[lit-ideas] Re: The Immanuel
- From: Jlsperanza@xxxxxxx
- To: lit-ideas@xxxxxxxxxxxxx
- Date: Tue, 28 Sep 2004 12:44:35 EDT
In a message dated 9/28/2004 11:01:21 AM Eastern Standard Time,
erin.holder@xxxxxxxxxxx writes:
I really thought there were
1) Maxims
2) Hypothetical imperatives
3) Categorical imperatives
Maxims, I thought, were practical precepts that assume purposes or
particular ends. Imperatives, I thought were practical laws and that
practical laws do _not_ assume purposes or particular ends, that they are
followed out of adherence to the law in and of itself. I thought that this
is how maxims and imperatives differ. Kant says it himself in the Remark on
Definition I (at 5:20 in my books) - "Thus maxims are indeed principles but
_not imperatives_". The difference I thought then between maxims and
hypothetical imperatives was this:
A maxim is merely a subjective desire dependent on sensibility - it can be
either physical or intellectual (I want pizza or I want truth). A
hypothetical imperative is _not_ a maxim insofar as it is not just a
particular desire. Yes, it is a desire, and it is subjective, but, unlike
in a maxim, there is a subjectively conditioned _necessity_ (you, as a
general rule, should work in your youth so as not to want in old age),
albeit a necessity that cannot be "presupposed in the same degree in all
subjects" (5:21).
-----
Okay. If you get a chance, you could consult Grice's second book, _The
Conception of Value_. In the chapter 2 he very much focuses on the logical form
of
a maxim _qua_ imperative. Examples discussed by Grice:
If let the cat be taken to the vet, then let it be put in a cage.
If let the cat be taken to the vet and there is no cage available, then let
Martha put it on her lap.
If the cat is sick, let it be taken to the vet.
Let the cat go to the vet, so let it be put in a cage.
Let the cat go to the vet; there isn't a cage, so let Martha put the cat on
her lap.
You should give up popcorn.
To get slim, you should give up popcorn.
If you want to get slim, you should give up popcorn.
---- Grice gets more formal in his _third_ book, where he proposes a
derivation of Kant's ideas: Ceteris paribus, for any creature x (for any A, B),
if x
wills A and judges that if A, then A as a result of B, then x wills b. x
wills that (for any A, B) if x wills A and judges that if A, A as a result of
B,
then x is to will that A. Etc., etc. I provide some further quotes below.
Cheers,
JL
-----
Grice writes: "Today's Special (2. Cats and Hypes) calls (even clamours) for
interpretation" (Aspects of Value p. 49).
In _The Conception of Value_ (2nd lecture) Grice explores at some length
_four_ alternative interpretations of the hypohtetical imperative.
The first is a Formal Interpreation:
"A blind logical nose might lead us (or be led) to the assumption of a link
between hypothetical imperatives and hypothetical statements
(propositions). Such a link no doubt exists, but the most obvious version
of it is plainly inadequate. At least one other philosopher besides myself
hs noticed that"
(1) If she molests the children, have her arrested!
"is unlike to express a hypothetical imperative; and that even if one
restricts oneself to caes in which the antecedent clause specifies a _will_
we find pairs of examples like":
(2) If you will to go to Chicago, travel by AA via Cleveland!
(3) 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 a
hypothetical imperative (I won't tell you which). A less easily eliminable
suggestion, yet one which would still interpret the notion of a
'hypothetical imperative' in terms of that particular logical form to which
the names 'hypothetical' and 'conditional' attach, would be the following.
Let us assume that it is established, or conceded, as legitimate to
formulate conditionals in which not only the consequents (_apodoses_) are
couched in some mode other than the indiative, as in conditional commands:
(4) If you see the whites of their eyes, shoot (fire)!
but also the antecedents (_protases_) or some part (clause) of them; in
which case all of the following might be admissible conditionals:
Judicatival Antecedent:
(5) If the cat is sick, take it to the vet.
