Скачать презентацию Argument structure and its realization 2 21 2005 Nianwen Xue Скачать презентацию Argument structure and its realization 2 21 2005 Nianwen Xue

bb4b1d0e4d293f530160c0ff10e490f9.ppt

  • Количество слайдов: 35

Argument structure and its realization 2/21/2005 Nianwen Xue Argument structure and its realization 2/21/2005 Nianwen Xue

outline • • • Syntactic argument vs semantic argument Selectional restriction Subject control Object outline • • • Syntactic argument vs semantic argument Selectional restriction Subject control Object control Raising

Semantic valency • Arguments can be viewed as central participants in a situation w Semantic valency • Arguments can be viewed as central participants in a situation w Aous laughed. Aous = the entity that does the laughing w Chris invited Dave. Chris = host Guest = Dave

Semantic valency • One-place predicates w Laugh, smile, • Two-place predicates w Invite, hit, Semantic valency • One-place predicates w Laugh, smile, • Two-place predicates w Invite, hit, • Three-place predicates w give, send, etc. • Four-place predicates w Accompany, • Five-place predicates w Rent,

Arguments defined in the frame files • http: //verbs. colorado. edu/framesets w w Smile Arguments defined in the frame files • http: //verbs. colorado. edu/framesets w w Smile Invite Accompany Rent

Syntactic argument (transitivity) • Arguments that are actually realized w Intransitive: takes one argument Syntactic argument (transitivity) • Arguments that are actually realized w Intransitive: takes one argument w Transitive: takes two arguments w Ditransitive: takes three arguments

Transitive to intransitive w He / interrupted / the meeting. He / interrupted. w Transitive to intransitive w He / interrupted / the meeting. He / interrupted. w Amy / knits / sweaters. Amy / knits. w They / are reading / a book. They / are reading.

Intransitive to transitive w Dennis / died / a peaceful death. w Lukas / Intransitive to transitive w Dennis / died / a peaceful death. w Lukas / laughed / an infectious laugh. w Mona Lisa / was smiling / a mysterious smile.

Ditransitive-intransitive w We teach college students syntax. We teach college students. We teach syntax. Ditransitive-intransitive w We teach college students syntax. We teach college students. We teach syntax. We teach. w He told me the whole story. He told me. He told the whole story. He better not tell.

Transitive to ditransitive w I baked a delicious cake. I baked my friends a Transitive to ditransitive w I baked a delicious cake. I baked my friends a delicious cake. w The lions killed a gazelle. The lions killed themselves a gazelle. w She sang a lullaby. She sang her baby a lullaby.

Alternation • Transitive to intransitive • Object to subject w Aous broke the window. Alternation • Transitive to intransitive • Object to subject w Aous broke the window. w The window broke. • Changes of (syntactic) arguments • Change of position Levin (1993)

Syntax/semantics mismatch • Semantic arguments are not always explicitly realized • Realized syntactic arguments Syntax/semantics mismatch • Semantic arguments are not always explicitly realized • Realized syntactic arguments are not necessarily semantic arguments • Semantic arguments are not always realized in the same syntactic positions

Modification • Properties of events, which are less central, are often realized as modifiers: Modification • Properties of events, which are less central, are often realized as modifiers: w w w w Manner: He read the letter carefully. Point in time: They discussed the proposal in the afternoon. Duration: You should keep your tax records for several years. Frequency: I read the Times quite often. Location: We met in my office. Origin: We set out from Bangalore. Destination: We arrived in Benares. Cause: He threw it away out of spite. w Purpose: I sent the message to warn everyone.

Predication and subject requirement • That they are corrupt is evident. Their corruption is Predication and subject requirement • That they are corrupt is evident. Their corruption is evident. • It is evident that they are corrupt. * Is evident that they are corrupt. • Does “it’ contribute to the semantics of this sentence? If not, why is it needed?

Pleonastic ‘there’ • Several vexing questions remain. There remain several vexing questions. • There Pleonastic ‘there’ • Several vexing questions remain. There remain several vexing questions. • There is a clean shirt right here. There is a clean shirt over there. • * Remain several vexing questions.

Selectional restrictions • Amy drank the { lemonade, #sandwich } Lukas drank a whole Selectional restrictions • Amy drank the { lemonade, #sandwich } Lukas drank a whole { quart, #piece } • { Two hours, #the shift, #two liters, #Larry } elapsed without further incident. • The { paramilitary, #bomb, #avalanche } murdered { her husband, #the olive tree, #her house }. • The { paramilitary, bomb, avalanche } killed { her husband, the olive tree, #her house }.

