[Lexicog] Re: Citation forms in Prefixing Languages

Linguistics Philippines Linguistics_Philippines at SIL.ORG
Wed Jan 21 06:52:40 UTC 2004


> Given that Young and Morgan have already done all the work of listing what
> the inflected prefix strings are, it seems like it wouldn't be that hard
> to make a separate index of all those strings and what they decompose
> to....
(from Jordan Lachler's note below)

My experience is with Philippine languages that involve a lot of prefixation
and hidden consonants similar to what Ron Moe described.

I like what Jordan is suggesting here.  Make a list of all word forms likely
to be encountered, and then for each word form show what lookup form
(perhaps the root; perhaps something else) to use in order to find it in the
dictionary.  Use that as an index to the dictionary.  It should then take no
more than two steps (index lookup, then main entry lookup) to find any word
in the dictionary, without any guesswork.  I haven't actually seen this
done, but it seems straightforward.  The parser Mike Maxwell mentioned could
be used by the lexicographer to identify the roots and create the index.

Allan Johnson, Philippines


----- Original Message -----
From: "Jordan Lachler" <lachler at unm.edu>
To: <lexicographylist at yahoogroups.com>
Sent: Tuesday, January 20, 2004 7:04 PM
Subject: Re: [Lexicog] Re: Citation forms in Prefixing Languages


Howdy folks,

>Lest there be any doubt, I agree with Ron (and others): there is no good
>solution for printed dictionaries (although for prefixing languages,
>alphabetizing from the end of the word would not be impossible).

I guess I'd want to be a little more optomistic than that.  Having worked
on dictionary projects for both Iroquoian and Athabaskan languages, I've
been struggling with this very issue of prefixing for a while, and I've
enjoyed and appreciated the discussion it has generated here.

In the Iroquoian verb template, you basically have two types of prefixes:
pronominal prefixes (around 60 of them), occurring immediately before the
verb stem, and prepronominal prefixes (modal, directional, negative, and
others) occuring before the pronominal prefix.  If you look at the most
recent series of Iroquoian dictionaries to come out -- for Oneida (by
Michelson et. al.), Onondaga (by Woodbury) and Cayuga (by Dyck et. al.),
all published in the last year or two by U of Toronto Press -- they tackle
the prefixing problem by exploiting this functional/positional difference.

They list as headwords each of the combinations of prepronominal prefixes
(there's around 120 of them, depending on the language).  For example,
here's an entry from the Onondaga dictionary (p. 359):

[ë here stands for a nasal-e written as e + nasal hook in the printed
version]

===============

dës-/dëj-/dëji-/dët-  prepronominal prefix: dualic, future mode, and
repetitive.
Note: The alternant dëj- occurs before pronominal prefixes beginning in y;
dëji- occurs before first person inclusive dual and plural agent prefixes,
second person dual and plural prefixes, and the second person singular
patient prefix; dët- occurs before the second person agent pronominal
prefix -hs- and before pronominal prefixes beginning in s, other than those
taking the dëji- alternant; dës- occurs elsewhere.

================

[Just as an aside, while the Proto-Athabaskans may have mastered the art of
making their language inscrutable to outsiders, the Proto-Iroquois were
clearly no slouches in this department, neither...]

So, what we have here is a headword which is neither a single
morpheme/prefix, nor is it a whole word... rather, it's just a "useful,
recurring chunk".  If a user was trying to identify a word that happened to
start with that particular combination of prepronominal prefixes, they're
able to get that whole chunk at once, instead of having to look up d- for
the Dualic prefix, ë- for the Future mode prefix, and s-/j-/ji-/t- for the
Repetitive prefix.

Once they've managed to strip off that prepronominal chunk, they can move
on to the pronominal prefixes, which have entries that look like the
following:

================

ethi-/ethiy-/-yethi-/-yethiy-  pronominal prefix.
Note: First person inclusive non-singular acting on third person
feminine-indefinite singular, or third person feminine-zoic or masculine
non-singular, interactive pronominal prefix: we (you and I, two or more)
act on her, or somebody, or them.  The alternants ethi-/ethiy- occur
word-initially or medially after ? [glottal stop], -yehti-/-yethiy- occur
medially in remaining environments.  The alternants ethi-/-yethi- occur
with bases that begin in consonants including R, or with bases that begin
in the vowel i, with the loss of the i, ethiy-/-yethiy- occur with bases
that begin in a, e, ë, o or u.

================

Once the user has stripped that off, they then have a shot at finding the
main verb entry... and so long as it doesn't start with one of those
vanishing-i's or an R (a reconstructed proto-consonant slot filler that
Iroquoianists are fond of using, and which is cogante with an
actually-occuring /r/ in Mohawk).

Now, clearly, this is still a pretty complex system, and it's hard to
imagine that an untrained speaker/learner would be able to use it easily --
although, to be fair, they were not the primary intended audience of these
dictionarires.  For them, a better solution might be to treat the whole
prefix string, prepronominal and pronominal together, as an entry, which
might look something like this:

================

dëjyethi-/dëjyethiy-  prefix string.
Note: Dualic, future mode, and repetitive prepronominal prefixes, plus
first person inclusive non-singular acting on third person
feminine-indefinite singular, or third person feminine-zoic or masculine
non-singular, interactive pronominal prefix: we (you and I, two or more)
act on her, or somebody, or them.  The alternant dëjyethi- occur with bases
that begin in consonants including R, or with bases that begin in the vowel
i, with the loss of the i, dëjyethiy- occur with bases that begin in a, e,
ë, o or u.

================

The obvious drawback to this, however, is that there are thousands and
thousands of these strings, which would take up a lot of space in a printed
dictionary... although the rather verbose entries exemplified above could
certainly be streamlined with a judicial use of abbrevations.


Young and Morgan took a similar sort of approach in their "Big Blue" Navajo
dictionary  from 1987.  Verbs in this dictionary are not listed by root as
they are in most other big Athabskan dictionaries, but instead by the third
person singular Imperfective (I) form.  This is followed by the third
person singular Iterative (R), Perfective (P), Future (F) and Optative (O)
forms of the verb -- so, in a sense, each verb entry has five
headwords!  For example [using umlauts for nasal vowels and circumflexes
for high-tone nasal vowels]:

================

ch'éhébääs (I), ch'ínáhábäs (R), ch'íheezbââz (P), ch'íhidoobäs (F),
ch'íhóbääs (O), biL --, to ride out (horizontally) one after another, to be
brought out one load after another (as passengers, in wheeled
vehicles).  Tó yíläädgo 'áLchíní tsékoohdêê' tsinaabääs biL ch'ídahaazbââz,
when the flood came the children were brought out of the canyon one (load)
after another by wagon. (*0bääs: to move by rolling -- a hooplike object.)
(ch'éhi-)

================

At the end of the entry, in parentheses, is a cross-reference to the entry
for ch'éhi-:

================

ch'éhi- (ch'íhi-): ch'í-, Pos. IVb, out horizontally + -hi-/yi-, Pos VIa-c,
seriative, one after another.

================

This is then followed by a *large* table (16x5) showing all the different
prefix strings one can get by interdigitating the inflectional person and
mode prefixes into the lexical/derivational prefix string ch'éhi-.  So, for
instance, it shows the different first person singular prefix strings based
on ch'éhi-, one for each of the five modes:

Imperfective: ch'éhésh-
Iterative: ch'ínáhásh-
Perfective: ch'éhé-
Future: ch'íhideesh-
Optative: ch'íhósh-

A quick scan of the table shows that, in fact, ch'éhi- itself is not an
occurring surface string... there's always some inflectional prefix or
other that changes that string into something else.  So, ch'éhi- is in fact
just a label for the table that holds all these related fully-inflected
prefix strings.

But in practice that actually works out pretty well.  For instance, if you
wanted to take make the first person dual imperfective form of this verb,
you could take the third person singular citation form given as a headword,
ch'éhébääs, strip off the prefix string (everything but the last syllable,
which is the verb root/stem), and then replace it with the first person
dual imperfective prefix string found in the table, giving you ch'íhiibääs.

During my time at Navajo Community College (now Diné College), I remember
talking with several of the Navajo language instructors, and they all
preferred the way verbs were treated in the Big Blue dictionary, compared
to the Big Red analytical dictionary where all the verbs were listed by
root.  Even though verb roots/stems are quite easy to identify in
Athabaskan (much much more so than in Iroquoian), these speakers and
teachers appreciated having actual words as headwords... to them, it seemed
more like what a dictionary is supposed to be like.

Although the system used in the Big Blue dictionary is "better", it still
has its shortcomings.  For instance, if a student were reading along and
came across the word ch'íhiibääs, he would have to do some hunting in order
to figure out what that verb means, since the inflected prefix string
ch'íhii- is not listed as a headword of its own -- it only occurs as part
of the table of inflected strings under the entry for ch'éhi-.  Given that
Young and Morgan have already done all the work of listing what the
inflected prefix strings are, it seems like it wouldn't be that hard to
make a separate index of all those strings and what they decompose to....
though again, as for Iroquoian, we're talking about thousands and thousands
of strings


So, in the end, lexicographers have come up with some pretty effective ways
of dealing with all this pesky prefixing, but it requires one to look
beyond listing just individual morphemes or whole words as headwords, and
to include the possibility of listing prefix strings, whether
actually-occurring ones as in Iroquoian, or somewhat abstracted ones, as in
Navajo.

===
Jordan Lachler
Sealaska Heritage Institute





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 Sponsor ---------------------~-->
Buy Ink Cartridges or Refill Kits for your HP, Epson, Canon or Lexmark
Printer at MyInks.com. Free s/h on orders $50 or more to the US & Canada.
http://www.c1tracking.com/l.asp?cid=5511
http://us.click.yahoo.com/mOAaAA/3exGAA/qnsNAA/HKE4lB/TM
---------------------------------------------------------------------~->

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/



More information about the Lexicography mailing list