Logical Gap

[Larry Trask]

Lloyd Anderson writes:

[snip]

> Here is an example of an exclusion which is an indirect consequence,
> not explicitly stated in a set of criteria, yet a real exclusion nonetheless.
> None of Trask's criteria refer explicitly or directly to verbs,
> yet given the structure of Basque, in which as I understand Trask
> no verbal word is monosyllabic (though verbal *roots* are),
> his criteria do end up excluding all verbs.

In modern Basque, a few non-finite forms of a few verbs are monosyllabic.
But it is clear that, in every case, the forms were originally
polysyllabic, and that their monosyllabic modern forms result from
regular changes.

So, historically, no ancient Basque verb-forms are ever monosyllabic, yes.
But why is this relevant?  My criteria say nothing about the number
of syllables.

More importantly, no ancient Basque verb-form is ever monomorphemic.
Since I am expressly looking only at monomorphemic lexical items,
then, yes, it follows that all verbs will be excluded.  Why should
this bother anybody?  Why should anyone be concerned that a list of
monomorphemic lexical items excludes a set of polymorphemic word-forms?

> In this case, we only need a single criterion, monosyllabicity of the word,
> to end up excluding all verbs.

No: monomorphemic structure is what excludes verbs.  Monosyllables are
not excluded, and quite a few monosyllables will make it into my list,
because they satisfy my criteria.

> Yet "monosyllabicity of the word"
> is a criterion which nowhere mentions anything directly related to
> "verb".  The exclusion is indirect, because of something else,
> a fact of verb structure (specific to, though not unique to, Basque).

Verbs are excluded, yes, because they never appear as monomorphemic
lexical items -- and it is monomorphemic lexical items I'm looking at.

> In the case of a criterion based on early attestation,
> there is nothing referring explicitly to excluding any particular
> strata of vocabulary or individual lexical items,
> yet an indirect consequence is that any strata of vocabulary
> which are selectively disfavored for written attestation,
> or any individual lexical items so disfavored,
> will be statistically excluded.

I'm not sure what Lloyd means here by 'strata'.  Normally, when we
talk about strata in a lexicon, we mean groups of words which have
entered the language at different times.  And, of course, the whole
point of my list is to get back to the earliest recoverable stratum
in the Basque lexicon, and hence to exclude all words which have
entered the language more recently.  So, excluding particular strata
is exactly what I'm up to.

But I suspect Lloyd has something else in mind when he mentions
'strata' -- perhaps something like 'words exhibiting particular
types of formation' or 'words in particular semantic domains'.
However, if such words fail to be native, ancient and monomorphemic --
as indeed they often do -- then they don't belong in my list.

> So it can happen, through using a combination of criteria,
> that several strata of vocabulary end up being excluded *statistically*
> (all I have ever claimed) by Trask's criteria.
> Since most of us know independently of Trask's criteria
> that most or all living languages contain vocabulary of such strata,
> it would normally be considered appropriate to include some of them
> as candidates for reconstruction of any pattern which is expected
> to have any sort of general validity for the language.
> ("general" does not mean "universal").

Not so, I'm afraid.

Again, the suggestion is that native and ancient words of *particular*
phonological structure have, for some reason, either failed to survive or
failed to be recorded.  And I still can't see any reason to worry about
this.

> And any set of criteria which had the effect, direct or
> indirect, of excluding such vocabulary strata systematically
> (statistically), would be judged as a poor set of criteria.

No.

If I'm interested in the most ancient stratum of native monomorphemic
lexical items -- as I am -- then criteria that have the effect of
excluding everything else strike me as a pretty good set of criteria.

Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK

larryt at cogs.susx.ac.uk