Re: CHAT: EU allumettes (was: Re: THEORY/CHAT: Talmy, Jackendoff and Matchboxes

Stephen Mulraney said:
> Mark P. Line wrote:
>
>>Tommie L Powell said:
>>>Actually, there are some natlangs that are very nearly
>>>impossible to make ambiguous statements in (without
>>>violating mandatory rules of grammar).  Czech is one
>>>-- so the Czechs have become adept at sneaking
>>>ambiguity into sentences by slurring key words (so that
>>>listeners can imagine that something else is being said)!
>>>
>>I assume that for a language to be unambiguous means that every possible
>>utterance either has a single reading or is "ungrammatical" (which term
>>would have to be given a referent by your favorite brand of theory). In
>>that case, the language would have to be devoid of polysemous lexemes,
>>even if the grammar were such that *syntactical* ambiguities do not
>> occur.
>>
>>I find it extremely hard to believe that Czech or any other natlang is
>>completely devoid of polysemous lexemes.
>>
>>
>
> Surely the truth of the issue lies somewhere between these two extremes.
One extreme must be Tommie's suggestion that a natlang can make it very
nearly impossible to make ambiguous statements.

What is the other extreme you're talking about?


> I don't think Tommie was intending to suggest that there exists some
> theoretical construct by means of which Czech could be judged genuinely
> unambiguous. He does point out an illustration (the slurring of words)
> which suggests, if true, that whatever is going on regularly affects
> speakers' ability to speak ambiguously, at least in some commonly met
> circumstances. Certainly very interesting, but hardly so unbelievable, no?

Tommie said that Czech made it very nearly impossible for speakers to make
ambiguous statements without violating grammatical rules.

I continue to find that unbelievable, yes.

And if somebody says something like "without violating mandatory rules of
grammar", then they're talking about a theoretical construct whether they
intend to or not; in particular, one that says that languages have
grammar, that grammars hvae rules, that rules can be mandatory and that
such mandatory rules can be violated by potential utterances.

Without some kind of proviso of grammaticality (which concept presupposes
a theoretical construct), the idea of a language being "unambiguous" is
meaningless: surely not every conceivable morpheme sequence would be
"unambiguous" in a language that we would otherwise want to judge
"unambiguous".


-- Mark