be-destined-for
be-destined-for.91 - reification of :destination
For AMR annotations.Reification added to support AMR annotations.
Aliases:
be-destined (v.)
be-destined-for (v.)
have-destination (v.)
be-destined-for (v.)
have-destination (v.)
Roles:
ARG1-PPT: entity
ARG2-GOL: destination
ARG2-GOL: destination
off to
IARG1
am offrel
torel
AtlantaARG2
.(o / off-09 :ARG1 (i / i) :ARG2 (c / city :name (n / name :op1 "Atlanta")))
be-from
be-from.91 - reification of :source
For AMR annotations.Reification added to support AMR annotations.
Aliases:
be-from (v.)
have-source (v.)
have-source (v.)
Roles:
ARG1-VSP: entity
ARG2-VSP: source
ARG2-VSP: source
from Ipanema
The girlARG1
is fromrel
IpanemaARG2
.(b / be-from-91 :ARG1 (g / girl) :ARG2 (c / city :name (n / name :op1 "Ipanema")))
be-located-at
be-located-at.91 - reification of :location
For AMR annotations.Aliases:
have-location (m.)
be-located-at (m.)
be-located-at (m.)
Roles:
ARG1-PPT: entity
ARG2-LOC: location
ARG2-LOC: location
not here
SheARG1
isrel
not hereARG2
.be-polite
be-polite.91 - reification of :polite.
For AMR, UMR annotations.comparison to AMR-2019: same roleset, same usage.
Aliases:
be-polite (m.)
have-polite (m.)
have-politeness (m.)
have-polite (m.)
have-politeness (m.)
Roles:
ARG1-VSP: entity
ARG2-VSP: value (typical value: +)
ARG2-VSP: value (typical value: +)
reification of English sentence
Pleaserel
goARG1
! [UMR: +ARG2
](g / go-02 :mode imperative :ARG0 (y / you) :ARG1-of (b / be-polite-91 :ARG2 +))
be-temporally-at
be-temporally-at.91 - reification of :time in AMR
For AMR annotations.Aliases:
be-temporally-at (m.)
have-temporal (m.)
have-temporal (m.)
Roles:
ARG1-VSP: event
ARG2-TMP: time
ARG2-TMP: time
on Saturday
The beach partyARG1
is on SaturdayARG2
.(b / be-temporally-at-91 :ARG1 (p / party :place (b2 / beach)) :ARG2 (d / date-entity :weekday "Saturday"))
belong
belong.91 - for nonprototypical clauses: predicative possession
For UMR annotations.comparison to AMR-2019: belong-01 in AMR has ARG0/ARG1, which is shifted here to ARG1/ARG2. Belong-01 will still be used in English AMR for lemma-specific instances.
Aliases:
be (v.)
Roles:
ARG1-VSP: possession
ARG2-VSP: possessor
ARG2-VSP: possessor
English: copular clause with possessive pronoun
The dog is his .
binding-affinity
binding-affinity.91 - quantification of how strong molecules bind to each other
Normalize Kd to dissociation-constant.Normalize Ki to inhibitor-constant.
Typical usage: bind-01 :ARG4 binding-affinity-91
Typical values for binding-affinity-91 :ARG1 include: acid-dissociation-constant (Ka), affinity-constant (Ka), dissociation-constant (Kd), inhibitor-constant (Ki), michaelis-constant (Km), power-of-dissociation-constant (pKd), power-of-inhibitor-constant (pKi)
Aliases:
binding-affinity (n.)
Roles:
ARG1-PPT: binding affinity constant measured (e.g. dissociation-constant (Kd), inhibitor-constant (Ki))
ARG2-PRD: binding affinity quantity (e.g. concentration-quantity such as '5 nanomolar')
ARG2-PRD: binding affinity quantity (e.g. concentration-quantity such as '5 nanomolar')
Kd
Gefitinib binds wild - type EGFR kinase with
KdARG1
= 53.5 nM.ARG2
(b / bind-01 :ARG1 (s / small-molecule :wiki "Gefitinib" :name (n / name :op1 "gefitinib")) :ARG2 (k / kinase :wiki "Epidermal_growth_factor_receptor" :name (n2 / name :op1 "EGFR") :mod (w / wild-type)) :ARG4 (b2 / binding-affinity-91 :ARG1 (d / dissociation-constant) :ARG2 (c / concentration-quantity :quant 53.5 :unit (n3 / nanomolar))))
Ki
TKI258 binds tightly to the FGFR1
KDARG1
V561 M ( Ki ~ 35ARG2
nM ) .(b / bind-01 :ARG1 (s / small-molecule :wiki - :name (n / name :op1 "TKI258")) :ARG2 (e / enzyme :wiki "Fibroblast_growth_factor_receptor_1" :name (n2 / name :op1 "FGFR1") :ARG1-of (k / knock-down-02) :ARG2-of (m / mutate-01 :value "V561M")) :ARG4 (b2 / binding-affinity-91 :ARG1 (i2 / inhibitor-constant) :ARG2 (a / approximately :op1 (c / concentration-quantity :quant 35 :unit (n3 / nanomolar)))) :ARG4 (t / tight-05))
but
but.91 - discourse relation for contrastive events underspecified for (un)expectedness
For UMR annotations.comparison to AMR-2019: but-91 is generic for pure-contrast/unexpected-co-occurrence in the discourse relation lattice. It may cover some of what contrast-01 was used for in AMR (no arg changes), but maybe be quite infrequent in English.
Aliases:
but (m.)
Roles:
ARG1-VSP: first event
ARG2-VSP: second event
ARG2-VSP: second event
byline
byline.91 - for byline information
For AMR, UMR annotations.comparison to AMR-2019: same roleset, same usage.
Aliases:
byline (n.)
by-line (n.)
dateline (n.)
date-line (n.)
by-line (n.)
dateline (n.)
date-line (n.)
Roles:
ARG0-VSP: news organization
ARG1-VSP: author
ARG2-VSP: photographer, illustrator
ARG3-VSP: translator
ARG4-VSP: means
ARG1-VSP: author
ARG2-VSP: photographer, illustrator
ARG3-VSP: translator
ARG4-VSP: means
Xinhua byline
Xinhua New AgencyARG0
, Shanghai:place
, February 10:temporal
, by wireARG4
( reporters Jinhu Xie and Chijian ZhangARG1
)(b / byline-91 :ARG0 (o / organization :name (n / name :op1 "Xinhua" :op2 "New" :op3 "Agency")) :ARG1 (a / and :op1 (p / person :name (n2 / name :op1 "Jinhu" :op2 "Xie")) :op2 (p2 / person :name (n3 / name :op1 "Chijian" :op2 "Zhang")) :ARG0-of (r / report-01)) :ARG4 (w / wire) :time (d / date-entity :month 2 :date 10) :location (c / city :name (n4 / name :op1 "Shanghai")))
cite
cite.91 - for citations
For UMR annotations.comparison to AMR-2019: Lemma-specific cite-01 was used in AMR, but in order to get the arguments to fit, additional intermediate nodes between the cited text and the cite-01 predicate had to be included: ([cited-text] :ARG1-of (a / attest-01 *or* d / describe-01 :ARG0 (p / publication :ARG1-of (c / cite-01 [...])))) This has been streamlined for UMR by reworking the arguments-- ARG1 is now the cited text, and arg2 is the publication cited.
Aliases:
Roles:
ARG1-PPT: text accompanied by a citation
ARG2-SRC: source publication (may use publication-91)
ARG2-SRC: source publication (may use publication-91)
French military defeats have been studied extensivelyARG1
( 1 - 7 )ARG2
.correlate
correlate.91 - for the X-er the Y-er constructions
For AMR, UMR annotations.comparison to AMR-2019: same roleset, same usage.
Aliases:
the_Xer_the_Yer (m.)
Roles:
ARG1-VSP: first item in correlation
ARG2-VSP: second item in correlation
ARG2-VSP: second item in correlation
the Xer the Yer
The lower the priceARG1
, the more you sellARG2
.(c / correlate-91 :ARG1 (h / have-degree-91 :ARG1 (p / price) :ARG2 (l / low-04) :ARG3 (m / more)) :ARG2 (h2 / have-quant-91 :ARG1 (t / thing :ARG1-of (s / sell-01 :ARG0 (y / you))) :ARG3 (m2 / more)))
contrast
contrast.91 - discourse relation: for pure contrast
For UMR annotations.comparison to AMR-2019: This roleset is identical to contrast-01 used in AMR, just renamed with -91.
Aliases:
but (c.)
Roles:
ARG1-VSP: first item in comparison
ARG2-VSP: second item in comparison
ARG2-VSP: second item in comparison
not unexpected cooccurrence
Petja staratel'nyiARG1
, arel
Vanja lenivyjARG2
.course
course.91 - for school course enrollment information
For AMR, UMR annotations.comparison to AMR-2019: same roleset, same usage.
Aliases:
course (n.)
Roles:
ARG1-VSP: topic
ARG2-VSP: number
ARG3-VSP: section
ARG4-VSP: institution
ARG5-VSP: instructor
ARG2-VSP: number
ARG3-VSP: section
ARG4-VSP: institution
ARG5-VSP: instructor
Economy 101
He enrolled in
EconomyARG1
101ARG2
at HarvardARG4
.(e / enroll-01 :ARG1 (h / he) :ARG2 (c / course-91 :ARG1 (e2 / economy) :ARG2 101 :ARG4 (u / university :name (n / name :op1 "Harvard"))))
distribution-range
distribution-range.91 - distribution range (incl. for a normal distribution)
For AMR, UMR annotations.comparison to AMR-2019: same roleset, same usage.
Aliases:
distribution-range (m.)
distribution (n.)
range (n.)
distribution (n.)
range (n.)
