Native speaker accessibility to Alg. dictionaries

Richard Rhodes rrhodes at COGSCI.BERKELEY.EDU
Mon Mar 29 17:56:02 UTC 2004


Wayne,
	One possible idea to deal with semi-sophisticated users who
forget to strip prefixes is to list forms with a given prefix and
cross-reference with dummy entries. In a dictionary of 10,000
entries, adding 100 or 2000 which get one past the prefix hurdle is
not excessive. So the Cheyenne dictionary would contain the following
entries, appropriately alphabetized:

naaa- see eaa-
naa'a- see ea'a-
na'aa- see e'aa-
na'a'a- see e'a'a-
naae- see eae-
naa'e- see ea'e-
na'ae- see e'ae-
na'a'e- see e'a'e-
naah- see eah-
na'ah- see e'ah-
naak- see eak-
na'ak- see e'ak-
naam- see eam-
na'am- see e'am-
naan- see ean-
na'an- see e'an-
naao- see eao-
naa'o- see ea'o-
na'ao- see e'ao-
na'a'o- see e'a'o-
naap- see eap-
naas- see eas-
na'as- see e'as-
na'ap- see e'ap-
naat- see eat-
na'at- see e'at-
naats- see eats-
na'ats- see e'ats-
naav- see eav-
na'av- see e'av-

naea- see eea-
nae'a- see ee'a-
na'ea- see e'ea-
na'e'a- see e'e'a-
naee- see eee-
nae'e- see ee'e-
na'ee- see e'ee-
na'e'e- see e'e'e-
naeh- see eeh-
na'eh- see e'eh-
naek- see eek-
na'ek- see e'ek-
naem- see eem-
na'em- see e'em-
naen- see een-
na'en- see e'en-
naeo- see eeo-
nae'o- see ee'o-
na'eo- see e'eo-
na'e'o- see e'e'o-
naep- see eep-
na'ep- see e'ep-
naes- see ees-
na'es- see e'es-
naet- see eet-
na'et- see e'et-
naets- see eets-
na'ets- see e'ets-
naev- see eev-
na'ev- see e'ev-

naha- see eha-
nahe- see ehe-
nahk- see ehk-
naho- see eho-
nahp- see ehp-
naht- see est-
nahts- see ests-

naka- see eka-
nake- see eke-
nako- see eko-

nama- see ema-
name- see eme-
namo- see emo-

nana- see ena-
nane- see ene-
nano- see eno-

	etc.

Of course, not all the theoretically possible root onset syllables
are realized, so those that aren't can be eliminated. (I actually did
this for the variants in b ~ p ~ bb, g ~ k ~ gg, and d ~ t ~ dd in
Ottawa.)

The thing about dictionaries is that you are not punished for
redundancy in moderation. And appropriate redundancy can pay big
dividends for users.

Rich



>  > Wayne, I think this is a really important issue.  Linguists and other
>>  scholars are likely to be very obedient, and will read the
>>  introductory material and follow the directions given.  But we can't
>>  expect community members to necessarily do that, and if a dictionary
>>  has community members as a target audience, this has to be addressed.
>
>Absolutely, Monica. It was a wakeup call for me when not a single Cheyenne
>in my class was able to locate the verbs because they all assumed that they
>needed to find the verbs starting with the pronominal prefixes (just as they
>locate words in English dictionaries starting with the first letter). And I
>did
>have a short explanation on the first page of the Intro which they were
>required to read. Since last week's class I have expanded that section a
>great deal, giving examples.
>
>>
>>  Does anyone have any suggestions?
>>
>>  With respect to your specific question - I don't know anything about
>>  Cheyenne
>
>Oh, it pretty much behaves like the other Alg. lgs.
>
>>  - does it have person-marking prefixes on all of the forms
>>  of the verb, making it impossible to choose a form without one for
>>  the entry?
>
>Yes, unlike some Alg. lgs. Cheyenne uses pronominal prefixes even for third
>person, so it has pronominal prefixes for each person of the Independent
>Order.
>
>Now, for AI verbs, the verb stem *is* the 2nd person plural addressee
>imperative form.
>
>Over the years, in my various classes, Cheyennes have had no problem being
>able to pull the verb stem out of the verb and give its meaning. For that
>matter, they usually can do this also for preverbs. So, either I just need
>to remind the students to use the same morpheme approach to locating verb
>stems in the dic'y (and I have full verb subentries under the stem
>headwords), or else I need to pick one of the person, probably either first
>or third as the dic'y headword, and give up wanting to have semantically
>related forms cluster together in the dictionary due to their having close
>spelling based on derivational relationship.
>
>Wayne
>-----
>Wayne Leman
>Cheyenne website: http://www.geocities.com/cheyenne_language
>
>>
>>  - Monica
>>
>>
>>  >There is one message totalling 97 lines in this issue.
>>  >
>>  >Topics of the day:
>>  >
>>  >   1. Cheyenne dictionary: the jury's verdict
>>  >
>>  >----------------------------------------------------------------------
>>  >
>>  >Date:    Tue, 23 Mar 2004 11:11:05 -0500
>>  >From:    Wayne Leman <wayne_leman at SIL.ORG>
>>  >Subject: Cheyenne dictionary: the jury's verdict
>>  >
>>  >Also posted to Lexicography List, March 23, 2004:
>>  >http://groups.yahoo.com/group/lexicographylist/
>>  >
>>  >This semester I am teaching a course at the reservation college called
>>  >Cheyenne Linguistics. Students are fluent speakers of the language, eager
>to=
>>  >
>>  >learn. I have had most of them in classes I have taught in the past. Last
>>  >week their home assignment was to work with the current release of our CD
>>  >dictionary of Cheyenne. I gave the students a homework sheet with
>specific
>>  >things to try to find in the dictionary. A key part of the homework was
>for
>>  >the students to locate three entries which happened to appear on the same
>>  >page of the dictionary, and all of which have the same stem -m=E9sehe
>meanin=
>>  >g
>>  >'eat.' This was the first field test of a stem-based Cheyenne dictionary.
>>  >
>>  >The first entry they were to find was the verb =E9m=E9sehe 'he is
>eating,'
>>  >composed of the third person prefix =E9- plus the verb stem.
>>  >
>>  >The second entry was n=E1m=E9s=EAh=E9t=E1no 'I want to eat,' composed of
>the=
>>  >  first
>>  >person prefix n=E1-, the verb stem, plus a suffix -t=E1no meaning 'want
>to.'=
>>  >
>>  >
>>  >The last entry was the noun m=E9s=EAh=E9stoto meaning 'potatoes,'
>composed o=
>>  >f the
>>  >steam 'eat', a nominalizer -htot and an animate plural suffix -o.
>>  >
>>  >We discussed in class how the students did interacting with the
>dictionary.
>>  >Although all of the students know well the verb prefixes, not one was
>able
>>  >to locate the verbs 'he is eating' and 'I want to eat'. The issue was
>that
>>  >they were trying to locate 'he is eating' by its first letter, the prefix
>>  >=E9-, and 'I want to eat' by its prefix n=E1-. Each of the verbs they
>were
>>  >looking was clearly listed as subentries under their "headword" verb
>stem,
>>  >but that did not help when they were looking for those verbs under their
>>  >pronominal prefixes.
>>  >
>>  >I thanked the students for helping me realize that something more needed
>to
>>  >be done so that they and other Cheyennes with far less class time than
>they
>>  >have had would be able to locate verbs. I then illustrated why the verbs
>>  >were arranged as they are in the dictionary, to group together words
>(both
>>  >nouns and verbs) which have the same stem. They quickly saw from examples
>I
>>  >put on the board that if we listed all forms of all verbs, with all
>person
>>  >prefixes, all pluralization suffix combinations, all modes, etc., that
>the
>>  >dictionary would mushroom in size to have millions of entries.
>>  >
>>  >This problem of native speakers locating entries in a dictionary for a
>>  >language which has pronominal prefixation (or even more difficult
>prefixal
>>  >allomorphy) is a common one, and has been discussed in the
>lexicographical
>>  >literature.
>>  >
>>  >There are various solutions, none of which are perfect. We can enter only
>>  >complete verbs, choosing one of the persons, such as third person, to be
>the=
>>  >
>>  >default form for entering verbs. This would obscure the stem relationship
>>  >between verbs and nouns derived from them which start with the first
>letter
>>  >of the stem. We can have a more detailed explanation in the Introduction
>to
>>  >the dictionary of how the verbs are composed (we have spent a lot of time
>on=
>>  >
>>  >this in previous classes and students understand verb construction well)
>and=
>>  >
>>  >how to locate verbs by their stems. I already had the following paragraph
>on=
>>  >
>>  >the first page of the dictionary, which they were asked to read in the
>>  >homework assignment:
>>  >
>>  >"This dictionary is arranged in alphabetical order according to the first
>  > >letter of the Cheyenne entry. Many entries are parts of Cheyenne words
>>  >called morphemes ("blocks") or roots. Most of these partial word entries
>>  >also include complete Cheyenne words as examples. There is also a
>separate
>>  >English index to the Cheyenne words in the dictionary."
>>  >
>>  >Obviously, this was not complete enough to help the students. At a
>minimum,
>>  >there needed to be examples illustrating how to locate verbs by the first
>>  >letter of their stem.
>>  >
>>  >I know that others of you have struggled with creating dictionaries which
>>  >are accessible by language speakers themselves, not just the linguistic
>>  >community.
>>  >
>>  >What solutions have you found which help language speakers use their own
>>  >dictionaries fairly easily?
>>  >
>>  >Thank you,
>>  >Wayne
>>  >-----
>>  >Wayne Leman
>>  >Cheyenne website: http://www.geocities.com/cheyenne_language
>>  >
>>  >------------------------------
>>  >
>>  >End of ALGONQDICT Digest - 11 Dec 2003 to 23 Mar 2004 (#2004-1)
>>  >***************************************************************
>>
>>
>>  --
>>  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>  Monica Macaulay
>>  Department of Linguistics
>>  University of Wisconsin
>>  1168 Van Hise
>>  1220 Linden Drive
>>  Madison, WI  53706
>>
>>  email:  mmacaula at wisc.edu
>>  phone:  (608) 262-2292
>>  web:  http://ling.wisc.edu/~macaulay/monica.html
>>  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


--
******************************************************************

  Richard A. Rhodes
  Department of Linguistics
  University of California
  Berkeley, CA 94720-2650
  Voice (510) 643-7325
  FAX (510) 643-5688

******************************************************************



More information about the Algonqdict mailing list