[Lexicog] part of speech for phrases

[Ron Moe]

You have two related problems, both caused by typological features also
found in other languages. The first problem is the choice of a citation
form. Prefixing languages like Igbo are especially problematic. In Bantu
languages virtually every word has a prefix. The result is that most of the
words are alphabetized under a few prefixes. You might as well just sort the
dictionary first by part of speech and then by alphabet. But of course that
isn't done, primarily because many users wouldn't know the part of speech,
especially of some of the smaller word classes. But in many languages with
heavy affixation, the user has to parse the word before he can figure out
what the citation form would be anyway. If he can parse the word, he
probably knows what the part of speech is.

Your second problem is how to indicate suprasegmental features. You can
sometimes do this on the citation form. However as you note, the citation
form may include affixes that obscure the tone class. Or the tone of the
word when spoken in isolation may undergo changes that obscure the
underlying tone. In many Bantu languages there are only two tone classes for
verbs. In Lugungu the tone is carried on the first mora of the stem. However
due to complex rules that are affected by affixation, syllable structure,
syntax, and phonological context, the actually surface tones display a
bewildering array of forms. All this complexity can be reduced merely by
indicating which tone class the verb belongs to. This can be done on the
citation form, in a separate pronunciation field, or in a separate tone
class field (the tone marks would actually go above the following vowel):

kw'ala v. To make a bed.
kw`ala v. To swim.

Or:

kwala [kw'ala] v. To make a bed.
kwala [kw`ala] v. To swim.

Or:

kwala H v. To make a bed.
kwala L v. To swim.

In isolation the two words are kw'al^a and kw`al'a. But this information
only tells you how the infinitive is spoken in isolation. It tells you
nothing about the overall tone pattern unless you know something of the tone
rules and can figure out the underlying tone from the surface tone. (The
underlying tone does not always surface on the first syllable as it does in
these examples.) If the orthography does not mark tone, the users may not
want diacritics on the citation form. So option two or three may be
preferred.

So choosing a prefixed form as the citation form will affect how words sort
and will affect the surface tones in a tonal language. There are other
problems with prefixed forms that I will not go into here.

Ron Moe
SIL, Uganda


-----Original Message-----
From: Chinedu Uchechukwu [mailto:neduchi at hotmail.com]
Sent: Monday, January 19, 2004 7:03 AM
To: lexicographylist at yahoogroups.com
Subject: Re: [Lexicog] part of speech for phrases


Ron,John and Rhodes,
I found your discussion has been very interesting, as it touches on
some of the issues that I am presently preoccupied with.
Ron initially pointed out the dictionary features that might be
connected with the typology of a particular language, and connected
it to the question: root or stem in dictionary organization.
POSSIBLE SOLUTION:
In the dictionaries of my native language Igbo (a tone language,
spoken in South Eastern Nigeria)two solutions have been profered:

ROOT ORGANIZATION
The first was root organization, and for the following reason:
(1) The language has only two vowels to distinguish between the
infinitives. This would mean organizing the whole verbs of the
language under two initial alphabets ("i" and "!")
(2) The infinitive forms can constitute a problem to the learner, as
they ALTER the tone of the root, which is needed to classify the
verb. For example -rí (eat), is a high tone verb, but with the
infinitve marker 'i', it becomes írí (to eat). The infinitive marker
has a high tone and the root tone is lowered to a downstep, although
the root has maintains a high tone in all its inflectional forms.

So to maintain the inherent tone of the verb a root organization was
prefered.


IMPERATIVE FORM ORGANIZATION
THis method was adopted by Echeruo in his Igbo-English dictionary. It
has been found quite confusing, and not as much helpful as the root
organization.

A possible solution that I see can more readily be realized through
an electronic dictionarry than a print dictionary. This solution is
first of all directed at the learner of the language. This means
making the verbal inflection and tonal variations part of the sub-
entries to a head word.

It is of course an entirely different issues to realize such a
dictionary in a book form.

Chinedu Uchechukwu
Otto-Friedrich University, Bamberg
Germany.


--- In lexicographylist at yahoogroups.com, "John Roberts"
<dr_john_roberts at s...> wrote:
> Ron Moe said:
>
> > What I've been hoping for is a set of criteria for
subcategorizing various
> > types of idiomatic phrases. They can be distinguished by their
internal
> > structure. They can be distinguished by their degree of setness
(can the
> > constituents be reordered, can other words be inserted, can any
of the
> words
> > be replaced, can any of the words be inflected, etc.). They can
also be
> > distinguished according to their semantics, but this is like
assigning the
> > part of speech 'verbal noun' to 'death', or 'verbal adjective' to
> 'morbid'.
> > Are there any other ways they can be subcategorized? I'm currently
> assigning
> > all idiomatic phrases the part of speech 'phr.', since I have no
well
> > motivated way of further categorizing them.
>
> But you do want to categorize idioms according to their syntactic
function
> as well as their internal structure. They are not just phrases, but
phrases
> that function as 'phrasal verbs', 'phrasal adjectives', 'phrasal
nouns',
> etc.
>
> So I would say that _pass up_ and _kick the bucket_ are both
phrasal verbs
> on the basis of their internal structure. Both _pass_ and _kick_
can be
> inflected for tense, for example, and both have objects. _going
over_ is a
> phrasal noun. You can pluralise the nominal in this phrase, as
in 'they gave
> him several goings over' (this identifies the lexical class) and
you can
> modify the expression with an adjective, as in 'they gave him a
good going
> over' (this identifies the syntactic function).
>
> John





Yahoo! Groups Links

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

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

Your use of Yahoo! Groups is subject to:
 http://docs.yahoo.com/info/terms/





Yahoo! Groups Links

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

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

Your use of Yahoo! Groups is subject to:
 http://docs.yahoo.com/info/terms/