Re: Dictionaries of agglutinating languages

On Tue, 3 Oct 2000, Robert Hailman wrote:

> dirk elzinga wrote:
> >
> > Instead of confining affixes to a separate section of the dictionary
> > (a very fine idea, IMO), you could also include derived lexical items
> > under the root from which they are derived in a tree-like structure.
> > So from the hypothetical root PAD there may be the derived lexical
> > items padit, pada, padismo, padismoti, padamaa, padakt, etc. They
> > would be arranged in this way
> >
> >         PAD ...
> >           padit ...
> >             padismo ...
> >               padismoti ...
> >           pada ...
> >             padamaa ...
> >             padakt ...
> >           &c.
> >
> > each with its own definition. This accomplishes a couple of things.
> > First, you acknowledge the relationships among derivationally related
> > words by including them in the same "macro" entry. Second, you provide
> > a partial analysis of the derivation by virtue of the visual
> > appearance of each macro entry; ie, the indentation shows you the
> > "derivational history" of a particular complex form. Third, each
> > lexical item gets a separate definition which can be especially useful
> > for irregular or exceptional meanings. The burden is then on the user
> > to know what the root of any given form is, though that would be true
> > for any dictionary which is organized according to roots.
>
> That's also a very good idea, but I could see a flaw if there were a
> good number of affixes, because then I could see repeating this pattern,
> and a similar one for every part of speech, for every entry could make
> the dictionary quite hefty.
This is true only if every derivational affix could be attached to
every root, something which is true of no language I've heard of.
Yes, there would be repetition of patterns, but these should be
tolerated (if not encouraged) for the sake of consistency. There would
be an unexpected benefit as well; suppose an enterprising linguist
wanted to study the distribution of a particular derivational affix or
set of affixes. It would be a rather trivial (if tedious) task to
search through the dictionary to see how many roots have been modified
with the affixes in question. However, if the affixes are confined to
an appendix, it's anyone's guess.

> It would make the dictionary easier to use:
> pick out the root, and the meaning of the derived term is right there,
> mit definition.
Yes. But it does require the user to be aware of the derivational
history of a word--something which won't always be true. Last year on
an exam I asked students to parse complex words into their component
morphs. One of the words was 'actively'. I was very surprised to see
how many students did not see the root 'act' in the word. They did
find 'active', but 'act' was, for all intents and purposes, invisible
to them. Just because a linguist is able to demonstrate a particular
morphological structure, it doesn't mean that linguistically innocent
speakers will see the same structure. So organizing a dictionary
according to roots will have problems.

> With the idea I proposed, it takes a little more work:
> find the root in the dictionary, get it's definition, go to a table in
> the back, find every affix used, and figure out how the combine to give
> the derived work it's meaning. That would be quite a laborious task.
But not always accurate. A favorite example I give to my students when
the subject of derivation comes up is the suffix -hood in English. It
attaches to nouns to create abstract nouns meaning something like 'the
property of being an X'. Thus, father -> fatherhood; knight ->
knighthood, etc. However, it won't work with all nouns: candle ->
*candlehood (although one could imagine what that might mean). And
there are nouns which don't conform to the "regular" pattern: neighbor
-> neighborhood (does *not* mean 'property of being a neighbor'). So
confining all of the derivational affixes to an appendix and relying
on the user to piece things together will fail to capture these kinds
of irregularities--especially if the user is an L2 learner.

> Basically, the same amount of work has to be done, but using my idea
> it's up to the reader to do most of the work, and with your idea, Dirk,
> it's up to the person writing the dictionary to do most of the work. I
> could see either working.
Suppose that the language community numbered 1000. Now, one person
could spend a great deal of time and energy creating a dictionary
which "spoon-fed" derivation in a manner similar to what I have
suggested, enabling the 1000 users to find words quickly and
efficiently. Or we could have the lexicographer spend not quite as
much time and energy (though still a lot) and simply provide the users
with an appendix of derivational affixes. Then 1000 people would have
to spend a lot of time and energy *every time* they wanted to look up
a word. I don't think that it would pay off in the long run to
organize the dictionary that way.

I do think that having the affixes in an appendix is a great idea. But
having derivational information *only* in such an appendix would
create many problems that could be avoided.

Gee, I really sound invested in this idea, don't I? I'm really not.
But it is fun to think about.

Dirk

--
Dirk Elzinga
dirk.elzinga@m.cc.utah.edu