Re: Unilang: the Morphology

On Mon, 23 Apr 2001, Andreas Johansson wrote:

> Then we should, I think, either use 'word' only in the "phonological" sense,
> or alternatively use other names for both senses you describe and leave
> 'word' as a non-technical word that can be used according to whatever
> traditions exist for a particular lang.
It is not uncommon in the linguistic literature to refer to "phonological
words" and "syntactic words" to differentiate the two most common uses.
There should probably be "semantic words" as well, though I've never seen
that used.

> Hm. 'Clitics' and 'words' could be sorted as "syntactic units" or something
> like that, without being equated themselves.
Indeed. The term 'clitic' is just one of many "trash cans" that linguists
through words/morphemes into whenever they don't fit into the phonological
or syntactic word categories. For example, the second position clitics of
Polish are not very similar to the pronominal clitics of Pima.

So the conclusion this leads us to so far is using:
'word' for phonological words
'clitic' for syntactic words that depend on phonological words

We just need something for syntactic words that do not fall into other
categories. (And something for semantic words, if that is something you
are worried about, but probably not.)

Marcus