Multiple Sluicing

1
Multiple Sluicing

• A purely syntactic account

2
Sluicing

• Sluicing is a phenomenon of IP-ellipsis, where
  wh-phrase moves out of IP, and IP is deleted at
  PF (Ross, 1969 Lasnik, 1999 Merchant, 2001). It
  represents a construction where there is an
  interrogative clause with only a wh-element
  pronounced - as in (1)
• (1)
• John loves somebody but I dont know who John
  loves.
  (English)

3
Sluicing

• MATRIX CLAUSE
• (2) A John loves somebody. (English)
• B Who John loves?
• (3) A Dima ljubit kogo-to.
  (Russian)
• B Kogo Dima ljubit ?
• EMBEDDED CLAUSE
• (4) Dima ljubit kogo-to, no ja ne znaju kogo
  Dima ljubit
• Dima loves someone, but I dont know who(m)
• Dima loves somebody but I dont know who

4
Overview of Multiple Sluicing

• Ross 1969, Lasnik 1999, Merchant 2001 among
  other
• wh-movement is to Spec,CP followed by the PF
• deletion of the category below CP (IP-deletion).
• (5) John bought something. I wonder CP what IP
  John bought t (single sluice)
• Multiple Sluicing following Takashi (1984)
• IS sluicing with multiple wh-remnants

5
Russian Multiple Sluicing Point 1

• MATRIX CLAUSES
• (9) A Kadyj priglasil kogo-to na tanec
• everyone invited someone to dance
• Everyone invited someone to a dance
• B Kto kogo?
• who whom
• (10) A Kogo-to kadyj priglasil na tanec.
• B Kogo kto?
• whom who

6
Where do sluices come from?

• a. Licensing
• what positions do wh-phrases occupy in Russian
  Sluices, considering the properties of
  wh-movement in Russian? What functional
  projection(s) license(s) Sluicing?
• b. The analysis of multiple interrogatives in a
  given language constrains the range of possible
  antecedents in multiple Sluicing in that language.

7
Brief overview of sluicing analyses

• Stjepanovic (2003)
• since wh-phrases must be in Spec,CP in Sluicing
  constructions, C0 must be present in overt
  syntax. Strong wh feature of C0 triggers
  Superiority effects.
• Grebenyova (2004) uses Stepanovs (1998) account
  of multiple wh in Russian
• wh-fronting in Russian is not driven by a
  wh feature of C0.
• The argument is based on the Economy approach to
  Superiority, where the strong wh feature C0
  attracts the closest element to Spec,CP for
  feature checking, as in Chomsky (1995), Bokovic
  (1998, 2002) among others.

8
Grebenyovas Analysis

• The apparent Superiority effects under Sluicing
  follow from Parallelism required in ellipsis.
• Adopting the notion of Parallelism of Fiengo and
  May (1994) variables in the elided and
  antecedent clauses are bound from parallel
  positions.

9
Grebenyovas data

• a.Kadyj priglasil kogo-to na tanec, no ja
  ne znaju kto kogo.
• Everyone invited someone to a dance, but I
  dont know who whom
• b. Kadyj priglasil kogo-to na tanec, no ja
  ne znaju kogo kto.
• c. Kogo-to kazdyi priglasil na tanec, no ja ne
  znayu kogo kto.
• d. ??Kogo-to kazdyi priglasil na tanec, no ja ne
  znayu kto kogo.
• No apparent explanation why (c ) is bad.

10
REVIEW

• In sluices
• Russian exhibits superiority effects in the
  subordinate clauses, but not in the matrix
  clauses.
• The existing account of semantic parallelism
  doesnt seem to explain the following
• Why matrix sluices with no superiority are ok?
• Why embedded sluices even with parallelism are
  not ok?
• Let us consider an alternative account

11
Multiple Sluicing in Russian

• EMBEDDED CLAUSE
• (6) Kadyj priglasil kogo-to na tanec, no
  ja ne pomnju
• everyone invited someone to dance but I not
  remember
• 
  kto kogo
• 
  who whom
• Everyone invited someone to a dance but I dont
  remember who (invited) whom
• (7) Kogo-to kadyj priglasil na tanec no ja ne
  pomnju kogo kto (whom who)
• (8) Kogo-to kadyj priglasil na tanec no ja ne
  pomnju kto kogo (whom who)

12
Russian multiple wh

• with multiple whs Russian exhibits superiority
  in embedded clauses but not in matrix clauses
• (11) Kto cto posovetoval Darii?
• whoNOM whatACC advised DariaDAT
• Who advised what to Daria?
• b. Cto kto posovetoval
  Darii?
• whatACC whoNOM advised DariaDAT
• "Who advised what to Daria?"