Mixed (Judicatival-cum-Volitival Antecedent)
(6) If you are to take the cat to the vet & there's no cage available,
put it on Martha's lap!
Volitival Antecedent:
(7) If you are to take the cat to the vet, put it in a cage!
"If this suggestion seems rebarbative, think of these quaint conditionals
(when they are quaint) as conditionalised versions of _arguments_, such as:
Volitival-cum-Judicatival Premises:
(8) i. Take the cat to the vet!
ii. There isn't a cage.
________________________________
iii. Put the cat on Martha's lap!
Volitival Premise:
(23) i. Take the cat to the vet!
_________________________________
ii. Put it in a cage!
"and then maybe the discomfort will be reduced".
Grice considers a second Formal Interpretation: "Among conditionals with an
imperatival or volitival consequent 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!).
I might now give a provisional definition of the terms categorical and
hypothetical imperative. A hypothetical imperative is _either_ a
conditional the consequent of which is imperatival and the antecedent of
which is volitival or mixed (partly judicatival, partly volitival) _or_ it
is an _elliptical_ version of such an imperative. A categorical imperative
is an imperative which is either _not_ conditional in form or else, if it
is conditional, has a purely judicatival antecedent." Grice makes three
'quick comments' on this second interpretation:
(i) REAL IMPERATIVES: The structures which are being offered as a way
of interpreting hypothetical and categorical imperatives 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. Grice writes: "the imperatives
suggested by me are _really_ imperatives": they conclude "Do such-&-such!"
not 'You/one ought to do such and such'. "But maybe my suggestion could be
modified to meet the demand for the appearance or occurrence of 'ought'
(etc) if such occurrence is needed.
(ii) "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.
(iii) NEUSTICAL vs TROPICAL ANTECEDENTS. "There are, I think, serious
doubts of the admissibility of conditionals with _non-judicatival_
antecedents, which will be to my mind 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. I do not know the answer to this 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".
(24) Give up popcorn!
(25) To get slim, give up popcorn!
(26) 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".
The Dynamics of Imperatives in Discourse.
Grice then gives three examples which I've discussed in the thread on
_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. Grice writes: "We should, I suggest, 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:
(27) 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!
Comments: Grice says he's "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_.
(28) i. Fight for your country!
ii. If you're to fight for your country, join up (one of the services)!
________________________________________
iii. Join up!
Comments: i and iii do not specify the _protasis_. If iii did, it would
repeat premise ii.
(29) i. Increase your holdings in oil shares!
ii. If you visit your father, he'll give you some oil shares.
______________________________________________________________
iii. Visit your father!
Comments: "this argument (purportedly) transmits value".
Let us explore these characterisations by Grice with the aid of Hare's
distinctions. Holdroft has Hare saying that, in a hypothetical imperative,
"the protasis contains a neustic/tropic" (_Language of Morals_, p.37).
Holdcroft makes a distinction 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!)): Holdcroft writes: "A hypothetical imperative can be distinguished
from a _conditional_ imperative:
(30) If you want to make bread, use yeast!
(31) If you see anything suspicious, telephone the police!
"by the fact that _modus ponens_ is not valid for it." Holdcroft writes
(p.93). "I 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 Holdcroft's view, treating the major premiss 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.
__________________________________________________
Give Peter drug D!
Holdcroft comments: "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,
(33) 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, Holdcroft says the following is "arguably valid" because
the major premise is a 'conditional' imperative and not a mere hypothetical
one:
(34) 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. Give Peter drug D!
Holdcroft explains 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
(Clarke). Holdcroft then tries to elucidate Hare's ideas on the logical
form of the hypothetical imperative proper.
Holdcroft writes: "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
(35) i. !p -> !q
ii. |-p
______________
iii. !q
But this violates a principle of MODE CONSTANCY (see Clarke): 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
(36) i. p -> !q
ii. !p
_______________
iii. !q
Holdcroft has a caveat here: "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". Holdcroft concludes that "if"
"does not always connect phrastic with phrastic but sometimes connects two
expressions consisting of a phrastic and a tropic" (p.87). Holdcroft
further considers:
(37) If you walk past the post office, post the letter!
