How to handle verb paradigms in lexicon

[stephanie villard]

Hi,
I am a new user of Language Explorer. I am documenting a variant of
Chatino, Otomanguean language of Oaxaca Mexico. I am wondering how I
could handle the verb paradigms in the lexicon. So far it seems that
all verb forms have to be specified because they show a lot of
irregularities. I have picked the Potential form as the citation form
because its aspectual morpheme shows less allomorphy, plus it does not
obscure the root as all other aspectual morpheme tend to do. So
ideally I would have the Potential form as the main entry and then I
would specify all 3 other aspectual forms in the same entry. How can I
do that and will they show up in the dictionary view? Also, if the 3
forms are not main entries will they be recognized in text parsing?
Thanks,
s**

[Ronald Moe]

When a verb has a set of inflected forms that are irregularly inflected, you
should create a "variant" entry for each of the irregularly inflected forms.
One way you can do this is by creating a new entry for the inflected form
and then linking it to the main entry in the "Variant of" field. You can
then specify how it relates to the main entry in the "Variant Type" field.

However it is usually easier to create the "variant" entry from within the
main entry. You do this in the "Variants" section toward the bottom of the
entry. Click somewhere on the "Variants" line and then click "Insert
Variant". This brings up a dialog box where you type the inflected form.
FLEx creates an new entry in the database for the inflected form and lists
the inflected form in the "Variants" section of the main entry. You will see
the inflected form in the "Variant Form" field. You can specify how the
inflected form relates to the basic form in the "Variant Type" field. The
list of "variant types" are maintained in the Lists area. You can add new
variant types there. I set up a new variant type for each tense/aspect form
that I need for my dictionary. For instance here is an example from my Greek
dictionary. The entry for the basic form looks like this:

Lexeme Form: rus
Citation Form: ruomai
Summary Definition: rescue
...
Definition: To rescue someone from (ek) danger.
------------
Variants
Variant Form: errusthen
Variant Type: Aorist Passive

The entry for the irregularly inflected form looks like this:

Lexeme Form: errusthen
...
Variant Type: Aorist Passive
Variant of: ruomai

I created a variant type in the Lists-Variant Types pane that looks like
this:

Name: Aorist Passive
Abbreviation: aor pass of
Reverse Abbr.: aor pass

The two entries are displayed as follows. First the main entry:

ruomai (aor pass errusthen) V To rescue someone from (ek) danger.

Then the minor entry:

errusthen (aor pass of ruomai 'rescue')

You can adjust the formatting (parentheses, quotes) as needed in the
Tools-Configure-Dictionary menu.

The parser uses the Lexeme Form field in the main entry and any allomorphs
that you have listed in the "Allomorphs" section of the main entry. For this
particular verb I need four allomorphs. The underlying/basic form "rus" I
put in the Lexeme Form field:

Lexeme Form: rus

I then create three allomorphs in the Allomorphs section:

Stem Allomorph: rrus
Environments: /e_th
Stem Allomorph: ru
Environments: /_s
Stem Allomorph: rru
Environments: /e_s

The parser divides errusthen into e-rrus-the-n and finds the root "rrus" in
the first Stem Allomorph field.

You can find more help and explanation about formatting your dictionary in
my paper "Intro to Lexicography" in the Help-Resources menu. You can find
more help on allomorphs and parsing in Andy Black's paper "Intro to Parsing"
also under Help-Resources.

Ron Moe

--
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
No virus found in this incoming message.
Checked by AVG - www.avg.com
Version: 9.0.801 / Virus Database: 271.1.1/2810 - Release Date: 04/13/10
23:31:00

--
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

[Andreas Joswig]

Dear Ron,
Thanks for your helpful response to Stephanie. One thing confuses me: You
create a minor entry for errusthen as an irregular form, and then you
provide the parser with sufficient information through your allomorphs to
actually parse the word. Isn't that overkill? My understanding was so far
that as long as the parser is happy, you don't need any irregular forms in
the lexicon. Or do you include the variant form only for the human
dictionary user, and the allomorphs for the parser?

My response to Stephanie's question would be to use different stem forms as
allomorphs, as you have also suggested, and variant forms only if they
cannot be constructed with any of the material provided by the rules of the
grammar.

