FLEx Comments and Several questions

[SaraB]

Hi,

I'm new to this group and to FLEx and to Bantu languages. At the
moment, I'm doing language learning in Lingala and using FLEx as a way
to help me do some language analysis and using my language data as a
means of become familiar with FLEx, lexicography and refresh my
analysis skills. I've read the lexicography and the morphological
parsing articles.

I'm struggling with how what's happening phonetically and
phonologically should be represented in FLEx, since it's using the
orthography of the language and there is not a one-to-one relationship
between the sounds and the characters. At the moment I have three
things which are troublesome.

1. Noun class 9/10 has an underlying nasal which becomes 'm' or 'n'
orthrographically and 'm', 'n' or eng phonetically depending on the
consonant which follows it. I need to enter a prefix for these
classes, but don't know how to represent and underlying nasal or if I
should choose 'n' (which is written the most frequently) and then do
something with the environments.

Speaking of environments, does anybody know if there's an article
similar to the parsing one that talks about how environments should be
used? I understand them linguistically, but haven't quite figured out
how they integrate with FLEx.

2. I have a few lexical roots that start with a tone and an empty
vowel slot, example 'friend' /V́níngá/. The vowel is supplied by the
noun class prefix, e.g. /báníngá/ 'friends.' What's the best way to
represent this in lexeme form field?

3. The issue in question two leads to a third question. For some words
whose root begin with a high-tone empty vowel slot, the high tone of
the root combines with low tone of the noun class prefix to form a
rising tone. Is there a way to help FLEx recognize when this is
happening (especially the parser)?

Thanks,
Sara

[Jeff and Peg Shrum]

Sara wrote:

1. Noun class 9/10 has an underlying nasal which becomes 'm' or 'n'
orthrographically and 'm', 'n' or eng phonetically depending on the
consonant which follows it. I need to enter a prefix for these
classes, but don't know how to represent and underlying nasal or if I
should choose 'n' (which is written the most frequently) and then do
something with the environments.

Speaking of environments, does anybody know if there's an article
similar to the parsing one that talks about how environments should be
used? I understand them linguistically, but haven't quite figured out
how they integrate with FLEx.

------------------------------------------------------------------------

Homorganic nasals are fairly common in noun class 9/10. I searched the help files for Flex and could not find where a class marker like [N] could be used in the lexicon. However, Sena (one of the sample databases that comes with Flex) has homorganic nasal as well. If you look at that sample database and look up the morpheme /im-/ in the citation field and /N/ in the lexeme field. This is the noun class marker for noun class 9. They have used environments quite extensively to define what the surface form of /N/ should be. I think you can get some good ideas from looking at the Sena data.

The help files on environments are pretty good. My one suggestion is that you define as many of the natural classes as you can before defining environments. To get to the natural classes area, first click on the Grammar button, then choose the "Natural Classes" option from the list that appears.

I have not had any experience with tone in Flex, so I can't help you there. I would suggest the program being developed in Africa Area called "P-test" for phonological and tonal analysis.
http://lingtransoft.info/

Jeff S.

[John Heins]

Thanks for your good answer Jeff.

Sara, the allomorphic variants I have in the Sena database for the
class 9-10 prefix are incomplete, but I think they will give you a
good start. (filter for im- at the beginning of the word.) Just a note
that the underlying prefix for both class 9 and 10 is in fact "im-".
(I believe that "impala" fits in that class which is perhaps either
from Zulu or Sotho. In Sena there are only 2 or 3 words which I've
found that preserve the old "im-" prefix, like, "imbu" "mosquito".
Could it be that the im- occurs in all single syllable 9-10 nouns in
Sena, need to check that.) Anyway, you are right that the class marker
is almost almost always realized as an homorganic nasal, which in the
Sena database I've entered as "N", but with the citation form of
"im-". Please note that in Sena these nasals are usually prenasalized,
not syllabic nasals, which are contrastive (eg. n'khondo (3/4) "animal
path" and nkhondo (9-10) "war" ). Please also note that the N- coming
before another nasal fuses to make one nasal. The only way I know to
describe it in a rule is "N > null/ _ N" but I'm not sure I remember
how FLEx deals with nulls.

