How do you handle idiomatic expressions in FLEx?

[Kevin Warfel]

My question seems so fundamental that there must be lots of FLEx users out there for whom the answer is obvious.  I am trying to understand how to handle an idiomatic expression in FLEx, and have not been able to find much written on the topic in any of the help files.  If someone for whom this is a “no-brainer” has the time to write a few lines and enlighten me, I’d greatly appreciate it.

 

Here is one example of what I’m dealing with:

ʋ̀    lɔ     wuló

he  falls   yell(n)

He lets out a yell.

 

The verb ‘fall’ is used in conjunction with the word ‘a yell’ and I want to record the idiomatic expression lɔ wuló “fall a yell” in my lexicon.

 

Here’s what I know how to do already:

1) link the two words in the Text pane of the Texts & Words view, to form a phrase

2) analyze and gloss the words that make up the phrase

 

What I don’t know is the following:

1) Whether to treat this as an expression in the lexicon, and if so, how.  Some possibilities:

   a) list ‘emit’ as a second(ary) sense of the verb lɔ in my lexicon?

   b) list lɔ wuló as some kind of sub-entry of the primary sense of the verb lɔ?

   c) list lɔ wuló as a separate entry in the lexicon?

My intuition tells me I should do option “b” above, but I don’t know exactly how to do that.

 

If anyone can point me to something in the help files on this topic, I would appreciate that as well.  I had expected to find it in the Introduction to Lexicography document by Ron Moe, but didn’t see anything in the Table of Contents that looked like it dealt with idioms or expressions consisting of multiple words, nor did a search for “idiom” lead me to anything like what I had expected to find.

 

Thanks,

Kevin Warfel

[Ronald Moe]

Kevin Warfel wrote:

“I had expected to find it in the Introduction to Lexicography document by Ron Moe…”

 

Sorry, I should have used the word ‘idiom’ somewhere in my paper. I call such things ‘lexical phrases’, since multi-word lexical items fall under a number of different categories including idioms, sayings, phrasal verbs, and a number of other terms that lexicographers employ. A common general term that I’ve seen in the literature recently is MWE (multi-word expression), but not all MWEs are lexical nor would you put them all in a dictionary. Idiomatic expressions are treated in my paper under the following sections:

3.2.4 Complex forms

3.4 Types of entries in a published dictionary

 

Unfortunately I can’t find an explanation in my paper of how to enter lexical phrases in FLEx. The way FLEx handled them changed between version 5 and 6. Apparently I didn’t explain the change anywhere. So here it is:

 

To handle a lexical phrase, such as an idiom, do the following:

1. Create a new entry and enter the form of the lexical phrase in the ‘Lexeme Form’ field.
2. In the ‘Complex Form Type’ field indicate what type of complex form it is (idiom, phrasal verb, etc.). You can add other types of complex forms to this list in the Lists-Complex Form Types area.
3. In the ‘Components’ field link the lexical phrase to each root. If you have not already created an entry for each root, FLEx will offer to create the entry for you. Alternatively you can link a lexical phrase to one or more of the words that it is composed of. For instance ‘peace treaty’ is composed of peace + treat-y. So you could link ‘peace treaty’ to ‘peace’ and ‘treat’ (the two roots) or to ‘peace’ and ‘treaty’. Unfortunately (as far as I know) FLEx does not currently permit hierarchical linking. So you cannot link ‘peace treaty’ to ‘treaty’ and then link ‘treaty’ to ‘treat’. If you attempt to do this, FLEx will not correctly export your data for publication.
4. In the ‘Show Subentry under’ field specify which root (or roots) you want the lexical phrase to appear under. This only applies in a root-based dictionary. In a stem-based dictionary all complex forms are presented as main entries. But in a root-based dictionary complex forms are presented under one or more of the roots. This field enables you to specify which root(s). So if you are producing a stem-based dictionary you can ignore this field.
5. In the ‘Show Minor Entry’ field put a check in the check box if you want FLEx to generate a minor entry for the lexical phrase. This only applies in a root-based dictionary. For instance if you choose to display ‘peace treaty’ under ‘treat’, the user would not find it if he looked under ‘peace’. So FLEx can create a minor entry ‘peace treaty’ that refers the user to the entry ‘treat’ where he will find the full subentry for ‘peace treaty’.

 

If you are producing a stem-based dictionary, it is not necessary to do steps 2-5. However I would highly recommend that you at least indicate the Complex Form Type and indicate the Components. This enables you to do research on all the complex forms based on a single root. (I have found it very helpful to investigate all the complex forms based on a single root at one time. The meanings of each complex form are often related to each other in systematic ways and I can gain insights into their meanings using this technique.)

 

By the way, it is not necessary to add secondary senses to a word in order to handle the meanings of idioms. For instance the meaning of your idiom “fall a yell” would be described in its own entry. Unless “fall” is used elsewhere with the meaning ‘emit’, you would not give ‘emit’ as a secondary meaning of “fall”. It would only have this meaning in the idiom “fall a yell”. Actually it would be more correct to say that it doesn’t have this meaning at all. Instead it would be more proper to say the entire expression “fall a yell” means ‘to yell’ or ‘to let out a yell’. An idiom is a multi-word expression, the meaning of which is not the sum of the meanings of each word. So the individual words in an idiom do not have separate meanings. Instead the entire idiom has a single meaning which can often be quite different from the meanings of the component words as they are used elsewhere.

