MORPHEME

MORPHEME. In LINGUISTICS, a minimal unit of form and meaning. There are many variations in how the term is used and understood, arising in the main from a distinction between language as arrangement and language as process: (1) As proposed by the US structural linguist Leonard BLOOMFIELD (Language, 1933), the morpheme is the unit of MORPHOLOGY and therefore grammatical. In this approach, language analysed as a static arrangement of data consists of minimal units of form and meaning, each of which can be physically identified. The sentence The cats were sitting unhappily in the rain is analysable into the morphemic string the + cat + s + were + sit(t) + ing + un + happy + ly + in + the + rain. The 8-word sentence consists of 12 morphemes, all of equal status. (2) As proposed by the French linguist Joseph Vendryes (Le Langage, 1921), the morpheme is one of two units, one grammatical, one semantic, and each in its own sense minimal. Language in this approach is the outcome of processes which may or may not all have observable forms, but which can be analysed as units of grammatical meaning (morphemes) and units of lexical meaning (for Vendryes semantemes, but now known, more or less, as LEXEMES). Here, morphemes are the glue that holds lexemes together, and the specimen sentence can be analysed as: the + CAT + s + (BE + past/plural) + SIT(T) + ing + un + HAPPY + ly + in + the + RAIN (in which the lower-case items are morphemes, the upper-case lexemes). The 8-word sentence in this analysis contains 8 morphemes and 5 lexemes.

There have been many variations on these themes. The US linguist Dwight Bolinger (Aspects of Language, 1968) divided Bloomfield's morpheme into a system morpheme (the glue) and a source morpheme (the lexical content), while the French linguist André Martinet (Éléments de linguistique générale, 1970) subsumed Vendryes's morpheme and lexeme under a unifying unit the moneme. In this approach, the specimen sentence has 13 monemes divided into 8 morphemes and 5 lexemes. Currently, whatever the terms used, linguists tend to agree on three points: (1) Grammatical and lexical units need to be distinguished: for example, the two elements cat and s in the word cats are different aspects or levels of language. (2) Not all the features in a stretch of language are physically realized: for example, cats may exhibit a marker of plurality, but sheep does not. (3) One unit of form may serve more than one end: for example, were in the specimen sentence above combines BE and past. The traditional structuralist approach assumes that all the morphemes of a language can in principle be listed in a morpheme inventory, like a PHONEME inventory, but because of the complexities involved, few such lists have been attempted and none exists for English.

In later STRUCTURAL LINGUISTICS, the morpheme has been defined as the abstraction behind a morph (a form that has semantic distinctiveness). It may subsume two or more allomorphs, morphs that have common semantic identity but differ in their pronunciation according to well-defined rules: for example, the prefixes in-, im-, il- are allomorphs of the same morpheme (in this case a negative prefix) in the words insincere, impolite, illogical, the choice of prefix being determined by the initial sound of the stem that follows the prefix. When it deals with morphs and morphemes, morphology is known as morphemics. See -EME, LEVEL OF LANGUAGE.