Andreas

Ron wrote:

The two entries are displayed as follows. First the main entry:

ruomai (aor pass errusthen) V To rescue someone from (ek) danger.

Then the minor entry:

errusthen (aor pass of ruomai 'rescue')

For this particular verb I need four allomorphs. The underlying/basic form
"rus" I put in the Lexeme Form field:

Lexeme Form: rus

I then create three allomorphs in the Allomorphs section:

Stem Allomorph: rrus
Environments: /e_th
Stem Allomorph: ru
Environments: /_s
Stem Allomorph: rru
Environments: /e_s

The parser divides errusthen into e-rrus-the-n and finds the root "rrus" in
the first Stem Allomorph field.

[Jeff and Peg Shrum]

Ron's explanation just about covers it. Thank you for taking the time to
give such a detailed answer. I too thought of Greek in the answer of
Stephanie's original question. I would add that Greek lexicons use the
first person singular indicative form of verbs as the lexeme forms. There
is much precedence for not using the infinitives for the Lexeme/headword in
lexicons. I would encourage you to use whatever form of the verb that is
most transparent in figuring out the morphology. I would add though you
will need to be consistent and use the same form for the headword of all
verbs.

Jeff S.
Milange, Mozambique

[Jeff and Peg Shrum]

Ron's explanation just about covers it. Thank you for taking the time to
give such a detailed answer. I too thought of Greek in the answer of
Stephanie's original question. I would add that Greek lexicons use the
first person singular indicative form of verbs as the lexeme forms. There
is much precedence for not using the infinitives for the Lexeme/headword in
lexicons. I would encourage you to use whatever form of the verb that is
most transparent in figuring out the morphology. I would add though you
will need to be consistent and use the same form for the headword of all
verbs.

Jeff S.
Milange, Mozambique

-----Original Message-----
From: flex...@googlegroups.com [mailto:flex...@googlegroups.com] On

Behalf Of Ronald Moe
Sent: Friday, April 16, 2010 1:01 AM
To: flex...@googlegroups.com

[stéphanie villard]

Thanks Ron,
it really helped. Now, another question:
Is there anyway to hide the variant forms as entries in the dictionary view? I am using one inflected form as the lexeme (potential aspect), and then all other 3 aspectual forms (Completive, Progressive and Habitual) appear in that same entry in the dictionary view. So I don't want the latter forms to reappear as independent entry in the dictionary.
Thanks,
s**

--
Stéphanie Villard
Doctoral Candidate
Linguistics Department
University of Texas at Austin
stephv...@mail.utexas.edu

[Beth (work) Bryson]

In Configure Dictionary, when you have Root-based view chosen (I
think that's what you have), you probably want to scroll to the very
bottom and configure it so Minor Entries don't show.

In FW 6.0.3 it was supposed to now be possible to uncheck that box
and have the minor entries not show (they would just show as
subentries of the main entry they are a derivation of, not as a
headword themselves). But I think that isn't totally working
correctly. Try it and see if it does. If it doesn't make the
separate entries for the derived forms go away, then open up the
section for Minor Entry and just uncheck all the boxes there. That
should make them not show.

-Beth

> unsub...@googlegroups.com

> For more options, visit this group at http://groups.google.com/
> group/flex-list
>
>
>

> --
> Stéphanie Villard
> Doctoral Candidate
> Linguistics Department
> University of Texas at Austin
> stephv...@mail.utexas.edu
>
> --
> 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-

> unsub...@googlegroups.com

[Robert Hedinger]

Yes, in 6.0.3 each minor entry can be unchecked to make it not appear as a
minor entry.

I don't understand your second to last sentence "...open up the

section for Minor Entry and just uncheck all the boxes there"

Robert

--------------------------------------------------
From: "Beth (work) Bryson" <Beth-wor...@sil.org>
Sent: Wednesday, April 21, 2010 5:43 AM
To: <flex...@googlegroups.com>
Subject: Re: [FLEx] How to handle verb paradigms in lexicon

[Ronald Moe]

stéphanie Villard wrote:
"Is there anyway to hide the variant forms as entries in the dictionary
view?"

