[Lexicog] Dictionary of a language with classifiers

[Ron Moe]

Kim,

It appears that I misunderstood your use of space in iyá 'a' and iyá kap, or
else we are having font problems. (Sometimes I hate computers. Too bad we
can't clarify what the other is saying online.) On my computer the two
appear similarly to iya 'a' and iya kap, each with a space in them and an
acute accent over the first a. In this light your statement, "Our
classifiers are technically "suffixes" rather than compounds, so they're not
written with a word break" is rather confusing. I took the space to indicate
that these were phrases. If the classifiers are suffixes (written without a
space), then my suggestions are "wrong". I'm not saying that particular
solutions are right or wrong. It is more that historical conventions lead
the average user to expect certain formats to indicate certain types of
data. Most users would expect "n." to mean "this word is a noun". Subentries
are normally used for complex forms, not collocations, lexical relations, or
examples.

There are different types of classifiers. In Japanese a classifier is used
with numerals (e.g. kami yon-mai 'paper four-CL' (four sheets of paper)
where -mai is the classifier for flat objects). In this case kami is
associated with -mai and you might want to indicate this in your dictionary.
But you would not enter 'kami mai' as a subentry. Similarly we would not
enter '(two) head of cattle' '(two) stalks of celery' '(two) pieces of
paper' as subentries in an English dictionary. Instead we would enter them
as lexical relations.

However in a Bantu language the noun class prefixes are obligatory. You can
say 'muntu' (mu-ntu 'C1.SG-person') and 'bantu' (ba-ntu 'C2.PL-person'), but
you can't say 'ntu' by itself. So we give muntu as the citation form for
this word. We would not give ntu as the citation form and then give muntu
and bantu as subentries. So even though mu- indicates that ntu is in the
human class, this is not the same sort of classifier as the Japanese '-mai'
or the English 'head'.

You mention that your classifiers overlap with roots. In that way they are
more like the English 'head' than the Japanese '-mai'.

If your classifiers are non-obligatory suffixes, then you are probably right
in making them subentries, especially if they can result in a change of
meaning. In that way they are more like derivational affixes. But we are
talking about a number of very different phenomena.  I would need to know
more how your classifiers are operating morphologically and semantically to
be able to say more.

Hopefully we can bring a little more clarity to this discussion.

Ron Moe
  -----Original Message-----
  From: lexicographylist at yahoogroups.com
[mailto:lexicographylist at yahoogroups.com]On Behalf Of Kim Blewett
  Sent: Wednesday, June 15, 2005 10:05 AM
  To: lexicographylist at yahoogroups.com
  Subject: RE: [Lexicog] Dictionary of a language with classifiers


  In response to Ron's comments about noun+classifier combinations,

    The problem with Kim's solution (below) is that using \se makes it
appear
    that iyá 'a' and iyá kap are phrasal lexemes. They are not, if I
understand
    you correctly. Also using \se causes each sense to begin a new line:

    iyá n. stone.
      iyá 'a' n. stone.
      iyá kap n. gravel.

    This is not at all what you want. At least it does not accurately
    communicate the nature of iyá 'a' and iyá kap.

  Actually, this format, beginning each compound form on a new line and
bolding it, is exactly what we want for Rapoisi. I guess it depends on who
your target audience is; we plan to produce a Rapoisi-English-Tokpisin
dictionary for local distribution. However, I don't see that either of these
solutions necessarily implies that these are "phrasal lexemes." They are
forms that need to be easy to locate in a bilingual dictionary, (1) because
the meaning can not always be deduced completely from the parts, especially
for a new learner of the language, and (2) the correct classifier can not
always be predicted for a form. Consistency in dictionary entries is
important, and perhaps an introduction explaining what certain fields
convey/contain, but I was not aware of rules for how a dictionary is allowed
to treat certain forms. A dictionary format should reflect the structure of
the language it's displaying.

  Our classifiers are technically "suffixes" rather than compounds, so
they're not written with a word break, but the set of "roots" overlaps
somewhat with the set of "classifiers", which resembles compounding. I have
seen languages where the orthography breaks up phonological words, either
compounds or clitic markers, for ease of reading. Actually, I was not sure,
from Nilson's example, how independent the classifiers in Karo are.

    Currently you are making each
    sense of iyá into a separate entry:

    iyá 'a' n. stone.
    iyá kap n. gravel.

    This format also makes it look like these are phrasal lexemes. There are
two
    possible formats that would convey the information correctly:

    iyá n. 1. iyá 'a' stone. 2. iyá kap gravel.

    Or:

    iyá n. 1. stone. iyá 'a'. 2. gravel. iyá kap.

    This second format is preferred because, by tradition, examples of usage
and
    collocation are placed after the definition.

  This looks good for a dictionary targeted for linguists, but my feeling is
that Rapoisi speakers, or other PNGans who would like to learn Rapoisi,
would be confused by the switching of position here.

    However some dictionary
    traditions place parts of the paradigm toward the beginning of the
entry:

    have v. had, having, has Used as an auxiliary verb before...

    We do the same sort of thing in Bantu dictionaries to give the plural of
    nouns:

    muntu n. bantu A person, someone...

  Your examples here demonstrate grammatical forms. There is a big
difference between these and noun+classifier forms, which in Rapoisi
function as a unit with respect to plural, possessive, and even can be
verbalized.

    Unfortunately MDF does not have a vernacular field that occurs between
the
    initial fields (\lx \lc \se \ps \sn) and the definition fields (\dv \de
    \dn). Some of us have complained about this. There are some solutions,
but
    they all involve modifying the MDF Consistent Changes tables or
modifying
    the Word document once MDF has processed the file. Neither solution is
easy.
    So I would recommend that you put the classifier in a lexical function
field
    and label it as a classifier:

    \lx iyá
    \ps n
    \sn 1
    \de stone
    \lf Classifier
    \lv iyá 'a'
    \sn 2
    \de gravel
    \lf Classifier
    \lv iyá kap

    You will get the following output (or something very similar):

    iyá n. 1. stone. Classifier: iyá 'a' 2. gravel. Classifier: iyá kap.

  The main problem I see with using the lexical field is with reversals. I
want the defs of these forms to show up in a reversed finder list, for the
benefit of language learners in the Rapoisi community. Under \se I can add a
\re field. With \lf, I can't sort out noun+classifier forms from all the
other things that \lf contains that do not need a reversal entry. (It's MDF
that wants to see a \re field under \se; if not using MDF you could probably
make the first word of the \de field appear, or something similar.

  So, I still feel that for some languages the Subentry field is the way to
go, although Ron's comments are good and have caused me to think things
through more carefully--thanks, Ron!

  Kim Blewett



----------------------------------------------------------------------------
--
  Yahoo! Groups Links

    a.. To visit your group on the web, go to:
    http://groups.yahoo.com/group/lexicographylist/

    b.. To unsubscribe from this group, send an email to:
    lexicographylist-unsubscribe at yahoogroups.com

    c.. Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lexicography/attachments/20050615/49e90afd/attachment.htm>