Paradigms in FLEx
Roland Fumey
Dear All
I realize a number of messages have already gone back and forth
in recent years about paradigms in FLEx, but I still wonder how to
handle this best. A few times I have also heard and read that FLEx
doesn't handle paradigms well yet but that it will sometime in the
future.
My first question therefore is, has paradigm handling somehow slipped into FLEx without me realizing it or is there anything in the works? (It seems that the former is not the case since FLEx help still says, “Currently, Language Explorer has no fields in which you can to collect and organize paradigm (inflected) forms of words.”)
Now, here is my problem: The verbs of the language I have been
working with for over 30 years can be conjugated into tens of
thousands of forms (and this is probably an understatement).
Whereas the portmanteau subject and other prefixes, as well as
many suffixes are quite regular and predictable, the absolutive
affixes together with the verb roots are definitely not. They need
to be listed with the verbs, otherwise one wouldn't know how to
conjugate them. (Similar to the verb forms ‘go, went, gone’
without them one wouldn't know how the verb ‘to go’ is used in
English.)
Here's an example. Stem forms of the verb ‘to take’:
| Subject sg | Subject pl | Conjugation Classes | Translation | |
| Collective Forms (when taken together): | ||||
| Abs 1 sg m/f | ndap -an | ndøp -øn -em | 1, 5c | take me |
| Abs 2 sg m/f | ndap -aq | ndøp -øq -em | 1, 5c | take you(sg) |
| Abs 3 sg m/f | ndaf | ndaf -em | 1, 5c_irreg | take him/her |
| Abs 1 pl m/f | upo -no | upo -no | 1, 6 | take us |
| or: | tonqo -no | tonqo -no | 1, 6 | (same as above) |
| Abs 2/3 pl m/f | upo -zo | upo -zo | 1, 6 | take you(pl)/them |
| or: | tonqo -zo | tonqo -zo | 1, 6 | (same as above) |
| Abs 3 sg n | ndap | ndap -em | 1, 5c | take it |
| Abs 3 pl n | up -a | up -ø -nem | 1, 6_irreg | take them (neuter) |
| or: | tonqo -a | tonqo -a -nem | 1, 6_irreg | (same as above) |
| Abs 3 sg p | ndap -az | ndøp -øz -em | 1, 5c | take it (4th ‘gender’) |
| Abs 3 pl p | upo -zo | upo -zo -nem | 1, 6 | take them (4th ‘gender’) |
| or: | tonqo -zo | tonqo -zo -nem | 1, 6_irreg | (same as above) |
| Distributive Forms (i.e. when taken individually): | ||||
| Abs 1 pl m/f | upø -ng -im | upø -ng -im | 1, 7 | take us |
| or: | tonqo -ng -im | tonqo -ng -im | 1, 7 | (same as above) |
| Abs 2/3 pl m/f | upø -g -im | upø -g -im | 1, 7 | take you(pl)/them |
| or: | tonqo -g -im | tonqo -g -im | 1, 7 | (same as above) |
Apart from the above there are also stative stem forms.
It would be great if there was a way to add paradigm lists or
tables like the above to individual entries. And preferably all
the involved roots and affixes should then also be linked to their
respective FLEx entries and the conjugation class numbers should
point to their respective entries in a conjugation table.
Years ago I created a rather complex Access data base (involving
quite a bit of VBA code) for verb stems, but if possible I would
like to incorporate this into my FLEx lexicon, where I would think
it should belong. (As a workaround, I would like to jump back and
forth between entries belonging together in Access and FLEx,
similar to the Jump2Toolbox.dll for Toolbox.)
As I said above, I'm in no way wanting to put all verb forms into FLEx, just the stem forms so that the verbs can be conjugated accurately.
Thanks a lot in advance!
Roland
PS: Just in case you might wonder why I call the affixes
belonging to the stem absolutive and not object
affixes: This is because although in transitive verbs they do
refer to the object, in intransitive verbs they refer to the
subject.
ron_lo...@sil.org
Hi Roland,
The first thing I would like to clarify is what you want to track in FLEx vs. what you would like your dictionary entries to look like. For the former, having various variant forms of the verb stem seems logical to me. For the latter, whereas FLEx does not have a specific field or way to handle paradigms, there is the ability to create a table for an entry (in FLEx Help, search for ‘table’ and look at the topic ‘Format a table’.) I’m wondering, though, if you would really want a table like below for each entry in your published dictionary.