I hope this helps,
John
Sena adviser for the FLEx team

[Emmanuel HABUMUREMYI]

You are true John, Sena can be a good model for most of Bantu languages in terms of allomorphic variants.  By the the way, I was wondering whether it is possible to import Sena list into the new database to guide new users who want to used such features in their own langageprojects?
Regards,
Emmanuel Habumuremyi

---

De : John Heins <john_...@sil.org>
À : FLEx list <flex...@googlegroups.com>
Envoyé le : Ven 23 Octobre 2009, 16 h 23 min 01 s
Objet : [FLEx] Re: FLEx Comments and Several questions

[SaraB]

Thanks for your reply. I took a look at the "N"/"im-" entry in the
Sena 3 database. I noticed that you entered eng as a phoneme in the
nasal natural class. Is there any reason to do that if I never have an
eng in the orthography (I'm not entering the IPA for any of my words
-- I'm sticking strictly with the orthography)? I notice that Sena 3
does not have any engs in any of the lexical forms. I'm struggling
with how/where/whether I should melding the linguistic aspects with
the lexicographic aspects (as in what point/benefit if any arises from
having eng as it never appears in except in a prenasalized consonant
and only syllable initially -- and only when it's not preceded by an
open syllable).

I'm pretty sure that my 9/10 prefixes are simpler. As far as I can
tell (which is only based my own limited analysis -- I'm in the
process of trying to find a publication(s) that covers Lingala
phonology and morphotactics), there's no distinction between nasal
consonants and prenasalized consonants. Additionally, the
prenasalization drops from pronunciation (but not from the written
form) for voiceless consonants.

Sara

[John Heins]

Hi Sara, nice to hear about your continued work with Lingala.

SaraB wrote:
> I noticed that you entered eng as a phoneme in the
> nasal natural class. Is there any reason to do that if I never have an
> eng in the orthography (I'm not entering the IPA for any of my words
> -- I'm sticking strictly with the orthography)? I notice that Sena 3
> does not have any engs in any of the lexical forms. I'm struggling
> with how/where/whether I should melding the linguistic aspects with
> the lexicographic aspects (as in what point/benefit if any arises from
> having eng as it never appears in except in a prenasalized consonant
> and only syllable initially -- and only when it's not preceded by an
> open syllable).

Sena actually has both an allomorphic "eng" and a full phonemic "eng".
As an allophone: N > velar nasal/ _ velar consonant. (where N is a
prenasalized N, I'm not positive about the syllabic nasal, I'd have to
go back and check it.) But as a phoneme it is expressed
orthographically as: ng' eg. "ng'ombe" "cow" (there is no stop (i.e.
"g") there). If you find that "eng" is always explainable as an
allomorph in Lingala then, no, you shouldn't put it in your phoneme
list.

SaraB wrote

> it never appears in except in a prenasalized consonant and
> only syllable initially -- and only when it's not preceded by

> an open syllable."
I'd say that you have proven that the phone: eng is predictable and so
does not need to be represented in the phoneme list or the
orthography, but that is also an excellent phonemic analysis which you
want to capture. So you should consider writing an allomorph rule for
it, or at least consider making a Phonology Note.

> I'm in the
> process of trying to find a publication(s) that covers Lingala
> phonology and morphotactics), there's no distinction between nasal
> consonants and prenasalized consonants. Additionally, the
> prenasalization drops from pronunciation (but not from the written
> form) for voiceless consonants.

Yes, there is often some kind of interference between nasals and
voiceless consonants in Bantu languages. There is a simple Lingala
grammar on the wikipedia. Here is a description of the process you are
talking about: http://en.wikipedia.org/wiki/Lingala_language#Prenasalized_consonants
Good luck on your continued research. Hope this helps a bit.

[SaraB]

John,

Thanks so much for your advice. I know that in some ways my questions
are very elementary, but I have no experience with Bantu languages so
I'm still looking for my footing. I will definitely check out the link
you posted.

Sara

[Mary Steele]

Dear John,

I am an SIL member serving in Ghana, and got your name from the FLEx google
group, also Hannes Hirzel has just told me that he met you once, and that
you work in a Bantu language.

The reason I am writing to you is that the language I work in has many noun
classes distinguished by prefixes and/or suffixes for sg. and pl. Just now I
am wondering how to number the classes. At the moment I have them numbered
in pairs (sg. / pl. pairs), but don't feel that is the best way. (There are
about twelve pairs.) My question is: do you have any suggestions as to the
best way of numbering?

Hoping to hear from you some time.

Mary Steele

[Jeff and Peg Shrum]

Mary Steele wrote:

 

 

The reason I am writing to you is that the language I work in has many noun

classes distinguished by prefixes and/or suffixes for sg. and pl. Just now I

am wondering how to number the classes. At the moment I have them numbered

in pairs (sg. / pl. pairs), but don't feel that is the best way. (There are

about twelve pairs.) My question is: do you have any suggestions as to the

best way of numbering?

 

--------------------------------------------------------------------------------

 

Below are the Bantu noun classes that are pre-programmed into Flex.  The list covers most of the

noun classes that you will commonly see in a Bantu language.  If you need to add to the list Flex

allows for that if it is necessary.  Noun classes are actually a form of gender, and they dictate

concordance on all the adjectives etc that are modifying the noun. Since there

are more than three most linguists prefer a different term than gender and “noun class” has become most widely used.

