Contents.
Introduction.
PART I: EXPLAINING ACTION.
1.1
Explanation.
1.2
Explaining Action.
1.3
Explanatory and Justifying Reasons.
1.4
Actions and Events.
1.5
Actions and Mistakes.
1.6
Intentions.
1.7
Explanation and Prediction.
1.8
Davidson on Reasons for Actions and Practical Reasoning.
1.9
Davidson on Desire.
1.10
Desire-states as Phenomenal States?
1.11
Direction of Fit and the Teleological Argument.
1.12
Shortcomings of Definitions by Direction of Fit?
1.13
Desires: Pro or Proper?
1.14
The Nature of Desire and Belief.
PART II: PRACTICAL REASONING.
2.1
Practical Reason and Introduction to the Dispute Between Humeans and
Rationalists.
2.2
Practical Reasoning.
2.3
The Conclusion of Practical Reasoning.
2.4
The Premises of Practical Reasoning.
2.5
Practical and Theoretical Reasoning.
2.6
Reasons for Actions.
PART III: RATIONALISM VS. HUMEANISM.
3.1
The Dispute Between Humeans and Rationalists.
3.2
Nagel's Position.
3.3
Motivated and unmotivated desires.
3.4
Schueler on Nagel's Distinction.
3.5
Nagel's strategy and the Genealogy of Desires.
3.6
Motivated Reasoning?
3.7
Reasons for Desires.
3.8
Critical Remarks.
3.9
Concluding Remarks.
Biblographical
References.
The main conclusion of this essay is that since
desires dispose us to act, desires are what motivate us to act. This view is
reflected in the title which alludes to the triple meaning of 'entertain': to
amuse and interest, to provide for, and, to think about or reflect on; a trio
that is further mirrored in the way I have divided my work into parts. In Part
I: Explaining Action, I investigate the concept of 'desire' as it is used to
explain actions by being a reason for which one acts; i.e., how desires provide
for actions. As a platform for this, I start by providing an account of
explanation in general before turning to the explanation of action by way of
reasons. This leads me to consider reasons, actions and events. Further, I have
a look at intentions and predictions before I again turn to reasons for action
and practical reasoning. Then I start considering desires as such, first Davidson's
conception of desires, before I turn to the question whether desire-states are
necessarily phenomenal states - whether they really are amusing - and to a well
known argument to prove that desires must always figure as the motivating
element for intentional action. The first part ends in two chapters, one on a
distinction that has been made between desires proper and pro attitudes, and
the other on the nature of desires and beliefs as I see it.
In Part II: Practical Reasoning, I
investigate the concept of 'desire' in relation to practical reasoning, how
desires function in our reflection on what to do. This part starts in a
brief introduction to the subject of practical reason and the ever raging
battle between Humeans and rationalists. Next, an overall conception of
practical reasoning is presented, as well as it's constituents: premises and
conclusion; before it's relation to action-explanation and theoretical
reasoning is considered.
In Part III: Rationalism and Humeanism, I
investigate how the concept of 'desire' is necessary to account for the interest
we have in that which motivates us to act. I start by complicating the debate
between Humeans and rationalists, before I turn to exposing Nagel's rationalist
conception. An important distinction introduced by Nagel, that between
motivated and unmotivated desires, is considered, before I turn to the
application of this on the present subject. This leads me to an account of
motivated desires that is Humean, a view I make some last critical remarks
upon. I end this third part and the paper with some concluding remarks.
When we explain our actions we give the reason why we
acted: 'For what reason did you call me?'; 'I called you because I wanted to
know whether or not you were home.' The actions for which we have reasons are
what we call intentional actions, actions done with some intention. Some
actions done with an intention, the action done is not the one we intended
done: 'For what reason did you call me?'; 'I am sorry, I wanted to call Peter
(not you) to find out whether he is home.' Other intentional actions are done
for no further reason: 'For what reason are you chewing gum?'; 'No reason; I
just feel like it.' Still other actions are done for no reason whatsoever, like
things that just happens: 'For what reason did you cut your hand?'; 'No reason
at all (you idiot!); I accidently nicked it while peeling potatos."[1]
We do not only attribute reasons to
ourselves when we act, we explain other's behaviour with the same means. We
typically rationalize the behaviour of other's and ourselves, and sometimes we
do it against what clear evidence there might be for the contrary. This sort of
mistakes set light on one important feature of action-explanation, a feature
that stands out more clearly here than in explanation in general - even though
we must accept certain context-dependency in all explanation, also in science
like van Fraassen (1980) has argued - its normativity. We attribute reasons to
people, we even attribute the ground for assigning reasons in that we assign
intentionality to the action at hand. Both explanandum and explanans are mainly
based on normative evidence rather than empirical evidence. The
philosophy of action or action theory is the study of this normativity; the
structure of human action, the process by which it originates, and the ways in
which it is explained. It provides us with an account of how we assign reasons
to agents, why we assign those reasons, etc.
Seeing someone act and wondering why, many answers may
be relevant per se. To the question why someone did visit his
grandmother at the hospital, a fitting, statistic, answer could be that humans
at all times have been closely connected to their relatives, both for comfort
and economical reasons, so the connection between relatives have had important
social consequences. Nevertheless, this was perhaps not what we wanted, rather
we wanted the personal story behind his visit. But still, not any story.
We must hold that to be an explanation of human action, as we must hold in
scientific explanations, our account must tell a story of how things did happen
and how the events hang together.
We may describe what we do and what
happens to us and their effects as a net of events. When we explain action we
pick certain events out to explain, and in doing this we realize that not any
way of drawing attention to a certain event residing in this net will explain,
anymore than any event will explain what we want to explain, even though it
certainly precedes it. We may speak of salient features that serve our
questions. The question is thus by what light do we find what we may refer to
as 'the reason why he did it' or 'the real reason why he did it'. This salient
feature we may call with van Fraassen the 'relevance relation', a relation that
is context dependent. Consider the following example:
Suppose a father asks his teenage son, 'Why is the porch light on?' and the son replies 'The porch swich is closed and the electricity is reaching the bulb through that switch.' At this point you are most likely to think that the sort of answer the father needed was something like: 'Because we are expecting company.' But it is easy to imagine a less likely question context: the father and the son are re-wiring the house and the father, unexspectedly seeing the porch light on, fears that he has caused a short circuit that bypasses the porch light switch. In the second case, he is not interested in the human expectations or desires that led to the depressing of the switch.[2]
One approach to explanation in general can be found by
focusing on why-questions. Consider the following question (1) 'Why did you
flip the switch?' The topic of this question, and what the answer to it
must relate to, is the proposition 'flipping of the switch'. The answer is also
dependent upon a contextual assumption that van Fraassen calls the contrast-class.
The requirement of a contrast-class is seen if we interpret the above question
in for instance these different ways:
(1a) Why was it you that flipped the switch
(and not your brother)?
(1b) Why was it the switch you flipped (and not
the other device vired to the lamp)?
(1c) Why did you flip the switch (instead of
leaving it in the position it was)?
The difference between these interpretations is that
they point to different contrasting alternatives that are indicated in the
paranthesis above. To (1a) and also (1b) the answer 'because I wanted to turn
the light on' is not a good answer, while to (1c) it is fine. On the other
hand, 'because my brother does not reach the switch' is inapropriate to (1b)
and (1c), but good to (1a). 'Because I have a liking for switches over
treadmills' is inapropriate to both (1a) and (1c) but apropriate to (1b). Such
considerations leads van Fraassen to conclude that the general, underlying
structure of why-questions must be: Why (is it the case that) P in
contrast to (other members of) X?, where P is a statement (the
topic) and where X, the contrast class, is a set of alternatives to
which P may belong or not. Very often the contrast class is tacitly
assumed and generally understood from the context in which the question
is uttered. Further, an answer to a question of the form indicated above must
adduce information that favours P in contrast to other members of X,
a source of information that is wholly dependent, in some cases, upon the coice
of contrast-class X. This have among its further consequences for the
understanding of the nature of explanations that we see that no event is
explanable qua particular event, only qua event of a certain kind
(qua (intentional) action, qua spastic reaction, qua
collission between magnitudes).
Above we saw the importance of context as
determining what features being salient, that is, the context determines
relevance of explanans to explanandum (events leading up to the flipping
(desires, beliefs, intentions), the standing conditions that made the flipping
possible (availability of switch, etc.), or what function the flipping has,
etc.), and discussing why-questions we were led to recognize a further
contextually determined factor, namely that of a contrast-class. The range of
hypotheses about the event which the explanation must weed out is not
determined solely by the interests of the discussants, but also by the range of
contrasting alternatives to the event.
"An explanation is an answer to a why
question," says van Fraassen.[3]
An explanation is essentially relative, like being a daughter is relative:
every woman is a daughter, and every daughter is a woman, yet being a daughter
is not the same as being a woman.
Thus
we have it that the why-question Q: 'Why (is it the case that) Pk?',
in a given context K, will be determined by three factors: the topic Pk,
the context-class X={P1, . . ., Pk,
. . .}, and the relevance relation R. Q is given as Q=<Pk,
X, R>.
A
proposition S is called relevant to Q exactly if S
bears relation R to the couple <Pk, X>.
The direct answer Sd to Q is given in the form:
Pk in contrast to (the rest of) X because Sd.
What is claimed in the answer 'because Sd' is the following:
(1) that Pk is true; (2) the other members of the
contrast-class X are not true; (3) Sd is true; and (4)
Sd is a reason.
Following
van Fraassen, the word 'because' signifies only that Sd is
relevant in this context, to this question, and hence the claim is merely that Sd
bears relation R to <Pk, X>. For example,
suppose you ask why I got up at seven o'clock this morning, and I say 'because
I was awakened by the sound of someone practising the violin'. In that case I
have interpreted your question as asking for a sort of reason that at least
includes events-leading-up-to my getting out of bed, and my word 'because'
indicates that the sound of the violin was that sort of reason. But I could
have construed your request as being specifically for a motive. In that case I
would have perhaps answered 'no reason really, I could easily have stayed in
bed, but . . .' and so on. In this case I do not say 'because', for the sound
of the violin does not belong to the relevant range of events, as I understand
your question.
A why-question
presupposes the following[4],
that: (a) its topic Pk is true; (b) in its context-class X,
only its topic Pk is true; (c) at least one of the
propositions S that bears its relevance relation R to its topic Pk
and contrast-class X is also true.
As
van Fraasen himself notes, even if all three of these presuppositions are true,
the question may still not have a telling answer, that is, the theory so far
does not tell us if the answer is good or better than any other. Suppose we are
in a context with background k of accepted theory plus information, and
the question Q arises here. Let Q have topic B, and a
contrast-class X={B, C, . . ., N}. How good is the
answer 'because S'? The objection could run like this: if 'because S'
does not only indicate that S is a reason, but that it is the
reason, or at least a good reason, how can we determine this? Here van
Fraassen takes the weakest of two relevance relations R, and says that
"the point is rather that anyone who answers a question is in some sense
tacitly claiming to be giving a good answer."[5]
But in either case, the determination of whether the answer is good or telling,
or better than other answers that might have been given must still be carried
out, that is, we need some criteria of evaluation to follow.
Van Fraassen proposes three ways in which
the answer S (the explanans) is evaluated:
1.
S is evaluated as acceptable or as likely to be true.
2.
S is evaluated in regards to what extent it favours the
topic B against other members of X.
3.
S is evaluated in regard to how 'because S'
compares with other possible answers to the same question. This comparison has
three respects: (a) is S more probable (in view of K); (b)
whether S favours the topic to a greater extent; (c) whether S is
made wholly or partially irrelevant by other answers that could be given, that
is, whether S gets screened off by another answer.
It must be noted that the concept of 'truth' have a
meaning in van Fraassen's work that it does not have when considering
action-explanation. He presupposes a distinction between theory and facts that
is not proper to introduce in the context of action-explanation because of the
normative aspect of the latter. Nevertheless, his account serves well, I think,
the purpose of introducing explanation in general and to point to its
context-dependent character. Further, the relevance of introducing why-question
as an approach to explanation is definetly not turned down by shifting focus to
action-explanation. 'Why did s/he do that?' is a frequently asked question and
one that is motivated by our desire to understand each other. Such questions
are perhaps the clearest expression of our rationalizing behaviour. Now, let us
turn to action-explanation and some important basic topics in that respect.
It is hard not to agree with Davidson (1987) when he
says that "there is an irreducible difference between psychological
explanations that involve the propositional attitudes and explanations in
sciences like physics and physiology."[6]
This feature has suggested for many that action-explanation is teleological -
determined by some goal of the agent (Smith (1987; 1994), Davidson (1975,
1984)). This appeal to intentional concepts is not unproblematic, because the
concepts used in the appeal are themselves problematic. Davidson notes this by
saying that,
if we were to ask for evidence that the explanation is correct, this evidence would in the end consist of more data concerning the sort of event being explained, namely further behaviour which is explained by the postulated beliefs and desires. Adverting to beliefs and desires to explain action is therefore a way of fitting an action into a pattern of behaviour made coherent by the theory. This does not mean, of course, that beliefs are nothing but patterns of behaviour, or that the relevant patterns can be defined without using the concepts of belief and desire. Nevertheless, there is a clear sense in which attributions of belief and desire, and hence teleological explanations of belief and desire, are supervenient on behaviour more broadly described.[7]
Action-explanations appeals to the concept of
'reason'. The belief and desire that explain an action, says Davidson,
"must be such that anyone who had that belief and desire would have a
reason to act in that way."[8]
We must understand this not as saying that the reasons for which agents act
must be good or moral in any sense, but rather as making the deep but obvious
point that the action to be explained must be reasonable in the light of
the assigned desires and beliefs, that rationalize the action; and also that
the assigned desires and beliefs must fit together. The cogency of
action-explanations rests on its ability to discover a coherent pattern in the
behaviour of an agent. To this we may add that the descriptions we
provide of desire and belief must exhibit the rationality of the action in the
light of the content of the belief and the object of the desire.
One
could remark that the methodological presumption of rationality may seem to
make it impossible to attribute irrational thoughts and actions to an agent.
Davidson denies this, but agree that it does impose a burden on such
attributions.
We weaken the intelligibility of attributions of thoughts of any kind to the extent that we fail to uncover a consistent pattern of beliefs and, finally, of actions, for it is only against a background of such a pattern that we can identify thoughts. If we see a man pulling on both ends of a piece of string, we may decide he is fighting himself, that he wants to move the string in incompatible directions. Such an explanation would require elaborate backing. No problem arises if the explanation is that he wants to break the string.[9]
I generally think there is no reason to reduce
intentional concepts and thereby intentional explanation to any other set of
concepts or explanation as long as they serve the aim they are set to serve,
something I think they do. Nevertheless, definitions of intentional concepts
and action-explanation is unavoidable (non-viciously) circular. Some
philosophers (for instance Stich (1983) and Churchland (1995)) have tried to
reduce mental or intentional concepts that explanation of action trades in
(neuro-) physiological concepts because they don't think intentional concepts
serve as well as I think. I deny this reductivism. Nevertheless I agree with
Davidson and others (like Dennett) that this denial of reduction does not force
us to conclude that mental concepts are themselves ontological real concepts
(referring to ontological entities) rather than conceptual ones, trading in
some conceptual realm. This means that mental objects and events (the referents
of mental concepts) such as beliefs, desires and intentions are at the same
time physical, biological and chemical objects and events, but that there is no
reduction or strict laws available for a description of the relation between
these objects and events. The reason for this, and the reason for the
(non-vicious) circular nature of action-explanation, is that mental concepts
are normative: they must be identified by their semantic contents in
reason-explaining practice. To say of an event like an intentional action that
it is mental is to say that we can describe it in a certain vocabulary, a
vocabulary whose mark is semantic intentionality. Thus, reason-explanations
differ from other explanations because "they are coached (in part) in an
intentional vocabulary".[10]
The semantic contents of attitudes and beliefs determine their relations to one another and to the world in ways that meet at least rough standards of consistency and correctness. Unless such standards are met to an adequate degree, nothing can count as being a belief, a pro-attitude, or an intention. But these standards are norms – our norms – there being no others.[11]
This is a very important point, and one that in one
sense seem to make the whole effort of trying to define conditions for
ascription of intentional events so difficult – there is no independent
standard to which we can appeal. In another sense it is just this fact that
makes it so easy, and (non-viciously) circular. The important point is not that
we as observers and explainers employ our norms in understanding the action of
others, since we do this whenever and whatever we study (something that is
evident from the account of explanation from van Fraassen). What is important
to see, according to Davidson, is that,
in explaining action we are identifying the phenomena to be explained, and the phenomena that do the explaining, as directly answering to our own norms; reason-explanations make other intelligible to us only to the extent that we can recognize something like our own reasoning powers at work. […] It is a sentral and irreplacable feature of the intentional.[12]
Our answers then to questions about reasons for action
are evaluated as acceptable and likely to be true.
Dennett
(1981, 1987) notes that we seem marvelously rational, but that however rational
we are "it is the myth of our rational agenthood that structures and
organizes our attributions of belief and desire to others and that regulates
our own deliberations and investigations. We aspire to rationality, and without
the myth of our rationality the concepts of belief and desire would be
uprooted."[13] The word
'myth' suggests there to be facts contrasting with it, and as Dennett notes,
there are, but not "functionally individuated […] the way belief/desire
psychology carves things up."[14]
He confirms the account presented above:
Folk psychology […] is idealized in that it produces its predictions and explanations by calculating in a normative system; it predicts what we will believe, desire, and do, by determining what we ought to believe, desire, and do.[15]
1.3 Explanatory and
Justifying Reasons.
Acting for a reason
is closely related to (a) acting intentionally (the concepts are interdefined),
(b) acting rationally, and (c) acting on the basis of practical reasoning. Any
account of acting for a reason needs to consider and explain how these elements
connect. This is what we shall do eventually. Now, it is time to consider what
it means to act for a reason. We often refer to people's desires, beliefs,
fears, feelings, etc. as reasons for which they act. Such states express
reasons we may have for acting. They may be classified into a conative
(motivating) element and a cognitive (guiding) element, though, so that we may
(simply) define a reason for which one acts as composed of these two elements.
We shall elaborate this conception later on. Now, the concept of reason is not
limited to this explanatory role; there are many kinds of reasons. To
get clear what we mean, we must consider these kinds.
According to
Schueler (1995), ordinary speech allows for at least four locutions involving
the term 'reason':
(1) The reason (that) q
was that p. 'The reason that Ralph left early was that he remembered he
had another appointment.'
(2) A's reason for j-ing was that p,
or A j-ed for the reason
that p. 'Ralph's reason for leaving early was that he remembered he had
another appointment.'
(3) There is (a) reason
for A to j. 'There is a
reason for Ralph to leave early.'
(4) A has (a) reason to j. 'Ralph has a
reason to leave early.'
It is evident from
the form of (1) that 'reason' has a quite distinct use here from the other
three. Instead of an example with an action, we could give the following
example: 'The reason the tree died was that it became infested with peach
bores.' It would make no sense to speak of 'the tree's reason to die' (2), or
'there is a reason for the tree to die' (3), or 'the tree has a reason to die'
(4). So (1) seems to be a clear explanatory use of 'reason'. The phrase
'the reason (that) q …' applies to any explanatory factor that explains p,
and have thus the same use as 'because'. This can in fact be used as a test for
whether the sentence before us is explanatory. If we apply the test on (2), we
see that this use also is explanatory, though differing from (1) in that it is
restricted to agents and what they do.[16]
Clearly (3) differs from (1) as already mentioned, because this use cannot be
applied on other than actions, but it also differs from (2) in that it makes
perfectly sense to put terms like 'good', 'strong', 'weak', 'sound' in front of
'reason' as a qualifier in (3) but not in (2). In addition, (2) carries with it
the assumption that A does indeed j, something that (3) does not.[17]
It makes no sense to say that 'Ralph's reason for leaving early was that he
remembered he had another appointment, but he did not leave early.' It makes
sense, however, to say that 'There was a reason for Ralph to leave early - he
remembered he had another appoinment - but he did not leave early.' Parallel
considerations argue for distinguishing (4) from (1) and (2); in fact in most
cases (3) and (4) can be seen as stylistic variants of each other. All this
means, since (1) is so obviously different from the other three, that the real
issue is to recognise that (2) is distinguishable from (3) and (4). While the
former locution allows of both an explanatory and a justifying role, the latter
two allow only a justifying role.
Audi (1993) gives
an account that bring forth other sides of the nature of reasons.
(5) A j's for a reason R,
or R is a reason for which A acts. His j-ing is motivated.
If A j's for R, he
has a desire for something; it is personal. A reason R for which one
acts can thus be expressed by an infinite clause giving the content of the
relevant desire. Let us say that a reason R for which A j's is, in part, a
state of affairs p which (i) expresses something he desires, and (ii) is
connected with his j-ing through an
appropriate belief, for example that his j-ing will achieve p.
(6) R is a reason why
A j's. Implied by (5),
one that at least partly explains his j-ing. But it need not be the
one for which he does so, since the action could simply be due to drugs or some
mad scientist controlling his actions.
(7) Reasons to j. Normative and
impersonal. There can be reasons to keep one's promises, even if nobody wants
to keep promises.
(8) Reasons for A
to j. Normative and
personal. That j-ing would fulfill A's
desire to visit Australia might be a reason for him to j, but not mine or
yours. Need not be potentially motivating since if he just believes j-ing will fulfill
his desire, he has a reason for j-ing.
(9) A reason R A
has for j-ing. Personal but
need not be normative, and is potentially motivating. It is the expression of a
desire by the agent, but need not lead to action.
All five kinds of
reasons can be involved at the same time. When A j's for a good
reason, e.g., writes a recommendation to keep a promise: (5) keeping his
promise is a reason for which he writes; (6) it is a reason why
he writes; (7) it is a reason to write; (8) it is a reason for him
to write; and (9) it is a reason he has to do so.
One distinction
expressed in both Schueler's and Audi's accounts that has gained wide recognition
and discussion is the one between explanatory (or motivating) reasons
and justifying (or normative) reasons. Some, like Smith (1987; 1994),
Pettit & Smith (1990), and Schueler (1995), have put much weight on this
distinction, too much in my opinion. One must be careful not to think that
reasons that explain actions do no justifying job. The
distinction is one of function or role, so the reasons that are justifying may
be explanatory, and the other way around. Also, as Schueler's list indicates,
justification ranges over good and bad reasons, something that
the explanatory role don't take into concern. I agree with Davidson when he
says that "the justifying role of a reason […] depends upon the
explanatory role, but the converse does not hold."[18]
Every explanatory reason for an action will justify in a "somewhat
anaemic" sense: "from the agent's point of view there was, when he
acted, something to be said for the action."[19]
By this he means that unless a reason explains some action, it can do no
justifying job. A reason cannot justify my action unless I did it for that
reason. This is a narrow understanding of reasons, however, that is only
connected to explanation. In a broader understanding of reasons, reasons can
give grounds for acting on them (it is a good reason for doing it) without the
reason ever being acted on (like Audi indicates). It is perhaps wrong to use
the term 'justifying' reason about this broad role of reasons, but it is being
used. The broad sense of reasons concerns reasons role in deliberation; when
considering reasons for or against some action, we consider reasons in a
justifying sense: whether it will justify (good or bad) our conduct.[20]
The only question about justifying
reasons interesting for the explanation of action is whether we always need a
further desire for some possible (external) reason, perhaps given to us by the
law, etc., for this reason to be a (explanatory) reason for which we act. This
does not mean that we may distinguish justifying and explanatory explanations
of actions. Schueler (1995) seems at times to think the latter, in that he
claims the two roles of reason in explanation of action to use not only two
different senses of reason, but also of desires. But how can a question of
desires be interesting for justifying reasons unless it is connected to the
reasons for which some agent acts? The important question is how, and if so, whether
an external, moral perhaps, reason may come to have a motivational role in the
explanation of action. We shall return to this later, but for now we shall be
concerned with reasons for which one acts - in the explanatory sense,
and where the narrow sense of justification applies. Let us first consider the
relation between actions and events.
In a famous passage
Wittgenstein (1953a) asks "was ist das, was übrigbleibt, wenn ich von der
Tatsache, daß ich meinen Arm hebe, die abziehe, daß mein Arm sich hebt?"[21]
I agree prima facie with Davidson (1987) when he answers 'nothing'. What
should this 'übrigbleibt' be, and why do we think there should be anything
'übrigbleibt'? But unless there is something left, how are we to distinguish
between intentional action and things just happening to us? What is the
difference between the sentence 'My arm goes up' and 'I raise my arm'? The
claim is that nothing needs to be added to the event of my arm going up to make
it a case of my raising my arm. Of course we have cases where my arm goes up without
me raising it;[22] so what is
meant in the above claim is not to deny that no further conditions have
to be satisfied to ensure that the rasing of my arm is a particular case of my
raising my arm. This is an addition to the description of the event, says
Davidson, and not to the event itself. This means that of all the events that
is a raising of my arm, some is rightly described as cases of my raising
my arm, but none of these cases are events that include more than my arm going
up. The event stands alone, so to say. That an event is an event whether it is
something that is done for some reason or for no reason at all means that great
stress is laid upon the question of what caused (or produced) the event;
its conditions. The distinction between what is done and what happens to us is
not weakened because we have freed the event itself from this burden of
evidence, since nothing is added to the event itself that makes it into an
action; nor is anything substracted to make it something just happening to the
agent. The evidence that it was an act of my raising my arm and no happening of
arm-rising must be on what caused the raising of the arm.[23]
So we are still free to distinguish between actions and mere happenings.
Yet, the answer that seems most
natural to the question of the difference between my arm going up and my
raising it is that I, the agent, made my arm go up, and that there were no such
making in the case where the arm merely went up. But this way of putting the
matter indicates the presence of two events, rather than one; that what
I did was not identical with the arm going up. On this alternative account, the
effect is the arm going up and the cause is what made it go up. So the 'making'
here must be the 'übrigbleibt'. Several suggestions have been given along these
lines; that some will is active, that acting is trying to perform, etc.
(vide for instance Hornsby (1980)). The obvious difficulty with any such
suggestion is that if some prior act of an agent is needed to make him act,
then, since this prior act is something the agent does, another prior-prior act
is needed for this act again, and so on, ad infinitum. Davidson
dismisses any such suggestions by noting that "it must in general be a
mistake to suppose that whenever an event is caused there must be something
called a causing."[24]
This initially cryptical remark gets its natural explanation in what follows.
Think of the following example (from
Davidson (1987)): the dropping of the egg caused it to break. Here we have two
events, the one causing the other: the dropping of the egg causing it to break.
If we compare this to the above alternative account, however, this would not be
enough, since we would in addition need a third event which is the causing
of the second event by the first. A causing qua causing is no event,
since a causing is just some special relation between two events. If such a
third event were required to relate the original cause and effect, two more
events were required to relate the original cause and effect with the causing.
For Davidson the difficulty which give rise to such alternative accounts is
only grammatical.
In the
sentence 'Smith kicked Jones,' the verb conceals reference to an event: the
logical form of such a sentence is made more clearly manifest, in my opinion,
by something like 'There was a kicking of which Smith was the agent and Jones
the victim.' This makes us think 'caused', as it appears between phrases
referring to events, must also conceal reference to still another event. But
'caused' relates events, as do the words 'before' and 'after'; it does not
introduce an event itself. Similarily, to say someone made his arm go up (or
caused his arm to go up) does not necessarily introduce an event in addition to
the arm going up.[25]
It may,
nevertheless, be the case that some event other than the event of the arm going
up is present causing the arm to go up. This can be the case if we for instance
have a scenario where a pulley and a rope is rigged up and I can raise my
(paralysed) left arm by pulling the rope with my right arm. In this case the
agent raises his left arm by doing something else causing it to happen. What is
the answer now to the question whether something need to be added to my left
arm going up to make it a case where I raised my (left) arm? The same; nothing.
Clearly, all this shows only that the pulling of the rope and the raising of
the (left) arm are two events and not identical events, so nothing need to be
added to the event of the (left) arm rising. It is obvious that my paralysed
left arm would never have gone up unless some other event caused it, in this case,
my pulling of the rope, but still nothing needs to be added to the event of the
arm going up.
The question now becomes whether
both are something I do, two acts of mine, instead of one act
causing another behaviour of mine; that is, the relation between my pulling the
rope and my raising my arm. (Coughing, puking, my heart beating and crying are
all behaviour of mine, but they are not something I do in a narrower
sense of doing, the sense that we may call proper action or intentional
action.) If what follows from this case of two events, the one causing the
other, that they both are acts of mine, the one act causing the other, then we
must conclude that something is added to the event of my (left) arm going up -
the pulling of the rope. My raising my arm is then the sum of two events, the
one causing the other.
I agree with Davidson when he
rejects this answer. Though one can say that both events, the pulling of the
rope and the raising of the arm are both behaviour of mine (of who else?) or
actions in a wide sense, they are not both acts of mine, they are not
both actions I perform. The only action I perform is the pulling of the
rope. Nevertheless, this act, says Davidson, belongs to at least two sorts
of actions, since I pull the rope and the conditions are as they are (the world
is willing) I also raise my arm. This is the same as when I flip the switch and
the light turns on, I do both the flipping of the switch and the turning on of
the light, if it is the case that my flipping the switch causes the light to
turn on. Neither in this case, even if there are two distinguishable events,
the flipping of the switch and the light turning on, do I perform two acts; the
only act I do is flipping the switch. This might be described both as a
flipping and a turning on of the light, however, thus as two sorts of actions.
In addition I might cause the event of some burglar not breaking into my house,
etc.; but I do not do this in addition to my act of flipping the switch,
because that is my action in the narrow (proper) sense.
We often identify actions by their
effects, and also by their intended effects, so they often allow of more than
one description; one that describes the action itself and one that describes
its effect. This is I think what motivates Davidson to say that we have one
act and two types of action, allowing (at least) two
descriptions. We must never miss out of sight, however, that there were two
events, the one caused by another and the first was the act of the agent. That
is, in such cases as above we can describe my action both as a pulling of the
rope, and as a rising of my arm in the one case, and both as a flipping of the
switch and as a turning on of the light in the other case. (My acts are
the ones properly described as 'pulling the rope' and flipping the switch'.)
The second case have something the first case does not have, however, and that
is an interval of time between the events. In the first case no interval is
possible no matter how long the rope is or whether it is tight or not. In the
other case, however, we may construe a device that delays the currency leading
from the switch to the lamp so that we understand it (it is always some delay
of course, but we seldom think of it), for instance by increasing the length of
the cable significantly. This might be seen as a difficulty to the above
thesis, but Davidson has an answer to this that I think plausible. The reply
comes in two parts.
To say that I turned on the light implies
that I did something that caused the light to get turned on, namely flipping
some switch (or whatever device connected to the lamp capable as conditioning
my causing of the light turning on). Thus, we can say that the verb 'to turn
on' is a causal verb: x turns on y iff there are two
events, e and e', so that x is the agent of e (the
turning on of y by flipping the switch), e' is being turned on by
x, and e caused e'. So the turning on and the being turned
on are distinct events with some time t between them. In other words, x
did something that turned y on after time t. The time-lapse makes
no difference to the fact that only one action was performed, namely the
flipping of the switch, and not two actions. The action done by x, the
flipping of the switch had not its desired consequence until time-interval t
had elapsed. Here we might introduce a useful distinction that Hornsby (1980)
has made concerning the class of verbs which occur both transitively and
intransitively, and which support the following inference (where 'V'
stands for such a verb, 'T' for transitive, and 'I' for intransitive): from a
VT b to b VI (but not the other way around,
mind). Examples of such verbs are 'melted', 'grows', 'move', 'raise', and 'turn
on'. Thus, we might from 'A turned onT the light' infer that
'the light turned onI'. Nominal phrases like 'the turning on of the
light' may create ambiguities since it can refer both to someone's action of
turning onT the light, and also the light turning onI.
Hornsby remarks to one of Davidson's thesis - that all (physical) actions are
bodily movements - that this statement is ambiguous between movementT
and movementI, and true only on the former reading. A successful trying
(an action) is on Hornsby's view an action because some bodily movementsI
occur. These bodily movementsI are effects of the trying (the trying
which in this case is an action), and it is the fact that this trying has such
effects which makes it appropriate to describe it as an action, a movementT
of the body. From this it follows that since she commits herself to the
Davidsonian thesis that events do not have their effects essentially, the
occurence of such effects is not essential for the trying to be the very thing
it is, and it follows that 'being an action' describes an accidental property
of the event, which is a trying and also an action. Since the actions (tryings)
precede and cause the bodily movementsI, which clearly are external
events, it follows, according to Hornsby, that actions must be internal events,
and occur inside the body - with accidental exteral effects. One can appreciate
the distinction between transitive readings and intransitive ones without
concluding like Hornsby does. Gjelsvik (1990) argues, by time considerations
mainly, that Hornsby's argument does not argue for the step from thinking of
two different continuants (a and b) that the inference from a
VT b to b VI applies, to thinking that it
applies for continuants that are identical, like it is when I move my arm. He
argues that there is an assymetry between thinking of bodily movements having
effects, and thinking of bodily movements as effects by these
transitive/intransitive means.[26]
To have this conclusion properly we should think of bodily movements as moving
directly; perhaps along lines of direct awareness or non-inferential knowledge,
as Gjelsvik suggests.
Now to the second point. To say I
turned on the light means that I did something that caused the light to be
turned on. And so we might both describe my action as a turning on of the light
and flipping of the switch, although my real action was the flipping of the
switch. But when I have flipped the switch, I cannot immediately say that I
have turned on the light. This means that the action cannot be described in
both ways until the latter event (the turning on of the light, the effect of my
act of flipping the switch) have taken place after time-span t. Consider
again the answer 'nothing' that we have defended above. One thing is to deny
that effects belongs essentially to actions. Another thing is to deny that
causes belongs essentially to actions (bodily movementsT). I have
argued with Davidson that both are true, something that is implied by the
answer 'nothing'. Under one reading, Gjelsvik's suggestions above, that
movementsT of the body do not cause movementsI of
the body (for instance clenchingT and clenchingI of the
fist) but that it is a conceptual connection, might be said to imply the view
that bodily movementsT are events with a certain type of causes
essentially. Consider the following well-known example. There is a thing we may
describe as a bronze-statue and as a piece of bronze. In one sense these two
objects are identical. But in another sense, they are not. If I melt the statue
I still have a piece of bronze, but I don't have a statue anymore. And if there
is a possible world in which there is a statue exactly similar to one I have made
in this world - one made by e.g. nature - they would not share all properties;
the one in this world would have the causal history of being made by me, the
other, in the possible world, by nature. The same considerations apply to
actions. It all depends upon what types of properties we will consider
belonging to the event essentially. If we take it that essentially to be an
action that it has the property of being caused by some intention, we must
conclude that events that are not actions and events that are actions are not
identical, since the latter events have the property of being caused by some
intention of some agent. We might say that the indiviualisation of events and
actions, since it is not given per se, is relative to what we want out
of the individualisation. The topic of essential properties is one especially
difficult, as the history of philosophy shows.
Now to another related topic, that of mistakes.
According to the view on actions and events above - that the effects of actions
are not essential to the action itself - we could claim that one should release
any murderer on the defence that he never really killed, because all he did
was, for instance, to fire the gun (or pull the trigger or contract the muscels
in the finger in contact with the trigger) or stab with a knife (move the
muscels in the arm). There is a sense we could perhaps defend a murderer in
this way, and that is if we could prove e.g. that the agent were practicing his
(rather obscure, though) hobby of stabbing his sofa pillow in the dark, and
that he under no circumstances could know that his dear wife, tired of his
snoring had gone down to sleep on the couch. If we could prove this bizarre
scenario, we could say that since he did not intentionally stab his wife, he
did not intentionally kill her. The events are there, so there is no doubt that
the man in some sense killed his wife, but to be precise (if our defence go
through) we should say that what the man did, accidentally caused (no causing
is accedental, it is only because we apply intentional concepts to the effects
of actions that we can speak of 'accidental' causing) his wife to get killed or
die as the event can be described as. Davidson (1985) writes about a similar
case of a trial following a tragic incidence on board of a ship at sea, where a
man fell overboard after a fight and died. The defence pleaded the accused not
guilty, because "jurisdiction was confined to the trial of persons
committing offences on board an Irish ship on the high seas where such offences
were committed and completed on the ship",[27]
and because there are two ingredients in the definition of manslaughter, first
some unlawful act which causes the death and the second is the death itself.
Davidson (1985) remarks that the crime was both committed and completed on
board of the ship, and the unlawful act does not include the consequence
that makes it unlawful. So according to Davidson, 'A killed B'
means 'There were two events such that A was the agent of the first, the
second was a death of B, and the first caused the second.' Thus A's
action had (at least) two properties (two events), but only one of them were A's
action. Consider the sentence: 'the unlawful act does not include the
consequence that makes it unlawful'. In my example above we have it that 'it is
not unlawful to stab one's sofa pillow', and we have that 'it is unlawful to
stab one's wife'. We can divide the act of 'stabbing one's sofa pillow' into
the two events (i) 'A stabbingT the sofa pillow' and (ii)
'the stabbingI of the sofa pillow'. The same with the stabbing of
the wife: (iii) 'A stabbingT the wife' and (iv) 'the stabbingI
of the wife'. Stabbing one's wife is illegal, but stabbing one's sofa pillow is
legal. Now, what A did in both cases was a stabbing movement, so we
could put (i) and (iii) together to designate the basic action of stabbing: (v)
'A stabbingT X'. (ii) and (iv) now becomes: (vi) 'the
stabbingI of X'. We have it that for some X, (vi)
makes (v) unlawful. Neither (vi) nor (ii) and (iv) is unlawful since they are
not actions, but merely events (all are events, only that the events that are
caused properly have the property of being actions). We have it that whether
the stabbing is illegal or not depends upon what is stabbed, what fills
in the X. Actions are individualized by their causes and effects, says
Davidson. Among the causes are the desired effect of the action believed to be
causing the effect. In the above case, A desires to stab his sofa
pillow, and believes that by moving his hand in a certain way with a knife in
it some place near the sofa pillow, he shall be stabbing his sofa pillow.
Unfortunately, he stabs his wife, something that makes his stabbing illegal.
But does the stabbing of the wife make his action intentional? Davidson
consider mistakes, and says that "although intention implies agency, the
converse does not hold"[28]
I don't think intention implies agency, since we can always intend to do things
we do not actually do. Does agency imply intention? All intentional
action (agency) are caused by intentions or rather reasons for action.
Intentional actions are often identified by their consequences and their
causes. This must mean, I think, that we must have complete identification
between reasons and actions and their consequences for something to be a proper
intentional action. So unless the effect of my action is the desired effect,
and the action of mine is the action I believed would fulfill my desire, the
action I did was not intentional. These are pretty hard criteria. Davidson says
that some mistakes such as misreading a sign, misinterpreting an order, etc.
are actions but not intentional ones, since "making a mistake must in each
case be doing something else intentionally. A misreading must be a reading,
albeit one that falls short of what was wanted; misinterpreting an order is a
case of interpreting it […]",etc.[29]
That is, we have a class of events we call actions, some of which are
intentional and some of which are not intentional, nevertheless actions. What
is the common element between events that are actions? Consider the example
above again. We must distinguish three situations in which it is correct to say
that A stabbed his wife: in the first he did it intentionally (desired
to stab his wife, and stabbed her); in the second A didn't do it
intentionally but it was still an action (he thought it was the sofa pillow);
in the third it was not A's action at all (someone else than A)
lifted his arm and used it to stab (intentionally or unintentionally) his wife.
We would convict A on the two first, premediated murder (or 2nd
degree murder) and manslaughter, but not on the last (at least in an ideal
world). The difference between the three events are, according to Davidson that
in the first two cases, but not in the third, A is intentionally doing something.
According to this, Davidson proposes the following criteria of agency: "a
man is the agent of an act if what he does can be described under an aspect
that makes it intentional."[30]
"We do not say that a person who snores acts," says Hornsby
(1980), "because there is no description of the event of his making the
sounds he makes that can be used to tell us of anything he intentionally
does."[31] A in
the above example, then, stabs what he thinks is the sofa pillow, but actually
he stabs his wife. The stabbing of his wife is his action (or the effect of his
action) for that action is identical (have the same effect as) his attempt at
stabbing the sofa pillow. After having stabbed his wife we could ask A
why he stabbed his wife to death. He would probably answer: 'For no reason at
all (you idiot!); I thought I was stabbing the sofa pillow.' This means, if
Davidson is right, that the example from Mele (1988) mentioned earlier that I
(with him) called an unintentional action - cutting my hand while peeling
potatos - really, if being redescribable as intentional, was intentional after
all. That is, that I intentionally cut my hand. The problem is that both the
potato peeling and the cutting of the hand have the same (causal) history; the
reason I cut my hand is the reason I have for peeling potatos. Recognizing
this, however, makes us see what Davidson means; since they have both the same
reason, we must look at what action is rationally connected to the reason. We
clearly see in this case that this is the action of peeling potatos, not the
cutting of the hand. (This would have been the other way around if what I
intended was the cutting of my hand and ended up peeling potatos.) So we may
have a redescription of the action of cutting my hand along the intended lines
that my actions were supposed to take.[32]
We have seen that actions are often
identified by their consequences and that we must distinguish between actions
and intentional actions, so that an action is identified not only by what
reasons caused it and the effect, but also by the fact that what he does can be
described as falling under an aspect that makes it intentional. Thus, the
criterion of agency is intentionality, but the expression of agency is itself
purely extensional. Intentionally A wants to stab his sofa pillow, but
extensionally the event that the agent is wanting (we cannot really say it this
way, because of the obliqueness of attitudes, but understand me right) is his
wife being stabbed, just like Oedipus married his mother.
What we have considered
above is mainly a conception of action as actions done for some purpose, for
some reason(s), with some intention. As we have already introduced the concept
of intention as something that can both explain and not explain some action, we
may come to think that there certainly is something that is added to an event
to make it an action, namely an intention. This is what we shall investigate
next.
If we
shall follow what we have said so far, the role and nature of intentions is
quite easy to state. Intentions are not part of an action – the event – it must
be what causes it. Intentional actions are actions done with some
intention, and for this intention to explain the action it must at least
produce the action and be connected to it properly as the intention with which
the action was performed. Supposedly, this connection is causal; that the
intention causes the action, and that intentions, and beliefs and desires must
be states of some sort able to cause action or intention. It is not the expression
of reasons (beliefs, desires, intentions) that cause, but the fact that these
states disposes us in some way. It is only in this sense that makes it
appropriate to talk of reasons causing action. I shall not argue for this basic
Davidsonian conception, however, but take it for granted when I need mention
it.
That we
must include intentions as states of some sort, along with desires and beliefs,
is evident when we consider intentions never being acted on; pure intentions.
Still, that an action was performed with some intention is not secured
by the action being caused by this intention. This is evident in what
has come to be known as deviant causal chains. Many such cases have been
treated and tried eliminated through the years, but with no success, according
to Davidson (1987). This makes him think "that the concepts of event,
cause and intention are inadequate to account for intentional action."[33]
Our interest becomes more focused
when we turn to the stages in the emergence of intentions. What is it
that produces intentions and intentional actions? Do we need both a desire and
a belief, or just a belief, which has been proposed by some to explain
intention and action? Further, if we concede that a desire is needed in
explanation of action and intention, we must ask about the emergence of this
desire, whether the explanation of this require a further desire or just a
belief. This story, it must be remembered, about how intentions – and later
beliefs and desires – cause action, is not arrived at by introspection or
experimentation on agents acting, but rather by reflecting on the nature
of intentions, beliefs, and desires on the one hand, and the nature of
actions on the other hand.
After having disclaimed his early
(1963) reduction of intentions to primary reasons, Davidson claims that an
intentional act requires both a belief, a desire, and an intention: a desire
toward some actions with a certain property, a belief that acting in a certain
way will promote such an outcome or situation, and an intention or judgement to
perform this action. If this is right, the emergence of an intention requires
two transformations on the belief-desire couple, one in which the belief
and the desire are brought together, the other in which some intention is
formed.[34]
These transformations are not undisputed, but they seem to be required through
how we reason about what to do. Unless the belief and desire are brought together,
a course of action (belief-content) is not seen by the agent as attractive in
the light of the fact that it promises to bring about a desired state of
affairs (desire-content). This transformation is what has been called practical
reasoning; that is, reasoning from a perceived value of some end to the
value of its means. I consider this below. Nevertheless, if we left it here,
leaving this account to explain action – like Davidson (1963, 1980) does – we
would have it that we perform every action that we believe would promote some
good or satisfy some obligation. This is not the case, we do not do everything
we consider good, etc., because in many cases the options we think of as good
are contradictory in a way that leaves us unable to perform all: for instance,
I want to go both to Paris and London the very same weekend. I cannot do both,
at least not at the same time, so I have to decide which I want most. In doing
this, if I shall ever be able to act, I must commit myself to act on the one or
the other. For the mature Davidson this is going from a stage in which we
perceive, or imagine we perceive, the attractions or the drawbacks of a course
of action to a stage in which we commit ourselves to act. This stage is perhaps
another desire (or pro attitude) but unlike other desires or pro-attitudes it
is not merely conditional or prima facie. If it shall be able to produce
action it must not only express some appreciation that some good would come of
acting in a certain way, but also be definite about it. In this sense,
intentions are agglomerate - which means that "if I know that joint
satisfaction of desires A and B is impossible, I cannot in a rational way both
intend A and also intend B. I can at most either intend A or intend B"[35]
- they permit of no contradicting intentions in the way that conditional or prima
facie desires do. I will specify these steps some more later on.
1.7 Explanation and Prediction.
When we explain
action, we explain it in retrospect. We are seldom very good at predicting
others actions, even though we are quite good at predicting our own, at least
when we have intended something.[36]
Intention does not logically imply agency. Thus, there is a certain assymetry
between prediction and explanation. This has been recognized in general
explanation too. Consider the following example: I use fertilizer to have nice
plants. To have nice plants is the reason why I fertilize; the fertilizer is
very strong though, so many plants die, and many lives on in a better condition
than before. Before I fertilize, I predict that the fertilizer will make my
plants better. And accordingly when someone asks me why some plant is so fine I
say it is so because of the fertilizer. The problem is that I give the same
reason to explain why other plants are dead. Something analogues is true also
in action explanation. One cannot from a piece of practical reasoning of some
agent predict what it is he will do in any definite way. Perhaps he has
performed another piece of practical reasoning that gives him a better reason
to perform some other action. Thus prediction from pieces of practical
reasoning to action is difficult. Only when the action is done can we say what
piece of practical reasoning, if any, led to the action.
Take a
look at Davidson's early (1963) account of how reason explain action:
C1. R
is primary reason why an agent [A] performed the action [j] under the description d only if R
consists of a pro attitude of the agent towards actions with a certain property
[a], and a belief of the agent that [j], under the description d, has that
property.[37]
This account is
evidently false in that it reduces intentions to primary reasons (a
belief and a desire). I take this for granted, so it is another feature of this
account that I will focus on. Take some simple event: A flips some
switch. His reasons are, we learn (some way or other) that he wants to
turn on the light, and he believes that flipping the switch will turn the light
on. Davidson imagines A's reasoning as follows: any action
of mine that turning on the light is desirable; my flipping the switch will be
such an action; so, my flipping this switch is desirable. Parallel to A's
reasoning, is our explanation: A wants to have the light turned
on. He thinks that flipping the switch will constitute such an action.
Therefore, he flips the switch. A want is, or entails, says Davidson, a
certain disposition to act to obtain what one wants, and that someone has a
certain disposition may be expressed as a generalization or law governing the
behaviour of that person.