 

There is more that could be said on this topic, but it’s long enough already.

Ron Moe

 

---

From: flex...@googlegroups.com [mailto:flex...@googlegroups.com] On Behalf Of Kevin Warfel
Sent: Monday, January 04, 2010 9:53 AM
To: flex...@googlegroups.com
Subject: [FLEx] How do you handle idiomatic expressions in FLEx?

 

My question seems so fundamental that there must be lots of FLEx users out there for whom the answer is obvious.  I am trying to understand how to handle an idiomatic expression in FLEx, and have not been able to find much written on the topic in any of the help files.  If someone for whom this is a “no-brainer” has the time to write a few lines and enlighten me, I’d greatly appreciate it.

 

Here is one example of what I’m dealing with:

?Ì    l?     wuló

he  falls   yell(n)

He lets out a yell.

 

The verb ‘fall’ is used in conjunction with the word ‘a yell’ and I want to record the idiomatic expression l? wuló “fall a yell” in my lexicon.

 

Here’s what I know how to do already:

1) link the two words in the Text pane of the Texts & Words view, to form a phrase

2) analyze and gloss the words that make up the phrase

 

What I don’t know is the following:

1) Whether to treat this as an expression in the lexicon, and if so, how.  Some possibilities:

   a) list ‘emit’ as a second(ary) sense of the verb l? in my lexicon?

   b) list l? wuló as some kind of sub-entry of the primary sense of the verb l??

   c) list l? wuló as a separate entry in the lexicon?

My intuition tells me I should do option “b” above, but I don’t know exactly how to do that.

 

If anyone can point me to something in the help files on this topic, I would appreciate that as well.  I had expected to find it in the Introduction to Lexicography document by Ron Moe, but didn’t see anything in the Table of Contents that looked like it dealt with idioms or expressions consisting of multiple words, nor did a search for “idiom” lead me to anything like what I had expected to find.

 

Thanks,

Kevin Warfel

--
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: 9.0.725 / Virus Database: 270.14.124/2598 - Release Date: 01/04/10 01:24:00

[Kevin Warfel]

Thank you, Ron.  That was very clear and helpful.  I understand your reasoning for adopting the term ‘lexical phrase’, but didn’t think of searching for those particular words when I was looking for help with what I wanted to do.  I think I’ve now got this entry set up correctly and will now know how to handle the others that I will inevitably encounter.

 

Kevin Warfel

[Kevin Warfel]

Now that I’ve gotten a satisfactory answer to my initial question, I have a follow-up on the handling of lexical phrases in FLEx, this time on the glossing/parsing of them when one of the elements of the phrase is inflected.

 

Now that I have an entry for lɔ wuló, how do I get the word glossing/parsing tool to correctly analyze and gloss the phrase when the verb is conjugated with a different tense or aspect than the one in the lexical entry?  More precisely, I have the past tense form of the phrase in one of my texts.  This past tense form is lɔɔ́ wuló.  My lexical entries allow for the correct parsing of this verb form (lɔɔ́) in isolation, but not when it is combined with a second word.  Likewise, the second word (wuló) is recognized by the parser when it occurs in isolation, but not when it is part of a phrase.  Is there a way for me to get the Morphemes line to show the verb root lɔ, plus the past tense suffix -ɔ́, or do I just have to settle for showing the “pastness” of the phrase in the Word Gloss line?

 

 

Kevin Warfel

 

From: flex...@googlegroups.com [mailto:flex...@googlegroups.com] On Behalf Of Ronald Moe

Sent: Monday, January 04, 2010 2:34 PM
To: flex...@googlegroups.com

ʋ̀    lɔ     wuló

he  falls   yell(n)

He lets out a yell.

 

The verb ‘fall’ is used in conjunction with the word ‘a yell’ and I want to record the idiomatic expression lɔ wuló “fall a yell” in my lexicon.

 

Here’s what I know how to do already:

1) link the two words in the Text pane of the Texts & Words view, to form a phrase

2) analyze and gloss the words that make up the phrase

 

What I don’t know is the following:

1) Whether to treat this as an expression in the lexicon, and if so, how.  Some possibilities:

   a) list ‘emit’ as a second(ary) sense of the verb lɔ in my lexicon?

   b) list lɔ wuló as some kind of sub-entry of the primary sense of the verb lɔ?

   c) list lɔ wuló as a separate entry in the lexicon?

[Heidi J. Rosendall]

Hi,

One of our projects came to me with this problem.

 

They began interlinearizing with a phonetic writing system before they had decided upon their orthographic characters and now they would like to interlinearize these texts orthographically. These seems to me to be a reasonable progression that should be supported.

