1
120213 FL@IITD/ Bhattacharya
Tanmoy Bhattacharya
Centre for Advanced Studies in Linguistics University of Delhi
[email protected]
120213 FL@IITD/ Bhattacharya
The standard view: CHL : ([F],LEX) EXP However, this has been called into question recently (Stroik 2000, 2009) from the perspective of conceptual necessity, so has been the associated machinery: Phases Phase Impenetrability Condition (PIC) Phase Complement Transfer (PCT) Transfer (Multiple) Spell Out Economy conditions like Shortest Move
2
120213
Similarly: Why Internal Merge (IM)?
FL@IITD/ Bhattacharya
Axioms and Derived Operations 2. Unifying External and Internal Merge 3. A model without IM is similar to a Lto-R parsing algorithm 1.
3
120213
Head as a postulate/ axiom should be regarded as a purely formal statement (and not necessarily as a fact based on experience) asymmetry Head Projections array of LIs from LEX for further computation Head basic local relations (where H-Complement is more fundamental, others being “elsewhere”) Head Chain Link
FL@IITD/ Bhattacharya
A Head gives rise to a
4
120213
A CFC Head allows an “extra” Spec [Wh(C), EPP(T), and OS(v)]: = [XP [(EA) H YP]] The set resulting from S-selection and Hasymmetry, has the following properties that hold between the members:
FL@IITD/ Bhattacharya
a)
HEAD [XP [(EA) H(=C) YP]]
b)
Relate to higher T
HEAD [(EA) H(=V) YP]
c)
ZP [XP [ H(=C) YP]]
Extended Locality [pseudo direction] 5
120213
Optimal
FL@IITD/ Bhattacharya
solutions to design principles: Filler-Gap Dependency Displacement External motivation: distinct semantic interpretation (Q-scope) Processing (reanalysis) Displacement and devices implementing it are only apparent imperfections 6
120213
FL@IITD/ Bhattacharya
Chain Link + Adjuncts DOMAIN Basic elements of representations (to be derived derivationally = CH) Basic transformation operation: FORM CHAIN (tension btween SM versus fewest steps)
[e seems [e to be likely [John to win]]] CH = (John, t’, t) in a single step via Form Chain
7
120213
With recursive PS rules, Generalized Transformation (GT) was given up, MP revives is through parallel derivations GT: Binary operation: (K, K1) K* Move : Singularly operation: K K* Unifying: Target external to K in both: Extension Condition [substitution operations always extend target]
FL@IITD/ Bhattacharya
1 (a)
[ seems [ is certain [John to be here]]] it 2
8
120213 FL@IITD/ Bhattacharya
[Bhattacharya, 2008] The problem of encoding movement in a left-to-right parser generally They also highlight the difficulty of implementing the syntactic finding that subjects in Bangla occupy a high topic-like position. Designing an efficient P-G parser requires reexamination of the idea of EPP/EFs in general
9
120213
FL@IITD/ Bhattacharya
The CP as a Wh-phrase raises from a postverbal SVO base-position to a Wh-position located below the subject, resulting in licensing of the Wh-phrase as follows: Sub [CP … wh …]i V ti Bangla has WH movement
10
120213 FL@IITD/ Bhattacharya
Position of subject hides Wh movement in Bangla. Thus only referentially specific elements occur as subjects preceding Wh-phrases in the “subjectposition”: Chele du-To [kon boi-Ta]i poRlo ti ? boy two-CL which book-CL read „Which books did the two boys read?‟ #Du-To chele [kon boi-Ta]i poRlo ti ? two-CL boy which book-CL read
[DP +Specific/ *-Specific] > WH 11
FL@IITD/ Bhattacharya
Bangla is a language with obligatory overt Wh-movement from an underlying SVO base position Non-Wh arguments (Subjects)/ adjuncts raise to a Topic-like position The Wh/Q-licensing position follows the regular surface position of the subject and is headed by a polaritytype functional head
120213
12
120213
FL@IITD/ Bhattacharya
13
120213 FL@IITD/ Bhattacharya
Since Chain Links are axiomatic, Minimal Domain is derivative and so is Agree Agree is an Operational explanation (and so is Phiagreement) Agree as valuation-driven rather than interpretability driven Pesetsky and Torrego (2007):
“[…] reason for assuming that the T-feature of Tns is unvalued, though interpretable: the fact that Tns appears to learn its value in finite clauses from the finite verb.” interpretable unvalued feature: [uF:val] values [iF: __]
Selected Modal/T/C is of that type (Wurmbrand)
14
120213 FL@IITD/ Bhattacharya
Zeijlstra (2010): iF [goal] » uF [probe] Haegeman and Lohndal (2010): α Agrees with β if α c-commands β, α and β both have a feature F, and there is no γ with the feature F such that α c-commands γ and γ c-commands β. o Adger (2003): An uninterpretable feature F on a syntactic object Y is checked when Y is in a c-command relation with another syntactic object which bears a matching feature F.
15
120213
FL@IITD/ Bhattacharya
16
120213
FL@IITD/ Bhattacharya
17
120213
FL@IITD/ Bhattacharya
18
