The process which is designated by the term grammaticalisation involves a shift in status of words from full lexical items to grammatical endings or words. There is usually a sequence of steps which the words pass through during grammaticalisation (development of grammatical endings).
| Status | Steps in process |
| lexical word | (full meaning available) |
| semi-lexical word | loss of meaning through semantic bleaching |
| clitic | frequent attachment to another lexical word |
| inflection | permanent loss of independence and retention of grammatical meaning only |
A few examples from the history of English can illustrate this principle clearly. Old English lic meant ‘form, body’ and existed as an independent word but later was lost in this capacity and was retained only as an ending -lice which later resulted in the modern adjectival suffix -ly. A northern pronunciation of the word without palatalisation of the final segment gave like which is both an independent grammatical word and an element of a compound as in childlike.
Another instance is provided by Germanic /au/ which became /ɑæ/ in Old English but remained /au/ in Old Norse. Leas /lɑæs/ is therefore an Old English form while lauss is the shape in Old Norse. This was borrowed into English and resulted in Modern English loose, whereas the Old English form is still to be seen in the suffix -less which has been grammaticalised to a privative ending as in homeless, hapless (the first element of the latter form is also a bound lexical morpheme, i.e. there is no hap in Modern English).
The Old English word hwīl ‘time’ occurred in the dative plural hwīlom meaning ‘during’ (cf. German bei Zeiten). This was reduced to while by inflectional loss and the original noun was lost also so that the adverb of time is the only survival. Again in Old English there was a word dōm meaning ‘judgement; condition’ (Modern English doom) which as an ending lost its definite meaning but survived as a suffix indicating a quality noun. Hence forms like kingdom and wisdom are taken to be simply stem + grammatical suffix.
Any word class can be subject to grammaticalisation. An example from that of verbs would be the development of the future with go in English. This was originally locative in meaning (and still is in cases like She goes to work at eight) but the lexical meaning of movement forward in space came to be interpreted as a movement in time and sentences like She's going to buy some new clothes show that both a locative and a temporal interpretation are possible while instances like She's going to say something about her new clothes show a purely temporal meaning.
The development of manner adverbs in Romance provide a clear instance of grammaticalisation. In Latin mente (ablative) < mens was used in the sense ‘with a state of mind’ as in lenta mente ‘slowly’, dulce mente ‘softly’. Later one finds (in Italian, Spanish and French) the ending -mente/-ment as a general ending indicating a quality: absolutamente ‘absolutely’.
An instance of grammaticalisation from German would be the intensifier aller- as in ihre allererste Wahl ‘her very first choice'; sein allerbester Film. This derives from die erste aller ‘the first of all-GEN’ but the genitive meaning of aller has been lost and it simply intensifies the adjective it is prefixed to.
The following is a standard textbook on grammaticalisation.
.gif)
Hopper, Paul and Elizabeth Closs Traugott 2003. Grammaticalization. Second edition. Cambridge: University Press.
