SlideShare a Scribd company logo
1 of 12
Download to read offline
Semantic Theory
                       Summer 2006
            Lexical Semantics II: Event Semantics
                                         M. Pinkal




         Examples Annotated with Thematic Roles


     – [The window]pat broke
     – [A rock ]inst broke [the window ]pat
     – [John ]ag broke [the window ]pat [with a rock ]inst

     – [Peter ]ag gave [Mary]rec [the book ]pat
     – [Mary ]rec received [the book ]pat [from Peter ]ag




Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik   2
An example for a Frame Instantiation


  • COMMERCIAL_TRANSACTION
         - SELLER: Airbus
         - BUYER: China Southern
         - GOODS: five A380 superjumbo planes
         - PRICE: 220 million Euro




Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik   3




                       Frame-to-Frame-Relations




Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik   4
Event Semantics: Donald Davidson's Problem

(1) The gardener killed the baron at midnight in the park
(2) The gardener killed the baron at midnight
(3) The gardener killed the baron in the park
(4) The gardener killed the baron




Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik   5




         The interpretation of adjunct constructions

• First attempt:
• (1) ⇒ kill4(g, b, m, p)
  (2) ⇒ kill3(g, b, m)
  (3) ⇒ kill2(g, b, p)
  (4) ⇒ kill1(g, b)




Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik   6
The interpretation of adjunct constructions

• A problem: How can the logical entailment relations
  between the different uses of kill be explained?

                                            (1)

                                     (2)            (3)

                                             (4)




Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik   7




         The interpretation of adjunct constructions

• Naïve FOL interpretation does not solve the problem:

     – kill4(g, b, m, p) I≠ kill3(g, b, m)
     – kill3(g, b, m) I≠ kill1(g, b)
     – etc.




Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik   8
Adjunct Interpretation: Second Attempt

• Fixed arity of the underlying predicate; syntactically not
  realized predicates are existentially bound:
    (1) ⇒ kill(g, b, m, p)
    (2) ⇒ ∃y kill(g, b, m, y)
    (3) ⇒ ∃x kill(g, b, x, p)
    (4) ⇒ ∃x∃y kill(g, b, x, y)




Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik   9




                                Another Problem

• The entailment relations are obtained, but:
• What is the correct arity of an event verb/ its underlying
  predicate?
    The gardener killed the baron at midnight in the park
    under cover of absolute darkness with a shotgun …
• Also the order of adjuncts (as compared to complements)
  (usually) has no impact on the truth conditions of the
  sentence.




Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik   10
Third Attempt: Higher-order Logic

• Adjuncts are analysed as sentence modifiers (type <t,t>):
    (1) ⇒ in the park(at-midight(kill(g, b)))
• The analysis solves the arity problem, but entailment
  relations arelost again:
    at-midnight(kill(g, b)) I≠ kill(g, b)
• Note also that the order of adjuncts (as compared to
  complements) (usually) has no impact on the truth
  conditions of the sentence.




Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik   11




                              Davidson's Solution

• Verbs expressing events have an additional argument
  position for an event variable, which is not realised at
  linguistic surface:
       kill ⇒ λxλyλe.kill(e,x,y), where kill: <e,<e,<e,t>>>
• Adjuncts express two-place relations between events and
  the respective quot;cirumstantial entitiesquot; (a time, a location,
  ...)
• In finite/tensed clauses, the event variable is existentially
  bound:
  The gardener killed the baron at midnight in the park
       ⇒ ∃e[ kill(e,g,b) ∧ time(e, m) ∧ location(e, p) ]

Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik   12
Davidson's Solution

• Event verbs are represented by relations of a fixed arity
  (number of syntactic complements +1)
• Event verbs have an argument position occupied by an
  event variable.
• Adjuncts are represented by two-place relations.
• Entailments follow straightforwardly, as well as the fact
  that adjunct semantics is order-independent:
•         ∃e[ kill(e,g,b) ∧ time(e, m) ∧ location(e, p) ]
          I= ∃e[ kill(e,g,b) ∧ time(e, m) ]
          I= ∃e[ kill(e,g,b) ]
Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik      13




 Compositional Semantics for Adjunct constructions

• Adjuncts are analysed as intersective modifiers for event predicates
  (type: <<e,t>,<e,t>>), in full analogy to intersective noun modifiers
  (adjectiveds, PPs):
     – red ⇒ λFλx[F(x) ∧ red(x)]
     – at midnight ⇒ λEλe[E(e) ∧ time(e, midnight)]


    The gardener killed the baron at midnight
    ⇒ λEλe[E(e) ∧ time(e, midnight)](λe.kill(e, g, b))
    ⇔ λe.kill(e, g, b) ∧ time(e, midnight)
    In finite clauses, the event variable is eventually bound:
    ⇒ ∃e.kill(e, g, b) ∧ time(e, midnight)


Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik      14
Uniform treatment of modifiers

• One semantic representation for the use of PPs as adjuncts and
  postnominal modifiers:
    in the park ⇒ λFλx[F(x) ∧ location(x, p)]
• Local adjunct /event modifier
    [[The gardener killed the baron ] in the park]
• Post-nominal modifier of an event-denoting deverbal noun:
    The [[murder] in the park]
• Post-nominal modifier of an standard common noun:
    The [[pavillon] in the park]
    Note: Event semantics provides a natural interpretation for deverbal
    common nouns.


Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik       15




               „Neo-Davisonian“ Event Semantics

• Complements can be treated analogously to adjuncts: Event verbs
  are represented as one-place event predicates. Thematic roles are
  two-place relations linking arguments to the event denoted by the
  verb:
    The gardener killed the baron at midnight in the park
    ⇒ ∃e [kill(e) ∧ ag(e,g) ∧ pat(e,b) ∧ time(e,m) ∧ location(e,p)]


• Neo-Davidsonian semantics allows the partioning of semantic
  information into minimal pieces, but:


• Proper interpretation of the role relations anticipates knowledge of the
  event predicate, to some extent.

Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik       16
Event anaphora in DRT

• The gardener killed the baron . It happened at midnight.
• Yesterday, I went by train from Hamburg to Saarbrücken.
  That was a boring trip.


• Event referents
     – a new kind of discourse referents
     – are introduced (e.g.) by finite clauses
     – and can be referred to by nominal anaphoric
       expressions


Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik   17




                          Event anaphora in DRT

 •The gardener killed the baron . It happened at midnight.


                e, g, b                           e, g, b,e'

                gardener(g)                       gardener(g)
                baron(b)                          baron(b)
                kill(e,g,b)                       kill(e,g,b)
                                                  midnight(m)
                                                  time(e',m)
                                                  e'=e


Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik   18
FOL Model Structure with Events

• Like standard FOL Model Structure M = <U,V>, except
  that the universe is subdivided into
     – a set of standard individuals US, and
     – a set of events UE, which is partially ordered by a
       quot;temporally precedesquot; relation.




Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik   19




         Temporal relations in an Event Semantics

• Event Semantics allows the explicit representation of
  tense and temporal relations in FOL/DRT
    John left        ⇒ ∃e[ leave(e, j*) ∧ e < eu ]
    where < is interpreted as temporal precedence, and is the
    utterance event.
    John left, after Peter had arrived
    ⇒ ∃e1 ∃e2[ leave(e1, j*) ∧ e1 < eu ∧ arrive(e2, p) ∧ e2 < e1 ]




Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik   20
Temporal relations in an Event Semantics

                John left, after Peter had arrived


                                j, e, p, e'

                                leave(e,j)
                                e< eu

                                arrive(e',p)
                                e'< e



Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik   21




                          Events, activities, states

• Davidsonian event semantics works well for verbs expressing
  individual events that have a specific temporal location (like in The
  gardener killed the baron or John left).
• Activities (John is walking, working), usually expressed by the
  progressive form in English, lack a precise temporal location: If I am
  working during a time interval, I am also working during all sub-
  intervals – representation via discourse referents is problematic.
• Events and activities are usually subsumed under the common
  concept of quot;eventualitiesquot;, in contrast to states (John lives in
  Saarbrücken, John likes Mary).
• Event-denoting expressions resemble ordinary quot;countablequot; common
  nouns. Activities and states are semantically similar to plurals and
  mass nouns.


Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik   22
What else?

• Semantics of tense and aspect is a large and important
  research area in natural language semantics.
• Unfortunately, we had no opportunity to look into it, as
  well as into many other interesting questions of NL
  Semantics, e.g. the semantics of
     – plurals, mass nouns, collective predicates
     – spatial prepositions
     – adjectives, comparatives, superlatives
     – vague expressions
     – metaphors and metonymies
Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik   23

More Related Content

More from Dr. Cupid Lucid

More from Dr. Cupid Lucid (20)

Content Analysis vs secondary analysis
Content Analysis vs secondary analysisContent Analysis vs secondary analysis
Content Analysis vs secondary analysis
 
Basic elements of scientific concepts
Basic elements of scientific  conceptsBasic elements of scientific  concepts
Basic elements of scientific concepts
 
Observational methods
Observational methodsObservational methods
Observational methods
 
Types of educational_research
Types of educational_researchTypes of educational_research
Types of educational_research
 
Types of research
Types of researchTypes of research
Types of research
 
Types of Research
Types of ResearchTypes of Research
Types of Research
 
Literature what is it
Literature what is itLiterature what is it
Literature what is it
 
History of english literature sajid
History of english literature sajidHistory of english literature sajid
History of english literature sajid
 
A guide to_writing_research_papers
A guide to_writing_research_papersA guide to_writing_research_papers
A guide to_writing_research_papers
 
What isliterature
What isliteratureWhat isliterature
What isliterature
 
101 masterpieces of literature in english
101 masterpieces of literature in english101 masterpieces of literature in english
101 masterpieces of literature in english
 
National Curriculum of English Grade I-XII
National Curriculum of English Grade I-XIINational Curriculum of English Grade I-XII
National Curriculum of English Grade I-XII
 
The Linguistic Variables
The Linguistic VariablesThe Linguistic Variables
The Linguistic Variables
 
Term Paper
Term PaperTerm Paper
Term Paper
 
Syllabus Designing
Syllabus DesigningSyllabus Designing
Syllabus Designing
 
Semiotics Final
Semiotics FinalSemiotics Final
Semiotics Final
 
Research Proposal Methoo
Research Proposal MethooResearch Proposal Methoo
Research Proposal Methoo
 
Questionnaire
QuestionnaireQuestionnaire
Questionnaire
 
Pygmalion
PygmalionPygmalion
Pygmalion
 
Model Of Cohesion
Model Of CohesionModel Of Cohesion
Model Of Cohesion
 

Recently uploaded

Apidays New York 2024 - The value of a flexible API Management solution for O...
Apidays New York 2024 - The value of a flexible API Management solution for O...Apidays New York 2024 - The value of a flexible API Management solution for O...
Apidays New York 2024 - The value of a flexible API Management solution for O...apidays
 
2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...Martijn de Jong
 
Strategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a FresherStrategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a FresherRemote DBA Services
 
Why Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire businessWhy Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire businesspanagenda
 
EMPOWERMENT TECHNOLOGY GRADE 11 QUARTER 2 REVIEWER
EMPOWERMENT TECHNOLOGY GRADE 11 QUARTER 2 REVIEWEREMPOWERMENT TECHNOLOGY GRADE 11 QUARTER 2 REVIEWER
EMPOWERMENT TECHNOLOGY GRADE 11 QUARTER 2 REVIEWERMadyBayot
 
Artificial Intelligence Chap.5 : Uncertainty
Artificial Intelligence Chap.5 : UncertaintyArtificial Intelligence Chap.5 : Uncertainty
Artificial Intelligence Chap.5 : UncertaintyKhushali Kathiriya
 
Apidays New York 2024 - Passkeys: Developing APIs to enable passwordless auth...
Apidays New York 2024 - Passkeys: Developing APIs to enable passwordless auth...Apidays New York 2024 - Passkeys: Developing APIs to enable passwordless auth...
Apidays New York 2024 - Passkeys: Developing APIs to enable passwordless auth...apidays
 
CNIC Information System with Pakdata Cf In Pakistan
CNIC Information System with Pakdata Cf In PakistanCNIC Information System with Pakdata Cf In Pakistan
CNIC Information System with Pakdata Cf In Pakistandanishmna97
 
Boost Fertility New Invention Ups Success Rates.pdf
Boost Fertility New Invention Ups Success Rates.pdfBoost Fertility New Invention Ups Success Rates.pdf
Boost Fertility New Invention Ups Success Rates.pdfsudhanshuwaghmare1
 