13

• (12) Maria ne znaet kto ctoi posovetoval
  ti Darii? Maria not knows whoNOM whatACCadvised
  DariaDAT
• "Maria doesn know who advised what to
  Daria?
• (13) ?Maria ne znaet, ctoi kto posovetoval ti
  Darii? Maria not knows whatACCwhoNOM
  advisedDariaDAT
• "Maria doesnt know who advised what to
  Daria?"

14
Towards the analysis through wh

• There is strong wh feature in Russian. In
  multiple wh
• fronting all the wh undergo CP checking. The
  Superiority
• can be violated because there is an escape path
  into the
• TopP position that is located higher than CP.
• This TopP position is ONLY available in matrix
  clauses. This
• creates the matrix/ embedded clause asymmetry for
  the
• availability of this position.
• I suggest the following structure for multiple wh
  constructions in Russian.
• TopP SpecCP wh whIP

15
Kto cto posovetoval Darii? Cto
kto posovetoval Darii?
There is other evidence for this structure I can
expand on it later
16

• If the above analysis is satisfactory for whs, we
  can assume the same analysis would be sufficient
  for sluicing constructions.
• Thus, we dont need Grebenyovas semantic
  parallelism account.
• Shes used it because she used Stepanovs wh-
  analysis for Russian no superiority analysis,
  no wh movement analysis.
• In addition, this analysis explains the asymmetry
  between the matrix and the embedded clauses in
  multiple wh-constructions, and in multiple
  sluices.

17
What do we get?

• (14) Kadyi xudonik vybral kakuju-to krasku no
• every painter chose some
  paint but
• ja ne zametil kto kakuju
• I not noticed who which
• Each painter chose some color paint but I
  havent noticed who chose which
• (15) Kakuju-to krasku vybral kadyi xudonik no
  ja ne zametil kakuju kto (which who) most
  informants

18
This somehow leads us to another tricky puzzleI
DO NOT KNOW THE ANSWER TO YET

• (16) a. Kto-to cto-to citaet, no ja
  ne uveren kto cto.
• Somebody something reads but I not
  sure who what
• Somebody reads something, but I am not sure who
  reads what
• b. Kto-to cto-to citaet, no ja
  ne uveren kto i cto.
• Somebody something reads but I not sure
  who and what
• Somebody reads something, but I
  am not sure who reads and what

19

• BUT
• (18) a.Kto-to kogo-to udaril, no ja ne
  znaju kto kogo.
• smbd. smbdACC. hit but I not
  know who whom
• Somebody hit someone but I dont know who
  whom
• b. Kogo-to kto-to udaril.
• smbd.ACC smbd. hit.
• c. Kogo-to kto-to udaril, no ja ne znaju
  kogo kto.
• smbd.ACC smbd. hit but I not
  know who whom

20

• These facts are similar to Polish where at least
  one wh was claimed to occur in CP (Citko, 2001?)
• (19) a. Kto kogo koca? (Polish)
• who whom loves
• b.Kogo kto koca?
• whom who loves

21

• (20) Kto kogo koca, ale ne vjem kto
  kogo.
• smbd. smbd.ACC loves but not know-I who
  whom.
• b. Kogo kto koca, ale ne vjem kogo
  kto.
• smbd.ACC smbd. loves but not know-I whom
  who
• (21)Kto co citall ale ne vjem kto co.
• Somebody something read, but I dont know
  who what
• Kto co cital, ale ne vjem kto i co.
• Somebody something read but I dont know who and
  what

22
Alternatives

• Kazenin (2002) analyzes similar constructions in
  pure multiple whs but in no reference to sluices
• Kazenin refers to it as coordinated sluicing
• This is the issue of further consideration.

23

• ANY QUESTIONS?
• COMMENTS?
• SUGGESTIONS?
• THANK YOU!!! ?