treament of circumfixes

[Bruno Estigarribia]

Brian,
As usual, you are a step ahead. It is true that at least in this corpus, 
the circumfix doesn't always appear "complete". Sometimes only the 
suffixal part -i is present, with "no" from Spanish before the root. A 
rather interesting case of mixing, I'd say.
We are doing the %mor tier, no question. I am just morphologizing the 
main tier also to help my coder, who doesn't know all that much Guaraní. 
My understanding is that mlu (and freq) will work on these old codes.
thanks!
Bruno
> Bruno,
>
>     Leonid is right in suggesting that you do your coding on the %mor 
> tier, even if you have to do it by hand.  However, I would like to add 
> a comment regarding circumfixes.  CLAN has no special method for 
> considering circumfixes as single morphemes.  In part this is because 
> trying to figure this out computationally would be quite difficult. 
>  But it is also motivated by the fact that it is difficult to argue 
> that children learn circumfixes or other discontinuous markers as 
> single morphemes.  Consider the case of French ne ….. pas.  Here, one 
> can trace the fact that the “ne” is often omitted in Quebécois and 
> even older French dialects, despite the linguistic claim that this a 
> single morpheme.  And children may omit either part of this “single” 
> morpheme.  I know nothing about Guarani, but Navajo has a rich set of 
> aspectualmorphemes linked across positions  2 and 6 in the verb that 
> one could analyze as discontinuous morphemes, but I don’t think one 
> gains that much through that type of analysis.
>     Of course there are always arguments on both sides of topics like 
> this.  So, if you wish to treat nd … i as a single morpheme for MLU, 
> then you can achieve that effect by coding on the %mor line in this way:
>
> neg#1s#v|walk&neg
>
> The ampersand will assure that the second part of nd….i is not 
> counted. This is a hack, but probably sufficient.  We could always 
> introduce another symbol for this purpose, but one would not gain much 
> through that.
>
> In terms of main line coding, you could use a parallel method, as in
>
> nd-a-guata&i
>
> you get the right number of morphemes, and you can always search for 
> &i.  The only problem is that you somewhat mess up easy retrieval of 
> verb stems, but then main line morphemic coding was never a great idea 
> for any language.
>
> —Brian MacWhinney
>
> On Nov 14, 2013, at 9:46 AM, Bruno Estigarribia <brunilda at gmail.com 
> <mailto:brunilda at gmail.com>> wrote:
>
>> Hello all!
>>
>> I am analyzing a corpus of code-switching between Spanish and Guaraní.
>> While we prepare a better version using a %mor tier (mostly done by 
>> hand, obviously, since I don't think we have lib files for Guaraní 
>> :-) ), I have been doing some quick "old-style" morphemization on the 
>> main tier (which I believe will still work with MLU as long as I 
>> include the -t%mor switch, right?).
>> Then a question came up: how does mlu treat circumfixes? Example, the 
>> Guaraní negation nd-...-i as in:
>> nd-a-guata-i
>> NEG#1s#walk-NEG
>> "I don't walk"
>> I obviously don't want to count nd# and -i separately...
>> I couldn't find the answer on either the CHAT or the CLAN manuals, or 
>> on this list. Sorry if I've missed something!
>> Thanks!
>> Bruno
>>
>> Bruno Estigarribia
>>
>> Assistant Professor of Spanish, Department of Romance Languages and 
>> Literatures
>>
>> Research Assistant Professor of Psychology, Cognitive Science Program
>>
>> Affiliate Faculty, Global Studies
>>
>> https://sites.google.com/site/brunoestigarribialing/
>>
>> http://roml.unc.edu/people/spanish/faculty/bruno-estigarribia/
>>
>> Dey Hall, Room 332, CB# 3170
>>
>> University of North Carolina at Chapel Hill
>>
>>
>> -- 
>> You received this message because you are subscribed to the Google 
>> Groups "chibolts" group.
>> To unsubscribe from this group and stop receiving emails from it, 
>> send an email to chibolts+unsubscribe at googlegroups.com 
>> <mailto:chibolts+unsubscribe at googlegroups.com>.
>> To post to this group, send email to chibolts at googlegroups.com 
>> <mailto:chibolts at googlegroups.com>.
>> To view this discussion on the web visit 
>> https://groups.google.com/d/msgid/chibolts/ec8d98cb-371e-4edd-b28e-6ab744f86b2f%40googlegroups.com.
>> For more options, visit https://groups.google.com/groups/opt_out.
>

-- 
You received this message because you are subscribed to the Google Groups "chibolts" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chibolts+unsubscribe at googlegroups.com.
To post to this group, send email to chibolts at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/5284F316.1040809%40gmail.com.
For more options, visit https://groups.google.com/groups/opt_out.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/chibolts/attachments/20131114/81b84f03/attachment.htm>