Biography Of Angeliki Cooney | Senior Vice President Life Sciences | Albany, ...
Biography Of Angeliki Cooney | Senior Vice President Life Sciences | Albany, ...Biography Of Angeliki Cooney | Senior Vice President Life Sciences | Albany, ...
Biography Of Angeliki Cooney | Senior Vice President Life Sciences | Albany, ...Angeliki Cooney
 
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...apidays
 
MINDCTI Revenue Release Quarter One 2024
MINDCTI Revenue Release Quarter One 2024MINDCTI Revenue Release Quarter One 2024
MINDCTI Revenue Release Quarter One 2024MIND CTI
 
Ransomware_Q4_2023. The report. [EN].pdf
Ransomware_Q4_2023. The report. [EN].pdfRansomware_Q4_2023. The report. [EN].pdf
Ransomware_Q4_2023. The report. [EN].pdfOverkill Security
 
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers:  A Deep Dive into Serverless Spatial Data and FMECloud Frontiers:  A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FMESafe Software
 
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...apidays
 
presentation ICT roal in 21st century education
presentation ICT roal in 21st century educationpresentation ICT roal in 21st century education
presentation ICT roal in 21st century educationjfdjdjcjdnsjd
 
Emergent Methods: Multi-lingual narrative tracking in the news - real-time ex...
Emergent Methods: Multi-lingual narrative tracking in the news - real-time ex...Emergent Methods: Multi-lingual narrative tracking in the news - real-time ex...
Emergent Methods: Multi-lingual narrative tracking in the news - real-time ex...Zilliz
 
Apidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, Adobe
Apidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, AdobeApidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, Adobe
Apidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, Adobeapidays
 
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerHow to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerThousandEyes
 

Recently uploaded (20)

Apidays New York 2024 - The value of a flexible API Management solution for O...
Apidays New York 2024 - The value of a flexible API Management solution for O...Apidays New York 2024 - The value of a flexible API Management solution for O...
Apidays New York 2024 - The value of a flexible API Management solution for O...
 
2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...
 
Strategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a FresherStrategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a Fresher
 
Why Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire businessWhy Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire business
 
EMPOWERMENT TECHNOLOGY GRADE 11 QUARTER 2 REVIEWER
EMPOWERMENT TECHNOLOGY GRADE 11 QUARTER 2 REVIEWEREMPOWERMENT TECHNOLOGY GRADE 11 QUARTER 2 REVIEWER
EMPOWERMENT TECHNOLOGY GRADE 11 QUARTER 2 REVIEWER
 
Artificial Intelligence Chap.5 : Uncertainty
Artificial Intelligence Chap.5 : UncertaintyArtificial Intelligence Chap.5 : Uncertainty
Artificial Intelligence Chap.5 : Uncertainty
 
Apidays New York 2024 - Passkeys: Developing APIs to enable passwordless auth...
Apidays New York 2024 - Passkeys: Developing APIs to enable passwordless auth...Apidays New York 2024 - Passkeys: Developing APIs to enable passwordless auth...
Apidays New York 2024 - Passkeys: Developing APIs to enable passwordless auth...
 
CNIC Information System with Pakdata Cf In Pakistan
CNIC Information System with Pakdata Cf In PakistanCNIC Information System with Pakdata Cf In Pakistan
CNIC Information System with Pakdata Cf In Pakistan
 
Boost Fertility New Invention Ups Success Rates.pdf
Boost Fertility New Invention Ups Success Rates.pdfBoost Fertility New Invention Ups Success Rates.pdf
Boost Fertility New Invention Ups Success Rates.pdf
 
Biography Of Angeliki Cooney | Senior Vice President Life Sciences | Albany, ...
Biography Of Angeliki Cooney | Senior Vice President Life Sciences | Albany, ...Biography Of Angeliki Cooney | Senior Vice President Life Sciences | Albany, ...
Biography Of Angeliki Cooney | Senior Vice President Life Sciences | Albany, ...
 
