R. M. Hare on H. P. Grice on the implicatum of the buletic "if"

Grice explores at some length _four_ alternative interpretations of the "if" imperative. The first is a horseshoe interpretation. A blind logical nose might lead us or be led to the assumption of a link between an "if" imperative and an "if" doxastic. Such a link no doubt exists, but the most obvious version of it is plainly inadequate. At least one other philosopher besides

Grice has noticed that "If he torments the cat, have 

him arrested!" is unlike to express a

"if" imperative; and that even if one

restricts oneself to this or that case in which 

the antecedent clause specifies a

_will_
we find pairs of examples like

"If you will to go to Chicago, 

travel by AA via Cleveland!" or 
"If you will to go to Philadelphia, 

see a psychiatrist!"
where it is plain that one is, and the other is not, 

the expression of an "if" imperative 

(for fun, Grice does not tell which!). A less easily
eliminable suggestion, yet one which would 

still interpret the notion of an 'if' imperative 

in terms of that particular LOGICAL FORM to
which "if", 'hypothetical' and 'conditional' 

attach, would be the
following. Let us assume that it is established, 

or conceded, as legitimate to
formulate an 'if' utterance in which not only 

the consequent (apodosis) is
couched in some mode other than 

the indiative, as in conditional
commands:
"If you see the whites of their eyes, shoot (fire)!"
but also the antecedent (protasis) or some part 

(clause) of them. In which case all of the 

following might be admissible conditionals.
doxastic protasis
"If the cat is sick, take it to the vet"
Mixed (buletic-cum-doxastic protasis
"If you are to take the cat to the vet AND 

there's no cage available,
put it on Martha's lap!
buletic protasis:
"If you are to take the cat to the vet, 

put it in a cage!
If this suggestion seems rebarbative, think of 

this or that quaint "if" utterance
(when it is quaint) as conditionalised 

versions of this or that "therefore"-sequence, such
as:
Volitival-cum-Judicatival Premises:
"i. Take the cat to the vet! ii. There isn't a cage.________________________________
iii. Put the cat on Martha's lap!"

Volitival Premise:
"i. Take the cat to the vet!_________________________________
ii. Put it in a cage!"
And then, maybe, the discomfort is reduced.

Grice considers a second Formal Interpretation: 

"Among 'if' utterances
with an
imperatival or volitival apodosis 

some will have, then, 'mixed'
antecedents (partly judicative, partly volitival) 

and some will have
purely
_judicatival_ antecedents 

(like 'If the cat is sick, take him to the
vet!).
Grice then gives a provisional 

definition of the 'if' and 'non-if' imperative. 

An 'if' imperative is _either_ an
'if' utterance the apodosis of 

which 

is buletic and the protasis 

of which is buletic or mixed 

(buletic-cum-dxastic) OR

it is an _elliptical_ version of such an "if" 

utterance. 

A "non-if"
imperative
is an imperative which is either _not_ 

"iffy" in form or else, if
it
is iffy, has a purely doxastic antecedent. 

Grice makes three 'quick comments' 

on this second interpretation:
First, re: REAL IMPERATIVES: 

The structures which are being offered as a way
of interpreting an iffy and a non-iffy 

imperative do not, as they
stand, offer any room for the appearance

 of practical modalities like
'ought' and 'should' which 

are so prominently visible in the standard
examples of those kinds of imperatives. 

The imperatives
suggested by Grice are _really_ imperatives": 

they conclude "Do
such-and-such!"
not 'You/one ought to do such and such'. 

Grice modifies this suggestion to meet 

the demand for the appearance or occurrence 

of 'ought' (etc) if such occurrence is needed.
Second, it would remain to be decided how 

close the preferred reading of
my 'deviant' conditional imperatives 

would be to the accepted
interpretation of standard hypothetical imperatives. But even if there
were
some divergence, that might be acceptable 

if the 'new' interpretation
turned out to embody a more preicse 

notion than the standard conception.
Third, NEUSTICAL vs TROPICAL 

ANTECEDENTS. 