I had them use Tools/Configure/Interlinear to add lines for the orthographic writing system for their Words, Morphemes, and Lexical Entries, believing that they could then assign and choose orthographic information by writing in the Orthographic word and then breaking it into morphemes. This seemed to work all the way up to the point of accessing the words (which already exist) in the lexicon but it fails at this point. The default writing system at this point is now orthographic rather than phonetic.

 

Is this a bug or am I doing something wrong? I know that there are warnings given about switching between writing systems and that there are limitations but it almost worked.

 

Thanks for your help,

 

 

Heidi J. Rosendall

Language Software and Typesetting Manager

Wycliffe Nigeria Group

Heidi_R...@wycliffe.org

Skype: heidi.james.rosendall

 

 

[Andy Black]

On 1/4/2010 2:22 PM, Kevin Warfel wrote:

Now that I’ve gotten a satisfactory answer to my initial question, I have a follow-up on the handling of lexical phrases in FLEx, this time on the glossing/parsing of them when one of the elements of the phrase is inflected.

 

Now that I have an entry for lɔ wuló, how do I get the word glossing/parsing tool to correctly analyze and gloss the phrase when the verb is conjugated with a different tense or aspect than the one in the lexical entry?  More precisely, I have the past tense form of the phrase in one of my texts.  This past tense form is lɔɔ́ wuló.  My lexical entries allow for the correct parsing of this verb form (lɔɔ́) in isolation, but not when it is combined with a second word.  Likewise, the second word (wuló) is recognized by the parser when it occurs in isolation, but not when it is part of a phrase.  Is there a way for me to get the Morphemes line to show the verb root lɔ, plus the past tense suffix -ɔ́, or do I just have to settle for showing the “pastness” of the phrase in the Word Gloss line?

This is a known issue for the parser.  Section 4.2 "Morpheme Types" includes this paragraph:

"One may also tag an allomorph as a phrase, although there are limitations. The FieldWorks Language Explorer parser treats such phrases like a stem. That is, any derivational or inflectional affixes must attach around the phrase, not within it (unless such an affix is an infix, of course). In particular, the FieldWorks Language Explorer parser cannot handle something like kicked the bucket where the past tense inflection occurs as a suffix on the first word of the phrase."

So, unless you want to add special infix allomorphs just for (exceptional?) cases like this, the parser is not going to be able to help you.  Unless you have a very large number of these phrasal idioms and unless you also have a lot of inflection, I would not recommend trying to make the parser deal with these phrases. 

Note that within the interlinear tool you can create variant entries for these inflected phrasal idioms.  You could either label these variants as "Irregularly Inflected Form" / "Past" or, since these are not actually irregular (right?), you could make a new variant type (in the Lists area) for, say, "Inflected Phrasal Idiom" / "Past".

To be more specific and assuming you have an entry already for lɔ wuló, do this in the interlinear:

1. Join lɔɔ́ wuló into a phrase by clicking on the link icon.
2. In the Lex. Entries line, click on the drop down button and choose "Variant of..."
3. Type lɔ wuló in the Variant of field (i.e. find the main idiom entry).
4. Select it in the Lexical Entries pane.
5. In the Variant Type drop down box, choose the appropriate variant type (e.g. "Inflected Phrasal Idiom" / "Past").
6. Click the "Add Variant" button.

The result in the interlinear may not be ideal because it will show something like the following screen shot where the inflection is appended to the end, rather than on the verb:



Unless I'm mistaken (and I'd be happy to find out I am), this is the best FLEx currently has to offer.

--Andy

--

[Kevin Warfel]

Hi Andy,

 

Good to hear from you.  I had read section 4.2 in the Introduction to Parsing before sending out my email, and in light of that, I fully expected to receive a response back saying that this simply cannot be done.  I am very pleasantly surprised to report that not only can it be done, but there are at least two ways to do it!

 

Your proposal is an interesting alternative to one that Marlon Hovland pointed me to yesterday.  (His message didn’t go through the FLEx list, but maybe this information would be of interest to some who subscribe to this list, so I’m including the details he sent me here for the benefit of everyone.)

 

Marlon directed me to the “going for broke” example near the bottom of “Insert or remove morpheme breaks” (Using_Tools/Texts_&_Words_tools/Interlinear_Texts/Insert_or_remove_morpheme_breaks.htm) in the Help files.  I found this page very quickly (after receiving his hint) by clicking on Help, then Language Explorer, then the Search tab.  I typed “going for broke” into the box labeled “Type in the word(s) to search for:” and there is only one page that comes up in response to that – the one Marlon was referring to.  (Prior to receiving Marlon’s email, I could find nothing in the Helps that indicated how I could do what I was wanting to do, as I did not realize that I needed to “insert morpheme breaks” in order to accomplish it, so didn’t think to search for that topic, and therefore didn’t find the one page in all the Help files that contained the answer to my question.)

 

There it is pointed out that one can use the Insert Morpheme Breaks option to cut the inflectional suffix from its normal position and paste it at the end, much as your proposal indicates needs to be done, inflecting, as it were, the entire phrase, rather than just the verb.  Rerunning the parser after having done that led to a successful parse – to my great surprise, frankly, because of all the morphophonemics that go on in Puguli verb inflection, and with the conditioning environment now being separated from the suffix by an intervening word.  Here is the screen shot of what I got after following the instructions Marlon pointed me to.

 

 

Since my glosses are in French, here are a few hints to facilitate understanding. 

tomber = fall

cri = yell

ACC = perfective

a poussé un cri très fort = let out a yell

 

I like this presentation on the Morphemes line (line 2) in spite of the displaced morpheme because it actually shows me the surface form of that morpheme, while the Word line (line 1) shows me where the suffix is actually appended, and the Lex. Entries line (line 3) shows me the underlying form of the past tense (perfective aspect) morpheme.  The explanation about why the suffix has to be displaced to the end of the phrase is sufficient to overcome my initial dislike of the perturbation of the Puguli morpheme/word order, and I am happy to have access to all of the information that would normally appear when interlinearizing a single word (e.g. this particular verb when used to literally mean “fall”).  I think the FLEx team has done a very good job of providing us with a way to handle this particular complexity of language.

 

Blessings!

Kevin

[Robert Hedinger]

Are there plans to improve FLEx capability on this? What I would like to see is that FLEx can analyse each individual part of MWEs (multi word expressions) for their internal structure (root, affixes, etc.) while also being able have these MWEs as lexical entries.

We have many phrasal verbs with a meaning that cannot be derived from the parts, but where the morphology is very transparent. The same is true for derivatives. Some are semantically and morphologically transparent. Others are not transparent from a semantic point of view, but the morphology is identical to that of the first group. I want to assign one meaning to the word as a whole, but would like FLEx to be able to parse it like the ones with a separate meaning associated with each morpheme.

 

I hope this can be done in future enhancements. Or is this impossible to program?

 

Robert

From: Andy Black

Sent: Tuesday, January 05, 2010 4:25 PM

To: flex...@googlegroups.com

Subject: Re: [FLEx] How do you handle the inflection of idiomatic expressions in FLEx?

[Andy Black]

On 1/5/2010 9:02 AM, Kevin Warfel wrote:

... one can use the Insert Morpheme Breaks option to cut the inflectional suffix from its normal position and paste it at the end, much as your proposal indicates needs to be done, inflecting, as it were, the entire phrase, rather than just the verb.  Rerunning the parser after having done that led to a successful parse – to my great surprise

My surprise, too.  I suspect that the automated morphological parser actually did *not* succeed.  Try right clicking on the word in the interlinear and choosing "Show in Word Analyses."  When you get there, look at the analysis given and I fully expect that the Parse result field will say "Failure."

Another way to test this is to key the phrase in Try a Word and parse it.  Again, I fully expect that you'll find that it does not parse.

What I think you have run across is that the *manual* parsing process (which merely looks morphemes up in the lexicon) has given you this analysis.

Nonetheless, it is nice to know of this other way of dealing with the issue.

--Andy

[Andy Black]

On 1/5/2010 9:07 AM, Robert Hedinger wrote:

Are there plans to improve FLEx capability on this? What I would like to see is that FLEx can analyse each individual part of MWEs (multi word expressions) for their internal structure (root, affixes, etc.) while also being able have these MWEs as lexical entries.

We have many phrasal verbs with a meaning that cannot be derived from the parts, but where the morphology is very transparent. The same is true for derivatives. Some are semantically and morphologically transparent. Others are not transparent from a semantic point of view, but the morphology is identical to that of the first group. I want to assign one meaning to the word as a whole, but would like FLEx to be able to parse it like the ones with a separate meaning associated with each morpheme.

 

I hope this can be done in future enhancements. Or is this impossible to program?

I doubt that it is impossible to program, but it is certainly not immediately clear how to do it efficiently in the general case (to me, anyway).

We would like to be able to offer such a capability, but it may take a while before it gets done...

--Andy

[Ronald Moe]

Robert Hedinger wrote:

“I hope this can be done in future enhancements. Or is this impossible to program?”

 

Andy will have to answer whether or not it is possible to program the parsing of inflected words within an idiom. What we have to do is tell him what we need and how we want the program to work. Just to stimulate the discussion, let me give an English example ‘It is raining cats and dogs.’ I’ve glossed the words in Japanese, just for fun. First I’ve merely analyzed each word. But because of the idiomatic nature of this sentence, there are some major problems. ‘It’ is a dummy subject. There is not a good match between the individual words and the meaning as expressed in the free translation.

 

It rained cats and dogs last night.

it     rain-ed          cat-s      and dog-s  last  night

sore ame.furu-Pst neko-Pl to    inu-Pl saku ban

Saku ya ame ga hageshiku futta. [lit. Last night rain hard fell.]

 

I would say the idiom is underlyingly ‘it rain cats and dogs’, but I would give the citation form as ‘it is raining cats and dogs’. Because ‘cats’ and ‘dogs’ are plural, we can’t just pull off all the inflectional affixes:

 

It rained cats and dogs last night.

it rain cat and dog        -ed  -s   -s  last  night

ame.ga.hageshiku.furu -Pst -Pl -Pl saku ban

Saku ya ame ga hageshiku futta. [lit. Last night rain hard fell.]

