Word Order

1
Word Order

• Linguistic Research on Second Language
  Acquisition Development Lecture 4

2
Todays lecture

• We will.
• critically review of studies on L2 acquisition of
  German verb placement
• examine data on L2 acquisition of direct object
  scrambling in Dutch (Unsworth 2003, to appear)

3
What is word order?

• Word order is the result of the interplay
  between the projection of structure from
  particular categories, as determined by X-theory
  and the feature specifications of particular
  functional heads like I .
• (Hawkins 2001 124)

4
Variation within languages

• Base order vs. surface order
• Cécile NegP (ne) pas VP boit du café
• Cécile I (ne) boiti NegP pas VP ti du café
• Cécile I (ne) V lej boiti NegP pas VP ti
  tj

5
Variation across languages

• Same surface orders .
• (4) Charlie Brown hates Lucy
• (5) Charlie Brown haat Lucy ti

6
Variation across languages

• Same surface orders .
• (4) Charlie Brown hates Lucy
• S V O
• (5) Charlie Brown haat Lucy ti
• S V O

7
Variation across languages

• but different base orders
• (6) I think that Lucy has kissed Charlie
  Brown
• (7) Ik denk dat Lucy Charlie Brown gekust
  heeft

8
Variation across languages

• but different base orders
• (6) I think that Lucy has kissed Charlie
  Brown
• S V O
• (7) Ik denk dat Lucy Charlie Brown gekust
  heeft
• S O
  V

9
Verb placement in German

• Descriptive facts
• declarative sentences SVO
• (8) Johann kaufte heute ein Buch
• non-finite verb sentence-finally
• (9) Johann hat heute ein Buch gekauft
• (10) Johann wird heute ein Buch gekauft
• (11) Johann nahm heute ein Buch auf

10
Verb placement in German

• Descriptive facts
• V2 i.e. finite verb in second constituent
  position
• (12) Heute kaufte Johann ein Buch
• simple and compound verb forms clause-finally in
  embedded clauses
• (13) Sie weisst dass Johann ein Buch gekauft hat
• (14) Sie weisst dass Johann ein Buch kaufen wird
• (15) Sie weisst dass Johann ein Buch aufnahm

11
Clahsen Muysken (1986)

• L2 acquisition of verb placement in German
• L1 Spanish, Portuguese, Italian
• SVO
• no V2
• L1 Turkish
• SOV
• no V2

12
CM (1986) Subjects

• 93 adults (Italian, Spanish, Portugese)
• 10 adolescents, aged 14-16 years (Turkish)
• cross-sectional and longitudinal data

13
CM (1986) Stage I

• S(Aux)VO
• (16) Ich studieren in Porto
• Adverb-fronting, i.e. no V2 
• (17) Vielleicht andere Kollege sagen

14
CM (1986) Stage II

• Verb separation, i.e. S Vfin O V-fin, but still
  no V2
• (18) Wir haben drei Feuer gesehen

15
CM (1986) Stage III

• V2, including in embedded clauses
• (18) Veilleicht wissen viele Leute nicht
• (19) Wann wir fahren hier in Deutschland, drei
  Feuer gesehen

16
CM (1986) Stage IV

• Verb-final embedded clauses
• (20) Wann wir hier in Deutschland fahren

17
CM (1986) Summary

• systematic stages across L2ers with different L1s
• similar stages found in other studies with
  different L1s

18
CM (1986) Analysis

• Why do L2ers go through such stages?
• L2ers use (phrase-structure) rules which are not
  constrained by UG
• cf. L1 acquisition of German verb placement

19
CM (1986) Problem

• CM specifically constructed their analysis for
  L2ers
• So its not so surprising that its not
  UG-constrained!
• ? Solution we need independently motivated
  principles of UG

20
DuPlessis et al (1989) Theory

• assume Travis (1984) analysis
• left-headed IP and right-headed VP
• inflection strong in German
• (21) IP Johann I kauftei VP heute ein Buch
  ti
• verb is adjoined to I (eg (14) in Hawkins)

21
DuPlessis et al (1989) Theory

• V2 substitution of V I into C and adverb or
  object into Spec, CP
• (22) CP Heutek C hati IP Johann ti VP VP tk
  ein Buch gekauft ti

22
DuPlessis et al (1989) Theory

• verb separation non-finite verb stays in V
  whilst finite verb moves to I
• (23) CP Johannj C hati IP ti tj VP VP heute
  ein Buch gekauft ti

23
DuPlessis et al (1989) Theory

• embedded clauses finite verb cannot move to I
  because C is filled by complementiser
• complementiser governs I, making its inflection
  weak, so verb doesnt have to move to I (see eg
  (19) in Hawkins)

24
CM (1986) Stage I

• S(Aux)VO
• (16) Ich studieren in Porto
• Adverb-fronting, i.e. no V2 
• (17) Vielleicht andere Kollege sagen

25
DuPlessis et al (1989) Analysis

• Stage I
• VP head-initial ? S(I)VO order
• Adverb fronting takes place by means of
  IP-adjunction (which is ok in L1, but not in L2)

26
CM (1986) Stage II

• Verb separation, i.e. S Vfin O V-fin, but still
  no V2
• (18) Wir haben drei Feuer gesehen

27
DuPlessis et al (1989) Analysis

• Stage II
• VP head-final ? SIOV
• V moves to I
• plus IP-adjunction

28
CM (1986) Stage III

• V2, including in embedded clauses
• (18) Veilleicht wissen viele Leute nicht
• (19) Wann wir fahren hier in Deutschland, drei
  Feuer gesehen

29
DuPlessis et al (1989) Analysis

• Stage III
• CP projection is activated
• V moves to C (via I)
• XP moves to Spec, CP
• filled C doesnt govern I ? strong inflection in
  I ? V2 in embedded clauses

30
CM (1986) Stage IV

• Verb-final embedded clauses
• (20) Wann wir hier in Deutschland fahren

31
DuPlessis et al (1989) Analysis

• Stage IV
• L2ers learn that C filled with complementiser
  governs I and thus, cancels strong inflection in
  I

32
Schwartz Tomaselli (1990)

• Problem with theoretical account used by
  DuPlessis et al(1987)
• Why should embedded C in German be able to
  license I, neutralising its strong inflections?
• Solution right-headed IP (SOVI) gets rid of this
  stipulation

33
Verb placement in German

• Right-headed IP and right-headed VP ..
• (den Besten 1983)
• (22) CP Johannj C kauftei IP tj VP heute ein
  Buch ti ti
• (23 CP Johannj C hati IP tj VP VP heute ein
  Buch gekauft ti ti
• (24) CP Heutek C hati IP Johann VP tk ein
  Buch gekauft ti ti

34
Schwartz Tomaselli (1990)

• Stages I-III
• more or less same as DuPlessis et al
• Stage IV
• I changes from left-headed to right-headed

35
Schwartz Tomaselli (1990)

• Advantages of this proposal
• right-headedness of IP is allowed under UG, just
  like right-headedness of VP
• no special mechanism necessary, i.e. filled C no
  longer needs to license I

36
Vainikka Young-Scholten (1994, 1996a, 1996b)

• Series of studies on L2 acquisition of German
  verb placement
• Focus initial state
• Contra claim by CM that L2ers with OV L1 start
  with VO

37
Vainikka Young-Scholten (1994, 1996a, 1996b)

• L1 Turkish, Korean
• (both SOV, no V2)
• L1 Italian, Spanish
• (both SVO, no V2)

38
Vainikka Young-Scholten (1994, 1996a, 1996b)

• Turkish
• (25) Helmut VP simdi Istanbul-da Türkçe
  ögren-iyor
• Helmut now Istanbul-part Turkish
  learn-part
• Helmut is now learning Turkish in Istanbul
• Korean
• (26) Helmut-ka VP Peter-eke chaek-ul
  chu-oss-ta
• Helmut-part Peter-part book-part
  give-part-part
• Helmut gave the book to Peter

39
Vainikka Young-Scholten Subjects

• L1 Turkish (n11)
• L1 Korean (n6)
• aged 28-60 years
• little or no formal instruction
• arrived in Germany after age of 20
• resident in Germany for between 18 months and 24
  years

40
Vainikka Young-Scholten Method

• spontaneous data
• some semi-elicited data
• longitudinal and cross-sectional
• acquisition criteria 60 suppliance in
  obligatory context
• use implicational scaling to determine
  development stages

41
Implicational scaling

• If a learner is accurate on a difficult
  morpheme, she will also perform accurately on
  some easier morpheme. accuracy on a
  difficult morpheme y implies accuracy on easier
  morpheme x. But the reverse is not the case.
  Another learner who is accurate on morpheme x may
  not be accurate on the more difficult morpheme
  y. (Hawkins 2001 47)

42
Vainikka Young-Scholten Stage I

• Utterances with verb and object
• 98 Korean/Turkish L2er OV
• cf. 81 Italian/Spanish L2ers VO

43
Vainikka Young-Scholten Stage I

• (27) Eine Katze Fisch alle essen
• a cat fish all eat
• A cat ate the entire fish
•   Changsu, 033, Korean
• Ich sprechen die meine Firma
• I speak the my firm
• I speak (to/at) my firm
• Salvatore, 3, Italian

44
Vainikka Young-Scholten Stage I

• Little overt evidence for IP projection no
  auxiliary verbs tense, person and number
  inflection generally absent finite verbs not
  raised over adverbs and negation
• ? VP projection only, transferred from VP

45
Vainikka Young-Scholten Stage II

• Utterances with verb and object
• 64 of utterances of slightly more advanced
  Italian/Spanish L2ers OV

46
Vainikka Young-Scholten Stage II

• (29) Vielleicht Schule essen
• maybe school eat
• Maybe (he/she) eats at school.
•   Salvatore, 6, Italian
• Italian/Spanish L2ers have switched VP from VO to
  OV

47
Vainikka Young-Scholten Stage III

• Auxiliaries and modal verbs used in position to
  left of VP
• When auxiliaries arent used, thematic verbs are
  raised to left of object, adverb or Neg
• this includes Turkish/Korean speakers, whose L1
  has right-headed IP

48
Vainikka Young-Scholten Stage III

• (30) Ich sehen Schleier
• I see veil
• I see (the) veil
•   Kemal, 604, Turkish
• (31) Mir machen nichts mehr
• me make nothing more
• (They will) do nothing more (to) me
•   Mehmet, 103,
  Turkish

49
Vainikka Young-Scholten Stage III

• Little evidence of tense, person/number agreement
  
• ? I position is projected but it remains
  underspecified F(unctional)P
• ? No L1 influence at IP level (because no
  right-headed IP in Turkish/Korean L2ers)

50
Vainikka Young-Scholten Stage IV

• subject-verb agreement
• obligatory subject
• verb raising

51
Vainikka Young-Scholten Stage IV

• (32) Ich kaufe dich Eis
• I buy you ice-cream
• I (will) buy you (some) ice-cream
• Gabho, 254, Korean
• FP is specified to AgrP

52
Vainikka Young-Scholten (1994, 1996a, 1996b)

• structure-building
• a.k.a. Minimal Trees
• cf. Full Transfer / Full Access approach
  (Schwartz Sprouse 1994, 1996)
• cf. Valueless Features approach (Eubank 1993/4,
  1996)

53
And now..

• Hawkins (2001 166-7) Exercise 3