Beth Bryson has mentioned one way. First I'll mention the other way and then
talk a bit about several underlying problems. You can control how variant
forms are presented either globally or on a one-by-one basis. Beth mentions
the global control: Go to Tools-Configure-Dictionary. Scroll down to the
bottom and check or uncheck the box next to Minor Entry. If the box is
unchecked, FLEx will not generate any minor entries for variants or
irregularly inflected forms. (This is probably not what you want.) The other
way is to control the minor entries on a one-by-one basis. Go to Lexicon
Edit-Entry pane. Scroll down to the Variants section. There are four fields
in the Variants section (if Show Hidden Fields is checked). The third field,
Show Minor Entry, allows you to specify whether you want a minor entry to be
generated for just this variant form. You can use Bulk Edit to check/uncheck
the Show Minor Entry boxes for lots of entries at once. One factor in
deciding whether you need a minor entry is whether the variant will
alphabetize next to the primary form or not.

But there is an underlying problem with what you are doing. I assume that at
some point you will want to make use of the variant system to handle true
variants and irregularly inflected forms (and not just paradigm forms). The
variant system was not designed to handle regular paradigm forms. In the
Toolbox/MDF package there was a system for handling paradigm forms. FLEx has
not yet implemented it. Let me illustrate the two phenomena (paradigm forms
vs irregularly inflected forms) from English and then explain how the two
are different. First, English has a set of verb paradigm forms that need to
be specified (at least for language learners) because they are not always
predictable:

turn, turns, turning, turned, turned
stop, stops, stopping, stopped, stopped
try, tries, trying, tried, tried
sing, sings, singing, sang, sung
hop, hops, hopping, hopped, hopped
hope, hopes, hoping, hoped, hoped
hit, hits, hitting, hit, hit
go, goes, going, went, gone

English nouns likewise need to have the plural specified:

boy, boys
man, men
child, children
sheep, sheep

Many forms are rule based, but language learners often do not yet know the
rules, which are a bit complicated. So (for at least some dictionaries)
these forms need to be specified for each verb and noun. For other
dictionaries just the irregular forms need to be specified. Dictionaries for
native speakers perhaps could ignore the irregular forms, but most create
minor entries for them (probably because the publishers hope language
learners will also buy them).

Second, English verbs and nouns have inflected forms that are irregular,
cannot be predicted, and alphabetize far from the primary form. The examples
above contain several examples (sang, sung, went, gone, men). So some
paradigm forms are regular and some are irregular. The need with the
irregular forms is not just to list them in the entry for the primary form,
but also to create minor entries for them. The problem is that when a
language learner encounters one of these irregular forms, he cannot figure
out the primary form and therefore cannot find the entry in the dictionary.
There are other forms that are regularly formed, but would alphabetize far
from the primary form (e.g. tried). Most dictionaries create minor entries
for irregular, unpredictable forms that would be hard to find (e.g. sang),
but expect the user to learn the rules that would enable him to find the
correct primary entry for rule based forms (e.g. tried).

So we need a system that will allow us to list paradigm forms for each
grammatical category. (This would also be a nice grammar discovery tool.) We
also need a system that will permit us to mark some of these forms as
irregular and therefore needing a minor entry. The variant system was not
designed to handle this, although you can use it if you want. I would prefer
for FLEx to provide us with a paradigm field. Until it does, I use a custom
field for each paradigm form. So, for instance, I would create a custom
field for the English -s form, another field for the -ing form, another for
the past tense, and another for the part participle. I would use Bulk Edit
to (semi)automatically generate each form and then correct the irregular
forms by hand. I would do the same for the noun plural. I would also use the
variant system to create an entry for each irregular form that needs a minor
entry in the published dictionary. I realize that this is not ideal. One of
the biggest problems currently with FLEx is that it combines variants and
irregularly inflected forms in one system. The only thing in common between
the two is that they often need minor entries. But the two are fundamentally
different and require different handling.

