Review of Modality, Aspect and Negation in Persian
| AUTHOR: Taleghani, Azita H.
TITLE: Modality, Aspect, and Negation in Persian
SERIES TITLE: Linguistik Aktuell/Linguistics Today 128
PUBLISHER: John Benjamins
Maziar Toosarvandani, Department of Linguistics, University of California, Berkeley
In this 179 page monograph, Azita H. Taleghani investigates in some detail the
syntax--and to a lesser extent the semantics--of modality, aspect, and negation
in Persian (specifically, the variety spoken in Iran). While she formulates her
analysis within a minimalist framework, much of the book comprises basic
description of Persian and should be relevant, not only to syntacticians of a
more theoretical bent, but also to students and scholars of Persian and other
Iranian languages. The monograph is divided into seven chapters.
Taleghani opens her study in Chapter 1 (14 pp.) with the question that she aims
to address: ''Does the morpho-syntactic structure of modals map on their root and
epistemic interpretations or vice versa?'' (p. 1). This is followed by some basic
discussion of Persian phrase structure, including its SOV word order, the
absence of a true passive or raising construction, and its lack of an expletive
Chapter 2 (32 pp.) introduces the class of modals Taleghani examines. She first
provides a morphosyntactic classification, dividing them into two categories:
adverbial modals (e.g. ''motma'enan'' 'certainly', ''hatman'' 'certainly', etc.) and
verbal modals. This latter group can itself be subdivided into the auxiliary
modal verbs (''shâyad'' 'may', ''bâyad'' 'must', ''tavânestan'' 'can') and the complex
modal verbs (complex predicates like ''majbur shodan'' 'to be forced', etc.) This
is followed by a semantic classification using basic descriptive categories
(e.g. root vs. epistemic, possibility vs. necessity). Having done this, the
author explores the interaction between the different types of modals and the
aspect of the clause they embed. While root modals are only compatible with
complements in the present subjunctive, epistemic modals can take complements
either in the present subjunctive or the present perfect subjunctive.
In Chapter 3 (53 pp.), the most substantive section of the monograph, Taleghani
uses the morphosyntactic and semantic typologies developed in Chapter 1 to
investigate their syntactic representation. Overall, there is no correlation
between whether a modal expresses root or epistemic modality and its syntactic
structure. Complex modal verbs with a root interpretation are all control verbs
(though some, in the sense of Wurmbrand 1998, exhibit ''syntactic control'', where
the infinitive has a PRO subject, while others exhibit ''semantic control'', in
which the infinitive lacks a subject altogether). In contrast, the auxiliary
modal verbs, which Taleghani argues realize the T head, take a verb phrase
complement (not a full clause as in a control structure). The auxiliary modal
verbs can also have epistemic interpretations with the same syntactic
representation, while all complex verbs with an epistemic interpretation occur
in a ''pseudo-raising'' structure (essentially, a CP embedded under a predicate
whose subject is null).
In Chapter 4 (29 pp.), Taleghani analyzes the syntax of negation in Persian. She
proposes that negation is realized syntactically as a Neg head that takes TP as
its complement. The morphological reflex of negation occurs on the clause's
predicate through Agree: in a clause with a simplex predicate, v Agrees with Neg
in the feature [negation], which is realized as ''na-'' or ''ne-'' depending on the
verb that raises to adjoin to v. This is the simple case; however, Taleghani
must also account for the position of negation in complex predicates and
temporal and aspectual constructions with more than one verbform (i.e. the
future, progressive, and perfect aspect constructions).
The author uses this analysis of Persian negation in Chapter 5 (21 pp.) to
investigate the scopal interaction between negation and modals. The scopal
ambiguity that exists in English between a modal like ''may'' and negation, does
not exist in Persian since, with a few gaps in the paradigm, the surface word
order of negation relative to the modal maps directly onto their scopal order.
When the negative prefix appears on the auxiliary ''bâyad'' 'must' itself,
negation takes wide scope over the modal, under either its root or epistemic
interpretations, whereas when it appears prefixed to the embedded predicate,
negation takes narrow scope.
In Chapter 6 (13 pp.), Taleghani returns to the issue of word order that she
introduced in Chapter 1. While Persian has SOV word order, with DP direct
objects occurring before the verb, bare sentential complements (complements that
are overtly CPs) occur after the verb. This is problematic for her structural
analysis of complex modal verbs like ''majbur budan'' 'to be obliged' since, in
the analysis of Folli et al. (2005), the light verb ''budan'' (lit. 'to be') is
the overt realization of v, while ''majbur'' 'obligation' heads the complement of
v: if the CP is sister to ''majbur'', then ''budan'', which in Farsi is
right-headed, should appear--contrary to fact--linearly after the sentential
complement. After considering a number of different analyses, Taleghani settles
on one in which the CP raises out of vP before vP itself undergoes remnant
movement to the left of CP.
Chapter 7 (4 pp.) is a short conclusion that states her answer to the question
posed in Chapter 1 (''No.'') and summarizes the high points of her analysis.
In addressing modality, aspect, and negation, Taleghani has chosen to analyze
quite a large chunk of Persian grammar. Even though the same areas of English
grammar have consumed a far greater number of pages than are found in this
monograph, Taleghani does an excellent job of condensing the essential insights
of this past research and applying them to the syntax of modality, aspect, and
negation of Persian, which, prior to the publication of her monograph, had
received scant attention. It is thus a major step toward a level of
understanding comparable to what we already possess for English.
At the very beginning of her monograph, Taleghani states that her goal is to
address the question of whether ''the morpho-syntactic structure of modals map[s]
on their root and epistemic interpretations or vice versa?'' (p. 1). This could
profitably be expanded into the two following subquestions. First, is the
semantic distinction between root and epistemic modality reflected in the
morphology and/or syntax of Persian modals? And, second, does negation interact
scopally with root and epistemic modals in the same way? I consider her answers
to these two questions in that order.
To the first question, Taleghani answers in the negative: modals in Persian
cannot be classified morphologically or syntactically in a way that lines up
with the semantic division between root and epistemic modality. While I
ultimately find this conclusion convincing, the so-called auxiliary modals do
not seem to me to form a natural class distinct from the so-called complex modal
verbs, either morphologically or syntactically. There are three purported
members in the class of auxiliary modals: ''shâyad'' 'may', ''bâyad'' 'must', and
''tavânestan'' 'can'. Taleghani provides some evidence that they are distinct from
the adverbial modals, but she does not explicitly argue that they are distinct
from the complex modal verbs (besides the trivial difference that they are not
complex predicates). One might think that a lack of verbal agreement is a
morphological property of the auxiliary verbs, since ''bâyad'' 'must' and ''shâyad''
'may' are both invariant in their realization. But ''tavânestan'' 'can' is not
defective in this way: it agrees fully in person and number (p. 19). Moreover,
there are some complex modal verbs, such as ''majbur budan'' 'to be obliged' that
only have a third person singular realization (p. 21).
Nor do the auxiliary modals seem to have a syntactic structure distinct from the
complex modal verbs. Taleghani herself analyzes ''tavânestan'' 'can' in the same
way as a complex modal verb; it embeds a CP (p. 97). She assigns a different
structure to ''shâyad'' 'may' and ''bâyad'' 'must'--they are the realization of T
and take a vP complement--but this predicts, incorrectly, that they should not
be able to embed a phrase headed by the complementizer ''ke'' 'that':
(1) tanhâ yek jomleh mi-gu-yam: bâyad ke bi-yâ-yi.
only one sentence PRES-say-1SG must that SUBJ-come-2SG
'I only say one thing: You have to come.'
(2) agar senf-e IT hamin ast ke mi-bin-am shâyad ke bi-gonâh-am
if trade-EZ just.this is that PRES-see-1SG may that man. I without-fault-am
'If the field of IT is just what it seems, it may not be my fault.'
Based on the presence of ''ke'' in these naturally-occurring examples, ''bâyad''
'must' (1) and ''shâyad'' 'may' (2) take CP complements.
Moving on to her second question, I first consider her analysis of negation.