 

Instead we need to pull off just the inflectional affixes which are not part of the underlying form of the idiom. Leaving the inflection affix “in place” doesn’t seem like a good option because there is no way to line up the affix and its gloss:

 

It rained cats and dogs last night.

it rain-ed cats and dogs      last  night

ame.ga.hageshiku.furu-Pst saku ban

Saku ya ame ga hageshiku futta. [lit. Last night rain Sbj hard fell.]

 

So putting the inflectional affix after the idiom seems better. But this displaces the affix from the stem it occurs on, requiring the reader to do a bit of analysis to understand the display:

 

It rained cats and dogs last night.

it rain cats and dogs     -ed  last  night

ame.ga.hageshiku.furu -Pst saku ban

Saku ya ame ga hageshiku futta. [lit. Last night rain Sbj hard fell.]

 

This means that the user has to manually separate the inflectional affixes from the idiom, leaving something that the parser can find in the lexicon. In the example above pulling the –ed suffix off leaves the lexeme form. But there are still problems here. The lexeme form contains affixed forms. So the text has not been fully analyzed. That may be OK. But there are idioms whose structure is less frozen. Idioms range from fully frozen forms (e.g. ‘of course’, ‘to and fro’) to very fluid forms with lots of variants. For instance ‘hang out one’s dirty laundry’ occurs in many forms: ‘hang your wash out to dry’, ‘let everyone see your dirty linen’, etc. This “idiom” is really more a scene that can be expressed in many ways, each of which has the same basic meaning, ‘to speak or act in such a way that you reveal your errors, faults, problems, etc. to the public’.

 

Getting a parser to handle and display frozen forms seems fairly trivial. Getting it to handle a form like ‘it’s raining cats and dogs’ is rather more complicated, because you have two inserted elements ‘-‘s’ and ‘-ing’. But it still seems to me to be doable. Getting a parser to handle the ‘dirty laundry’ type idiom is probably not doable because there is no underlying lexeme form that appears in all manifestations of the idiom. It would probably be easier to parse each word as straight text and add an explanatory comment to the free translation.

 

So how would we handle the idiom ‘(be) on/at sb’s heels’ as in the following (examples from Longman Language Activator)?

 

The rebels headed for the border but government troops were still at their heels.

Just 15 minutes into the race Lawson was already hot on the champion’s heels.

He rushed out of the theatre with a pack of reporters at his heels.

 

Here we have an idiom with inflected, optional and replacement parts. It can also have other words inserted in the middle of the idiom. I have a hard time imagining a display that a reader could make sense of.

 

Ron Moe

 

 

---

From: flex...@googlegroups.com [mailto:flex...@googlegroups.com] On Behalf Of Robert Hedinger
Sent: Tuesday, January 05, 2010 9:07 AM
To: flex...@googlegroups.com
Subject: Re: [FLEx] How do you handle the inflection of idiomatic expressions in FLEx?

 

Are there plans to improve FLEx capability on this? What I would like to see is that FLEx can analyse each individual part of MWEs (multi word expressions) for their internal structure (root, affixes, etc.) while also being able have these MWEs as lexical entries.

We have many phrasal verbs with a meaning that cannot be derived from the parts, but where the morphology is very transparent. The same is true for derivatives. Some are semantically and morphologically transparent. Others are not transparent from a semantic point of view, but the morphology is identical to that of the first group. I want to assign one meaning to the word as a whole, but would like FLEx to be able to parse it like the ones with a separate meaning associated with each morpheme.

 

I hope this can be done in future enhancements. Or is this impossible to program?

 

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

No virus found in this incoming message.

Checked by AVG - www.avg.com

Version: 9.0.725 / Virus Database: 270.14.126/2601 - Release Date: 01/05/10 00:35:00

[Beth]

On Jan 4, 2010, at 8:52 AM, Kevin Warfel wrote:

> Here is one example of what I’m dealing with:
> ʋ̀ lɔ wuló
> he falls yell(n)
> He lets out a yell.
>
> The verb ‘fall’ is used in conjunction with the word ‘a
> yell’ and I want to record the idiomatic expression lɔ wuló
> “fall a yell” in my lexicon.
>

Kevin, the other thing I wonder about this is how productive it is.

In Urdu, they have a lot of "light verbs". These are verbs that do
have their own semantic meaning, but in a lot of cases are used with
bleached semantics to indicate something about aspect or nature of
the action, rather than anything about their original semantics.
They are like auxiliaries.

For instance, there is a verb that means something like "hit" or
"strike", but when used with another verb, it often means, "did it
immediately" or in a particular manner. The specific meaning often
varies depending on what verb it's used with, but it's highly
productive in that many verbs can co-occur with it, and there are
similarities in the range of meaning it can express. In that, sense,
this doesn't feel like an idiom or a frozen form.

If your lɔ is anything like that, then I might go ahead and just make
lɔ its own entry, with explanations of the range of meanings it can
have, and lots of links to complex forms built from it. But those
complex forms likely wouldn't be inflected.

I don't know if that really helps; it's just that there is a whole
continuum from auxiliary to light verb to idiom, and each might be
treated somewhat differently.

-Beth

[Allan Johnson]

Ronald Moe wrote:

To handle a lexical phrase, such as an idiom, do the following:

1. Create a new entry and enter the form of the lexical phrase in the �Lexeme Form� field.
2. In the �Complex Form Type� field indicate what type of complex form it is (idiom, phrasal verb, etc.). You can add other types of complex forms to this list in the Lists-Complex Form Types area.
3. In the �Components� field link the lexical phrase to each root. If you have not already created an entry for each root, FLEx will offer to create the entry for you. Alternatively you can link a lexical phrase to one or more of the words that it is composed of. For instance �peace treaty� is composed of peace + treat-y. So you could link �peace treaty� to �peace� and �treat� (the two roots) or to �peace� and �treaty�. Unfortunately (as far as I know) FLEx does not currently permit hierarchical linking. So you cannot link �peace treaty� to �treaty� and then link �treaty� to �treat�. If you attempt to do this, FLEx will not correctly export your data for publication.

Hi Ron,

The limitation mentioned here about hierarchical linking - you're talking about root-based dictionaries here, right? It's been a while since I've looked at this detail, but my recollection is that chains of links such as �peace treaty� to �treaty� and �treaty� to �treat� do work fine in the database, and display correctly in a stem-based format, but a root-based format has problems with it. The root-based format would have to offer multiple levels of subentries in order to handle it. Is this the same issue that you have in mind? It's one of my pet issues, since we can't yet achieve the needed format for some of our Philippine dictionaries. I hope it can be resolved before too long.

Allan

[Allan Johnson]

Heidi J. Rosendall wrote:

Hi,

One of our projects came to me with this problem.

 

They began interlinearizing with a phonetic writing system before they had decided upon their orthographic characters and now they would like to interlinearize these texts orthographically. These seems to me to be a reasonable progression that should be supported.

I had them use Tools/Configure/Interlinear to add lines for the orthographic writing system for their Words, Morphemes, and Lexical Entries, believing that they could then assign and choose orthographic information by writing in the Orthographic word and then breaking it into morphemes. This seemed to work all the way up to the point of accessing the words (which already exist) in the lexicon but it fails at this point. The default writing system at this point is now orthographic rather than phonetic.

 

Is this a bug or am I doing something wrong? I know that there are warnings given about switching between writing systems and that there are limitations but it almost worked.

Hi Heidi,

I think there might be an easy fix. If they go to the lexicon and adjust the display to show the Lexeme Form in both writing systems, what do they see? I'm guessing that there will be a lexeme form in the phonetic writing system, but nothing in the orthographic writing system. They should be able to copy the phonetic lexeme forms to orthographic lexeme forms using Bulk Edit. Once this is done, I think the interlinearizer will be able to find the lexemes.

Allan J.

[Heidi J. Rosendall]

Thanks, I’ll try it and pass it on.

 

From: flex...@googlegroups.com [mailto:flex...@googlegroups.com] On Behalf Of Allan Johnson

Sent: Wednesday, January 06, 2010 12:27 PM
To: flex...@googlegroups.com

[Kevin Warfel]

You are right, Andy.  The Word Analysis showed a failure by the parser and the Try a Word parser option failed to make sense of it as well. 

 

I can understand how the “manual parser” could find ‘fall’ and ‘yell’ in the lexicon, so your explanation is transparent for those elements, but I don’t understand how it was able to correctly analyze the verbal suffix without help from the parser, which looks at conditioning environments, since this surface form of this suffix assimilates in several ways to the vowel of the verb root to which it is suffixed.  Are you suggesting that this “manual parser” simply looks through the lexicon for a match to any Lexeme Form or any Allomorph, without regard to the Environment(s) in which they are stated to occur?

 

At the end of the day, one thing still stands true, however, and that is that I like the result that FLEx gave me.  The way it did that is apparently different than what I originally thought, but still helpful.

 

Thanks for your input as I continue my efforts to broaden my understanding of how FLEx works.

 

Kevin Warfel

 

From: flex...@googlegroups.com [mailto:flex...@googlegroups.com] On Behalf Of Andy Black

Sent: Tuesday, January 05, 2010 4:53 PM
To: flex...@googlegroups.com

Subject: Re: [FLEx] How do you handle the inflection of idiomatic expressions in FLEx?

 

On 1/5/2010 9:02 AM, Kevin Warfel wrote:

[Kevin Warfel]

Beth,

Thanks for your input. I've never heard the term "light verb" before, but Puguli also has particles that trace their origins to specific verbs, but which now function more as auxiliaries to a main verb. These auxiliaries cannot take the full range of inflectional or derivational suffixes that a normal verb can (certain ones are completely immutable) and their semantic content is difficult to define precisely, so they sound a bit like Urdu's light verbs. I would certainly agree with you that these should be handled differently than idioms and other MWE's.

