C-NNLA output measures

[Leonid Spektor]

Hi,

	The way C-NNLA computes columns "% sentences produced", "% sentences with correct syntax, semantics*", "% sentences with flawed syntax" and "% sentences with flawed semantics*" has been fixed. The bug had to do with word contractions. For example, when computing those particular columns 
for the word "she's" C-NNLA did not see the "~cop|be&3S" part of the word.

This bug has been fixed and new CLAN has been updated on the web.


Leonid.

> On Oct 7, 2022, at 12:17, mara.st... at gmail.com <mara.steinberg at gmail.com> wrote:
> 
> Hi chiblots,
> 
> We were wondering if there was a way to determine which sentences/utterances are  included in the output measures from C-NNLA. Specifically, we are looking at the % sentences with flawed syntax. We are getting a mismatch between our manual coding and the automated output and are wondering if we are coding something incorrectly or different than the program. We see the formula in the CLAN manual (p.119), but were unsure of exactly how the calculation works, especially in terms of which utterances/ sentences get included (i.e., abandoned utterances, complete, etc.).
> 
> In the example below, C-NNLA reported 100% as the output for % sentences with flawed syntax. Manually, we counted 11 utterances total from the participant, 5 were excluded using [+exc] since they were unrelated to the picture description and 3 were marked as ungrammatical [+gram]. That would leave 3 utterances that could be counted as grammatically correct but we are not sure if single word or fragmented utterances were counted. It seems like there’s at least one grammatically correct sentence (“she's just out of it” - but includes a revision) making the reported value of 100% possibly incorrect.
> 
> 
> *PAR: &-right now <where do I> [/] where do I put the <another> [>] +/. [+ exc] 
> *INV: <you need> [<] to tell me about it . 
> *PAR: or just tell you about it . [+ exc] 
> *PAR: sorry &=laughs . [+ exc] 
> *PAR: the mum's not very clever because she's xxx xxx sink &=laughs . [+ gram] 
> *PAR: &-uh and the boys [: boy] [* m:+s] are [: is] [* m:vun:a] very naughty because they're [: he's] [* s:r:gc:pro] xxx the cookie jar and he's gonna fall off the +... [+ gram] [+ gcc] 
> *PAR: I've done that before . [+ exc] 
> *PAR: &-um yes . 
> *PAR: and [e] &-yep xxx &+fai <the kid> [//] <the boy> [//] the &-um (.) girl was [: is] [* m] gaving [: giving] [* m] xxx her [//] him and her +... [+ gram] 
> *PAR: and [e] &-you_know they're in the kitchen +... 
> *PAR: she [//] she's just out of it . 
> *PAR: &=laughs and that's about it . [+ exc] 
> *INV: yep !
> Thanks for your help!
> 
> ~Mara Steinberg Lowe & Kirrie Ballard
> 
> 
> -- 
> 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 view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/03fb5c3c-4050-46fa-94b5-bc1710f45173n%40googlegroups.com <https://groups.google.com/d/msgid/chibolts/03fb5c3c-4050-46fa-94b5-bc1710f45173n%40googlegroups.com?utm_medium=email&utm_source=footer>.

-- 
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 view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/5EBFAEE3-BC44-4640-A7A0-7E2D75786CEC%40andrew.cmu.edu.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/chibolts/attachments/20221008/a2ad0f6b/attachment.htm>