Dissociation cliticsdeterminers in a group of Italian SLI children

1
Dissociation clitics-determiners in a group of
Italian SLI children

• An elicited production experiment

2
Introduction

• Specific Language Impairment is a developmental
  pathology that seems to selectively affect
  language abilities.
• Roughly 5 of population below 6-7 years of age
  is affected.
• It is more common in males than in females.
• It shows a genetic bias.

3
Criteria for inclusion and exclusion

• Non verbal IQ over 85 (Verbal IQ NV IQ-20)
• At least 1.25 SD less than the age-matched
  controls in standard language tests (TROG, PFLI,
  TCGB, BNT, PEABODY).
• Absence of relational problems.
• Absence of neurological damages/malformations.
• Absence of hearing impairment/loss (reiterated
  Otitis Media).
• Absence of vocal tract malformations.

4
SLI general profile

• Common Features across languages
• Late talkers
• Comprehension gt Production
• Lexicon gt Syntax
• Omissions gt Substitutions

5
SLI affects

• Specifically language abilities (van der Lely,
  Wexler, Clahsen, Davies)
• Auditory/phonological abilities (Tallal, Bishop,
  Gathercole)
• General cognitive abilities (Kail, Locke)
• General processing abilities, expecially of brief
  elements (Leonard) the surface hypothesis

6
The Surface Hypothesis

• Leonard et al. (1989,1992,1998,2000) acustic
  features of linguistic elements make some
  grammatical elements particularly problematic in
  SLI.
• Equally non salient brief elements are
  predicted to be equally easy to omit.
• The Surface Hypothesis is assumed to hold in
  normal language development as well.

7
Predictions of the Surface Hypothesis

• The Surface Hypothesis can be easily tested in
  Italian (and in French).
• In these languages, definite determiners and
  clitics are predicted to be equally absent from
  SLI productions, because that they are brief in
  duration and phonetically identical
• La bambina, la vedo / Lo squalo, lo
  mangio
• The girl her I see / The shark him
  I eat
• Determiners and clitics can not stand on their
  own, nor can they be conjoined.
• Determiners precede the noun, clitics precede the
  finite verb.
• Definite determiners encode specificity, clitics
  do too but..
• Determiners are always present, clitics are not
  (pronouns, NPs).
• Clitics have the same interpretation of strong
  pronouns (arguments).

8
Determiners and clitics inprevious studies
Italian

• OMISSIONS
• Leonard et al.(1992) 15 SLI children (4-6 yrs
  old). (Elicited Production data)
• Bottari et al. (1998) 11 SLI children (M 63)
  (Spontaneous Production data)

9
Determiners and clitics in previous studies
French

• Hamann et al. (1996) 2 SLI children with
  opposite profiles.
• -Rafaelle non finite main verbs (RI)
• clitic omissions
• - determiner omissions
• -Loris few RI
• - clitic omissions
• determiner omissions
• Result determiners and clitics do not behave
  alike in pathology. Clitic omissions correlate
  with problems within the verbal domain

10
French, again

• Jakubowicz et al.(1998) 13 French SLI (M811).
• Determiners (90)gtSubject Clitics
  (73)gtReflexives (59)gtObject Clitics (26)
• Leonards answer in French the determiner, even
  if phonologically weak, is semantically salient
  because it carries gender information.
• If carrying (grammatical) gender information
  makes an element more salient, then the clitic in
  Italian should be more salient (and hence better
  preserved)

11
My experiment

• SLI 4 subjects, 4 males (M 97)
• 3 control groups (age, vocaboulary, syntax)
• 10 obligatory contexts per morpheme
• LA, LO determiners (NP is both SUB and OBJ
  position)
• LA, LO accusative clitics (all present tense
  contexts)
• SI reflexive clitic

12
SLI
13
Age and vocaboulary controls
14
Syntactic controls
15
Comparison

• WITHIN SLI
• - cliticsdeterminers is significant. Clitics
  are omitted more frequently.
• pronominalreflexive is not significant. Both
  omitted.
• SLI 9-6 CONTROLS
• presenceomission of clitics is significant. SLI
  omit clitics.
• presenceomission of determiners is not. SLI do
  not omit them.
• SLI 4 CONTROLS
• presenceomission of pronominal clitics and
  determiners is not significant. Both groups omit
  pronominal clitics. Both groups do not omit
  determiners.
• presenceomission of reflexive clitics is
  significant. SLI omit the reflexive, ND do not.

16
Intermediate summary

• Italian SLI in production there is a
  dissociation between clitics and determiners.
  Determiners are better preserved. (contra
  Leonard, ? Bottari)
• SLI behave like ND children for what concerns
  pronominal clitics but not for reflexive clitic.
• Italian SLI are different from French SLI
  (Jakubowicz reflexivegtpronominal)
• - Difference in design/material?
• - Difference in codification?

17
Codification

• OBJ Que fait Kiki à Nounours? I _ passe
  le mouchoir
• What does Kiki to Nounours ? He passes
  the handkerchief
• Jakubowicz omission of object clitic
• Present Experiment not coded
• It is a DATIVE clitic OMISSION (le/ le in French
  la/le lo/gli in Italian)
• REFL Que fait Kiki ? e _ lave un mai
• What does Kiki do? she
  washes a hand
• Jakubowicz Non Target Correct
• Present Experiment Omission of a reflexive.
• It is REFLEXIVE OMISSION, but DATIVE (si/si).
• Back to the experiment DATIVE-ACCUSATIVE

18
Examples

• OBJ Cosa fa il cane al gatto?
• What does the dog do to the cat?
• Target Lo gratta
  SLI gli toglie le pulci
• Him scratches
  to him takes away fleas
• REFLCosa fa Cenerentola?
• What does Cinderella do?
• Target Si pettina
  SLI si pettina i capelli
• herself brushes
  to herself brushes the hair

19
SLI
20
A closer look
21
To sum up

• Determiners and Clitics are DIFFERENT (Bottari et
  al, Hamann et al, Jakubowitz, present study)
• French SLI reflexive gt pronominal
• Italian SLI reflexive gt pronominal gtdative
• ND controls (French and Italian) reflexive gt
  pronominal

22
Some Speculations

• SLI have problems with operation Move a (as
  proposed by van der Lely 1998,2002 Davies,
  2004). This would account for the difference
  clitics-determiners
• If what is problematic is movement, then why
  dont we ever find the clitic after the verb?
  Vedolo (see him) is unattested
• Along the lines of Davies (2004) and van der Lely
  (2005) I take movement to be optional in SLI
  grammars.
• Along the lines of Sportiche (1992), I assume
  that the clitic is the head of CliticVoiceP. The
  complement of the verb is pro. pro and the clitic
  have to be in a Spec-Head relation (Clitic
  Criterion, EPP).
• Given that movement is optional, pro does not
  always get to Spec, CLP. When this happens,the
  whole projection is eliminated and we get
  omission.
• Optionality might be at work even in the grammars
  of ND children.

23

• Special thanks to
• Giuseppe Cossu
• Marcel den Dikken
• Alexia Ioannidou
• Miho Nagai
• Luigi Rizzi
• Ji Shim Young
• Erin Quirk
• The kids AL,DC,LM,TM