440 likes | 590 Views
Lecture 4: Double Objects and Datives. Advanced Syntax. Universal Theta role Assignment Hypothesis Every argument bearing the same theta role is in the same structural position in all constructions at D-structure Therefore, we can identify what the agent, theme, etc. positions are.
E N D
Lecture 4: Double Objects and Datives Advanced Syntax
Universal Theta role Assignment Hypothesis • Every argument bearing the same theta role is in the same structural position in all constructions at D-structure • Therefore, we can identify what the agent, theme, etc. positions are Recap: VP structure theme agent experiencer
Verbs which have only theme arguments have a simple VP structure • John arrived • = unaccusatives • As lexical verbs do not assign Case and there is no abstract verb, the theme will move to the subject position Recap: VP structure
Verbs which have only an agent argument will have a more complex structure • John laughed • They have: • an abstract agentive verb with an agent specifier • A lexical verb with no argument • = unergative • The abstract verb assigns Case to the specifier of its complement • The agent does not get Case and therefore moves to subject position • The verb moves to support the abstract verb Recap: VP structure
Some verbs have an agent and a theme argument: • John killed Bill • They have: • An abstract agentive verb with agent argument • A lexical verb with a theme argument • = transitive • The abstract verb assigns Case to the theme • The agent does not get Case, so moves to subject • The verb moves to support the abstract verb Recap: VP structure
Verbs which have experiencer and theme arguments will have the same structure • John saw Bill • The abstract verb will have a different meaning (‘experience’ rather than ‘do’) • Also transitive • Case and movement relations will be the same Recap: VP structure
Some verbs have both agent and experiencer arguments • The clown scared the children • Must have two abstract verbs (agentive + experiencer) • The agentive verb will assign Case to the experiencer • The agent will not get Case and so will move to subject • The verb will move to support both abstract verbs • The experiencer is therefore the object Recap: VP structure
But there are verbs which have three arguments: • Complex transitives – agent + theme + location • [He] put [the book] [on the shelf] • Datives – agent + theme + goal • [He] gave [the money] [to his lawyer] • Double object verbs – agent + goal + theme • [He] wrote [Mary] [a letter] ThRee place predicates
Traditionally a double object verb is said to have two objects: • Indirect object = goal or beneficiary • He sent Mary a message • He knitted his granny a scarf • Direct object = theme • He sent Mary a message • He knitted his granny a scarf • The indirect object always precedes the direct object: • * they built a house him (they built him a house) Double Object (DO) construction
The indirect object shows more object properties than does the direct object: • It immediately follows the verb: • I saw her yesterday * I saw yesterday her • I owed her the money * I owed the money her • It moves to subject in passives: • They awarded her a medal • She was awarded a medal • * a medal was awarded her Double Object (DO) construction
The dative construction expresses something very close to the double object construction, but has a number of syntactic differences: • Both have similar arguments (agent, theme, goal/beneficiary) • Dative verbs have one DP object (theme) and one PP argument (goal/beneficiary) • They delivered the package to the shop • The goal is expressed as a to PP and the beneficiary as a for PP – to him/for him • The order is theme before goal/beneficiary The Dative construction
The direct object has more object properties: • It is closest to the verb: • I sent the letter to the manager • * I sent to the manager the letter • It moves to subject in the passive: • The letter was sent to the manager • * the manager was sent the letter to The Dative construction
As the two constructions mean similar things, it is often supposed that both are related • Perhaps one is a more basic form and the other is derived from it • But which way round? • DO Dative • Dative DO The relationship between do and dative constructions
Arguments for DO dative • There are some DO constructions which have no dative counterparts • so how could they have been formed from a dative? • That gave John a shock • They made Mary president • I spared the court the details • I envied John his good looks • We gave the car a new door • * that gave a shock to John • * they made president to/for Mary • * I spared the details to the court • * I envied his good looks to John • * we gave a new door to the car
In the case of give (give him a shock) this is a light verb construction • I had a look = I looked • I took a walk = I walked • I gave him a shock = I shocked him • So it isn’t really a DO construction Problems
In the case of make (make her president) this is related to the secondary predicate construction • They painted the barn red (the barn is red) • They made her president (she is president) • They gave him a pie (* he is a pie) • The two predicates seem to form a single complex predicate: • Paint-red ( to colour) • Make-president ( to elect) • So it isn’t really a DO construction Problems
In the case give the car a new door this involves inaliable possession: • John’s car => John has a car • John’s leg > John has a leg • Just like this distinction between a ‘possessor’ in a DP and the subject of the possessive verb, it seems that inaliable possession only works in the DO construction • Give a new heart to John • Is not ungrammatical • It just does not mean the heart is part of John • If anything, this indicates that the DO and dative construction have properties of their own • So perhaps neither is formed from the other Problems
If the dative were derived from the DO, the fact that there are two datives (to and for) would be hard to account for • If the DO is derived from the dative, it is easy: • The process involves the loss of the preposition • I V-ed something to someone • I V-ed someone something • I V-ed something for someone Arguments for dative DO
Arguments for dative DO • There are some datives with no DO counterpart: • I donated the money to charity • He said something to you • He reported the crime to the police • I sent the parcel to London • * I donated charity the money • * he said you something • * he reported the police the crime • * I sent London the parcel • The first two cases are difficult to explain as there are very similar verbs (give and tell) which do allow both dative and DO constructions • The last case shows that the two constructions can mean different things • The goal in the dative does not have to be the recipient in the dative • Hence one may not be derived from the other
