CLAN SUGAR analysis

[Amy Wilder]

Hi all,

I have a question about the SUGAR profile that can be run automatically with the "sugar +t*CHI" command in CLAN. In the CLAN manual (Section 8.10) it says:

 "MLU-S: This measure is the same as the MLU currently in CLAN"

Does this mean that the program computes a "regular" MLU (as described in section 7.18 MLU of the CLAN manual)? So the MLU-S calculated by CLAN would not include the extra SUGAR derivational morphemes (e.g., -er, -est, -tion, -ly, etc.)?

Thanks for your help clarifying this.

Amy Wilder

[Brian Macwhinney]

Dear Amy,
We configured CLAN’s SUGAR program in accord with the sample text in Pavelko & Owens 2017. For that sample, CLAN ends up with the same numbers as SUGAR. Because the CLAN method runs the text through MOR, it pulls out all the derivational morphemes mentioned in Table 3 of that article, along with some more. It is possible that the -sion and -tion suffixes noted in that table could occasionally be “missed:. For example, CLAN would not analyze “mission” as having the -sion suffix. A mission is not somethat that you miss.
I’ve added a paragraph at the end of the material in the manual on SUGAR to point to the two screencast tutorials at https://talkbank.org/screencasts/ that could help you see what is involved. This is what I added:

> There are two tutorial screencasts on the web at https://talkbank.org/screencasts/ that describe two different ways of preparing a file for SUGAR analysis in CLAN. The first screencast assumes that you have created a file using the SUGAR methodology in MS-Word. In that case, you save your transcripts as text only and then use CLAN's TEXT2CHAT program to create a CHAT file. The second screencast assumes that you have created the file from the beginning using the CLAN editor. In both cases, you end up with a CHAT file which you run through MOR and then SUGAR to automatically produce spreadsheet output.

— 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 view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/b6cfd570-4c29-41b7-bad1-3e7bc175ff7bn%40googlegroups.com.

[Amy Wilder]

Hi Dr. MacWhinney,

Thank you for the clarification regarding MLU calculation. After reading the MLU section in the CLAN manual, I had thought MLU was calculated mostly following Brown’s rules with a few exceptions like diminutives. I didn’t realize that most derivational morphemes were being counted. If I wanted to track which derivational morphemes are being included in MLU, would it be accurate to say that all the derivational morphemes in the English “0affix.cut” file are being counted for MLU as long as the root word they attach to does not have a [block pre] or [block post] code? Also, since MLU is calculated from the %MOR tier in KIDEVAL, this means the derivational morphemes are being counted in that program as well, correct?

Thanks again for your help!

Amy Wilder