Fairy Tale about Rationality and Reasons

advertisement
Tale about Rationality and Reasons
Once upon a time philosophers lived happily in their forest, firmly believing in
Davidsonian truth which claims that our mental states, propositional attitudes, constitute
reasons, and very probably, that those reasons are also causes of our actions. In the real
fairy tale, at this point, the witch would come in order to interrupt our hero’s Arcadian
peace. In our story, however, some people appeared, those that are never satisfied with
what they have, asking awkward questions and thus disturbing the long lasting peace.
One of those was John Broome.
What did philosophers happily believe? Human actions are based on reasoning
and deliberating about the things of the world. Reasoning itself is divided into its
theoretical and practical part. Reasoning is governed by rules of rationality belonging
respectively to the theoretical as well as the practical way of reasoning. Rationality
governs reasoning in the way that the reasoning process (if it is at all conceived as a
process) results with the action that is appropriate or right, in the sense that elements of
reasoning, beliefs, desires, and so on, explain why this and not same other action is
chosen. In other words, elements of the reasoning process are reasons for action.
Therefore, in the field of theoretical deliberation, rules of rationality sort out the
"right" kind of beliefs; in the field of practical deliberation, they give us the "right" kind
of actions. Here are some examples of theoretical and practical rules of rationality that
govern reasoning, respectively:
a) "if you at t believe p than you should not at t believe non-p";
"If you believe p and you believe that p implies q, than you should believe q".
1
b) "if you intend to F at t than you should not at t intend not to F".
"if you intend to F and believe you cannot achieve F unless you do G, you should do
G".
Our philosophers felt safe in thinking that elements of reasoning, attitudes,
provide reasons for beliefs and actions and also explain why people act in the way they
do. For example, if Peter wants to travel to the place P and believes that the only way to
do that is to take the train, then Peter ought to take the train. Peter's desire or intention to
travel, together with the relevant belief, makes his intention to take the train rational,
provides him reasons to take the train, and explains why he intends to take the train.
Rationality, reasons and the explanations, in this theoretical view, come in the same
package. That is why our philosophers believed that rationality goes along with reasons
or even that rationality consists in responding to reasons.
However, the first indication of uneasiness enters this Arcadian serenity with
Bernard Williams' famous "gin" example. Imagine a person who wants to drink gin and
tonic and believes that the way to get it is to take the glass containing the desired liquid.
Yet, unknown to him, somebody put petrol in the glass instead of gin. Does the person in
question have the reason to take the glass? Intuitions widely differ. Williams, (reason)
internalist himself, did not give (and could not give) the definite answer. According to
our philosophers and their doctrine, the story would go this way: since reasons are
internal, and it has been beyond the person's power to know about petrol in the glass, he
is rational, his desire and belief explain his action, and, finally, he does have a reason to
take the glass. Having a reason, he ought to take the glass, which means that doing
otherwise he would be irrational. Same other people disagree.
2
Let me first briefly expose basic concepts on which the standard or Arcadian
doctrine is founded. These concepts are rationality, reasons and normativity.
I will start with the concept of rationality. Customarily conceived, rationality is a
property that a particular kind of agents are endowed with. For the property of rationality
to be actualized, a number of faculties or capacities are required. The capacity is
understood as aptness to attain what rationality demands. Therefore, the necessary
condition to have a property of rationality is to be endowed with capacities to perform
what property of rationality demands. The argument sometimes goes as follows:
1. According to our evidence or experience, there are rational people.
2. To be rational, one ought to satisfy necessary conditions for rationality.
 Rational people satisfy those conditions.
The above form of the argument goes back to Kant and is known as the transcendental
argument. The first premise is the empirical one. The second premise is the
transcendental one. Let us assume that one of the necessary conditions for rationality is:
"If one believes p and one believes that if p then q, than one believes q". So according to
this instance, the premise would be "To be rational, one ought to, If one believes p and
one believes that if p then q, then one believes q".
According to more recent interpretations, the second premise has the following form:
2’. Necessarily, if one has the property of rationality, then "if one believes p and one
believes that p implies q, then one believe q";
Or, shorter:
Necessarily, if one has P, then one p.
3
This approach to rationality Broome calls the property theory of rationality (p. 118) and,
as we will see later, confronts to it the theory he endorses, the source theory of
rationality.
Let us look at reasons. Reasons can be different things, but for something to be a
reason, according to the Arcadian view, it necessarily has to be recognised and accepted
by the agent, grounded on beliefs, desires or some other attitudes. A desirable object can
give me a reason, but only if I desire it. The rushing car can be the reason for my jumping
aside, but only on condition that I believe it1. It is clear that reasons are understood as
subjective and internal. I have reasons only if they can move me2. Otherwise, I have no
reasons.
Finally, the normativity is to be mentioned. If one is rational, rationality puts on
him some obligations or requirements. What is the nature of those requirements? When
rationality requires something from me, should I respond to it, and if I should, how? This
kind of question is known as a normative question. Broome offers a very meticulous
taxonomy of normativity. When rationality requires not to have contradictory beliefs,
does it mean that one ought not to have contradictory beliefs because rationality requires
so? The positive answer to this question Broome (p. 184) formulates as
Strong Normativity: Necessarily, if rationality requires N to F, then N ought to F
because rationality requires N to F.
The normativity of rationality can be understood as a weaker claim (p. 185).
1
See: Jonathan Dancy, (2000), Practical reality, p. 157., Oxford U.P
The above description seems to refer to the reason internalism only. However, it is not
so. If Kantian duty is supposed to be a reason for action, than one should have a believe
attitude about this duty. In this case, again, attitude is a reason for action.
2
4
Weak Normativity: Necessarily, if rationality requires N to F, there is a reason for
N to F.
In this formulation "the fact that rationality requires you to F may not itself be a reason to
F" (p. 185). The weak rationality "associates rationality with normativity, but it does not
say that rationality is a source of normative requirements". (p. 185).
There is, however, one more approach besides two previous approaches to
normative question, According to this one, rationality consists in responding correctly to
reasons. This view of normativity of rationality puts the thing differently: "rationality
would not require something from you unless there was a reason for you for achieve that
thing" (187). The traditional (Arcadian) view very often endorses either strong
normativity or the view that normativity consists in responding correctly to reasons. My
concern in this paper, as well as my objections to Broome's theory, has to do primarily
with this latter aspect of normativity, namely, that it consists in responding correctly to
reasons.
Going back to the exposition of the Arcadian theory, we can say that, whatever
normativity is supposed to be (either strong normativity or requiring responding correctly
to reasons), it is somehow intrinsically built into the concept of rationality connecting it
with the concept of reasons. So, let us summarise.
Attitudes, beliefs and intention (for simplicity reasons, I will mention only these
two attitudes), if they satisfy rationality demands they are reasons for actions. In the
theoretical reasoning, if you believe p and you believe that p implies q, you have a reason
to believe q. Otherwise, you are irrational. In the case of practical rationality, if you form
an intention to F, you have a reason to F and you are irrational if you fail to do F. Or,
5
taking the case of instrumental reasoning, if you intend to F, your reason to F should be
transmitted to your reason to take the means to your end. If you fail to take the means to
your end, you are irrational. In both cases irrationality issues because either a) you fail to
believe q or do F in spite of the fact that rationality was a reason for that, or b) although
the rationality was a reason for you to do that, you fail to respond correctly to it. In both
cases, rationality is based on reasons. This reasons-based account of rationality seems
also to explain why we do what rationality requires from us.
That is an outline of the seemingly plausible and more or less convincing picture
that our slumbering philosophers have been dreaming. Now, we come to the dramatic
turn in our story, the appearance of the agitating philosophers with their awkward
questions.
Broome has critical objections to all three points I have sketched above and to
each point he offers an alternative account. However, before entering into a more detailed
elaboration of his view, let me first present the sketchy and intuitive idea that I assume
underlines Broome's theory. Although it is neither accurate nor precise enough, it can
help us (in very broad strokes though) understand what it all is about.
In this anticipating outline I will concentrate on Broome's scrutinizing the alleged
quality of the reasons-based account of rationality. Firstly, that requirements of
rationality are normative (either in the strong sense or in the way that rationality consists
in responding correctly to reasons) and, secondly, that those accounts explain our actions.
The intuitive idea, at least as I see it, is that our cognitive, rational endowment and
the world do not stay in such a perfect harmony as the reasons-based account would
presuppose. Sometimes, in the hostile environment, our cognitive set up matches the
6
reasons for accomplishing what rationality requires. However, it is certainly not the case,
Broome holds, that:
necessarily, if rationality requires from you that (if you belief that p and you believe that
p implies q, you have a reason to believe that q).
Or,
necessarily, if rationality requires from you that (if you intend to F and believe
that you cannot do F unless you do G, you have a reason to do G).
Why it is so? The following examples will illuminate the point. The first one is Broome's
example:
One can believe that the world was created in six days. That obviously implies that the
world was created in less then seven days. The conditional that contains these two
sentences is: "If one believes that the world was created in six days, then one believes
that the world was created in less then seven days". Does it mean that one ought to or has
a reason to believe that the world was created in less then seven days?
On the other hand, I am intending to travel to Dubrovnik and I believe that I
cannot reach there in time unless I take a plane. Does it mean that, necessarily, given the
intention and belief I have, I ought to (or have a reason to) take the plane?
In the first example, it is clear that one has no reason to believe that the world was
created in less then seven days. In the second example, I have no reason to take a plane if
I have no reason to travel to Dubrovnik. If it is the case, the intention for the end (to
travel to Dubrovnik) ought not to be transferred to the taking of the means for it. In this
example, my intention alone cannot be the reason for action, as in the first example, my
7
belief in the antecedent of the conditional (that the world was created in six days), if
false, cannot be a reason for believing the consequent.
