Gradience in Split Intransitivity

1
Gradience in Split Intransitivity
- The Turkish Pattern -
Mara van Schaik
2
The Threads

• Current views on split intransitivity ( SI)
• The gradience approach
• Telicity vs agentivity
• Turkish diagnostic tests
• Unaccusativity mismatches
• A possible solution
• Conclusions

3
Current views on SI

• The Unaccusative Hypothesis
• unaccusative vs unergative predicates
• The role of semantics
• projectionist approaches
• thematic/aspectual structure ? syntax
• constructional approaches
• syntax ? semantic interpretation
• Gradience in SI
• verbs have various degrees of unaccusativity
• or unergativity (within and across languages)

4
The Gradience Approach

• Arguments pro
• Soraces hierarchy (ASH/SIH)
• Change of location
  selects BE (least variation)
• Change of state
• Continuation of a pre-existing state
• Existence of state
• Uncontrolled process
• Controlled process (motional)
• Controlled process (non-motional) selects
  HAVE (least variation)

5
The Gradience Approach

• Typological predictions
• 1. Other tests in Romance Germanic obey the
  ASH
• 2. Tests in languages without AS obey the ASH
• 3. Core verbs pass more tests than non-core
  verbs
• Problems
• 1. Impersonal passive constructions in German
• 2. Quantifier floating in Japanese
• 3. Turkish diagnostic tests

6
Telicity vs agentivity

• Semantic features underlying the ASH
• telicity (predicate feature) ? aspectual
  structure
• determines the upper half of the ASH
• agentivity (argument feature) ? thematic
  structure
• determines the lower half of the ASH
• Problems
• the impersonal passive test
• some active-inactive languages

7
Telicity vs agentivity

• An alternative proposal
• Foleys Actor-Undergoer hierarchy
• Actor volitional performer
• causing an event or change of state
• sentience
• movement A U
• stationary
• causally affected
• incremental theme
• Undergoer undergoing a change in state

8
Turkish diagnostic tests

• The -mIs participle ( -mIs)
• The impersonal passive ( IP)
• The -(y)ArAk gerund ( -(y)ArAk)
• Other tests

9
Turkish diagnostic tests

• The -mIs test
• prenominal participle
• postterminative / stative-resultative value
• Examples
• a. çürü-müs yiyecek rotten food
  (unaccusative)
• rot-mIs food
• b. kos-mus çocuk run child
  (unergative)
• run-mIs child

10
Turkish diagnostic tests

• The IP test
• attaches the passive suffix
• good diagnostics past future (not aorist -
  generic)
• Examples
• a. Gösteri boyunca
  bagir-il-di. (unergative)
• demonstration throughout
  shout-PASS-PAST.3per
• It was shouted throughout the
  demonstration.
• b. Ay, dün burada çok fena
  kay-il-di. (unaccusative)
• oh yesterday here very badly
  skid-PASS-PAST
• Oh, yesterday it was skidded here very
  badly.

11
Turkish diagnostic tests

• The -(y)ArAk test
• denotes simultaneous or consecutive action
• the two verbs must be both unacc. or both unerg.
• Examples
• a. Kiz (top) oyna-y-arak sarki söyle-di.
  (unergative)
• girl ball play-Y-ArAk sing-PST
• The girl, while playing (ball),
  sang.
• b. Kiz (top) oyna-y-arak kay-di.
  (unaccusative)
• Girl ball play-Y-ArAk slip-PST
• The girl, while playing (ball),
  slipped.

12
Turkish diagnostic tests

• Other tests
• et- do vs ol- be compounds (unerg. vs unacc.)
• -Irken while vs -IncE when (the same)
• -GAn, -tI vs. -mA nominalizations (the same)
• -Ik adjectives (unaccusative)

13
Mismatches -mIs

• Change of location
• arrive, go, exit, enter occur with -mIs
  plus an adverbial (of path/manner/time)
• rise, descend, advance freely occur with
  -mIs
• State
• position Vs (sit, lie) and psychological
  states (be scared) freely occur
  with -mIs
• exist must take an adverbial

14
Mismatches -mIs

• Uncontrolled process
• some bodily function Vs (sweat, blush,
  sleep) freely occur with -mIs, some (shiver,
  cough, dream) take an adverbial, some
  (bleed) do not occur with -mIs
• Controlled process
• manner of motion Vs (run, swim, climb)
  occur with -mIs plus a path adverbial
• non-motional Vs (talk, work) occur with -mIs
  plus an adverbial

15
Mismatches IP

• Change of location
• occur in IPs with agentive implicit subject (e.g.
  come)
• Change of state
• psychological changes (get bored) occur in IPs
• Uncontrolled processes
• body function Vs (sneeze, shiver, hiccup,
  sweat) occur in IPs

16
Mismatches -(y)ArAk

• Uncontrolled processes
• cry may coocur with come (about a child)
• scream may coocur with be born (the same)
• sway may coocur with slip (about a drunkard)
• sparkle may coocur with go out (about light)
• foam may coocur with pull back (about the
  sea)
• sweat may coocur with die (about a person)

17
Mismatches

• Soraces ASH is challenged by -mIs IP tests
• definite change of location Vs cannot occur with
  -mIs without an adverbial (non-core behaviour)
• (agentive) change of location Vs occur in IPs
  (the same)
• manner of motion Vs can occur with -mIs if
  accompanied by a path adverbial
• non-motional controlled processes can occur with
  -mIs and an adverbial
• psychological change of state verbs occur in IPs

18
Mismatches

• Foleys hierarchy is challenged by the -mIs test
• both agentive and non-agentive motion Vs occur
  with -mIs
• psychological states and positions also occur
  with -mIs

19
A possible solution

• The IP depends on the degree of agentivity of the
  implicit subject (confirmed by Foleys hierarchy)
• (telic) change of location Vs may occur in IPs
• The -mIs test seems to depend on the degree of
  telicity agentivity of the construction
• (telic) change of state Vs with non-agentive
  arguments pass the test as unaccusatives (e.g.
  melt)
• telic change of location Vs pass the test as
  unaccusatives (e.g. fall, flee)

20
A possible solution

• Most Turkish motion Vs are not inherently telic.
• in the company of a path adverbial they become
  telic and pass the -mIs test
• Problem
• non-motional controlled processes occur with -mIs
  and an adverbial
• An alternative solution
• Turan -mIs does not test unaccusativity, but is
  rather associated with verbs with a consequent
  state

21
Conclusions

• Turkish -mIs, IP, -(y)ArAk test results
• change of state Vs are core unaccusatives in
  Turkish
• controlled process (non-motional) Vs are core
  unerg.
• intermediate classes exhibit variation
• agentive motion Vs are underspecified for
  telicity
• telicity plays no role in the IP test, but seems
  to be relevant in the -mIs and -(y)ArAk tests
• there are doubts as to the validity of the -mIs
  test

22
(No Transcript)