Agree (1) The children agreed to dance. How many clauses does this sentence have? Agree (1) The children agreed to dance. How many clauses does this sentence have?

agree (2) • The childreni agreed [that theyi would get wet ]. # The agree (2) • The childreni agreed [that theyi would get wet ]. # The { horses, trees, rocks }i agreed [that theyi would get wet ]. • The childreni agreed [that theyi would speak Twi ]. # The childreni agreed [that theyi would { elapse, evaporate } ]

agree (3) • The children agreed [*PRO* to get wet ]. # The { agree (3) • The children agreed [*PRO* to get wet ]. # The { horses, trees, rocks } agreed [*PRO* to get wet] . • The children agreed [*PRO* to speak Twi ]. # The children agreed [*PRO* to { elapse, evaporate } ]. • The subject control verb and the lower verb impose separate selectional restrictions on their subject. • Therefore two clauses

Representing subject control structure S NP-1 VP S VV NP VP TO VP The Representing subject control structure S NP-1 VP S VV NP VP TO VP The children agreed *PRO*-1 to get wet.

Properties of subject control • *PRO* can’t be replaced with an overt NP • Properties of subject control • *PRO* can’t be replaced with an overt NP • *PRO* must be coreferential with the subject NP in the matrix clause

Arbitrary *PRO* • [*PRO* to err] is human. • [*PRO* to forgive] is divine. Arbitrary *PRO* • [*PRO* to err] is human. • [*PRO* to forgive] is divine.

Small *pro* in Chinese S VP NP NP VC CP IP NP QP ADJP Small *pro* in Chinese S VP NP NP VC CP IP NP QP ADJP NP DEC VP VV NP 建筑 是 *pro* 开发 浦东 的 一 项 主要 经济 活 construction be develop Pudong DE one CL main economic activity “Construction is a main economic activity in developing Pudong. ”

Seem (1) • The children seemed [ to dance ]. • * There agreed Seem (1) • The children seemed [ to dance ]. • * There agreed to be a problem. ok There seemed to be a problem. • The children agreed [ PRO to get wet ]. # The { horses, trees, rocks } agreed [ PRO to get wet ]. • ok The children seemed to get wet. ok The { horses, trees, rocks } seemed to get wet.

Seem (2) • Subject idiom chunks w # The cat agreed [ PRO to Seem (2) • Subject idiom chunks w # The cat agreed [ PRO to be out of the bag ]. w [The cat]i seems [ ti to be out of the bag ]. • Weather it: w # It agreed to be raining. w It seems to to be raining. Seem does not impose selectional restrictions

Seem (3) • It seems that he is happy. • *That he is happy Seem (3) • It seems that he is happy. • *That he is happy seems. “Seem” does not take an argument at its subject position.

Representing raising S VP NP-1 S VBP NP-1 TO VP VP VB The children Representing raising S VP NP-1 S VBP NP-1 TO VP VP VB The children seems *-1 to dance.

Representing raising S VP NP VBP VP VP TO VB There seems to be Representing raising S VP NP VBP VP VP TO VB There seems to be NP a problem.

Properties of raising verbs • Do not impose selectional restrictions on its subject. Selectional Properties of raising verbs • Do not impose selectional restrictions on its subject. Selectional restrictions on its subject imposed by the lower verb • Are not associated with a thematic role

Persuade vs expect We persuaded the children to dance. • We expected there to Persuade vs expect We persuaded the children to dance. • We expected there to be a problem. • # We persuaded there to be a problem. • We expected it to rain. • #We persuaded it to rain. • We expect the cat to be out of the bag. • #We persuaded the cat to be out of the bag.

Representing object-control S VP NP VBD NP S NP We persuaded the children *PRO* Representing object-control S VP NP VBD NP S NP We persuaded the children *PRO* VP to dance.

Exercise • Is “tend” a control verb or raising verb? • What about “promise”? Exercise • Is “tend” a control verb or raising verb? • What about “promise”?

tend • There tend to be huge traffic jams during rush hour. • It tend • There tend to be huge traffic jams during rush hour. • It tends to rain at night. • The cat tends to be out of the bag. (? )

promise • Control w The { children, #horses } promised [ to eat their promise • Control w The { children, #horses } promised [ to eat their oatmeal ]. w The { children, #horses } promised [ that they would eat their oatmeal ]. w The children softly promised [ to eat their oatmeal ]. w The children obediently promised [ to eat their oatmeal ]. • Raising w This filly promises to win the race. w All available evidence indicated that this filly will win the race. w There promises to be a new version by spring.

Homework • Treebank the following sentences w Criminal charges are not filed in shootings Homework • Treebank the following sentences w Criminal charges are not filed in shootings that authorities determine to be accidental. w Two Republican governors on Monday questioned a Bush administration decision allowing an Arab-owned company to operate six major U. S. ports, saying they may try to cancel lease arrangements at ports in their states.