UniversalDependencies / UD_English-PUD

Parallel Universal Dependencies.
Other
10 stars 2 forks source link

Make Historian the head of Historian David Crouch #52

Closed AngledLuffa closed 4 months ago

nschneid commented 4 months ago

This is a special construction for combining a role or title with the name of an individual. We tentatively decided that the role/title should be a modifier (because it can be omitted), and a good name for the deprel would be nmod:desc ("descriptor"). But we have not implemented this policy yet.

AngledLuffa commented 4 months ago

Here are a few instances of this in EWT:

# sent_id = weblog-blogspot.com_dakbangla_20050311135387_ENG_20050311_135387-0169
# text = A report by the Center for Disease Control of interviews with AMI employees (as well as detailed interviews by author Leonard Cole) ...
21      author  author  NOUN    NN      Number=Sing     22      compound        22:compound     _
22      Leonard Leonard PROPN   NNP     Number=Sing     19      nmod    19:nmod:by      _
23      Cole    Cole    PROPN   NNP     Number=Sing     22      flat    22:flat SpaceAfter=No

# sent_id = weblog-blogspot.com_dakbangla_20041119231111_ENG_20041119_231111-0010
# text = According to Professor S D Muni, ...
3       Professor       Professor       PROPN   NNP     Number=Sing     36      obl     36:obl:according_to     _
4       S       S       PROPN   NNP     Number=Sing     3       flat    3:flat  _
5       D       D       PROPN   NNP     Number=Sing     3       flat    3:flat  _
6       Muni    Muni    PROPN   NNP     Number=Sing     3       flat    3:flat  SpaceAfter=No

# sent_id = email-enronsent31_01-0005
# text = After talking with Professors Titman, Ronn, Brown and Jemison, ...
1       After   after   SCONJ   IN      _       2       mark    2:mark  _
2       talking talk    VERB    VBG     Tense=Pres|VerbForm=Part        15      advcl   15:advcl:after  _
3       with    with    ADP     IN      _       4       case    4:case  _
4       Professors      Professor       PROPN   NNPS    Number=Plur     2       obl     2:obl:with      _
5       Titman  Titman  PROPN   NNP     Number=Sing     4       flat    4:flat  SpaceAfter=No

# sent_id = email-enronsent31_01-0015
# text = Professor Ronn:
1       Professor       Professor       PROPN   NNP     Number=Sing     0       root    0:root  _
2       Ronn    Ronn    PROPN   NNP     Number=Sing     1       flat    1:flat  SpaceAfter=No
3       :       :       PUNCT   :       _       1       punct   1:punct _

# sent_id = newsgroup-groups.google.com_alt.animals_02c2d614bfbf6b20_ENG_20050223_232900-0022
# text = OSU forestry professors William Ripple and Robert Beschta ...
1       OSU     OSU     PROPN   NNP     Number=Sing     3       compound        3:compound      _
2       forestry        forestry        NOUN    NN      Number=Sing     3       compound        3:compound      _
3       professors      professor       NOUN    NNS     Number=Plur     4       nmod:desc       4:nmod:desc     _
4       William William PROPN   NNP     Number=Sing     10      nsubj   10:nsubj        _
5       Ripple  Ripple  PROPN   NNP     Number=Sing     4       flat    4:flat  _

# sent_id = newsgroup-groups.google.com_humanities.lit.authors.shakespeare_0018a7697318f71f_ENG_20031006_163200-0046
# text = <<Doctor RODERIGO LOPEZ (1525 -1594) was the Portugese Jewish doctor accused by Essex of trying to poison Queen Elizabeth.
2       Doctor  Doctor  PROPN   NNP     Number=Sing     14      nsubj   14:nsubj        _
3       RODERIGO        Roderigo        PROPN   NNP     Number=Sing     2       flat    2:flat  _
4       LOPEZ   Lopez   PROPN   NNP     Number=Sing     2       flat    2:flat  _

# sent_id = reviews-170650-0008
# text = Doctor Gonzales and his entire staff are the most professional people I have ever dealt with.
1       Doctor  Doctor  PROPN   NNP     Number=Sing     11      nsubj   11:nsubj        _
2       Gonzales        Gonzales        PROPN   NNP     Number=Sing     1       flat    1:flat  _

# sent_id = reviews-170650-0012
# text = Thank you Doctor Gonzales, Doctor Stout, Eva Marie and the entire staff!
1       Thank   thank   VERB    VBP     Mood=Ind|Number=Sing|Person=1|Tense=Pres|VerbForm=Fin   0       root    0:root  _
2       you     you     PRON    PRP     Case=Acc|Person=2|PronType=Prs  1       obj     1:obj   _
3       Doctor  Doctor  PROPN   NNP     Number=Sing     1       vocative        1:vocative      _
4       Gonzales        Gonzales        PROPN   NNP     Number=Sing     3       flat    3:flat  SpaceAfter=No
5       ,       ,       PUNCT   ,       _       6       punct   6:punct _
6       Doctor  Doctor  PROPN   NNP     Number=Sing     3       conj    1:vocative|3:conj:and   _
7       Stout   Stout   PROPN   NNP     Number=Sing     6       flat    6:flat  SpaceAfter=No

# sent_id = reviews-288100-0002
# newpar id = reviews-288100-p0002
# text = I began seeing Dr. Romanick back in 2000...
4       Dr.     Dr.     PROPN   NNP     Number=Sing     3       obj     3:obj   _
5       Romanick        Romanick        PROPN   NNP     Number=Sing     4       flat    4:flat  _

# newdoc id = reviews-148971
# sent_id = reviews-148971-0001
# newpar id = reviews-148971-p0001
# text = Dr. Strzalka at Flagship CVTS is not a good doctor
1       Dr.     Dr.     PROPN   NNP     Number=Sing     10      nsubj   10:nsubj        _
2       Strzalka        Strzalka        PROPN   NNP     Number=Sing     1       flat    1:flat  _

So, more often than not, the title is the head, at least for the titles I searched for. Certainly could expand that search. Would rather just focus on the one issue now and do a more global fix later

nschneid commented 4 months ago

Fair enough to make it flat for now. See: UniversalDependencies/UD_English-EWT#59

AngledLuffa commented 4 months ago

We will just have to remember that this exists if'n the new standard is eventually adopted!