Error Message: Current BEG time is smaller than previous tier BEG time

[Nan Ratner Lab]

Hello,

Our lab has been cleaning up transcription files using CHECK in order to run some analyses and a couple of the members have been receiving error messages that read, "Current BEG time is smaller than previous tier BEG time". We were not sure if this would require them to rebullet the entire file or if there was an easier way to fix this.  Any help is appreciated!

Thank you,

Allison Catalano 

[Brian MacWhinney]

Dear Allison,

   This is probably the most frustrating error type to correct, because finding exactly which @Bg is supposed to align with with @Eg by paging through transcripts is frustrating.  The best approach is to run GEMLIST which will show you which gem markers are where and that should make it clear which ones are missing.  Sometimes the error involves misspelling of the field after the @Bg or @Eg,  as in 

@Bg: book reading

*TXT: the little dog barked.

@Eg: book reeding

—Brian MacWhinney

--
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/4a0c8536-5480-44f1-8559-e7c6c595004f%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Leonid Spektor]

Allison,

This error only effects two tiers. There is no need to rebullet the whole file. The error is the result of time information within a bullets violating normal time progression. Following example will demonstrate the problem:

*FAT: and what are you asking for ? •0_2312•

*ROS: alert [!] alert ! •2312_3715•

*FAT: alert means like it's time for a fire alert . •2310_6937•

The last *FAT tier has BEG time within it's bullet set to 2310 msecs, but the previous *ROS tier has BEG time set to 2312 msecs which is 2 msecs later than *FAT tier's 2310 msecs. This means that either those two speakers should be placed in different order like this:

*FAT: and what are you asking for ? •0_2312•

*FAT: alert means like it's time for a fire alert . •2310_6937•

*ROS: alert [!] alert ! •2312_3715•

or the time BEG value of either speaker's bullets adjusted slightly to make sure that it complies with normal time flow, like this:

*FAT: and what are you asking for ? •0_2312•

*ROS: alert [!] alert ! •2312_3715•

*FAT: alert means like it's time for a fire alert . •2312_6937•

You can open data file that has this error with CLAN, then locate the tier that causes the error and select menu "Mode->Expand bullets" or press ESC-a keys. This will display the time values within bullets and it will let you edit them by hand. 

Leonid.

[Nan Ratner Lab]

Thank you for your quick response!  I think the problem is that the transcripts we are working with are a compilation of different trials copied and pasted into one file.  After reading your message, we tried to edit the bullet numbers, but it seems that there are many bullet times that do not correspond since the files were copied and pasted in.  Is there an easier solution than re-numbering all of the bullets from the copied file?  I have included an example of the issue below.

Example:

*SUB: <he was> [//] he loved going to the casino and playing the slot

machines.  •797965_803733•

*SUB: he was on the slot machines three hours 0 [=! vocalizes]. •803733_806725•

@Eg: Run1

@Bg: Run2

*SUB: so Butch and his sister Sally went to the playground uh@fp looking

for some fun. •123817_158243•

*SUB: and they decided to play on the <&see> [//] seesaw so Butch got on

one side.

*SUB: Sally got on the other side and immediately they saw

the problem. •158243_165974•

[Leonid Spektor]

Allison,

The transcript's tiers have to be in an increasing time order. It is not allowed to have different parts from different time trials  Also, since the bullet times are different, I have to assume that those different trial tiers are associated with different media files. Each CHAT transcript can only have one media file associated with it. So, what you did will not let you playback bullets correctly. If different trials from which you copied and pasted tiers are all referencing the same media file, then you need to assemble those trial samples in the right order within your compilation file.  If you do not care about playing bullets in this compilation of different trials file, then you can just delete all bullets from the file using  "Edit->Replace" menu. In the "Replace String" dialog put text cursor in the "Replace" field and click on "insert Bullet" button. Next click "Replace all" button and CLAN will remove all bullets from the file. Now CHECK will not complain about bullets.

Leonid.

To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/25c31c92-d86b-4982-b8aa-35cfd845a8a4%40googlegroups.com.

[Brian MacWhinney]

In addition to the idea of deleting all bullets which Leonid mentions, there is a way to tell CHECK to ignore your problems with bullets. This is to include an Options line that reads:

@Options:     bullets

This line has to precede the first @ID line, after the @Participants line.  If you include this exemption, CHECK will be happy, but your data will not play correctly in sequence.  I think this only applies in the case that your are at least conforming with the idea that each transcript links to a single media file.  

The best approach is to fix the misalignment either by changing the time values in the bullets and reordering the transcript, if necessary.

—Brian MacWhinney

To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/6547F623-376B-49B3-B447-A47C864FB380%40andrew.cmu.edu.

[Nan Ratner Lab]

Thank you so much for both of your help!

[Nan Ratner Lab]

Another member of our lab is having problems with this error, but with a different set of data.  This file was not copied and pasted together like the other ones.  In this file, most of the bullets are showing up as the same time frame.  We were not sure how to address this other than manually changing the numbers as you suggested or rebulleting the file.

49    *MOT: five times. •61461_64037•

50    *MOT: one two three four five. •2011_5921•

51    *MOT: five you do yellow. •2011_5921•

52    *MOT: yellow. •2011_5921•

53    *MOT: one two three four five. •2011_5921•

54    *MOT: yay! •2011_5921•

55    *MOT: now we turn the page and what happens? •2011_5921•

56    *MOT: look at that! •2011_5921•

57    *MOT: one two three four five. •2011_5921•

58    *MOT: 0 [=! laughs]. •11447_13572•

This happens consistently throughout the transcript.

120   *CHI: 0 [=! crying]. •189190_190633•

121   *MOT: that's distracting. •2011_5921•

122   *CHI: xxx. •2011_5921•

123   *MOT: no. •195456_196189•

Any help is appreciated, thank you!

[Brian MacWhinney]

Apparently someone did a lot of cutting and pasting of individual bullets.  It is not clear why, but that is the only way that this pattern could have arisen.  It is good that CHECK catches these errors, because they are pretty serious.  I think this file just has to be relinked.  If there are big stretches that do not have these problems, then you can just use CHECK to locate the problem areas to be fixed.

—Brian MacWhinney

To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/c4eec44b-3638-47cf-b865-88ca07ad8277%40googlegroups.com.

[Leonid Spektor]

Nan,

CLAN will allow you to just relink the bullets that are erroneous. There is no need to relink the whole transcript file. You just need to place text cursor on the last tier with the correct bullet and press F5 key and just continue from that point until you redo all bad bullets. At that time you can just stop CLAN and save file.

Leonid.

To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/71DF11CE-1BB7-473A-89BA-C7BA193888F2%40cmu.edu.