There are, Grice thinks, serious
doubts of the admissibility of conditionals 

with _non-judicatival_
antecedents, which are for Grice 

connected with the very difficult

question whether the Judicative and the 

Volitival Mode are co-ordinate
or
whether the Judicatival Mode is in some 

crucial sense _prior_ to the
Volitival. Grice confesses he 

does not know 

the answer to that question.
A third formal interpretation 

"links the categorical-hypothetical
distinction to the absolute-relative 

value distinction. Hypothetical
imperatives would be _end-relative_ and 

might be analogous to
evidence-relative probabilities. 

Categorical imperatives would not be
end-relative.
Finally, a fourth Interpretation is not formal, 

but _material_. This is
close to part of what Kant says on the 

topic. It is a distinction
between
an imperative being _escapable_ (hypothetical), 

through the absence of a
particular _will_ and its not being 

escapable (categorical). If we
understand the idea of escabability 

sufficiently widely, the following
imperatives are all escapable, 

even though their logical form is not in
every case the same.
"Give up popcorn!," "To get slim, give up popcorn!", 

If you" will to get slim, give up popcorn!"
Suppose I have no will to get slim. One might say 

that the first
imperative is 'escaped', 

provided giving up popcorn has nothing else to
recommend it, by _falsifying_ 

'You should give up popcorn'. The second
and
the third imperatives would not, 

perhaps, involve _falsification_ but
they
would, in the circumstances, be _inapplicable_ 

to me -- and
inapplicability, too, counts, as escape. 

Categorical imperatives
however,
are in no way escapable.
ReThe Dynamics of Imperatives in Discourse.
Grice then gives three examples which 

had discussed in 
_Aspects of Reason_, which concern 

_arguments_: This we may 

see as an
elucidation to grasp the logical form 

of an hypothetical imperative in
its
dynamics in argumentation. 

We should, Grice suggests, consider
not merely imperatives of each sort, 

together with the range of possible
characterisations, but also the possible forms 

of _argument_ into which
_particular_ hypothetical imperatives might 

enter". His examples being:
"i. Defend the Philosophy Department!
ii. If you are to defend the philosophy department,
learn to use bows and arrows!______________________________________________________
iii. Learn to use bows and arrows!"
Grice says he is using the dichotomy of 

original-derived value. Re i, it is not specified 

whether the will is original or derived. ii
specifies 'conducive to' (means), iii. would 

involve a 'derived' will,
provided ii. is _true_.
"i. Fight for your country!
ii. If you're to fight for your country,

 join up (one of the services)=
!
________________________________________
iii. Join up!
Here, i and iii do not specify the _protasis_. 

If iii did, it would
repeat premise ii.
"i. Increase your holdings in oil shares!
ii. If you visit your father, he'll give 

you some oil shares.
_____________________________________________________________
iii. Visit your father!
This argument (purportedly) transmits value.
Let us explore these characterisations by Grice 

with the aid of R. M. Hare's
distinctions. For Hare in a hypothetical

imperative,
"the protasis contains a neustic/tropic" 

(_Language of Morals_, p.37).
A distinction may be made between 'hypothetical 

imperative' and a
term
used by Grice in his first interpretation of 

the hypothetical
imperative,
that of 'conditional command' 

('If you see the whites of their eyes,
shoot!)) A hypothetical imperative can be
distinguished from a _conditional_ imperative:
"If you want to make bread, use yeast!"
"If you see anything suspicious, telephone 

the police!"
by the fact that _modus ponens_ is not valid for it.