Ron
--
"FLEx list" messages are public. Only members can post.
flex_d...@sil.org
http://groups.google.com/group/flex-list.
---
You received this message because you are subscribed to the Google Groups "FLEx list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to flex-list+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/flex-list/be089925-fd40-52cd-8516-066793df5a80%40sil.org.
Mike Maxwell
> I’m wondering, though, if you would really want a table like
> below for each entry in your published dictionary.
inflected forms--just enough irregular forms to predict other irregular
forms (and therefore regular words list only the citation form). That's
ok for dictionary users who understand the language's morphology, but
less good for those who are not linguistically sophisticated. I don't
know whether there are dictionaries that take a more liberal view of
forms that should be included.
Electronic dictionaries are a different beast. If you have to type in a
word to search for, and there's no smart fuzzy search, or morphological
parser front end, then it can be very hard for a linguistically
unsophisticated user to find words. There are various ways to help with
this, although short of using a morphological parser as a front end,
they work better with suffixing languages than with prefixing languages.
Mike Maxwell
University of Maryland
Roland Fumey
Thanks Ron and Mike for your replies.
As to what I want to track in FLEx vs. what I would like the
dictionary entries to look like:
If at all possible I'd like to have a system that allows me to
list all stem forms with grammatical description, inflection
classes and possibly also translation in FLEx and then from that
extract all relevant forms to a published dictionary, hopefully
in such a way that I could try out various formats without
needing to do a lot of changes in the FLEx entries themselves.
As you point out, Ron, a table under each entry might not be a
good idea in a printed dictionary. Maybe I could just list the
stem forms without too much explanation in the printed entries
and have paradigm tables in an appendix. I have actually typeset
a dictionary for colleagues that had over 100 pages of such verb
paradigms in an appendix. But that had all been done manually in
MS Word. However, I really would like to have all that
information in FLEx itself.
I don't think that FLEx allows certain parts of an entry to be
printed in an appendix (or does it?). But maybe such parts could
be marked somehow and then be extracted automatically or
semi-automatically.
As you mention, Mike, electronic dictionaries are a different story again. Here we might have a paradigm table right in oder under an entry or connected by a link, but this would probably depend a lot on the software used. (I'm currently not very familiar with that.)
You also mention searching for a word ‘work[s] better with
suffixing languages than with prefixing languages’. This is
definitely true and another challenge, as all finite verbs in
the language are prefixed. The stem forms after that very often
can't occur by themselves and infinitives sometimes have a
completely different form.
For example the verb ‘to go’: root –av; absolutive stems 1sg:
–nav, 2sg: –qav, 3sg: –wav, 1pl: –niav, 2/3pl: –zav; infinitive:
okho (Examples of full verb forms would be, among thousands,
‘I'll go’ etc: 1sg: nǿnáv, 2sg: qóqáv, 3sg: géwáv, 1ple:
níniáv, 1pli: zíniáv, 2pl: zózáv, 3pl: gézáv.)
I'm currently using the absolutive stem 1sg: –nav as the
citation form. The root –av by itself doesn't mean anything to
anybody. (Absolutive prefixes and/or suffixes are an integral
part of the verb stems. Verbs that have them can never occur
without.)
BTW: What is the correct way to mark roots or stems that always
need to be affixed?
Thanks for any insights,
Roland
PS: I'll have a closer look at table formatting in FLEx.
On 6/13/2023 5:48 AM, ron_lo...@sil.org wrote:
I’m wondering, though, if you would really want a table like below for each entry in your published dictionary.
In my experience, print dictionaries include a minimal number of inflected forms--just enough irregular forms to predict other irregular forms (and therefore regular words list only the citation form). That's ok for dictionary users who understand the language's morphology, but less good for those who are not linguistically sophisticated. I don't know whether there are dictionaries that take a more liberal view of forms that should be included.
Electronic dictionaries are a different beast. If you have to type in a word to search for, and there's no smart fuzzy search, or morphological parser front end, then it can be very hard for a linguistically unsophisticated user to find words. There are various ways to help with this, although short of using a morphological parser as a front end, they work better with suffixing languages than with prefixing languages.
