Stress-interchange

|l7Definition of conversion, its synonymous terms]

The term conversion (originally 'the process of bringing over from one belief, view, or party to another') [ME fr. MFr fr. L conversion-, conversio, fr. converses, pp. of convertere fr. con 'with, together, thorougly' + vertere 'to turn around'] was first mentioned in 1900 by H. Sweet /Sweet: 1900/. It refers to numerous cases of phonetic identity of two words (primarily in their initial forms) belonging to different parts of speech (round adj, n, v, adv; back n, adj, adv, v; idle adj, v; water n,v; eye n,v; up prep, v.).

Some of the new names derived by conversion are used regularly and become lexicalized, enter the lexicon. Some of the uses, however, remain nonce words, or occasional words.

The term conversion when applied to word-formation is not quite adequate because source words are neither converted nor transformed in contrast to, for example, water when it is converted into ice. To avoid polysemy of the term conversion some other terms were coined to denote this specific process of naming.

Thus, the term 'affixless word-derivation' is used to underline the formation of a new word without a derivational affix. But this term does not permit us to distinguish it from sound- or stress-interchange (shift) that derived words without adding affixes, either.

The term 'zero-derivation' stresses that a new word is derived by means of a special affix called the zero affix because its absence in a word is meaningful. But the existence of such an affix is still debatable

The term 'root-formation' is used to point out that root words participate in the process. But other complexes may participate in the process, too, as in to machine-gun, to fire-gun, to wireless and the term turns out to be inadequate.

The term 'functional change' stresses that it is a phenomenon of usage, not word-formation. But this view can hardly be accepted because in fact a new word is derived with its own paradigm and system of meanings.

Some linguists regard conversion as a kind of polysemy because it is regularly patterned and derived units are semantically related like the senses of a polysemantic word. But in contrast to polysemy, the new naming units created by conversion belong to different parts of speech - they are different words and not just new senses. Conversion, therefore, is rather a kind of homonymy, though a very specific kind - a patterned lexical-grammatical homonymy where the old and new lexemes are semantically related.

So, conversion may be regarded as a lexical-semantic or morphological or even a syntactic means of word derivation by means of a functional change. In any event, conversion is one of the most productive ways of extending the English vocabulary. Here, following the view of professor A.I. Smirnitsky, conversion will be treated as a morphological way of word-formation where the only word building means is a change of a word's paradigm (cf.: the morphological paradigms of the word eye as a noun: eye -eyes and that of the word eye as a verb: to eye, eyes, eyed, will eye).

;2. Reasons for high productivity of conversion in modern English]

While affixation has always been a productive means of word-formation in English, conversion became active only in the Middle English period and it is widely used in modern English.

There was no homonymy between initial forms of words belonging to different parts of speech in Old English having a complex system of inflections. Due to loss of inflections in Middle English many of these words became lexical-grammatical homonyms (cf.: love n — love v in present-day English and their inflected equivalents lufu n and lufian v in Old English; see also different inflected forms for modern English work: OE were, weorc as a noun and wyrcan as a verb, or answer n and v that had in OE a form andsawru as a noun and andswarian as a verb).

Another reason for the existence of conversion pairs in modern English is assimilation of borrowings. The modern English verb and noun cry, for example, had different forms in Old French from which they were borrowed: crier v and cri n.