To illustrate how messy all this gets currently in FLEx, let me explain what
I am having to do in my Greek dictionary. For paradigm forms I have a custom
field for the genitive form of nouns. I have another for the feminine and
neuter forms of adjectives. (I combine both forms in a single field.) I have
several custom fields for verb forms--one for future, one for aorist, one
for perfect, one for aorist passive, and one for future passive. I also put
all irregularly inflected forms in the Variants section and create a Variant
Type for each. I currently have 108 unique variant types for irregularly
inflected forms. (The morphology of Greek is rather complicated and there
are many potential forms.) I have the dictionary configured to display the
custom noun genitive form and the custom adjective feminine/neuter field.
But I can't display the custom fields for verb forms because it conflicts
with the Variants system. (All the irregular forms would appear twice.
<sigh>) Ah well, a lot of things work very well in FLEx.

I hope this clarifies the problem somewhat.

Version: 9.0.801 / Virus Database: 271.1.1/2824 - Release Date: 04/20/10
13:14:00

--
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

[Ronald Moe]

Andreas Joswig wrote:
"One thing confuses me:"

Don't feel bad. We are all confused by these things. They sometimes get so
complicated that I despair of ever knowing what is really going on in the
brain and how we can model it in a dictionary or parser. There have also
been some changes to FLEx recently, so I can't speak with any certainty. (In
order to write this email I had to test certain points in FLEx. <sigh>)
First I will speak to some general principles. Then I will talk about the
difference between the "Variants" section and the "Allomorphs" section. Then
I will give some "how to" instructions about how to handle various kinds of
forms.

There are three general principles that we need to keep in mind:
(1) What the parser needs and what the (published) dictionary needs are
often two different things. But we maintain everything in a single database.
(Result: occasional confusion, but it is far more efficient to maintain a
single database.)
(2) The way we capture information in the database and how it is presented
in a published dictionary are often very different. (Result: occasional
confusion, but it is far better to keep the information in a standardized
database and enable you to publish multiple dictionaries in a variety of
formats.)
(3) FLEx is an attempt to model linguistic behavior in a way that enables
you to investigate, analyze, and document a language's lexicon, morphology,
texts, etc. But obviously no one can completely model linguistic behavior
with a computer program. (Result: occasional limitations and inadequacies,
but FLEx is still an excellent tool for these purposes.)

Next an explanation of the "Variants" section and "Allomorphs" section":
The "Variants" section is where you create entries in the database for
variants, such as dialectal and spelling variants (isn't/ain't,
color/colour) and for irregularly inflected forms (were, went, ran). These
entries are generally formatted in a published dictionary as minor entries.
(Note that these entries are full entries in the FLEx database. I call them
"Variant" entries merely because they are for variants and irregularly
inflected forms, and are linked to the entry for the primary form. But as
far as the database is concerned, they are all just "entries".) Once you set
up a "Variant" entry, FLEx can do some nice things with it. For instance it
can use it to create a minor entry in your published dictionary. In addition
the parser will find the variant 'colour' and use the sense information from
the primary entry 'color'. You don't have to duplicate all the senses for
both entries. However if you want, you can also add one or more senses
(gloss, definition, example sentences, etc) for the variant 'colour' and the
parser will give you the option of using either the gloss(es) for 'color' or
the gloss(es) for 'colour'. It doesn't make much sense to do this for
variants, but it can be very useful for certain kinds of irregularly
inflected forms.

The "Allomorphs" section is where you enter allomorphs of a morpheme. You
can also enter "allomorphs" of a stem if you only want to parse down to the
stem level (e.g. re-gen-er-ate-d is parsed to the root level, regenerate-d
is parsed to the stem level). FLEx does not create an entry for an
allomorph. In order for the parser to deal with "errusthen", it must split
it into morphemes and label each morpheme. Since e-rrus-the-n contains an
allomorph "rrus" of the root "rus", I have to add "rrus" in the "Allomorphs"
section.

So the "Variants" section is used for both the published dictionary and the
parser. "Variant" entries are used to produce minor entries in the published
dictionary and the parser looks in them to find forms. On the other hand the
"Allomorphs" section is only for the parser. I don't know of any published
dictionary that includes allomorphs. (Although someone could include them.)

Next, how to handle XYZ:

How to handle an allomorph:
Enter the form of the allomorph in the Allomorph section of the Lexicon
Edit-Entry pane. Specify the conditioning environment in the Environments
field (e.g. / _ i). For instance for the English verb bend/bent you would
enter the primary form 'bend' in the Lexeme Form field and the allomorph
'ben' in the Allomorphs section.