Mike Maxwell
University of Maryland
Asigwan
Roland,
One thing to think about is how you want a finished entry to look. In a printed dictionary, I think the most you could ask for (due to how much space this will take up in print) is the actual verb forms under, for example, the rest of the entry. Even that might be too much unless you are going to have a several-volume dictionary. Would something like this below work for you? The verb forms, based on my language, would be in a predescribed pattern of 1st sg, 2nd sg, 3rd sg m, 3rd sg fem, 1st pl, 2nd pl m, 2nd pl fem, 3rd pl m, and 3rd pl fem divided up under predescribed labels like Aorist, Perfective, Imperfective, etc. Then there are participles and imperatives. We even have negative perfectives, imperfectives, particles, etc. So the list after each verb entry would be pretty long. Make it a smaller font for sure.
Ecc, v.t., to eat......rest of entry... Aorist: cceɣ, ṯecceḏ, icc, ṯecc, necc, ṯeccem, ṯeccent, ccen, ccent; Perfective: cciɣ, ṯecciḏ, icca, ṯecca, necca, ṯeccim, ṯeccint, ccan, ccant; Imperfective: tetteɣ, .....
This would, I'm sure, at least double the size of a printed dictionary, but it is pretty cool info to have. A sophisticated user of the dictionary might be able to figure out most verb entries without them all being listed, but that person would probably be one in 100,000 in our language - if that many. And we also have some really wild verbs that, though regular, the schwas and other vowels are a nightmare to try and figure out. Longer ones start looking like this: ṯseqrujjuƹem, ṯseḵmummusem, ṯserkukuƹem and worse. The negative forms can also change the vowels.
Once you know what you want, a FLEx expert would have to see how close you could come to an entry like that using the fields FLEx has. We just have an electronic dictionary using LexiquePro. Having all these forms is invaluable for spelling in our NT and such. I do not do this in FLEx. So don't ask me. :)
Crockett
--
"FLEx list" messages are public. Only members can post.
flex_d...@sil.org
http://groups.google.com/group/flex-list.
---
You received this message because you are subscribed to the Google Groups "FLEx list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to flex-list+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/flex-list/757735c7-9b3e-8483-814d-06ac70b1444e%40sil.org.
Kevin Warfel
Crockett,
I can imagine at least two ways of entering these kinds of data in FLEx.
1) A series of custom fields
2) Irregularly inflected variants
Irregularly inflected variants would create a minor entry for each form, so that a user could look up that form and find information that referred them to the correct main entry. In an electronic publication, there would be a clickable link to the main entry, enabling the user to jump to it. Configuration of the irregularly inflected variants in the format shown in your sample entry (or something quite close to it) should be possible, though I’d have to actually try it to be absolutely certain of that.
Custom fields could easily be configured to display as shown in your sample entry, but they would not a generate minor entry for each form, nor would there be any sort of clickable link between a particular form and the main entry it relates to.
If Roland finds the format of your entry to be a pattern acceptable (or adaptable) to his situation, these options might work for him. This is different than a table, however. The tabular presentation might be in an appendix, as has been suggested. If he goes that route, I’m not sure how to provide a link in each entry to the appropriate table, though. Maybe someone else has an idea of how to do that.
Kevin Warfel
Associate Dictionary & Lexicography Services Coordinator
Rapid Word Collection workshop consultant
To view this discussion on the web visit https://groups.google.com/d/msgid/flex-list/000001d99e9a%2463f568c0%242be03a40%24%40yahoo.com.
Jan
--
"FLEx list" messages are public. Only members can post.
flex_d...@sil.org
http://groups.google.com/group/flex-list.
---
You received this message because you are subscribed to the Google Groups "FLEx list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to flex-list+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/flex-list/757735c7-9b3e-8483-814d-06ac70b1444e%40sil.org.
Roland Fumey
Thanks Jan! I didn't realize this.
If I choose bound root or bound stem as Morph Type, then FLEx places an asterisk in front of the headword, unless I use an entire word as the citation form.
That's very helpful to know.
Roland
To view this discussion on the web visit https://groups.google.com/d/msgid/flex-list/CAFmu-PARTevwRXkqh_g7TBVux5o9oMouTB5Fvj38%2B2J8D1Cu4A%40mail.gmail.com.
ron_lo...@sil.org
Hi again Roland,
I think creating variants for the stems that are irregular might be a good way to go. Choose an irregular form that is representative and make it an irregularly inflected form variant and assign it a variant type that you have created. Here’s a stab at what it might look like. (I created a dummy FLEx project to illustrate this. I’m sure have things wrong in places, but I hope you see the principle.)
Lexical entry:
Each of the variant forms are clickable.
Variant types list:
Dictionary Preview:
You can choose which of these variants show up as a minor entry.
FLEx Entry:
If you need help with the mechanics of doing this I’d by happy to meet offline.
As far as creating an appendix with full paradigms. I think myself or others with a bit of programming skills could make a FlexTools module that would do this for you.
Ron
From: flex...@googlegroups.com <flex...@googlegroups.com> On Behalf Of Roland Fumey
Sent: Thursday, June 15, 2023 10:24 AM
To: FLEx list <flex...@googlegroups.com>
Subject: Re: [FLEx] Paradigms in FLEx
Thanks Jan! I didn't realize this.
To view this discussion on the web visit https://groups.google.com/d/msgid/flex-list/9f313e18-b168-1c1f-1a64-eb100db940b6%40sil.org.
Beth-docs Bryson
On Jun 15, 2023, at 5:33 AM, ron_lo...@sil.org wrote:
Hi again Roland,I think creating variants for the stems that are irregular might be a good way to go. Choose an irregular form that is representative and make it an irregularly inflected form variant and assign it a variant type that you have created. Here’s a stab at what it might look like. (I created a dummy FLEx project to illustrate this. I’m sure have things wrong in places, but I hope you see the principle.)Lexical entry:
<image004.png>
Each of the variant forms are clickable.Variant types list:
<image005.png>Dictionary Preview:<image001.png>
You can choose which of these variants show up as a minor entry.FLEx Entry:
<image002.png>
To view this discussion on the web visit https://groups.google.com/d/msgid/flex-list/03b501d99f74%24c4738cb0%244d5aa610%24%40sil.org.
Roland Fumey
Hi Beth
That sounds very interesting! I suggested
such a word form generator many years ago. I wonder if/how it
will work since I need to use ToneParse for accurate parsing.
I have now requested access to the new FLEx Parsing List.
Roland
To view this discussion on the web visit https://groups.google.com/d/msgid/flex-list/92A0DADF-D2EF-471E-B44F-44EC4511407A%40sil.org.
Roland Fumey
Hi Ron,
Thank you very much again for your suggestions.
I played around a bit with variants but then
realized that this only solves part of the problem. Whereas it
show us all the root variants of that verb, it doesn't tell us
how the verb stem is formed, i.e. what absolutive affixes it can
take and to what verb class these belong to. (I have currently
14 of such classes, but there are probably more to come.)
The verb ‘to take’ we looked at is actually complex in several different ways: It has several different verb roots, absolutive affixes of more than one class, a couple of irregular forms that don't fit in any class, plus vowel adaptation (‹upa› for example, is lenited to ‹upø› [‹ø› is a central vowel] before certain suffixes and adjusted to ‹o› before C‹o›]).
Let's maybe look at a simpler verb first. The verb ‘to go’ that I mentioned in one of my earlier emails has the following forms: root *av; absolutive stems 1sg: *nav, 2sg: *qav, 3sg: *wav, 1pl: *niav, 2/3pl: *zav; infinitive: okho (Examples of full verb forms would be, among thousands, ‘I'll go’ etc: 1sg: nǿnáv, 2sg: qóqáv, 3sg: géwáv, 1ple: níniáv, 1pli: zíniáv, 2pl: zózáv, 3pl: gézáv.)
Here I would like the dictionary entry to
look something like this:
*nav (Abs 1sg: *nav, Abs 2sg: *qav,
Abs 3sg:
*wav, Abs
1pl: *niav, Abs 2/3pl: *zav; inf: okho)
vi go, leave
In addition there should also be a way to
indicate the inflection class either for each stem or for a
group of stems. For this verb it is inflection class 2, which
consists of the following absolutive prefixes:
The single consonant allomorphs occur before vowels (except for ‹ni› in 1pl, only in the verb *nav).
Now, how can I link the root -av with the above absolutive prefixes so that the stem forms are created and automatically appear in the entry? These stem forms are not variants but combinations of a root with affixes. (In addition the root -av must also somehow be restricted to the above absolutive prefixes for the parser.)
I assume the absolutive affixes could be
linked to their respective verbs in Grammar–Category but the
problem is that my verb Affix Template is already extremely
complex with 14 slots and the Subject Prefix having 350 entries
and the Tone pattern (which is treated as a suffix by ToneParse)
having about 60 entries (both with more to come). Now, if I
wanted to create individual affix tables for each verb class it
would mean that I would need at least 14 such tables which would
need to be accurately updated whenever additional affixes turn
up, which is bound to happen. This would just not be manageable
any more.
Once I find a solution how to go about for
this relatively easy verb ‘to go’, I'll try to tackle the verb
to ‘to take’ and others.