With two arguments following the verb, three argument verbs have always been problematic for analysis The structure of the two constructions
One old analysis assumes both arguments are in complement position • But this means the structures have three branches and no other structure has this • Another places the first argument in complement and the second in an adjunct position • But adjuncts are recursive and arguments are not The structure of the two constructions
The two arguments together form a constituent • I gave [a rose to Mary] and [deadly nightshade to Bill] • This suggests the structure • But this does not conform to X-bar theory • XP has no head • And what is XP? Reasons why both of these are wrong
It seems that the first argument is higher than the second • The subject can be the antecedent of the object • But the object cannot be the antecedent of the subject • John1 likes himself1 • * himself1 likes John1 • Subjects are structurally higher than objects • Antecedents have to be structurally higher • John1 wants [himself1 to win] • * himself1 wants [John1 to win] Further observations
Consider: • The analyst revealed Bill1 to himself1 • * The analyst revealed himself1 to Bill1 • John showed Bill1 himself1 (in the mirror) • * John showed himself1 Bill1 (in the mirror) • So it seems that the first argument is higher than the second in both cases Further observations
This suggests the structure • This conforms to X-bar principles • The second argument is lower than the first • The two arguments are in the same constituent • We still don’t know what XP is • But it looks familiar • Similar to the transitive structure Further observations
Comparison to the transitive structure agent agent theme theme goal dative transitive This is compatible with the UTAH
Comparison to the transitive structure agent agent goal theme theme Double object transitive This is not compatible with the UTAH
The fact that the DO construction is not compatible with the UTAH, suggests that it is derived and that the dative is the basic structure DO is derived
Consider the Case relations in the dative • Nothing assigns Case to the agent • So it moves to subject position • The agentive verb assigns Case to the theme • The preposition assigns Case to the goal • Everything is as it should be How to derive the DO construction
If the DO construction has a structure similar to the dative, it should look like this • The theme is in specifier of the lexical verb • The goal is in the complement of the lexical verb • The agent gets no Case and so moves • The theme gets Case from the agentive verb • The goal does not get Case • Lexical verbs don’t assign Case • There is no preposition How to derive the DO Construction theme goal
The goal must move to a Case position • This is in front of the theme • We might assume the presence of another abstract verb • This would provide a specifier for the goal to move to • And provide an extra Case assigner • The agentive verb assigns Case to the moved goal • The extra verb assigns Case to the theme How to derive the DO construction
There are a number of problems which face this analysis: • Motivation for the extra abstract verb • Case assignment • The movement Problems with this analysis
We have assumed abstract verbs in other structures • E.g. Transitives and unergatives • But there was motivation for these in terms of their semantic contribution • Melt (transitive) = make + melt • Hit = do + hit • Smile = do + smile • There is little semantic contribution for the proposed abstract verb • I sent it to him = I sent him it The identity of the extra verb
The abstract agentive verb assigns agent thematic role and accusative Case (Burzio’s generalisation) The passive morpheme does not assign any thematic role and it does not assign Case The extra abstract verb does not appear to assign a thematic role It is strange therefore that it can assign Case Case assignment
The proposed analysis involves the movement of the goal argument from the complement of the lexical verb to the specifier of the abstract verb: • [VP DP1absV [VP DP lexV t1 ]] • We know movements have to be short • RelativisedMinimality = move to the nearest appropriate position • The proposed movement moves the goal DP over the theme DP • This seems to violate RelativisedMinimality Movement
If the goal cannot move over the theme, then it must originate in this position • Thus, either • The dative is derived from the double object, or • Both are basic and neither derives from the other Possible solutions
Both these solutions require a viable structure for the DO construction in which • The theme is in the specifier of the lexical verb (not in its complement position) • The goal is in a valid position consistent with X-bar theory The structure of the DO Construction
The only viable analysis is: • An agentive verb with agent argument • A goal verb with a goal argument • The lexical verb with a theme argument • This is similar to the agent-experiencer verbs • There are two abstract verbs The structure of the DO Construction
The goal verb assigns Case to the theme • The agentive verb assigns Case to the goal • The agent is not Case marked, so it moves • The verb moves to support both abstract verbs • So both the goal and the theme are objects • Following the verb The structure of the DO Construction
It might be possible to derive the dative construction from this structure But there is so much evidence against the dative being derived from the DO that it is unlikely So we are left with the alternative that both structures are unrelated So we might as well assume the simplest structure for the dative So ... What about the dative?
But this appears to violate the UTAH • There are two positions for goals • The specifier of an abstract goal verb • The complement of a preposition • But recall that the two structures don’t have the same meaning • The goal of the dative is not necessarily a recipient • I sent the message to London • The goal of the DO may be an inaliable possessor • They gave John a new heart So ... What about the dative?
Certain arguments of verbs can appear as objects of prepositions and maintain their meaning • John killed Bill • Bill was killed by John • He climbed the hill • He climbed up the hill • He loaded hay onto the cart • He loaded the cart with hay • Perhaps the UTAH can allow one argument position for each argument in the VP and another one in the PP So ... What about the dative?
Conclusion The Dative Construction The Double Object Construction