How to handle a variant:
Enter the form of the variant in the Variants section of the Lexicon
Edit-Entry pane. Specify the type of variant it is in the Variant Type field
(e.g. spelling variant). For instance you would create an entry for primary
form "color". Then in the Variants section you would enter "colour" and
specify that it is a "British spelling variant". FLEx automatically creates
an entry for "colour" and links it to the primary entry "color". If you had
"colour" in your text corpus, the parser would find the form "colour" in the
entry for "colour" but use the sense information (grammatical category and
gloss) from the entry for "color".

How to handle an irregularly inflected form:
Enter the inflected form in the Variants section of the Lexicon Edit-Entry
pane. Specify the inflectional category in the Variant Type field (e.g. past
tense). For instance for feel/felt you would create a (main) entry for
"feel". Then in the Variants section you would enter "felt" and specify that
it is the "past tense" in the Variant Type field. (You would first have to
add "past tense" to the list of "Variant Types" in the Lists area.) FLEx
automatically creates an entry for "felt" and links it to the primary entry
"feel". Next you need to add an allomorph "fel" in the Allomorphs section
and specify the environment (/ _ t). If you had "felt" in your text corpus,
the parser would find the form "felt" in the entry for "felt" and would
suggest the analysis "feel + pst" (assuming that you gave "pst" as the
abbreviation for the Variant Type "past tense". The parser would also find
the allomorph "fel" of "feel" and the allomorph "-t" of "-ed", and would
suggest the analysis "feel -ed" with the appropriate glosses for each
morpheme. For the English word "felt" I would pick the second analysis.

How to handle suppletion:
Enter the inflected form in the Variants section of the Lexicon Edit-Entry
pane. Specify the inflectional category in the Variant Type field (e.g. past
tense). For instance for go/went you would create a (main) entry for "go".
Then in the Variants section you would enter "went" and specify that it is
the "past tense" in the Variant Type field. FLEx automatically creates an
entry for "went" and links it to the primary entry "go". (I would not add an
allomorph "wen" in the Allomorphs section, but you could. If you did, the
parser would offer two analyses as with "felt" above.) If you had "went" in
your text corpus, the parser would find the form "went" in the entry for
"went" and would suggest the analysis "go + pst". (The parser finds "went"
and sees that it is linked to "go". It finds that the Variant Type is "past
tense" and supplies the abbreviation "pst".)

How to handle a portmanteau morpheme (root + affix):
Enter the form of the portmanteau morpheme in the entry for the root in the
Variants section of the Lexicon Edit-Entry pane. Specify the inflectional
category in the Variant Type field (e.g. past tense). For instance you would
handle "were" under the main entry "be". You would enter "were" in the
Variants section and specify that it is the "past tense" in the Variant Type
field. (I don't know how to indicate that "were" must agree with "you(sg)"
"you(pl)" "we", or "they"). You would not enter "were" as an allomorph
because you cannot break "were" into morphemes and link one allomorph to
"be" and another to "pst". If you had "were" in your text corpus, the parser
would find the form "were" in the entry for "were" and would suggest the
analysis "be + pst". (The parser finds "were" and sees that it is linked to
"be". It finds that the Variant Type is "past tense" and supplies the
abbreviation "pst".) You would do the same for all the other forms of "be"
(was, are, am, is, etc). The parser uses the sense information from the main
entry for "be". It is best just to set up one primary entry for "be",
describe all the senses there, and create "Variant" entries for all the
other forms.

How to handle a portmanteau morpheme (affix + affix):
Enter the portmanteau morpheme as a main entry. Specify the combined meaning
in the gloss field. (Sorry, but I cannot illustrate this from English.) In
Greek the noun case suffixes are single morphemes (e.g. logo-i
'word-dative'). The plural is formed by suffixing the plural morpheme "-s"
after the case suffixes (e.g. logo-i-s 'word-dative-plural'). However the
genitive plural "-on" is a portmanteau morpheme. It cannot be split into two
morphemes, one meaning 'genitive' and the other meaning 'plural'. So it must
be entered as a main entry and given the gloss 'gen.pl'. (You might need to
set up a separate Affix Template in the Grammar-Category Edit area in order
to handle this.)