Holdcroft writes: 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 "!(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", Holdcroft writes: "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." (I think this 'strangeness' is aptly explained away by Hare
in terms of Gricean implicature). Interestingly, Holdcroft quotes Dummett
(1958) as 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:
(38) If you go out, wear your coat!
Holdcroft is not so much concerned with how to _escape_ this, as Grice was,
but how to _conform_ it. He writes: "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.
(39) If anyone tries to escape, shoot him!
it is, Holdcroft thinks, "indifferent whether we treat it as a conditional
imperative or not", so why bother. There's a small caveat by Holdcroft
here: If no one tries to escape, the imperative is _not violated_. He asks:
"might there not be an important practical difference bewteen saying that
an imperative has not been violated and that it has been complied with?"
(Holdcroft thinks Dummett ignores this distinction). Honestly, I don't
think there is much of a practical difference there (Am I an intuitionist?).
Holdcroft writes: "Suppose that you are a frontier guard and the antecedent
of (65) 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...
Holdcroft here quotes from Dummett.
Holdcroft quotes 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" (Dummett 1973:343). Personally, I
think this complicates the issues.
Holdcroft cites Geach, who with Grice and Hare, defends imperative
inference against Kenny and B. A. O. Williams. Holdcroft writes: "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. In some respects, Holdcroft is more conservative
than Hare. Holdcroft considers:
(40) i. If you stand by Jane, don't look at her!
ii. You stand by Jane
__________________________________________
iii. Don't look at her!
This is valid. However, the following, obtained by anti-logism, is _not_:
(41) i. If you stand by Jane, don't look at her!
ii. Look at her!
_____________________________________________
iii. You don't stand by Jane.
"Honestly, it seems more reasonable", Holdcroft says, 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 Holdcroft's example:
(42) i. Varnish every piece of furniture you make!
ii. You are going to make a table.
_______________________________________________
iii. Varnish it!
This is, Holdcroft writes, "_prima facie_ valid". The following, however,
switching the order of the neustics in the premisses is not:
(43) 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": Holdcroft then considers A. Ross's famous example
(44) i. Post the letter!
_______________________________
ii. Post the letter or burn it!
as 'invalid' (Ross 1944:38 -- endorsed by B. Williams). Holdcroft starts by
quoting H. Kamp: "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". Holdcrot reviews 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.
Holdcroft quotes Hare's passage 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, he maintains 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"". As Holdcroft notes, 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':
(45) i. Put on your parachute & jump out!
____________________________________
ii. Jump out!
Holdcroft comments: "Someone who _only_ jumps out of an aeroplane does not
fulfil 'Put on your parachute & 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"
(cfr. _Language of Morals_, p. 20). "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"
(Holdcroft here refers to R. Edgley's treatment of this). Holdcroft
discussses Hare's test of cancellability in the case of the transport
officer who says:
(46) Go via Coldstream or Berwick!
-- analysed in first post of this thread. Holdcroft comments: "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
(47) _Prepare_ to go via Coldstream or Berwick!
As for Hare's application of Grice's cancellability thesis here, to yield,
in the circumstances:
(48) 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 thus
"reminiscent 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.
Grice writes considers Kant's example in GRUNDLEGUNG (the only fully
explicitly 'stated' technical imperative, Grice writes, to be found in Kant's
writings):
(1) It is _necessary_, given that one is to bisect a line on an unerring
principle that I'm to draw from its extremities 2 intersecting arcs.
Obviously some notion of hypothesis is understood here. Grice writes:
"Though Kant does not express himself very clearly [for a similar claim:
i.e. an Oxonian complaining of Kant's obscurity -- see Hare's remarks in
the interview with B. Magee] I am certain
that his claim is that this imperative [(1) above] is validated in virtue
of the fact that it is, ANALYTICALLY, a consequence of an INDICATIVE
statement which is TRUE, and, viz. the statement vouched for by geometry,
that:
(2) If one bisects a line on an unerring principle,
one does so _ONLY_ as a result of having
drawn from its extremities two intersecting arcs.
