error of irregular past participal POS tagging

[Brian MacWhinney]

Dear Y. J.,
    We currently have a problem with the tagging of past participles after auxiliaries  by POST.  We believe that it may be arising from a problem in the way that POST is processing clitic structures including auxiliaries, and I have asked Christophe Parisse to review that issue.  Until we are able to resolve this problem, I would recommend examining this particular issue in detail by searching for both patterns.  Of course, the particular test you are using is going to oversample the problem because all of the contexts are basically the same.  In full corpora, there will still be a problem, but the numbers will be different.

-- Brian MacWhinney
On May 31, 2012, at 11:28 AM, Huang Y.J. wrote:

> Dear all,
> 
> I am currently using CLAN to do grammar pattern search. And I
> encountered an error tagging problem could use some help.
> 
> I want to search a grammatical pattern BeV + p.p. in a CLAN mor tagged
> corpus. To build expressions for testing, I first make a list of a set
> of sentences in a file. And then I tagged it with mor to do
> observations so that I can generalize searching expressions from it
> for ‘combo’ search. After tagged the file with mor, I found that the
> past participles of the irregular verb 'hurt' in the sentences are
> systematically mis-tagged as the following examples:
> 
> E.g.
> (correctly tagged as 'part|hurt&PERF'  )
> *TXT:	He has been hurt by the policeman .
> %mor:	n:prop|He aux|have&3S aux|be&PERF part|hurt&PERF prep|by det|the
> n|+n|police+n|man .
> 
> (incorrectly tagged as adj|hurt)
> *TXT:	He had been hurt by the policeman .
> %mor:	n:prop|He aux|have&PAST v:cop|be&PERF adj|hurt prep|by det|the n|
> +n|police+n|man .
> 
> (incorrectly tagged as adj|hurt)
> *TXT:	He will be hurt by the policeman .
> %mor:	n:prop|He mod|will v:cop|be adj|hurt prep|by det|the n|+n|police
> +n|man
> 
> Only two out of eighteen occurrences are correctly tagged as past
> participle, the rest of them are mis-tagged as adj. I wonder if I did
> something wrong?
> 
> Many thanks.
> Huang Y.J.
> ---
> In attachment of my sample sentences and processed pst file as the
> following:
> 
> He is hurt by the policeman.
> He was hurt by the policeman.
> He has been hurt by the policeman.
> He had been hurt by the policeman.
> He will be hurt by the policeman.
> He will have been hurt by the policeman.
> He would have been hurt by the policeman.
> 
> He is going to be hurt by the policeman.
> He was going to be hurt by the policeman.
> 
> He is badly hurt by the policeman.
> He was badly hurt by the policeman.
> He has been badly hurt by the policeman.
> He had been badly hurt by the policeman.
> He will be badly hurt by the policeman.
> He will have been badly hurt by the policeman.
> He would have been badly hurt by the policeman.
> 
> He is going to be badly hurt by the policeman.
> He was going to be badly hurt by the policeman.
> 
> And the pst file after running MOR and POST:
> 
> @UTF8
> @Begin
> @Languages:	eng
> @Participants:	TXT Text
> @Options:	heritage
> @ID:	eng|text|TXT|||||Text||
> *TXT:	He is hurt by the policeman .
> %mor:	n:prop|He aux|be&3S part|hurt&PERF prep|by det|the n|+n|police+n|
> man .
> *TXT:	He was hurt by the policeman .