The expression I'm dealing with here has nothing to do with this kind of verb. The verb 'fall' is not only a fully-productive verb, taking all of the normal suffixes that a Puguli verb can take, but it is extremely *regular* (= predictable) in the way that it forms those various inflected and derived forms. I don't have any hesitation about labeling this expression an idiom.

Kevin Warfel

-----Original Message-----
From: flex...@googlegroups.com [mailto:flex...@googlegroups.com] On Behalf Of Beth
Sent: Tuesday, January 05, 2010 6:52 PM
To: flex...@googlegroups.com

[Andy Black]

On 1/6/2010 7:21 AM, Kevin Warfel wrote:

I can understand how the “manual parser” could find ‘fall’ and ‘yell’ in the lexicon, so your explanation is transparent for those elements, but I don’t understand how it was able to correctly analyze the verbal suffix without help from the parser, which looks at conditioning environments, since this surface form of this suffix assimilates in several ways to the vowel of the verb root to which it is suffixed.  Are you suggesting that this “manual parser” simply looks through the lexicon for a match to any Lexeme Form or any Allomorph, without regard to the Environment(s) in which they are stated to occur?

Yes, that is correct.  When you are manually breaking a word into its constituent morphemes in FLEx's interlinear, FLEx does a simple lexicon lookup.  Any form that matches the characters *and that has the same morpheme type* will match.  This "manual" parser does not pay attention to anything else in the lexical entry or in the grammar.  When breaking words by hand, FLEx assumes that you know what you are doing and that you can (and will) correct it if its guess is wrong.

The automated parser, on the other hand, pays attention to all of the environments, categories, etc.  The results of the automated parser are thus what your lexicon and grammar predict the analysis of the word should be.

--Andy

[Kevin Warfel]

Thanks, Andy.  That certainly helps me understand more clearly why my result was not due to the parser correctly analyzing the Puguli word structure, but rather to a simple blind (or nearly so) lookup in the lexicon.

 

Kevin Warfel

 

From: flex...@googlegroups.com [mailto:flex...@googlegroups.com] On Behalf Of Andy Black

Sent: Wednesday, January 06, 2010 11:04 AM
To: flex...@googlegroups.com

Subject: Re: [FLEx] How do you handle the inflection of idiomatic expressions in FLEx?

 

On 1/6/2010 7:21 AM, Kevin Warfel wrote:

[Ronald Moe]

Allan Johnson wrote:

“Is this the same issue that you have in mind?”

 

Yes. Linking complex forms to their roots is irrelevant for a stem based format. It only affects the display in a root based format. A hierarchical display would be something like the following. (I’ve inserted hyphens to make sure the indentation comes out correctly. Each entry should be on a separate line.)

 

ceive  (root)

--deceive v.

----deception n.

----deceptive adj.

------deceptively adv.

--perceive v.

----perception n.

----perceptive adj.

------unperceptive adj.

--receive v.

----reception n.

----receptive adj.

 

This can only be achieved if we can link ‘deceptively’ to ‘deceptive’, then link ‘deceptive’ to ‘deceive’, and then link ‘deceive’ to ‘ceive’. Currently all the derivatives have to be linked to ‘ceive’. Notice that if the subentries are alphabetized, ‘unperceptive’ would occur at the end. If your language has derivational prefixes, sorting the subentries would scatter the related subentries. At least in English most of our derivational affixes are suffixes. But FLEx isn’t designed for English. Philippine languages would suffer major disorganization because many derivational affixes are prefixes.

 

Ron Moe

 

---

From: flex...@googlegroups.com [mailto:flex...@googlegroups.com] On Behalf Of Allan Johnson

Sent: Wednesday, January 06, 2010 4:01 AM
To: flex...@googlegroups.com

Subject: Re: [FLEx] How do you handle idiomatic expressions in FLEx?

 

Ronald Moe wrote:

To handle a lexical phrase, such as an idiom, do the following:

1. Create a new entry and enter the form of the lexical phrase in the ‘Lexeme Form’ field.

1. In the ‘Complex Form Type’ field indicate what type of complex form it is (idiom, phrasal verb, etc.). You can add other types of complex forms to this list in the Lists-Complex Form Types area.

1. In the ‘Components’ field link the lexical phrase to each root. If you have not already created an entry for each root, FLEx will offer to create the entry for you. Alternatively you can link a lexical phrase to one or more of the words that it is composed of. For instance ‘peace treaty’ is composed of peace + treat-y. So you could link ‘peace treaty’ to ‘peace’ and ‘treat’ (the two roots) or to ‘peace’ and ‘treaty’. Unfortunately (as far as I know) FLEx does not currently permit hierarchical linking. So you cannot link ‘peace treaty’ to ‘treaty’ and then link ‘treaty’ to ‘treat’. If you attempt to do this, FLEx will not correctly export your data for publication.

Hi Ron,

The limitation mentioned here about hierarchical linking - you're talking about root-based dictionaries here, right? It's been a while since I've looked at this detail, but my recollection is that chains of links such as ‘peace treaty’ to ‘treaty’ and ‘treaty’ to ‘treat’ do work fine in the database, and display correctly in a stem-based format, but a root-based format has problems with it. The root-based format would have to offer multiple levels of subentries in order to handle it. Is this the same issue that you have in mind? It's one of my pet issues, since we can't yet achieve the needed format for some of our Philippine dictionaries. I hope it can be resolved before too long.