Roles:
ARG1-VSP: center, mean
ARG2-VSP: lower bound
ARG3-VSP: upper bound
ARG4-VSP: radius, distance from center to bounds
ARG5-VSP: confidence that value is in range, typically a percentage, e.g. 95%
ARG6-VSP: deviation covered by range, e.g. standard-deviation, standard-error-of-the-mean
ARG7-VSP: type of distribution, e.g. normal-distribution
ARG2-VSP: lower bound
ARG3-VSP: upper bound
ARG4-VSP: radius, distance from center to bounds
ARG5-VSP: confidence that value is in range, typically a percentage, e.g. 95%
ARG6-VSP: deviation covered by range, e.g. standard-deviation, standard-error-of-the-mean
ARG7-VSP: type of distribution, e.g. normal-distribution
center and radius
142.0ARG1
± 15.7 nMARG4
(d / distribution-range-91 :ARG1 (c / concentration-quantity :quant 142.0 :unit (n / nanomolar)) :ARG4 (c2 / concentration-quantity :quant 15.7 :unit (n2 / nanomolar)))
center and bounds
a
rangerel
of 16.4 %ARG2
- 19 %ARG3
( 17.7 %ARG1
being the average )(d / distribution-range-91 :ARG1 (p / percentage-entity :value 17.7) :ARG2 (p2 / percentage-entity :value 16.4) :ARG3 (p3 / percentage-entity :value 19))
with confidence percentage
With
95 %ARG5
confidence , 42ARG2
- 45 %ARG3
of voters will voterel
for John Smith .(i2 / include-91 :ARG1 (p / person :ARG0-of (v / vote-01 :ARG1 (p2 / person :name (n / name :op1 "John" :op2 "Smith")))) :ARG2 (p3 / person :ARG0-of (v2 / vote-01)) :ARG3 (d / distribution-range-91 :ARG2 (p4 / percentage-entity :value 42) :ARG3 (p5 / percentage-entity :value 45) :ARG5 (p6 / percentage-entity :value 95)))
with standard-deviation
The gestational
agerel
was determined to be 28.5 weeksARG1
± 10 daysARG4
( 2 standard deviations )ARG6
.(d / determine-01 :ARG1 (a / age-01 :ARG1 (p / person :ARG1-of (g / gestate-01)) :ARG2 (d2 / distribution-range-91 :ARG1 (t / temporal-quantity :quant 28.5 :unit (w / week)) :ARG4 (t2 / temporal-quantity :quant 10 :unit (d3 / day)) :ARG6 (p2 / product-of :op1 2 :op2 (s / standard-deviation)))))
with standard-error-of-the-mean
65ARG1
± 3 %ARG4
( meanrel
± SEMARG6
)(d / distribution-range-91 :ARG1 (p / percentage-entity :value 65) :ARG4 (p2 / percentage-entity :value 3) :ARG6 (s / standard-error-of-the-mean))
emit_sound
emit_sound.91 - for transcription of noises that occur outside of quotative predicates
For UMR annotations.comparison to AMR-2019: AMR provided (s / string-entity), but no abstract roleset for use when no quotative verb appears in connection with the noise transcription.
Aliases:
sound (m.)
Roles:
ARG1-PAG: entity emitting sound
ARG2-PPT: the sound (may use (s / string-entity))
ARG2-PPT: the sound (may use (s / string-entity))
sound emission
What do you call those little ... um ... paddleball ? You know , with the ball on a string ?
Chong chong chong chongARG2
.exist
exist.91 - nonprototypical clauses: for thetic location
For UMR annotations.comparison to AMR-2019: AMR used the :location/:location-of roles or their reification, be-located-at-91, for these. This roleset is not the same as exist-01.
Aliases:
be (m.)
Roles:
ARG1-LOC: location
ARG2-PPT: entity
ARG2-PPT: entity
thetic location
On the rockARG1
was a symbolARG2
Chinese example
并且 还 有 很多 高层 的 人物 哦 !
gesture
gesture.91 - a preliminary roleset for incorporating gestures into multimodal UMR graphs
For UMR annotations.comparison to AMR-2019: No handling of multimodal gestures in AMR 2019. This roleset is preliminary, needed to handle certain instances in the Pear Story for the UMR release. Will likely be updated/replaced as we work on the official schema for multimodal UMR.
Aliases:
{gesture} (m.)
Roles:
ARG0-PAG: gesturer
ARG1-PPT: the gesture
ARG2-GOL: audience
ARG1-PPT: the gesture
ARG2-GOL: audience
ELAN transcription with gesture reference
And he goes ... {
countingARG1
gesturerel
}have
have.91 - reification of :poss; nonprototypical clause: thetic possession
For UMR annotations.comparison to AMR-2019: This roleset replaces have-03 as the reification for :poss in UMR. Argument numbers have been shifted from 0/1 to 1/2. have-03 is still to be used for English in lemma-specific cases.
Aliases:
's (m.)
Roles:
ARG1-VSP: possessor
ARG2-VSP: possession
ARG2-VSP: possession
English: reification of :poss
the man 'sARG1
horseARG2
Kukuma example
MijiritinARG1
ɨarayaraARG2
have-actor
have-actor.91 - reification of :actor
For UMR annotations.comparison to AMR-2019: This is a new reification for a new role; neither existed in AMR, as these roles were included as numbered arguments in PropBank rolesets.
Aliases:
have-actor (m.)
Roles:
ARG1-VSP: event
ARG2-PAG: actor
ARG2-PAG: actor
reified version of a standard English sentence
The enemyARG2
sank the shipARG1
.have-affectee
have-affectee.91 - reification of :affectee
For UMR annotations.comparison to AMR-2019: UMR replaces :beneficiary with :affectee, which includes maleficiaries in addition to beneficiaries. For reification, UMR introduces this new roleset which is to be used instead of benefit-01. Benefit-01 had a different argument structure (ARG0-giver, ARG1-benefactive).
Aliases:
have-affectee (m.)
Roles:
ARG1-VSP: event
ARG2-GOL: affectee
ARG2-GOL: affectee
reified version of an English sentence
I went to the storeARG1
for my motherARG2
.reified version of a Falam Chin sentence
ParteARG2
in Thangte hrang ahARG1
hmeh a suangC-ARG2
.have-age
have-age.91 - reification of :age
For UMR annotations.comparison to AMR-2019: This replaces age-01 as the reification of :age. Age-01 is Aspectually different from this role/roleset and contains an extra Arg0-agent that does not apply here. Args 1 and 2 are the same here as in the old roleset.
Aliases:
have-age (m.)
Roles:
ARG1-VSP: entity
ARG2-VSP: age
ARG2-VSP: age
Formerly annotated as age-01
My grandmotherARG1
isrel
85ARG2
.have-apprehensive
have-apprehensive.91 - reification of :apprehensive
For UMR annotations.comparison to AMR-2019: This roleset is new. It covers cases where there are 2 mutually exclusive events and one is carried out with the aim of preventing the other.
Aliases:
have-apprehensive (m.)
Roles:
ARG1-VSP: main event
ARG2-VSP: apprehensive event
ARG2-VSP: apprehensive event
reified English example
Grab a stick
lestrel
he attack youARG2
.have-cause
have-cause.91 - reification of :cause
For UMR annotations.comparison to AMR-2019: :cause was used in AMR as well, but in UMR, what was previously covered by :cause has been split into :cause and :reason. have-cause-91 is now the reification for :cause, and have-reason-91 is now the reification for :reason. Argument structure has been shifted from 0/1 to 1/2 compared to cause-01. Lemma-specific uses ('caused', etc) should still be annotated in English with cause-01.
Aliases:
have-cause (m.)
Roles:
ARG1-VSP: event
ARG2-CAU: cause
ARG2-CAU: cause
reification of :cause
He has no right to beARG1
asrel
we are not togetherARG2
.Causal argument fragment
Becauserel
I said soARG2
.have-causer
have-causer.91 - reification of :causer
For UMR annotations.comparison to AMR-2019: This is a new reification for a new role; neither existed in AMR, as these roles were included as numbered arguments in PropBank rolesets.
Aliases:
have-causer (m.)
Roles:
ARG1-VSP: event
ARG2-PAG: causer
ARG2-PAG: causer
Reified version of a Kukama sentence
naiARG2
kuratata churan ui unipuARG1
Reified version of an English sentence
The rulesARG2
forbade us from smokingARG1
.have-companion
have-companion.91 - reification of :companion
For UMR annotations.comparison to AMR-2019: Accompany-01 was used to reify :accompanier in AMR. :companion replaces :accompanier in UMR and has slightly different boundaries: :companion is limited to animate entities, while :accompanier was not. Have-companion-91 is a new roleset, not just a new name for accompany-01. The annotation strategy is also different, as have-companion-91 includes 2 arguments-- one for the modified event, and one for the accompanier. AMR modified the event with :manner (a / accompany-01 :ARG0-accompanier :ARG1-companion).
Aliases:
have-companion (m.)
Roles:
ARG1-VSP: event
ARG2-VSP: accompanier
ARG2-VSP: accompanier
Sentence fragment
She wentARG1
with her motherARG2
.have-concession
have-concession.91 - reification of :concession
For AMR, UMR annotations.comparison to AMR-2019: same roleset as in AMR, but its coverage is split in UMR between have-concession-91 and have-concessive-condition-91.
Aliases:
have-concession (m.)
Roles:
ARG1-VSP: main event
ARG2-VSP: concession
ARG2-VSP: concession
although it rained
The game continuedARG1
althoughrel
it was rainingARG2
.(h / have-concession-91 :ARG1 (c / continue-01 :ARG1 (g / gaming-02)) :ARG2 (r / rain-01))
have-concessive-condition
have-concessive-condition.91 - reification of :concessive-condition
For UMR annotations.comparison to AMR-2019: new roleset, covers cases where there is an underlying expectation that the conditioning event may not lead to the consequent event.
Aliases:
have-concessive-condition (m.)
Roles:
ARG1-VSP: main event
ARG2-VSP: concessive condition
ARG2-VSP: concessive condition
Concessive
Evenrel
ifrel
you arrive only five minutes lateARG2
, you will be firedARG1
.have-condition
have-condition.91 - reification of :condition
For AMR annotations.comparison to AMR-2019: same roleset.
Aliases:
have-condition (m.)
Roles:
ARG1-VSP: main event
ARG2-VSP: condition
ARG2-VSP: condition
if it rains
The game will be canceledARG1
ifrel
it rainsARG2
.(h / have-condition-91 :ARG1 (c / cancel-01 :ARG1 (g / gaming-02)) :ARG2 (r / rain-01))
have-degree
have-degree.91 - reification of :degree (in AMR), and used for comparatives and degree-based result constructions (AMR/UMR)
For AMR, UMR annotations. Do not use this frame for intensifiers and downtoners (e.g. very tall, a little sad).comparison to AMR-2019: same roleset, same usage.
Aliases:
have-degree (m.)
have-comparison (m.)
have-comparison (m.)
Roles:
ARG1-ADV: domain, entity characterized by attribute (e.g. girl)
ARG2-PPT: attribute (e.g. tall)
ARG3-EXT: degree itself (e.g. more, most, less, least, times, equal, enough, too, so, to-the-point, at-least)
ARG4-PPT: compared-to
ARG5-PRD: superlative, reference to superset
ARG6-PRD: reference, threshold of sufficiency
ARG2-PPT: attribute (e.g. tall)
ARG3-EXT: degree itself (e.g. more, most, less, least, times, equal, enough, too, so, to-the-point, at-least)
ARG4-PPT: compared-to
ARG5-PRD: superlative, reference to superset
ARG6-PRD: reference, threshold of sufficiency
basic comparative
The girl is taller than the boy .
(h / have-degree-91 :ARG1 (g / girl) :ARG2 (t / tall) :ARG3 (m / more) :ARG4 (b / boy))
'as X as' construction example
He runsARG1
asrel
fastARG2
asrel
lightningARG4
. [ equalARG3
](h / have-degree-91 :ARG1 (r / run-02 :ARG1 (h2 / he)) :ARG2 (f / fast) :ARG3 (e / equal) :ARG4 (l / lightning))
polarity examples, with arg6 consequence
It is
toorel
earlyARG2
torel
reach any conclusion as to the motive and identity of the attackers .(h / have-degree-91 :ARG2 (e / early-01) :ARG3 (t / too) :ARG6 (c / conclude-01 :ARG2 (a / and :op1 (t2 / thing :ARG0-of (m / motivate-01 :ARG1 (p / person :ARG0-of (a / attack-01)))) :op2 (i / identity :poss p))))
superlative example, implicit superset made explicit in AMR
She works for UBS , the biggest [UMR:
mostARG3
bigARG2
] bankARG1
in SwitzerlandARG5
.(w / work-01 :ARG0 (s / she) :ARG2 (o / organization :name (n / name :op1 "USB") :domain (b / bank :ARG1-of (h / have-degree-91 :ARG2 (b2 / big) :ARG3 (m / most) :ARG5 (b3 / bank :location (c / country :name (n2 / name :op1 "Switzerland")))))))
another superlative, implicit superset made explicit in AMR
the
mostrel
intelligentARG2
girlARG1
on the teamARG5
(g / girl :ARG1-of (h / have-degree-91 :ARG2 (i / intelligent) :ARG3 (m / most) :ARG5 (g2 / girl :ARG0-of (h2 / have-org-role-91 :ARG1 (t / team)))))
measurable degrees, see guidelines/dict for guidance on distinguishing from have-quant-91
The total of 18 deaths from malignant mesothelioma , lung cancer and asbestosisARG1
was far higher [UMR: moreARG3
highARG2
] than expectedARG4
, the researchers said .(s / say-01 :ARG0 (p / person :ARG0-of (r / research-01)) :ARG1 (h / have-degree-91 :ARG1 (t / thing :ARG2-of (d / die-01 :ARG0-of (c2 / cause-01 :ARG1 (o / or :op1 (m / mesothelioma ARG1-of (m2 / malignant-01)) :op2 (c / cancer-03 :location (l / lung)) :op3 (a / asbestos)))) :ARG1-of (t3 / total-01 :ARG2 18)) :ARG2 (h2 / high-02) :ARG3 (m3 / more) :ARG4 (t2 / thing :ARG1-of (e / expect-01))))
have-degree.92 - reification of :degree for intensifiers/dowtoners/equals
comparison to AMR-2019: same roleset, but in UMR, the values for ARG2 are limited to 'intensifier', 'downtoner', and 'equal'.Aliases:
have-degree (m.)
have-downtoner (m.)
have-intensifier (m.)
have-downtoner (m.)
have-intensifier (m.)
Roles:
ARG1-ADJ: attribute (e.g. tall)
ARG2-VSP: 'Intensifier', 'Downtoner', or 'Equal'
ARG2-VSP: 'Intensifier', 'Downtoner', or 'Equal'
intensifier
I know that he is of
highrel
intelligenceARG1
. [UMR: intensifierARG2
](k / know-01 :ARG0 (i / i) :ARG1 (h / have-degree-92 :ARG1 (i / intelligent-01 :ARG1 (h / he)) :ARG2 Intensifier))
have-degree-of-resemblance
have-degree-of-resemblance.91 - Comparison of competing resemblances; for constructions such as 'X is more like Y than Z'
Aliases:
have-degree-of-resemblance (v.)
be-more-like (v.)
more-like (v.)
more-like-x-than-y (v.)
be-more-like (v.)
more-like (v.)
more-like-x-than-y (v.)
Roles:
ARG1-PPT: Thing resembling things
ARG2-PPT: First resemblance under comparison
ARG3-PPT: Second resemblance under comparison
ARG4-EXT: Degree word comparing ARG2 to ARG3
ARG2-PPT: First resemblance under comparison
ARG3-PPT: Second resemblance under comparison
ARG4-EXT: Degree word comparing ARG2 to ARG3
degree of resemblance with explicit event
They danceARG1
moreARG4
likerel
the natives here thanARG2
normal peopleARG3
.(h / have-degree-of-resemblance-91 :ARG1 (d / dance-01 :ARG0 (t / they)) :ARG2 (d2 / dance-01 :ARG0 (n / native :location (h2 / here))) :ARG3 (d3 / dance-01 :ARG0 (p / person :ARG1-of (n2 / normal-02))) :ARG4 (m / more))
entity arg1
He 'sARG1
moreARG4
likerel
his brother NeilARG2
thanrel
his fatherARG3
.(h / have-degree-of-resemblance-91 :ARG1 (h2 / he) :ARG2 (p / person :wiki "Neil_Bush" :name (n / name :op1 "Neil") :ARG0-of (h3 / have-rel-role-91 :ARG1 h2 :ARG2 (b / brother))) :ARG3 (p2 / person :ARG0-of (h4 / have-rel-role-91 :ARG1 h2 :ARG2 (f / father))) :ARG4 (m / more))
adjectival
And the city decided to
treat its guestsARG1
moreARG4
likerel
royalty or rock starsARG2
thanrel
factory ownersARG3
.(a / and :op2 (d / decide-01 :ARG0 (c / city) :ARG1 (h / have-degree-of-resemblance-91 :ARG1 (t / treat-01 :ARG0 c :ARG1 (g / guest :poss c)) :ARG2 (t2 / treat-01 :ARG1 (o / or :op1 (r / royalty) :op2 (s / star :mod (r2 / rock)))) :ARG3 (t3 / treat-01 :ARG1 (p / person :ARG0-of (o2 / own-01 :ARG1 (f / factory)))) :ARG4 (m / more))))
have-direction
have-direction.91 - reification for :direction
HAVE DIRECTION NOTES: Added for SpatialAMR.comparison to AMR-2019: standard AMR never created a roleset for :direction, but if they had, it would have looked just like this.
Aliases:
have-direction (m.)
Roles:
ARG1-VSP: entity/event
ARG2-DIR: direction
ARG2-DIR: direction
reified English sentence
Move that block
the other wayARG2
.(m / move-01 :mode Imperative :ARG0 (y / you) :ARG1 (b / block :mod (t / that))) :ARG1-of (h / have-direction-91 :ARG2 (w / way :ARG1-of (o / other-01))))
have-duration
have-duration.91 - reification of :duration
For UMR annotations.comparison to AMR-2019: This replaces last-01 as a reification of :duration in UMR. The argument structure and annotation strategies are identical.
Aliases:
have-duration (m.)
Roles:
ARG1-VSP: thing that lasts
ARG2-TMP: period of time
ARG2-TMP: period of time
Reification of English Sentence
Patrick Makau finished the marathonARG1
in 2 hours , 3 minutes and 38 secondsARG2
.have-example
have-example.91 - reification of :example
For UMR annotations.comparison to AMR-2019: replaces exemplify-01 as the reification for :example. Argument structure has been shifted from 0/1 to 1/2.
Aliases:
have-example (m.)
for_example (m.)
for_instance (m.)
for_example (m.)
for_instance (m.)
Roles:
ARG1-VSP: example
ARG2-VSP: of what
ARG2-VSP: of what
example fragment
For example ,
some proteins function as enzymesARG2
.have-experience
have-experience.91 - for reifiying experiential phenomena
For UMR annotations.comparison to AMR-2019: This is a new roleset.
Aliases:
have-experience (m.)
Roles:
ARG1-PAG: experiencer
ARG2-PPT: the experience itself
ARG3-CAU: stimulus
ARG2-PPT: the experience itself
ARG3-CAU: stimulus
clausal experience without lemma-specific roleset
I 'm all verklempt .
'to them, X'
To themARG1
, all men are subjectsARG2
.have-experiencer
have-experiencer.91 - reification of :experiencer
For UMR annotations.comparison to AMR-2019: This is a new reification for a new role; neither existed in AMR, as these roles were included as numbered arguments in PropBank rolesets.
Aliases:
have-experiencer (m.)
Roles:
ARG1-VSP: event
ARG2-PAG: experiencer
ARG2-PAG: experiencer
reified version of an English sentence
YouARG2
had seen the movieARG1
.have-extent
have-extent.91 - reification of :extent
For AMR, UMR annotations.comparison to AMR-2019: same roleset, same usage.
Aliases:
have-extent (m.)
Roles:
ARG1-VSP: event/entity
ARG2-EXT: extent
ARG2-EXT: extent
long trip
I 'm driving across Canada ;
2500 miles and countingARG2
.(d / drive-01 :ARG0 (i / i) :ARG1 (a / across :op1 (c / country :name (s1n / name :op1 "Canada"))) :ARG1-of (h / have-extent-91 :ARG2 (s / sum-of :op1 (d / distance-quantity :quant 2500 :unit (m / mile)) :op2 (d2 / distance-quantity :ARG2-of (c2 / count-01)))))
have-frequency
have-frequency.91 - reification of :frequency
For AMR, UMR annotations.comparison to AMR-2019: same roleset, same usage.
Aliases:
have-frequency (m.)
Roles:
ARG1-PPT: entity
ARG2-TMP: frequency
ARG2-TMP: frequency
earthquakes
In Germany ,
earthquakesARG1
arerel
rareARG2
.(h / have-frequency-91 :ARG1 (e / earthquake-02 :place (c / country :name (n / name :op1 "Germany"))) :ARG2 (r / rare))
reified English sentence
I 've been going to the gymARG1
every other WednesdayARG3
.(g / go-02 :ARG0 (i / i) :ARG4 (g / gym) :ARG1-of (h / have-frequency-91 :ARG2 (r / rate-entity-91 :ARG3 (t / temporal-quantity :quant 2 :unit (w / week)) :ARG4 (d / date-entity :weekday (w2 / wednesday)))))
have-force
have-force.91 - reification of :force
For UMR annotations.comparison to AMR-2019: This is a new reification for a new role; neither existed in AMR, as these roles were included as numbered arguments in PropBank rolesets.
Aliases:
have-force (m.)
Roles:
ARG1-VSP: event
ARG2-CAU: force
ARG2-CAU: force
Reified version of an English sentence
The stormARG2
damaged the power linesARG1
.have-goal
have-goal.91 - reification of :goal
For UMR annotations.comparison to AMR-2019: AMR used :destination for end points and recipients; UMR splits those into :goal and :recipient. The old reification was be-destined-for-91, the new reifications are have-goal-91 and have-recipient-91. Argument structure preserved.
Aliases:
have-goal (m.)
Roles:
ARG1-VSP: entity in motion
ARG2-GOL: goal
ARG2-GOL: goal
Batmobile; elided motion predicate
To the BatmobileARG2
!have-group
have-group.91 - reification of :group
For UMR annotations.comparison to AMR-2019: The old AMR role :consist-of has been split with some of its coverage being funnelled into :group and some being funnelled into :material. The reification for :consist-of was consist-01, whose args were 1/2 but were flipped compared to this roleset..
Aliases:
have-group (m.)
Roles:
ARG1-VSP: members
ARG2-VSP: group
ARG2-VSP: group
swarm
a swarmARG2
of beesARG1
have-instrument
have-instrument.91 - reification of :instrument
For AMR, UMR annotations.comparison to AMR-2019: same roleset, same usage.
Aliases:
have-instrument (m.)
Roles:
ARG1-VSP: entity
ARG2-MNR: instrument
ARG2-MNR: instrument
Reified English sentence
I ate my soupARG1
with a ladleARG2
.(e / eat-01 :ARG0 (i / i) :ARG1 (s / soup :poss i) :ARG1-of (h / have-instrument-91 :ARG2 (l / ladle)))
have-li
have-li.91 - reification of :li
For AMR annotations.Aliases:
have-li (m.)
Roles:
ARG1-PPT: entity
ARG2-VSP: list-item
ARG2-VSP: list-item
list item
1ARG2
. Go to workARG1
.have-list-item
have-list-item.91 - reification of :list-item
For UMR annotations.comparison to AMR-2019: For UMR, the roleset name has been expanded to have-list-item-91 to make it easier to identify.
Aliases:
have-list-item (m.)
Roles:
ARG1-PPT: entity
ARG2-VSP: list-item
ARG2-VSP: list-item
list item
1rel
.rel
Go to workARG1
.have-manner
have-manner.91 - reification of :manner
For AMR, UMR annotations.comparison to AMR-2019: same roleset, same usage.
Aliases:
have-manner (m.)
Roles:
ARG1-VSP: event/entity
ARG2-MNR: manner
ARG2-MNR: manner
reified English sentence
The problem was solvedARG1
with a handshakeARG2
.(s / solve-01 :ARG1 (p / problem) :ARG1-of (h / have-manner-91 :Arg2 (h2 / handshake-00)))
have-material
have-material.91 - reification of :material
For UMR annotations.comparison to AMR-2019: The old AMR role :consist-of has been split with some of its coverage being funnelled into :group and some being funnelled into :material. Have-material-91 is a new roleset that preserves the argument structure of consist-01, which was used as a reification for :consist-of.
Aliases:
have-material (m.)
Roles:
ARG1-VSP: entity
ARG2-VSP: material
ARG2-VSP: material
Minecraft inspired
He mashed together a little mountainARG1
out of potatoARG2
.have-medium
have-medium.91 - reification of :medium
For UMR annotations.comparison to AMR-2019: AMR never made a reification for this role. If one had existed, it would have looked just like this.
Aliases:
have-medium (m.)
Roles:
ARG1-VSP: entity
ARG2-VSP: medium
ARG2-VSP: medium
French
a
FrenchARG2
songARG1
have-mod
have-mod.91 - reification of :mod (and :domain, in AMR); nonprototypical clauses: property predication
For AMR, UMR annotations.Comparison to AMR-2019: Same roleset used by AMR. As the reification of :mod, it has the same scope in UMR as in AMR: It is the catch-all role for arguments that don't seem to fit any of the other roles, including but not limited to determiners, adverbials, discourse markers, and other hard-to-categorize relations. (However, it is NOT used to represent as-yet undefined core-participant roles, for languages undergoing Stage 0 annotation; see note on have-other-role-91.) It is also used for property predication, when the property being predicated does not have its own roleset for the language being annotated; see example. (Note this would have been annotated with :domain in AMR.) Finally, note that “nonprototypical clause” refers to the same thing as “nonverbal clause” in the UMR guidelines.
Aliases:
have-mod (m.)
Roles:
ARG1-PPT: entity
ARG2-VSP: mod
ARG2-VSP: mod
shirt quirk
My shirtARG1
isrel
quirkyARG2
.(h / have-mod-91 :ARG1 (s / shirt :poss (i / i)) :ARG2 (q / quirky))
have-modal-strength
have-modal-strength.91 - reification of attribute :modal-strength
NOTES: added by Julia during :domain retrofits for UMR.Aliases:
as_soon_as_possible (m.)
possible (j.)
possibility (n.)
possibly (r.)
potential (j.)
potential (n.)
have_potential (l.)
potentially (r.)
probable (j.)
probability (n.)
probably (r.)
possible (j.)
possibility (n.)
possibly (r.)
potential (j.)
potential (n.)
have_potential (l.)
potentially (r.)
probable (j.)
probability (n.)
probably (r.)
Roles:
ARG1-VSP: event
ARG2-VSP: modal-strength value
ARG2-VSP: modal-strength value
have-modal-strength-91
But when it is a bad plant , one must
destroy itARG1
as soon as possiblerel
, the very first instant that one recognizes it .be_potential
The
potentialrel
is there for bothARG1
.nested modalities
Of course money is a problem for him so he
mayrel
HAVErel
torel
drop out.have-mode
have-mode.91 - reification of :mode
For AMR, UMR annotations.comparison to AMR-2019: same roleset, same usage.
Aliases:
have-mode (m.)
Roles:
ARG1-PPT: entity
ARG2-VSP: mode
ARG2-VSP: mode
reification of 'Go!'
GoARG1
! [UMR: imperativeARG2
](h / have-mode-91 :ARG1 (g / go-02) :ARG2 Imperative)
have-name
have-name.91 - reification of :name
For AMR, UMR annotations.comparison to AMR-2019: same roleset, same usage.
Aliases:
have-name (m.)
Roles:
ARG1-VSP: entity
ARG2-VSP: name
ARG2-VSP: name
to support a former name
The archaeologists visited
IstanbulARG1
, the formerARGM-TMP
ConstantinopleARG2
(v / visit-01 :ARG0 (a / archaeologist) :ARG1 (c / city :wiki "Istanbul" :name (n / name :op1 "Istanbul") :ARG1-of (h / have-name-91 :ARG2 (n2 / name :op1 "Constantinople") :time (f / former))))
have-ord
have-ord.91 - reification of :ord and :range
For AMR, UMR annotations.comparison to AMR-2019: AMR had a have-ord-91 roleset, but it did not include a numbered arg for :range, which was a sub-role under :ord. SpatialAMR also added :range-start and :range-trajectory sub-roles. All are included as numbered args now.
Aliases:
have-ord (m.)
have-ordinality (m.)
have-ordinality (m.)
Roles:
ARG1-VSP: main entity
ARG2-VSP: ord
ARG3-SRC: range in which counting has meaning
ARG4-SRC: start of the range
ARG5-DIR: direction of counting within range
ARG2-VSP: ord
ARG3-SRC: range in which counting has meaning
ARG4-SRC: start of the range
ARG5-DIR: direction of counting within range
reification of :ord, :range; temporal usage
ourARG1
firstARG2
visitC-ARG1
in ten yearsARG3
.(h / have-ord-91 :ARG1 (v / visit-01 :ARG0 (w / we)) :ARG2 1 :ARG3 (t / temporal-quantity :quant 10 :unit (y / year)))
predication of ordinal
I do n't know whether
itARG1
was the firstARG2
.(k / know-01 :polarity - :ARG0 (i / i) :ARG1 (t / truth-value :polarity-of (h / have-ord-91 :ARG1 (i2 / it) :ARG2 1)))
reified English phrase, with range and trajectory; spatial
The secondARG2
of the rows with yellow blocksARG3
, from the bottom of the structureARG4
.(h / have-ord-91 :ARG1 (r / row) :ARG2 2 :range (r2 / row :consist-of (b / block :ARG1-of (y / yellow-01))) :source (b / bottom :part-of (s / structure)))
(h / have-ord-91 :ARG1 (r / row) :ARG2 2 :ARG3 (r2 / row :plural + :configuration-of (b / block :plural + :color (y / yellow))) :ARG4 (t / thing :ARG1-of (b2 / bottom-03 :ARG2 (s / structure) :ARG3 (c / cartesian-framework-91 :ARG1 (b3 / build-space)))))
have-org-role
have-org-role.91 - X is in organization Y as Z (AMR)
For AMR, UMR annotations.comparison to AMR-2019: same roleset used by AMR.
Aliases:
have-org-role (m.)
have-organization-role (m.)
have-organizational-role (m.)
org-role (m.)
have-role (m.)
have-organization-role (m.)
have-organizational-role (m.)
org-role (m.)
have-role (m.)
Roles:
ARG0-VSP: office holder
ARG1-VSP: organization
ARG2-VSP: title of office held
ARG3-VSP: description of responsibility
ARG1-VSP: organization
ARG2-VSP: title of office held
ARG3-VSP: description of responsibility
with organization
United NationsARG1
Secretary - GeneralARG2
Ban Ki - moonARG0
(h / have-org-role-91 :ARG0 (p / person :name (n / name :op1 "Ban" :op2 "Ki-moon")) :ARG1 (o / organization :name (n2 / name :op1 "United" :op2 "Nations")) :ARG2 (s / secretary :mod (g / general)))
without organization
PresidentARG2
ObamaARG0
(h / have-org-role-91 :ARG0 (p / person :name (n / name :op1 "Obama")) :ARG2 (p2 / president))
member
SpainARG0
is a memberARG2
of NATOARG1
.(h / have-org-role-91 :ARG0 (c / country :name (n / name :op1 "Spain")) :ARG1 (m / military :name (n2 / name :op1 "NATO")) :ARG2 (m2 / member))
with :ARG3
As janitorARG2
, heARG0
cleans bathrooms and empties the trashARG3
for the schoolARG1
.(h / have-org-role-91 :ARG0 (h2 / he) :ARG1 (s / school) :ARG2 (j / janitor) :ARG3 (a / and :op1 (c / clean-01 :ARG0 h2 :ARG1 (b / bathroom)) :op2 (e / empty-01 :ARG0 h2 :ARG2 (t / trash))))
have-org-role
have-org-role.92 - X is in organization Y as Z (UMR-- note arg number changes)
For AMR, UMR annotations.comparison to AMR-2019: Roleset args have been shifted up from 0/1/2/3 to 1/2/3/4. Roleset has been renamed with -92 suffix to distinguish version with new arg numbers.
Aliases:
have-org-role (m.)
have-organization-role (m.)
have-organizational-role (m.)
org-role (m.)
have-role (m.)
have-organization-role (m.)
have-organizational-role (m.)
org-role (m.)
have-role (m.)
Roles:
ARG1-VSP: office holder
ARG2-VSP: organization
ARG3-VSP: title of office held
ARG4-VSP: description of responsibility
ARG2-VSP: organization
ARG3-VSP: title of office held
ARG4-VSP: description of responsibility
with organization
United NationsARG2
Secretary - GeneralARG3
Ban Ki - moonARG1
without organization
PresidentARG3
ObamaARG1
member
SpainARG1
is a memberARG3
of NATOARG2
.with :ARG3
As janitorARG3
, heARG1
cleans bathrooms and empties the trashARG4
for the schoolARG2
.have-other-role
have-other-role.91 - reification of :other-role
For UMR annotations.This is the reification of :other-role, which is used for as-yet undefined core-participant arguments that don't seem to fit any of the other general participant roles from UMR’s inventory. This relation will typically only be used for languages undergoing Stage 0 annotation, where many of the events may not have rolesets yet and therefore the strategy is to use general roles like :actor, :undergoer, etc. If there is a core participant role that doesn’t fit one of these relations from UMR’s general participant role inventory, :other-role used. It is ONLY used for these core participant roles (and as a replacement for :cost); it is NOT used as a catch-all for other relations such as adverbs, discourse markers, etc. :mod is used for these, just like it was in AMR. Finally, note that “nonprototypical clause” refers to the same thing as “nonverbal clause” in the UMR guidelines.
Aliases:
have-other-role (m.)
Roles:
ARG1-VSP: event or entity
ARG2-VSP: other role
ARG2-VSP: other role
reified English sentnce, instead of cost-01
The bananaARG1
was ten dollarsARG2
.have-part
have-part.91 - reification of :part
For AMR, UMR annotations.comparison to AMR-2019: same roleset, same usage.
Aliases:
have-part (m.)
Roles:
ARG1-VSP: main entity
ARG2-VSP: part
ARG2-VSP: part
part
ItARG2
is allC-ARG2
partrel
ofrel
UPS 's new " Logistics " advertisement campaignARG1
.(h / have-part-91 :ARG1 (c / campaign-01 :ARG0 (c2 / company :wiki "United_Parcel_Service" :name (n / name :op1 "UPS")) :ARG1 (a2 / advertise-01 :ARG1 (l / logistic)) :ARG1-of (n2 / new-01)) :ARG2 (i2 / it :mod (a / all)))
prepositional
He repaired
the roofARG2
of the houseARG1
.(r / repair-01 :ARG0 (h / he) :ARG1 (r2 / roof :ARG2-of (h2 / have-part-91 :ARG1 (h3 / house))))
have-path
have-path.91 - reification of :path
For UMR annotations.comparison to AMR-2019: AMR never created a reification for :path. If it had, it would have looked just like this.
Aliases:
have-path (m.)
Roles:
ARG1-VSP: event/entity
ARG2-DIR: path
ARG2-DIR: path
to grandmothers's house
Over the river and through the woodsARG2
.have-location
have-place
have-place.91 - reification of :place, nonprototypical clauses: predicative location
For UMR annotations.comparison to AMR-2019: In UMR, :place replaces :location as the general semantic role for locative phrases. UMR also includes a new roleset, have-location-91, for predicative expressions of location. Have-place-91 uses the same argument structure as be-located-at-91 did in AMR.
Aliases:
have-place (m.)
be (v.)
be (v.)
Roles:
ARG1-PPT: entity
ARG2-LOC: location
ARG2-LOC: location
movie trialer; with reification and inverse
In a world where the sun burns coldARG2
...predicative location
The symbolARG1
wasrel
on the rockARG2
have-polarity
have-polarity.91 - reification of :polarity
For AMR, UMR annotations.comparison to AMR-2019: same roleset, same usage.
Aliases:
have-polarity (m.)
Roles:
ARG1-VSP: entity
ARG2-VSP: polarity (typical value: -)
ARG2-VSP: polarity (typical value: -)
elliptic occurrences of not
Will they bite ?
So farARGM-TMP
, NOTrel
. [UMR: -ARG2
](h / have-polarity-91 :ARG2 - :time (s / so-far))
reified English sentence
Pleaserel
goARG2
!(h / have-polarity-91 :ARG2 - :time (s / so-far))
Spanish example of double negation for emphasis
No le dijo nada a nadie
(d / decir-01 :polarity - :ARG0 (p / person :refer-person 3rd :refer-number Singular) :ARG1 (n / nada) :ARG2 (p2 / person :refer-person 3rd :refer-number Singular))
have-pure-addition
have-pure-addition.91 - reification of :pure-addition
For UMR annotations.comparison to AMR-2019: new roleset, covers cases where two events form a complex figure, cannot occur separately from each other, but have no sequence specification. This covers some of what was covered by (a / and) in AMR. Use reification if the two events are split across sentences in a corpus.
Aliases:
have-pure-addition (m.)
Roles:
ARG1-VSP: first event
ARG2-VSP: second event
ARG2-VSP: second event
fragment, separated across sentence boundary
You have to
have your hand stampedARG1
. Andrel
show your ticketARG2
.have-purpose
have-purpose.91 - reification of :purpose
For AMR, UMR annotations.comparison to AMR-2019: same roleset, same usage.
Aliases:
have-purpose (m.)
Roles:
ARG1-VSP: event/entity
ARG2-VSP: purpose
ARG2-VSP: purpose
to eliminate bugs
We tested the programARG1
to eliminate bugsARG2
.(h / have-purpose-91 :ARG1 (t / test-01 :ARG0 (w / we) :ARG1 (p2 / program)) :ARG2 (e / eliminate-01 :ARG1 (b / bug)))
have-quant
have-quant.91 - reification of :quant, and quantifiable version of have-degree-91
For AMR, UMR annotations. Do not use this frame for simple quantities (e.g. We saw two horses and many cows.)comparison to AMR-2019: same roleset, same usage.
Aliases:
have-quant (m.)
Roles:
ARG1-PPT: entity, thing being quantified
ARG2-ADV: numerical quantity or quantifier; how much/ how many
ARG3-EXT: degree mention modifying quantity, thing quantified (e.g. more, most, less, least, times, equal, enough, too, so, to-the-point, at-least)
ARG4-PPT: compared-to
ARG5-PRD: superlative, reference to superset
ARG6-PRD: consequence, result of quantity
ARG2-ADV: numerical quantity or quantifier; how much/ how many
ARG3-EXT: degree mention modifying quantity, thing quantified (e.g. more, most, less, least, times, equal, enough, too, so, to-the-point, at-least)
ARG4-PPT: compared-to
ARG5-PRD: superlative, reference to superset
ARG6-PRD: consequence, result of quantity
with consequence; 'enough' is arg3
I had
scarcely enoughARG3
drinking waterARG1
torel
last a week .(h / have-03 :ARG0 (i / i) :ARG1 (w / water :purpose (d / drink-01 :ARG0 i) :ARG1-of (h2 / have-quant-91 :ARG3 (e / enough :mod (s / scarce)) :ARG6 (l / last-03 :ARG1 w :ARG2 (t / temporal-quantity :quant 1 :unit (w2 / week)) :ARG3 i))))
with consequence and polarity -
I had
toorel
muchARG2
waterARG1
torel
carry .(h / have-03 :ARG0 (i / i) :ARG1 (w / water :ARG1-of (h2 / have-quant-91 :ARG2 (m /much) :ARG3 (t / too) :ARG6 (c / carry-01 :ARG0 i))))
'as many X as' example for quant
He sold
asrel
manyARG2
carsARG1
asrel
his competitor . [UMR: equalARG3
](s / sell-01 :ARG0 (h / he) :ARG1 (c / car :ARG1-of (h2 / have-quant-91 :ARG2 (m / many) :ARG3 (e / equal) :ARG4 (c2 / car :ARG1-of (s2 / sell-01 :ARG0 (p2 / person :ARG0-of (c3 / compete-01 :ARG1 h)))))))
cars
He has
therel
mostrel
carsARG1
ofrel
anyone in the yacht club(h / have-03 :ARG0 (h2 / he) :ARG1 (c / car :ARG1-of (h3 / have-quant-91 :ARG3 (m / most) :ARG5 (a / anyone :ARG0-of (h4 / have-org-role-91 :ARG1 (c2 / club :mod (y / yacht)))))))
cats
He knew that his neighbor had adopted
fourARG2
moreARG3
catsARG1
.(k / know-01 :ARG0 (h / he) :ARG1 (a / adopt-01 :ARG0 (p2 / person :ARG0-of (h2 / have-rel-role-91 :ARG1 h :ARG2 (n / neighbor))) :ARG1 (c / cat :ARG0-of (h3 / have-quant-91 :ARG2 4 :ARG3 (m / more)))))
have-reason
have-reason.91 - reification of :reason
For UMR annotations.comparison to AMR-2019: in UMR, what was previously covered by :cause has been split into :cause and :reason. Have-reason-91 is now the reification for :reason. Arg structure is 1/2 rather than the 0/1 used by the original cause-01.
Aliases:
have-reason (m.)
Roles:
ARG1-VSP: event
ARG2-CAU: reason
ARG2-CAU: reason
reified English sentence
They got marriedARG1
because they were in loveARG2
.have-recipient
have-recipient.91 - reification of :recipient
For UMR annotations.comparison to AMR-2019: AMR used :destination for end points and recipients; UMR splits those into :goal and :recipient. The old reification was be-destined-for-91, the new reifications are have-goal-91 and have-recipient-91. Argument structure preserved.
Aliases:
have-recipient (m.)
Roles:
ARG1-VSP: event
ARG2-GOL: recipient
ARG2-GOL: recipient
Reified version of an English sentence
I showed the picturesARG1
to herARG2
.have-rel-role
have-rel-role.91 - A and B (of same type) are related, e.g. as family or friends (AMR)
For AMR annotations.Aliases:
have-rel-role (m.)
have-relation-role (m.)
have-relational-role (m.)
rel-role (m.)
have-role (m.)
have-relation-role (m.)
have-relational-role (m.)
rel-role (m.)
have-role (m.)
Roles:
ARG0-VSP: entity A
ARG1-VSP: entity B
ARG2-VSP: role of entity A (must be specified)
ARG3-VSP: role of entity B (often left unspecified)
ARG4-VSP: relationship basis (contract, case; rarely used)
ARG1-VSP: entity B
ARG2-VSP: role of entity A (must be specified)
ARG3-VSP: role of entity B (often left unspecified)
ARG4-VSP: relationship basis (contract, case; rarely used)
family
hisARG0
motherARG2
(p / person :ARG0-of (h / have-rel-role-91 :ARG1 (h2 / he) :ARG2 (m / mother)))
family (both roles specified)
JohnARG0
and MaryARG1
are husbandARG2
and wifeARG3
(h / have-rel-role-91 :ARG0 (p / person :name (n / name :op1 "John")) :ARG1 (p2 / person :name (n2 / name :op1 "Mary")) :ARG2 (h / husband) :ARG3 (w / wife))
friend
myARG1
old friendARG2
MichaelARG0
(p / person :name (n / name :op1 "Michael") :ARG0-of (h / have-rel-role-91 :ARG1 (i / i) :ARG2 (f / friend :mod (o / old))))
lawyer (with ARG4)
SheARG0
is hisARG1
lawyerARG2
in the patent infringement suitARG4
.(h / have-rel-role-91 :ARG0 (s / she) :ARG1 (h / he) :ARG2 (l / lawyer) :ARG4 (s / sue-02 :ARG2 (i / infringe-01 :ARG1 (p3 / patent-01))))
sister city
BerlinARG0
is a sister cityARG2
of Los AngelesARG1
.(h / have-rel-role-91 :ARG0 (c / city :name (n / name :op1 "Berlin")) :ARG1 (c2 / city :name (n2 / name :op1 "Los" :op2 "Angeles")) :ARG2 (c3 / city :mod (s / sister)))
sister cities
BerlinARG0
and Los AngelesARG1
are sister citiesARG2
.(h / have-rel-role-91 :ARG0 (c / city :name (n / name :op1 "Berlin")) :ARG1 (c2 / city :name (n2 / name :op1 "Los" :op2 "Angeles")) :ARG2 (c3 / city :mod (s / sister)) :ARG3 (c4 / city :mod (s2 / sister)))
have-rel-role
have-rel-role.92 - A and B (of same type) are related, e.g. as family or friends (UMR version-- not arg number shift)
For AMR, UMR annotations.comparison to AMR-2019: This roleset shifts the arg structure of have-rel-role-91 from 0/1/2/3/4 to 1/2/3/4/5. Because of this, the roleset ID numeric suffix has been changed for this version to -92. For UMR, the two sister-city examples moved into have-role-91, which is better for inanimates now.
Aliases:
have-rel-role (m.)
have-relation-role (m.)
have-relational-role (m.)
rel-role (m.)
have-role (m.)
have-relation-role (m.)
have-relational-role (m.)
rel-role (m.)
have-role (m.)
Roles:
ARG1-VSP: entity A
ARG2-VSP: entity B
ARG3-VSP: role of entity A (must be specified)
ARG4-VSP: role of entity B (often left unspecified)
ARG5-VSP: relationship basis (contract, case; rarely used)
ARG2-VSP: entity B
ARG3-VSP: role of entity A (must be specified)
ARG4-VSP: role of entity B (often left unspecified)
ARG5-VSP: relationship basis (contract, case; rarely used)
family
hisARG1
motherARG3
family (both roles specified)
JohnARG1
and MaryARG2
are husbandARG3
and wifeARG4
friend
myARG2
old friendARG3
MichaelARG1
lawyer (with ARG4)
SheARG1
is hisARG2
lawyerARG3
in the patent infringement suitARG5
.have-role
have-role.91 - for nonprototypical clauses of object predication that don't fall under have-rel-role-92 or have-org-role-92
For UMR annotations.comparison to AMR-2019: This is a new roleset. AMR used :domain or have-mod-91 for most of these.
Aliases:
have-role (m.)
be (v.)
be (v.)
Roles:
ARG1-VSP: first entity
ARG2-VSP: optional second entity
ARG3-VSP: object category for ARG1
ARG4-VSP: optional object category for ARG2
ARG2-VSP: optional second entity
ARG3-VSP: object category for ARG1
ARG4-VSP: optional object category for ARG2
object predicational nonprototypical clause
PandaARG1
isrel
a catARG3
have-source
have-source.91 - reification of :source
For UMR annotations.comparison to AMR-2019: :source was a role in AMR. Some of what it covered stays under :source in UMR, and some moves into :start. The reification for the old :source role was be-from-91, which is now replaced with have-source-91 and have-start-91 in UMR. Argument structure preserved.
Aliases:
have-source (m.)
Roles:
ARG1-VSP: entity
ARG2-SRC: source
ARG2-SRC: source
Reified English sentence
He picked some berriesARG1
from the bushARG2
.have-start
have-start.91 - reification of :start
For UMR annotations.comparison to AMR-2019: :source was a role in AMR. Some of what it covered stays under :source in UMR, and some moves into :start. The reification for the old :source role was be-from-91, which is now replaced with have-source-91 and have-start-91. Argument structure preserved.
Aliases:
have-start (m.)
Roles:
ARG1-VSP: event/entity
ARG2-VSP: start
ARG2-VSP: start
reified English sentence
She walked homeARG1
from the storeARG2
.have-stimulus
have-stimulus.91 - reification of :stimulus
For UMR annotations.comparison to AMR-2019: This is a new reification for a new role; neither existed in AMR, as these roles were included as numbered arguments in PropBank rolesets.
Aliases:
have-stimulus (m.)
Roles:
ARG1-VSP: event
ARG2-PPT: stimulus
ARG2-PPT: stimulus
reified version of an English sentence
YouARG2
had seen the movieARG1
.have-subevent
have-subevent.91 - reification of :subevent
For AMR annotations.Reification added to support AMR annotations.
Aliases:
have-subevent (m.)
Roles:
ARG1-VSP: main event
ARG2-VSP: subevent
ARG2-VSP: subevent
soccer
Inrel
the 1974 soccer World CupARG1
finalARG2
, Germany beat the Netherlands 2 - 1 .(b / beat-03 :ARG0 (c / country :wiki "Germany" :name (n / name :op1 "Germany")) :ARG1 (c2 / country :wiki "Netherlands" :name (n2 / name :op1 "Netherlands")) :ARG2 (f / final :ARG2-of (h / have-subevent-91 :ARG1 (g / game :wiki "FIFA_World_Cup" :name (n3 / name :op1 "World" :op2 "Cup") :mod (s / soccer) :time (d / date-entity :year 1974)))) :quant (s2 / score-entity :op1 2 :op2 1))
have-substitution-91
have-substitution.91 - A instead of B, reification of :substitute in UMR
For AMR, UMR annotations.comparison to AMR-2019: new roleset, replaces include-91 and acts as reification for new role version, :substitute .
Aliases:
instead-of (m.)
instead (m.)
as-opposed-to (m.)
rather-than (m.)
not-x-but-y (m.)
instead (m.)
as-opposed-to (m.)
rather-than (m.)
not-x-but-y (m.)
Roles:
ARG1-VSP: this
ARG2-VSP: instead of that
ARG2-VSP: instead of that
instead of, in lieu of, in place of, rather than, as opposed to
Tim went for an early morning bike rideARG1
instead of sleeping inARG2
.not A, but B
I 'm not angryARG2
, butrel
disappointedARG1
.have-subtraction
have-subtraction.91 - reification of :subtraction; discourse relation for cases where one event is characterized by the absence of a second event that is normally part of it
For UMR annotations.comparison to AMR-2019: New role/roleset. AMR used except-01 for some of these cases.
Aliases:
except (p.)
Roles:
ARG1-VSP: main event
ARG2-VSP: event/entity that does not occur
ARG2-VSP: event/entity that does not occur
reified English sentence
People do n't own tigers , except for Joe Exotic .
have-temporal
have-temporal.91 - reification of :temporal in UMR
For UMR annotations.comparison to AMR-2019: be-temporally-at-91 is the same roleset used in AMR. It corresponded to :time in AMR, and corresponds to :temporal in UMR, which replaced :time everywhere except within (d / date-entity).
Aliases:
be-temporally-at (m.)
have-temporal (m.)
have-temporal (m.)
Roles:
ARG1-VSP: event
ARG2-TMP: temporal
ARG2-TMP: temporal
on Saturday
The beach partyARG1
is on SaturdayARG2
.have-theme
have-theme.91 - reification of :theme
For UMR annotations.comparison to AMR-2019: This is a new reification for a new role; neither existed in AMR, as these roles were included as numbered arguments in PropBank rolesets.
Aliases:
have-theme (m.)
Roles:
ARG1-VSP: event
ARG2-PPT: theme
ARG2-PPT: theme
Reified version of an English sentence
I showed the picturesARG1
to herARG2
.have-topic
have-topic.91 - reification of :topic
For UMR annotations.comparison to AMR-2019: This replaces concern-02 as the reification for :topic in UMR. The argument structure is shifted from 0/1 to 1/2.
Aliases:
have-topic (m.)
Roles:
ARG1-VSP: entity/event
ARG2-VSP: topic
ARG2-VSP: topic
reified topic modifier
informationARG1
about the caseARG2
have-undergoer
have-undergoer.91 - reification of :undergoer
For UMR annotations.comparison to AMR-2019: This is a new reification for a new role; neither existed in AMR, as these roles were included as numbered arguments in PropBank rolesets.
Aliases:
have-undergoer (m.)
Roles:
ARG1-VSP: event
ARG2-PPT: undergoer
ARG2-PPT: undergoer
Reified version of an English sentence
The enemy sank:actor
the ship:undergoer
.have-unit
have-unit.91 - reification of :unit
For UMR annotations.comparison to AMR-2019: :unit didn't have a reification in AMR, and it was only used as a sub-role under X-quantity types. In UMR, :unit is used for ad-hoc mensural constructions, so the reification is more likely to be needed.
Aliases:
have-unit (m.)
Roles:
ARG1-VSP: event
ARG2-VSP: unit of measurement
ARG2-VSP: unit of measurement
no mensural head mentioned, unit in fragment
How many did you take ? A handful .
have-value
have-value.91 - reification of :value.
For AMR, UMR parsing.comparison to AMR-2019: Same roleset, same usage.
Aliases:
have-value (m.)
Roles:
ARG1-VSP: entity
ARG2-VSP: value
ARG2-VSP: value
reification of value
The phone numberARG1
is 1-800-555-1223ARG2
.(h / have-value-91 :ARG1 (p / phone-number-entity) :ARG2 "1-800-555-1223")
have-vocative
have-vocative.91 - reification of :vocative
For UMR annotations.comparison to AMR-2019:
Aliases:
have-vocative (m.)
Roles:
ARG1-VSP: proposition
ARG2-VSP: vocative
ARG2-VSP: vocative
reified English sentence
DadARG2
, I need more moneyARG1
.hyperlink
hyperlink.91 - for hyperlinks
For AMR, UMR parsing.comparison to AMR-2019: same roleset, same usage. Roleset was taken from hyperlink-01, minus args0 and 2.
Aliases:
hyperlink (m.)
Roles:
ARG1-PPT: text being hyperlinked
ARG3-VSP: url-entity
ARG3-VSP: url-entity
Declaration of Independence hotlink
<a
href="https://www.monticello.org/sites/default/files/uploaded-content-images/Declaration_Engrav_Pg1of1_doctored_0.jpg"<ARG2
Declaration of IndependenceARG1
</a>(d / declare-01 :ARG2 (d2 / depend-01 :polarity -) :ARG1-of (h / hyperlink-91 :ARG3 (u / url-entity :value "https://www.monticello.org/sites/default/files/uploaded-content-images/Declaration_Engrav_Pg1of1_doctored_0.jpg")))
identity
identity.91 - nonprototypical clauses: equation
For UMR annotations.comparison to AMR-2019: This is a new roleset. AMR has a lemma-specific identical-01 roleset, but it is semantically not quite the same so the two should not be merged. AMR used :domain or have-mod-91 for these.
Aliases:
be (v.)
Roles:
ARG1-VSP: theme
ARG2-VSP: equated referent
ARG2-VSP: equated referent
equational object nonprototypical clause
That SubaruARG1
isrel
my carARG2
equated referent
She works for
UBSARG1
, the biggest bank in SwitzerlandARG2
.include
include.91 - For relative size of subset compared to superset
For AMR, UMR annotations.comparison to AMR-2019: UMR only uses :subset in the document level annotation, so reification isn't possible in that context. However, this roleset is still useful when describing relative sizes between subsets/supersets.
Aliases:
include (v.)
of (p.)
of (p.)
Roles:
ARG1-VSP: subset
ARG2-VSP: superset
ARG3-VSP: relative size of the subset compared to superset
ARG2-VSP: superset
ARG3-VSP: relative size of the subset compared to superset
percentage
10 %ARG3
ofrel
smokersARG2
die of lung cancerARG1
.(i / include-91 :ARG1 (p / person :ARG1-of (d / die-01 :ARG1-of (c2 / cause-01 :ARG0 (d2 / disease :wiki "Lung_cancer" :name (n / name :op1 "lung" :op2 "cancer"))))) :ARG2 (p2 / person :ARG0-of (s2 / smoke-02)) :ARG3 (p3 / percentage-entity :value 10))
26 of 44
26ARG1
ofrel
the 44 countries in EuropeARG2
are members of NATOC-ARG1
.(i / include-91 :ARG1 (c / country :quant 26 :ARG0-of (h / have-org-role-91 :ARG1 (m3 / military :wiki "NATO" :name (n2 / name :op1 "NATO")) :ARG2 (m2 / member))) :ARG2 (c2 / country :quant 44 :location (c3 / continent :wiki "Europe" :name (n / name :op1 "Europe"))))
infer
infer.91 - for presentation of inferences
For UMR annotations.comparison to AMR-2019: Roleset is a renamed copy of infer.01, which was used lemma-specifically and for implicit inference in AMR. Some inferences (like 'I think, therefore I am') that were covered with this roleset in AMR are now covered with :reason/have-reason-91 in UMR. This roleset is still to be used in dialogues where one speaker says something and the other speaker presents an inference based on that information.
Aliases:
infer (v.)
Roles:
ARG0-PAG: entity inferring something
ARG1-PPT: inference
ARG2-DIR: source of inference
ARG1-PPT: inference
ARG2-DIR: source of inference
inference in dialogue
Sorel
, make it longerARG1
.instead-of
instead-of.91 - A instead of B
For AMR annotations.Aliases:
instead-of (m.)
instead (m.)
as-opposed-to (m.)
rather-than (m.)
not-x-but-y (m.)
instead (m.)
as-opposed-to (m.)
rather-than (m.)
not-x-but-y (m.)
Roles:
ARG1-VSP: this
ARG2-VSP: instead of that
ARG2-VSP: instead of that
instead of, in lieu of, in place of, rather than, as opposed to
Tim went for an early morning bike rideARG1
instead of sleeping inARG2
.(r / ride-01 :ARG0 (p / person :wiki - :name (n / name :op1 "Tim")) :ARG1 (b / bike) :time (e / early :op1 (d / date-entity :dayperiod (m / morning))) :ARG1-of (i / instead-of-91 :ARG2 (s / sleep-00 :ARG0 p)))
not A, but B
I 'm not angryARG2
, butrel
disappointedARG1
.(d / disappoint-01 :ARG1 (i / i) :ARG1-of (i2 / instead-of-91 :ARG2 (a / angry :domain i)))
mean
mean.91 - for cases where an element signifies a meaning
For UMR annotationcomparison to AMR-2019: mean-01 was used both lemma-specifically and for symbols like '*' and '=', as well as for implicit cases. UMR will use this roleset for the latter types. This may be revisited later.
Aliases:
mean (v.)
* (s.)
= (s.)
* (s.)
= (s.)
Roles:
ARG0-PAG: agent
ARG1-MNR: element containing meaning
ARG2-PPT: meaning
ARG3-GOL: perceiver, beneficiary
ARG1-MNR: element containing meaning
ARG2-PPT: meaning
ARG3-GOL: perceiver, beneficiary
typing correction usage
shapeARG2
*rel
misspeaking correction
sorry
greenARG2
speaker is explaining shorthand used previously
purpARG1
=rel
purpleARG2
proverb
proverb.91 - for fixed/semi-fixed expressions that are allegorical or convey wisdom
For UMR annotations.comparison to AMR-2019: no such roleset existed.
Aliases:
proverb (n.)
Roles:
ARG1-VSP: proverb
ARG2-VSP: 2nd part of a two-part allegorical exp. Gives a focused interpretation of ARG1
ARG2-VSP: 2nd part of a two-part allegorical exp. Gives a focused interpretation of ARG1
standard proverb
The mountains are high and the emperor is far away .
standard proverb
山 高 皇帝 远
publication
publication.91 - for publication information, citations
For UMR annotations.comparison to AMR-2019: same usage, added roles for publisher and url.
Aliases:
publication (m.)
Roles:
ARG0-VSP: author(s)
ARG1-VSP: title
ARG2-VSP: abstract
ARG3-VSP: full text
ARG4-VSP: venue (journal, conference)
ARG5-VSP: volume
ARG6-VSP: issue
ARG7-VSP: pages
ARG8-VSP: id (e.g. PubMed)
ARG9-VSP: editor(s)
ARG10-VSP: publisher
ARG11-VSP: url
ARG1-VSP: title
ARG2-VSP: abstract
ARG3-VSP: full text
ARG4-VSP: venue (journal, conference)
ARG5-VSP: volume
ARG6-VSP: issue
ARG7-VSP: pages
ARG8-VSP: id (e.g. PubMed)
ARG9-VSP: editor(s)
ARG10-VSP: publisher
ARG11-VSP: url
Special Relativity
Albert EinsteinARG0
: Zur Elektrodynamik bewegter KörperARG1
( 1905ARGM-TMP
) Annalen der PhysikARG4
, 322ARG5
( 10ARG6
) : 891 - 921ARG7
(p / publication-91 :ARG0 (p2 / person :wiki "Albert_Einstein" :name (n / name :op1 "Albert" :op2 "Einstein")) :ARG1 (p3 / publication :wiki "Annus_Mirabilis_papers#Special_relativity" :name (n2 / name :op1 "Zur" :op2 "Elektrodynamik" :op3 "bewegter" :op4 "Körper")) :ARG4 (j / journal :wiki "Annalen_der_Physik" :name (n3 / name :op1 "Annalen" :op2 "der" :op3 "Physik")) :ARG5 322 :ARG6 10 :ARG7 (v / value-interval :op1 891 :op2 921) :time (d / date-entity :year 1905) :medium (l / language :wiki "German_language" :name (n4 / name :op1 "German")))
range
range.91 - not a reification of :range-- this is for instances not involved in ordinals
For UMR annotations.comparison to AMR-2019: Some languages use have-range-91 for instances like this in AMR, but the :range role was created as a subrole under :ord, and so its reification is also part of have-ord-91 in UMR. This roleset provides a home for ranges outside of ordinals.
Aliases:
from (p.)
Roles:
ARG1-SRC: start of the range
ARG2-GOL: end of the range
ARG2-GOL: end of the range
translation from Chinese corpus
Ahasuerus ascended the throne in Shush and ruled over 127 provinces
from IndiaARG1
to CushARG2
.rate-entity
rate-entity.91 - for frequencies
For UMR annotations.comparison to AMR-2019: same roleset, same usage.
Aliases:
rate-entity (m.)
Roles:
ARG1-VSP: quantity (with an implied default of 1)
ARG2-VSP: per quantity
ARG3-VSP: regular interval between events
ARG4-VSP: entity on which recurring event happens
ARG2-VSP: per quantity
ARG3-VSP: regular interval between events
ARG4-VSP: entity on which recurring event happens
with :ARG1 and :ARG2 (temporal)
twiceARG1
a yearARG2
(e / event :frequency (r / rate-entity-91 :ARG1 2 :ARG2 (y / year :quant 1)))
with :ARG1 and :ARG2 (price per volume)
$3ARG1
per gallonARG2
(r / rate-entity-91 :ARG1 (m / monetary-quantity :quant 3 :unit (d / dollar)) :ARG2 (v / volume-quantity :quant 1 :unit (g / gallon)))
with :ARG3
every other weekARG3
(e / event :frequency (r / rate-entity-91 :ARG3 (t / temporal-quantity :quant 2 :unit (w / week))))
with both :ARG3 and :ARG4
every otherARG3
WednesdayWednesday
(e / event :frequency (r / rate-entity-91 :ARG3 (t / temporal-quantity :quant 2 :unit (w / week)) :ARG4 (d / date-entity :weekday (w2 / wednesday))))
full set of args
$2ARG1
a personARG2
every 5 milesARG3
whenever it snowsARG4
(e / event :frequency (r / rate-entity-91 :ARG1 (m / monetary-quantity :quant 2 :unit (d / dollar)) :ARG2 (p / person :quant 1) :ARG3 (d2 / distance-quantity :quant 5 :unit (m2 / mile)) :ARG4 (s / snow-01)))
regardless
regardless.91 - without consideration of or in spite of
For AMR annotations. Use Concessive-Conditional in UMR.Aliases:
regardless (m.)
Roles:
ARG1-VSP: main event
ARG2-VSP: irrelevant thing
ARG2-VSP: irrelevant thing
regardless
The match will be playedARG1
regardlessrel
of weatherARG2
.(p / play-01 :ARG1 (m / match) :ARG1-of (r / regardless-91 :ARG2 (w / weather)))
no matter whether
The match will be playedARG1
norel
matterrel
whetherrel
it rains or the sun shinesARG2
.(p / play-01 :ARG1 (m / match) :ARG1-of (r / regardless-91 :ARG2 (o / or :op1 (r2 / rain-01) :op2 (s / shine-01 :ARG1 (s2 / sun)))))
rain or shine
The match will be playedARG1
rain or shineARG2
.(p / play-01 :ARG1 (m / match) :ARG1-of (r / regardless-91 :ARG2 (o / or :op1 (r2 / rain-01) :op2 (s / shine-01))))
without regard
All employees will be treated equallyARG1
withoutrel
regardrel
torel
race , gender , national origin , color , ethnicity , religion , disability or sexual orientationARG2
.(t / treat-01 :ARG1 (p / person :ARG1-of (e / employ-01) :mod (a / all)) :manner (e2 / equal-01) :ARG1-of (r / regardless-91 :ARG2 (o / or :op1 (r2 / race) :op2 (g / gender) :op3 (o2 / originate-01 :ARG2 (n / nation)) :op4 (c / color) :op5 (e3 / ethnic-group) :op6 (r3 / religion) :op7 (d / disability) :op8 (o3 / orient-01 :ARG3 (s / sex)))))
request-confirmation
request-confirmation.91 - for tag questions, or rephrasing received instructions in order to request confirmation of understanding
For AMR annotations.Aliases:
request-confirmation (m.)
tag-question (m.)
isn't (m.)
aren't (m.)
won't (m.)
right (r.)
eh (m.)
tag-question (m.)
isn't (m.)
aren't (m.)
won't (m.)
right (r.)
eh (m.)
Roles:
ARG0-PAG: entity requesting confirmation
ARG1-PPT: thing to be confirmed
ARG2-GOL: entity requestion is being made to
ARG1-PPT: thing to be confirmed
ARG2-GOL: entity requestion is being made to
tag question: isn't _ ?
He 's hungryARG1
, isrel
n'trel
herel
?rel
(h / hunger-01 :ARG0 (h2 / he) :ARG1-of (r / request-confirmation-91))
tag question: will X ?
Come hereARG1
, willrel
yourel
?rel
(c / come-01 :mode imperative :ARG1 (y / you) :ARG4 (h / here) :ARG1-of (r / request-confirmation-91))
tag question: right
YouARG2
must be Jack , rightrel
?rel
(i / infer-01 :ARG1 (p / person :wiki - :name (n / name :op1 "Jack") :domain (y / you) :ARG1-of (r / request-confirmation-91)))
requesting confirmation with intonation/question mark
A green oneARG1
?rel
(h / have-color-91 :ARG1 (o / one) :ARG2 (g / green) :ARG1-of (r / request-confirmation-91))
request-response
request-response.91 - for requests for response
For AMR annotations.Aliases:
request-response (m.)
well (r.)
so (r.)
well (r.)
so (r.)
Roles:
ARG0-PAG: entity requesting response
ARG1-PPT: thing to be responded to
ARG2-GOL: entity requestion is being made to
ARG1-PPT: thing to be responded to
ARG2-GOL: entity requestion is being made to
well?
wellrel
?rel
(r / request-response-91)
resemble
resemble.91 - be like
For UMR annotations.comparison to AMR-2019: This roleset is the same as resemble-01 except for the name change to resemble-91; AMR used resemble-01 for similarity constructions (e.g., 'A is like B'). This will take the place of resemble-01 in UMR for these situations. UMR will not use have-degree-of-resemblance-91, but will use a combo of have-degree-91 and resemble-91. ARG3 has been added to capture "A is like B in terms of feature C"
Aliases:
like (j.)
Roles:
ARG1-PPT: resembler
ARG2-PPT: resemble what
ARG3-DOM: aspect in which :ARG1 and :ARG2 are similar
ARG2-PPT: resemble what
ARG3-DOM: aspect in which :ARG1 and :ARG2 are similar
very typical
LifeARG1
is likerel
a box of chocolatesARG2
.say
say.91 - for vocatives and unattributed speech in dialogues
For UMR annotations.comparison to AMR-2019: say-01 in AMR has ARG3, which is specific to the English verb so is dropped here. say-01 will still be used in English AMR for lemma-specific instances.
Aliases:
Roles:
ARG0-PAG: sayer
ARG1-PPT: utterance
ARG2-GOL: hearer
ARG1-PPT: utterance
ARG2-GOL: hearer
English: vocative. (Note this UMR example does not include aspect annotation.)
The dog has been found , Penelope !
English: ONLY to be annotated with say-91, as shown, if it occurs in a dialogue. (Note this UMR example does not include aspect annotation.)
The dog has been found !
score-on-scale
score-on-scale.91 - for score ranges
For AMR, UMR annotations.comparison to AMR-2019: ARG4 added during UMR framing. AMR users can choose whether or not to use it for future annotation. Using :scale in AMR was also questionable, since it was only included as a sub-role under the quantity concepts.
Aliases:
score-on-scale (m.)
Roles:
ARG1-VSP: score
ARG2-VSP: worst score on scale
ARG3-VSP: best score on scale
ARG4-VSP: name of scale
ARG2-VSP: worst score on scale
ARG3-VSP: best score on scale
ARG4-VSP: name of scale
French scale
Jacques scored a
18ARG1
/ 20ARG3
in math .(s / score-01 :ARG1 (p / person :wiki - :name (n / name :op1 "Jacques")) :ARG2 (s2 / score-on-scale-91 :ARG1 18 :ARG3 20) :ARG3 (m / mathematics))
on a scale of 1-10 with all args
On a scale of
1ARG2
to 10ARG3
, how muchARG1
do you agree with X ?(f / feel-01 :ARG0 (y / you) :ARG1 (p / pain :quant (s / score-on-scale-91 :ARG1 (a / amr-unknown) :ARG2 1 :ARG3 10)))
letter grades, with arg 4
John received a
BARG1
on the Amrican five - letter grade scale , AARG3
- FARG2
.(r / receive-01 :ARG0 (p / person :name (n / name :op1 "John")) :ARG1 (s / score-on-scale-91 :ARG1 "B" :ARG2 "F" :ARG3 "A" :scale (s2 / scale :mod (c / country :name (n2 / name :op1 "America")) :mod (l / letter :quant 5))))
statistical-test
statistical-test.91 - for p-values
For AMR, UMR annotations.comparison to AMR-2019: same roleset, same usage.
Aliases:
statistical-test (n.)
p-value (n.)
p-value (n.)
Roles:
ARG1-VSP: finding
ARG2-VSP: p-value (probability of null-hypothesis)
ARG3-VSP: r-squared (coefficient of determination)
ARG4-VSP: significance test (e.g. t-test, Chi-square test, Fisher's exact test)
ARG5-VSP: alternative hypothesis (vs. ...)
ARG2-VSP: p-value (probability of null-hypothesis)
ARG3-VSP: r-squared (coefficient of determination)
ARG4-VSP: significance test (e.g. t-test, Chi-square test, Fisher's exact test)
ARG5-VSP: alternative hypothesis (vs. ...)
(p < 0.05)
Protein expression increased 10 - foldARG1
( prel
< 0.05 )ARG2
.(i / increase-01 :ARG1 (e / express-03 :ARG2 (p / protein)) :ARG2 (t / times :quant 10) :ARG1-of (s / statistical-test-91 :ARG2 (l / less-than :op1 0.05)))
correlation
Raf-1 also correlated with nuclear activated MAPKARG1
( prel
= 0.0022ARG2
, r2rel
= 0.1525ARG3
) .(c / correlate-01 :ARG1 (e / enzyme :wiki "C-Raf" :name (n / name :op1 "Raf-1")) :ARG2 (e2 / enzyme :wiki "Mitogen-activated_protein_kinase" :name (n2 / name :op1 "MAPK") :location (n3 / nucleus) :ARG1-of (a / activate-01)) :ARG1-of (s / statistical-test-91 :ARG2 0.0022 :ARG3 0.1525) :mod (a2 / also))
street-address
street-address.91 - for street address information
For UMR annotations.comparison to AMR-2019: same roleset, same usage
Aliases:
street-address (m.)
Roles:
ARG1-VSP: street number
ARG2-VSP: street
ARG3-VSP: unit (apartment, suite, etc.)
ARG4-VSP: city
ARG5-VSP: state, province, district
ARG6-VSP: zip code
ARG7-VSP: country
ARG2-VSP: street
ARG3-VSP: unit (apartment, suite, etc.)
ARG4-VSP: city
ARG5-VSP: state, province, district
ARG6-VSP: zip code
ARG7-VSP: country
White House address
1600ARG1
Pennsylvania AveARG2
, WashingtonARG4
DCARG5
, 20006ARG6
(s / street-address-91 :ARG1 1600 :ARG2 (r / road :wiki "Pennsylvania_Avenue" :name (n / name :op1 "Pennsylvania" :op2 "Ave")) :ARG4 (c / city :wiki "Washington,_D.C." :name (n2 / name :op1 "Washington")) :ARG5 (d / district :wiki "Washington,_D.C." :name (n3 / name :op1 "DC")) :ARG6 20006)
unexpected-co-occurrence
unexpected-co-occurrence.91 - discourse relation between events where the second event is unexpected if the first occurs
For UMR annotations.comparison to AMR-2019: This is a new roleset. These were handled with contrast-01 or :concession in AMR.
Aliases:
unexpected-co-occurrence (m.)
but (c.)
but (c.)
Roles:
ARG1-VSP: first event
ARG2-VSP: second event (unexpected)
ARG2-VSP: second event (unexpected)
unexpected co-ocurrence
Vanja caught a coldARG1
, butrel
went to schoolARG2