These groupings like English [t], [th] and [t], with respect to their simultaneous unity and diversity, have traditionally been dealt with in terms of two levels of representation. That is to say that at a concrete physical level the members of these groups of sounds are different phonetically – they have different phonetic properties – but that abstractly it is useful to group them together as being related. In fact, grouping them together this way reﬂects the intuition of the native speaker that these sounds are ‘the same’
in some sense. Taking this view we can say that abstractly English has a ‘t’ and that concretely the pronunciation of this ‘t’ depends on the context in which it occurs. That is, if the ‘t’ of English appears at the beginning of a word it is pronounced as [th], if it appears as part of a consonant cluster following [s] it is pronounced as [t], if it appears at the end of a word it may be pronounced as [t] (or indeed as [ʔ] or [t]). In the same way, we can say that English ‘p’ has several concrete representatives: [p], [ph] and [p].