> %mor:	n:prop|He v:cop|be&PAST&13S adj|hurt prep|by det|the n|+n|police
> +n|man
> 	.
> *TXT:	He has been hurt by the policeman .
> %mor:	n:prop|He aux|have&3S aux|be&PERF part|hurt&PERF prep|by det|the
> n|+n|police+n|man
> 	.
> *TXT:	He had been hurt by the policeman .
> %mor:	n:prop|He aux|have&PAST v:cop|be&PERF adj|hurt prep|by det|the n|
> +n|police+n|man
> 	.
> *TXT:	He will be hurt by the policeman .
> %mor:	n:prop|He mod|will v:cop|be adj|hurt prep|by det|the n|+n|police
> +n|man
> 	.
> *TXT:	He will have been hurt by the policeman .
> %mor:	n:prop|He mod|will aux|have v:cop|be&PERF adj|hurt prep|by det|
> the n|+n|police+n|man
> 	.
> *TXT:	He would have been hurt by the policeman .
> %mor:	n:prop|He mod|will&COND aux|have v:cop|be&PERF adj|hurt prep|by
> det|the
> 	n|+n|police+n|man .
> @Blank
> *TXT:	He is going to be hurt by the policeman .
> %mor:	n:prop|He aux|be&3S part|go-PROG inf|to v:cop|be adj|hurt prep|
> by det|the
> 	n|+n|police+n|man .
> *TXT:	He was going to be hurt by the policeman .
> %mor:	n:prop|He aux|be&PAST&13S part|go-PROG inf|to v:cop|be adj|hurt
> prep|by
> 	det|the n|+n|police+n|man .
> @Blank
> @Blank
> *TXT:	==== He is badly hurt by the policeman .
> %mor:	? n:prop|He aux|be&3S adv|bad&dadj-LY adj|hurt prep|by det|the n|
> +n|police+n|man .
> *TXT:	He was badly hurt by the policeman .
> %mor:	n:prop|He v:cop|be&PAST&13S adv|bad&dadj-LY adj|hurt prep|by det|
> the n|+n|police+n|man .
> *TXT:	He has been badly hurt by the policeman .
> %mor:	n:prop|He aux|have&3S aux|be&PERF adv|bad&dadj-LY adj|hurt prep|
> by det|the n|+n|police+n|man .
> *TXT:	He had been badly hurt by the policeman .
> %mor:	n:prop|He aux|have&PAST v:cop|be&PERF adv|bad&dadj-LY adj|hurt
> prep|by 	det|the n|+n|police+n|man .
> *TXT:	He will be badly hurt by the policeman .
> %mor:	n:prop|He mod|will v:cop|be adv|bad&dadj-LY adj|hurt prep|by det|
> the n|+n|police+n|man .
> *TXT:	He will have been badly hurt by the policeman .
> %mor:	n:prop|He mod|will aux|have v:cop|be&PERF adv|bad&dadj-LY adj|
> hurt prep|by 	det|the n|+n|police+n|man .
> *TXT:	He would have been badly hurt by the policeman .
> %mor:	n:prop|He mod|will&COND aux|have v:cop|be&PERF adv|bad&dadj-LY
> adj|hurt 	prep|by det|the n|+n|police+n|man .
> @Blank
> *TXT:	He is going to be badly hurt by the policeman .
> %mor:	n:prop|He aux|be&3S part|go-PROG inf|to v:cop|be adv|bad&dadj-LY
> adj|hurt 	prep|by det|the n|+n|police+n|man .
> *TXT:	He was going to be badly hurt by the policeman .
> %mor:	n:prop|He aux|be&PAST&13S part|go-PROG inf|to v:cop|be adv|
> bad&dadj-LY adj|hurt prep|by det|the n|+n|police+n|man .
> @Blank
> @End
> 
> -- 
> You received this message because you are subscribed to the Google Groups "chibolts" group.
> To post to this group, send email to chibolts at googlegroups.com.
> To unsubscribe from this group, send email to chibolts+unsubscribe at googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/chibolts?hl=en.
> 
> 

-- 
You received this message because you are subscribed to the Google Groups "chibolts" group.
To post to this group, send email to chibolts at googlegroups.com.
To unsubscribe from this group, send email to chibolts+unsubscribe at googlegroups.com.
For more options, visit this group at http://groups.google.com/group/chibolts?hl=en.