LINGUIST List 7.596

Mon Apr 22 1996

Sum: Use of MUST + HAVE + participle

Editor for this issue: Ann Dizdar <>


  1., Summary: Must have been submitted

Message 1: Summary: Must have been submitted

Date: Mon, 22 Apr 1996 20:15:46 +1000
From: <>
Subject: Summary: Must have been submitted

On 4th March, I posted the following example for evaluation by
Linguist subscribers on the usage of MUST:
	All forms must have been submitted by February 16.

This example appeared on a home page of a local university
administration Web service. At the time at which I read it, it was
prior to February 16 (some time around February 12, if I recall.) The
question I asked was whether the use of MUST + HAVE + participle could
be used deontically as a 'future perfect' form, and whether it could
be considered deviant or ambiguous.

I'd like to thank the following people who replied to this posting:

Donna Andrews, Robert Beard, Ginny Brennan, Annabel Cormack, Dick
Crouch, Peter Daniels, Susanne Dopke, Bethany Dumas, Lynne Hewitt,
Graham Katz, John Lawler, Steve Nicolle, Eduardo-Jose Pechorro, Heidi
Quinn, John Reighard, Raphael Salkie, Dan Slobin, Larry Trask, Theo
Venneman, and John Verhaar.

Of the 20 responses received, 6 found the example perfectly acceptable
(Group 1), 8 thought it was OK but rare, or confined to formal usage
(Group 2), and 3 considered it either impossible or unacceptable (Group
3). Two respondents considered it acceptable, but not deontic (see
below). One respondent did not evaluate the sentence.

Some of the comments of those who accepted it are as follows:

"It's totally grammatical, and in context not even the least bit
"Although Hofmann claimed in his early work that the perfect aspect was
incompatible with root modality, I don't think this is strictly true.
Especially when the reference time is explicitly provided by an adjunct
or in the discourse ... "
"I understand it to mean unambiguously that when Feb. 16 gets here, all
forms must already have been submitted."
"It just seems to be a strong deontic modal in a future perfect
" ... there must be an explicit or implicit reference time, and that
time must be in the future with respect to the time at which your
example is produced, and also with respect to the date cited." (speaker
of US English).

Amongst the second group, some of the comments were:

"I think the kind of deontic interpretation of MUST + HAVE + past
participle that you noticed is quite common in bureaucatic texts/
instructions. While these are spearsely represented in the LOB corpus of
British English, the following ambiguous example did come to light:
'Second --- for benefit to be paid at the standard rate the husband must
have paid or have been credited with a yearly average of 50
" ... 'must have been submitted' looks at the act of submitting as an
event in the past by projecting the position of the writer into a point
in the future after the due date for the submissions, ... "
" In this sense, the sentence seems to be taking a deictically shifted
point of view, wherein the actual present is projected with reference to
a future time as past."
"It's foolishly pompous, but then that's what we expect from deontic
Authorities, right?
"It strikes me as potentially ambiguous, and I'm sure I've come across
deontic uses of MUST + HAVE + past participle here in New Zealand."
" ... the sentence seems quite natural to me (raised in Minnesota, USA)
in its deontic sense, although maybe smacking somewhat of a written
style ... "
" ... it is rare for MUST + have-en to have a deontic sense."

Amongst those who rejected it, some comments were:

"As a speaker of the standard of standard English dialects (central
North Carolina to Central Pennsylvania, USA) I can state categorically
that the assertion 'All forms must have been submitted by February 16'
cannot be made before February 16."
"My immediate reaction was that the sentence is impossible, but I would
not be convinced by this unless I knew why."
" ... there must be some context around it to throw it into a

Comments from the two respondents who did not consider the
deontic/epistemic distinction relevant in this case included the
following, from Steve Nicolle:

"If we treat must as monosemous and view the identification of intended
modality type (deontic, dynamic etc) as inferred by an addressee,
innovations like your example (I assume this is innovative) can be
viewed as pragmatic rather than semantic (encoded) reanalyses.";

and from John Verhaar:

" ... I find that in English use of _must_ as possibly indicative of
obligation is indicative of necessity, or perhaps inevitability instead.
... Thus your Web example just means that, unless the addressee submits
those forms by February 16, the following stage in whatever is supposed
to happen (enroll, e.g.) will prove impossible. The epistemic sense you
find in _must have_ + past part. also has the connotation of

The dichotomy of stative verbs restricted to epistemic senses and active
verbs to deontic senses was addressed by only one respondent, Graham
Katz, who discussed the 'derived' stativity of the predicate in my
example, with the following explanation for the possibility of deontic

" ... imperatives (and perhaps deontic modals) are implicitly and by
default requests for immediate action. Second for a change in state to
occur an event needs to happen. It is impossible for a state to 'become
true' immediately, because some event must intervene. So unmodified
statives cannot be interpreted deontically. On the other hand, if we add
temporal modification, making it clear that the state should not 'become
true' immediately, but at some remote time, then there is time for the
appropriate events to occur."

References supplied:

Groefsema, Marjolein 1995. (paper) JL
Hofmann 1966. 'Past tense replacement and the modal system'. In
McCawley, J. (ed.) _Syntax and Semantics_ 7 (NY: Academic Press)
Picallo, M. Carme. 1990. 'Modal verbs in Catalan'. _Natural Language and
Linguistic Theory_ 8: 285-312.
Sandstroem, G. (1993 dissertation)

John Lawler referred to a modal page on the Web, for those interested.
The address is:
Also, Eduardo-Jose Pechorro has suggested setting up a
multi-disciplinary mailing list on the area of modality. His address is:

Finally, I did in fact run the same example (with '1997' added to the
date) past a number of Australians, who were neither linguists nor
university students. Of the 20 who replied, only 2 did not want to
change the form of the sentence. Of the others, 2 changed the form to
SHOULD HAVE BEEN, 1 to SHOULD BE, and the remaining 15 changed the form
to 'ALL FORMS MUST BE SUBMITTED BY FEBRUARY 16, 1997. My intuitions at
the present stage are that the matter is at least partly a dialectal
one, although naturally more formal statistics would be required to
confirm this. If anyone has more ideas or comments to make on this (or
similar uses of MUST), I'd be happy to hear about them. Meanwhile,
thanks again to all those who replied.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue