Morphophonemics (henceforth MP) is an area of considerable complexity. The basic subject matter is easy to grasp; the ways of treating it in linguistic and psycholinguistic theory are what generate the complexity. All MP is a problem for (structural) linguistics, for reasons that will be spelt out below. As an instance of an MP problem, consider the fact that in English, the word nature "changes" its pronunciation when the suffix -al is added; the vowel of the first syllable of natural is short, low and not a diphthong, unlike the corresponding one in nature. The scare-quotes around the word changes hint at one of the problematic issues that will be dealt with below. It must be emphasized at the outset that there is no single agreed approach to MP among linguists, and much remains very controversial. It is therefore impossible to avoid a substantial amount of technicality in this article. Languages show a great deal of typological difference with respect to MP. Some languages have an immense amount of such phenomena, for instance classical Sanskrit and Latin, the Celtic languages, and some Amerindian languages such as Tubatulabal and Yawelmani Yokuts. At the other extreme, there is remarkably little in the indigenous languages of Australia. Others show next to none; a case in point is Vietnamese, it is often claimed. However, even these preliminary comments are problematic, for the very definition of MP, and therefore of what is an instance of it, is linked to the theoretical assumptions one makes. This account roughly follows the historical order of developments within the subject since the 1930's.
This paper is not available online