One may use 'conditional imperative' for an 

imperative which is
_grammatically_ conditional, and reserve

 'conditional command' for a
command which is conditional on the 

satisfaction of the antecedent".
Thus,
on thisview, treating the major premise of the 

following
argument
as a hypothetical imperative, turns the 

argument invalid':
Major Premise as Hypothetical Imperative
(32) i. Major Premise:
If you will to make someone mad, 

give him drug D!
ii. Minor Premise:
You will to make Peter mad.
__________________________________________________
Therefore, Give Peter drug D!
The hypothetical imperative tells one only 

what
_means_ to adopt to achieve a given _end_ in 

a way which does not
necessarily _endorse_ the adoption of that end, 

and hence of the means
to it. Thus someone might say,
"If you will to make someone mad, give 

him drug D!"
But, of course, even if you will to do that,
you must _not_ try to do so.
On the other hand, 

the following is arguably valid
because
the major premise is a 'conditional' imperative 

and not a mere
hypothetical
one:
Major Premise as Conditional Imperative:
i. Major Premise:
I you will to make someone mad, give him drug D!
ii. Minor Premise:
Make Peter mad!
__________________________
iii. Therefore, give Peter drug D!
We can explain this in terms of the 

presence of the neustic in the
antecedent of the imperative working

as the major premise. The
supposition
that the antecedent of a hypothetical imperative 

contains a neustic, as
Hare proposes, neatly explains 

why the argument with the major premise
as a
hypothetical imperative is not valid, 

but the argument with the major
premise as a conditional imperative is, 

as well as helping to
differentiate
a hypothetical imperative from a _conditional_ 

one. For, if the
_antecedent_ of the major premise in 

the hypothetical imperative is
volitival, the mere fact that you will to make 

Peter mad does not
license
the inference of the imperative to give 

him the drug; but this _can_ be
inferred from the major premise of the 

hypothetical imperative _together
with an imperative -- the minor premise 

in the conditional imperative
-- to
make Peter mad. In other words, 

whether the subordinate clause
contains a
neustic thus does have have a consequence 

as to the validity of
inferences into which the complex sentence 

enters.
The Principle of Mode Constancy in Imperative 

and Indicative Inference
One may tries to elucidate Grice's ideas on the 

logical
form of the hypothetical imperative proper.
Hare's suggestion is, admittedly, rather tentative.
But it might be argued, in the spirit of it, that a 

hypothetical imperative
is
of the form
"i. If !p, !q
ii. |-p
______________
iii. Therefore, !q
But this violates a principle of MODE 

CONSTANCY.

A phrastic
must remain _in the same mode_ 

(within the scope of the same _tropic_)
throughout an argument. 

A conditional imperative does not violate the
principle of Modal Constancy, since it is of 

the form
i. If p, !q
ii. !p
_______________
iii. Therefore, !q". The question of the logical 

form of the
hypothetical imperative is too 

obscure to base much on arguments
concerning it. Holdcroft mentions an 

alternative to Hare's account of the
validity of
an argument featuring a conditional imperative. 

This is to treat the
major
premise of a conditional imperative, 

"as some have urged it should be"
as
an _indicative_ tantamount to 

"In order to make someone mad, you have to
give him drug D". 

Then someone who _asserts_ the major 

premise of a
conditional imperative and _commands_ 

the second premise is in
consistency
committed to commanding the conclusion.

"If" does not always connect phrastic with 

phrastic but sometimes connects
two
expressions consisting of a phrastic and a tropic.

Consider:
"If you walk past the post office, post the letter!"
The antecedent of this imperative states, it seems,
the _CONDITION_ under which the imperative 

expressed becomes operative, and
so
can _not_ be construed imperatively, since 

an imperative cannot itself
state a condition. Hence, the antecedent 

ought not be within the scope
of
the imperative modal operator "!", and whatever 

we take to represent the
form of the utterance above we 

must not take "!(if p, q)"
to do so. One way out. On certain interpretation 

of the Isomorphism
Thesis between Indicative and Imperative 

Inference the utterance has to
be
construed as an imperative (in the generic sense) 

to make the indicative
conditional "If you will walk past the post office, 

you will post the
letter" _true_.

Leaving aside issues of the implicature of "if", 

that
the utterance can _not_ be so construed 

seems to be shown by the fact
that
the imperative to make the associated 

indicative conditional true is
conformed with by one who does not walk 

past the post office. But it
seems
strange at best to say that the utterance is 

conformed with in the same
circumstances. This 'strangeness' 

is aptly explained away by
Hare
in terms of Gricean implicature.

At Oxford, Dummett
was endorsing this idea that a conditional 

imperative be
construed as
an imperative to make an indicative 

material conditional true (also
Dummett
1973:339) 

Dummett urges to divide conditional

 imperatives into those
whose
antecedent is "within the power of the 

addressee" --- like the
utterance in
question -- and those in which it is not). 

Consider:
"If you go out, wear your coat!"
One may be not so much concerned with how 

to _escape_ this, as Grice
was,
but how to _conform_ it.

A child may choose not to go out in

order to comply with the imperative". 

For an imperative whose
antecedent is
_not_ within the power of the addressee, e.g.
"If anyone tries to escape, shoot him!
it is indifferent whether we treat it as a
conditional
imperative or not, so why bother. A 

small caveat 
here: If no one tries to escape, the imperative 

is _not violated_. One might ask,
might there not be an important practical 

difference bewteen saying
that
an imperative has not been violated and 

that it has been complied with?
Dummett ignores this distinction.

One may feel
think there is much of a practical difference 

there (Is Grice an
intuitionist?)=
.Suppose that you are a frontier guard and the
antecedent has remained unfulfilled. Then, 

whether we say that you complied
with it, or simply did not _violate_ it will 

make a great deal of
difference if you appear before a war 

crimes tribunal. But then I don't
see why I would be in the war crimes 

tribunal in the first place.
For Dummett, the fact that in the case of 

an imperative
expressed by a conditional imperative 

in which the antecedent is not
within
the agent's power, we should not say 

that the agent had obeyed just on
the
ground that the antecedent is false, is no 

ground for construing an
imperative as expressing a conditional 

command: for there is no
question of
fixing what shall constitute obedience 

independently of the
determination
of what shall constitute disobedience.

This complicates the issues.
One may

with Grice and Hare, defend imperative
inference against A. J. P. Kenny and 

B. A. O. Williams. What
is
questioned by the sceptics about 

imperative inference is whether if each
one of a set of imperatives is used with the 

force of a command, one can
infer a _further_ imperative with that force 

from them". Cfr. Aristotle
on
the practical syllogism. One may

be more conservative
than Hare.

Consider:
" i. If you stand by Jane, don't look at her!
ii. You stand by Jane
__________________________________________
iii. Therefore, Don't look at her!
This is valid. However, the following, 

obtained by anti-logism, is
_not_:
"i. If you stand by Jane, don't look at her!
ii. Look at her!
_____________________________________________
iii. Therefore, You don't stand by Jane."
Honestly, it seems more reasonable, 

one may say, to deny Hare's 

thesis
and maintain that Anti-logism is valid in 

imperative inference than it
is
to hold onto Hare's thesis and deny that 

antilogism is valid in the
case in
question.
The ordering of tropics and neustics.
Consider:
"i. Varnish every piece of furniture you make!
ii. You are going to make a table.
_______________________________________________
iii. Therefore, Varnish it!
This is _prima facie_ valid. The following,
however,
switching the order of the neustics in the 

premisses is not.
"i. You are going to varnish every piece
of furniture that you make.
ii. Make a table!
___________________________________________
iii. Varnish it!
Conversational Implicature at the Rescue.
Problems with "or": Consider

A. Ross's infamous example
"i. Post the letter!
_______________________________
ii. Therefore: Post the letter or burn it!
as 'invalid' (Ross 1944:38 -- endorsed by 

B. Williams). To permit to do p or q is 

to permit to do p and to
permit
to do q. Similarly, to give permission 

to do something is to lift a
prohibition against doing it. Admittedly, 

Williams does not need this
so
I'm stating his claim more strongly than he 

does. One may review
Hare's
way out (defense of the validity of 

the utterance above in terms of
Gricean
implicature. Hare claims that whilst 

the premise's "permissive
presupposition" (to use the term 

introduced by Williams) is entailed by
it,
the conclusion's is 

only _conversationally implicated_. 

Typically for an
Isomorphist, Hare says this is something 

shared by indicative
inferences.
As Hare puts it in 'Some alleged differences 

between
imperatives and indicatives:
If, being absent-minded, I ask my wife, 

'what have I done with the
letter?' and she replies that I have
posted it or burnt it, she 

conversationally implicates that she is not
in a
position to say which I have done. She 

also conversationally
implicates that I may not have post it, 

so long as I have burnt it.
Similarly, the future 

tense indicative, "You are going to
post
the letter" has the conversational implicature 

"You may be not going to
post the letter so long as you are going to burn it". 

But this surely
does
not validate "p, ergo p or q"". 

One _can_,
similarly,
say: "Eclipse will win. He may not, of course, if it rains. And I
_know_ it
will not rain".
Problems with "and": Holdcroft considers 

Hare's example in 'Imperative
Sentences':
"i. Put on your AND jump out!
____________________________________
ii. Therefore, Jump out!
Someone who _only_ jumps out of an aeroplane 

does
not
fulfil 'Put on your parachute and jump out!' 

He has done only what is
necessary, but not sufficient to fulfil it. 

Imperatives do not differ
from
indicatives in this respect, except, Hare notes 

that fulfilment takes
the
place of belief (which is the form of 

acceptance apprpriate to
statements.
(Language of Morals). 

Someone who is told
"Jones put on his parachute & jumped out" 

is entitled to believe that
Jones
jumped out. But if he believes that this is _all_ 

Jones did he is in
error"
Cf. R. Edgley's 

treatment of this). One may discuss Hare's test of 

cancellability in the case of the transport
officer who says:
"Go via Coldstream or Berwick!"
It seems
the
transport officer's way of expressing 

himself is extremely _eccentric_.
If
he's not sure if a storm may block one of 

the routes, what he should
say is
"_Prepare_ to go via Coldstream or Berwick!"
As for Hare's application of Grice's cancellability 

thesis here, to
yield,
in the circumstances:
"Go either via Coldstream or Berwick! 

But you may not go via
Coldstream
if you do not go via Berwick, & 

you may not go via Berwick if you do
not go
via Coldstream."
Such qualifications seem to empty the 

imperative of all content and is
thusreminiscent of Henry Ford's 

utterance that people can choose what
colour
car they like provided it is black. But then 

I don't think Ford was
being
illogical, was he.
Refs.
ARISTOTLE. Practical syllogism.
CLARKE D. Mode constancy in mixed 

inferences. _Analysis_ 30.
DUNCAN-JONES AE. Symposium on 

imperatives. See MITCHELL.
FOOT P. Morality as a system of a 

hypothetical imperatives. In _Moral
Philosophy_.
Discussed by Grice in _The Conception of 

Value_.
GEACH PT. Imperative inference. _Analysis_ 23
GREGOR MJ. _The categorical imperative_.
GRICE HP. The conception of value. Clarendon
Aspects of Reason. Clarendon
HARE RM. _The language of morals_. 

Clarendon.
Some alleged differences between imperatives 

& indicatives.
_Mind_ 76, repr. in _Practical Inferences_.
HOLDCROFT D. _Words & deeds_. 

Ch V: Hare's theses: phrastics & neustics.
KENNY AJP. Practical inferences. _Analysis_ 26
MAYO B. Symposium on imperatives. 

See MITCHELL.
MITCHELL B. Imperatives. 

Aristotelian Society 52.
PATON HJ. _The categorical imperative_
WILLIAMS BAO. Imperative inference. 

_Analysis_ 23
WILLIAMS T. The concept of the categorical 

imperative.