Grice's own example:
(3) To preserve a youthful complexion, one should
smear one's face with peanut butter before
retiring at night.
Grice distinguishes between "should" -- unqualified acceptability --,
"ought" -- ceteris paribus acceptability -- and "must" -- unyielding
subscription, as R. M. Hare would have it.
Curiously, it's only "must" which has
an analogue ("may") which allows a Deontic "Square of Opposition".
Grice smartly notes that the person who will accept (3) should better 'buy'
it provided it's based on something _true_ (not just 'nicey'):
"There is some initial plausibility in the idea that the _practical_
acceptability statement in (3) is satisfactory
iff the following ALETHIC [indicative. JLS] acceptability statement is
acceptable:
(4) It should be, given that
i. it is the case that one smears one's skin
with peanut butter before retiring.
and
ii. it is the case that one has a relatively
insensitive skin,
it is the case that one preserves a youthful complexion.
(Cf. D. Potter's _Pennies from Heaven_.
Episode between Gemma Craven and Nigel Havers. London: Faber).
Grice reconstructs, on Kant's behalf, the argument to prove the
acceptability in the geometry example as resting on the 'analyticity' of
(5) The agent who wills the end wills the means.
Grice writes:
"It seems to be to be very meritorious on Kant's part, _first_ that he saw
a need to justify hypohtetical imperatives of this [problematic] sort [what
elsewhere, Grice says, Kant refers to as 'technical imperatives' JLS] which
it is
ONLY TO EASY TO TAKE FOR GRANTED
and _second_ that he invoked the principle that (5)".
Grice proposes to remedy Kant's obscurity of style (and sloppy thinking and
incomplete premisses) in the Stanford lectures.
Grice's chain of reasoning is much better than Kant's (as anyone who is not
familiar with Kant will agree) and involves nice seven steps: Here they go.
Comments (easy ones) welcomed.
Step I:
It is a fundamental law of Human Psychology that, ceteris paribus, for any
rational creature -- call it R -, for any P and Q,
if R wills that P &
R judges that if P, P is a result of Q,
R wills Q.
Step II. Placing this law within the scope of a "willing" operator:
R wills for any P & Q,
if R wills that P
&
R judges that if P, P is a result of Q,
R wills Q.
Step III. Turning "will" to "should" [this is done via 'will' = 'shall'.
And 'shall'
= 'should']
If rational, R will have to block unsatisfactory (literally)
attitudes:
R should (qua rational) judge
for any P & Q,
if it's satisfactory to will that P
&
it's satisfactory to judge that if P, P as a result of Q,
Ergo it's sastisfactory to will that Q.
Step IV. Expliciting mode specifications:
A 'should' statement is transformed into an utterance in the imperative
mode (symbolised !p). Utterances in the indicative mode are symbolised by
Grice, a la Frege, as "|-p".
R should (qua rational) judge for any P & Q,
if it's satisfactory that !P
&
that if it |-P, |-P only as a result of Q,
it's satisfactory that !Q.
Step V. Via (p & q -> r) -> (p -> (q -> r)).
R should (qua rational) judge
for any P & Q,
if it's satisfactory that if |-P, |-P only because Q,
it's satisfactory that,
if let it be that P, let it be that Q.
Step VI.
R should (qua rational) judge
for any P & Q,
if P, P only because p _yields_
if let it be that P, let it be that Q.
Step VII.
For any P & Q if P, P only because Q _yields_
if let it be that P, let it be that Q.
------------------------------------------------------------------
To change your Lit-Ideas settings (subscribe/unsub, vacation on/off,
digest on/off), visit www.andreas.com/faq-lit-ideas.html
Other related posts: