Aristotle on Rational Action

ALEXANDER BROADIE

consideration of the many pages Aristotle devoted to a close

painstaking scrutiny of theoretical reasoning gives rise
and
to regret that he did not also write a Prior and Posteri'or
AJxalytics on practical reasoning. In sharp contrast to his work on
the theoretical syllogism, which is investigated systematically and
in great depth, his work on the so-called practical syllogism' is un-
systematic and unsustained. Yet this work, for all its fragmentariness,
has hardly, if at all, been overtaken by subsequent philosophers.
Certainly the huge strides taken by logicians in recent decades, in
the course of which they have tucked the Prior A nalytics into a
corner of the lower predicate calculus2, compare impressively with
the occasional shuffling that goes on within the field of practical
reasoning3.
One tempting explanation of this very uneven progress made
by logicians is that Aristotle did after all say all there was to be said
about practical reasoning, but left substantial gaps in his work on
theoretical reasoning. Yet this view prompts the question as to why
there is so very much to say about the one kind of reasoning and so
very little about the other.
Another explanation, which I think is closer to the mark, is that
the entities with which practical reason deals are at once more complex
and more elusive than are the entities which are the concern of
theoretical reasoning. One consequence of this would be to make
practical reasoning more difficult than its theoretical counterpart
to symbolise. This is important, because logic is severely restricted so
long as it remains unsymbolic. The purpose of this paper is to scrutinise
one particular obstacle to the symbolisation of practical reasoning,
in the light of Aristotle's own writings on the subject. At the end of

70
the scrutiny we shall at least be a little clearer about one of the
sticking points in a murky but central area of philosophy.

It might be supposed that in order to determine the structure of

practical reasoning it would be sufficient to consider the various
examples of it that are scattered about in Aristotle's works.4 But
it is a notorious fact that hardly two of those examples have the same
structure, and in certain cases the dissimilarities are more conspicuous
than the similarities. Furthermore, when Aristotle states in general
terms the structure of practical reasoning, which he does in several
places5, his general descriptions do not in fact cover all the examples
he offers that are apparently intended to fall under those .general
descriptions.
The wide differences in structure exhibited by the different examples
of practical reasoning can easily be shown. Perhaps the most complete
piece of reasoning offered by Aristotle is the following6 - let us call
it P.S.I: "I need a covering, a coat is a covering: I need a coat. What
I need I ought to make, I need a coat: I must make a coat
And the conclusion I must make a coat is an action [xon
io re The structure of the
premisses in P.S.I contrasts strongly with the structure of the premisses
in the following argument: P.S.II "I ought to create a good
yoi a house is a good: straightway I make a house
otxlcxv (De Motu 701 a 16-7).
Furthermore neither of the above arguments has the form of the
practical syllogism as that form is described by Aristotle. In De
Anima 434 a 15-20 he says of the practical syllogism: "The one
premiss or judgment is universal and the other deals with the partic-
ular (for the first tells us that such and such a kind of man should
do such and such a kind of act, and the second that this is an act of
the kind meant, and I am a person of the type indicated)". The
premisses of neither P.S.I nor P.S. II answer to this description.
Let us, however, allow that Aristotle recognised in practice if not in
theory that the premisses in practical reasoning can exhibit many
structures - just as he recognised in practice (but also in theory)
that the premisses in theoretical reasoning can exhibit many structures.

71
 Whatever the structure of the premisses, Aristotle's position seems
clear on the nature of the conclusion. It is apparently an action.
At the end of the argument P.S. I concerning the making of the coat
Aristotle asserts: "That the action is the conclusion is quite clear -
To
pLEvolv fi 7tpOCçf.C; (701 a 22-3).
. It is probable that Aristotle is also asserting that an action is the
conclusion of a practical syllogism when he writes, in a much disputed
passage: "... pLi« oc,U,76v,&vcX.yx1J To
¡..tÈv
(EN 1147 a 26-8). RackhaM7 translates, or rather glosses the passage
as: "When the two premisses are combined, just as in theoretic
reasoning the mind is compelled to a f f iym the resulting conclusion,
so in the case of practical premisses you are forced at once to do
it". There are however notorious problems about this passage, and
it will be necessary to revert to it later. But Aristotle's assertion in
De Motu "0'.Tt tl-Èv To seems sufficiently
univocal to entitle us, at least for the present, to assume that he did
indeed hold that the conclusion of a practical syllogism is an action.
If this assumption is correct then Aristotle's position is open to
an objection that he nowhere counters. Indeed, he nowhere seems
alive to the objection. The problem is this: what concept of validity
is invoked when it is said that a given action, or an action of a given
kind, is a valid conclusion from a set of premisses? In theoretical
reasoning a given conclusion C is validly drawn from a given set of
premisses P if the correct application of a given set of canons of
inference permits the inference of C from P. The canons of inference
are so designed that they ensure that if the premisses are true
then so also must be any conclusion drawn from those premisses by
the correct application of those canons.
Unfortunately this account of the validity of theoretical reasoning
cannot be applied en bloc to account for the validity of Aristotelian
practical reasoning, since there seems no relevant sense of "true"
and "false" according to which we can with justification describe
actions as true or false. It can be conceded that in ordinary parlance
actions are sometimes described as true, or false. But when so used
the terms have an essentially moral import, as they have when used
to describe someone as a true friend or as a false lover. If we use
"true" and "false" in this moral sense then an action might be a
valid conclusion from a set of premisses where the action is in the

72
moral sense false and the premisses are in the non-moral sense true.
But a valid inference cannot have true premisses and a false conclusion.
Consequently, the terms "true" and "false" will be used hereinafter in
a non-moral way.
. The inability of an action to take a truth value does not however
force us to an immediate rejection of Aristotle's claim that in a
practical syllogism an action can be the conclusion of a set of premisses.
One conspicuous development that has taken place in logic in recent
years has been the construction of systems of logic employing values
other than truth-values in their account of validity. This development
has been particularly marked in the field of imperative logic. There
is a good deal yet to be learned about imperatives and their logic,
but there is at least a wide-spread belief among logicians that systems
of imperative logic can be constructed even though imperatives
do not take a truth-valueg, but instead take an obedience-value9
or a satisfaction-value' or a satisfactoriness-valueli. That logicians
are not agreed about the precise kind of semantic value imperatives
take does not alter the fact that the concept of valid inference em-
ployed by formal logicians is no longer just the narrow concept
defined in terms of the conclusion being true if the premisses are
true, and a premiss being false if the conclusion is false. Consequently
the fact that the conclusion of a practical syllogism cannot take a
truth-value does not imply that such a syllogism cannot be valid,
and cannot even be well-formed. But if the relationship between
premisses and conclusion in the Aristotelian practical syllogism is not
truth functional, what is it?
A view that has some plausibility is that it is one of psychological
causation. This view has been defended by R. D. Milo in Aristotle
on Practical Knowledge and Weakness of Will (The Hague, 1966).
His position, which I have considered elsewherel2, is that an agent
who accepts the premisses of a practical syllogism is "psychologically
constrained" (ibid. p. 51) to embody those premisses in action. My
argument against Milo was that psychological constraint is a form of

73
compulsion. We are psychologically constrained to perform those
actions we cannot help performing, or cannot stop ourselves per-
forming. The trouble with position, I claimed, was that rational
action, understood as action that is explained in terms of processes
of reasoning involving means-ends relations, and principles of conduct,
is precisely action that is freely done in the sense that we can help
doing it, we can stop ourselves doing it.
I now think that this criticism of Milo will not do, for psychological
constraint need not be nearly so tough a form of pressure as I had
suggested; there are degrees of constraint. We often can help doing
what we are psychologically constrained to do, for though the
constraint forms an obstacle to action the obstructed agent may
well recognise that the obstacle is not insurmountable, and could
be surmounted if he made an effort that he is in fact willing to make.
Thus an action may be rational where the agent's recognition of the
validity of the premisses of his action constrains him to perform that
action.
What I wish to argue now is that Milo was, though perhaps not
in the way he had intended, correct. He held that the relation between
premisses and conclusion in the practical syllogism is one of psy-
chological causation. I wish to argue that the relation is indeed a
causal relation, though not the kind of causal relation that Milo
had in mind.
The clue to the nature of the relation is provided by consideration
of a possible parallel between theoretical and practical reasoning.
There is one place in particular, namely De ll?Z otu.701 a 8-14, where
Aristotle seems to suggest such a parallel. The passage occurs at the
point where Aristotle raises the question why thought is sometimes,
though not always, followed by action. He writes: "What happens
seems parallel to the case of thinking and inferring about the im-
movable objects of science. There the end is the truth seen (for,
when one conceives the two premisses, one at once conceives and
comprehends the conclusion - Yap 'raq 8vo 7tpo't'cX.cre:?ç; To
xai avv€.??xEV), but here the two premisses result in a
conclusion which is an action - lv«a6Ja 8' ex 8uo 7tpo't'cX.cre:Cù'J
To

There is a second passage, subject to dispute, that has been taken

to make much the same point, namely EN 1147 a 25-30. Donald
Allanl3 translates the passage as follows: "Further, the cause (of

74
<xxp<x<7Kx) may be viewed in this way. One judgment is universal,
the other bears upon particulars which belong to the sphere of per-
ception. When these judgments coalesce into one, it is necessary
that, just as elsewhere [sc. in the sphere of theory] the mind assents
to a conclusion, so here, when the premisses are practical, it should
at once act". Allan's scilicet "in the sphere of theory" follows Ross's
interpretation of the passagel4, and Rackham's15 and Dirlmeier'sls.
Gauthier and Jolif (p. 92) take much the same view of the passage:
"Lorsque les deux premisses se sont combinees pour engendrer une
proposition, 1'ame doit, dans le cas du raisonnement speculatif,
affirmer aussitot la conclusion, et, dans le cas de premisses pratiques,
aussit6t 1' accomplir" ..
Dr. A. Kennyl7 objects to Allan's account of the matter, on the
ground that "The reference to theory is quite irrelevant to the context"
(p. 177). Allan himself has some qualms about it since "it might suggest
that in demonstration as well as in action, there are minor premisses
which have a particular, perceptible subject; and this, we know,
is not Aristotle's doctrine" (p. 328). Kenny's own suggestion on the
matter, namely, that the contrast is not between theoretical and
practical reasoning but, rather, between practical syllogisms with
positive and with negative conclusions, has the merit of fitting the
context, since Aristotle follows the disputed passage with two examples
of practical syllogisms, the first with a positive, and the second with
a negative conclusion. Against Kenny, however, it can be maintained
that reference to theoretical reasoning is not in this context irrelevant.
Aristotle is here concerned to clarify the nature of practical reasoning,
and light is certainly shed on this question by the information that
practical reasoning parallels theoretical.
Kenny also has the above De Motu passage (701 a 13ff) in his
sights, on the grounds that "according to the same paragraph of
the same book the words are the conclusion of a practical
syllogism (701 a 20)" 182). This would give us some ground for
saying that Aristotle's claim that the conclusion of a practical syllogism
is an action can be interpreted as meaning that such a conclusion
merely states what is to be done, but for the fact that Aristotle does

