LINGUIST List 10.1710

Wed Nov 10 1999

Sum: For Query: 10.1655/ Persian Data and EPP

Editor for this issue: Lydia Grebenyova <>


  1. Lotfi, For Query: 10.1655/ Persian data and EPP

Message 1: For Query: 10.1655/ Persian data and EPP

Date: 9 Nov 1999 12:37:08 EDT
From: Lotfi <>
Subject: For Query: 10.1655/ Persian data and EPP

For Query: 10.1655

Dear Linguists,
On 1 Nov 1999, I posted a query about 'Persian data and EPP', to
which I received responses from Linguists below. While thanking
them all, I post a summary to the list. I re-post the query itself
for those who didn't see it.

Dear Linguists,
The Persian verb 'bayestan' (meaning 'to have to') is deficient in the
sense that contrary to other Persian verbs it cannot be inflected for
person. It may assume one of three morphologically isolated forms: (a)
'bayad' mainly used for present and future events (and even for past
ones if the verb to follow is already inflected for past tense), and
(b) 'bayest/bayesti' for past (but again not necessarily as they can
be used with other verbs inflected for present tense in order to refer
to a present/future tense event). Then when used in a phrase, it is
the verb to follow that must be inflected for person and tense
(although 'bayad' is not atemporal itself as mentioned above):
1. a. (Man) bayad beravam.
 I must go-1st-sing-present.
 "I must go"
 b. (To) bayad beravi.
 You must go-2nd-sing-present.
 "You must go"
 c. ...
2. a. (Man) bayad/bayest miraftam.
 I must go-1sr-sing-past
 " I had to go"
 b. (To) bayad/bayest mirafti.
 You must go-2nd-ding-past
 "You had to go"
 c. ...
'Bayad' cannot be an auxiliary verb because Persian doesn't use
AUXs. Perhaps 'khastan' is the only verbial element that one may
consider as a future-tense auxiliary in Persian:
3. a. (Man) khaham raft.
 I want-1st-sing.-future go-nonfinite
 "I will go"
 b. (To) khahi raft.
 You want-2nd-sing.-future go
 "You will go"
 c. ...
Anyway, here it is 'khastan' that is inflected for tense and
agreement; the main verb (raft) remains non-finite. 'Bayad' cannot be
an auxiliary because it is the verb to follow and not 'bayad' itself
that is inflected for tense and agreement. Interestingly enough, in
Persian subjectless sentences like (4) below there's no inflection for
agreement and tense on the second verb either:
4. Bayad raft.
 must go-nonfinite
 roughly meaning "To go is a must/ one must go"
'Bayad' is not inflected either. Perhaps it is just some temporal
feature of 'bayad' that makes it still a (present-time) sentence.
Although Persian is a prodrop language, one cannot assume that pro is
the subject of such sentences. Actually, the sentence is neutral with
regard to the subject, and whatever personal subject pronoun one
inserts in the subject position will make the sentence ungrammatical:
5. *a. Man bayad raft.
 *b. To bayad raft.
 *c. ...
The sentence will be still ungrammatical if a universal quantifier
occupies the subject position:
6. * Har kas bayad raft.
 every one must go
The grammatical version of (6) is:
7. Har kas bayad beravad.
 every one must go-3rd-sing-present
This seems to be a violation of the Extended Projection Principle
because it is grammatically necessary for the predicate 'bayad raft'
NOT to have a subject.
Does anyone know of similar phenomena in other languages?
"As far as your question is concerned, it is quite common in a lot of
languages to have an impersonal construction of "must", like French il
faut, Italian bisogna, Spanish hay que etc."
Johannes Reese <>
- -------------------------------------------------
"I am not sure how relevant this example is but the French
 il faut partir (roughly equivalent to bayad raft)
has a dummy subject insofar as 'il', although it is the third person
singular masculine subject pronoun, cannot refer to any masculine
singular entity nor can it be replaced by any other pronoun.
However, the verb 'falloir' is fully conjugated in all tenses (but only
in the third person singular). It is followed by the infinitive.
I am not an expert in either Russian or Turkish but I wonder
whether the Russian 'mozhno' (sorry, this is only a transcription) or
the Turkish 'm|mk|n' (I'm not totally sure about the diareses) might
parallel the Persian construction, although they are possibly non-
Jeremy Whistle
- -------------------------------------------------
"It is unclear to me why you are assuming that
pro cannot occupy the subject position of the sentence below:
4. Bayad raft.
 must go-nonfinite
 roughly meaning "To go is a must/ one must go"
Since the subject of that sentence is interpreted as indefinite or
arbitrary, it is not clear that it needs to satisfy the same
as referential pro. In fact, according to the most accepted theory
developed from Rizzi 1986) person markings are not necessary to license
non-referential null subjects (only number agreement is necessary). It
be the case, as Persian shows, that arbitrary pro can be licensed even
without number marking (in fact any marking at all)."
Marcello Modesto.
- ----------------------------------------------------------
"This strikes me as a description of an impersonal verb, 'it is
necessary/required' ~ 'it was necessary/required'.
An analogy that occurs to me is Spanish hay/habia 'there is/was', also
used in the sense of 'must', though my introduction to Spanish preceded
introduction to linguistics and my abilities with it are far from
native-speaker level. I think that hay que SUBJ VERB is the
(que 'that'), and there are no person restrictions.
Perhaps also analogous would be the imperative construction favored in
Pacific Northwest languages in the US (Indian languages). It is my
understanding that they use a construction 'good that X' as an
Somewhat different, but resulting in the same situation would be the
modern Greek future in tha na X. I gather that tha na is uninflected,
whereas X is. The formant tha na is documented as having evolved from
thelo (I wish) na 'that', etc., which was inflected, but has 'worn down'
to an uninflected form. I assume that the Iranian situation is
since I suppose the -d is a reflect of the typical Indo-European third
person marker *-t(i)?
I don't see the application here, but another historical source of
uninflected (not personally inflected) verbs is be + participle
constructions with the inflected be form omitted, e.g., the Slavic pasts
in -l- + gender/number instead of person, which evolve from past active
participles in l."
John Koontz
- ----------------------------------------------------------
"Modern Standard Persian uses two finite modals, XASTAN
and TAVANESTAN, and two non-finite modals, BAYESTAN & SHAYESTAN. It is
generally the case in diachronic change that some modals may develop,
emantically and syntactically, into other forms. Therefore, BAYAD &
SHAYAD are functioning as adverbs now.
The fact that sentence (4) does not take a subject may be explicated in
terms of agreement checking. As a first good approximation, the verb
an infinitive at D-Structure, preceded by an adjunct (adverb) at
S-Structure. Thus, the infinitival element with null agreement checks
case and PRO will appear in such constructions."
A. Soheili
- ---------------------------------------------------------
"What you describe appears to me reminiscent of what has been called
'serial verbs'. This is a phenomenon that occurs in various languages
(e.g., Japanese, some West African languages like Yoruba, and --
much closer to Persian -- in Sanskrit based Indic languages.)
In general, in verb serialization the first verb either remains
uninflected or can be inflected only in very limited ways, whereas the
second verb has the full morphological inflection.
In Gujerati, which is closely related to Hindi, constructions which
correspond to relative clauses in e.g., English, French, German, appear
as something like conjoined sentences with serial verbs at the end.
Hindi-type languages are more-or-less ergative and verb last. Come to
think of it, *all* languages with serial verbs I know of are verb-last
- -------------------------------------------------------
Ahmad. R. Lotfi, Ph. D.
Chair of English Dept.
Graduate School
Azad University at Khorasgan
Esfahan, Iran.
Mail To:
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue