Causatives have been analysed as a single clause with a split-VP.

If the binding of reflexives is only possible intra-clausally, why is there this contrast ingrammaticality?

[Mary]i let [John]j kill [herself]i*
[Mary]i let [John]j kill [himself]j

It seems that precedence does play a part after all.

  • No, they are not a single clause with a split VP. "Let" is a catenative verb and the subordinate clause "kill herself" is its catenative complement. The intervening NP "John" is the syntactic object of "let", but only the understood subject of "kill". The fact that the pronoun is thus in a different clause to the intended antecedent blocks the reflexive, whereas in [2] the pronoun and the antecedent are in the same clause and the reflexive becomes mandatory.
    – BillJ
    Nov 19 '17 at 19:35
  • @BillJ If John is the syntactic object of ‘let’, then that should mean that it is in the matrix clause. Binding of the reflexive ‘himself’ to John then shouldn’t be possible because they are in different clauses, unless there’s a PRO subject. Nov 19 '17 at 20:05
  • On the surface there is only one clause and John is raised to object, but the true 'object' of the verb "let" is a second 'event' with subject John and verb "[did] kill".
    – amI
    Nov 20 '17 at 23:17
  • Yes, but "John" is the semantic subject of "kill". If the intended meaning in [1] is that John killed Mary, then you can't say that "John killed herself", since "the antecedent is in another clause, i.e. the matrix. Thus it has to be the non-reflexive "her".
    – BillJ
    Nov 23 '17 at 18:58
  • @BillJ If a causative construction is so easily analysed as being composed of 2 clauses, as supported by data on binding, why are there generative syntacticians (e.g. Radford, 2011) who say that it’s a single clause? Nov 24 '17 at 0:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.