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
 
MINDCTI Revenue Release Quarter One 2024
MINDCTI Revenue Release Quarter One 2024MINDCTI Revenue Release Quarter One 2024
MINDCTI Revenue Release Quarter One 2024
 
Ransomware_Q4_2023. The report. [EN].pdf
Ransomware_Q4_2023. The report. [EN].pdfRansomware_Q4_2023. The report. [EN].pdf
Ransomware_Q4_2023. The report. [EN].pdf
 
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers:  A Deep Dive into Serverless Spatial Data and FMECloud Frontiers:  A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FME
 
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...
 
presentation ICT roal in 21st century education
presentation ICT roal in 21st century educationpresentation ICT roal in 21st century education
presentation ICT roal in 21st century education
 
Emergent Methods: Multi-lingual narrative tracking in the news - real-time ex...
Emergent Methods: Multi-lingual narrative tracking in the news - real-time ex...Emergent Methods: Multi-lingual narrative tracking in the news - real-time ex...
Emergent Methods: Multi-lingual narrative tracking in the news - real-time ex...
 
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
 
Apidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, Adobe
Apidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, AdobeApidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, Adobe
Apidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, Adobe
 
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerHow to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
 

Lect13

  • 1. Semantic Theory Summer 2006 Lexical Semantics II: Event Semantics M. Pinkal Examples Annotated with Thematic Roles – [The window]pat broke – [A rock ]inst broke [the window ]pat – [John ]ag broke [the window ]pat [with a rock ]inst – [Peter ]ag gave [Mary]rec [the book ]pat – [Mary ]rec received [the book ]pat [from Peter ]ag Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 2
  • 2. An example for a Frame Instantiation • COMMERCIAL_TRANSACTION - SELLER: Airbus - BUYER: China Southern - GOODS: five A380 superjumbo planes - PRICE: 220 million Euro Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 3 Frame-to-Frame-Relations Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 4
  • 3. Event Semantics: Donald Davidson's Problem (1) The gardener killed the baron at midnight in the park (2) The gardener killed the baron at midnight (3) The gardener killed the baron in the park (4) The gardener killed the baron Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 5 The interpretation of adjunct constructions • First attempt: • (1) ⇒ kill4(g, b, m, p) (2) ⇒ kill3(g, b, m) (3) ⇒ kill2(g, b, p) (4) ⇒ kill1(g, b) Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 6
  • 4. The interpretation of adjunct constructions • A problem: How can the logical entailment relations between the different uses of kill be explained? (1) (2) (3) (4) Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 7 The interpretation of adjunct constructions • Naïve FOL interpretation does not solve the problem: – kill4(g, b, m, p) I≠ kill3(g, b, m) – kill3(g, b, m) I≠ kill1(g, b) – etc. Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 8
  • 5. Adjunct Interpretation: Second Attempt • Fixed arity of the underlying predicate; syntactically not realized predicates are existentially bound: (1) ⇒ kill(g, b, m, p) (2) ⇒ ∃y kill(g, b, m, y) (3) ⇒ ∃x kill(g, b, x, p) (4) ⇒ ∃x∃y kill(g, b, x, y) Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 9 Another Problem • The entailment relations are obtained, but: • What is the correct arity of an event verb/ its underlying predicate? The gardener killed the baron at midnight in the park under cover of absolute darkness with a shotgun … • Also the order of adjuncts (as compared to complements) (usually) has no impact on the truth conditions of the sentence. Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 10
  • 6. Third Attempt: Higher-order Logic • Adjuncts are analysed as sentence modifiers (type <t,t>): (1) ⇒ in the park(at-midight(kill(g, b))) • The analysis solves the arity problem, but entailment relations arelost again: at-midnight(kill(g, b)) I≠ kill(g, b) • Note also that the order of adjuncts (as compared to complements) (usually) has no impact on the truth conditions of the sentence. Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 11 Davidson's Solution • Verbs expressing events have an additional argument position for an event variable, which is not realised at linguistic surface: kill ⇒ λxλyλe.kill(e,x,y), where kill: <e,<e,<e,t>>> • Adjuncts express two-place relations between events and the respective quot;cirumstantial entitiesquot; (a time, a location, ...) • In finite/tensed clauses, the event variable is existentially bound: The gardener killed the baron at midnight in the park ⇒ ∃e[ kill(e,g,b) ∧ time(e, m) ∧ location(e, p) ] Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 12
