phrasal verbs

[Janet Y Bang]

Hello, 

We have a question about linkages/compounds. We have been transcribing samples of LENA recordings in English (caregiver-child interactions with 2-year-old children) and we have compiled a list now of different phrasal verbs in transcripts. We defined phrasal verbs as units of "verb + adverb or preposition" where the meaning is attributed to the unit more than the meaning of each individual word. 

Our main interests for now are to collect measures of tokens, types (lemma), and MLUw (not morphemes) of caregiver and child speech. Determining the linkages are important for our measure of types, since we will be collapsing across word families with the lemmas. 

I noticed in the English MOR the n+v+prep.cut file has some phrasal verbs (e.g., dress up), but I wanted to check if there was some school of thought on how they should be treated? There are always situational issues we've come across with these. For example, in the case of particle verbs, sometimes these are transitive, so we weren't sure on how these should/would be transcribed (e.g., clean up your toys vs. clean that up). Also, it's not always clear from the usage if it is following our definition of "meaning is attributed to the unit more than the meaning of each individual word".

We were thinking of the following solutions: 

1. Link all phrasal verbs only when the key words are combined next to each other (e.g., clean up)

2. Link all phrasal verbs. In the case of transitive particle verbs, link those variations as well and count these all towards their respective 'particle verb family' (e.g., "clean that up" would be counted towards a type of "clean up")

3. Do not link any phrasal verbs (e.g., for "clean up", this would separated into the type "clean" and the type "up"). 

ate_up

back_up

calm_down

catch_up

chill_out

clean_up

come_back

come_here

come_on

come_out

come_over

drank_up

dress_up

get_down

get_out

get_up

giddy_up

go_ahead

go_on

hang_on

here_we_go

here_you_go

hold_on

hold_up

hurry_up

lay_down

let_go

light_up

lights_out

make_up

pick_up

pull_up

slow_down

stand_up

stay_out

take_care

throw_up

wake_up

watch_it

watch_out

work_out

zoom_in

Any thoughts would be much appreciated!

Janet

--

Janet Y. Bang, Ph.D.

Postdoctoral Fellow

Department of Psychology

Stanford University

jb...@stanford.edu

[Brian MacWhinney]

Dear Janet,

    For computational analysis two most difficult aspects of English grammar are phrasal verbs and the distinction between auxiliaries and the copula.  The problem with phrasal verbs is that the particle or preposition is often separated from the verb.  I agree that the solution you propose of marking the verb and its particle together is not a bad one.  In that case, the object of the preposition in a sentence such as "He calmed_down the crowd" will be interpreted as a direct object which is also not that bad.  Some of these combinations, such as chill_out,  don't take objects and many allow for omission of the "object".  

    So, if your goal is to get a nice dependency graph this could work well.  However, if your goal is to calculate MLU, then joining the particle with the verb would warp that computation.  

    By the way, the combinations you found in n+v+prep.cut are not verbs, but nouns, as in "he is a burnout".  This is a different pattern.

    In general, dealing with English verb+particle constructions is not an easy problem to solve.  I'm working now on the similar problem in Chinese which has remarkably similar issues, including the verb separation pattern.

-- 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/BYAPR02MB5285EA0D6F11967F870B4F9AD7D40%40BYAPR02MB5285.namprd02.prod.outlook.com.

[Janet Bang]

Hi Brian, 

Thank you for your response. It's comforting to know that this wasn't just us! Since we are not concerned with MLU (morphemes), we will likely move forward with our solution (2), which is to link all phrasal verbs and also do our best to group phrasal verbs and their variants together (e.g., "clean_that_up" and "clean_up" as two of the same type). Since we are working with types, this means that noun and verb forms will be collapsed (e.g., "let's play dress_up" and "let's dress up the doll").

Janet

To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/5077B49D-A010-49CB-9700-51875C2BD755%40andrew.cmu.edu.

[Brian MacWhinney]

It's an interesting approach, but you will have a lot to group together if you want to pull in the object too.  Consider "please clean_all_the_mess_that_you_have_left_in_your_room_up before you go out"

--Brian

To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/CAC5V4hhyayRCczSxgj3hqDXZPEj7YSAr2N7jVdQep%2BN_3SU-Tw%40mail.gmail.com.

[Janet Bang]

Hi Brian,

Yes you bring up a good point! After conferring with my colleagues we are now thinking of limiting the linkages of such phrases to the following cases: 

1) The linkage is used as communicators in a formulaic way

For example, "come on" is often used as a replacement for "come" and the preposition "on" is not used in a meaningful way, and "hold on", "hang on" or "hold up" can be used as synonyms for "wait" (e.g., "hold on a minute" when the caregiver is telling the child to wait) whereas in "hold on tight to that there" the combination of "hold" and "on" does not mean "wait". We noticed that many of these were also in the co-under.cut files.

2) The linkage is used to refer to a noun

For example, "let's play dress up" (noun) versus "let's dress up the doll" (verb), or "put on your pull up" (noun) versus "pull up your pants" (verb)

This seems like it best allows us to capture usage in a way that is consistent with respect to its function and meaning.

Janet