So if a language (such as Stephanie's) has a set of stem forms for a
particular verb, you have to ask several questions. (1) Can the parser
correctly identify the stem? If not, do I need to add allomorphs? (2) Can
the user of the published dictionary find the correct entry? If not, do I
need to add a minor entry (for errusthen) that will direct the user to the
main entry (ruomai)? (3) Is this a portmanteau morpheme, that is, does it
combine two morphemes into a single form that cannot reasonably be divided
(e.g. were, was)? If so, I need to create a "Variant" entry for the
portmanteau morpheme and may need to provide grammatical information and
semantic information for it.

This problem is also apparent in English verbs such as choose/chose and
run/ran (known as "vowel replacement" verbs or "ablaut"). A non-native
English speaker would not know where to find the past tense forms "chose"
and "ran". So you must create "Variant" entries for them and format them as
minor entries in the published dictionary. Since we cannot make up rules
that account (synchronically) for these forms, the parser has to look at the
"Variant" entries in order to find them.

[Note on go/went. Technically "wen" is not an allomorph of "go". It is not
derived from it historically and is not related to it by some morphophonemic
rule, that is, it is not related to "go" phonologically and you cannot
specify a phonological environment that governs its use. You can make up an
"ad hoc" rule, but that is a different matter. Historically "went" was the
past tense of "wend", but today "wended" is the past tense of "wend" and
"went" is the past tense of "go". So in cases of suppletion such as this you
have to somehow get the parser to interpret "went" as "go" + "Pst". Notice
that it is not "wen" + "-t", or even "wend" + "-ed". This is why I recommend
handling "went" as a "Variant" (i.e. an irregularly inflected form) rather
than an allomorph. Morphology gets really complicated sometimes. Isn't it
fun?]

Ron Moe

-----Original Message-----
From: flex...@googlegroups.com [mailto:flex...@googlegroups.com] On

No virus found in this incoming message.
Checked by AVG - www.avg.com

Version: 9.0.801 / Virus Database: 271.1.1/2814 - Release Date: 04/15/10
23:31:00

[Michael Galant]

This thread is extremely relevant and useful for me.

 I have a couple of follow-up questions (this is for some colleagues who are working on a variety of Mixtec), related to what to give as lexeme vs. citation form and how to make the right information visible during interlinear analysis.

I’ll use an example from (a hypohetical variety of) Zapotec, since the same type of issue arises.  Let’s say that there’s a verb that has the habitual stem ruun, the perfective stem been, and the irrealis stem guun.  Let’s also assume that the citation form is given as ruun. I want to make sure that ruun, been and guun are visible in the dictionary output/view and that they are visible to the parser  in interlinear analysis.

 

Here are some possibilities:

 

• Although you suggest otherwise in this thread, I wonder if there's any merit to this possibility: create separate lexical entries for ruun, been, and guun (treating each as a lexeme and a citation form), and in each entry, give the gloss as asp-root or asp.root (depending on beliefs about segmentability), e.g., ruun should be hab-do (or hab.do if we weren’t sure about the segmentability).  Would FLEx be able to see the aspectual information in the glosses when parsing?

• create just one lexical entry, as you suggest, with the lexeme being an abstract root -uun, make ruun the citation form, and make been and guun as variants, as per your suggestions - concerns:

• how would the aspectual information for ruun (which is given only as the citation form, not a variant) be visible for the purposes of interlinear analysis? [ If I understand this thread correctly, the aspectual information for been and guun would be visible for the purposes of interlinear analysis by labelling them as the appropriate variant types under irregularly inflected forms (perfective and irrealis) - Is that correct?]

• Would it be a good idea to also create custom fields for the perfective and irrealis stems, just for the sake of the lexicographer (not for the dictionary output, not for the parsing), i.e., to be able to eyeball the "principal parts"in a convenient way?  I assume that if this were done, the parser would have no way to see this information, but if so, how would that be done, and could that prevent the need for listing these forms as variants?

