Re: Marking addressee and excluding from MLU counts

[Rowland, Caroline]

Hi Leonid and Brian,

 

We’ve tried using [+ mlue] postcodes on the mainline (see attached for sample test file) and it doesn’t seem to be excluding the utterances automatically. Note that we don’t have a %mor line so we are using simply:

 

mlu –t%mor *.cha

 

Note also that we are transcribing in ELAN (for various reasons) and converting to CHAT using the ELAN export to chat command, so the attached file is a conversion from an eaf file.

 

Thanks

 

Caro

 

From: chib...@googlegroups.com <chib...@googlegroups.com> On Behalf Of Leonid Spektor
Sent: 20 January 2019 01:22
To: 'heba salama' via chibolts <chib...@googlegroups.com>
Cc: McLaughlin, Paula <Paula.Mc...@liverpool.ac.uk>
Subject: Re: Marking addressee and excluding from MLU counts

 

Caro,

 

            If you want to exclude utterance and its dependent tiers, then it is best to use post-codes. For example, instead of %add: SIB dependent tier specify post-code [+ sib] on the main speaker tier. If you want to always exclude utterance from MLU, then specify post-code [+ mlue]. The advantage of using this post-code is that it will effect results of MLU, KIDEVAL and few other CLAN commands automatically without you having to specify anything on the command line. If you choose to use your own post-code, then on command line you need to add option -s"[+ sib]" to exclude utterances with that post-code from any CLAN analyses.

Leonid.

 

On Jan 19, 2019, at 10:27, Rowland, Caroline <crow...@liverpool.ac.uk> wrote:

 

Hi all,

 

We want to both transcribe utterances addressed to siblings of the target child but also to sometimes exclude them from MLU counts.  This is because sometimes we will want a measure of how much CDS a particular adult is using (no matter the addressee) and sometimes we will want a measure of how much CDS is addressed to a particular target_child.

 

Normally we would mark utterances addressed to a non-target child (e.g. a sibling) using a dependent tier; e.g.

 

*MOT: what do you want?

%add:   SIB

 

If we transcribe like this, is there then an easy way to then exclude these utterances from an MLU count; or indeed from any other CLAN command?

 

Or do we need to use a different notation?

Thanks

 

Caro

 

[Leonid Spektor]

Caro,

MLU is supposed to be used with %mor tier. If you are running MLU on just speaker tier, then you will get fewer morpheme counts than if you had %mor tiers in the file, because morphemes are identified by special symbols on %mor tier only. I will change MLU to not ignore [+ mlue] post-code when running MLU on speaker tier.

Leonid.

-- 
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+u...@googlegroups.com.
To post to this group, send email to chib...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/86c331348476448581f0c777eb3bcf5b%40mpi.nl.
For more options, visit https://groups.google.com/d/optout.

[Rowland, Caroline]

Thank you. That’s great.

Best wishes,

To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/586F71B6-8FD4-4962-9A8E-A1A0055AB885%40andrew.cmu.edu.

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

 

--

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+u...@googlegroups.com.
To post to this group, send email to chib...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/586F71B6-8FD4-4962-9A8E-A1A0055AB885%40andrew.cmu.edu.

[Leonid Spektor]

Caro,

New CLAN is on the web.

Leonid.

To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/8fa242e38ee24dfda3a7c02260d94e8c%40liverpool.ac.uk.