convergence. However, Michael Vaughan-Rees (1996: 61) proposed a distinction-table for the referring and proclaiming tones in so far as he believes the fall-rise to indicate an aspect of commonality between speaker and hearer; “..an excursion into assumed common ground, rather than a step towards the greater convergence..” (Brazil: 115).


î ì



  • associated with old / shared information (previous text, the situation, shared knowledge)


  • associated with new information
  • more is implied, more could be said
  • no more needs saying; this part of the message is complete
  • lifting a barrier, nearing, creating solidarity ® ¬
  • creating a barrier, distancing

¬ ®

Brazil says, however, that when asking questions, for instance, we can choose either a fall (for seeking unknown information) or a fall-rise (for confirmation of what we already believe). “The result will be much the same whichever tone you use. It is very common for people to behave as if they just needed confirmation…we quite often ask questions in order to be friendly, rather than because the answer is of any real importance, or even interest, to us” (1994: 41). Such statements clearly demonstrate an attitudinal aspect in his writings, which will be addressed in the next section.


According to Brazil, we might nevertheless be tempted to abstract from a range of reactions such as surprise, pleasure, annoyance and seek an attitudinal label which we could attach to some highest common factor. The key system enables a speaker to project an existentially valid contrast: to bring into sharp opposition a pair of possibilities and simultaneously exclude one of them. In doing this on a particular occasion, they may betray feelings, or anticipate feelings in their hearer, to which a more-or-less specific attitudinal label can be applied, but such a specific interpretation depends on a multitude of factors.

It is at this point that his attitudinal approach comes out quite clearly. Actually, attitudes are in fact quite often expressed in Brazil’s work: “…Choice of r+ tone…heard as a warmer offer…”; “…the r+ version is more pressing and cordial…”; “…over dominance introduces a superimposed note of irritation or exasperation into the relationship…”; “…of something we might perhaps characterise locally as ‘forcefulness’ or ‘insistence’ can be traced… “

Cruttenden also argues that “… another part of the problem lies in Brazil’s choice of the labels ‘proclaiming’ and ‘referring’, which are heavily discourse-oriented, when, as we have seen, at least some part of the local meanings of almost all tones is attitudinal.

Jenner (1996: 42) nevertheless believes that “…a new breed of pioneers, such as David Crystal, Gillian Brown and, of course, David Brazil, have forced us to attend to the role of intonation in communicative interaction”. Barbara Bradford has shown that intonation can - and perhaps must - be taught in a communicative context. Moreover, Brazil has even proposed an intonation programme as the basis for the teaching of all aspects of pronunciation (Jenner : 43). Before making final remarks about the kind of intonation model which should be taught to EFL students, however, we should perhaps have a look at the position of English as an international language of business and other transactions.


What needs to be addressed at this point is the importance of intelligibility, advocated by various phoneticians and phonologists, since they believe it the most important communicative goal. The realisation that English is becoming, or even already is, a lingua franca, used on a daily basis especially in business and other transactions and international interactions, makes it rather clear that English is no longer the sole property native speakers. Jenkins (1996: 34) even claims that ”…native-like pronunciation can no longer be the ultimate goal of teaching”. It is true that certain errors impede intelligibility, errors that concern not only pronunciation, but also prosody. We will eventually perhaps also need to take the intelligibility of international English into account when creating new programmes for the teaching of English intonation to non-native learners of English. The scope of this paper, however, does not allow me to address this issue in great detail.

The traditional descriptions of pronunciation seemed to have little relevance to the structure of messages, and in any case, work on phonological details got in the way of communicative practice and threatened learners’ self-confidence. Although we need to consider communicational efficiency, we ought also to seek some degree of agreement among present-day pronunciation teachers that communicative, interactional functions are the most interesting aspects of prosody; at the same time “the expression of emotions and attitudes remains as an extra function, ill-defined, but important”, claims Roach (1996: 48).


Certainly, the gap between attitudinal features and communicative approach suggests a move away from a typical intonation model, where students found themselves being regularly taught attitudinal features of intonation and had to use drills as the main means for intonation practice. If this situation is now in changing, it may also be worth changing our methodology in the teaching of English intonation to speakers of English as a foreign language and adopt a new, second-language model of teaching.

If we wish to give a proper portrayal of intonation in English to EFL students, we perhaps ought to combine the proposed models of intonation, and equip speakers of English as a foreign language with the knowledge of different aspects of intonation. Moreover, students should become observers of naturally occurring speech, so that they can become not only better listeners, but also better users of intonation. According to Roach, we ought to show EFL students how people make use of what we understand as emotional or attitudinal expression in order to achieve some interactional result. It is a model of intonation where ideas proposed by O’Connor and Arnold are combined with those advocated by David Brazil that may achieve the best results in the teaching of intonation.



O’Connor, J. D. and Arnold, G. F. (1973) Intonation of colloquial English, Longman Group Ltd., Bristol, U.K.

Brazil, D. (1985) The Communicative Value of Intonation in English, English Language Research, University of Birmingham, U.K.

--- (1994) Pronunciation for Advanced Learners of English, CUP, U.K.

Cauldwell, R. & Hewings, M. Discourse intonation and listening, in: Speak Out! Changes in Pronunciation (Summer 1996), ed. Vaughan-Rees, M. , CUP, Cambridge, U.K., pp. 49-57

Cruttenden, A. (1986, 1997) Intonation, CUP, Cambridge, U.K.

Jenner, B. Changes in objectives for pronunciation teaching, in: Speak Out! Changes in Pronunciation (Summer 1996), ed. Vaughan-Rees, M. , CUP, Cambridge, U.K., pp. 41-45

Roach, P. Emotions, attitudes and the English speaker, in: Speak Out! Changes in Pronunciation (Summer 1996), ed. Vaughan-Rees, M. , CUP, Cambridge, U.K., pp. 45-49

Vaughan-Rees, M. Discourse Intonation: Extending the Definitions, in: Speak Out! Changes in Pronunciation (Summer 1996), ed. Vaughan-Rees, M. , CUP, Cambridge, U.K., pp. 57-62

Tench, P. (1996) The Intonation System of English, Cassell, London, U.K. pp 107-115.