• as a similar alternative to #2 above, create just one lexical entry, as you suggest, but enter as the lexeme (and citation form) the habitual stem ruun, and been and guun as variants.  In this case, how would the aspectual information for ruun be made visible for the purposes of interlinear analysis?  Could this work by making the gloss hab.do (or hab-do)?  In this case, there would be no lexeme in the entire dictionary consisting merely of 'do'.  Would that be a problem?

Thank you in advance for any feedback you may have.

Sincerely,
Mike

[Jeff Shrum]

Michael,

 

The abstract form field is useful for configuring one’s dictionary, but for the parser to work correctly I suggest you configure “stem names” for the perfective, and irrealis tense-aspect.  You can read about this feature of Fieldworks in the help.  I have pasted one relevant section below but there is more information in Help.

 

Insert a Stem Name

Stem Names control stem allomorphy that is dependent not on phonological issues, but on the presence of certain inflection features. Make sure the necessary inflection features are available before you do the steps in this topic.

1. In the Navigation Pane, click Grammar and then click Category Edit.
2. In the center column, select the category.

Subcategories will inherit any Stem Names which their parent categories have, so you will probably want to insert the Stem Name at the highest level category.

3. In the Category (or Part of Speech) pane, click the Stem Names field label, and then do one of the following:
• Click the Insert Stem Name link that appears.
• Click the menu button that appears, and then select Insert Stem Name.

Stem Name, Abbreviation, Description and Feature Set fields appear.

4. Enter a name, abbreviation and description for the Stem Name.
5. In the Feature Set field, choose a set of inflection features that are relevant to this Stem Name.
6. If you need more than one Feature Set for this Stem Name, insert a Feature Set for each one.

Important

• Stem Names may only be used for stem allomorphs, not affix allomorphs.
• When a particular allomorph of a stem is tagged as belonging to a Stem Name, that allomorph is only valid when the word it is in contains certain inflectional features. The main way that the word can have those certain features is if one or more inflectional affixes have those features. (It is also possible for one or more derivational affixes in the word to add inflectional features to the word.)

A "matching" process happens: when a particular stem name is used with a given stem allomorph, it means that this allomorph is good only when there are one or more inflectional affixes in the word with at least those inflectional features. (Some or all of the required features can also be from the To Inflection Features field of a derivational affix.) Therefore, for each feature in each Feature Set, make sure there is at least one inflectional affix entry with that feature in its Inflection Features field (Lexicon Edit) or at least one derivational affix entry with that feature in its To Inflectional Features field. Choose inflection features as necessary.

• If a computational parser gives invalid analyses, do the following:

·          

• Review your affix template tables. Make sure that you have at least one inflectional template defined for each category that uses Stem Names.
• Review your inflectional features used with Stem Names. Be aware that you need to do three things for valid parses:

o     

• Insert feature sets for Stem Names.
• Choose Stem Names for stem allomorphs.
• Choose inflection features for inflectional affixes.
• Category Edit overview has a discussion about category hierarchy.
• For more information and examples, point to Resources on the Help menu, and then click Introduction to Parsing.

Jeff Shrum

SIL International

Language Technology Consultant

Dallas, TX USA

--
You are subscribed to the publicly accessible group "FLEx list".
Only members can post but anyone can view messages on the website.
To change your status, please write to flex_d...@sil.org.
You can join this group by going to 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 post to this group, send email to flex...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/flex-list/b17e8c9a-c2f9-4aa6-bc88-435b657a2558%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Michael Galant]

Thanks so much for this helpful information, Jeff.

 

I’ll give this a try.

 

Best,

Mike

 

From: flex...@googlegroups.com [mailto:flex...@googlegroups.com] On Behalf Of Jeff Shrum

Sent: Friday, March 31, 2017 6:13 AM
To: flex...@googlegroups.com

• Insert feature sets for Stem Names.

You received this message because you are subscribed to a topic in the Google Groups "FLEx list" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/flex-list/UZ0mgMRM-CA/unsubscribe.
To unsubscribe from this group and all its topics, send an email to flex-list+...@googlegroups.com.

To post to this group, send email to flex...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/flex-list/99e18f271128fb2cd4e80b5c99c5cdb0%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.

Virus-free. www.avast.com