Briefly, the interpretive content of negation is housed in a Neg head, which
must Agree with a verbform bearing the [+negation] feature, realized as the
prefix ''na-'' or ''ne-'' (p. 124). (I interpret Taleghani to mean that Neg bears an
interpretable [negation] feature, while the verbform bears an
uninterpretable/unvalued [negation] feature that must be checked in order for
the derivation to converge.) This works for simplex verbs as well as complex
predicates (Neg Agrees with v), but runs afoul of Persian's complicated tense
and aspect constructions. Consider the past perfect construction, illustrated in
(3), which is comprised of a participle followed by the past tense copula (here,
(3) sârâ davâ-sh-o na-xorde bud.
Sara medicine-her-OBJ NEG-eat.PART was
'Sara hadn't taken her medicine.' (p. 127)
In the past perfect, negation shows up prefixed to the participial component.
But assuming Taleghani's analysis of this construction, where the copula
realizes v while the participle heads the complement of v, there is no reason
not to expect negation to also show up on the copula:
(4) *sârâ davâ-sh-o xorde na-bud.
Sara medicine-her-OBJ eat.PART NEG-was
Intended: 'Sara hadn't taken her medicine.'
The null Neg head in (4) checks the [negation] feature on the copula rather than
on the participle (note that ''nabud'' is perfectly well-formed morphologically).
This problem is more general: anytime there are two or more verbforms in a
sentence, Neg will be able to Agree with any head bearing the [negation] feature
that it c-commands. This means that the different scopal interpretations of
negation cannot reflect variation in the position of the Neg head. Taleghani
proposes that the wide scope interpretation of negation over a modal auxiliary
is derived when Neg occurs high, above TP, which is headed by the modal; Neg
agrees in the [negation] feature with the modal, which is realized as the
negative prefix ''na-/ne-''. The narrow scope interpretation derives from a
structure in which Neg occurs above vP but below TP, with Neg now Agreeing with
the embedded verb. But there is no reason that Neg cannot occur high (above TP)
and Agree with the embedded verb: this would produce the ungrammatical result in
(5), where negation is realized morphologically on the embedded verb while
taking wide scope over the modal.
(5) *sârâ bâyad be in konferâns na-r-e.
Sara must to this conference NEG-go-3SG
Intended: 'It is not the case that Sara must go to this conference.'
Setting this difficulty aside, Taleghani's analysis does derive the fact that
root and epistemic modals do not behave differently as regards the position of
negation. Both root modals like ''majbur budan'' 'to be obliged' and epistemic
modals like ''momken budan'' 'to be possible' allow either wide or narrow scope
with respect to negation. Even the ambiguous auxiliary modal ''bâyad'' 'must'
allows both interpretations under both its root and epistemic understandings.
This follows from the fact that Agree is blind to anything but the featural
makeup of the probe and goal: whether or not a modal receives a root or
epistemic interpretation is irrelevant.
The writing is generally clear, though there are a few opaque passages and a
large number of abbreviations. The plentiful tables summarizing the results of
Taleghani's investigations were extremely useful in aiding understanding.
Besides several typos that should have been caught in proof, the Persian data
and their accompanying interlinears are all too frequently incorrectly parsed or
glossed: e.g. in example 2 (p. 18), ''biyâd'' 'come (subj. 3sg.)' is broken down
as bi-y-âd, when the correct parse is bi-yâ-d (SUBJ-come-3SG).
These criticisms should not detract from Taleghani's achievement: the book is a
foundational work in the study of Persian syntax.
Folli, Raffaella, Heidi Harley, and Simin Karimi. 2005. Determinants of event
type in Persian complex predicates. _Lingua_ 115:1365-1401.
Wurmbrand, Susi. 1998. Infinitives. Ph.D. dissertation, Massachusetts Institute
ABOUT THE REVIEWER
Maziar Toosarvandani is a doctoral student in linguistics at the University of
California, Berkeley. The core of his research focuses on phenomena that
elucidate how the syntax, semantics, and pragmatics interact, primarily in the
Iranian languages (Persian and Zoroastrian Dari), the Numic languages (Northern
Paiute), and Germanic (English and Danish).