  • 7. Davidson's Solution • Event verbs are represented by relations of a fixed arity (number of syntactic complements +1) • Event verbs have an argument position occupied by an event variable. • Adjuncts are represented by two-place relations. • Entailments follow straightforwardly, as well as the fact that adjunct semantics is order-independent: • ∃e[ kill(e,g,b) ∧ time(e, m) ∧ location(e, p) ] I= ∃e[ kill(e,g,b) ∧ time(e, m) ] I= ∃e[ kill(e,g,b) ] Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 13 Compositional Semantics for Adjunct constructions • Adjuncts are analysed as intersective modifiers for event predicates (type: <<e,t>,<e,t>>), in full analogy to intersective noun modifiers (adjectiveds, PPs): – red ⇒ λFλx[F(x) ∧ red(x)] – at midnight ⇒ λEλe[E(e) ∧ time(e, midnight)] The gardener killed the baron at midnight ⇒ λEλe[E(e) ∧ time(e, midnight)](λe.kill(e, g, b)) ⇔ λe.kill(e, g, b) ∧ time(e, midnight) In finite clauses, the event variable is eventually bound: ⇒ ∃e.kill(e, g, b) ∧ time(e, midnight) Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 14
  • 8. Uniform treatment of modifiers • One semantic representation for the use of PPs as adjuncts and postnominal modifiers: in the park ⇒ λFλx[F(x) ∧ location(x, p)] • Local adjunct /event modifier [[The gardener killed the baron ] in the park] • Post-nominal modifier of an event-denoting deverbal noun: The [[murder] in the park] • Post-nominal modifier of an standard common noun: The [[pavillon] in the park] Note: Event semantics provides a natural interpretation for deverbal common nouns. Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 15 „Neo-Davisonian“ Event Semantics • Complements can be treated analogously to adjuncts: Event verbs are represented as one-place event predicates. Thematic roles are two-place relations linking arguments to the event denoted by the verb: The gardener killed the baron at midnight in the park ⇒ ∃e [kill(e) ∧ ag(e,g) ∧ pat(e,b) ∧ time(e,m) ∧ location(e,p)] • Neo-Davidsonian semantics allows the partioning of semantic information into minimal pieces, but: • Proper interpretation of the role relations anticipates knowledge of the event predicate, to some extent. Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 16
  • 9. Event anaphora in DRT • The gardener killed the baron . It happened at midnight. • Yesterday, I went by train from Hamburg to Saarbrücken. That was a boring trip. • Event referents – a new kind of discourse referents – are introduced (e.g.) by finite clauses – and can be referred to by nominal anaphoric expressions Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 17 Event anaphora in DRT •The gardener killed the baron . It happened at midnight. e, g, b e, g, b,e' gardener(g) gardener(g) baron(b) baron(b) kill(e,g,b) kill(e,g,b) midnight(m) time(e',m) e'=e Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 18
  • 10. FOL Model Structure with Events • Like standard FOL Model Structure M = <U,V>, except that the universe is subdivided into – a set of standard individuals US, and – a set of events UE, which is partially ordered by a quot;temporally precedesquot; relation. Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 19 Temporal relations in an Event Semantics • Event Semantics allows the explicit representation of tense and temporal relations in FOL/DRT John left ⇒ ∃e[ leave(e, j*) ∧ e < eu ] where < is interpreted as temporal precedence, and is the utterance event. John left, after Peter had arrived ⇒ ∃e1 ∃e2[ leave(e1, j*) ∧ e1 < eu ∧ arrive(e2, p) ∧ e2 < e1 ] Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 20
  • 11. Temporal relations in an Event Semantics John left, after Peter had arrived j, e, p, e' leave(e,j) e< eu arrive(e',p) e'< e Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 21 Events, activities, states • Davidsonian event semantics works well for verbs expressing individual events that have a specific temporal location (like in The gardener killed the baron or John left). • Activities (John is walking, working), usually expressed by the progressive form in English, lack a precise temporal location: If I am working during a time interval, I am also working during all sub- intervals – representation via discourse referents is problematic. • Events and activities are usually subsumed under the common concept of quot;eventualitiesquot;, in contrast to states (John lives in Saarbrücken, John likes Mary). • Event-denoting expressions resemble ordinary quot;countablequot; common nouns. Activities and states are semantically similar to plurals and mass nouns. Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 22
  • 12. What else? • Semantics of tense and aspect is a large and important research area in natural language semantics. • Unfortunately, we had no opportunity to look into it, as well as into many other interesting questions of NL Semantics, e.g. the semantics of – plurals, mass nouns, collective predicates – spatial prepositions – adjectives, comparatives, superlatives – vague expressions – metaphors and metonymies Vorlesung Semantik 2006 © M. Pinkal/A. Koller UdS Computerlinguistik 23