Using Dictionaries (was Re: Greek question (night?))

[Robert Whiting]

On Mon, 15 Mar 1999, Yoel  L. Arbeitman wrote:

> Point well taken. But, dictionaries of "Classical Languages", Lat. and
> Gk., will always give both the nom. citation form AND the stem, usually
> in the form of the genitive sg. Case under discussion: Lat. nox, noctis
> will always be the listing from which one knows that the nom. nox /noks/
> represents a reduced stem /nok-/ + the desisence -s. And OIndic (Skt.)
> dictionaries give for noun and verb the stem rather than either resp.
> the nom. sg. or the third sg. active present.

Yes, dictionaries of different languages and different kinds of
dictionaries will give different information both with respect to types
and tokens.  But the primary point was that regardless of how much
information the dictionaries provide with respect to lexicon, they
usually provide either limited or no systematic information on phonology
or morphology.  And the real point behind it all, which I was making
rather obliquely, is that for comparative work you want to have the
best philological information available and that dictionaries only
provide a part of this.

> For the verb Classical Dictionaries give first sg. act. present. In
> Semitic the verb is given in the past/perfect(ive) 3 sg. masc., etc.

True for West Semitic, but dictionaries of East Semitic give the so-called
"infinitive" (nomen actionis) as the lemma for verbs.  The Akkadians
themselves generally used this form in their lexical lists so this is
a lexicographical tradition with a very long history.  But using this
form hides information about the stem vowel of the verb from the user
and this information has to be given separately in the header or excavated
from the article by the user.

> Thus there is variation in what is given as the basic datum. Probably
> what is here said about Finnish I am sure is true irrespective of
> whether or not I want to compare it to the hypothetical IE cited. And
> this is probably true for most modern language dictionaries. The
> Classical, OIndic, Hebrew, Arabic, etc., cases I am most familiar with,
> have a very long grammatical tradition.

All true, but I expect that all dictionaries anticipate that the user is
familiar with the mechanics of the language involved and therefore focus
on lexical usage rather than morphophonology.

But in raising the point that "using a dictionary without being aware of
the phonotactical rules of the language is a recipe for disaster when
doing comparative work" on this list I was hoping that I was preaching to
the converted, and that such a comment would be more appropriate on
something like sci.language.  Surely no one with linguistic training would
base language comparison solely on dictionary entries without a knowledge
of the morphophonology of the language.  But we have all seen the results
of doing just that.  A good example can be found at
http://members.aol.com/IrishWord/akkadian.htm which presents a comparison
between Akkadian and Irish words and shows no knowledge whatsoever of
the Akkadian language.

Bob Whiting
whiting at cc.helsinki.fi