Thanks again for any help!
Roland
PS: A FLExTools module to create an appendix with paradigms
sounds very good!
To view this discussion on the web visit https://groups.google.com/d/msgid/flex-list/03b501d99f74%24c4738cb0%244d5aa610%24%40sil.org.
ron_lo...@sil.org
Hi Roland,
Here’s another go at it now that I understand a little more.
I’m wondering if you could tread the absolutive affixes as derivational affixes. With derivational affixes you can choose one or more inflection classes they attach to. These don’t need to be (in fact, aren’t allowed) in your templates. Then you can create the derived form entries like qav, wav & zav. Set these forms to have components. The dictionary configuration system is all set up to show complex forms of a given entry. Additionally you can create complex form types that correspond to the different absolutive forms. Here are the ones I defined:
Here’s my result. The parsing works for these as well (I didn’t do them all). Note: I set the derived entries as ‘Is Abstract Form = yes’.
To view this discussion on the web visit https://groups.google.com/d/msgid/flex-list/f6163dd9-10a1-9602-953c-93127fa9c55f%40sil.org.
Roland Fumey
Hi Ron,
Thank you very much for that. We're heading in the right direction, I think!
I was writing you a longish email because I just couldn't figure out how to add an inflection class to the verb stems, but suddenly it worked, possibly after I checked the ‘Is Abstract Form’ box in the derived entries. Great!
My dictionary entry currently looks like this:
which is pretty good, but I haven't found out how to add the verb
class to it. In the lexical entry it is defined as follows:
How do I add the verb class in the dictionary? (Checking
Grammatical Info and Category Info only added ‘v’ to each Complex
Form, which is not what I had in mind.)
I haven't been successful with parsing full forms of this verb
yet. But this is probably TonePars related. (The verb ‹gewav›
actually comes with seven different tone patterns!)
Parsing had all worked well in Toolbox and CarlaStudio (together with TonePars) but I'm still struggling to figure out things in FLEx. (The reason why I moved from Toolbox/Carla to FLEx was because I was really struggling to keep things consistent there.)
Thanks again for all your help!
Roland
To view this discussion on the web visit https://groups.google.com/d/msgid/flex-list/07c801d9a36d%24abf4f320%2403ded960%24%40sil.org.
ron_lo...@sil.org
Hi Roland,
Looking good. I should have given you more details about Inflection classes. You define them in the Grammar section and the category view. As you likely found out, you can add Classes and SubClasses. E.g.
Also for the derived forms to parse correctly, you need to select the appropriate inflection classes for the derivational suffixes in these fields:
To get the inflection class on the main stem to show up. Right click on the preview pane of the entry on the grammatical category. (circled in red here) Then select Configure Dictionary
Click the check box for Inflection class:
To view this discussion on the web visit https://groups.google.com/d/msgid/flex-list/48ae4ce6-6bc1-7171-99e7-5ce7678fbc4d%40sil.org.
Roland Fumey
Hi Ron
Thanks a lot. This is very helpful.
Now the verb class is displayed in the
dictionary as it should:
Just to make sure I understood you correctly in your previous email about ‘derivational affixes ... don’t need to be (in fact, aren’t allowed) in your templates’: I originally had ‘AbsolutivePf’ and ‘AbsolitiveSf’ defined as slots in the Verb Affix Templates, filled with all those individual affixes. But now I removed those two slots both in the actual table as well as in the Affix Slots list underneath it. Is this correct?
I then defined Class 2 in the Inflection Class Info underneath that Affix Slots (with more classes to be added later).
And now for each of the Absolutive prefixes I defined
‘Derivational, Verb → Verb’ in the Category Info dialog box of the
Grammatical Info Details. (I need to do that later for all
absolutive suffixes too.)
If I understand things correctly the parser normally assumes that
all inflectional affixes are defined in the template (in this case
the verb template). But now by having defined those absolutive
affixes as ‘Derivational, Verb → Verb’ the parser should still
find and include those affixes when it parses a verb. Is that so?
Roland
To view this discussion on the web visit https://groups.google.com/d/msgid/flex-list/00c001d9a3b0%248c253300%24a46f9900%24%40sil.org.
ron_lo...@sil.org
Hi Roland,
Yes. Everything you said is right.
To view this discussion on the web visit https://groups.google.com/d/msgid/flex-list/55e49f67-54eb-fcf4-7894-d325ab6a7fc9%40sil.org.