In the
above simple account, our description of A's reasons provides us with a
classical nomological-deductive explanation of his action. Our description of A's
desire expands into a law of some sort, through the entailment to a
disposition, while the description of his belief provides the singular premise
that triggers the wanted conclusion. If we take it that the relationship
between an agent's reasons and his (intended) action is causal, the singular
causal claim 'A caused B' explains the occurrence of B
provided A only if there are general laws connecting the former with the
latter in such a way that, according to Hempel, "given a description of
the antecedent events [A], the occurrence of the effect [B] can
be deduced with the help of the laws."[38]
Davidson (1976, 1980) interprets this not as meaning there is a true law
connecting A and B under some descriptions of A and B,
but rather the law copes with the very expressions 'A' and 'B',
which means that the descriptions of cause and effect must come close to giving
the form of the required law.
The above account is evidently false
when we consider the truth of the law – the description of desire (giving the
form of the law). People often fail to perform any action at all to achieve a
desired end, although they believe the means at hand. Further, no one ever
performs all the actions he believes will lead to the end, if it is more
than one. What we can say, according to Davidson, is that someone who has a
desire for some goal will tend to behave in certain ways under specific
circumstances. These kind of reflections led Davidson to reject his own early
view (1963) and import the concept of prima facie account of reasoning
(1970) into action-explanation. The desire is now represented by a probabilistic
law that together with some appropriate belief provides us with a low-grade
(not zero probability) statistic explanation. To have real laws we must
transform these sorts of low-grade tendency statements into serious empirical
laws. The way this can be done is perhaps by making a fuller statement of the
conditions under which the agent will act. Hempel among others have suggested
that the laws of action state what a rational agent will do: the rational
agent will do what, in the light of his beliefs and desires, is his optimal
course of action. Explaining the agent's actions we describe his attitudes in
terms of their contents, we refer to laws that specify how any rational agent
acts in various circumstances, and add as a premise that the agent was
rational. Decision theory is provided as a solution to the well-known problem
of the wants for doing several incompatible things. Davidson opposes this idea
for the following reasons. Besides questions about the adequacy of decision
theory (that he discusses in Davidson (1976a, 1980: pp. 269-273)), there
remains the following oddity in Hempel's proposal, according to him:
the 'law', so-called, of decision theory (or any other
theory of rationality) are not empirical generalizations about all agents. What
they do is define what is meant (or what someone means) by being rational.
Application of Hempel's scheme depends on knowing that the agent is rational,
but there is no way of determining this except by establishing that the 'laws'
fit the agent. The 'explanation' in terms of rationality therefore lacks
explanatory force.[39]
What we want to
know about is the beliefs, desires and intentions of the individual person
having these, and not how anyone (or everyone) with certain dispositions and
beliefs would act under certain circumstances. Davidson further provides us
with an argument showing that the main empirical thrust of action-explanation,
does not come from any assumption of rationality, but "rather from the
attributions of desires, preferences, or beliefs", themselves.[40]
As such, a 'law' like 'Someone who wants to j has a tendency to do what he
believes will result in j', "adds
nothing to what we already understand if we know what a want or a desire is;
and […] the relevant belief and desire explain the action only if the
belief-desire pair caused the action in the right way."[41]
We now see a connection to the related point noted before, that when we explain
action we identify the phenomena to be explained, and the phenomena that do the
explaining, as directly answering to our own norms, but not from any assumption
of rationality in the premises of the explanans. In addition, this point marks
the importance of getting to grips with the nature of beliefs and
desires, since they are the determinants of explanation and the providers of
empirical thrust. So Davidson allows a certain prediction available for action
explanation, one that is limited to individual agents and their history of
action. To specify the conditions for such predictions are almost impossible,
but getting to know people means getting better at predicting their behaviour.
Below we shall consider in some detail Davidson's two conceptions of desire.
There
might be seen to be a conflict between attributing attitudes and a functionalist
account of states. A typical functionalist account of desires and beliefs,
etc., fixes the meaning of such folk psychological or normative concepts as one
does with theoretical terms, by the set of laws and generalizations in which
they figure. Beliefs and desires, then, enter into a multitude of causal and
nomological relations. My belief that the mail has arrived, together with my
desire to have the mail, and other beliefs, may cause my doing so as to get the
mail. We might on a functionalist account formally describe regularities such
as the above by formulating them as laws of typical causation; for instance, a
belief that p and a belief that if p then q will typically
cause a belief that q; a desire that p and a belief that p
iff I bring it about that q will typically cause an intention to q;
and several other more specific laws. One must not apply such generalizations
too mechanically, however, because there are indefinitly many circumstances,
not exhaustively specifiable in advance, in which these general or specific
regularities fail to hold. Now, the question is what role such nomological
regularities play in the actual attribution of attitudes; that is,
whether or not the methodological context for such attributions is
nomological reasoning. At least two possibilities seem plausible here; either
one may think that if the methodological context is nomological reasoning this
requires the laws to be internalized to be able to use them, and on this ground
think that an altogether different methodology should be the basis. Or, one may
say that whereas the correct explication of the concepts of attitudes requires
that one cite such laws, mastery of the concepts (the ability to use them) does
not require that one have internalized such laws. Gordon (1987) suggests that
we take the first possibility here, since he thinks the second to be involving
"an uneasy alliance".[42]
He introduces the concept of 'practical simulation' to represent the
methodological ground and elaborates it by insights from the much discussed
field of radical interpretation, issued and developed by Quine and Davidson,
and others. This is interestingly enough, but shall not concern us here. I
propose we choose the first possibility mentioned here, but have a sober
understanding of what type of 'laws' we are concerned with here. I think this
way of taking the issue does not go against any suggestions by Gordon, only
where he thinks mastery of laws or rules implies mastery of laws or rules qua
laws or rules. I deny this. We may certainly explicate our state-attributions
both through how states are related by causal laws, thus explicate attitudes qua
states, and through methodological considerations, thus explicate
attitudes qua attributions.
As noted above, when we explain
actions and try to predicate actions, we try to "see" a coherent
pattern in the behaviour of some agent. Further, since the concepts used cannot
properly have referents in any neuro-psychological realm, this coherent pattern
and all the concepts therein, also the concept of action itself, are
interdefined, as well as the laws connecting them. It is this fact that makes
it inessential that we distinguish between the attribution and the
explication in action explanation; although it might bring forth points we
otherwise would miss. Attribution is part of what explaining action is,
as well as states and dispositions. They all point in the same direction and
from the same vantage point; there is no underlying onthology displayed in
functionalist conceptions.
1.8 Davidson on Reasons for
Actions and Practical Reasoning.
Davidson's (C1)
states that there are some actions ("an indefinitely large number"[43])
with a certain property that the agent have a pro attitude towards, and that
this, together with some connecting belief that some action has that property,
is what constitutes a reason for action. Among the intutions behind this
conception is that the expression of a desire (or pro attitude) must cite some
"feature, consequence, or aspect of the action" for which it is a
reason.[44]
We cannot
explain why someone did what he did simply by saying the particular action
appealed to him; we must indicate what it was about the action that appealed.[45]
Desires have a
certain generality which makes them unsuited (in most cases) to connect
directly with particular actions. This aspect is taken care of by introducing
an instrumental belief that directs, suggests, channels, or specializes
the desire, so that one action comes up as (possibly) contributing to the
desired end. We can give the following example (from Davidson (1963) to
enlighten (C1). A flips the switch (to the on-position), i.e.
does j, and turns on the
light, illuminating the room. Flipping the switch have the property of turning
on the light, that is, flipping the switch is (or has the effect of) turning on
the light, i.e., j is a. A believes
or knows that j is a, and has a pro
attitude toward actions with property a.
It is perhaps awkward to say that
what we desire are an indefinitely large number of actions with some property
that is the effect of those actions. We are accustomed to say that what we
desire is that effect or property itself. This is right, and Davidson does not
deny it. Rather, his conception of desire in (C1) expresses a deductive
approach by spelling it out, enabling us to see straightaway an underlying
dispositional analysis of desires - that having a desire is being disposed
to act so as to obtain what one wants.[46]
It is
the above considerations upon the relations of belief and desire to action that
point the way for Davidson (1963, 1980) to see how reasons explain actions;
lending its pattern from practical syllogisms:
Aristotle held
that an action is like the conclusion of a syllogism of which the belief and
desire which are the reason for the action are the premises. A reason explains
an action because, given the reason, the action somehow follows logically. For
instance, the desire, functioning as the major premise, could be expressed as
'any action of mine which results in a state of affairs S is desirable.' The
belief, the minor premise, would be 'jing would result in S.' The
conclusion which can be deductively drawn from these two premises is that jing is desirable, but Aristotle simply
identified the conclusion with the performance of a j-type action.[47]
The
nomological-deductive conception of reasons has seem problematic for Davidson.
Especially because it takes uncritically all actions being good as long as it
contributes to the desired end; but also because it seems to collapse in the
face of cases of pure intending. On the old (1963) account of reasons for
action, the proposition corresponding to the action would have to be something
of the form 'this action is desirable',[48]
and this suggests that we should also identify this proposition as the
expression of the intention to j. When an action is performed
there is no need to separate the intention from the action itself, and the
practical syllogism could be taken to explain the intention or the action
indifferently; that is, the primary reason for doing j is the same as the
primary reason for intending to j. And, given this,
when we have an intention only, and no action, we will have an explanation for
the intention which is the same as that we would have had for the action if it
had been performed. But as Davidson (1970) saw, there is some serious difficulty
with this.
Evnine (1991) asks us to consider
the following two syllogisms, concerning one piece of poisoned chocolate, to
make evident the difficulty.[49]
First, any
action of mine which is an eating of something sweet is desirable; eating this
chocolate would be eating something sweet; therefore eating this chocolate is
desirable. And, secondly, any action of mine which is an eating of something
poisonous is undesirable; eating this chocolate would be eating something
poisonous; therefore eating this chocolate is undesirable. We appear to have
two valid arguments, both with true premises, but yielding contradictory
conclusions, that eating this chocolate is both desirable and undesirable.[50]
Something has gone
wrong with the account of practical reasoning, and having identified the
conclusion with an intention, we would have to suppose that I did both intend
to eat and not to eat the chocolate. But since intentions are agglomerative,
this cannot be so. The problem can be diagnosed by noting that on the one side,
since practical reasoning is meant to justify the idea of weighing up different
considerations in deciding how to act, it would be wrong to suppose that we
cannot go through both the arguments about the poisoned chocolate. On the other
hand, this is something we cannot countenance as things stand, since we cannot
have two valid arguments from true premises which have contradictory
conclusions. The cure might be to relativize the conclusions. One solution
might be to use 'prima facie' as an attributive adverb, a predicate,
helping to form such predicates as 'x is better, prima facie,
than y' as a conclusion, and that this does not contradict 'y is
better, prima facie, than x'.[51]
This leaves the conclusion we can draw in every case of conflict that 'x
is better, prima facie, than y, and y is better, prima
facie, than x', which comes down quite empty as 'There is something
to be said for, and something to be said against, doing so and so - and also
for and against not doing it.' We must perhaps introduce a third factor, an
argument adding up the conflicting arguments, the conclusion of which expresses
one's better judgement. In this way we know not only the reasons on each side,
but also how they add up. It is important to note that that something is one's
better judgement is not given through the "most" moral argument of
the conflicting ones, at least not in a sense interesting here, but as the
judgement about which is the best of the two. Not being a syllogism at all,
however, this third argument will not solve anything, not even when we again
introduce 'prima facie' in suitable places.
The real source
of difficulty is now apparent: if we are to have a coherent theory of practical
reason, we must give up the idea that we can detach conclusions about
what is desirable (or better) or obligatory from the principles that lend those
conclusions colour. The trouble lies in the tacit assumption that moral
principles have the form of universalized conditionals; once this assumption is
made, nothing we can do with a prima facie operator in the conclusion will save
things.[52]
Davidson's solution
is to introduce a prima facie sentential operator into the major
premise of the practical syllogism and thus introducing a new conception of
desire. A sentential operator makes a complex sentence out of one or more
sentences, and the special operator in question here, the prima facie
operator, makes a complex sentence out of two sentences, one stating that
some actions are desirable, and the other stating in which respect they
are desirable. Practical reasoning, containing all weighing of pros and cons,
takes us only so far as a judgement of prima facie desirability -
"it is practical in its subject, not in its issue"[53]
- it never issues in action. An action or intention is reflected by a judgement
of unconditional desirability. The problem now is to say how we get from
judgements of the first kind to those of the second. When we act, says
Davidson,
we have got
beyond the stage of thinking that the action we perform is desirable in so far
as it has some feature or other. Our very acting shows that we formed a
judgement that not only is it desirable in this or that way, it is
unconditionally desirable. Judging an action unconditionally desirable doesn't
mean not having any reservations about it. The point is that in acting
intentionally, we have finished our deliberation and made our decision.[54]
Thus, we have on
the one hand practical reasoning, the conclusion of which are judgements about prima
facie desirability; on the other hand actions that reflects unconditional
or all-out judgements in favour of doing something. In cases where we do
in fact act there is no need to separate this judgement from the action itself,
and in cases where we do not act, and have gone beyond our judgements of
prima facie desirability, having made a judgement of unconditioned
desirability, we have a case of pure intending.
Of course I can
still say of the completed action that it is desirable in so far as it is this
or that, but in choosing to perform it I went beyond this; my choice
represented, or perhaps was, a judgement that the action itself was desirable.
[But in] pure intending […] there is no action to judge simply good or
desirable. All we can judge is the desirability of actions of a sort, and
actions of a sort are generally judged on the basis of the aspect that defines
the sort. Such judgements, however, do not always lead to reasonable action, or
we would be eating everything sweet we could lay our hands on (as Anscombe
pointed out).[55]
The connection
between a prima facie, all-things considered judgement and an all-out
judgement is provided by a third element: the Principle of Continence, a
principle which, as a rational man, the agent accepts. This principle tells me
to "perform the action judged best on the basis of all available relevant
reasons",[56] that is,
the all-things considered best action to perform.[57]
There are several problems with the above account. How we are to account for
the Principle of Continence is one thing, the nature of the probabilistic law
is another. I shall not consider this here.
The radicality of
Davidson's new conception of desires beyond what is said above is not well
understood in many respects. It has great implications for discussions about
practical reasoning, action explanation, and the understanding of desires in
general. Nevertheless, I will not try to explicate Davidson's new conception of
desire any further here. There were no way to avoid mentioning this conception
in an essay considering mainly desires, but that does not necessarily mean we
cannot avoid mentioning it any further. At least, so I hope. I hope what will
be said further on desire will not contradict Davidson's new conception of
desires. I think the following considerations will make this possible.
First, the claim that a desire is a
disposition to act so as to obtain what one wants can be held without
specifying the law governing the behaviour of that person expressing this
disposition.
A want is, or
entails, a certain disposition to act to obtain what one wants. That someone
has a certain disposition may be expressed as a generalization or law governing
the behaviour of that person. […] One trouble is, of course, that the 'law' is
false. […] But we can say of someone who has a desire or end that he will tend
to behave in certain ways under specified circumstances.[58]
Thus, we need not
specify these conditions. Second, when we explain actions, we ordinary give
reasons with the conditions not mentioned. Typically we explain some action by
citing someone's desire for some state of affairs that is his desired goal.
Specifying the conditions further is definetly important, but that does not
mean we are wrong in giving the reason with which someone acted as a desire
that a. Third, the form I
will give about how we shall take practical reasoning will be so general that
it can be developed so as to incorporate Davidson's analysis of desires.
1.10 Desire-states as Phenomenal States?
Above, we have
taken it for granted that for any action, among the reasons for that action, we
find a desire (or pro attitude) as the motivating element. This view may be
termed a Humean one. The Humean view on motivation for action has been
challenged, however, by a rationalist view on motivation, that claims belief
alone being able to spark off action. There is nothing odd about this, because
often what we do seem to be wholly dependent upon what we think we should
do. Think of someone, let us call him Peter, who is bying free his son from
kidnappers wanting ransom. It may seem odd to say that this person desires
to pay the ransom. Rather, what we should say, according to the rationalists,
is that Peter belives he should pay the ransom to get his son back, and acts on
that belief. In such a case then, accordingly, beliefs may spark off action
itself. We shall consider this claim more thoroughly later.
Desires, beliefs, suppositions, hopes,
etc. are propositional attitudes. Propositional attitudes have propositional
contents. The propositional content can be identical between different
attitudes; the only variation being the attitude itself. What makes the Humeans
think we must have a desire motivating our conduct? What is it about desires,
in opposition to beliefs, that motivates us? What is the nature of desires? How
do we differentiate a desire from other propositional attitudes?
One of the classical accounts of the
nature of desire and motivation is partly phenomenological. About passions,
Hume (1740, 1978) says:
'Tis obvious, that when we have the prospect of pain or pleasure from any object, we feel a consequent emotion of aversion or propensity, and are carry'd to avoid or embrace what will give us this uneasiness or satisfaction.[59]
As Shaw (1989) correctly points out, this passage
suggests that the way we "know" about the presence of the separate
desire factor that is always needed to motivate actions, is by being directly
aware of some desire-feeling. On Hume's view, desire moves us because it
is a passion, an affect, a felt impulse or urge, something which has a
phenomenal character. If desire were not an affect, it could no longer, on this
view, move us to act.[60]
[T]he impulse [to action] arises not from reason, but is only directed by it. 'Tis from the prospect of pain or pleasure that the aversion or propensity arises towards any object […]. Where the objects themselves do not affect us, their connexion can never give them any influence; and 'tis plain, that as reason is nothing but the discovery of this connexion, it cannot be by its means that the objects are able to affect us.[61]
That the phenomenal character of desires (or passions)
is important is evident from Hume's list of passions; one can not
"feel" love without feeling it.[62]
The original (but incomplete) Humean view then, have two claims about reasons
for action: first, that there will be no action without any desire (or
passion); second, that the desire in question has a phenomenal quality, that it
is felt. In some cases both claims seem to be correct, but in other cases the
desire present seem to have no phenomenal character at all. In still other
cases, it seems to be no desire component present at all. Consider an example
from Schueler (1995):
I would say, for instance, that I had no desire to attend a meeting at my son's school the other evening. I would much rather have stayed home and read. But I attend the meeting because I believed I had a responsibility to do so, a responsibility mostly to my son but partly to my community.[63]
Here the agent actually goes to the meeting, he
manages to act even having no felt desire to do what he does. How can this be?
All we need to have a case against Hume's view as exposed above, is a possibility
for there being no felt desire at the time of the decision, and there is
certainly such a possibility present in many cases. If we opens up for this
possibility, one might claim that the desire present have no phenomenal
character at all, and thereby deny Hume's second claim. One could go further
and deny on the same grounds that there is no desire present at all. Hume would
handle such denials by claiming that the agent in the above example is
afflicted by a "calm desire", the sort of desire that fools the
unwary into thinking that it is a belief (or product of Reason).
Now 'tis certain, there are certain calm desires and tendencies, which, tho' they be real passions, produce little emotion in the mind, and are more known by their effects than by the immediate feeling or sensation. [. . .] When any of these passions are calm, and cause no disorder in the soul, they are very readily taken for the determination of reason [. . .][64]
But this certainly has an air of the ad hoc,
and can presumably be replaced by a more cogent explanation if available,
presumably one that denies desires a phenomenal character essentially. The
problem is how.
According
to Hume, a typical quality of passion is an emotion of propensity, or one of
aversion. The first is a positive emotion, the other is a negative one. The
question is in what sense are the first positive and the other negative.
Answering this question will also distinguish passion (and therein desire) from
other propositional attitudes like belief. We should not let the basis for the
negative-positive distinction rest in the pleasant-unpleasant or the
attractive-aversive dimensions, and thus not in this way distinguish passions
from beliefs. One reason for this is that some qualifier, such as 'typical'
used above, seems unavoidably to come creeping back in when we consider
passions. That is, it may be typical that one finds desire pleasant overall,
but that does not make it necessarily so. A second reason is that it is
possible to make sense out of the notion of unconscious desire, as well
as non-conscious desire, or at least desiring without feeling the
desire. In this case we would not be able to call up phenomenal qualities to
explain the positivity of desire and negativity of aversion. Finally, the
pleasantness of desire and and unpleasantness of aversion seems itself to call
for explanation. Why does desire feel pleasant? Isn't it because one has a
positive attitude, a pro attitude towards some propositional content? It
seems that what is called for is nothing but such a positive (or negative in
the case of aversions) attitude.
One
might alternatively explain the distinction between positive and negative
attitudes, by saying that the positive attitudes, like desires, involve a
positive evaluation of the object or content of the attitude, whereas the
negative attitudes, like aversion, involve a negative evaluation of the object
or content of the attitude. Thus as stated, a desire is for something one
believes to be good, whereas aversion is for something one believes to
be bad. This is all right as long as believing something to be good (or
bad) is understood to involve having the corresponding attitude - and nothing more.
But it is arguable that evaluative beliefs do not necessarily involve the
corresponding attitudes; that one may believe something to be (justifiably)
good (or bad) while being quite indifferent to it oneself. Also, it is arguable
that unlike the concept of attitude, the concept of 'evaluation' is closely
connected to evaluative utterance; to attribute to someone an evaluation
of something is to say something about what he would be disposed (sincerely) to
say, if he were to say anything one way or the other. As Gordon (1987) remarks,
this affinity for utterance makes it more troubling to attribute evaluations to
non-linguistic animals, although one might be willing - like for instance
Dretske (1988) - to talk about their attitudes, desires or preferences. For
these two reasons it might be thought preferable to speak of attitudes. Stampe
(1987) has nevertheless proposed an evaluative way of accounting for desires,
but one that builds on the attitudinal. We shall consider this later.
Now,
the above definition is so broad that it not only incorporates desires and
aversions proper but also other, what Hume calls, direct passions, and indirect
passions; attitudes that have other properties not shared by desires, and
therefore ought to be kept separat.[65]
A further argument is thus needed. Such an argument, the teleological argument,
has been provided by Smith (1987). We shall have a look at that next.[66]
1.11 Direction of Fit and
the Teleological Argument.
Smith finds the origin of his conception of desire in
one of Hume's remarks (cited above), that some desires are "more known by
their effects than by the immediate feeling or sensation".
According to this alternative conception, desires are states that have a certain functional role. That is, according to this conception, we should think of the desire to f as that state of a subject that grounds all sorts of his dispositions: like the disposition to f in conditions C, the disposition to f in conditions C', and so on (where, in order for conditions C and C' to obtain, the subject must have, inter alia, certain beliefs). For Hume's suggestion about how the calm passions are known may then be translated into the thought that the epistemology of desire is simply the epistemology of such counterfactuals.[67]
It is the effect of desires that makes us
recognize that we have them, that is, the action we do or are disposed to do.
Smith's account can be applied to
distinguish between desires and beliefs, a distinction that metaphorically can
be made in terms of their different directions of fit. The distinction
between (what has later been known as) two directions of fit is presented in
Anscombe (1963).[68] She
describes a man going shopping with a shopping list while being tailed by a
private detective listing the man's purchases, and asked what distinguishes the
shopping list from the detective's list, she answers the question thus:
It is precisely this: if the list and the things that the man
actually buys do not agree, and if this and this alone constitutes a mistake,
then the mistake is not in the list but in the man's performance […]; whereas
if the detective's record and what the man actually buys do not agree, then the
mistake is in the record.[69]
Platts (1979) uses the distinctions capacity on mental
states thus:
Beliefs aim at the true, and their being true is their fitting the world; falsity is a decisive failing in a belief, and false beliefs should be discarded; beliefs should be changed to fit the world, not vice versa. Desires aim at realisation, and their realisation is the world fitting with them; the fact that the indicative content of a desire is not realised in the world is not yet a failing in the desire, and not yet any reason to discard the desire; the world, crudely, should be changed to fit with our desires, not vice versa.[70]
Since the concept of 'direction of fit' is just
metaphorical, we must specify it, and Smith's shot at this is the following
argument, called the teleological argument:
(1) Having a motivating reason is, inter alia, having a goal
(2) Having a goal is being in a state with which the world must fit
and
(3) Being in a state with which the world must fit is desiring.[71]
It should be added to this argument the following
assumption:
(4)
Desiring is being in a state that disposes the agent to act under certain
conditions.
The dispositional analysis of desire is consistent
with the claim that certain desires have phenomenological content essentially,
and with the claim that certain desires lack phenomenological content
altogether.
For, according to this conception, desires have phenomenological content just to the extent that the having of certain feelings is one of the things that they are dispositions to produce under certain conditions. Some desires may be dispositions to have certain feelings under all conditions: these have phenomenological content essentially. Other desires, though they are dispositions to behave in certain ways, may not be dispositions to have certain feelings at all: these lack phenomenological content altogether.[72]
If we now return to how the specification of the
metaphor of 'direction of fit' actually is a specification, Smith says that,
the difference between beliefs and desires in terms of direction of fit comes down to a difference between counterfactual dependence of a belief and a desire that p, on a perception that not p: roughly, a belief that p is a state that tends to go [out] of existence in the presence of a perception that not p, whereas a desire that p is a state that tends to endure, disposing the subjects in that state to bring it about that p. Thus, we may say, attributions of beliefs and desires require that different kinds of counterfactuals are true of the subject to whom they are attributed. We may say that this is what a difference in their direction of fit is.[73]
Here Smith talks of desires as the state of desiring
that p - some goal. Being in such a state means/is identical with being,
or grounds being, disposed to bring it about that p, thus do any action
that will contribute to p, the desired end under certain conditions. A
desire can both be described as a state and what it disposes the agent to
(thus, it is a dispositional state). According to Smith, the dispositional
analysis of desire is an account of what a desire is that explains how it can
be that desires have propositional content, "for the propositional content
of a desire may then simply be determined by its functional role."[74]
He notes that there need be nothing simple about this determination,
however. What is meant by the claim that the functional role determines
content? It must just be that it determines the direction of the content, since
the content itself is a proposition that can equally well be held by a belief
or some other propositional attitude. We must distinguish the desire from its
content and define it qua attitude. The attitude that is a desire then
is defined first, as determining the direction of the content, and second, as
disposing the agent to act so that the content comes through. As such, desires
are often recognised by their effect, rather than by their phenomenal
character. This account ties desires closely to actions in that desires are
defined by their (possible) effects.
1.12 Shortcomings of
Definitions by Direction of Fit?
Humberstone (1992) notes two shortcomings of Smith's
(1987) account. First an observation made also by Pettit (1987) and Price
(1989). One may concede to Smith the claim, according to his cited
specification above, that no propositional attitude can have opposite
directions of fit in respect of its propositional content, without being forced
to deny that, say, desires are beliefs. For the concession means only, it is
argued, that a desire that p isn't a belief that p, not that,
say, a desire that p is not a belief that q, for some distinct
proposition q. And this q might be the proposition that it is, in
some suitable attuneated sense, according to Humberstone (1992), desirable that
p.[75]
Second, recall that according to Smith
(1987) a belief that p is (roughly) "a state that tends to go [out]
of existence in the presence of a perception that not p, whereas a
desire that p is a state that tends to endure [in the presence of a
perception that not p], disposing the subjects in that state to bring it
about that p." According to Humberstone (1992) there are at least
two ways of interpreting the "perception that" locution here. If the
phrase is read (as it most naturally would be) as meaning "belief
that", the characterization is vulnerable to a certain objection, that
Humberstone calls the mutatis mutandis objection, whereas, if the phrase
is read non-doxastically, the characterization does not meet a plausible
requirement of universality.[76]
On the first interpration, it is claimed that it is too restrictive to count
the referance of "perception that not p" as specifically to
sensory perception, and we should understand by this phrase, either a coming to
believe that not p or, since "perceive that" is a
factive construction, a coming to know that not p. Either way, it
is clear that we are here explicating the cognitive direction of fit by
reference to states with the cognitive direction, that is, it may seem
circular. Indeed, even if we took "perceive" in "perceive
that" to allude to specifically sensory perception, the point would
remain.[77]
No asymmetry has been characterized by the observation that whereas a belief that p tends to disappear on the acquisition of a belief that not p, a desire that p does tend instead to persist on the acquisition of a belief that not p. To disclose an asymmetry, all the relevant mutanda must be mutated.[78]
Humberstone claims that it is not a circularity
objection that has been urged:
The worry is not that some would-be analysis of the concept of
belief fails in virtue of employing, in disguise, that very concept; for
clearly no such analysis was being offered. The point is rather that you cannot
informatively characterize a fundamental disanalogy between the ways in which
beliefs and desires relate to their objects by contrasting them in a respect
itself specified by reference to one of those ways.[79]
On the second interpretation of Smith's talk of
perception, taking the subject's perceiving that not p to amount to no
more than that its perceptually appearing to the subject that not p,
where this is related to any belief that may or may not be formed as the ground
of that belief. Thus, perceptions are now not associated specifically with the
cognitive direction of fit in the unwarranted way of actually themselves
possessing that direction. The mutatis mutandis objection lapses in the
face of this. From this fact, however, the need to talk of states tending
to go out of existence or to endure on a perception that not p, arises.
S may well take appearances at face value, and come to believe on the basis of those appearances, that that is how things are. But S may suspend belief, perhaps out of suspicion that the current circumstances are not conductive to reliable belief-formation, or indeed retain or form a belief that things are not how they seem in the respect in question. The talk of a tendency for beliefs that p to disappear on perceptions that not p is a way of registering the default status of taking apperances at face value and the special or unusual nature of the circumstances prompting the other (suspension, disbelief) responses.[80]
The account of what direction of fit consists in, thus
rests on tendencies; but can we rest content with this? What we aimed at in the
beginning was not an account that rendered beliefs with a certain
characteristic direction of fit, one which it might be deemed to have on
the basis of how its instances typically behave, the existence of a minority of
instances behaving otherwise notwithstanding. What we wanted was an account
that rendered every individual case of believing as a case of attitudizing
which itself has the cognitive direction of fit. Likewise, this must be so with
desires and the conative direction. If this is right, we should look for an
account of direction of fit which is universal in the sense that it adresses
something shared by all cases of belief and all cases of desire. I will not
consider any positive suggestions to this effect, but only note that there are
problems with defining beliefs and desires having opposite directions of fit in
this respect.[81]
A third critique of the distinction
between beliefs and desires according to their direction of fit comes from
Schueler (1991). Schueler recognizes levels of problems with the direction of
fit-metaphor when considering 'hope'; an attitude that Smith himself seems to
want included among those states with which the world must fit. If I hope that p
then I must want it to be the case that p, so it seems that hopes should
be included among desires. But while I can hope that p without believing
that p and without it being the case that p, I cannot hope that p
when I know that not-p, or when I believe p is impossible, says
Schueler. He notes that Smith's conception overlooks an important cognitive
component in 'hope'; "such as the fact that such phrases as "grounds
for hope that p" or good reason to hope that p"
typically refer to exactly the same thing as would such phrases as
"grounds for belief that p" or "good reason to believe
that p"."[82]
Consider the man who hopes that he remembered to turn off the bath water last
week, before he left on his vacation. What is the direction of fit here? The
point comes down a lot heavier when we consider the grounding of Smith's
metaphor in dispositions to act. The hope of the man above seem not to
"endure" - as Smith wants desires in general to do when confronted
with a perception that not-p - when he gets a phone call from his home
neighbours telling him that their kids are scuba-diving in his living-room.
There is no way that one can hope that p once one discovers that not-p.
Thus, it seems like hopes have the same direction of fit as beliefs; they tend
to go out of existence in the presence of a perception that not-p. It
cannot be the case then that for all hopes, at least, one is disposed to act in
such a way so that what one hopes be realised. This is evident considering
hopes about the past. There is nothing that the man in the example above can do
now to make it the case that he turned off the bathwater before he left on his
vacation last week, but it does not stop him from hoping he did. This is not
only true for hopes, though, but for any desire about the past: 'I wish that I
didn't sell my Munch in 1967', 'I want it to be the case that I didn't drink
that last bottle of wine yesterday.' As his account stands, Smith cannot cope
with such objections as these, and he cannot just drop the dispositional claim
either - that a desire disposes the object to bring it about that p -
because this would leave the account saying only that a desire that p is
a state that tends to endure on a perception that not-p', which would
make desire, as Schueler says, "a form of irrational disbelief."[83]
Hume includes hope among the direct
passions together with desire, aversion, fear, despair and security, passions
that arise directly from good or evil, pain or pleasure. The indirect passions,
passions that arise only indirectly through other qualities from good or evil,
pain or pleasure, are, according to Hume: pride, humility, ambition, vanity,
love, hatred, envy, pity, malice and generosity. We might add several more on
both types. We may obviously divide between positive and negative
attitudes. Now, it must be recognized that desire and aversion in the sense I
have used them, can be distinguished from the other passions we might call emotions.
The above argument against the direction of fit definition by Schueler plays on
the fact that some emotions like hope and fear (they are complimentary: if I hope
that a, I fear that not-a) are forward-looking, while others (most) are
backward-looking. Gordon (1987) has remarked that to distinguish emotions into
these two clases does not hold, since no emotion is exclusively forward- or
backward-looking.[84]
Instead, Gordon wants a distinction made between factive emotions and epistemic
emotions.[85] If a person
is glad or unhappy that p then he knows that p, whereas a
person hopes or fears that p only if he does not know that p.
Thus, a person emotes that p iff he knows that p;
i.e., 'A emotes that p' presupposes 'A knows that p'.
On the other hand, a person emotes that p iff he does not
know that p. Gordon argues further for this distinction, arguments we
need not go into here. It is sufficient for us to note the following. When A
is delighted or angry about the fact that p, A already knows it
to be the case that p. This leaves A with no possibility of instrumental
or aversive action, like he would if he desired that p (or not-p).
This makes the possibility of subsuming emotions under desires difficult. Most
obviously, as Gordon notes, where the object phrase is in the past tense the
verb 'desire' gives us a deviant utterance: 'I desire that I didn't sell my
Munch in 1967'. For this and other reasons, he says, it is preferable to speak
of 'wishing' something to be so. When A is pleased (etc.) about the fact
that p, then A wishes it to be the case that p; when A
is angry (etc.) about the fact that p, A wishes it not to be the case
that p.
We can now put the factive-epistemic
condition and the negative-positive distinction made above together. We have it
then that when A is uncertain whether p or not-p
(epistemic), he fears, etc. (negative) when he wishes that not-p, and he
hopes, etc. (positive) when he wishes that p. When A is certain
(and knows) that p (factive), he is angry, etc. (negative) when he
wishes that not-p, and he is delighted (positive) when he wishes that p.
Epistemic emotions like fears, hopes and worries that p (or not-p)
can be said to be confirmed or borne out when there come to be
strong reasons to believe that p (or not-p), something that we
cannot say of factive emotions. There are no conditions under which one's anger
or regret that p (or not-p) may be confirmed or borne out. In
this respect, expressions like 'fear' or 'hope' differ from expressions like
'anger' or 'regret' and resemble 'believes', 'suspects', and other epistemic
or cognitive predicates. This should be enough for us to see the
cognitive element mentioned by Schueler, and to carefully distinguish emotions
from desires and beliefs. Schueler's critique hits target only if this
distinction is not made.
We have seen in this chapter that the
direction of fit metaphor is quite fit to account for the distinction between
desires and beliefs as long as a certain mutatis mutandis objection is
handled with, and the distinction between desires and emotions is made.
We shall take a
closer look at the concept of 'desire', and see how broad it must be understood
in general. The concept of 'desire' as I use it includes, first, what we
ordinary call desires or wants: appetites, hungerings, cravings,
yearnings, urges, wants, and would likes.[86]
Second, we must include, like Davidson says, "a great variety of moral
views, aesthetic principles, economic prejudices, social
conventions, and public and private goals and values".[87]
What makes such different attitudes all desires, according to Davidson (1963,
1980), is that "they can be interpreted as attitudes of an agent directed
towards actions of a certain kind."[88]
Later (1987) he has given the same point somewhat differently, that a desire
(or pro-attitude) is "a disposition to act under specified conditions in
specific ways."[89]
This latter definition is similar to the one given by Smith (1987) above, and
presumably takes into account the difficulties connected to the earlier version
of desires.
According
to the above view, for any intended action we must attribute a desire to the
agent for some goal.[90]
Schueler (1995) acknowledges this point, but he says that it "entails that
there are at least two distinct senses (or uses) of the term
"desire."[91]
The two senses comes to the fore when we ask ourselves whether it is possible
for someone to intentionally do something that one has no (proper)
desire whatsoever to do. In one sense, this is simply not possible. In another
sense, however, there is nothing mysterious about people doing things they have
no desire to do. Remember George intending to attend the school meeting,
and therefore have a desire to attend the school meeting, while really wanting
to stay home. The distinction is very important for Schueler (1995) since it is
"one of the main conclusions of [his] book".[92]
He analyses the situation above as follows:[93]
On the one
side, when I deliberate about what to do, I have (as it seems to me) a positive
desire to stay home and read. The book I am reading is interesting; the chair
by the fire is comfortable; and so on. On the other side, I have no desire (or
at least none that I can detect) to go to the meeting. It involves going out on
a cold nigth, driving to the school, sitting through lots of boring talk, not
reading the book I want to read, etc. The only argument for going, I believe,
is that parents have a responsibility to attend such meetings and I am a
parent. But this argument, I decide, is the stronger one. So I decide in the
end to go to the meeting. (Had I come to believe that I didn't have a
responsibility to go after all, I would have decided to stay home.)[94]
The fact that
George wants to do what he has responsibility to do, doesn't (or at
least needn't) come into the content of his reasoning at all. Perhaps, says
Schueler, can this only be deduced retrospectively from the fact that in
situations like this, George give considerations of responsibility a lot of
weight in his deliberations. Even if George is aware of a desire to do what is
his responsibility, this desire need not come into his reasoning in this
situation. That is, it is not necessarily the fact that he wants to do
what he has a responsibility to do that gives him a reason to go to the
meeting, but rather that he has this responsibility that he believes
gives him reason to go. This is what he weigh up when deliberating.
That is, in the
course of my practical deliberation here and elsewhere, I don't usually weigh
up my own desires and then do what I think I have the strongest desire to do. I
weigh up [. . .] whatever seems to me to provide reasons for or against
doing one or another of the alternatives I think I have.[95]
So if we are not
prepared to hold that reasoning of this sort always involve a kind of
systematic delusion about one's own desires, says Schueler, and if we are to
make sense of the content of George's deliberation in the situation described,
we have to do so using a sense of "desire" in which it is possible
for George to do things he has no desire to do. What we need is a distinction,
according to Schueler, that can accomodate this difference.
The terms
"desire" and "want" [. . .] are, when simply used without
further explanation, ambiguous between pro attitudes and desires
proper, the latter of which will presumably include such things as
cravings, urges, wishes, hopes, yens, and the like, as well as at least some
motivated desires, but not such things as moral or political beliefs that could
appear in practical deliberation as arguing against the dictates of ones urges,
cravings, or wishes.[96]
A pro attitude
can be describes as a kind of placeholder which has the only function of
indicating a motivational direction, and no psychological state of any
kind. If it can be said to refer to anything it is whatever led the
agent to perform the action he did, something that could just as easily be a
moral belief as a craving for chocolate. Schueler regards as analytic that if A
intentionally performs action j under description d,
A has a pro attitude towards performing j under description d.
This allow us to reserve the term 'desire poper' for cases where it is possible
for somone intentionally to perform some action j under description d
without having any desire, in this sense, to perform j under d. [97]
Now, according to Schueler, the question whether a particular use of 'desire'
or any other related expression refers to a pro attitude or a desire proper, if
there in fact is such a distinction, turns on the answer to the question of
whether it would be possible to perform the action in question without having a
desire in the sense used. In the case of pro attitudes the answer we are left
with is that they are just indicators of motivational direction and nothing
more.
What is
the relation between pro attitudes and desires proper? One place Schueler says:
"It might be thought that there will never be any cases of pro attitudes
without some sort of desire proper also present,"[98]
while he at another place says: "On [Dretske's (1988)] account then, there
seems to be no room for the possibility of an agent having a pro attitude that
is not a proper desire after all."[99]
In the first sentence, he denies that there must always be a desire
proper in addition to a pro attitude. What he denies then is that whenever
there is a pro attitude to j, there must also
be a desire proper to j in addition. The
second sentence indicates that pro attitudes are to desires proper as genus to
species. We find evidence for this interpretation in a footnote where he says
that "the term "pro attitude" is intended to cover much more
than desires proper."[100]
In short, when
I say that since someone did something intentionally, he or she must have
wanted to do it, the term "want" as used here, if it refers to
anything, refers to whatever led the agent to perform that action, but this
could just as easily be a moral belief as a craving for sweets.[101]
The overall aim
of Schueler's distinction can be seen in the following:
[…] there is
just no way that a deliberating agent can consider his or her pro
attitudes as pro attitudes in his or her rational deliberation about
what to do.
[…] from the
deliberating agent's point of view, his or her own pro attitudes are
"invisible".[102]
This means that
when we consider in practical reasoning (or deliberation) reasons for or
against some action, what we consider is either some desire proper or some
moral, political, social, etc., belief that function as a reason:
[T]here is no
conflict between the idea that a pro attitude is always involved in the
explanation of intentional action and the idea that a belief alone might have
led to an action, since at least some beliefs (e.g., moral views, aesthetic
principles, economic prejudices, social conventions, etc.), if they move people
to act, would count as pro attitudes.[103]
The pro attitude we
analytically attribute being a reason for some intentional action is not
necessarily considered qua desire (proper) in practical reasoning, but
it may. So the philosopher's sense of desire in explanation of action does not
always include a desire proper that can be treated qua desire in
practical reasoning.
The
conflation of these two senses of desire creates problems in both explanation
of action and practical reasoning, according to Schueler. He uses this
distinction to go against the view that desires are somehow essential in the
rational justification of actions, and against the view that sees desires as
essential for the explanation of actions. He argues for a sharp distinction
between justifying and explanatory reasons - the first figuring in practical
reasoning, the other in explanation of action - and claims against among others
Williams (1979; 1989) that a desire-element in addition to the justifying
reason itself is not needed in practical reasoning. The claim that having a
desire of some sort is a necessary condition for having a justifying reason to
perform an action will only seem plausible if we are using the pro-attitude
sense of 'desire'. Schueler also argues against Goldman's (1970) and Dretske's
(1988) accounts of desires qua causally operative states. This he does
on the basis of one claim:
If it is the
deliberative process that explains the resulting action, then it is not the
desire (proper), with its alleged causal powers, that is essential to the
explanation but rather the agent's awareness of it, that is, the agent's belief
or judgement that he or she has this desire.[104]
According to this,
Goldman's causal account fails because he sees desires as being involved in two
radically different explanatory schemes.
That is, he
tries to see desires both as things about which an agent can deliberate
(and hence as things that can be ignored, things about which the agent can be
mistaken, and the like, in short, what I have called desires proper) and
as the basic explanatory factors (possibly even, as Goldman holds, causal ones)
in explanations or predictions of behavior (that is, as what I have called pro
attitudes).[105]
Dretske's account
fails, according to Schueler, because he is committed to the reward thesis -
that all purposive behaviour is undertaken to get some reward or pleasure - and
there seems to be no room for the possibility of an agent having a pro attitude
that is not a proper desire at all. He cannot then account for practical
reasoning.
Schueler concludes his book by
distinguishing between a direct and a reflective model of
explanation of action. The first is susceptible to a causal analysis like
Dretske's blind-force model, while the second can be given by a deliberative
model runned by the principle of practical rationality.[106]
In principle, this is a distinction that is the same as Pettit & Smith
(1990) holds, between desires in the foreground and desires in the background.
We
shall return to some of Schueler's claims later. To conclude this section, I
think we must see desires as a broad category ranging over attitudes that shows
us seeing some goal in a favourable light. As such I follow Schueler and
Davidson, and others. This does not mean that we must go further with Schueler
and conclude that we should distinguish between a deliberative conception,
concerned with practical reasoning and desires proper, and an explanatory one,
concerned with reasons for actions and pro attitudes.
1.14 The Nature of Desire and Belief.
We have seen above
that the nature of desire must lay not in its propositional content, nor in its
phenomenal qualities, but in its functional role - having a certain direction
of fit - disposing the subject to bring it about that p on a perception
(or belief) that not-p.[107]
The teleological argument claims that a desire is a desire iff (a) it is
a state with which the world must fit, and (b) it disposes the agent to act so
that the world fit the desired goal. This determines both the nature of desire
and how a desire may function as a reason for action.
Are there any ways for a belief to
slip in here and serve as the motivating source? Not qua belief. But if we
consider not only beliefs that represents the way how things are, but
how they should be, can such a belief slip in? It seems to me it can be
thought of as being a state with which the world must fit, but does it dispose
the agent to act so as to make the world fit? There seems to be no real way to
answer this question unless some action is performed, and the motivating
element may be deemed a desire because of its (by now) evident disposing power.
There seems thus to be no way we can account for the deliberating agent's
options between which he chooses; whether it is a desire or a belief. Schueler
(1995) thinks we can properly decide between desires and beliefs, between what
we desire (properly) to be the case and what we believe should be the case, when
we reason practically. Only after the agent has acted may we say that he had a
pro attitude for what he believed should be the case, if that is what he
choosed. Considering this, Schueler criticises Williams (1979) internalism
because he cannot distinguish it from externalism.[108]
The internal account of reasons is an account of what it is for an agent to
have a justifying reason, says Schueler, and its internality comes from the
fact that to be a reason for some agent at all it must be "derived from
something in the agent that is capable of moving the agent to perform the
action for which it is a reason."[109]
What this means, he says next, is that "the intentional account of reasons
can make no use at all of pro attitudes."[110]
Why not? According to Schueler, it is because,
pro attitudes
cannot possibly be the motivating factors that the internal account claims are
required for something actually to be a reason for someone to do something if
we want to maintain that there is a difference between an internal and an
external account of justifying reasons.[111]
The externalist
claim is that a belief might serve as motivating an agent without the agent
having any prior motivation for the content of the belief. It must be remarked
that Williams himself does not think we should distinguish between internal and
external accounts of reasons, but rather that there is in fact no possibility
that we may have an external account of reasons for action. He claims that
whenever we act for a reason, the reason statement must display a relativity to
"the agent's subjective motivational set," which "can
contain such things as dispositions of evaluation, patterns of emotional
reaction, personal loyalties, and various projects, as they may be abstractly
called, embodying commitment of the agent."[112]
What this relativity is, we shall be concerned with later. I cannot see why pro
attitudes does not work with internalism, however. If we describe both the
motivation set of the agent and the reason relative to this set as pro
attitudes (something I think we should do) why cannot this hold externalism
away?[113] Schueler
may give us an answer in the way he thinks of pro attitudes:
The fact that
I don't feel any sympathy for the hungry people in Afrika (or more
generally, the fact that I can't detect in myself any motive that might lead me
to try to help them) in no way shows that I have no pro attitude toward
helping them, since anything that could lead me intentionally to help
them, including the most complex Kantian arguments, would thereby demonstrate
that I had a pro attitude of the required sort after all. And merely surveying
my emotions, desires and other internal states (including my evaluative
beliefs), we cannot rule out as pro attitudes all the possible philosophical
arguments about justice, altruism, etc., that I might read or hear or think up.[114]
The point is that a
reason cannot motivate me to act unless there is something about it that makes
up my pro attitude against it. The externalists are closed out on this
"pro-attitude reading of the internal-reasons account"[115]
just because I do not act on every philosophical argument I read or hear to be
good. Such arguments may motivate me just in case it make me see it as
good, i.e., if they are internalised.
I am
inclined to think that the way we have defined desires above, through the
teleological argument, is sufficient to deal with the present problem. The way
we are disposed makes it evident that the desire (or pro attitude) is
internalised. As such the ground for a distinction between desires proper and
pro attitudes falls away. However, if one wants some principle to distinguish a
belief that p is good (for instance following a complex Kantian
argument) from a desire that p - possibly expressed as its being good
that it be the case that p - we could on some suggestions from Stampe
(1987) make a logical filter to decide whether a given reason is a desire or a
belief. Stampe asks what it is that is lacking in a person who simply has no
desire that p - e.g., that he takes a new job - although he thinks, even
knows, that it would be a good thing were he to do so? According to Stampe, one
is tempted to describe the difference in some way such as this:
in desire, one
is somehow struck by, affected by, the merits of the thing wanted, or
the prospect of having it, in a way one needn't be if one merely knows it would
be good; so one may say, "I know it would be best to take that job, but
the prospect of doing so simply leaves me cold, does not strike me as something
I want to do; it somehow doesn't seem to me the good prospect I know it
is." Sometimes one is thus affected through a feeling, as in thirst; sometimes
not. In either case, there is something characteristic of desire: if one wants
a thing it seems to one as if the thing wanted would be good. This is
not necessarily the case when one merely believes (or knows) that it would be a
good thing: knows it, say, because one has been reliably told it, or because
one has reasoned it out. For in that case, there is, I think, a sense inwhich
it need not seem to one at all a good thing, and thus one may say,
"Although it doesn't seem so to me, it is my belief that it would
be good to do it" (just as one might say, consistently, "Although it
doesn't seem to me to be so, the President is a great man"). But
one may not say, "I want it, though it in no way seems to me as if it
would be good to have it"; if one wants the thing, it must in some way seem
good, when one wants it.[116]
We can then,
although before we have defined 'seem', formulate a logical filter with this
concept. Negatively we would have it that for any propositional attitude P,
if it is true that A P's that a and it does not necessarily seem
to A that it would be good that a, then P is not a
desire. Positively we would have it that:
For any
propositional attitude P, P is a desire if A P's that a and it necessarily
seems to A that it would be good that a.
One could object to
this that one may certainly express a belief so that we have for instance 'It seems
to me that it would be good that a.' Stampe dismisses
this objection on the ground that 'seem' here have only a parenthetical use
that need not refer to the way things genuinely seem to the speaker. This is
only true if 'seem' is properly defined, however, something it is not yet. Let
us for the moment take it for granted that it is properly defined and see how
Schueler's distinction between desires proper and pro attitudes comes out on
this desire-filter.
Schueler found there is a difference
between George's desire to stay home and his obligation to attend a school
meeting concerning his son. Nevertheless, he acts on his obligation, and thus
have a pro attitude for going to the meeting. The sense in which he desires the
first option is very different from the second option, says Schueler, because
in his reasoning (if there is a such one) he does not weigh between two
desires, a desire to stay home, and a desire to attend the school meeting. What
he weighs between is a desire (proper) to stay home and read, and an obligation
to his son. Only after the attending of the school meeting have happend may we
talk of a desire (pro attitude) of George to attend the school meeting. The
claim seems like a Davidsonian, and also one I hold, but is not properly such,
because in the deliberation of George, on Schueler's account, what is
entertained (not necessarily conscious) is not a pro attitude for any
obligation to his son, but just a belief that he should attend the
school meeting. As such, Schueler comes down as a Nagelian rationalist.[117]
This seems to fit badly into the above filter, since for Schueler it is not
necessary that something seems good about an option to choose it; what
is chosen between is not two desires, but one desire and an obligation qua
belief, i.e., as a belief that something is good about the option. Now, if the
filter above is right, we have a case against an analysis of desires like the
one given by Schueler. This is however dependent upon how we define 'seem'. We
shall see how this might be done next. As I said above, I am inclined to
believe that we have a case against Schueler's distinction because of the
teleological argument alone, so we may only see the above filter as confirming
my point of view.
Unfortunately,
Stampe's definition of 'seem' fails. He thinks desiring is analogous to perceiving.
It is a special perceiving, however, because although it is (wrongly, I think)
seen as conscious or a form of awareness on Stampe's account, it is
non-sensuous and whereas what you see or hear or feel is the very thing you
perceive, what you desire cannot be identified, in general, with the thing you
are perceiving. It is enough I think to dismiss this further theory of Stampe
on the ground that he needs the concept of awareness to have his analogue with
perception work. I think we should dismiss then also his concept of 'seem' if
it is defined as a perceiving. But I don't think we should dismiss the concept
'seem' itself, because it may work, as I have shown above, as a filter
distinguishing between such beliefs that something is good and desires. We thus
need some other way to define this concept. One way to go might be to define it
along some internalist lines, as something relative to a given set of
motivations, like Williams (1979) does. This is perhaps difficult because it
seems to keep out appetites as possible seemings. I will not, however, provide
any further development of 'seem' here, but leave that for another time.
Above we have
considered explanation, events, actions, intentions, what it means having a
reason for action, as well as the nature of desires. Now we shall consider an
account that brings us to see the close connection between having reasons for
action and practical reasoning.
2.1 Practical Reason and Introduction to the Dispute
Between Humeans and Rationalists.
Above, we have given an account of explanation of
action according to desire and belief, and thereby in some sense explicated the
nature of beliefs and especially desires. The view that desires must stand
behind every action is not undisputed, however. One question concerns the role
of especially moral reasons in this picture: need we have a desire for doing
anything we feel morally obliged to do? Accounts of practical Reason or
reasoning divide on this central question that has been much debated in recent
philosophy.[118] The proponents fall roughly into rationalist
(or anti-Humean) and Humean camps. Thus, connected with the subject of
practical Reason - the faculty of argument and demonstrative inference
considered in its role of prescribing or selecting conduct (or action) - is
what role, if any, practical Reason plays in determining norms of
conduct.
Traditionally,
the faculty (or capacity) of Reason has been divided into among others the
faculty of theoretical Reason and practical Reason. Aristotle identifies
mathematics, theology and physics as the subject matter of theoretical
Reason; practical Reason on the other hand is concerned with
determining guides to good conduct and deliberating about proper courses of
action.[119] According
to Hume, the faculty of Reason is divided in its practice between theoretical
Reason and practical Reason according to content or subject matter:
logical and mathematical in the one respect, empirical relations in the other.
For
Kant the faculty of Reason in a broad sense is the faculty of Judgement. In the
narrow sense of Reason it is distinguished from the faculty of Understanding.
Both consist in the capacity to judge, but Understanding is the faculty of
distinct cognition and mediate judgement, and Reason (in the narrow sense) is
the faculty of syllogistic reasoning by which one draws inference. The goal of
theoretical Reason is explanatory completeness and unconditionedness of being -
transcending what is possible in experience (Understanding). It breaks all
limits and are as such allied with freedom, but it is nevertheless limited by
'what is'. Practical Reason, on the other hand, is free in both respects, and
determines what 'ought to be'.
Reason,
as a faculty or capacity, may be regarded as a hybrid composed (broadly
construed) of theoretical and practical Reason, or as a unity having both
theoretical and practical functions. Aristotle can be seen to embrace the
former conception, while Hume and Kant the latter, in that they talk of the
practical or theoretical employment of Reason. Reason is broadly contrasted
sometimes with experience (Understanding), sometimes with emotion and passion
(desire), sometimes with faith. Its presence has been regarded by many as the
primary difference between human and non-human animals.
Humeanism is recognized in
Hume's claim that practical Reason is and ought only to be the slave of the
passions. This is because practical Reason by itself is incapable of
influencing action directly; although it can do so indirectly by
disclosing facts that arouse motivational impulses, and by discerning means-end
relations by which our goals may be attained. But none of these goals are set
by Reason; all are set by the passions - the deciderative and aversive impulses
aroused in us by what our cognitive faculties apprehend.
It does not follow from this view per
se that ethical motivation reduces to mere desire and aversion, based on
the pleasure and pain different courses of action might afford. One could argue
for the existence of specifically ethical passions, or that independently based
moral injunctions have in themselves a special capacity to provoke ordinary
desire and aversion. Nevertheless, instrumentalism is often associated with the
view that pleasure and pain, happiness and unhappiness, are the sole objects of
value and disvalue, and therefore the only possible motivations of conduct.
Hence, it is claimed, moral injunctions must be grounded in these motives, and
practical Reason is of interest only as subordinated to inclination.
Rationalism stands as an
alternative to the above instrumentalism, a view historically connected to
Kant, that practical Reason is an autonomous source of normative principles,
capable of motivating behaviour independently of ordinary desire and aversion.
On this view it is the passions that lack intrinsic moral import, and the
function of practical Reason is to limit their motivational role by formulating
universalizable normative principles founded upon the operation of the
practical Reason itself.
Practical reasoning
is the inferential process - constituting the role we call practical Reason -
by which considerations for or against envisioned courses of action are brought
to bear on the formation and execution of intention. There are several
important issues concerning practical reasoning, including how it can be
evaluated, how it relates to theoretical reasoning, whether it is a causal
process, etc. These issues I think is wholly or partly answered by
investigating the nature of the concept itself. This is not so easy since it
has been applied by many philosophers with many different meanings. Other
concepts such as 'practical syllogism', practical thinking', 'practical
inference' and 'deliberation' have all been meant as referring to practical
reasoning. First then, we must establish what an account of practical reasoning
must capture. Audi (1982) has provided us with a simple account of this.
Central to our account is the
following questions related to practical reasoning: (a) to explain what it is
for an agent A to act for a reason, (b) to specify a structure in terms
of which intentional actions can be explained, (c) to exhibit the structure of
acting for a reason so that such actions can be seen to follow from, in some
sense, or at least to be reasonable in the light of, the reasons. Following
these questions is another important one (d) whether acting for a reason is equivalent
to action based on practical reasoning. If this is the case then A j's for a reason iff
A's j-ing is based on
practical reasoning. Below, I will argue, with Audi, and in agreement with
Davidson (1978, 1980) that all we can claim is that there is a correspondence
between acting for a reason and acting on the basis of practical reasoning;
"if someone acts with an intention, he must have attitudes and beliefs
from which, had he been aware of them and had the time, he could have reasoned
that his action was desirable".[120]
Two assumptions and intuitions of
mine will be presupposed in the following discussion, first, that a piece of
reasoning must have a conclusion, and second, that drawing the conclusion in
practical reasoning must be an appropriate response to the practical question
"What should I do?".[121]
Another main question we shall find answer to in what follows is what must be
added to some piece of practical reasoning to make it not only practical in
subject, but also in issue? Intuitively I think there is but one answer to
this, and that is 'nothing', if we talk philosophy, not psychology. In other
words, the agent in question can give no further reasons why some piece of
practical reasoning ended up in action and not another one. We shall see
whether my intuitions are correct.
Audi
(1982) notes an important distinction between two uses of the concept
'practical reasoning': practical reasoning
and practical argument; i.e., between a process and the
corresponding abstract structure to which we directly refer when we
logically appraise someone's practical argument (or reasoning) as valid or
invalid. A related distinction is between the conclusion of a person A's
argument (the judgement that A should j) and A's concluding that argument (by judging,
on the basis of the premises, that he should j). It is essential that we distinguish between a
practical argument as a structure of propositions, and the process of passing,
from the premises of such a structure to its conclusion.
Audi uses the following terms to
distinguish these elements: 'practical argument' is certain structures of
propositions that can be deemed valid or invalid; and 'piece of practical
reasoning' is an instance of the corresponding process, i.e. appropriate
tokenings of such structures. Accordingly, we must distinguish between the
conclusion of a practical argument - a proposition (a judgement) - and
what corresponds to it in A's reasoning - his concluding that reasoning,
by inferring the conclusion from the premises (an instance of judging).
The conclusion (a proposition) of A's practical argument is A's
answer to the relevant practical question. His drawing that
conclusion is his (specific) response to that question. His general
response is the entire piece of reasoning, of which the drawing of the
conclusion is the terminating element.
2.3 The Conclusion of Practical Reasoning.
According to
Aristotle (at one point at least), practical reasoning is as follows:
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 make a coat. And the conclusion 'I must make a coat' is an
action. And the action goes back to a starting point. If there to be a coat,
there must first be this, and if this then this - and straightaway he does
this. Now that the action is the conclusion is clear.[122]
This kind of view
(Aristotle's considered meaning or not) we may call the action-as-conslusion
view. Another view that is also attributed to Aristotle is that the
conclusion of a practical syllogism is a practical judgement.[123]
Anscombe (1963) thinks this, as well as Audi. I agree with and will argue for
this latter way of seeing the conclusion of a piece of practical reasoning,
what we may call the judgement-as-conclusion view. Let us first consider
the action-as-conclusion view.
Does it constitute a plausible sufficient
condition for A's completing a piece of practical reasoning that it is
an action? Consider the following example from Audi (1982):
Suppose that
it occurs to [A] that he wants some gin, and he has the thought that he
must open the new bottle to get some. He might conclude: I should open it. But
imagine that [A] does not infer this. Instead, he resolutely adopts the
belief that he should not open it. If, being weak willed, he opens it, should
we say he completed his practical reasoning by doing so?[124]
What seems to me is
that he avoided drawing the conclusion, that is, completing the reasoning,
because if he completed it one should think, if he is rational, that he would
arrive at the judgement that he ought to open the bottle. On the
action-as-conclusion view, however, we must, since A did what the
premises favoured, say that he drew the right conclusion after all. But this is
surely implausible as Audi remarks, because although A may have opened
the bottle for a reason, his want for gin, it is certainly not clear that if A
does something for a reason expressed in practical premises which he
accepts, it follows that he has thereby concluded a piece of practical reasoning.
This is not sufficient to prove that actions are not the element with which
practical reasoning concludes, but some further considerations may confirm the
impression that the action-as-conclusion view is misguided.
Consider cases in which some agent A's
acting is not necessary for his concluding a piece of practical
reasoning. Again Audi provides us with a fine example. Suppose A says,
"I want to surprise Sue; I could do it best by bringing her live lobsters;
so that's what I should do." He has certainly completed his reasoning
whether he brings Sue lobsters or not. One initial objection to this would be
to say that what A does is to decide to bring Sue lobsters, so
that the act of decision is the conslusion of the practical reasoning.
Then it is an act anyway. Confronted with this objection Audi holds that surely
one can complete ones practical reasoning without even deciding to do
something, as one can just judge that one should do something. I believe
that when we talk of the doings of practical reasoning, rather than the form
practical arguments must take, we must say that judging is as much of an
act as deciding and the actual bringing of the lobsters.[125]
Actually I think Audi agrees with this, he just presupposes that the
acting-as-conclusion view claims that actions are movement of the body. Only if
this is the case does his arguments work as intended. I will take it that this
is the case, because it is a simple and plausible way to bring forth the
arguments I agree with. We may thus make the following distinction, between the
concluding of a practical reasoning and the acting on a practical
reasoning. It seems to me that this brings forth Audi's point, and the fact
that it is one thing to conclude a practical reasoning, another thing to act on
the practical reasoning. When we conclude we do not seem to act on the
practical reasoning, we just conclude it. This is evident when we
consider further the following example from Audi about the logic of certain
questions of reasoning:
Consider the
question "What reasoning of S's justifies his A-ing?"
This seems to presuppose that the reasoning, as providing justification, is
separable from A. If A were its concluding element, the above
question would be roughly equivalent to "What justifying premises did S
hold on the basis of which he A-ed?" It is not. For one thing, the
former presupposes that S reasoned; the latter does not. It does
apparently presuppose that S acted for a reason; but surely that does
not entail acting out of a piece of reasoning. Moreover, it makes clear sense
to say that S and S' went through exactly similar practical
reasonings, yet S A-ed and S' - say, through weakness of
will - did not. If the reasonings are exactly similar, they cannot have
different sorts of conclusions.[126]
Now, we must halt
for a minute. Audi (1986) thinks he in Audi (1982) provides a defence of the
claim mentioned above, that I agree to, that not all actions for a reason are
based on practical reasoning. In the account I have defended above - an account
based on Audi (1982) - this claim has not been defended but rather presupposed.
In what follows there will be put forth a conception that in fact defends this
presupposition, but this presupposes the above initial distinctions and claims.
I do not think the argument circular for this reason, since one can surely
conclude from conditionals.
If the above account is correct,
actions (considered as movement of the body) are not the conclusions of pieces
of practical reasoning. I mentioned above that I presuppose that drawing a
conlusion of practical reasoning must be an appropriate response to the
practical question "What should I do?". According to this, then,
drawing the conclusion of a practical reasoning can be seen as answering that
question. Given this, the best candidate for an answer to this question is the
agent A's forming a practical judgement in favour of j, i.e., that he should
j. We must, however,
construe the practical judgement as allowing that A may form one without
accepting it, that is, without putting all his weight behind it.
Thinking about Davidson's conception of practical reasoning, we should
accomodate with our conception such relativized judgements as "Given my
information (r), j-ing is the best
course." This means that we include those judgements favouring j such that normally
A's accepting them on the basis of the premises can explain his j-ing, and can bring
about his j-ing straightaway.
So, all this is consistent with the Aristotelian idea that successful practical
reasoning concludes in action, just as theoretical reasoning typically
concludes in belief. But as Audi claims: "what concludes in action need
not have action as its conclusion."[127]
Accordingly, then, we have the
following conception of practical reasoning, well put in Audi (1982), that,
there are
practical arguments as structures of propositions; pieces of practical
reasoning as appropriate tokenings of such structures; and actions, decisions,
intentions, etc. based on practical reasoning. On this view, the conclusion of
a practical argument is an appropriate answer to a practical question; [A]'s
drawing that conclusion is an appropriate response to his asking a practical
question; and his doing the thing endorsed by this conclusion is, for him, an
appropriate solution to the problem motivating the question.[128]
Consider the claim
from Audi that practical reasoning is an appropriate response to the question
'What shall I do?' In what contexts do we ask ourselves that question?
Typically in two contexts: first, thinking we should, in some sense of morally
good, etc., make it the case that p; and second, when confronted with a
more typical (proper) desire that p. Do we reason practically when
confronted with an aversion that p? Gordon (1987) uses many
examples in which aversions motivate action. One is the following:
Mary is aversive
that her wedding receive publicity.
Mary believes
getting married in Mexico could prevent the wedding from receiving publicity.
Mary judges that
she should marry in Mexico.
I think we should
allow aversions motivate actions in the present account. Davidson (1963) seems
to indicate that only pro-attitudes, not con-attitudes be
included. The problem is that aversion disposes us not to act so that
the desired effect will come true, rather not doing it. To mend this
proplem we could just describe an aversion as a desire that not-…. Mary's
aversion would then be that she desires that her wedding will not receive
publicity.
2.4 The Premises of Practical Reasoning.
Having (mostly)
considered the concluding element of practical reasoning, we shall now consider
the other elements, the premises. We should construe the major premise in
practical reasoning, as expressing some desire of A, in the broad
sense of desire, designating any pro-attitude by A for some goal
that is the content of the desire. What is essential is that the content of the
expressed desire can be taken to imply the presence of motivation. But since
one need not necessarily act on one's conclusion of a practical reasoning, the
motivation expressed in the major must not be seen as implying that the agent
will necessarily act on his desire. Even further, to take an example of
Audi (1982), if A says, "I've got an obligation to help Peter and
have to j to help him, so I
had better j", this should
be taken as practical reasoning, according to Audi, even if A in no way
wants to fulfill the obligation (or that it seems to A to be good to
fulfill the obligation). Therefore, though we should not expect A to j, he may
nevertheless think the premises right. "In any case," says Audi,
"it seems reasonable to allow, as a limiting case of practical reasoning,
the counterpart of theoretical reasoning from suppositions to further
suppositions."[129]
This is a very important observation, but it is questionable. We shall consider
this claim later, to see whether we may really go that far.
Whatever
the answer to the above question is, the major premise of a practical reasoning
needs not express what the agent desires most, the preference of
the agent, or the strongest desire. Many writers on practical reasoning
think that A must desire some a most, that he must think he
wants a on balance, i.e.,
the desire for a is a stronger
desire than the incompatible desire for a'. But, again A may do
practical reasoning as a result of wanting to realise a even if he does
not want it on balance or want it most; his purpose may be to savour the
prospect of realising a. A might
also engage in practical reasoning when, unable to decide between two
incompatible options which he desires equally much – walking in the mountains or
sunbathing at the beach – he constructs a practical argument regarding each,
hoping thereby to establish a preference by putting each means-end prospect
before him.
There
has been proposed that if the major premise expresses a want then it must
contain a characterization of what is wanted which shows its
desirability. This suggestion is countered by the following example: suppose A
reasons that he wants some poisonous berries, and to get some he must go into
the woods, so he'll do that.[130]
Surely, the major premise does not show the berries desirability. If we talk
about justified desire, however, we could deem the desire not to show it's
desirability, but then again, that is not what we are concerned with here. This
point does not go against a description of desires as 'it seems good to A
that a', because there
may be something good that A sees about the berries. The present
conception does not explicate in what respect A desires a, that is, what
property about a that he finds
desirable, and thus what actions he finds desirable. This problem was noted in
connection with Davidson's mature account on desires above, and I will not
consider it any more here. I think, as I have said above, the present account
is compatible with this.
The
minor premise is generally taken to be to the effect that j-ing is necessary
to realise a, sufficient
to realise it, or might realise it. The premise may be that (a) getting a requires j-ing, (b) that by j-ing A will
realise a, (c) that A's
j-ing is a good way
to bring about a, (d) that j-ing will
contribute to a, (e) that j-ing is identical
with a, etc. The last
possibility mentioned (e) is a case where the desire is for the action itself,
a limiting case. The minor need not say that j-ing will definitely realise a; reasoning may
surely count as practical even if there is only the weakest belief that the j-ing will succeed
to contribute to a.
Concerning
the conclusion I have already argued that we must take it as the formation of a
judgement. This judgement in turn causes or contributes to the formation of an
intention decision or action, etc. Many have argued that the conclusion is the
expression of an intention, however, and not just a judgement. We took pains
above to divide the conclusion from actions, and I think we must conclude the
same way in the case of intentions. This must still be true in cases where the
conclusion is an expression of intention or an apparent report of action; like:
"I'll do it!" or "So here goes!" We must think that
practical reasoning may be presented in a form that does not correctly display
its real structure, often elements are presented enthymematically. Davidson
often points this out when he claims that we need both beliefs and desires as
reasons for action, only that it is often awkward to mention both. Audi gives
the following example:
Wanting relief from an ache, S may say,
"The only way to get relief is to take this awful pill, so here
goes!" The structure of his reasoning is roughly: I want relief; to get it
I must take this awful pill; so I should take it. But S need not utter
"I want relief" or "I should take it." However, since
taking the pill is unpleasant, the exhortatory "Here goes" is natural,
and it might also serve to tell someone else that he is acting on his
reasoning.[131]
Still, the
practical upshot of practical reasoning is often the agent j-ing, deciding to j, or intending to j, and we must
concede even that this upshot is what the agent regards as solving the problem
generating the practical question to which his reasoning is directed. This,
however, should not force us to assimilate conclusion and upshot of the
reasoning. So we can allow the natural tendency we have to express a decision
to j, an intention to j or the actual j-ing as the
conclusion of the practical reasoning, but we must not forget that the
judgement that is the conclusion of the reasoning is presented
enthymematically.
Now we
are in a position to present a conception of what practical reasoning is. I
will roughly follow Audi (1982) in this. We have argued above that the
constituents of a piece of practical reasoning has minimally three
constituents: a motivational premise as major, a connecting premise (minor)
that relates the content of the motivational premise to action, and as
conclusion, the forming of a practical judgement favouring the action specified
by the connecting premise. Thus, major expresses a desire of the agent, minor a
belief of the agent that some action of his will contribute to the fulfilment
of this desire, and a practical judgement expressing the conclusion, and by
that connecting and confirm the belief and the desire of the premises. The
simplest basic schema for practical reasoning is then as follows:[132]
(Major) A desires that a.
(Minor) A believes that j-ing will
contribute to a.
(Conclusion) A judges that he should j.
Here a is a state of
affairs, ranging also over events. j is an action type
which A normally believes he can perform. In the minor I use 'j-ing will
contribute to a' instead of 'a is the property of
j-ing' or 'j-ing is a', because 'will
contribute to' is broader and have the other concepts as implicates. So that 'j is necessary for a' will be included,
as well as 'j is a way of
achieving a' rather than as a
means to it, like playing computer-games just for fun. The 'should' in the
conclusion does not only point to a specific moral or prudential force, but
should be taken to include any overall sense in which A thinks relevant.
It is thus a statement in favour of A's j-ing, to the effect that he
ought to j, that it would be
good to j, that it would be
desirable to j in the
circumstances, etc.
We must also allow, as Audi claims,
that practical reasoning can take more than two premises. This may happen when A
seeks a conclusive arguement for his j-ing.[133]
Further, practical reasoning may occur in chains of arguments, so that the
simple schema proposed represent only part of such reasoning.
As the above schema is a
representation of the structure of practical reasoning, a structure of
propositions (it is right that what is represented in the propositions is in
the premises a proposition related to a propositional attitude, and this again
is related to an agent, but what concerns us now is the propositions as
expressed in the structure above), then the tokening of this into pieces of
practical reasoning may be attributed to the reasoner himself, or to someone
else. Thus I think we must allow besides first-person accounts, both second-
and third-persons accounts of practical reasoning. This can come into conflict
with another claim given above that for any intentional action we must not
attribute a piece of practical reasoning. Thus, we must find some different
criteria other than that of the first-person answer to a question whether or
not he did some practical reasoning. Audi actually prefers the first-person
view on practical reasoning, because of the above reason I believe, although he
allows second- and third-persons accounts.
The above schema, as we have noted,
allows for a manifold of linguistic expressions, but the structure of reasoning
will correspond to the structure of the schema. Audi notes that,
the structure
will either have the specified form or embody the three crucial elements
[major, minor, and conclusion] in a form that represents a departure from it
[…], most notably by the addition of further premises or the conditionalization
of the conslusion (e.g., where [A] concludes that he should [j] if asked to).[134]
The above account
allows us perhaps to make a distinction between two types of practical
reasoning; one that allows the major premise to be a mere supposition that a, another that
requires the agent to desire it to be the case that a. We may distinguish the two
according to the dispositional criteria, so that the agent is disposed only in
the latter case to do so as to have it that a. The former case of practical
reasoning should be considered as an exploration of what one has a justifying
reason to do (and not motivated accordingly), etc. The expression of the major
premise of practical reasoning may be neutral to this distinction, while the
expression of a reason for action should not, I think, be neutral on this. This
latter is an internalist claim we shall discuss more later.
2.5 Practical and Theoretical Reasoning.
How are we to
account for the difference between practical and theoretical reasoning, now
that we have denied that practical reasoning is always practical in its issue,
understood as concluding in an action (of bodily movement)? First we must clear
away something that may contribute to a misunderstanding. One could think that
since some reasoning are practical and other theoretical, both cannot be
something we do. This is not right, we may do both. Still, however, we can
distinguish between different uses of the types of reasonings at hand.
If it is the use of the reasonings that distinguish between practical and theoretical
reasoning - that is, practical use of reasoning and theoretical use
of reasoning - we may think that it still is the issuing not the subject
that determines whether or not a piece of reasoning is practical or
theoretical. This is just what I want to maintain. The typical conclusion of
practical reasoning is the forming of a practical judgement. The proposition
judged may also be the conclusion of a theoretical argument, however, but
theoretical reasoning need not issue in the forming of a practical judgement.
Considering the premises, the subject (or content) of the premises of practical
reasoning must not have a content that is appropriate for practical reasoning,
but it may. As such, the distinction between the two types of reasoning is
mainly pragmatic. Logically, both kinds of reasoning may be taken to
express arguments, in the sense, very roughly, of propositional structures
whose premises are put forward as grounds for their conclusions. Indeed,
'practical argument', used for the propositional structures expressed in
practical reasoning, may be misleading. The typical uses of such
structures are practical; but the arguments themselves, as abstract structures,
are not intrinsically practical and can be used for theoretical purposes. So we
have no firm distinction between practical and theoretical reasoning along the
subject (or content) line, but a distinction is kept along lines of issue or use
of the reasoning in question; practical reasoning is reasoning done with a practical
purpose - to have an answer to the question "What shall I (you, s/he)
do?" This latter does not apply to theoretical reasoning. As Audi claims:
Even if [A]'s
conclusion in such reasoning that he should [j], his concluding this will not
contitute his responding to a practical question with a practical judgement,
and the reasoning will not have been undertaken to determine what to do. […] It
is one thing to undertake a piece of reasoning […] in order to determine what
to do. It is quite another thing to undertake it in order to determine what
it is true that one should do.[135]
When the judgement
is arrived at, there is a further difference between practical and theoretical
reasoning: what is implied in A's assenting to a conclusion of his
practical reasoning is normally (to say it weak) that he has some motivation to
act accordingly, something that is not implied in his assenting to a conclusion
of his theoretical reasoning. And further, this pragmatic point goes up the
reasoning-chain, so that when a desire is presented as issuing wondering what
to do, it is properly a desire. But this contradicts what was said above. The distinction
between finding out what to do and considering what it is true that one should
do coinsides with the distinction made between two types of practical
reasoning; a distinction actually building on a suggestion from Audi. Now,
certainly not all theoretical reasoning are finding out what it is true that
one should do. I think we should consider this also as practical reasoning, in
line with what was said above, only if the piece of reasoning that is a finding
out what it is true that one should do, is in a pragmatic context wherein the
agent is considering what to do. This does not make a clearcut distinction, but
it should be sufficient for the present purpose.
Having claimed that the major
difference between practical and theoretical reasoning is one of pragmatics,
there seems to be no limitation on what logic can be applied to both. We should
expect that practical arguments, like theoretical arguments, may or may not be
deductive, valid, inductively strong, etc. When confronted with this Audi claims
that he prefers to accommodate practical reasoning within assertoric logic.
This means that we need not represent practical reasoning as always valid. It
is also strange to think of practical reasoning as inductive, since there are
few if any examples that suggests that this is the natural way to understand
them. Rather we should perhaps divide between deductive and non-deductive.[136]
I have already made it clear that these questions will not concern us in a great
deal here.
Let us
return to the problems mentioned above, that (a) to explain what it is for an
agent A to act for a reason, (b) to specify a structure in terms of
which intentional actions can be explained, (c) to exhibit the structure of
acting for a reason so that such actions can be seen to follow from, in some
sense, or at least to be reasonable in the light of, the reasons, and (d)
whether acting for a reason is equivalent to action based on practical
reasoning. We are now in a position to answer these. Let us first consider the
first challenge. The above view indicates the answer to this by suggesting two
points. First, having separated the reasoning (and the conclusion of the
reasoning) from the action (or intention), the above view does not claim that
for any piece of practical reasoning an action will follow for which the
practical reasoning is the reason for the action. A may have a reason
for j-ing in virtue of
having completed some piece of practical reasoning, but nevertheless not act on
it. Also, in a weaker sense of reason, A may accept the premises of a
sound practical argument favouring j-ing, but fail to
draw the conclusion (perhaps because he have failed to grasp the premises
together). The weakest sense of reason, that of having a justifying reason in
the sense expressed by the form: 'there is a reason to j', is considered on
the above account to be included if it occurs in a pragmatic context in which
the agent seeks to find out what to do. It may not provide a reason for which
the agent acts, as will become clear below. Second, it should be clear from the
above account how acting on the basis of a piece of practical reasoning,
whether immediately or later, can be a case of acting for a reason. It is not
the only, however. I have above indicated that we should perhaps understand the
'on the basis of' as a causal claim, but that this is difficult because of
deviant causal chains, among other things. If we can hold that it is causal, we
need to have a causing in the right way. The latter is problematic as
Davidson notes.
I shall now propose
a conception of reason for action, based on the discussion above. This account
shall explain what it is for an agent to act for a reason, and in doing this I
shall specify a structure in terms of which intentional actions can be
explained, exhibit this structure of acting for a reason so that such actions
can be seen to follow from, in some sense, or at least to be reasonable in the
light of the reasons, and deny that acting for a reason is equivalent to
action based on practical reasoning, only to claim that we must see them as
both following the same structure of practical argument.
The view that follows, then, is one
in which the claim 'A j's for a reason iff
his j-ing is based on
practical reasoning' is false. There is a correspondence, however, between
acting for a reason and acting on the basis of practical reasoning, and that is
the fact that we may construct a practical argument that covers both by
being a common structure. We may say on this account then, that acting for a
reason is a concrete realisation of a practical argument, just like I
argued above that a piece of practical reasoning is that too. That it is a
concrete realisation means the following:
1. A instantiates the
argument, that is, he has the desire and belief expressed in its premises and
does the thing indicated by the conclusion.
2. The premises
represent the structure of the (presumably) causal and explanatory basis of the
action, a desire and a belief state.
3. The explanatory
relation which the desire and belief bears to the action, mirrors a kind of
supportive or justificatory relation which the premises of the argument bears
to its conclusion. It does in some (perhaps weak) sense justify the action.
4. A is at the time
disposed to appeal to the argument if asked to explain or justify his j-ing, rather in a
way one appeals to a rule or practice one has been automatically following.
This disposition expresses a normative dependence.
5. When A's j-ing is a
realisation of a practical argument, then even if the action is not based on
practical reasoning the explanatory relation between his reason and action is
just what it would have been if (ceteribus paribus) he had j-ed on the basis of
actual reasoning from its premises to its conclusion. The same reason but
perhaps more deliberately so.
When A
simply infers the conclusion of a practical argument from its
premises, the argument is inferentially realised. It may be at once non-inferentially
and behaviourally realised, through a spontaneous action expressing A's
motivating desire and guided by a belief which A need not entertain.
Practical reasoning can occur without A's acting on it. This is why we
above argued that one must not identify the conclusion of a practical reasoning
with some (external) action. I shall argue for the above points in what comes.
The
following is what Audi (1982) calls the pervasiveness view, a view that
implies that practical reasoning is a pervasive element at least in the most
common kind of intentional action, the view that,
all
intentional actions, or at least those performed with a further intention, are
based on practical reasoning, where [A] [j]'s with a further intention if and
only if [A] [j]'s with the intention to realize
(or at least contribute to realizing) [a], and believes that realizing [a] is not constituted by [j]-ing.[137]
Audi gives several
reasons by way of examples why this view is wrong, that is, why it is wrong to
identify having a reason for action with having performed some piece of
practical reasoning:
Suppose that [A]
wants water, believes that by turning on the tap he will get it, and does so
with the intention of getting it. Need he have done so on the basis of
practical reasoning? Surely he may have been simply standing in the kitchen
when, feeling thirsty, he just reaches for the tap. Or, suppose [A]
immediately hangs up his coat upon going home. [A] may do this both with
the intention to get it out of the way and quite "automatically" as
part of a standing practice. It is unlikely that he here does it out of practical
reasoning.[138]
Goldman (1970),
among others, would count the above instances as cases of practical reasoning
(or practical inference, as he calls it), only that (if A is I),
I do not recite
these propositions [those of the practical inference] to myself as I assent to
them; things happen much to quickly for that. But I do have occurent wants and
beliefs for these propositions during the brief time interval in question.[139]
Schueler (1995) has
remarked, however, that by his definition of 'occurent want' - "It is part
of our notion of an ordinary (occurent) desire that an agent is aware of this
desire."[140] - Goldman
must hold that the agent have some experience of the want in question.
And if that is so, it seems much to hard a requirement upon all reasons for actions.
Goldman is right in that we need not recite a practical argument for our
actions to be based on practical reasoning, but, as Audi holds, the doing of
practical reasoning by an agent "must be constituted by something occuring
in his consciousness; and surely what must occur in consciousness in cases like
the above need not count as practical reasoning."[141]
Nelkin (1993) has argued at length disconnecting intentionality from
consciousness, so we must - if we believe him, something I think we should do
(following my arguments above against an exclusively phenomenological
understanding of desires) - think that intentional states are not necessarily
conscious states.
I think the above considerations and
examples confirms our intuitions about the subject at hand. We experience that
many actions, that is, actions we would not at all deny being intentional, are
spontaneous or automatic. Nevertheless, if asked whether we have any reasons we
would say that certainly we have. Think of driving; would you deny that you had
any reason to shift gear from second to third, if acting intentionally? And
even, would you deny that you did it for that reason? Desiring something, we
often simply see how to get it and act accordingly.
Having said this much, however, we
must also hold (not necessarily, though) that if every intentional action is
held for some reason, that is, is explainable by desires and beliefs of the
agent, then for every intentional action, there is a corresponding practical
argument. This corresponding practical argument expresses a desire of the
agent, that he wants a; a connecting
belief that doing j will contribute to
a; and a relevant
practical judgement that he should do j. In other words, the same
structure respresented above as fitting practical reasoning. Thus, the
structure of the motivation and cognition explaining an intentional action can
always be expressed as a practical argument. Often the corresponding
practical argument is simply a reconstruction of what the agent's
reasoning would have been if he had reasoned. Any intentional
action then is the result of a (spontaneous) response to a "question"
we have called 'What shall I do?', denoting a pragmatic context.
To sum
up a bit, what I claim is (a) that we must not identify (external) action and
the conclusion, that is a judgement, of practical reasoning (even though the
judging is something we do, an action); (b) the conclusion of practical
reasoning does not entail that the agent performed the favoured action; (c) the
desired end, a, expressed in the
major premise need not be desired by A more than he desires anything, a', he believes
incompatible with a; thus (d) A
might conclude that he ought to j but then goes
through another practical argument, grounded in a desire for a', and concludes in
favour of j'; this means that
(e) A may when comparing this prospect with that of j-ing realize
(judge) that it would be better (all things considered) to j, yet fails to
resist a' and j'. In this case we
have two pieces of practical reasoning, one that is favoured and the other
acted on.[142] We may now
consider the questions we started out with as answered.
PART III: RATIONALISM VS. HUMEANISM.
What shall concern us now is the question whether
practical reasoning (or Reason) only plays an instrumental role, or whether it
also plays a generating role. Nagel claims the latter, and we shall consider
what this means, and whether it goes against the view exposed above.
3.1 The Dispute Between
Humeans and Rationalists.
The significance of the dispute between the Humeans
and the rationalists about the role of practical Reason - instrumentalist or
generative - lies mainly in its connection with a wider issue, that of the
nature and scope of moral requirements. It is a common and ancient thought that
moral requirements, if they are to provide reasons for action, must be capable
of guiding behaviour, leading those aware of them to be motivated in accordance
with them. Such a thought may be termed with the ambiguous concept
'internalism'. Wallace (1990) tells us that on such an internalist position
combined with the Humean picture of practical Reason,
it seems to follow that moral requirements can only
provide an agent with reason to act if they are appropriately related to the
agent's antecedent desires; for all motivation and behaviour, on the Humean
view, must be explained by reference to the agent's given, prior desires. The
resulting account represents moral behaviour as dependent on the agent's
existing dispositions, in a way that could restrict the scope of moral reasons
(since the required prior desires may not be universally distributed).[143]
By contrast the rationalist
picture shows us some altogether different psychological basis of moral
behaviour, suggesting that moral requirements are universal or inescapable in
their scope. This opens up for the possibility that there are processes of pure
practical reasoning (as generative) which can explain moral behaviour by
themselves, and which would lead all agents, regardless of their background
desires, to be motivated to act on moral requirements.
An internalist
moral theory is a theory according to which the knowledge or acceptance of a
moral judgement implies the existence of a motive for acting on that judgement.
A judgement that some action is right or good implies that the judgee has and
acknowledges some motive or reason for performing that action. It is part of
the judgement that a motive is present. This means that if someone agrees that
an action is right or good, but fail to see any motive or reason for doing it,
we must suppose, on this view, that he does not quite know what he means when
he agrees that the action is right or good. An externalist moral theory,
on the other hand, claims that such a conjunction of moral comprehension and
total unmotivatedness is perfectly possible, since knowledge is one thing and
motivation another. The simplest example of such a view would perhaps be one
according to which the motives for moral action come from something wholly
separate from a grasp of the correctness of the judgements, for instance God's
demanding that Abraham should kill his son.
Internalists
appeal to various types of motivation to secure that some motivational factor
is present in and guaranteed by ethical propositions, such as self-interest or
self-preference,[144]
sympathy, benevolence, self-respect,[145]
approval, or pro-attitude.[146]
Their appeal derives from the conviction that one cannot accept or assert sincerely
any ethical proposition without accepting at least prima facie
motivation for action in accordance with it.[147]
Those that think there is no room for rational assessment of the basic springs
of motivation tend to be internalists, at the cost of the objectivity of
ethical propositions. Nagel (1970) mentions two ways in which this may be done,
first, one can "build motivational content into the meaning of ethical
assertions by turning them into expressions of a special sort of inclinination,
appropriate only when that inclination is present, and rooted only in the
motivations of the speaker."[148]
The result of this is a anti-rationalist theory that is grounded in some
noncognitive element, be it a commitment, inclination, feeling, or desire, that
is simply given. A second way one can construct an internalist view is
by tying the motivation to the cognitive content of ethical claims, and
requiring the postulation of motivational influences which one cannot reject
once one has become aware of them. The task set for these is to show such
influences belonging to the content of ethics, thereby craving that someone
recognizing the truth of an ethical claim will have to accept the corresponding
motivation. The most influental anti-rational view is Hume's. His view is,
following Nagel,
that any justification of action must appeal to an inclination in the individual to whom it is offered, and that the justification proceeds by drawing connections between that inclination and other things (notably actions) which are means to its satisfaction. The inclination then becomes transferred to these by association, which is what makes persuasive justification possible. If we cast this view in terms of reasons, it will state that among the conditions for the presence of a reason for action there must always be a desire or inclination capable of motivating one to act accordingly.[149]
Hume's starting point is psychology, and from that
elaborates ethics, meaning that there is no way through ethics to define basic
psychological concepts. This together with his restriction on rational
assessment of the passions and of preferences, makes the possibility of
justifying morality limited, according to Nagel. When explanation comes to an
end, all justification are grounded in the presence of the emotion of sympathy,
and if this condition were not met, we would have no reason to be moral.
Whatever other passions we may be actuated by; pride, ambition, avarice, curiosity, revenge or lust; the soul or animating principle of them all is sympathy; nor wou'd they have any force, were we to abstract entirely from the thoughts and sentiments of others.[150]
For Hume, according to Nagel, the motivational basis
is prior to, and independent of the ethical system which results from it.[151]
To reverse this sort of priority one would need a quite different theory. Kant
has provided us with an example of such a theory, an ethical system not
dependent upon desires that must be taken for granted. The insistence that the
moral imperatives are categorical is an insistence of the independence of
morality from motivational factors prior to ethics. Ethical principles are
themselves put as the absolute source of moral conduct, they are themselves the
motivational impulse; one need not go to any other source, any independently
comprehensible desire or feeling.
The position Nagel defends is as follows. Central to
his position are two features that resembles Kant's ethical theory, first, that
it "provides an account of ethical motivation which does not rely on the
assumption that a motivational factor is already present among the conditions
of any moral requirement",[152]
and second, that it assigns a central role in the operation of ethical motives
to the moral agent's metaphysical conception of himself as merely a person
among other equally real persons. I will only consider the first feature, since
that is the most interesting one for my purpose. There is a set of reasons that
are specifically moral and it is because they represent moral requirements that
they motivate, and not vice versa, according to Nagel. This, if true, is
what makes it possible to think that ethics can reveal something about human
motivation. The information it reveals is not just peoples wants, since if
ethics is not to presuppose any motivations and rather reveal their
possibility, the discovery must be of something more fundamental. Some ethical
principles are "themselves propositions of motivation theory so fundamental
that they cannot be derived from or defined in terms of previously understood
motivations."[153]
The claim is, thus, that there are some ethical principles that cannot be
reduced to motivations already known, because they are themselves of such a
fundamental character. They specify how reasons for action follow from certain
given conditons, defining motivational possibilities rather than presupposing
them. What Nagel searches, then, are principles which belong both to ethics and
motivation theory, stating structural conditions "on the forms and
interrelations of reasons for action."[154]
This involves a fusion of what, among others, Smith (1994) wants separated,
namely explanatory and justifying reasons. Nagel is perfectly aware of this and
claims that
a close connection between the two is already embodied in the ordinary concept of a reason, for we can adduce reasons either to explain or to justify action. We may explain what a man does by referring to his reasons. On the other hand we may assert that circumstances provide someone with reason to act in certain ways, without implying that he will be accordingly motivated (if only because of the possibility of his ignorance). But though the explanatory and the normative claims can diverge, this does not mean that we are faced with two disparate concepts finding refuge in a single word.[155]
Central to Nagel's view is whether or not some prior
motivation besides some moral reason must be present to motivate. He attributes
the former view, wrongly I think, to Hume, and the latter to himself. I don't
agree with Nagel if he claims that the mere presence of an objectively good
(justifying) reason will motivate one to act on it. I don't think he means
that, however. But I do think Nagel means that a person who knows what is good
will do what is good. That is right, I think, only if one thinks the good qua
good (if only because of something else) and thereby will be disposed to act on
it. This might seem a mere play with words, and perhaps is so too. We shall
nevertheless consider an important distinction Nagel draws to defend this view.
3.3 Motivated and unmotivated desires.
Nagel (1970) deems
the Humean view - which accounts for all motivation in terms of agent's desires
- unfit. What desires can explain, he says, is limited; "even in simple
cases they produce action by a mechanism which is not itself explicable in
terms of desires."[156]
The claim that a desire underlies every act is true only if desires are taken
to include motivated as well as unmotivated desires.[157]
Accordingly, to explain actions by reference to some motivated desires are not
enough, because further reasons can be given that explain both the desire and
the action. While motivated desires are arrived at by decision and after
deliberation, that is, given through reasons, unmotivated desires
simply assail us like appetites and certain emotions do. Although
unmotivated desires, unlike motivated desires, cannot be given a rational or
motivational explanation, they can be explained. Hunger is produced by lack of
food, but is not motivated thereby. A desire to shop for groceries,
after discovering nothing eatable in the refrigerator, is on the other hand motivated
by hunger. Rational explanation is just as much in order for that desire as for
the action itself. If a desire is motivated, the explanation of it will be the
same as the explanation of the agent's intentional pursuit of the goal. So for
each kind of desire, there is, in addition to the way they arrive, a
corresponding difference in how they are explained: unmotivated desires are produced
(presumably causally) by states such as lack of food, but motivated desires
require rational or motivational explanations (presumably
causally too). That the two differences are closely connected can be seen by
considering for instance the contrast between a simple thirst and someone's
want to drink some water because he has decided to run the marathon in a few
hours. The first desire is an unmotivated desire and the second a motivated
one. In what follows I will discuss this distinction specifically to see
whether it holds. Then I shall, whatever the conclusion will be, go on to
consider Nagel's overall view upon the subject introduced above, i.e. the
dispute between Humeans and rationalists.
What is
involved in the claim that motivated desires are arrived at by decision and
after deliberation, says Schueler (1995), is neither that one decides to
have such desires, nor that they simply (causally) arrive after deliberation.
Rather, the
idea is that they are there in virtue of the deliberation; what explains them
is one's having deliberated and decided as one did. My decision to run the
marathon this afternoon, combined with my belief that this requires drinking
water a few hours before the race, generates (as I will put it), and hence
explains, a desire on my part to drink water a few hours before the race, a
desire that, as I said, I might therefore have without now being thirsty or
thinking that a few hours before the race I will be thirsty.[158]
Motivated desires
are generated by beliefs and/or desires, that explains them; they have
reason(s). For Schueler, Nagel's distinction has a significant point in that it
"marks a difference in the nature of the explanation required to explain
why the agent has the desire in question. Motivated desires are ones that must
be explained in terms of the agent's reasons for having them,"[159]
and it is for this reason that he claims Nagel's distinction being
"crucial for understanding what desires are."[160]
Looking
to the debate between Humeans and anti-Humeans, Lenman (1996) calls this way of
accounting for desires 'Nagel's strategy', a strategy that involves
inducing rationalism through the backdoor:[161]
The trick is to
concede that desire plays a part in all motivation of action but then to argue
that the motivation of that desire in turn, where it is motivated by reasons at
all, may be purely cognitive. The rational antecedents of action may always include
desires but the rational antecedents of motivated desires need not. So
practical reason can after all originate in mere unaided belief.[162]
Thus, on this
account a mere unaided belief can be enough to motivate one to action, but it
is conceded to the Humeans that it can do so only via a detour through desire.
The source of our actions may be identified in beliefs that are held to
motivate both the action and the desire. Desire is thus conceded
to lie behind all rationally motivated action, but not necessarily,
crucially, all the way back. Nagel (1970) sums up his view like this:
Though all
motivation implies the presence of desire, the sense in which this is true does
not warrant us in concluding that all motivation requires that desire be
operative as a motivational influence. To that extent it remains open
that there can be motivation without any motivating desire.
Some desires are themselves motivated by
reasons. Those desires at any rate cannot be among the conditions of the reasons
which motivate them. And since there may in principle be motivation without
motivating desires, those reasons may be motivationally efficacious even
without the presence of any further desires among their conditions.
There are two ways in which this might be so;
either some other motivating factor besides desire may be present among the
conditions for the existence of those reasons; or else their motivational
efficacy may derive not from the conditions themselves, but rather from the
principle which governs the derivation of reasons from those conditions.
In the latter event, the motivational
efficacy of reasons for action would be due only to the system by which they
are derived from their conditions. This would be explained by a connection
between the structure of a system of reasons and the structure of human
motivation. In that sense it would still be true that a reason is necessary
capable of motivating.[163]
According to the
above, as Platts (1991) claims, there are two subclasses of motivated desires;
one that makes a further reference to a desire, and another that does not.
Another, closely related point Nagel makes, Nagel's entailment point,[164]
is that whatever may be the motivation for someone's intentional pursuit
of a goal, it becomes in virtue of his pursuit ipso facto appropriate to
ascribe to him a desire for that goal.[165] He continues by saying that if this
desire is a motivated one, "the explanation of it will be the same as the
explanation of his pursuit" (call it Nagel's explanation point).[166]
Accordingly, if (i) A does j intentionally so
that a, (ii) A
desires that a (by (i) and
Nagel's entailment point). Now, what is the explanation of his pursuit? We
often say someone did something because they wanted some goal, in this case a. We can then say
(simply) that (iii) A did j because he
desired that a, his desire for a was the reason why
he did j. This explanation
of the desire that a, if it is a
motivated desire, will be the same as the explanation of his doing j. But this becomes:
(iv) A desire a because he desires
a (by Nagel's
explanation point), which is quite empty. Already this simple analysis shows
Nagel to be dealing with a concept of 'reason' that is not intimately related
to the one we have called a reason for which one acts. The denial of any
distinction between explanatory and justifying reasons cited above, then, comes
down on the justifying or normative side for Nagel. Reason is always a
normative reason, and explain actions as the reason for which the agent
acted if some motivating element is attached to it. The reason for which
A does j then is not his
desire that a (and a connecting
belief), but it is rather the content of the desire, a, that is the
reason.
Take
the example from Schueler of George who goes to a school meeting - even if what
he really wants is to stay home and read - because he has a responsibility to
his son. On Nagel's account, the reason why George goes (does j) is not that he desires
that he fulfills his responsibility to his son, but rather that there is this
reason that is fulfilling ones responsibility to one's son; that is, George
does j because of a. According to
Nagel's entailment point, then, since George does j because of a, he must have a desire
for a.
Now we
can reconsider the points above. The two first ((i) and (ii) is alright, but
the third becomes: (iii') A did j because of a; and the fourth
comes out true (in Nagel's sense) as (iv') A desires a because of a.[167]
But given this understanding of reasons, what is it left to argue for against
the Humeans? The trick has already been performed. Nagel next questions whether
it is obvious that "a desire must enter into this further
explanation", that is, the explanation of the motivated desire. Clearly,
this is rhetorical, but we will consider the question later on. Now, to the
question of how we shall understand the distinction between motivated and
unmotivated desires.
According
to Nagel, two conditions distinguish between motivated and unmotivated desires.
First, there is a distinction between in what way they are arrived at - by
decision and after deliberation, or by simply assailing us - and, second, there
is a distinction between the way they are explained - as rationally generated or
produced. Thus, we may give further reasons for motivated desires, but not
unmotivated ones. For the latter we can give no reason. Psychologically
we can say that they are produced by some internal drive-state of the agent.
Unmotivated desires can provide reasons for further (motivated) desires,
however.
The
distinction is supposed to be exclusive; desires are either motivated or
unmotivated. But it seems that this cannot be true if taken in one sense. If we
take "by decision and after deliberation" to mean consciously arrived
at, then it seems to be desires that cannot be subsumed under any of these
categories. This way to understand motivated desires is not correct, however.
Nagel says that beliefs also need not simply assail us "for often, as when
we simply perceive something, we acquire a belief without arriving at it by
decision."[168]
We can often explain a belief by citing the reason why we believe it. So as we
might cite any belief that is not the result of perception as inferred (or
motivated), we may cite any desire that is not produced by some drive like
hunger or thirst, etc., as motivated. This may at first seem too broad, having
the seemingly result that all motivated desires may be explained by a belief
(reason) alone. But this is so only if we forget that not all motivated desires
can be such explained, according to Nagel; there are those that need to be
explained by a further motivating desire.
There
is a problem about how to express unmotivated desires; how are we to express
'hunger' for instance? If we express such a desire as a propositional attitude,
we seem to be able to say that the content of the desire might explain
the desire. Thus, if 'hunger' means 'desire that I have some food', the reason
why I want food is that I am hungry. Your reason for wanting to eat may be to
get nourishment. In such a case, your desire to get nourishment both explains
and motivates your desire to eat. There may be reasons for your having a
desire to eat (or say, craving for food), such as food-deprivation, but we do
not have reasons for having such a desire to eat (or craving for food). To see
this, compare the question: 'Why do you have an appetite?', with 'What are your
reasons for having an appetite.' The first is a legitimate question, since it
can be answered for instance with 'Because I have been deprived of food for 5
days.' This is an explanation of why you have the appetite, but not a
reason. This is why the second question is illegitimate, since it asks for a
reason, but can't get it.
3.4 Schueler on Nagel's Distinction.
Schueler (1995)
finds the requirement that unmotivated desires must simply assail us too
restrictive. He gives no real reason for this suspicion, but examples like
'having a desire to visit my sister and her family next summer' seems to
indicate that if this desire is unmotivated it does not assail us like desires
as hunger and thirst does. In the end, after some refinement of Nagel's
definition, he defines the distinction between motivated and unmotivated
desires thus:
The distinction
between motivated and unmotivated desires can be made exhaustive if we simply
drop Nagel's definition (if such it is) of unmotivated desires as ones that
"simply assails us" and instead define unmotivated desires as ones
not arrived at by decision and after deliberation (and thus not subject to
rational or motivational explanation). This will be to understand unmotivated
desires simply as desires that are not motivated desires.[169]
All the burden of
evidence is thereby put upon the shoulders of motivated desires and the
features of being arrived at by decition and deliberation, being an act of
will.
One can understand desires
deliberated and decided upon in two ways, according to Schueler. In one sense
it seems to follow the old pattern noted by Kant that who wills the end, wills
also the means which are indispensably necessary and in his power. This
principle of practical inference has been repeated through the years. Goldman
(1970) expresses it like this: "if a man wants x and believes y
is a means to x, then, characteristically, he will want y."[170]
Smith (1987-1988) also notes that an irreducible norm of practical reason (when
suitably qualified) is: "If A desires that he fs and he can f by ying then he desires
that he ys."[171]
Schueler notes some problems with this principle, but I think we should
consider it true (when suitably qualified) in connection with practical
reasoning. When we have some desire and we think we must do something to have
it the way we desire, we have a desire for that means insofar as we conclude
that we should do it. The problematic point about this is whether this is the
way we should understand Nagel's view, or that whether we shall understand it
in a sense that is closer to how beliefs may follow from each other, that is,
outside practical reasoning. We can, according to Schueler, distinguish between
intention-generated motivated desires and motivated desires
respectively. We can now say that every motivated desire can be
explained rationally, and that some motivated desires are arrived at by
deliberation and after decision. Accordingly we must draw the distinction
between motivated and unmotivated desires like this: A desire is motivated if
and only if A wants whatever it is that A wants for a reason, and
a desire is unmotivated if and only if A does not want whatever A
wants for any reason.
There seems to be a great difference
between these two types of motivated desires though, a difference that makes
the one type true in cases where the other is false, especially in connection
with claims against the Humeans. The question is; can Nagel be taken to mean
that motivated desires are intention-generated motivated desires? If
deliberated and decided upon is taken to mean that an intention is formed, it
must be wrong, because then we talk of a distinction between desires that are
intentions and desires that are not. The distinction is thus between my
intending to j and my desire that
a. We shall see
whether I am right about this.
3.5 Nagel's strategy and the Genealogy of Desires.
Above I mentioned
what Lenman (1996) calls Nagel's strategy; to concede (to the Humeans) that
desire play a part in all motivation of action and then to argue that the
motivation of that desire in turn, where it is motivated by reasons at all, may
be purely cognitive. Thus the rational antecedents of action may always include
desires but the rational antecedents of motivated desires need not (not
considering motivated desires), since they are only recognised in retrospect - at least this is what Foot (1978) claims Nagel to
mean. The motivated desire is then only an empty placeholder-concept, and the
real motivating element (or source) is a belief expressing some prudential
consideration, some altruistic belief, some moral obligation, etc.[172] So practical
reason can after all originate in mere unaided belief. It must be remarked again that this claim does not
range over all (motivated) desires, since some may be further motivated
by desires, and presumably (in the end at least) by some unmotivated desire.
Consider
the following example taken from Darwall (1983).[173]
A young woman, Roberta, watches a documentary film about the plight of some
exploited workers, and as a result is motivated to do j, that is, join a
certain boycott. It is conceded from the rationalists that her motive for j-ing includes at
least one desire, the desire to j. The question is,
according to Lenman (1996), whether the motivation of that desire must
likewise involve a desire. Darwall claims that Roberta may have no desire
prior to viewing the film that explains her desire to do j, and that whatever
desire she does have after the film seems itself to be the result of her
becoming more aware of considerations that motivate her decision, that is, the
unjustifiable suffering of the workers. An explanation of her decision might
invoke some overall desire to relieve suffering, but, Darwall urges,
need not, for her to come to a general desire to relieve suffering she would
have to become actively concerned about the fact of suffering itself, conceived
independently of who suffers, and she may never have done that.
One possible Humean solution is to find
some desire-element as necessary for the motivation of desire as for the
motivation of action.[174]
Lenman (1996) does this by reapplying the teleological argument. This reapplication,
as Lenman notes, is not that straightforward. According to Wallace (1990):
"The point of the teleological argument is that it is not coherent
to suppose that intentional action could take place in the absence of a state
of desire."[175]
Taking this insight as our guiding star for an argument against Nagel's
strategy cannot be done unemended, following Lenman, because "we do not
speak of arriving at desires as a result of intentions so to arrive in the same
way as we speak of performing actions as a result of intentions so to perform.
It's natural enough to speak of intentional action but intentional desire is a
rather peculiar notion."[176]
This is a view that follows my claims above. We must halt here and consider the
question we ended with in the above chapter. How do Lenman and Darwall
interpret Nagel's distinction? Darwall's Roberta can be seen to be reasoning
practically, ending in some desire to do j, join a boicott. This may be
just what she does, and if she does that we may say that the reason why she did
it is that she believed by doing j she would have it
that p, and, according to the teleological argument, she desired that p.
If the case is seen like this, we do not need any reapplication to show that
her desire to do j was motivated by a
further desire, the desire that p, since on this interpretation, her
desire to do j is an expression
of her practical judgement that she should do j on the basis of her
reasoning. Thus, what we are concerned with here, as I will claim, is not
intention-derived motivated desires, if there are such things.
What we
need to get an analogue of the original teleological argument to be effective
at this level, Lenman proposes, is a distinction between desires that are rationally
and those that are nonrationally arrived at. Presumably, the motivated
desires that Nagel's strategy invokes, are distinguished just because they are
being rationally rather than nonrationally arrived at. It is important to notice
that the distinction we want is not that between rationality and irrationality,
however. Depending on what contrast we want to make talk of rationality can be
taken in two ways; all practical reasoning is rational as opposed to
nonrational just in virtue of being reasoning but it counts as rational only
when certain standards are met. Much humean reasoning is irrational, however,
meaning defectively rational. "Rationality in the sense opposed to
nonrationality is not only compatible with irrationality but is presupposed by
it," says Lenman, and continues:[177]
The central
point I want to make here is that not just any old thing can be a reason, even
a defective one, that there are constraints on what is admissible in the way of
rational explanation of action. Given this, we may defend the claim that
Roberta's decision insofar as it is open to rational explanation, in the sense
of contrasting with nonrational, must indeed require a contribution from desire
[…] all the way back. Or at least all the way back to the point where it ceases
to be an explanation in terms of reasons at all (and of course all explanation
in terms of reasons must eventually so cease). Insofar as the decition is
nonrational, on the other hand, the question of the relative significance of
belief and desire in the rational motivation of what she does, does not arise.[178]
The point here is
that there is no problem at all about seeing Roberta's conviction as a causal
chain, or psychological movement, from belief to desire, and so forth, the
difficulty just is with seeing the beliefs as constituting reasons for
the desire, that is, being caused in the right way.
My coming to
believe that P only begins to provide a rational explanation for my coming to
desire that not-P if there is something about P which I am averse. This is just
the teleological argument over again. Because desires have a common direction
of fit that beliefs do not share, it follows that not all the rational
psychological antecedents of desire are strictly cognitive states: no beliefs
can motivate any desire all by themselves, just as no amount of information about
my beliefs would tell you anything about what I was liable to do unless you had
some additional data on what I wanted.[179]
For Lenman, a
reason laying behind some motivated desire might then be as follows:
I am aversive about p.
I believe that p.
I desire that not-p.
This example seems
a bit constructed since the aversion about p is logically equivalent to
desiring that not-p. So for any desire we could describe it as an
aversion that not-p. I cannot see how this is the teleological argument
all over again. For any motivated desire we could construct a logically
equivalent desire that could be said to motivate its equivalence, but that
couldn't amount to much as a critique of rationalism.
One is
perhaps fooled by the example of Roberta, because it is so natural, says
Lenman. He proposes we instead should look on examples like the one given by
Nagel (1970); hearing that there is no parsley on the moon, so I come to
desire that some be transported there.[180]
What does the "so" mean here? It might just mean a causal connection
between the belief that there is no parsley on the moon and the desire that
some be transported there. This is alright, because, or so it seems, any
thought can cause another one (surely you have experienced that when your
thoughts are flying and after a while you stop and think how far you have gone
off the track you startet at), and, thus, one can come to learn that there is
no parsley on the moon, and so think that there are no pink parsley-eating
elephants there either, or desire to have some parsley, or whatever. If
"so" is taken to indicate that a full explanation in terms of reasons
has been given it has no business in such a sentence as "James believes
there is no parsley on the moon and so he wants parsley". The link
indicated by "so" in the foregoing examples may be a causal link but
cannot be of a kind appropriate to the antecedents motivated desire.
What makes
certain body movements actions is just that they are rationally motivated,
motivated by reasons. Precisely this is also true of motivated desire if we are
to understand by "motivated" something more interesting than merely
caused, if this story is to have anything to say to an account of practical
reason. So the inappropriateness of any talk of "intentional desire"
does not prevent the teleological argument from generalizing. If it works at
all, it works here. Nagel's strategy fails.[181]
But this is simply
not true unless the motivating desire motivating the motivated desire is more
substantially defined than the logically equivalent and negative description of
the motivated desire.
Another
point is that it may seem as the Humeans, as exposed by Lenman, mean that we
need an antecedent desire for doing everything. This would presumably make it
impossible for me to eat my first meal of Icelandic shark. Lenman invokes the
following example to first expose the wrong picture and then shake it off
Humean hands. Someone, A, comes into an art gallery by accident to get
shelter from the rain. A has no interest in paintings at all upon this
visit - he is an aesthetic neophyte - but sees a certain painting, by Leonardo,
say, and decides to buy it. Lenman imagines the anti-Humeans to tell a story
like this: the painting have certain features, a, b, c, …,
and in virtue of noticing these features, A decides to buy it. Perhaps A
reasons like this: This painting has features a, b, c, …, so I want to buy
it. It would be foolish, following Lenman, for the Humean to insist that
the reasoning is more plausibly represented by: I want to buy a painting
with features a, b, c, …, here is one, so I buy it. "It may be far
more natural here to suppose that the process of thought from seeing the
painting to wanting it is simply not a rational process at all.",
says Lenman, and according to this we might have: This painting has features
a, b, c, …, I want to buy it.
Now if the
painting is by Leonardo, the desire to buy it, albeit perhaps unrealistic, will
make enough sense to an outlooker of educated taste for the "so" to
have a justificatory force it would lack were the painting by, say, Lenman. But
as far as explanatory force is concerned, the cases are on a level. For we have
supposed our neophyte to have no antecedent reason that informs the
formation of his desire.[182]
Even though I have
no antecedent commitments to Icelandic shark before I have ever eaten it, I
might take with me, after the encounter, a disposition to desire or reject
Icelandic shark in the future. As Lenman says, "it is, at least sometimes,
through discovering what we are disposed to like or desire in particular cases
that we arrive at more general partialities for particular sorts of
things we like or desire."[183]
Lenman places Roberta and the aesthetic neophyte (and by implication also me),
at the beginning of such a process, and only further from this beginning can
the Humean invoke general and antecedent desires to explain rationally the
genesis of the particular desires in question. At the outset, however, our
agents "may simply be so constituted as to respond to certain experiences
by forming certain desires, where this most plausibly is not a rational process
at all."[184] What
prevents the teleological argument to attribute a desire for having a painting
with features a, b, c, … is that the desire was not
rational.
Still one can be a bit confused
about what Lenman says here. I see three possibilities here, and Lenman seems
to be confusing two of these in what he says. My desire to have some Icelandic
shark may come up just causally from any belief; there is nothing problematic
about this, and we might say that this desire is unmotivated. But, if I act
intentionally (or forms an intention to act) eating (or intending to eat) the
shark, describable as (for instance) 'I (desire to) eat this shark', following
a practical judgement 'I shall eat this shark'; then, according to the
teleological argument, I must have had a desire for something the eating of the
shark would fulfill. Perhaps a desire to eat something exotic, or to show off
to my friends, or just to eat Icelandic shark. These desires disposes me to act
so that their content comes true. The third possibility as I sees it, is that I
may have a motivated (not intention-derived motivated) desire to eat Icelandic
shark. The arrival of this desire cannot be explained the way that my judgement
that I shall eat the shark arrives. It is only in this third sense that
Lenman's claim that the Humeans need not induce a desire for anything - meaning
any desire (or belief) or unintentional act - is true. Although I need not
desire the eating of the Icelandic shark because of some essential property of
that meal (that it is hot, tastes something, etc.), since I have never eaten it
before, I could be disposed to eat exotic meals for instance. Lenman, however,
gives no good solution how to understand how desires be motivated, so this we
have to provide. I shall turn to this after considering another important
point.
As Lenman applies
the teleological argument - on the antecedents of desire - it may seem to prove
too much. Wallace (1990) has remarked this:
[The
teleological argument] turns on the claim that reason explanations are
necessarily teleological, in the sense that the psychological states which
constitute one's reasons are always goal-directed states. But if this were
true, then to have a reason for believing something would equally be to have some
desire which enters into the explanation of the belief. This is clearly not the
case, however, since rationalizing explanations of beliefs are ordinary given
exclusively in terms of further beliefs that the agent holds (together wih
principles or norms of theoretical rationality).[185]
How then are we to
stop the teleological argument from generalizing to the explanation of belief?
Lenman (1996) answeres this question with the question: Why should we?
Pace Wallace, there is surely nothing
parenthetical about the norms either of theoretical or of practical
rationality. Thus the governing norm of practical reasoning is, at least
arguable, that one's desires be satisfied, the governing norm of
theoretical reasoning that one's beliefs be true. These are constitutive
norms, giving point respectively to the more specific norms of theoretical and
practical reasoning and content to the distinction between the two. To reason,
theoretically or practically, we must, it is plausible to suppose, accept some
such norms, general or specific; and in my present broad characterization of
"desire" in terms of direction of fit, the acceptance of such norms
counts as having desires. So we do indeed need to have some desire if we are
ever to have a reason to do, or think, anything.[186]
According to
Lenman, the sentences 'Pierre believes that elephants exists so he wants a hot
bath' and (unless it is read enthymematically) 'I know how much this person is
suffering so I want to help' are on par in that neither give a proper
explanation in terms of reasons. The explanation of this reasoning may be
purely psychological, the belief causing the desire to occur. This is more
evident in the first case than in the second, because we take the second to
express, by the proposition before the 'so', a reason for the proposition after
'so'. This is not the case unless there is some motivating element behind the
first proposition that can explain the reasoning to the want to help, Lenman
claims. As such, all reasoning, practical or theoretical, is governed by
the acceptance of norms that may not be identified with beliefs.
Consider
the claim that practical reasoning is accepting a norm that one's desires be
satisfied, and that theoretical reasoning is accepting a norm that one's
beliefs be true, and that these are among the premises of the reasonings. This
cannot be right, because if I desire in theoretical reasoning that my belief be
true, I must also desire my desire to be satisfied; in this case that my
beliefs be true. The same goes for practical reasoning where I desire that my
connecting belief be true. I think it is important that we do not bring
pragmatic considerations into the premises to motivate a conclusion.
They must be kept as contextual conditions, not premises.[187]
It is one thing to claim that for any motivated desire there must be a further
desire motivating it, but another thing to say that what motivates it is some
general desire that one wants one's desires satisfied. Anyway, I can't see how
such a desire can motivate the desire in question. I cannot explain my
(motivated) desire by saying I want it to be satisfied. The same applies to the
motivation of beliefs. Consider the following:
I
desire that any desires of mine be satisfied.
I believe
that my desire that a will satisfy my
desire that any desires of mine be satisfied.
I
should desire that a.
This way of
representing motivated desires does not care about the content a of the motivated
desire at all. Nevertheless I agree with Lenman in a sense, because I think it
right that we should include among the pragmatic conditions of practical
reasoning and reasons for action a motivation to have a certain question
answered; about what to do. I am not sure about beliefs, but if one reason that
p and if p then q, q, one should perhaps include a
desire for the motivation of that too.[188]
Perhaps then, Wallace's critique does point to a mistake in the
reapplication of the teleological argument, if we must include conditions
properly held outside the premises of reasoning into the premises.
It
seems to me that we cannot think of reaching a (motivated) desire by reasoning practically,
at least not if we follow what I have said above. Finding out what to desire
seems to be doing theoretical reasoning, but then we have perhaps no reason to
think that the reasoner is really desiring what he has a reason to desire, and
thus we might just as well consider the desire as empty and exchange it for the
belief that motivated it. If I reason so that I believe I should desire a, then I should
desire a. But do I desire a because of this?
Certainly it may constitute some reason for me desiring it, and according to
Lenman there must be some added desire that my desire be satisfied for my
desire for a to be a motivated
desire. Given this conception, first, there is no reason, I think, to
characterize this as practical reasoning, and second, what stops me from
denying the conclusion that I should desire a? It seems like any moral view
might slip in and becoming a desire of mine as long as I find that I should
desire it. But then again, what stops the rationalist from being right in his
claim that we may skip the (added) desire and account for the motivated desire
by the belief that I should desire a? In the above
conception of reasons for actions and (typical cases of) practical reasoning,
what made me committed - in a non-necessary sense - to conclude that I should
do some action j, was my desire
that a. Through my
practical reasoning out what to do, my desire that a comes off on my judging that
I should j, and this makes it
motivating for me to intend to do, or do at once, j.[189]
Also, that my desire to a is a desire in a
motivating sense for action is that it disposes me to do any action that has
the effect that a. What makes me
committed in the case of motivated desires? We need a motivating desire that
will come off on the motivated desire both motivationally and in content.
According to the distinction made
above, between (i) mere causally generated (unmotivated) desires; (ii)
intention-derived motivated desires; and (iii) motivated desires proper, we may
now understand the Humean-rationalist debate more properly, or so I claim. We
need not concern us about (i) any further. If Nagel claims that there are intention-derived
motivated desires that are not motivated by desires, he is wrong because of the
teleological argument. This argument says that whatever motivates us to action
it must be for some goal, and we must consider it a desire accordingly. We must
consider the major premise in typical practical arguments qua desire.
Thus, (ii) is handled on the present theory. Now, how are we to account for the
major premise in any practical argument; need we introduce a further desire if
that desire is motivated, or may we explain it by a belief only, that
presumably makes it no point mentioning the (motivated) desire? This I consider
being the only thing the rationalists may question about the Humean conception.
But does this hold?
When one acts for a
reason, one's action is caused (I suppose) by desires and beliefs that are
related (roughly) in the following way; given the desire, what is believed (the
content of the belief) says something in favour of, or argues for the action
that may be judged to be what one should do in the conclusion. The desire is
connected with some type of behaviour by a cognitive bridge, an instrumental,
means-end, or connecting belief: by acting this way one will (at least)
contribute so that what one desire should come true, might come true. The
belief thus functions as an instruction for gaining, or at least possibly
helping to gain, the satisfaction of the desire.[190]
Actions performed for a reason are not only caused by our desires and beliefs;
they are in a sense, prescribed, dictated, or at least justified by the desires
and beliefs that cause them. They make the action intentional. Now, is this how
reasons and (motivated) desires are related? Is it true that given some desire,
what is believed says something in favour of, or argues for desiring something?
The underlying desire and belief (if so) seems to concern the object or content
of the resulting desire, i.e., what one is desiring, rather than the state
of desiring itself. This makes explicit Lenman's point that desires are not
intentional. Gordon (1987) has pointed this out in connection with emotions. He
construes emotions so that for any emotion there is a desire or wish and
a belief that causes it and is the reason for it.
Thus, if Mary
is embarrassed by the publicity about her wedding, then among the
"sustaining causes" of her embarrassment is her wishing there not to
be such publicity and her believing that there is or has been such publicity.[191]
If we believe the
distinction between motivated and unmotivated desires we cannot construe every
desire as issuing from some further desire and belief; and unless we argue for
it, we cannot even say that every motivated desire must have a desire as a
reason either. Nevertheless, I think we should accomodate the point that what
the explanation of a motivated desire is, it is an explanation not of the state
itself, but of the content of the desire. Mary's negative attitude toward
publicity - her wish that the wedding not be publicized - and her belief that
the wedding has been publicized do not cause embarrassment by saying something
in favour of being embarrassed. Her embarrassment was not dictated or called
for by this belief with her attitude toward publicity. The belief that enters
into the analysis of her embarrassment - the belief in virtue of which
she is embarrassed by or about the publicity, that is, embarrassed that the
wedding received publicity - is not the belief that her embarrassment will
or might somehow do some good, much less a belief that will, specifically undo
the state of affairs she is embarrassed about.[192]
Embarrassment is not an answer to the question: 'What shall I do to get
embarrassed?', like actions are the answer to the question: 'What shall i do to
fulfill my desire?' In the same sense, I claim, a (motivated) desire is not an
answer to the question: 'What shall I do to desire that a?'
Take a desire we might assume to be
motivated, one that can further motivate me to action: 'I desire that I have an
apple.' Is there any way in which we may construe a story like Mary's
embarrassment above? We may take it that if I jump up and down under a tree,
the reason why I do this might be explained by saying that I have a desire that
I have an apple, and I believe that by jumping up and down under an apple tree
will contribute to this state of affairs, so I jump up and down. Now, according
to Nagel, one could here cite as a reason for my action, the very same reason
that I desire an apple, if that desire is motivated. If we construe the
generation of the desire as we do with the action, we must say for instance
that 'I have hunger' (presumably describable as: 'I desire that I have some
food') and connect my hunger with my desire for an apple with a belief like: 'I
believe that by desiring an apple it will contribute to me having some food'.
But this sounds pretty awkward. We should instead, I think, construe the
reasoning behind a motivated desire, in this case, as:
I
desire that I have some food.
Apples
are food.
I
desire that I have an apple.
Still we may
explain my jumping by citing my desire for food. But the belief that apples are
food does not call for my desire or dictate it. Rather, it is, together with my
hunger, the belief in virtue of which I have a desire for an apple.
Gordon notes that there is a crucial
difference not only in the logic of emotions but also in the form of arguments
used to talk a person into (or out of) actions and emotions. In the case of
actions a fully explicit reason (desire and belief) entails a positive
evaluation of the action itself (under some descripion).[193]
He gives the following example of Mary. First, as reasons for action:[194]
Mary
desires that her wedding will receive publicity.
Mary
believes that getting married in New York will have the wedding receive
publicity.
Mary
judges she should get married in New York.
The following would
be the analysis of Mary being glad that her wedding has received
publicity.
Mary
desires that her wedding will receive publicity.
Mary
believes the wedding has received publicity.
Mary is
glad the wedding has recevied publicity.
A (fully explicit)
reason for action entails a positive evaluation of the action itself, a reason
for an emotion such as gladness does not entail a positive evaluation of the
emotion itself, the gladness, for instance, rather it entails an evaluation
of the object or content of the emotion.
In virtue
of its causal structure alone, we can expect intentional actions to be responsive to reasoning
concerning the goodness or badness of so acting. This does not hold for
emotions. It cannot be said in virtue of the causal structure alone, the
embarrassment will respond to reasoning concerning the goodness or badness of
being embarrassed. Instead, we can expect embarrassment to be responsive to
reasoning concerning the "object" or "content" of the
embarrassment: where someone is embarrassed "about a certain state of
affairs S," reasoning as to whether S actually obtains, and
reasoning as to whether S is bad.[195]
My claim is that
something similar must be said about motivated desires. We get from one desire
to another, not by evaluation of the desire itself, but by evaluation of the
content of the (motivated) desire. This content is given by the belief that the
agent believes instantiates, the original desire. It is not so that this
belief contributes to a positive evaluation of some action that must be done to
have it that one desires. It is not a belief the content of which will have the
effect of generating a (motivated) desire. We may, as is the case about
emotions, say that having a (motivated) desire is having such and such a
reason. Having such and such a reason, is having a desire. We may still
talk of states causing each other though, and thus have the relevancy of the
question 'Why …' still intact. What we must understand is that the questions
'Why did you do that?' and 'Why do you want that?' are different questions
requiring different answers. Intentional actions and (motivated) desires are
each produced by systems responsive to our desires and capable of listening to
reason (contrasted with non-reason, not ir-reason), particularily to arguments,
and refutations of arguments, that apply norms, standards, or values. But they
respond to arguments of significantly different forms. In virtue of
intentionally j-ing we are in
general, responsible agents, answerable to the norms concerning j-ing; whereas we
are not answerable to norms concerning desiring that a, merely in virtue
of desiring that a.
The subtle difference I have tried
to explicate here between the causal structure of (motivated) desires and
actions, based on insights from Gordon (1987) on the difference between
emotions and actions, is very important to understand several problems with
which we have been concerned above. The question of what is the nature of
motivated desires have partly been answered, though we have still left to consider
the upshot of Nagel's strategy, the difficult subclass of motivated desires,
those that presumably is motivated only by beliefs.
Can we allow beliefs motivating
desires? I don't think so. If someone tells me I should take care of my body,
this belief would not affect me so that it would seem to me good to take
care of my body, unless some desire was operative. This may happen the way
belief follow belief in the well known reasoning schema; p, if p
then q, q, or some other one. In the case of desires I only
follow this scheme if I desire that p, believe that if p then q,
and as a result desire that q. If I only believe that p, I would
not end up desiring q. In sum then, my view is a sort of subjective
internalism, making motivation dependent upon desires in a way. How this
dependency works, I have not explicated. What I have given is a structure
of how this can be.
The view defended
in this essay, a sort of subjectivism, counting only reasons as motives when
they are internalised or have aquired a motivational role, has been criticised,
because it seems to undermine the objectivity of especially moral claims. This
worry is supposedly what drove Kant as well as Nagel into denying Hume's
subjectivism ("'Tis not contrary to reason to prefer the destruction of
the whole world to the scratching of my finger").[196]
The subjectivism defended above
reflects a broadly instrumentalist theory of practical rationality that
includes finding suitable means to one's goals. Pro (or con) attitudes,
desires, then, have the power to rationalize or justify choice of action
themselves. I have defined such rationalizing attitudes above as functional
states that disposes the agent having them towards certain actions. To say that
someone has a pro attitude towards world peace is to say, among other things,
that his psychological setup disposes him to do what he believes will make
world peace more likely. But as Quinn (1993) has remarked,
how can a
noncognitive functional state whose central significance in this context is to
help explain our tendency to act toward a certain end, or in accordance with a
certain principle, rationalize our pursuit of the end or our deference
to the principle? How can the fact that we are set up to go in a certain
direction make it (even prima facie) rational to decide to go in that
direction? How can it even contribute to its rationality? Even if a past
decision is part of the cause of the psychological setup, there still remains
the question whether to continue to abide by it. […] Moral attitudes, whatever
their special moral earmarks, rationalize because they are dispositive
functional states and not because they are moral. The underlying neo-Humean
theory of rationalization is completely general. […] The basic issue here is
more fundamental: whether pro- and con-attitudes conceived as functional states
that dispose us to act have any power to rationalize those acts.[197]
He proceeds by
giving arguments to prove that whether or not a mere functional state is
sufficient, it cannot ground reasons for action. "What does this," he
says, "is another element (of necessity) typically present in basic
desire, namely, some kind of evaluation of the desired object as good - for
example, pleasant, interesting, advantageous, stature-enhancing, as
decent."[198] In this he
resembles Stampe (1987) and his claim that, what constitutes the
"authority of desire" is that some state of affairs seems good
to the agent. It is the supposedly deprivation of desires of any evaluative
thought, reducing them to functional states that "may show what we will do
under certain conditions, but not what we should do", that is the problem
about (neo-) Humeanism.[199]
Only on a view that takes such evaluative thoughts into consideration can
"the mystery of how the goodness of action […] provide reasons […]
disappear."[200]
I think Quinn's arguments rest upon a misunderstanding of the view under
attack. I will not defend this claim properly, only indicate that my view can
account for the evaluative thought necessary for desires to rationalize.
Nothing I have said excludes
evaluative thought from being internal to the desire on my view. I have in a
sense divided the analysis of action between what it is being a state of
reason, and what it disposes us to do. The claim is not that we by this have
exhausted the concept of desire, only that we have a minimally necessary and
sufficient condition for something to be a desire. Surely, if the functionalist
analysis of desires can account for the phenomenology of some desires in
disposing the bearer to feel something (as Smith (1987) claims), then I see no
obstacle to claim on the same view that desires disposes the bearer to express
or hold evaluative thoughts (beliefs). But more accurate, the view I favour is
one that holds that the state that is a desire just is an evaluative
relation between some proposition expressing a state of affairs and a pro
attitude, and that this evaluative relationship minimally is being
disposed to act so as to make the desired state of affairs come true. The
analysis presented is short on accounting for in what respect a certain
state of affairs is desired, but as I have said, I think this is compatible
with what is accounted for, namely that to act for a reason, and to
judge that one should do something (typically), means being motivated, and being
motivated means minimally having a desire that something should be the case and
being disposed to act accordingly. As such, morality and the objective value of
moral principles are, if not accounted for properly, at least not excluded.
This does not, however, reduce the challenge such considerations give a view
like mine.
Any kind of explanation is pattern-observation. What
is special about action-explanation is that the vantage point from which this special
type of pattern is seen, is, as Dennett (1987) notes, an intentional
stance and the nodes of this pattern are intentional concepts. "If you
"look at" the world in the right way, the patterns are obvious. If
you look at (or describe) the world in any other way, they are, in general
invisible."[201]
Intentional concepts are special in that they are normative - the pattern we
lay out is directly answering to our norms. This normativity is baked into
questions we ask about our own or other's conduct: 'What shall I do?', 'Why did
you do that?' Reasons are at work in both ends; the sending and the receiving
alike.
Basically,
reasons may serve two roles, a justifying role - a role connected to
responsibility, moral judgements upon conduct, and deliberation - and an
explanatory role, the reason for which someone acted. Actions are primarily
movements of the body, and individuated by their causes and effects. What
causes actions are desires (broadly seen) and beliefs, both qua states,
causing in the right way so that they are the reasons for which the agent acted.
Under ideal circumstances the desired state of affairs that is represented as
the content of the propostional pro attitude is (one of) the effect(s) of the
action performed. Besides the motivating desire, the intentional action comes
forth as the result of a practical argument where a connecting (instrumental)
belief channels the desired goal onto some action that is judged upon to be the
one the agent shall perform. Following this is the formation of an intention to
act or the action itself. The step from the concluding judgement to the
intention or action is arguably one from a prima facie judgement to an
unconditional one. Nevertheless, reasons for actions and practical reasoning
are built on the same practical argument, answering to the same pragmatic
question: 'What shall I do?' This pragmatic question is determined by
contextual features of the agent's situation, features that melts together with
the contextual features of a question after explanation, after why the agent
did what he did. These features goes beyond van Fraassen's general
considerations and points further to the normative constitution of the
implicated concepts - a constitution we may call folk-psychology.
Even
if desires are attributed (normative) entities we may say something about their
nature. Of special importance for the above account is that desires are
dispositional and teleological states. Teleological in that they are states
with which the world must fit, and dispositional by disposing the holder of the
state to act so that the world is made to fit. The dispositional analysis
(through the teleological argument) shows the close relationship between
actions and desires, something which in turn displays the normative and
interdefined character of intentional concepts. At the same time by being
defined as dispositional states, desires becomes essentially person-dependent. We must, so it
seems, agree with Anscombe when she says: "The primitive sign of wanting
is trying to get."[202]
On these considerations a
distinction from Schueler (1995) between desires proper and pro attitudes were
deemed unnecessary, because he fails to see the "internalising" point
of the teleological argument and its dispositional ground. Something is a
desire-state if it has the conative direction of fit and dispose the agent to
act. That is what it is being a desire-state, whether it is felt or not.
All this points to a Humean account
of motivation. I think in many cases that to distinguish between Reason and
desires (or passions) is unnatural and confusing. Hume's claim that Reason is
and ought only to be the slave of the passions, must be seen to take into
consideration several more passions than the one's he mentions if it is to be
true. Or rather only point to a function with an undefinitely large number of
instances, instead of a given set of passions. Whatever it means limiting
practical Reason to an instrumental role, it does not mean much more than
limiting fish to being in water. Whatever thought we may come up with, if it shall
motivate us to act, we must have a desire (or pro attitude) for the content of
the thought, and from that on, Reason works only instrumentally towards a
judgement that some action should be performed.
·
Anscombe (1963) - Anscombe, G. E. M.: Intention, Blackwell,
Oxford 1963.
·
Aristotle - Aristotle: Movements of Animals, in The
Complete Works of Aristotle, Vol. 1, pp. 1086-1096.
·
Aristotle - Aristotle: Nicomachean Ethics, in The
Complete Works of Aristotle, Vol. 2, pp. 1729-1867.
·
Audi (1982) - Audi, R.: "A Theory of Practical
Reasoning", American Philosophical Quarterly 19, Jan. 1982, pp.
25-39.
·
Audi (1986) - Audi, R.: "Acting for Reasons", The
Philosophical Review XCV, Oct. 1986, pp. 511-546.
·
Austin (1953) - Austin, J. L.: "How to Talk - Some Simple
Ways", Procedings of the Aristotelian Society 53, 1953, pp.
227-246.
·
Carroll (1895) - Carroll, L.: "What the Tortoise Said to
Achilles", in Gödel, Escher, Bach: an Eternal Golden Braid,
Hofstadter, D. R., Vintage Books, New York 1979, pp. 43-45.
·
Charles (1984) - Charles, D.: Aristotle's Philosophy of Action,
Cornell University Press, Ithaca 1984.
·
Churchland (1995) - Churchland, P. M.: The Engine of Reason, the
Seat of the Soul: A Philosophical Journey into the Brain, The MIT Press,
Cambridge 1995.
·
Darwall (1983) - Darwall, S.: Impartial Reason, Cornell
University Press, London 1983.
·
Davidson (1963, 1980) - Davidson, D.: "Actions, Reasons, and
Causes", in Essays on Actions and Events, Clarendon Press, Oxford
1980.
·
Davidson (1970, 1980) - Davidson, D.: "How is Weakness of the Will
Possible?", in Essays on Actions and Events, Clarendon Press,
Oxford 1980.
·
Davidson (1971, 1980) - Davidson, D.: "Agency", in Essays on
Actions and Events, Clarendon Press, Oxford 1980.
·
Davidson (1975, 1984) - Davidson, D.: "Thought and Talk", in Inquiries
on Truth and Interpretation, Clarendon Press, Oxford 1984.
·
Davidson (1976, 1980) - Davidson, D.: "Hempel on Explaining Action",
in Essays on Actions and Events, Clarendon Press, Oxford 1980.
·
Davidson (1978, 1980) - Davidson, D.: "Intending", in Essays
on Actions and Events, Clarendon Press, Oxford 1980.
·
Davidson (1980) - Davidson, D.: Essays on Actions and Events,
Clarendon Press, Oxford 1980.
·
Davidson (1985) - Davidson, D.: "Adverbs of Action", Essays
on Davidson: Action and Events, (ed.) Vermazen, B. & Hintikka, M. B.,
Clarendon Press, Oxford 1985.
·
Davidson (1987) - Davidson, D.: "Problems in the Explanation of
Action", in Metaphysics and Morality: Essays in Honour of J. J. C.
Smart, (ed.) Pettit, P. & Sylvan, R. & Norman, J., Oxford 1987.
·
Davis (1984) - Davis, W. A.: "Two Senses of Desire", Philosophical
Studies 45, 1984, pp. 181-195.
·
Dennett (1981, 1987) - Dennett, D. C.: "Three Kinds of Intentional
Psychology", The Intentional Stance, Cambridge, MA: The MIT Press/A
Bradford Book, 1987.
·
Dennett (1987) - Dennett, D. C.: The Intentional Stance,
Cambridge, MA: The MIT Press/A Bradford Book, 1987.
·
Dretske (1988) - Dretske, F.: Explaining Action: Reasons in a
World of Causes, The MIT Press, Cambridge Mass. 1995.
·
Evnine (1991) - Evnine, S.: Donald Davidson, Polity Press,
Cambridge 1991.
·
Foley (19?) - Foley, R.: "Prudence and the Desire Theory of
Reasons", The Journal of Value Inquiry, pp. 68-73.
·
Foot (1978) - Foot, P.: "Reasons for Action and
Desires", Virtues and Vices and Other Essays in Moral Philosophy,
Oxford 1978.
·
Frankfurt (1991) - Frankfurt, H. G.: "The Faintest Passion",
unknown, 1991.
·
Gjelsvik (1990) - Gjelsvik, O.: "On the Location of Actions and
Tryings: Criticism of an Internalist View", Erkenntnis 33, 1990,
pp. 39-56.
·
Gjelsvik (1995) - Gjelsvik, O.: "Valuing and Desiring",
unpublished paper, 1995.
·
Goldman (1970) - Goldman, A.: A Theory of Human Action,
Princeton University Press, Princeton 1970.
·
Gordon (1987) - Gordon, R. M.: The Structure of Emotions:
Investigations in Cognitive Psychology, Cambridge University Press, New
York 1987.
·
Hempel (1965) - Hempel, C. G.: Aspects of Scientific Explanation,
Free Press, New York 1965.
·
Hornsby (1980) - Hornsby, J.: Actions, Routledge & Kegan
Paul, London, 1980.
·
Humberstone (1987) - Humberstone, I. L.: "Wanting as
Believing", Canadian Journal of Philosophy 17, 1987, pp. 49-62.
·
Humberstone (1992) - Humberstone, I. L.: "Direction of Fit", Mind
101, Jan. 1992, pp. 59-83.
·
Hume (1740, 1978) - Hume, D.: A Treatise of Human Nature, Second
Editon, Clarendon Press, Oxford 1978.
·
Lenman (1996) - Lenman, J.: "Belief, Desire and Motivation: An
Essay in Quasi-Hydraulics", American Philosophical Quarterly 33,
July 1996, pp. 291-301.
·
Mele (1988) - Mele, A. R.: "Effective Reasons and
Intrinsically Motivated Actions", Philosophy and Phenomenological
Research XLVIII, June 1988, pp. 723-731.
·
Myers (1995) - Myers, R. H.: "On the Explanation, the
Justification and the Interpretation of Action", Noûs, 1995, pp.
212-231.
·
Nagel (1970) - Nagel, T.: The Possibility of Altruism,
Princeton 1970.
·
Nelkin (1993) - Nelkin, N.: "The Connection between
Intentionality and Consciousness", Consciousness: Psychological and
Philosophical Essays, (ed.) Davies, M. & Humphreys, G. W., Blackwell,
Oxford 1993.
·
Pears (1984) - Pears, D.: Motivated Irrationality, Oxford
University Press, Oxford 1984.
·
Pettit & Smith
(1990) - Pettit, P. & Smith, M.:
"Backgrounding Desire", The Philosophical Review XCIX, Oct.
1990, pp. 565-592.
·
Pettit (1987) - Pettit, P.: "Humeans, Anti-Humeans, and
Motivation", Mind XCVI, Oct. 1987, pp. 530-533.
·
Platts (1979) - Platts, M.: Ways of Meaning, Routledge &
Kegan Paul, London 1979.
·
Platts (1991) - Platts, M.: Moral Realities, Routledge,
London 1991.
·
Price (1989) - Price, H.: "Defending Desire-as-Belief", Mind
98, 1989, pp. 119-127.
·
Quinn (1993) - Quinn, W.: Morality and Action, Cambridge
University Press, Cambridge 1993.
·
Ross (1930) - Ross, W. D.: The Right and the Good,
Clarendon Press, Oxford 1930.
·
Schueler (1991) - Schueler, G. F.: "Pro-Attitudes and Direction
of Fit", Mind C, Apr. 1991, pp. 277-281.
·
Schueler (1995) - Schueler, G. F.: Desire: Its Role in Practical
Reason and the Explanation of Action, The MIT Press, Cambridge, Ma. 1995.
·
Searle (1979) - Searle, J.: Expression and Meaning,
Cambridge University Press, Cambridge 1979.
·
Searle (1983) - Searle, J.: Intentionality: An Essay in the
Philosophy of Mind, Cambridge University Press, Cambridge 1983.
·
Shaw (1989) - Shaw, D.: "Hume's Theory of Motivation", Hume
Studies 15, April 1989, pp. 163-183.
·
Smith (1987) - Smith, M.: "The Humean Theory of
Motivation", Mind XCVI, Jan. 1987, pp. 36-61.
·
Smith (1994) - Smith, M.: The Moral Problem, Blackwell,
Oxford 1994.
·
Stampe (1987) - Stampe, D. W.: "The Authority of Desire",
The Philosophical Review XCVI, July 1987, pp. 335-381.
·
Stich (1983) - Stich, S.: From Folk Psychology to Cognitive
Science: The Case Against Belief, The MIT Press, Cambridge, Ma. 1983.
·
Stroud (1977) - Stroud: Hume, Routledge & Kegan Paul,
London 1991.
·
Vadas (1984) - Vadas, M.: "Affective and Non-affective
Desire", Philosophy and Phenomenological Research XLV, Dec. 1984.
·
Van Fraassen (1980) - van Fraassen, B. C.: The Scientific Image,
Clarendon Press, Oxford 1980.
·
Wallace (1990) - Wallace, R. J.: "How to Argue about Practical
Reason", Mind XCIX, July 1990, pp. 355-385.
·
Warner (1987) - Warner, R.: Freedom, Enjoyment and Happiness,
Cornell University Press, Ithaca 1987.
·
Williams (1979) - Williams, B.: "Internal and External
Reasons", in Rational Action: Studies in Philosophy and Social Science,
(ed.) Harrison, R., Cambridge University Press, Cambridge 1979, pp. 17-28.
·
Williams (1989) - Williams, B.: "Internal Reasons and the
Obscurity of Blame", Logos 10, 1989, pp. 1-11.