 

 

Much has been written about Bantu noun classes, and there are many books and articles

that would be helpful for you.  I can summarize the above list for you to help you get started.

Noun class 1/2 is typically human/animate/kinship terms ( common prefixes are mu-, bu-, for singular and a-, ba- for plural)

Noun class 3/4 is typically plants, other animals and animates not included in ncl 1/2. (common prefixes are mu-/mi-)

Noun class 1a is a sub type of ncl 1/2 where the singular has no prefix on the noun but all the concords are the same as class 1.

Noun Class 5/6 are typically body parts, diseases, and more general things in the environment like sand, rocks and so on. (ni-/ma- are common prefixes for this noun class.)

Noun Class 9/10 is where borrowed words usually go. (common prefixes are syllabic nasals, or nulls)

Noun class 14 often has the same prefix for both singular and plural, only the concords change.  The concords actually are often the same as those for ncl 6, the name “noun class 14/6”.

Noun Class 15 is the infinitives used as nouns and are typically abstract things like “friendship” or “suffering”  

(this prefix is the same as the infinitive marker prefix, in Swahili it is ku-, but I have seen u-, and o- among others)

Noun Classes 16, 17, 18 are the locative classes.  Typically these prefixes concatenate on the original prefixes.

You can find books that give a fuller description of the various noun classes and what their typical surface forms are, I hope what I have written will give you a good start.

 

When deciding which numbers to use with which group of nouns in your language, it would be good to be consistent with other languages in your area and language family.  Is there a Linguistics consultant available to help you with these decisions?

 

From a software point of view, I would encourage you to try and make the set of “inflection features” for Bantu in Flex work, since it will give you greater consistency in your data and the automatic parser will give you better results.  Why reinvent the wheel?

 

Jeff S.

Milange, Mozambique

 

[Robert Hedinger]

I don't know what language you are working on. I guess it is one from West Africa, so it would be quite distant and different from Bantu. What I would do is find out how the linguists in your area do their numbering of noun classes in your language sub-family. I would recommend following their convention. This would make it easier for comparing systems across languages. There might be some similarities to Bantu, but probably not much.

 

Robert

--
You received this message because you are subscribed to the discussion group "FLEx list". This group is hosted by Google Groups and is open for anyone to browse.
To post to this group, send email to flex...@googlegroups.com
To unsubscribe from this group, send email to flex-list-...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/flex-list

[Kevin Warfel]

As someone who has worked in Burkina Faso, just to the north of Ghana, I would like to reinforce what Robert wrote.  I worked in a Gur language (Phuien/Puguli) where the noun class system is all but extinct, so I am not extremely familiar with how they are handled in languages in that part of the world, but I seem to remember that, unlike what seems to be the case in Bantu languages (based on Jeff’s response), there is not a nice, predictable correlation between the singular class and the plural class of a particular noun.  (That is, nouns that take class suffix -o, for example, in the singular, will not all – either unanimously, or even in the great majority – take class suffix -a in the plural, for example.  Rather, half of the nouns that take suffix -o in the singular may take -e in the plural and the other half -a.)  For this reason, each suffix is numbered individually (rather than numbering a singular/plural pair) and a noun is classified according to the COMBINATION of singular/plural class markers it takes.

 

Thus, if the following class suffixes have been identified and numbered (and these are completely hypothetical, but serve to illustrate my point):

 

1) -o SG

2) -i SG

3) -ma SG

4) -bo SG

5) -a PL

6) -e PL

7) -mo PL

8) -ru PL

 

a noun might be classified as 1/5, 1/6, 1/7, 1/8, 2/5, 2/6, 2/7, 2/8, 3/5, 3/6, 3/7, 3/8, 4/5, 4/6, 4/7, or 4/8, depending on the suffixes it takes in the singular and plural.  The numbering is done separately for the singular and plural suffixes because there are so many different combinations attested in West African noun class systems.  (A nice, neat system would be one where nouns of class 1 in the singular would always be of class 5 in the plural, but this is not typical of West African noun class systems.)  As you can see, my system has 16 pairs of possible combinations, but the numbering system is made up of only 8 elements.  Numbering the individual suffixes keeps the number of elements to a minimum and provides a way to indicate which ones accompany each particular noun.

 

But, as Robert suggested, West African linguists have already worked on these systems and have adopted a numbering system, and that is the numbering system that I would also suggest that you use, for exactly the reason that he cited.  I’m sorry I am not familiar enough with it to be able to supply you with more details, though.

 

Kevin Warfel