75
say that the conclusion is an action I propose here to
take Aristotle's statement literally.
Thus far I have made the following claims about Aristotle's position,
first, that the relation between premisses and conclusion in the
theoretical syllogism is the same as that relation in the practical
syllogism, and secondly, that the conclusion of the practical syllogism
is an action. I wish now to argue that these two claims are not mutually
inconsistent. My argument is based on my account of what Aristotle
takes to be the relation holding between premisses and conclusion in
the theoretical syllogism. The following two passages in particular
in the Posterior Analytics yield up information on the nature of the
relation: "We only have knowledge of a thing when we know its
cause. There are four kinds of cause: the essence, the necessitating
conditions, the efficient cause which started the process, and the
final cause. All these are exhibited through the middle term" (94 a
20-4); "Demonstrative knowledge must proceed from premisses
which are ... causative of the conclusion ... They must be causative...
because we only have knowledge of a thing when we know its cause"
(71 b 20-35).
What the premisses may give us information about is the formal
cause of the conclusion. In that case it may seem that we are entitled
to regard the relation of premisses to conclusion as the relation of
form to matter. But the thesis that the conclusion is the matter of
the premisses clashes with an important statement in the Physics
195 a 15-9. There we are told that "The premisses are [the causes of]
the conclusion, in the sense of 'that from which"'. We are immediately
told (19-20) that the premisses are the substratum and the conclusion
the essence. Thus the doctrine of the Physics is that the relation of
premisses to conclusion is that of matter to form.
This difficulty can be resolved by making a distinction between
on the one hand the premisses and conclusion, and on the other what
the premisses and conclusion are about. Since the conclusion consists
of subject and predicate terms, whose copulation is, within the con-
clusion, unexplained, and since the premisses provide the explanatory
principle, namely, the middle term, the conclusion must be regarded
as the matter that is structured by the premisses, and in particular
by the middle term. Knowledge of the conclusion is secured by a
consideration of what it is that structures it. Since what takes the

76
structure is the matter, the conclusion must be the matter of the
premisses.
On the other hand if we have regard to what the judgments in
the syllogism are about we get a different picture, and the one
Aristotle presumably had in mind in the Physics. Consider "If all
animals are mortal and all men are animals then all men are mortal".
Aristotle seems to conceive of the mortality of animals and the
animality of men as the two building blocks out of which the mortality
of men is constructed. The two blocks, considered not as judgments
but as elements in the world, are the "that from which" the mortality
of men is built. In this sense the premisses can be considered as the
matter of the conclusion. According to this view the conclusion
follows from the premisses in the sense that it is caused by them,
where the causal relationship is understood to be material.
From the fact that the relationship in question is classifiable as
one of material causality it follows that more than one conclusion
can be validly inferred from a given set of premisses, since although
a material cause (unless it is prime matter) is structured to some
degree, it is not so structured as to permit actualisation in only one
kind of way. As a non-prime material cause it is limited as to what
it can become, but it can still become any one of several things.
This point will become important when we apply our findings to the
practical syllogism.
Thus far it has been argued that from one point of view, what may
be called the ontological point of view, the premisses in a demonstra-
tion are the matter of the conclusion, while from another viewpoint,
the epistemological, the conclusion is the matter of the premisses.
(This point, it is worth noting, is one aspect of the Aristotelian doctrine
that the ratio essendi and the ratio cognoscendi move, in general, in
opposite directions).
Let us now consider whether these general remarks about the
demonstrative syllogism hold good when applied to the practical
syllogism. What we have shown is that the relation between premisses
and conclusion admits of a metaphysical interpretation, an inter-
pretation in terms of the relation between matter and what structures
that matter. This interpretation enables us to bypass the problem
raised near the start of this paper concerning how we are to cope with
the fact that actions, which constitute the conclusions of practical
syllogisms, cannot take a truth-value. For the line now available
to us is that the practical syllogism is amenable to a metaphysical

77
interpretation, according to which the relation between premisses
and conclusion can be understood as the relation between a material
cause and its effect. To say that the conclusion follows from the prem-
isses is to say that the action imposes a structure on the principles
and beliefs about means ends relations and about the situation at
hand as this is seen by the agent. In imposing a structure on this
material it unifies it. And of course this is exactly what the action
does do. So far as an action is performed as the outcome of a process
of practical reasoning it constitutes the embodiment of the elements
in that reasoning; the agent's principles, desires and beliefs are all
drawn together, given a unity by being expressed in the single action.
To use the earlier metaphor, the agent's principles, desires and beliefs
are the building blocks out of which the action is constructed.
Now, from a particular set of principles, desires and beliefs more
than one action can rationally follow - the agent may not be tied
down to a single action even if he is tied down to a single set of
principles. The factors he provides from within himself that will
determine him to act may leave him with considerable leeway as to
how he gives expression to those principles, desires and beliefs.
This does not mark a distinction however between the practical
syllogism and demonstration, for, as we saw, the premisses of a de-
monstrative syllogism need not tie one down to only one conclusion.
Thus from "All 111 is P and all S is ¡via" both "All S is P" and "Some
S is P" can be concluded. Likewise from "People in distress ought to
be aided by those able to help. That man is in distress and I am able
to help him" several possible actions could follow rationally. There
may, after all, be many ways of giving aid to a person in a specific
kind of distress. From one point of view, of course, all the alternative
actions are really the same action, since in so far as they are all
constructed out of the same material they can all be given the same
description. They can, that is, all be described in terms of the one
piece of matter of which each of those actions would be the material
effect. Each, for example, would be a case of "helping someone in
distress".
Even though, from the point of view just considered, the premisses,
of the practical syllogism are the matter of the concluding action,
from another point of view the action is the matter that is structured
by the premisses. In An. Post. 71 b 35 we are told that "we only have
knowledge of a thing when we know its cause". This is true of action
as it is of scientific phenomena. We know an action only when we

