Universal 20 and Intralinguistic Variation

1
Universal 20 and Intra-linguistic Variation

• Niina Zhang
• National Chung Cheng University
• 2007 International Conference on Linguistics in
  Korea
  
• Seoul, Jan. 19-20, 2007

2
1. Introduction

• Greenbergs (1963 87) Universal 20
• In a nominal containing the three types of
  elements
  
• Demonstratives, Numerals, Adjectives,
• If any or all of them precedes the N,
• they are always found in the order in (1a)
• If they follow the N,
• the order is either the same, as in (1b),
• or its exact opposite, as in (1c).
• (1)
• a. Dem Num Adj N
• b. N Dem Num Adj
• c. N Adj Num Dem

3
(2)
4
Three Aspects of Universal 20

• (4)
• Every element of the set Adj, Num, Dem can
  occur either to the left or right of N
  
• If any two elements of the set occur on the same
  side of N, they should be ordered according to
  the closeness relation to N illustrated in (2).
  
• c. N may undergo a leftward movement, but not
  rightward movement.

5
The Issue

• How are the constraints flexibilities of
  Universal 20 exhibited in one language?

6
Word orders of sign languages are flexible.
In verbal domains nominal domains

Bouchard (1997) and Bouchard Dubisson (1995)
sign languages do not have any hierarchical
structure.

• The empirical issue of this talk
• the word order of nominal domains of TSL.

vs.
Neidle et al. (2000 60-61)
7

• Lai (2005) presents a comprehensive description
  of the possible and impossible nominal-internal
  word orders of TSL, i.e., the orders of
  demonstratives, adjectives, numerals, and head
  nouns.
• --------------------------------------------------
  ---------------------------------------
  
• The data in Lai (2005) are all from a corpus
  established in the project A Study of Taiwan
  Sign Language Phonology, Morphology, Syntax and
  Digital Graphic Dictionary (NSC
  90-2411-H-194-025 (I), NSC 91-2411-H-194-030
  (II), NSC 92-2411-H-194-007 (III), and NSC
  93-2411-H-194-001 (IV)), headed by James H-Y Tai,
  and sponsored by Taiwan National Science Council
  (Aug. 2001-July 2005).

8

• But it remains a puzzle why certain orders are
  possible and others are not.
  
• The goal of this talk is to show that TSL
  exhibits Universal 20.
  
• This fact indicates that like any oral language,
  sign languages have hierarchical structures.

9
How is a hierarchical structure linearized?

• If elements that are organized hierarchically can
  be linearized symmetrically, linear order is not
  part of syntax.
  
• If the hierarchical structure correlates with any
  linear order, linear order is part of syntax.

10

• Since Reinhart (1979), it has been considered
  that while a hierarchy is crucial to Narrow
  Syntax, linear ordering in language is not.
  
• In this consideration, linear ordering
  sequentializes what is primarily only ordered
  hierarchically.
  
• (Marantz 1984 7-8, Abels Neeleman 2006)
• Linear ordering might be a PF phenomenon forced
  by the requirements of the articulators, or
  parsing.
  
• (see Chomsky 1995, 2005 5, 2006 7)

11

• However, works represented by
• Kayne (1994, 2004),
• Cinque (2005)
• argue for the narrow syntax status of linear
  ordering.

12

• Lillo-Martin (2001 304)
• future researchers may ask whether SLs may
  offer any new insights into the recent idea that
  while hierarchical structure is part of Narrow
  Syntax, ordering is not.
• SLs Sign Languages

13
Flexibility Rigidity

• Left and Right Positions
• Hierarchical Structures

14

• I will account for the rigidity in the TSL
  nominal-internal word orders in terms of purely
  syntactic hierarchy,
  
• and
• the flexibility there in terms of the freedom
  of constituent ordering.

15
If hierarchy rather than word order is part of
narrow syntax, (3) can be true in some
language.(3)

a. Every element of the set Adj, Num, Dem can
occur either to the left or right of N

• If any two elements of the set occur on the same
  
  
• side of N, they should be ordered according
  to
  
• the closeness relation to N.

16
2. Identifying possible orders of nominals in TSL

• This section demonstrates the flexibility in TSL.
  
  
• This is the first aspect of Universal 20.

