Many topics have been concerned with the description of phonemes, and it was pointed out that the subject of phonology includes not just this aspect (which is usually called segmental phonology) but also several others. We studied stress. Clearly, stress has linguistic importance and is therefore an aspect of the phonology of English that must be described, but it is not usually regarded as something that is related to individual segmental phonemes; normally, stress is said to be something that is applied to (or is a property of) syllables, and is therefore part of the suprasegmental phonology of English. (Another name for suprasegmental phonology is prosodic phonology or prosody.) An important part of suprasegmental phonology is intonation.

What is intonation? No definition is completely satisfactory, but any attempt at a definition must recognize that the pitch of the voice plays the most important part. Only in very unusual situations do we speak with fixed, unvarying pitch, and when we speak normally the pitch of our voice is constantly changing. One of the most important tasks in analyzing intonation is to listen to the speaker's pitch and recognize what it is doing; this is not an easy thing to do, and it seems to be a quite different skill from that acquired in studying segmental phonetics. We describe pitch in terms of high and low, and some people find it difficult to relate what they hear in someone's voice to a scale ranging from low to high. We should remember that "high" and "low" are arbitrary choices for end- points of the pitch scale. It would be perfectly reasonable to think of pitch as ranging instead from "light" to "heavy", for example, or from "left" to "right", and people who have difficulty in "hearing" intonation patterns are generally only having difficulty in relating what they hear (which is the same as what everyone else hears) to this "pseudo spatial" representation.

It is very important to make the point that we are not interested in all aspects of a speaker's pitch; the only things that should interest us are those which carry some linguistic information. If a speaker tries to talk while riding fast on a horse, his or her pitch will make a lot of sudden rises and falls as a result of the irregular movement; this is something which is outside the speaker's control and therefore cannot be linguistically significant. Similarly, if we take two speakers at random we will almost certainly find that one speaker typically speaks with lower pitch than the other; the difference between the two speakers is not linguistically significant because their habitual pitch level is determined by their physical structure. But an individual speaker does have control over his or her own pitch, and may choose to speak with a higher than normal pitch; this is something which is potentially of linguistic significance.

A word of caution is needed in connection with the word pitch. Strictly speaking, this should be used to refer to an auditory sensation experienced by the hearer. The rate of vibration of the vocal folds - something which is physically measurable, and which is related to activity on the part of the speaker - is the fundamental frequency of voiced sounds, and should not be called "pitch". However, as long as this distinction is understood, it is generally agreed that the term "pitch" is a convenient one to use informally to refer both to the subjective sensation and to the objectively measurable fundamental frequency.

We have established that for pitch differences to be linguistically significant, it is a necessary condition that they should be under the speaker's control. There is another necessary condition and that is that a pitch difference must be perceptible; it is possible to detect differences in the frequency of the vibration of a speaker's voice by means of laboratory instruments, but these differences may not be great enough to be heard by a listener as differences in pitch. Finally, it should be remembered that in looking for linguistically significant aspects of speech we must always be looking for contrasts; one of the most important things about any unit of phonology or grammar is the set of items it contrasts with. We know how to establish which phonemes are in contrast with b in the context -in; we can substitute other phonemes (e.g. p, s) to change the identity of the word from 'bin' to 'pin' to 'sin'. Can we establish such units and contrasts in intonation?