78
know its cause, when we know the agent's motives and intentions
and beliefs about his situation. Such knowledge not only affords
us insight into the action but even gives us warrant for regarding
the action as an action at all, that is, as an expression of human
agency rather than merely a set of physical movements. This knowl-
edge provides us with the form of the action, the matter being the
physical movements structured by that form.
The final cause of the action, which is also expressed in the prem-
isses, is the end envisaged by the agent, for the sake of which the
action was performed. But there is a problem as to whether also the
efficient cause of the action is expressed in the premisses. It might
seem that if we act as a result of a piece of deliberation, the delibera-
tion is the efficient cause of action. Yet Aristotle is explicit on the
matter: "The origin of action - its efficient, not its final
cause - is choice, and that of choice is desire and reasoning with a
view to an end" (EN l139 a 32ff). Choice itself is not, seemingly,
part of the practical syllogism, and consequently the efficient cause
of action does not figure in the practical syllogism.
Aristotle's position is not entirely free of difficulty, for the sentence
just quoted implies that all actions are chosen, though we know that
elsewhere he denies this. We learn in EN 1111 b 5 f f that the actions
of children and animals, though voluntary, are not chosen. It is
clear why Aristotle wants to restrict the field of chosen action in this
way. Choice, TrpofXLpEo?, is rational desire or desiderative reason.
Hence those incapable of reasoning are incapable of choosing. Animals
and children (or at least infants) cannot choose because they cannot
reason. However, in telling us that the efficient cause of action is
choice, Aristotle can reasonably be taken to be saying that in the case
of actions that are chosen (not 'actions simpliciter') their efficient
cause is choice.
When Aristotle states that the efficient cause of choice is desire
and reasoning with a view to an end (EN 1139 a 33) he clearly has
in mind as the efficient cause the premisses of the deliberative process,
for it is on the basis of these premisses, which express the reasoning
prompted by desire, that the choice is made. Thus choice can be seen
as the efficient effect of the premisses as well as the efficient cause of
the conclusion. The full practical syllogism is a chain of efficient
causes with choice acting as the link. The premisses, as the efficient
cause of the efficient cause of the conclusion, are the efficient cause of
the concluding action.

79
 Thus an action is the conclusion in a practical syllogism from a set
of premisses in the sense that those premisses are the formal, efficient,
final and material causes of the conclusion. Milo was thus correct
in holding that the relation of premisses to conclusion in the Aristote-
lian practical syllogism is a causal relation. But what needs to be added
is that implicit in Aristotle's account of the practical syllogism is a
full-blown metaphysics of action, involving his theory of four-fold
causation. Milo's explanation of the relation in terms of psychological
constraint clearly leaves out of account a good deal that is essential
to Aristotle's position.
What has emerged from the foregoing analysis is that the relation
of premisses to conclusion in the practical syllogism can be seen,
not as a logical but as a metaphysical relation. This suggests that
advances in the theory of practical reasoning must be attendant on
advances in the philosophy of action, and perhaps especially in what
I have termed the metaphysics of action. Till then, symbolic logicians
waiting for employment in this field will have little useful to do.

University of Glasgow

80