ALLOMORPHS

The combining form allo- from Greek allos ‘other’ is used in linguistic terminology to denote elements of a group whose members together constitute a structural unit of the language (allophones, allomorphs). Thus, for example, -ion/-sion/-tion/-ation in §5.6. are the positional variants of the same suffix. To show this they are here taken together and separated by the sign /. They do not differ in meaning or function but show a slight difference in sound form depending on the final phoneme of the preceding stem. They are considered as variants of one and the same morpheme and called its allomorphs. Descriptive linguistics deals with the regularities in the distributional relations among the features and elements of speech, i.e. their occurrence relatively to each other within utterances. The approach to the problem is consequently based on the principles of distributional analysis.

An allomorph is defined as a positional variant of a morpheme occurring in a specific environment and so characterised by complementary distribution. Complementary distribution is said to take place when two linguistic variants cannot appear in the same environment. Thus, stems ending in consonants take as a rule -ation (liberation); stems ending in pt, however, take -tion (corruption) and the final t becomes fused with the suffix.

Different morphemes are characterised by contrastive distribution, i.e. if they occur in the same environment they signal different meanings. The suffixes -able and -ed, for instance, are different morphemes, not allomorphs, because adjectives in -able mean ‘capable of being’: measurable ‘capable of being measured’, whereas -ed as a suffix of adjectives has a resultant force: measured ‘marked by due proportion’, as the measured beauty of classical Greek art; hence also ‘rhythmical’ and ‘regular in movement’, as in the measured form of verse, the measured tread.

In some cases the difference is not very clear-cut: -ic and -ical, for example, are two different affixes, the first a simple one, the second a group affix; they are said to be characterised by contrastive distribution. But many adjectives have both the -ic and -ical form, often without a distinction in meaning. COD points out that the suffix -ical shows a vaguer connection with what is indicated by the stem: a comic paper but a comical story. However, the distinction between them is not very sharp.

Allomorphs will also occur among prefixes. Their form then depends on the initials of the stem with which they will assimilate. A prefix such as im- occurs before bilabials (impossible), its allomorph ir- before r (irregular), il- before l (illegal). It is in- before all other consonants and vowels (indirect, inability).

Two or more sound forms of a stem existing under conditions of complementary distribution may also be regarded as allomorphs, as, for instance, in long a : : length n, excite v : : excitation n.

In American descriptive linguistics allomorphs are treated on a purely semantic basis, so that not only [ız] in dishes, [z] in dreams and [s] in books, which are allomorphs in the sense given above, but also formally unrelated [n] in oxen, the vowel modification in tooth : : teeth and zero suffix in many sheep, are considered to be allomorphs of the same morpheme on the strength of the sameness of their grammatical meaning. This surely needs some serious re-thinking, as within that kind of approach morphemes cease to be linguistic units combining the two fundamental aspects of form and meaning and become pure abstractions. The very term morpheme (from the Greek morphē ‘form’) turns into a misnomer, because all connection with form is lost.

Allomorphs therefore are as we have shown, phonetically conditioned positional variants of the same derivational or functional morpheme (suffix, root or prefix) identical in meaning and function and differing in sound only insomuch, as their complementary distribution produces various phonetic assimilation effects.

§ 5.8 BOUNDARY CASES BETWEEN DERIVATION, INFLECTIONAND COMPOSITION

It will be helpful now to remember what has been said in the first chapter about the vocabulary being a constantly changing adaptive system, the subsets of which have blurred boundaries.

There are cases, indeed, where it is very difficult to draw a hard and fast line between roots and affixes on the one hand, and derivational affixes and inflectional formatives on the other. The distinction between these has caused much discussion and is no easy matter altogether.

There are a few roots in English which have developed great combining ability in the position of the second element of a word and a very general meaning similar to that of an affix. These are semi-affixes treated at length in Chapter 6.¹ They receive this name because semantically, functionally, structurally and statistically they behave more like affixes than like roots. Their meaning is as general. They determine the lexico-grammatical class the word belongs to. Cf. sailor : : seaman, where -or is a suffix, and functionally similar, -man is a semi-affix.

¹ On the subject of semi-affixes see p.p. 116-118. 102

Another specific group is formed by the adverb-forming suffix -ly, following adjective stems, and the noun-forming suffixes -ing, -ness, -er, and by -ed added to a combination of two stems: faint-hearted, long-legged. By their almost unlimited combining possibilities (high valency) and the almost complete fusion of lexical and lexico-grammatical meaning they resemble inflectional formatives. The derivation with these suffixes is so regular and the meaning and function of the derivatives so obvious that such derivatives are very often considered not worth an entry in the dictionary and therefore omitted as self-evident. Almost every adjective stem can produce an adverb with the help of -ly, and an abstract noun by taking up the suffix -ness. Every verbal stem can produce the name of the doer by adding -er, and the name of the process or its result by adding -ing. A suffix approaching those in productivity is -ish denoting a moderate degree of the quality named in the stem. Therefore these words are explained in dictionaries by referring the reader to the underlying stem. For example, in “The Concise Oxford Dictionary” we read: “womanliness — the quality of being womanly; womanised a or past participle in senses of the verb; womanishly — in a womanish manner; womanishness — the quality or state of being womanish”.

These affixes are remarkable for their high valency also in the formation of compound derivatives corresponding to free phrases. Examples are: every day : : everydayness.

Other borderline cases also present considerable difficulties for classification. It is indeed not easy to draw the line between derivatives and compound words or between derivatives and root words. Such morphemes expressing relationships in space and time as after-, in-,¹ off-, on-, out-, over-, under-, with- and the like which may occur as free forms have a combining power at least equal and sometimes even superior to that of the affixes. Their function and meaning as well as their position are exactly similar to those characteristic of prefixes. They modify the respective stems for time, place or manner exactly as prefixes do. They also are similar to prefixes in their statistical properties of frequency. And yet prefixes are bound forms by definition, whereas these forms are free. This accounts for the different treatment they receive in different dictionaries. Thus, Chambers’s Dictionary considers aftergrowth a derivation with the prefix after-, while similar formations like afternoon, afterglow or afterthought are classified as compound nouns. Webster’s Dictionary does not consider after- as a prefix at all. COD alongside with the preposition and the adverb on gives a prefix on- with the examples: oncoming, onflow, onlooker, whereas in Chambers’s Dictionary oncome is treated as a compound.

The other difficulty concerns borrowed morphemes that were never active as prefixes in English but are recognised as such on the analogy with other words also borrowed from the same source. A strong protest against this interpretation was expressed by N.N.Amosova. In her

¹ Not to be mixed with the bound form in-/im-/il-/ir- expressing negation.

opinion there is a very considerable confusion in English linguistic literature concerning the problem of the part played by foreign affixes in English word-building. This author lays particular stress on the distinction between morphemes that can be separated from the rest of the stem and those that cannot. Among the latter she mentions the following prefixes listed by H. Sweet: amphi-, ana-, apo-, cata-, exo-, en-, hypo-, meta-, sina- (Greek) and ab-, ad-, amb- (Latin). The list is rather a mixed one. Thus, amphi- is even productive in terminology and is with good reason considered by dictionaries a combining form. Ana- in such words as anachronism, anagram, anaphora is easily distinguished, because the words readily lend themselves for analysis into immediate constituents. The prefix ad- derived from Latin differs very much from these two, being in fact quite a cluster of allomorphs assimilated with the first sound of the stem: ad-/ac-/af-/ag-/al-/ap-/as-/at-/. E. g. adapt, accumulation, affirm, aggravation, etc.

On the synchronic level this differentiation suggested by N.N. Amosova is irrelevant and the principle of analysis into immediate constituents depends only on the existence of other similar cases as it was shown in § 5.3 for the suffixes.