Re: Unilang: the Morphology

Andreas wrote:
>Marcus Smith wrote:
>>>I am, of course, operating on the theory that a 'word' is an speech unit
>>>that can be pronounced by itself and is "complete". According to this
>>>definition, the reduced forms of the English copula (-'m, -'re, -'s)
>>>aren't
>>>proper words - you won't say [z] in isolation if asked what the 3rd sg
>>>present of 'to be' is. If that Russian {s} is pronounced as part of the
>>>preceeding or following word, I won't consider it a 'word' on its own, but
>>>rather as an affix.
>>
>>So, English possessive -'s is an affix that attaches to any part of speech
>>what-so-ever, providing that the phrase containing the word is headed by a
>>noun, and that the noun is the possessor of another noun?
>
>Something like that, yes. It would perhaps be better to desribe it as
>attaching to the end of a phrase functioning as a noun in relation to the
>rest of the sentence. If we consider it a 'word' we have pretty much
>abolished the distinction between 'word' and 'morpheme'.
Two completely different responses came to my head immediately.

1) There is nothing necessarily wrong with the abolishment of the
word/morpheme distinction. A "morpheme" is a purely theoretical notion that
can be redefined anytime linguists (as a whole) feel the need to do so.
"Word" has a pre-theoretical meaning, but that meaning runs afoul of all
kinds of problems so that it is unworkable in a formalized context. The
relatively popular theory Distributed Morphology claims that words do not
exist -- only morphemes do, so abolishing the distinction between 'word'
and 'morpheme' is an idea that some linguists are taking seriously.

2) Possessive -'s and plural -s behave differently, and these differences
need to be captured somehow. Calling possessive -'s a phrasal affix will
cause more problems than it solves (not necessarily for English, but
certainly cross-linguistically). The best thing to do is call it a clitic.
As a clitic, it is a word (in a syntactic sense); but it lacks the prosodic
requirements of a word (in a phonological sense), so the clitic will lean
on (attach to) whatever word happens to precede it.

>>A serious problem in linguistics that few people have addressed is that
>>there is no decent definition of what a "word" is. Phonological definitions
>>run afoul of the syntactic data, and vice versa.
>
>I didn't claim my defintion was perfect, but it seems workable to me. Any
>better suggestion?
I don't know if this is better, but I would suggest not ignoring the
syntactic side of words. Syntax overtly manipulates words but not affixes.
(I have a professor or two who would probably skewer me for saying that.)
If something can appear in different positions relative to other relevant
words, it should probably be considered a word (or clitic) even if it can't
stand on it own.


Marcus Smith

"Sit down before fact as a little child,
be prepared to give up every preconceived notion,
follow humbly wherever and to whatsoever abysses Nature leads,
or you shall learn nothing."
                -- Thomas Huxley