SemCor | BBN | NomBank | VerbNet/SemLink | PropBank | OntoNotes 4 (5) | FrameNet Full Text | |
---|---|---|---|---|---|---|---|
values (times, quantities) | |||||||
named entities | CNE: The Fold/org | BNE: The Fold/org:other | ONNE: The Fold/org | CNE: The Fold/org | |||
nouns | WNS: fold.n.01 | BED | NBF: folding.01 | ~ONS: fold-n.01, (ONF) | FNF:
Reshaping, Endeavor_failure |
||
verbs | WNS: fold_up.v.01 | VNC: bend-45.2 | PBF: fold-v.03 | ONS: fold-v.01, ONF: fold-v.03 | |||
anaphoric coreference | |||||||
noun coreference |
A number of large corpora of English text include semantic annotations, such as named entities, word senses, predicate-argument structures, and coreference. This page provides an overview of several such corpora. (Syntactic annotations are not discussed here.)
Click on a corpus name for details and examples. Many of the table entries and annotations have explanatory tooltips that can be revealed by hovering your mouse over the text.
(stats are from this page)
TheDT
materialsNNmaterial.2
forIN
compromiseNNcompromise.1
areVBbe.3
at handJJat_hand.1
:p
TheDT
NationNNother
,p
Walter LippmanNNPperson
andCC
otherJJother.1
soberJJsober.2
commentatorsNNcommentator.1
(p
seeRBsee.1
Alan ClarkNNPperson
onIN
p.NN
367CD
)p
haveVBP
spelledVBspell_out.1
themPRP
outIN
again and againRBagain_and_again.1
.p
The directories brown1 and brown2 contain the portion of the corpus with full annotations; brownv has the verbs-only portion. From semcor/brown1/tagfiles/br-b20.xml in the NLTK release of SemCor 3.0:
<s snum="47"> <wf cmd="ignore" pos="DT">The</wf> <wf cmd="done" pos="NN" lemma="material" wnsn="2" lexsn="1:10:00::">materials</wf> <wf cmd="ignore" pos="IN">for</wf> <wf cmd="done" pos="NN" lemma="compromise" wnsn="1" lexsn="1:04:00::">compromise</wf> <wf cmd="done" pos="VB" lemma="be" wnsn="3" lexsn="2:42:05::">are</wf> <wf cmd="done" pos="JJ" lemma="at_hand" wnsn="1" lexsn="5:00:02:close:01">at_hand</wf> <punc>:</punc> <wf cmd="ignore" pos="DT">The</wf> <wf cmd="done" rdf="other" pos="NN" pn="other" ot="notag">Nation</wf> <punc>,</punc> <wf cmd="done" rdf="person" pos="NNP" lemma="person" wnsn="1" lexsn="1:03:00::" pn="person">Walter_Lippmann</wf> <wf cmd="ignore" pos="CC">and</wf> <wf cmd="done" pos="JJ" lemma="other" wnsn="1" lexsn="3:00:00::">other</wf> <wf cmd="done" pos="JJ" lemma="sober" wnsn="2" lexsn="5:00:00:serious:00">sober</wf> <wf cmd="done" pos="NN" lemma="commentator" wnsn="1" lexsn="1:18:01::">commentators</wf> <punc>(</punc> <wf cmd="done" pos="RB" lemma="see" wnsn="0" lexsn="4:02:00::">see</wf> <wf cmd="done" rdf="person" pos="NNP" lemma="person" wnsn="1" lexsn="1:03:00::" pn="person">Alan_Clark</wf> <wf cmd="ignore" pos="IN">on</wf> <wf cmd="done" pos="NN" ot="notag">p.</wf> <wf cmd="done" pos="CD" ot="notag">367</wf> <punc>)</punc> <wf cmd="done" pos="VBP" ot="notag">have</wf> <wf cmd="done" rdf="spell_out" pos="VB" lemma="spell_out" wnsn="1" lexsn="2:32:00::">spelled</wf> <wf cmd="ignore" pos="PRP">them</wf> <wf cmd="done" pos="IN" ot="notag">out</wf> <wf cmd="done" pos="RB" lemma="again_and_again" wnsn="1" lexsn="4:02:00::">again_and_again</wf> <punc>.</punc> </s>
There appear to be POS errors: see/RB and out/IN. Not sure why "p." (page) isn't sense-tagged.