Allan

No virus found in this incoming message.

Checked by AVG - www.avg.com

Version: 9.0.725 / Virus Database: 270.14.127/2603 - Release Date: 01/06/10 00:35:00

[Allan Johnson]

Thanks for your good presentation of this root-based display issue. I believe there are issues to be dealt with both in FLEx itself and in the programs to which it exports dictionaries - Lexique Pro & SIL Publishing Solution - though I haven't done enough testing to say exactly where the ball is dropped in the processing of subentries. I suspect that up to now most testing has been done using just stem-based formats, so sub-entry issues have slipped by unnoticed. Do any of you have fuller info on the state of things for root-based dictionaries - on exactly what issues need attention?

One additional but related issue that your writeup brings to mind is that we need a way to configure the ordering of complex forms within an entry in root-based format. As far as I know, this has not yet been dealt with. Do any of you have fuller info on the status of this issue?

Allan

[Ronald Moe]

Allan Johnson wrote:

“Do any of you have fuller info on the status of this issue?”

 

I’ve written this issue up in some detail for the FLEx developers. There are two basic issues. (1) We need to be able to order and arrange the subentries under a particular root. (2) We need both global and entry-by-entry choices for where and how a complex entry is displayed. This includes (2a) full subentries under one or more of the roots, (2b) a minor entry (which would be alphabetized), and (2c) minor subentries under one or more of the roots. FLEx currently does not handle (1). It gives full control of (2a) and (2b). It does not permit (2c), but you can create cross references. I give examples of these types of entries in section 3.4 of my paper Introduction to Lexicography in the FLEx Helps-Resources folder. The way FLEx handles complex forms was changed in version 6.0. The changes gave us some new functionality, but unfortunately we lost the ability to create minor subentries.

 

Ron Moe

 

---

No virus found in this incoming message.

Checked by AVG - www.avg.com

Version: 9.0.725 / Virus Database: 270.14.129/2605 - Release Date: 01/07/10 00:35:00

[Ronald Moe]

Ron Moe wrote:

“Linking complex forms to their roots is irrelevant for a stem based format.”

 

Allan Johnson kindly (off-line) pointed out to me that this statement is wrong. (My brain must have been in neutral.) It is possible and useful to link complex forms to their roots and display the morphological makeup of a complex form in a stem-based dictionary. I give an example of this in section 3.4.4 of my paper Intro to Lexicography:

 

make peace (id. of make, peace) If two people or countries make peace, they stop fighting and agree to end a conflict.

 

You can do the same thing with derivatives:

 

peaceful (der. of peace -ful) adj. A place or time when there is no conflict or when not much is happening.

 

These would be entries in a stem-based dictionary. FLEx enables you to do this. For ‘peaceful’ you would have to have entries for ‘peace’ and ‘-ful’ and link ‘peaceful’ to both. Then you use Tools-Configure-Dictionary to tell FLEx to display this information. To do this you select “Component references.”

 

Ron Moe

 

---

From: flex...@googlegroups.com [mailto:flex...@googlegroups.com] On Behalf Of Ronald Moe

Sent: Wednesday, January 06, 2010 12:07 PM
To: flex...@googlegroups.com

[Kemp Pallesen]

This is an issue for Central Sinama also (a southern Philippine language with canonical bisyllabic roots), for which a root-based format is more suitable. I would settle at the moment for a default alphabetical ordering of complex forms within a main entry, with some way of overriding the default in particular cases. I would expect the dictionary user who has identified a root to cope with any ordering of subentries.

 

Kemp

---

From: flex...@googlegroups.com [mailto:flex...@googlegroups.com] On Behalf Of Allan Johnson
Sent: 07 January 2010 15:50
To: flex...@googlegroups.com
Subject: [FLEx] Root-based display of complex forms

 

Thanks for your good presentation of this root-based display issue. I believe there are issues to be dealt with both in FLEx itself and in the programs to which it exports dictionaries - Lexique Pro & SIL Publishing Solution - though I haven't done enough testing to say exactly where the ball is dropped in the processing of subentries. I suspect that up to now most testing has been done using just stem-based formats, so sub-entry issues have slipped by unnoticed. Do any of you have fuller info on the state of things for root-based dictionaries - on exactly what issues need attention?

[Beth]

On Jan 7, 2010, at 3:04 PM, Kemp Pallesen wrote:

> This is an issue for Central Sinama also (a southern Philippine
> language with canonical bisyllabic roots), for which a root-based
> format is more suitable. I would settle at the moment for a default
> alphabetical ordering of complex forms within a main entry, with
> some way of overriding the default in particular cases. I would
> expect the dictionary user who has identified a root to cope with
> any ordering of subentries.
>

The FLEx team is working on the "default alphabetical ordering of
complex forms within a main entry". Allowing the user to specify the
order is clearly desired, but that won't be able to be implemented
until much later.

-Beth