17
2.1 Two layers

• (4) ?
• N
• ? A, Num, Dem

18

• (5) a. CUTE CAT IXpro1s LIKE. A N
  (Lai 2005 15)
  
• b.        CAT CUTE IXpro1s LIKE. N A
• Both I like cute cats.
• (6) a. IXdet AIRPLANE REACH AMERICA. (Lai
  2005 67) Det N
  
• b.       AIRPLANE IXdet REACH
  AMERICA. N Dem Both This plane is flying
  to America.
  

19
(7) a. TEACHER TABLE THREE BOOK PUT. (La
i 2005 44) Num N b.   TEACHER TABLE BOOK THREE P
UT. N Num Both The teacher put thre
e books on the table.
20
The two possible orders in other SLsASL

• Neidle et al. (2000 103) and Sandler
  Lillo-Martin (2006 308, 341) mention both A N
  and N A orders in ASL.
  
• However, MacLaughlin (1997) claims that left APs
  and right APs are hierarchically different (see
  Sandler Lillo-Martins 2006 341 for a review).

21
The two possible orders in other SLsHKSL

• Tang and Sze (2000) mention that both Num N and N
  Num orders are found in HKSL.

22
The two possible orders in other SLsASL

• Bahan, Kegel, MacLaughin, Neidle (1995),
• Neidle et al. (2000 89),
• Sandler Lillo-Martin (2006 339)
• They all claim that post-N determiner-like
  elements are not determiners in ASL.

23
2.2 Three layers

• (9) ?
• ?
• N
• ? A, Num
• ? Num, Dem

24

• 4 possible orders if ? A and ? Num (Lai
  200573)
  
• (10) a. IXpro3s FIVE CUTE CATS HAVE. Num
  A N
  
• She has five cute cats.
• b.  IXpro3s TELL-ME HAVE ONE MAN GOOD
  INTRODUCE TO-ME. Num N A
  
• She said shed like to introduce a
  good guy to me.
  
• c. IXpro3s RAISE CUTE CATS FIVE.
  A N Num
  
• She raises five cute cats.
• d. IXpro3s HIGH-HEELS BLACK TWO HAVE.
  N A Num
  
• I have two pairs of black high
  heels.
  

25

• 4 possible orders if ? A and ? Dem (Lai 2005
  81)
  
• (12)
• a. IXdet CUTE CAT IXpro1s BELONG-TO.
  Det A N
  
• b. IXdet CAT CUTE IXpro1s BELONG-TO. Det
  N A
  
• c. CUTE CAT IXdet IXpro1s BELONG-TO. A
  N Dem
  
• All That cute cat belongs to me.
• d. CAT CUTE IXdet BELONG-TO. N A
  Dem
  
• That cute cat belongs to me. (Hsin-Hsien Lee,
  p.c.)

26

• 4 possible orders if ? Num and ? Dem (Lai
  2005 84)
  
• (13)
• a. IXdet FOUR CAR IXpro1s FRIEND
  BELONG-TO. Det Num N
  
• b. IXdet CAR FOUR IXpro1s FRIEND
  BELONG-TO. Det N Num
  
• c. FOUR CAR IXdet IXpro1s FRIEND BELONG-TO.
  Num N Dem
  
• d. CAR FOUR IXdet IXpro1s FRIEND BELONG-TO.
  N Num Dem
  
• All Those four cars belong to my friend.

27
2.3 Four layers

• (14) Dem
• Num
• A

N
28

• Mathematically, four elements in a shell
  structure allow 8 orders (23).
  
• However, since Dem cannot occur at the right
  edge, we have found only four orders (23/222).
  
• At this moment, I cant account for this
  constraint.
  
• Note that the constraint is found in ASL on all
  nominals (see slide 22).

29
(15) a. IXdet FIVE NAUGHTY BOY IXpro1s BELONG-TO
STUDENT. Det Num A N b. IXdet NAUGHTY BOY
FIVE IXpro1s BELONG-TO STUDENT.
Det A N Num c. IXdet FIVE BOY NAUGHTY IXpr
o1s BELONG-TO STUDENT. Det Num N A d. IXd
et BOY NAUGHTY FIVE IXpro1s BELONG-TO STUDENT.
Det N A Num All These five naughty boy
s are my students. (Lai 200586)
30
2.4 Section conclusion

• All possible orders stated in Universal 20 are
  attested in TSL.
  
• The first aspect of Universal 20
• Every element of the set Adj, Num, Dem can
  occur either to the left or right of N
  
• The significance of this conclusion will be
  discussed in section 5.

31
3. Accounting for the impossible orders by the
hierarchical structures

• This section demonstrates the second aspect of
  Universal 20 in TSL
  
• If any two elements of the set A, Num, Dem
  occur on the same side of N, they should be
  ordered according to the closeness relation to N.

32
(2)
33
3.1 Three Element Nominals

• (9) ?
• ?
• N
• ? A, Num
• ? Num, Dem

34
The order of ??N is not allowed

• If we consider the N-final orders

(17) ??N a. A Num N b. A
Dem N
c. Num Dem N
(16) ??N a. Num A N (10a) b. Dem
A N (12a)
c. Dem Num N (13a)
35
3.2 Four Element Nominals

• (14) Dem
• Num
• A

N
We have seen four possible orders in (15), as
predicted by (14).
36

• Mathematically, if four elements do not have a
  shell structure, they should have 24 orders
  (4x3x2x1).
  
• However, according to Lai (2005), the orders in
  (15) are the only possible orders for four
  element nominals.
  
• Orders like the following are not acceptable

37

• (18) a. Det A Num N
• b. Det N Num A
• c. A Dem Num N
• d. A Num Dem N
• e. N Num A Dem
• f. N Dem A Num
• g. N A Dem Num
• In all of these unacceptable orders, the
  hierarchy requirement is violated.

38
3.3 Section conclusion

• If the order of nominal-internal elements in TSL
  is absolutely free, the restrictions are
  unexpected.
  
• The second aspect of Universal 20 is attested in
  TSL
  
• If any two elements of the set A, Num, Dem
  occur on the same side of N, they should be
  ordered according to the closeness relation to N

39
4. Accounting for other possible orders by
N-Raising

• All N-initial orders are fine for three element
  nominals.
  
• This is the third aspect of Universal 20.

N may undergo a leftward movement, but not
rightward movement.
40
Both N?? and N?? are possible

• N??
• a. N Num A
• b. N Dem A
• c. N Num Dem

• N??
• a. N A Num (10d)
• N A Dem (12d)
• c. N Num Dem (13d)

41

• (20)
• IXpro1s CAT FIVE FAT. N Num A
• I have five fat cats. (Hsin-Hsien Lee p.c.)
• b. IXpro3s ASK TEACHER QUESTION SOME DIFFICULT.
  N Quan A
  
• He asked the teacher some difficult
  questions.
  
• (Lai 200575 (24e))
• c. CAT IXdet CUTE IXpro1s BELONG-TO. N
  Dem A
  
• That cute cat belongs to me. (Lai 2005 80
  (34c))
  
• d. CAR IXdet FOUR IXpro1s FRIEND BELONG-TO.
  N Dem Num
  
• Those four cars belong to my friend. (Lai
  2005 84 (37e),(38b))

42
N-Raising in the Left-Merger Constructions

• (21) N ? ?
• ? A, Num ? Num, Dem

43
Rightward N-Raising is impossible

• (21) ? ? N
• X
• ? A, Num ? Num, Dem
• See the restrictions in 3.1.

44
5. Discussion

• Universal 20 and hierarchical structures of sign
  languages
  
• Representing the flexibility of Universal 20
• Intra-linguistic variation

45
5.1 Universal 20 and hierarchical structures of
sign languages

• TSL strictly follows the restrictions expressed
  by Universal 20.
  
• The restrictions are formalized in Cinque (2005)
  and Abels Neeleman (2006).
  
• Bouchard Doubissons (1995) theory that sign
  languages do not have any hierarchical structure
  cannot be maintained.

46
5.2 Representing the flexibility of Universal 20

• We have seen that TSL also exhibits the word
  order flexibility expressed by Universal 20
  
• Every element of the set Adj, Num, Dem can
  occur either to the left or right of N.
  
• This flexibility can be achieved by either of the
  freedom of merger directions (allowing
  right-merger), or movement.

47
Deriving the order of Dem N Adj Num (15d)
48

• One way to reduce descriptive complexity of the
  derivations of nominal internal orders is to
  assume that the elements can be merged at either
  right or left positions.
• This freedom suggests that syntax does not
  dictate the linear order of constituents.
  