Multiword units: In the example, note the 3 named entities, the idiomatic expressions, and the discontinuous verb particle construction spell_out. There does not appear to be much use of multiword expressions for values (a few dates had it; most did not) or light verb - object constructions like 'make - request'.
The cxtfile documentation in the (old) SemCor 1.6 release explains the markup format. The most pertinent details are as follows:
… The semantic tagging was done by hand, using various tools to annotate the English text with WordNet senses. The "raw" data were reformatted and syntactically tagged before semantic tags were assigned. …
In a letter to Georgia Gulforg:corporation
Presidentper_desc
Jerry R. Satrumperson
,
Mr. Martinperson
asked Georgia Gulforg:corporation
to answer its offer by Tuesdaydate:date
.
It was n't clear how NLorg:corporation
and Mr.
Simmonsperson
would respond if
Georgia Gulforg:corporation
spurns them again .
Excerpt from Coref-EntityType-Corpus/data/WSJtypes-subtypes/wsj00d.qa (line wrapping added for readability):
<DOC> <DOCNO> WSJ0080 </DOCNO> ... In a letter to <ENAMEX TYPE="ORGANIZATION:CORPORATION">Georgia Gulf</ENAMEX> <ENAMEX TYPE="PER_DESC">President</ENAMEX> <ENAMEX TYPE="PERSON">Jerry R. Satrum</ENAMEX> , Mr. <ENAMEX TYPE="PERSON">Martin</ENAMEX> asked <ENAMEX TYPE="ORGANIZATION:CORPORATION">Georgia Gulf</ENAMEX> to answer its offer by <TIMEX TYPE="DATE:DATE">Tuesday</TIMEX> . It was n't clear how <ENAMEX TYPE="ORGANIZATION:CORPORATION">NL</ENAMEX> and Mr. <ENAMEX TYPE="PERSON">Simmons</ENAMEX> would respond if <ENAMEX TYPE="ORGANIZATION:CORPORATION">Georgia Gulf</ENAMEX> spurns them again . ... </DOC>
ANIMAL 396 CONTACT_INFO:ADDRESS 4 CONTACT_INFO:OTHER 3 CONTACT_INFO:PHONE 33 DISEASE 317 EVENT:HURRICANE 104 EVENT:OTHER 220 EVENT:WAR 47 FAC:AIRPORT 34 FAC:ATTRACTION 13 FAC:BRIDGE 41 FAC:BUILDING 154 FAC:HIGHWAY_STREET 116 FAC:HOTEL 1 FAC:OTHER 87 FAC_DESC:AIRPORT 53 FAC_DESC:ATTRACTION 33 FAC_DESC:BRIDGE 40 FAC_DESC:BUILDING 1808 FAC_DESC:HIGHWAY_STREET 205 FAC_DESC:OTHER 430 FAC_DESC:STREET_HIGHWAY 1 GAME 90 GPE:CITY 5601 GPE:COUNTRY 5079 GPE:OTHER 192 GPE:STATE_PROVINCE 2694 GPE_DESC:CITY 377 GPE_DESC:COUNTRY 992 GPE_DESC:OTHER 69 GPE_DESC:STATE_PROVINCE 397 LANGUAGE 84 LAW 382 LOCATION 5 LOCATION:BORDER 1 LOCATION:CITY 1 LOCATION:CONTINENT 256 LOCATION:LAKE_SEA_OCEAN 80 LOCATION:OTHER 181 LOCATION:REGION 526 LOCATION:RIVER 39 NORP:NATIONALITY 3238 NORP:OTHER 91 NORP:POLITICAL 677 NORP:RELIGION 88 ORGANIZATION:CITY 2 ORGANIZATION:CORPORATION 23436 ORGANIZATION:EDUCATIONAL 366 ORGANIZATION:GOVERNMENT 4628 ORGANIZATION:HOSPITAL 23 ORGANIZATION:HOTEL 60 ORGANIZATION:MUSEUM 14 ORGANIZATION:OTHER 1254 ORGANIZATION:POLITICAL 413 ORGANIZATION:RELIGIOUS 44 ORGANIZATION:STATE_PROVINCE 1 ORG_DESC:CORPORATION 15186 ORG_DESC:EDUCATIONAL 238 ORG_DESC:GOVERNMENT 2501 ORG_DESC:HOSPITAL 55 ORG_DESC:HOTEL 56 ORG_DESC:MUSEUM 20 ORG_DESC:OTHER 1191 ORG_DESC:POLITICAL 151 ORG_DESC:RELIGIOUS 51 PERSON 13740 PER_DESC 26346 PLANT 194 PRODCUT:OTHER 1 PRODUCT:DRUG 2 PRODUCT:FOOD 1 PRODUCT:OTHER 519 PRODUCT:VEHICLE 382 PRODUCT:WEAPON 21 PRODUCT_DESC:OTHER 26 PRODUCT_DESC:VEHICLE 1223 PRODUCT_DESC:WEAPON 132 SUBSTANCE:CHEMICAL 529 SUBSTANCE:DRUG 439 SUBSTANCE:FOOD 888 SUBSTANCE:NUCLEAR 3 SUBSTANCE:OTHER 850 WORK_OF_ART:BOOK 123 WORK_OF_ART:OTHER 515 WORK_OF_ART:PAINTING 13 WORK_OF_ART:PLAY 31 WORK_OF_ART:SONG 39
In a letter to Georgia Gulf President Jerry R. Satrum ,
Mr. Martin asked Georgia Gulf
to answer its offer by Tuesday .
It was n't clear how NL
and Mr.
Simmons
would respond if
Georgia Gulf spurns them again .
In Coref-EntityType-Corpus/data/BBN-wsj-pronouns/WSJ.pron:
(WSJ0080 ... ( Antecedent -> S20:15-16 -> Georgia Gulf Pronoun -> S20:19-19 -> its ) ( Antecedent -> S21:9-9 -> Simmons Antecedent -> S21:6-6 -> NL Pronoun -> S22:10-10 -> their Pronoun -> S22:4-4 -> they Pronoun -> S21:16-16 -> them ) ... )
This indicates, for instance, that "Simmons" and "NL" are jointly antecedents of the designated third person plural pronouns. Note that some of these pronoun-antecedent links cross sentence boundaries. Note also that the "it" in S21 is non-referential.
(The marking of "Georgia Gulf" as the antecedent of "its" in S20 actually appears to be incorrect when considered in context.)
VerbNet is an English verb lexicon organized by Levin's taxonomy of semantic classes. It uses thematic roles (such as Agent, Instrument, and Destination) rather than verb- or frame-specific roles. Thematic roles are sometimes decorated with selectional restrictions like [+animate]. The lexicon files include mappings to WordNet entries and OntoNotes senses ("groupings").
The SemLink project builds upon VerbNet, providing:
In a letter to Georgia Gulf President Jerry R. SatrumARGM-LOC
,
Mr. MartinAgent
askedvp--atransfer_mesg-37.1-1
Georgia GulfRecipient
to answer its offer by TuesdayTopic
.
In a letter to Georgia Gulf President Jerry R. Satrum ,
Mr. Martin
asked
Georgia GulfARG0
to
answeri---aanswer.01
its offerARG1
by TuesdayARGM-TMP
.
Note that the verb answer has no VerbNet mapping, so it falls back to the PropBank annotation.
Example frame annotation from verbnet-3.2/transfer_mesg-37.1.1.xml (an inconsistency: VerbNet 3.2 does not list ask as a member of this frame, even though it appears in an example):
<VNCLASS ID="transfer_mesg-37.1.1" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="vn_schema-3.xsd"> <MEMBERS> <!--REMOVED "cite" misclassified--> <MEMBER name="demonstrate" wn="demonstrate%2:39:01" grouping="demonstrate.02"/> <MEMBER name="elucidate" wn="elucidate%2:32:00 elucidate%2:31:00" grouping=""/> <MEMBER name="explain" wn="explain%2:32:00 explain%2:32:01" grouping="explain.01"/> <MEMBER name="explicate" wn="explicate%2:32:00" grouping=""/> <MEMBER name="expound" wn="expound%2:32:00" grouping=""/> <MEMBER name="justify" wn="justify%2:32:02" grouping="justify.01"/> <MEMBER name="narrate" wn="narrate%2:32:00" grouping="narrate.02"/> <MEMBER name="pose" wn="?pose%2:42:00" grouping="pose.01"/> <MEMBER name="preach" wn="preach%2:32:02" grouping="preach.01"/> <MEMBER name="recite" wn="recite%2:32:00 recite%2:36:00 recite%2:32:01" grouping="recite.01"/> <MEMBER name="relay" wn="relay%2:32:00" grouping="relay.01"/> <MEMBER name="sniff" wn="" grouping="sniff.03"/> <MEMBER name="illustrate" wn="illustrate%2:32:00" grouping=""/> <MEMBER name="verify" wn="verify%2:31:01 verify%2:32:03" grouping=""/> <MEMBER name="corroborate" wn="corroborate%2:32:00" grouping=""/> <MEMBER name="communicate" wn="communicate%2:32:01 communicate%2:32:00" grouping=""/> <MEMBER name="outline" wn="outline%2:32:00" grouping=""/> <MEMBER name="summarize" wn="summarize%2:32:00" grouping=""/> <MEMBER name="sum_up" wn="sum_up%2:32:00" grouping=""/> <MEMBER name="boil_down" wn="boil_down%2:30:03" grouping=""/> <MEMBER name="snitch" wn="snitch%2:32:09" grouping=""/> </MEMBERS> <THEMROLES> <!--intersective members use Theme for Topic--> <THEMROLE type="Agent"> <SELRESTRS logic="or"> <SELRESTR Value="+" type="animate"/> <SELRESTR Value="+" type="organization"/> </SELRESTRS> </THEMROLE> <THEMROLE type="Topic"> <SELRESTRS/> </THEMROLE> <THEMROLE type="Recipient"> <SELRESTRS logic="or"> <SELRESTR Value="+" type="animate"/> <SELRESTR Value="+" type="organization"/> </SELRESTRS> </THEMROLE> <THEMROLE type="Source"> <SELRESTRS/> </THEMROLE> </THEMROLES> <FRAMES> ... </FRAMES> <SUBCLASSES> <VNSUBCLASS ID="transfer_mesg-37.1.1-1"> <MEMBERS> <MEMBER name="dictate" wn="dictate%2:31:00" grouping="dictate.02"/> <MEMBER name="quote" wn="quote%2:32:00" grouping="quote.01"/> <MEMBER name="read" wn="read%2:31:01" grouping="read.04"/> </MEMBERS> <THEMROLES/> <FRAMES> ... </FRAMES> <SUBCLASSES> <VNSUBCLASS ID="transfer_mesg-37.1.1-1-1"> <MEMBERS> <MEMBER name="show" wn="show%2:32:01" grouping="show.01"/> <MEMBER name="teach" wn="teach%2:32:00" grouping="teach.01"/> <MEMBER name="tell" wn="tell%2:32:04 tell%2:32:00 tell%2:32:01" grouping="tell.01"/> <MEMBER name="write" wn="write%2:32:00 write%2:32:08 write%2:32:01" grouping="write.01"/> </MEMBERS> <THEMROLES/> <FRAMES> ... <FRAME> <DESCRIPTION descriptionNumber="8.1" primary="NP V NP S_INF" secondary="NP-TO-INF-OC" xtag="0.2"/> <EXAMPLES> <EXAMPLE>Wanda asked me to present a new proposal.</EXAMPLE> </EXAMPLES> <SYNTAX> <NP value="Agent"> <SYNRESTRS/> </NP> <VERB/> <NP value="Recipient"> <SYNRESTRS/> </NP> <NP value="Topic"> <SYNRESTRS> <SYNRESTR Value="+" type="oc_to_inf"/> </SYNRESTRS> </NP> </SYNTAX> <SEMANTICS> <PRED value="transfer_info"> <ARGS> <ARG type="Event" value="during(E)"/> <ARG type="ThemRole" value="Agent"/> <ARG type="ThemRole" value="Recipient"/> <ARG type="ThemRole" value="Topic"/> </ARGS> </PRED> <PRED value="cause"> <ARGS> <ARG type="ThemRole" value="Agent"/> <ARG type="Event" value="E"/> </ARGS> </PRED> </SEMANTICS> </FRAME> ... </FRAMES> <SUBCLASSES/> </VNSUBCLASS> </SUBCLASSES> </VNSUBCLASS> </SUBCLASSES> </VNCLASS>
The corpus annotations (semlink1.1/vn-pb/vnpbprop.txt) augment those of PropBank:
wsj/00/wsj_0080.mrg 19 13 auto ask.02;VN=37.1-1 vp--a 0:1-ARGM-LOC 11:1-ARG0[Agent] 13:0-rel 14:1-ARG2[Recipient] 16:2-ARG1[Topic] wsj/00/wsj_0080.mrg 19 18 auto answer.01;VN=None i---a 14:1*16:0-ARG0 18:0-rel 19:1-ARG1 21:1-ARGM-TMP
In a letter to Georgia Gulf President Jerry R. SatrumARGM-LOC
,
Mr. MartinARG0
askedvp--aask.02
Georgia GulfARG2
to answer its offer by TuesdayARG1
.
In a letter to Georgia Gulf President Jerry R. Satrum ,
Mr. Martin
asked
Georgia GulfARG0
to
answeri---aanswer.01
its offerARG1
by TuesdayARGM-TMP
.
JohnARG0
translatedvp--atranslate.01
his dissertationARG1
from
EnglishARG3-from
into
Swahili, Chinese, Russian, and Yiddish.ARG2-into
Example frame annotation from propbank/frames/translate.xml:
<frameset> <predicate lemma="translate"> <note> Frames file for 'translate' based on sentences in financial subcorpus. No Verbnet entry. Comparison with 'change'. </note> <roleset id="translate.01" name="change (language)" vncls="-"> <roles> <role descr="causer of transformation" n="0"/> <role descr="thing changing" n="1"/> <role descr="end state" n="2"/> <role descr="start state" n="3"/> </roles> ... <example name="all args"> <inflection aspect="ns" form="full" person="ns" tense="past" voice="active"/> <text> John translated his dissertation from English into Swahili, Chinese, Russian, and Yiddish. </text> <arg n="0">John</arg> <rel>translated</rel> <arg n="1">his dissertation</arg> <arg f="from" n="3">English</arg> <arg f="into" n="2">Swahili, Chinese, Russian, and Yiddish.</arg> </example> </roleset> </predicate> </frameset>
Note that the actual annotations (propbank/prop.txt) are with respect to Penn Treebank trees and use a terse machine-readable notation:
wsj/00/wsj_0080.mrg 19 13 gold ask.02 vp--a 0:1-ARGM-LOC 11:1-ARG0 13:0-rel 14:1-ARG2 16:2-ARG1 wsj/00/wsj_0080.mrg 19 18 gold answer.01 i---a 14:1*16:0-ARG0 18:0-rel 19:1-ARG1 21:1-ARGM-TMP
Inflections are specified with 5 fields; their possible values are:
Thus, pnp3a indicates a present perfect third person active participle, e.g. has gone. (A filler of - means the inflectional feature is unspecified.)
InSupp a
letterletter.01
to Georgia Gulf President Jerry R. SatrumARG2
,
Mr. MartinARG0
askedSupp Georgia Gulf to answer its offer by Tuesday .
In a letter to Georgia Gulf President Jerry R. Satrum ,
Mr. Martin
asked
Georgia GulfARG3
to answerSupp
itsARG0
offeroffer.01
by Tuesday .
Mr. Wilbur 'sARG0
translationARG2-REFtranslation.01
Example frame annotation from nombank.1.0/frames/translation.xml:
<frameset> <predicate lemma="translation"> <roleset id="translation.01" name="change (language" source="verb-translate.01"> <roles> <role descr="causer of transformation" n="0"></role> <role descr="thing changing" n="1"></role> <role descr="end state" n="2"></role> <role descr="start state" n="3"></role> </roles> ... <example name="autogen2"> <text> Mr. Wilbur 's translation </text> <arg n="0">Mr. Wilbur 's</arg> <arg f="REF" n="2">translation</arg> <rel>translation</rel> </example> </roleset> </predicate> </frameset>
Note that the actual annotations (nombank.1.0/nombank.1.0) are with respect to Penn Treebank trees and use a terse machine-readable notation:
wsj/00/wsj_0080.mrg 19 2 letter 01 0:0,13:0-Support 2:0-rel 3:1-ARG2 11:1-ARG0 wsj/00/wsj_0080.mrg 19 20 offer 01 14:1*16:0-ARG3 18:0-Support 19:0-ARG0 20:0-rel
In a letter to Georgia Gulforg
President Jerry R. Satrumperson
,
Mr. Martinperson
asked Georgia Gulforg
to answer its offer by Tuesdaydate
.
OntoNotes includes updates to PropBank that have been incorporated since the standalone PropBank release. In addition to updating the proposition annotations to be compatible with the OntoNotes-style syntactic trees, the WSJ PropBank data in OntoNotes incorporates some changes: see the new PropBank annotation manual. Most of the other subcorpora have verb propositions as well, though coverage varies.
Expected in OntoNotes 5.0; will be loosely based on NomBank.
There is an OntoNotes-specific sense inventory for content words that was designed so as to achieve high inter-annotator agreement; essentially, closely related WordNet senses for lexical items were merged until the level of granularity was conducive to consistent tagging. Resulting senses contain mappings to the original WordNet senses as well as to PropBank rolesets and entries in other lexical resources (FrameNet and VerbNet). It is important to note that OntoNotes provides both PropBank annotations and OntoNotes sense annotations for verbs, and that these use a different numbering scheme.
While reported sense coverage is high for some subcorpora, this includes unannotated instances of monosemous nouns and verbs; annotated sense coverage is considerably lower.
Some related OntoNotes senses are further grouped into sense pools which are then mapped to nodes in the Omega ontology. However, the number of senses belonging to a sense pool is currently low, especially for nouns.
Thedt hordesnns ofin peoplenns whowp attendedVVD HooverNP DamNP 'sPOS 1935cd dedicationnn setVVD thedt citynn 'sPOS now-formidablejj pubicjj relationsnns machinenn intoin actionnn .sent
The hordes of people who attended
Hoover Damlocation
's 1935date
dedication set the city 's now-formidable public relations machine into action .
The hordesAgentAggregate (horde.n)
of people who attended Hoover Dam 's 1935 dedicationIndividualsGF=Dep PT=PP
set the city 's now-formidable public relations machine into action .
The hordes of peoplePersonPeople (people.n)
who attended Hoover Dam 's 1935 dedicationContext_of_acquaintanceGF=Dep PT=Srel
set the city 's now-formidable public relations machine into action .
The hordes of people AgentGF=Ext PT=NP Ant
whoAgentGF=Ext PT=NP Rel
attendedAttending (attend.v)
Hoover Dam 's 1935 dedicationEventGF=Obj PT=NP
set the city 's now-formidable public relations machine into action .
The hordes of people who attended Hoover Dam 's 1935 dedication set
the city 'sUserGF=Gen PT=Poss
now-formidableDescriptorGF=Dep PT=AJP
public relationsUseGF=Dep PT=N
machineGizmoGizmo (machine.n) metaphor
into action .
The hordes of people who attended Hoover Dam 's 1935 dedication set
the city 's now-formidable public relations machineAgentGF=Ext PT=NP
intoSupp
actionActIntentionally_act (action.n)
.
The hordes of people who attended Hoover Dam 's 1935 dedication set
the cityLocalePolitical_locales (city.n)
's now-formidable public relations machine into action .
From ANC__HistoryOfLasVegas.xml:
<sentence corpID="195" docID="23691" sentNo="1" paragNo="25" aPos="0" ID="4101098"> <text>The hordes of people who attended Hoover Dam 's 1935 dedication set the city 's now-formidable public relations machine into action .</text> <annotationSet cDate="03/12/2007 02:47:03 PDT Mon" status="UNANN" ID="6541891"> <layer rank="1" name="PENN"> <label end="2" start="0" name="dt"/> <label end="9" start="4" name="nns"/> <label end="12" start="11" name="in"/> <label end="19" start="14" name="nns"/> <label end="23" start="21" name="wp"/> <label end="32" start="25" name="VVD"/> <label end="39" start="34" name="NP"/> <label end="43" start="41" name="NP"/> <label end="46" start="45" name="POS"/> <label end="51" start="48" name="cd"/> <label end="62" start="53" name="nn"/> <label end="66" start="64" name="VVD"/> <label end="70" start="68" name="dt"/> <label end="75" start="72" name="nn"/> <label end="78" start="77" name="POS"/> <label end="93" start="80" name="jj"/> <label end="100" start="95" name="jj"/> <label end="110" start="102" name="nns"/> <label end="118" start="112" name="nn"/> <label end="123" start="120" name="in"/> <label end="130" start="125" name="nn"/> <label end="132" start="132" name="sent"/> </layer> <layer rank="1" name="NER"> <label end="51" start="48" name="date"/> <label end="43" start="34" name="location"/> </layer> <layer rank="1" name="WSL"> <label end="2" start="0" name="NT"/> <label end="12" start="11" name="NT"/> <label end="23" start="21" name="NT"/> <label end="39" start="34" name="NT"/> <label end="43" start="41" name="NT"/> <label end="46" start="45" name="NT"/> <label end="51" start="48" name="NT"/> <label end="70" start="68" name="NT"/> <label end="78" start="77" name="NT"/> <label end="123" start="120" name="NT"/> <label end="132" start="132" name="NT"/> </layer> </annotationSet> <annotationSet cDate="03/04/2008 03:26:24 PST Tue" luID="2886" luName="horde.n" frameID="103" frameName="Aggregate" status="MANUAL" ID="6551750"> <layer rank="1" name="Target"> <label cBy="361" end="9" start="4" name="Target"/> </layer> <layer rank="1" name="FE"> <label cBy="361" feID="745" bgColor="9400D3" fgColor="FFFFFF" end="9" start="4" name="Aggregate"/> <label cBy="361" feID="744" bgColor="FF0000" fgColor="FFFFFF" end="62" start="11" name="Individuals"/> </layer> <layer rank="1" name="GF"> <label end="62" start="11" name="Dep"/> </layer> <layer rank="1" name="PT"> <label end="62" start="11" name="PP"/> </layer> <layer rank="1" name="Other"/> <layer rank="1" name="Sent"/> <layer rank="1" name="Noun"/> </annotationSet> <annotationSet cDate="03/04/2008 03:26:51 PST Tue" luID="5511" luName="people.n" frameID="304" frameName="People" status="MANUAL" ID="6551751"> <layer rank="1" name="Target"> <label cBy="361" end="19" start="14" name="Target"/> </layer> <layer rank="1" name="FE"> <label cBy="361" feID="2618" bgColor="FF0000" fgColor="FFFFFF" end="19" start="14" name="Person"/> <label cBy="361" feID="3561" bgColor="FFA500" fgColor="000000" end="62" start="21" name="Context_of_acquaintance"/> </layer> <layer rank="1" name="GF"> <label end="62" start="21" name="Dep"/> </layer> <layer rank="1" name="PT"> <label end="62" start="21" name="Srel"/> </layer> <layer rank="1" name="Other"/> <layer rank="1" name="Sent"/> <layer rank="1" name="Noun"/> </annotationSet> <annotationSet cDate="03/04/2008 04:05:21 PST Tue" luID="14821" luName="attend.v" frameID="2030" frameName="Attending" status="MANUAL" ID="6551755"> <layer rank="1" name="Target"> <label cBy="361" end="32" start="25" name="Target"/> </layer> <layer rank="1" name="FE"> <label cBy="361" feID="12330" bgColor="0000FF" fgColor="FFFFFF" end="62" start="34" name="Event"/> <label cBy="361" feID="12329" bgColor="FF0000" fgColor="FFFFFF" end="19" start="0" name="Agent"/> <label cBy="361" feID="12329" bgColor="FF0000" fgColor="FFFFFF" end="23" start="21" name="Agent"/> </layer> <layer rank="1" name="GF"> <label end="62" start="34" name="Obj"/> <label end="19" start="0" name="Ext"/> <label end="23" start="21" name="Ext"/> </layer> <layer rank="1" name="PT"> <label end="62" start="34" name="NP"/> <label end="19" start="0" name="NP"/> <label end="23" start="21" name="NP"/> </layer> <layer rank="1" name="Other"> <label end="23" start="21" name="Rel"/> <label end="19" start="0" name="Ant"/> </layer> <layer rank="1" name="Sent"/> <layer rank="1" name="Verb"/> </annotationSet> <annotationSet cDate="03/04/2008 04:22:03 PST Tue" luID="10127" luName="machine.n" frameID="106" frameName="Gizmo" status="MANUAL" ID="6551756"> <layer rank="1" name="Target"> <label cBy="361" end="118" start="112" name="Target"/> </layer> <layer rank="1" name="FE"> <label cBy="361" feID="9677" bgColor="0000FF" fgColor="FFFFFF" end="118" start="112" name="Gizmo"/> <label cBy="361" feID="525" bgColor="800080" fgColor="FFFFFF" end="110" start="95" name="Use"/> <label cBy="361" feID="11959" bgColor="FF69B4" fgColor="FFFFFF" end="93" start="80" name="Descriptor"/> <label cBy="361" feID="526" bgColor="FF0000" fgColor="FFFFFF" end="78" start="68" name="User"/> </layer> <layer rank="1" name="GF"> <label end="110" start="95" name="Dep"/> <label end="93" start="80" name="Dep"/> <label end="78" start="68" name="Gen"/> </layer> <layer rank="1" name="PT"> <label end="110" start="95" name="N"/> <label end="93" start="80" name="AJP"/> <label end="78" start="68" name="Poss"/> </layer> <layer rank="1" name="Other"/> <layer rank="1" name="Sent"> <label end="0" start="0" name="Metaphor"/> </layer> <layer rank="1" name="Noun"/> </annotationSet> <annotationSet cDate="03/04/2008 04:23:07 PST Tue" luID="10611" luName="action.n" frameID="198" frameName="Intentionally_act" status="MANUAL" ID="6551757"> <layer rank="1" name="Target"> <label cBy="361" end="130" start="125" name="Target"/> </layer> <layer rank="1" name="FE"> <label cBy="361" feID="1609" bgColor="0000FF" fgColor="FFFFFF" end="130" start="125" name="Act"/> <label cBy="361" feID="1610" bgColor="FF0000" fgColor="FFFFFF" end="118" start="68" name="Agent"/> </layer> <layer rank="1" name="GF"> <label end="118" start="68" name="Ext"/> </layer> <layer rank="1" name="PT"> <label end="118" start="68" name="NP"/> </layer> <layer rank="1" name="Other"/> <layer rank="1" name="Sent"/> <layer rank="1" name="Noun"> <label end="123" start="120" name="Supp"/> </layer> </annotationSet> <annotationSet cDate="04/29/2011 02:48:39 PDT Fri" luID="3764" luName="city.n" frameID="195" frameName="Political_locales" status="MANUAL" ID="6562548"> <layer rank="1" name="Target"> <label cBy="571" end="75" start="72" name="Target"/> </layer> <layer rank="1" name="FE"> <label cBy="571" feID="1583" bgColor="00008B" fgColor="FFFFFF" end="75" start="72" name="Locale"/> </layer> <layer rank="1" name="GF"/> <layer rank="1" name="PT"/> <layer rank="1" name="Other"/> <layer rank="1" name="Sent"/> <layer rank="1" name="Noun"/> </annotationSet> </sentence>
Nathan Schneider, 2012-08-25. Feel free to contact me with fixes, updates, and additions. Thanks to Daniel Bauer and Liane Guillou for elucidating some of the nitty-gritty details.