Comments on Jakub Dotlačil’s presentation, ‘Dynamic properties of question words’ Matthew Gotham University of Oslo and Centre for Advanced Study PALLMYR-XI, UCL, 10 November 2017 Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 1 / 9
Making ICDRT Inquisitive semantics Who is walking? for questions Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 2 / 9
Making ICDRT Inquisitive semantics Who is walking? for questions + Someone 1 is walking. He 1 is singing. CDRT for anaphora Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 2 / 9
Making ICDRT Inquisitive semantics Who is walking? for questions + Someone 1 is walking. He 1 is singing. CDRT for anaphora = Who 1 is walking? Is he 1 singing? ICDRT for anaphora to wh-words Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 2 / 9
� � � � ‘Someone is walking’ standard static a proposition λ w s . ∃ x e . walk ( x , w ) inquisitive CDRT a set of propositions a state–state relation λ i c .λ o c . i [ x 1 ] o ∧ walk ( x 1 ( o )) λ p s � t . ∃ x e . p ⊆ ( λ w s . walk ( x , w )) ICDRT a proposition–state–state relation λ p s � t .λ i c .λ o c . i [ x 1 ] o ∧ p ⊆ ( λ w s . walk ( x 1 ( o ) , w )) Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 3 / 9
Anaphora to wh-words Why does it work? Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 4 / 9
Anaphora to wh-words Why does it work? Basically, because in ICDRT wh-words words are treated the same as indefinites, semantically.* Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 4 / 9
Anaphora to wh-words Why does it work? Basically, because in ICDRT wh-words words are treated the same as indefinites, semantically.* � Someone 1 is walking. You know him 1 . � = � Who 1 is walking? You know him 1 . � = λ p s � t .λ i c .λ o c . i [ x 1 ] o ∧ p ⊆ ( λ w s . walk ( x 1 ( o ))) ∧ p ⊆ ( λ w s . know ( you , x 1 ( o ) , w )) *To be qualified. Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 4 / 9
Why can’t non-inquisitive semantics just do the same thing? I.e., treat constituent questions just like existential statements Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 5 / 9
Why can’t non-inquisitive semantics just do the same thing? I.e., treat constituent questions just like existential statements It’s a distinctive feature of inquisitive semantics that statements and quesitons have the same semantic type. Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 5 / 9
Why can’t non-inquisitive semantics just do the same thing? I.e., treat constituent questions just like existential statements It’s a distinctive feature of inquisitive semantics that statements and quesitons have the same semantic type. To adapt a non-inquisitive system to treat indefinites and wh-words the same, you would have to either raise statements to the type of questions ...in which case you would have an inquisitive system, or Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 5 / 9
Why can’t non-inquisitive semantics just do the same thing? I.e., treat constituent questions just like existential statements It’s a distinctive feature of inquisitive semantics that statements and quesitons have the same semantic type. To adapt a non-inquisitive system to treat indefinites and wh-words the same, you would have to either raise statements to the type of questions ...in which case you would have an inquisitive system, or lower questions to the type of statements ...in which case you the treatment of questions would be inadequate. Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 5 / 9
Why can’t non-inquisitive semantics just do the same thing? I.e., treat constituent questions just like existential statements It’s a distinctive feature of inquisitive semantics that statements and quesitons have the same semantic type. To adapt a non-inquisitive system to treat indefinites and wh-words the same, you would have to either raise statements to the type of questions ...in which case you would have an inquisitive system, or lower questions to the type of statements ...in which case you the treatment of questions would be inadequate. Inquisitive semantics gives you the notion of answers to the question ( resolutions to the issue ). p s � t resolves φ ( s � t ) � t ⇔ φ ( p ) E.g. ( λ w s . walk ( john , w )) resolves ( λ p s � t . ∃ x e . p ⊆ ( λ w s . walk ( x , w ))) Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 5 / 9
But still... There has to be some semantic difference between indefinites and wh-words. ( I know someone’s walking vs. I know who’s walking .) Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 6 / 9
But still... There has to be some semantic difference between indefinites and wh-words. ( I know someone’s walking vs. I know who’s walking .) Proposal (from ms. Jakub sent me): unlike Someone is walking. , Who is walking? presupposes that someone is walking. Q # Who is walking? A # Someone. Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 6 / 9
But still... There has to be some semantic difference between indefinites and wh-words. ( I know someone’s walking vs. I know who’s walking .) Proposal (from ms. Jakub sent me): unlike Someone is walking. , Who is walking? presupposes that someone is walking. Q # Who is walking? A # Someone. As far as I know, this proposal hasn’t been formalized. Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 6 / 9
� � � � A suggestion ‘Someone is walking’ standard static a proposition λ w s . ∃ x e . walk ( x , w ) inquisitive CDRT a set of propositions a state–state relation λ i c .λ o c . i [ x 1 ] o ∧ walk ( x 1 ( o )) λ p s � t . ∃ x e . p ⊆ ( λ w s . walk ( x , w )) var-ICDRT a proposition–proposition–state–state relation λ p s � t .λ q s � t .λ i c .λ o c . i [ x 1 ] o ∧ q ⊆ p ∩ ( λ w s . walk ( x 1 ( o ) , w )) Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 7 / 9
The constrast in a trivalent semantics Someone 1 is walking. λ p s � t .λ q s � t .λ i c .λ o c . i [ x 1 ] o ∧ q ⊆ p ∩ λ w s . walk ( x 1 ( o ) , w ) � � Who 1 is walking? λ p s � t .λ q s � t .λ i c .λ o c . i [ x 1 ] o ∧ q ⊆ p ∩ λ w s . walk ( x 1 ( o ) , w ) � � p ⊆ λ w s . ∃ x e . walk ( x , w ) � � ∧ ∂ Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 8 / 9
Definitions These haven’t been properly checked yet... p s � t supports Φ ( s � t ) � ( s � t ) � c � c � t ⇔ ∀ i c . ∃ o c . Φ( λ w s . ⊤ )( p )( i )( o ) walks λ d c � e .λ p s � t .λ q s � t .λ i c .λ o c . i = o ∧ q ⊆ p ∩ λ w s . walk ( d ( o ) , w ) � � someone n λ P .λ p s � t .λ q s � t .λ i c .λ o c . ∃ k c . i [ x n ] k ∧ P ( x n )( p )( q )( k )( o ) who n λ P .λ p s � t .λ q s � t .λ i c .λ o c . ∃ k c . i [ x n ] k ∧ P ( x n )( p )( q )( k )( o ) � ∧ ∂ p supports λ r s � t .λ s s � t .λ j c .λ l c . ∃ x e . P ( λ m c . x )( r )( s )( j )( l ) � Where P :: ( c � e ) � ( s � t ) � ( s � t ) � c � c � t Matthew Gotham (Oslo and CAS) Comments on Dotlačil PALLMYR-XI, UCL, 10.11.2017 9 / 9
Recommend
More recommend
