in Persian Roya Kabiri Ali Darzi University of Arizona University of Tehran First North American Conference in Iranian Linguistics NACIL1 April 2830 2017 Outline Provide definition of conditional propositions ID: 601372
Download Presentation The PPT/PDF document "A syntactic analysis of Conditionals" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
A syntactic analysis of Conditionals in Persian
Roya
Kabiri
, Ali
Darzi
University of
Arizona, University of Tehran
First North American Conference in Iranian Linguistics (NACIL1)
April 28-30, 2017Slide2
Outline
Provide definition of conditional propositions.
Introduce two distinct views on the structural position of the conditionals in the literature:
Bhatt and
Pancheva
(2006)
Valmala
(2009)
Introduce
sentence-initial, sentence-final and
seemingly sentence-medial conditional clauses in Persian.
Examine
Persian conditionals in order to understand how these constructions are structured, and whether or not the existing theories can account for
them.Slide3
Conditional structures
Conditionals
are linguistic expressions expressed by means of syntactically complex forms which consist of a conditional
clause
(
protasis
or antecedent) and a main clause (apodosis or consequent).
Conditionals involve
an adverbial clause which
makes
the occurrence of one
event
dependent on the occurrence of another (
Inchaurralde
2005
).Slide4
Conditionals have been one of the most significant topics in the areas of semantics, pragmatics and philosophy of language, and have been studied within different approaches (
Kratzer
1986,
Kaufmann 2001
).
T
hey
have not been analyzed syntactically in
detail,
specifically in Persian
. Slide5
Structural position of conditionals
Syntactically, there is a debate as to how these expressions are
constructed:
First approach due to Bhatt
and
Pancheva
(2006):
It is an
adjunction-based
approach
(proposed as equivalent to external
merge).
T
he
sentence-initial conditional clause in English adjoins to TP and in some cases to CP (when preceding
wh
-arguments in
questions).
T
he
sentence-final conditional clause involves VP-adjunction to the right. Slide6
Second approach due to Valmala
(2009):
Spanish
and English sentence-initial conditional clauses are in the
Spec
of
TopP
or
FocP
.
T
he
sentence-final conditional
clauses are
in the Spec of a functional projection,
CondP
.
This
analysis is based on movement (
proposed
as equivalent to internal merge) in some cases.
The
sentence-initial conditional
clause has a
topic or focus interpretation, and it is usually derived via movement from a post verbal position to the front of the sentence. Slide7
Conditionals in Persian
C
onditional marker:
ægær
‘
if’ or more informally
æge
.
Three different positions of the conditional clause
: sentence-initial (1a), sentence-final (1b) and seemingly sentence-medial (1c): (1) Slide8
We show that the adjunction-based approach (proposed by Bhatt and
Pancheva
(2006)) rather than the
other approach advocated by
Valmala
(
2009) best
accounts for the data in Persian.
I
ndependent
syntactic properties such as the interaction of scrambling and principle C of Binding Theory, the structural position of focused
wh-arguments, distribution of higher and lower adverbs
and
vP
deletion
provide evidence for our analysis.Slide9
The sentence-initial conditional clauseT
he
movement-based analysis cannot account for
a sentence-initial
conditional
clause
containing a
referential expression
,
coindexed
with a
pronominal in the matrix clause. Taking into account the interaction of scrambling and principle C of Binding Theory, the sentence in (2) would be predicted to be ill-formed under the movement-based analysis due to the principle C violation, contrary to facts.
(2)Slide10
If the conditional clause is generated in a position following the main clause and then moved to its surface position, it should be
ungrammatical since:
R
econstruction
is well-known to be obligatory for principle C at LF (See
Iatridou
1991,
Sportiche
2005,
Valmala
2009)
Scrambling does not bleed principle C in Persian (Karimi, 2005: 179).
Thus
, the sentence-initial conditional clause is externally merged as an adjunct to TP which is not c-commanded by the subject in the main clause. Slide11
Taking into account principle C and the distribution of lower adverbs in Persian (lower adverbs are adjoined to the
vP
(
Karimi
, 2005
:
125)):
The
sentence-initial conditional clause is
adjoined
to TP.(3)Slide12
If a matrix clause containing a focused wh
-argument (occupying the Spec of
FocP
upon movement in Persian (
Karimi
2005)),
is preceded
by a conditional
clause, it can be argued:
If the
wh
-argument ki ‘who’ is in the Spec of FocP
, it is not possible for the conditional clause to simultaneously occupy this
position.
Persian allows two elements bearing contrastive focus in the same sentence only if at least one of them bears an inherent focus.
Focus position is argued to be unique in
Karimi
(2005
). Following what Bhatt and
Pancheva
(2006) proposed for English, the conditional clause may adjoin to FocP as well. (4) Slide13
It may be argued that they are in the multiple specifiers of
FocP
:
Karimi
(2005
):
two focused elements moved to Specs of
FocP
may not be separated by other syntactic
objects.
Færda ‘tomorrow’ may be construed as a temporal adverb for both the conditional clause and the matrix clause, indicating that there is no adjacency requirement between the
wh
-argument and the conditional
clause.
(5)Slide14
Syntactic position of the sentence-initial conditional
clause:
It
involves
TP-adjunction and
FocP
-adjunction (when preceding the focused
wh
-arguments which have moved to Spec of
FocP
).
This analysis offers support for the Bhatt and Pancheva’s (2006
) analysis
.
It is also consistent with some other studies
(
Haegeman
2004,
Iatridoue
1991)
which consider conditional clauses as adverbial clauses in adjunct positions as they are not triggered by the need to check features (Chomsky 1995).Slide15
The sentence-final conditional clause
Sentence-final conditional clauses involve
vP
-adjunction:
Principle
C of the Binding Theory
vP
deletion
Since
the adjunction-based approach accounted for the sentence-initial conditionals, to maintain consistency I
extend the adjunction-based approach to sentence-final conditionals as well. Slide16
Ill-formedness of sentence (6)
is due to a principle C violation because the pronominal subject of the main clause binds the subject of the embedded
clause adjoined to
vP
.
The sentence-final conditional clause is adjoined to a position lower than TP (the most likely candidate being the
vP
)
inducing principle C
violation.
(6)Slide17
The asymmetry observed in the behavior of sentence-initial and sentence-final
conditional clauses:
It is possible to have
a referential expression in the
sentence-initial conditional
clause
coreferent
with a pronominal subject in the main
clause.
It is impossible to
have a referential expression in the sentence-final conditional clause coreferent with a pronominal subject in the main clause.
Sentence-initial and sentence-final conditional clauses are merged in different positions in the derivation. Slide18
vP deletion shows the sentence-final conditional clause is in a position lower than
NegP
, the most likely candidate being the
vP
. In
(7),
all constituents below the
NegP
, including the conditional clause are elided, leaving the subject and the head
Neg
intact.
(7) Slide19
Syntactic position of the sentence-final
conditional clause:
It
is adjoined to
the right of the
vP
.
It is compatible with Bhatt and
Pancheva’s
(2006) analysis. Slide20
Our proposal for the TP and
FocP
adjunction of sentence-initial conditional clauses can
also account for:
The
grammaticality and ungrammaticality of all the sentences in which the conditional clause appears in the medial
position.
(
8
)Slide21
The sentence in
(
9
)
with two topicalized
DPs preceding
the conditional clause is also explained under
our
proposal.
More
than one syntactic object may be
topicalized in Persian (Karimi 2005).(
9
)Slide22
The ungrammaticality of (
10) is due to:
Topicalization
of the pronominal
un
‘he’
from the main clause subject to the front of the
sentence, places
it in a structural position from which it binds the
referential expression in the conditional clause, inducing principle C violation. Scrambling feeds principle C in Persian (Karimi 2005: 180).
(10)Slide23
Syntactic position of the seemingly sentence-medial conditional clause:
The
merge position of the
seemingly sentence-medial
conditional clause is in no way different from
the
sentence-initial position.
The
relative ordering between the conditional clause and the constituents preceding
it,
is due to scrambling of syntactic constituents to the sentence-initial
position.Slide24
ConclusionsThe
adjunction-based approach
of conditionals proposed
by Bhatt and
Pancheva
(2006) best
accounted
for the data in Persian
over
the other approach advocated by
Valmala
(2009). The conditional clause was argued to be an adverbial clause in Persian adjoined to the TP or FocP in sentence-initial position, and to the right of the vP
in sentence-final position.Slide25
Future Research
Why
left-adjunction is restricted to TP or higher projections such as
FocP
whereas right-adjunction is restricted to
vP
?
Is there
syntactic differences among various kinds of conditional clauses classified in terms of syntax/semantic
grounds?
Is there any
connectivity effects that hold relative to tense/aspect, especially with counterfactuals that may have implications for the syntactic analysis?How does conditional clause interact with
negation?
How phase theory can account for
the
syntactic position
of the conditionals?Slide26
Thanks for your attention!
متشکرمSlide27
Conditional clauses may be focused.
U
nder my
analysis nothing prevents a conditional clause to be modified by
f
æ
q
æ
t
as in
(6).
It doesn’t mean that it is base-generated in that position. Actually, it is base-generated in the adjunct position of the TP and then may optionally move to Spec of FocP.
(6)