The description of English intonation has developed along different lines in the US and in Britain (and, indeed, different US and British dialects may follow different intonation patterns).
British approaches British descriptions of English intonation can be traced back to the 16th century. Early in the 20th century the dominant approach in the description of English and French intonation was based on a small number of basic "tunes" associated with intonation units: in a typical description, Tune 1 is falling, with final fall, while Tune 2 has a final rise. Phoneticians such as H. E. Palmer broke up the intonation of such units into smaller components, the most important of which was the
nucleus, which corresponds to the main accented syllable of the intonation unit, usually in the last lexical word of the intonation unit. Each nucleus carries one of a small number of nuclear tones, usually including fall, rise, fall-rise, rise-fall, and possibly others. The nucleus may be preceded by a
head containing stressed syllables preceding the nucleus, and a
tail consisting of syllables following the nucleus within the tone unit. Unstressed syllables preceding the head (if present) or nucleus (if there is no head) constitute a
pre-head. This approach was further developed by Halliday and by O'Connor and Arnold, though with considerable variation in terminology. This "Standard British" treatment of intonation in its present-day form is explained in detail by Wells and in a simplified version by Roach. Halliday saw the functions of intonation as depending on choices in three main variables:
Tonality (division of speech into intonation units),
Tonicity (the placement of the tonic syllable or nucleus) and
Tone (choice of nuclear tone); these terms (sometimes referred to as "the three T's") have been used more recently. Research by Crystal emphasized the importance of making generalizations about intonation based on authentic, unscripted speech, and the roles played by prosodic features such as tempo, pitch range, loudness and rhythmicality in communicative functions traditionally attributed to intonation alone. The transcription of intonation in such approaches is normally incorporated into the line of text. A typical example would be: :::::We ˌlooked at the ↗sky | and ˈsaw the ↘clouds In this example, the | mark indicates a division between intonation units. An influential development in British studies of intonation has been Discourse Intonation, an offshoot of
Discourse Analysis first put forward by David Brazil. This approach lays great emphasis on the communicative and informational use of intonation, pointing out its use for distinguishing between presenting new information and referring to old, shared information, as well as signalling the relative status of participants in a conversation (e.g. teacher-pupil, or doctor-patient) and helping to regulate conversational
turn-taking. The description of intonation in this approach owes much to Halliday. Intonation is analysed purely in terms of pitch movements and "key" and makes little reference to the other prosodic features usually thought to play a part in conversational interaction.
American approaches The dominant framework used for
American English from the 1940s to the 1990s was based on the idea of pitch phonemes, or
tonemes. In the work of
Trager and Smith there are four contrastive levels of pitch: low (1), middle (2), high (3), and very high (4). (The important work of
Kenneth Pike on the same subject had the four pitch levels labelled in the opposite way, with (1) being high and (4) being low). In its final form, the Trager and Smith system was highly complex, each pitch phoneme having four pitch allophones (or allotones); there was also a Terminal Contour to end an intonation clause, as well as four stress phonemes. Some generalizations using this formalism are given below. The American linguist
Dwight Bolinger carried on a long campaign to argue that pitch
contours were more important in the study of intonation than individual pitch levels. • Normal conversation is usually at middle or high pitch; low pitch occurs at the end of utterances other than yes–no questions, while high pitch occurs at the end of yes–no questions. Very high pitch is for strong emotion or emphasis. Pitch can indicate attitude: for example,
Great uttered in isolation can indicate weak emotion (with pitch starting medium and dropping to low), enthusiasm (with pitch starting very high and ending low), or sarcasm (with pitch starting and remaining low). • Declarative sentences show a 2–3–1 pitch pattern. If the last syllable is prominent the final decline in pitch is a glide. For example, in
This is fun,
this is is at pitch 2, and
fun starts at level 3 and glides down to level 1. But if the last prominent syllable is not the last syllable of the utterance, the pitch fall-off is a step. For example, in
That can be frustrating,
That can be has pitch 2,
frus- has level 3, and both syllables of
-trating have pitch 1.
Wh-questions work the same way, as in
Who (2) will (2) help (3↘1)? and
Who (2) did (3) it (1)?. But if something is left unsaid, the final pitch level 1 is replaced by pitch 2. Thus in ''John's (2)
sick (3↘2) ...
, with the speaker indicating more to come, John's
has pitch 2 while sick'' starts at pitch 3 and drops only to pitch 2. • Yes–no questions with a 2↗3 intonation pattern usually have subject-verb inversion, as in
Have (2) you (2) got (2) a (2) minute (3, 3)? (Here a 2↗4 contour would show more emotion, while a 1↗2 contour would show uncertainty.) Another example is
Has (2) the (2) plane (3) left (3) already (3, 3, 3)?, which, depending on the word to be emphasized, could move the location of the rise, as in
Has (2) the (2) plane (2) left (3) already (3, 3, 3)? or
Has (2) the (2) plane (2) left (2) already (2, 3, 3)? And for example the latter question could also be framed without subject-verb inversion but with the same pitch contour:
The (2) plane (2) has (2) left (2) already (2, 3, 3)? •
Tag questions with declarative intent at the end of a declarative statement follow a 3↘1 contour rather than a rising contour, since they are not actually intended as yes–no questions, as in ''We (2) should (2)
visit (3, 1) him (1),
shouldn't (3, 1) we (1)?
But tag questions exhibiting uncertainty, which are interrogatory in nature, have the usual 2↗3 contour, as in We (2) should (2)
visit (3, 1) him (1),
shouldn't (3, 3) we (3)?'' • Questions with
or can be ambiguous in English writing with regard to whether they are either-or questions or yes–no questions. But intonation in speech eliminates the ambiguity. For example,
Would (2) you (2) like (2) juice (3) or (2) soda (3, 1)? emphasizes
juice and
soda separately and equally, and ends with a decline in pitch, thus indicating that this is not a yes–no question but rather a choice question equivalent to
Which would you like: juice or soda? In contrast,
Would (2) you (2) like (2) juice (3) or (3) soda (3, 3)? has yes–no intonation and thus is equivalent to
Would you like something to drink (such as juice or soda)? Thus the two basic sentence pitch contours are rising-falling and rising. However, other within-sentence rises and falls result from the placement of prominence on the stressed syllables of certain words. For declaratives or wh-questions with a final decline, the decline is located as a step-down to the syllable after the last prominently stressed syllable, or as a down-glide on the last syllable itself if it is prominently stressed. But for final rising pitch on yes–no questions, the rise always occurs as an upward step to the last stressed syllable, and the high (3) pitch is retained through the rest of the sentence.
The ToBI system A more recent approach to the analysis of intonation grew out of the research of
Janet Pierrehumbert and developed into the system most widely known by the name of
ToBI (short for "Tones and Break Indices"). The approach is sometimes referred to as
autosegmental. The most important points of this system are the following: :*Only two tones, associated with pitch accents, are recognised, these being H (high) and L (low); all other tonal contours are made up of combinations of H, L and some other modifying elements. :*In addition to the two tones mentioned above, the phonological system includes "break indices" used to mark the boundaries between prosodic elements. Breaks may be of different levels. :*Tones are linked to stressed syllables: an asterisk is used to indicate a tone that must be aligned with a stressed syllable. :*In addition, there are phrasal accents which signal the pitch at the end of an intermediate phrase (e.g. H− and L−), and
boundary tones at full phrase boundaries (e.g. H% and L%). :*A full ToBI transcription includes not only the above phonological elements, but also the acoustic signal on which the transcription is based. The ToBI system is intended to be used in computer-based transcription. A simplified example of a ToBI transcription is given below. In this example, two phrases "we looked at the sky" and "and saw the clouds" are combined into one larger intonational phrase; there is a rise on "sky" and a fall on "clouds": :::::L*L*H−H*H* L−L% :::::we looked at the sky and saw the clouds Because of its simplicity compared with previous analyses, the ToBI system has been very influential and has been adapted for describing several other languages. ==French==