Re: Clitics
From: | Matt Pearson <jmpearson@...> |
Date: | Friday, October 29, 1999, 21:17 |
Ed Heil:
>I don't know about clitics in the languages of the world; I only know
>the definition of "clitic" in the Classical languages, and that is the
>one I gave -- a short word which becomes part of another word for
>accent purposes.
>I suppose one could look at other suprasegmental features rather than
>accent to determine what is or isn't a clitic, but I'm not sure
>exactly what. Perhaps some linguists have already figured out a
>reliable definition of "clitic" in English, and it was on such bases
>that you were criticized before. I'm afraid I'm not aware of them
>though.
I'm aware of at least four separate (but heavily overlapping) definitions
of "clitic". These are:
(1) Phonological (metrical) definition: A clitic is a morpheme which,
although not an affix, is short enough that it fails to receive full
word-level stress, and is thus treated as part of an adjacent word
for stress purposes. (This is the definition that Ed gives above...)
(2) Morphological definition: A clitic is a morpheme which tends to
form a tight morphological bond with another word, but is more
'detachable' than an affix.
(3) Morpho-syntactic (phrase structure) definition: A clitic is an
affix-like morpheme which attaches to a phrase-level constituent
rather than a word-level constituent.
(4) Syntactic (word order) definition: A clitic is a morpheme which
occupies a fixed linear position in the sentence (e.g. after the first
constituent).
The English article "the" appears to fit definition (1), but not the
others. French unstressed pronouns appear to fit definitions (1) and (2),
and perhaps (3) (depending on your analysis). The English possessive
marker "'s" (as in "The king of England's hat") fits definition (1), and
probably (3). Unstressed auxiliaries and second-position particles in
Serbo-Croatian fit definition (4), and perhaps also definitions (1), (2),
and (3). Tense markers and pronouns in Sally Caves's Teonaht could
be classified as clitics according to definition (2), and perhaps (1) as
well...
Matt.