• Actually, right remerger of wh-elements in ASL
  has been argued for in Neidle (2002), where
  SpecCP is claimed to be projected rightward in
  the language.
• The assumed freedom in merger direction in all
  levels of a nominal have the following two
  significances ?

49
A. Merger direction is not configuration-specific

• We have seen that Dem, Num, and Adj all can occur
  either to the left or right of N.
  
• Takano (2003 524)
• adjuncts vs. non-adjuncts
• ?
• subject to antisymmetry and thus
• their positions are fixed to the relevant
  selecting element (left specifiers and right
  complements).
  
• Since it is generally recognized that Dem and Num
  are not adjuncts whereas Adjectives are, the free
  order of both types is not compatible with
  Takanos claim.

50
B. Merger direction is not category-specific

• Based on a study of word order of clauses in ASL,
  Romano (1991)
  
• lexical categories head-initial
• functional categories head-final
• Since the scope of our study includes both
  lexical and functional cats, our conclusion of
  the TSL nominals is not compatible with Romanos
  claim on ASL clauses.
• Syntactic variation is not restricted to
  functional items (contra Chomsky 1995, Kayne
  2005).

51
5.3 Intra-linguistic variation
for all nominal-internal categories
in the same language
Both head-initial and head-final orders

• The co-existence of two settings of a parameter
  has been found in other cases.

52
Case A Null argument licensing

• agreement discourse reference
• Romance lgs -
• East Asian lgs -
• English - -
• Parameter setting
• However,
• Warlpiri (Legate 2003) (either)
• ASL (either)
• (see Sandler Lillo-Martin 2006 16)

53
Case B Moved WH-phrases and WH-in-situ

• In languages like English, wh-movement is
  obligatory.
  
• In Chinese, wh phrases remain in situ.
• Parameter setting
•  
• ASL has moved WH-phrases as well as WH-in-situ
  (Lillo-Martin 1990, Sandler Lillo-Martin 2006
  16, ch. 23)

54
Case C Two Types of Noun Incorporation

• In compounding NI, the valence of the clause is
  decreased
  
• In classifying NI, the valence of the clause is
  not decreased
  
• Many languages with NI consistently use only one
  type Parameter setting
  
• However, some languages have both types (Hopkins
  1988, Mithun 1984).

55

• As stated by Sandler Lillo-Martin 200616),
• An explanation for this tendency to have it
  both ways is still to be determined.

56
Kayne (2005 3)

• The parametric variation that occurs within
  languages is of exactly the same sort as the
  parametric variation that occurs across
  languages.
• The elements subject to it are the same in both
  kinds of cases, and the features/properties in
  question are, too.

57
Parameters or Feature Values?

• The intra-linguistic variation studied here
  casts doubts on the assumption that languages
  themselves are the immediate locus of parametric
  variation (see also Newmeyer 2004, Kayne 2005
  1).
• What the parameter approach tries to explain can
  be simply covered by values of features.
  
• Different values of any feature can be attested
  either cross-linguistically or intra-linguisticall
  y.

58
6. Summary

• I have shown that both the flexibility and the
  restrictions generalized in Greenbergs (1963
  87) Universal 20 are basically attested in Taiwan
  Sign Language.
• Theoretically, this study concludes that like
  oral languages, sign languages have hierarchical
  structures.
  
• Moreover, if elements that are organized
  hierarchically can be linearized symmetrically,
  to reduce descriptive complexity, Kayne-Cinques
  correlation between syntactic hierarchical
  structures and linear ordering might need
  reconsideration.
• Furthermore, the intra-linguistic variation of
  the universal calls for reconsideration of our
  current theory of parameter setting.

59

• Abels, Klaus. Ad Neeleman. 2006. Universal 20
  without the LCA. Paper presented at the 29th GLOW
  Colloquium, Barcelona, March 5-8, 2006.
  
• Bahan, Benjamin, Judy Kegl, Dawn MacLaughlin and
  Carol Neidle. 1995. Convergent evidence for the
  structure of Determiner Phrases in American Sign
  Language. In FLSM VI Proceedings of the Sixth
  Annual Meeting of the Formal Linguistics Society
  of Mid-America, Volume 2. Bloomington, Indiana,
  Indiana University Linguistics Club, 1-12.
• Bouchard, Denis. 1997. Sign Languages and
  Language Universals The Status of Order
  Position in Grammar. Sign Language Studies 91
  Summer 1996.101160.
• Bouchard, Denis and Colette Dubuisson. 1995.
  Grammar, Order Position of Wh-Signs in Quebec
  Sign Language. Sign Language Studies 87.99139.
  
• Chomsky, N. 1995. The Minimalist Program.
  Cambridge, MA MIT Press.
  
• Chomsky, N. 2005. On Phases. Ms. MIT.
• Chomsky, N. 2006. Approaching UG From Below. Ms.
  MIT.
  
• Cinque, G. 2005. Deriving Greenberg's Universal
  20 and Its Exceptions. Linguistic Inquiry 36 (3)
  315-332.
  
• Greenberg, Joseph H. 1963. "Some universals of
  grammar with particular reference to the order of
  meaningful elements." In Greenberg, Joseph H.
  (eds.) Universals of grammar, 73-113. Cambridge,
  Mass. MIT Press.
• Hopkins, A. W. 1988. Topics in Mohawk grammar.
  PhD. Diss., University of New York.
  
• Kayne, Richard. 1994. The Antisymmetry of Syntax.
  Cambridge, Mass. The MIT Press.
  
• Kayne, Richard. 2004. Antisymmetry and Japanese,
  in Lyle Jenkins, ed., Variation and Universals in
  Biolinguistics, Elsevier.
  
• Kayne, Richard. 2005. On parameters and on
  principles of pronunciation. Ms.
  
• Lai, Yu-Ting. 2005. Noun Phrase in Taiwan Sign
  Language. MA thesis, National Chung Cheng
  University.
  
• Legate, Julie Anne. 2003. Arguments and Adjuncts
  in Warlpiri. Invited talks, Linguistics
  Colloquium Series, New York University, and
  University of Connecticut.

60

• Lillo-Martin, Diane. 1990. Parameters for
  questions evidence from WH-movement in American
  Sign Language. In Sign Language Research
  Theoretical Issues, ed. C. Lucas, 211-222.
  Washington, DC Gallaudet University Press.
• Lillo-Martin D. 2001. One syntax or two? Sign
  language and syntactic theory. Glot
  International, November 2001, vol. 5, no. 9-10,
  pp. 297-310.
• MacLaughlin, Dawn. 1997. The structure of
  determiner phrases evidence from American Sign
  Language. PhD dissertation, Boston University.
  
• Marantz, Alec. 1984. On the Nature of Grammatical
  Relations. Cambridge, MA. MIT Press.
  
• Mithun, M. 1984. The evolution of noun
  incorporation. Language 60 847-94.
  
• Neidle, C. 2002. Language across modalities.
  Linguistic Variation Yearbook 2 71-98.
  
• Neidle, Carol, Judy Kegl, Dawn MacLaughlin,
  Benjamin Bahan and Robert G. Lee. 2000. The
  Syntax of American Sign Language Functional
  Categories and Hierarchical Structure. Cambridge,
  MA MIT Press.
• Newmeyer, F. 2004. Against a parameter-setting
  approach to typological variation. Linguistic
  Variation Yearbook 4. 181-234.
  
• Reinhart, T. 1979. The Syntactic Domain of
  Syntactic Rules. In F. Guenther and S. Schmidt
  (eds.), Formal Semantics and Pragmatics. Reidel.
  
• Romano, Christine. 1991. Mixed headedness in
  American Sign Language evidence from functional
  categories. In Papers from the Third Student
  Conference in Linguistics, 1991, 241-254.
  Cambridge, MA MIT Working Papers in
  Linguistics.
• Sandler, Wendy and Diane Lillo-Martin. 2006. Sign
  Language and Linguistic Universals. Cambridge
  University Press.
  
• Takano, Yuji. 2003. How antisymmetric is syntax?
  Linguistic inquiry 3433, 516-526.
  
• Tang, Gladys and Felix Sze. 2000. Analysis of
  preverbal nominal expressions in signed and
  spoken languages A case between Hong Kong Sign
  Language and spoken Cantonese. Paper presented at
  the TLSC "The Effects of Modality on Language and
  Linguistic Theory, Amsterdam, Feb. 25-27, 2000.