What is so wrong with the reasons-based theory? Firstly, Broome argues, it
simply is not the case that the intention can bring about that you ought to F. In the same
way, your belief in p, given that if p than q, cannot make it the case that you ought to
believe that q. If this were the case, your attitudes would create reasons out of nothing.
The argument based on this claims is known as the bootstrapping argument, and was first
raised by Micheal Bratman3.
Secondly, if the first claim is true, then rationality neither a) provides reasons for
that what is required by rationality, nor, b) does rationality consist in responding correctly
to reasons.
Finally, if the first and second claim hold, the reasons-based theory does not
explain what needs to be explained. What needs to be explained is: why you are irrational
if you do not F. The answer is: you are irrational if you do not satisfy what rationality
requires from you, but you can be rational even if you have no reason to F. This answer
makes clear that being rational and having reasons or responding to reasons are not the
same things.
How is this to be explained?
As I said above, Broome offers alternative accounts on all three basic concepts on
which the reasons-based theory is founded. Let us tackle the concept of rationality first.
He chooses the source theory of rationality instead of the property theory. According to
this, rationality is not a property in the first place. Rationality is a source of requirements,
3
Bretman, Intention, Plans and Practical reason, (pp. 24 -7). Kolodny, Broome.
8
among other possible sources of requirements, as morality or prudence are. The
fundamental feature of rationality is the fact that it ultimately depends only on the
property of your mind. Rationality supervenes on our mind.
If your mind in situation s1 contains particular mental states and particular
relations among them, and their contents are true, your mind has the property of
rationality. Given that your mind contains the same mental states in situation s2, but some
of them are false, your mind must have the same properties, including rationality. So, you
are rational in situation s1 if and only if you are rational in situation s2.
In accounting for rationality, Broome is a strong internalist. But, he is not an
internalist accounting for reasons. Again, reasons could be different things, but some of
them should be firmly rooted in the world out there. Because of this kind of reasons,
rationality and reasons cannot come in the same package. You can be rational in
believing q or in intending F, even if you have no reason to believe p or to intend F.
Rationality requires many things from us, but those requirements may not be normative.
So, the story can finish here. The supremacy of a great part of Broome's
arguments over the Arcadian view can be proclaimed. But it is not quite so. It seems to
me that a certain aspect of the theory of rationality conceived as responding correctly to
reasons still can be vindicated.
I will start with Broome's distinction of the stronger and weaker variant of this
theory of rationality. The stronger variant is the thesis of equivalence.
Equivalence: Necessarily, you are rational if and only if you respond correctly to
reasons. That means that, ultimately, rationality can be reduced to reasons.
The weaker variant is formulated as a thesis of entailment.
9
Entailment: Necessarily, if you are rational you respond correctly to reasons.
Broome strongly denies the equivalence thesis and allows for the possibility of the
entailment thesis.
I will try to show that the equivalent thesis can be vindicated, and I will neglect
the entailment thesis. Of course, my claim concerning the equivalent thesis is in need of
further clarification. Let me say how.
Rationality is a complex concept applying to a vast area of activities. There are
several kinds (theoretical, practical, instrumental …) and many aspects of rationality
(coherence, consistency, inference and so on). Requirements of rationality are tailored
according to the aspects of rationality. Therefore, for example, the requirement "if you at
t believe p, than you should not at t believe non-p" corresponds to the coherence aspect of
rationality. If one is fully rational, one is rationally satisfying all requirements of
rationality4. The equivalence thesis demands that, necessarily, N is (fully or less than
fully) rational if and only if he satisfies requirements of rationality in the way that he
responds correctly to reasons. In addition, at least some of the requirements of rationality
are such that satisfying them, you respond correctly to reasons.
The one who denies the equivalence thesis, as Broome does, must argue that there
is no any requirement of rationality such that, satisfying it, you respond correctly to
reasons.
Accordingly, if the equivalence thesis is understood in the way I have ascribed to
Broome, the one who wants to vindicate the equivalent thesis needs to show that there is
4
I am neglecting here the possibility that requirements can be different in different
possible worlds.
10
at least one aspect of rationality with a corresponding class of requirements that satisfy
the equivalence. That will be my strategy.
Before I put forward my claim, let us see whether in Broome's theory there is a
possibility for supporting the equivalence thesis. Broome indeed allowed, hypothetically,
for such a possibility (although he denies it a bit later). Nevertheless, it is worthy to
examine this possibility. "If attitudinal reasons exist, they might provide support for the
equivalence". The possibility is exposed as follows: forming a belief (based on a direct
evidence) "it is not raining", this belief constitutes a reason for you not to believe it is
raining (see Ch. 4). Such a belief that it is not raining may be the result of an unconscious
process, or it can be obtained by conscious reasoning. If it is an unconscious process, "the
operation of this process counts as responding to your reason, which is your belief that it
is not raining" (Ch. 4., p?). If this is a conscious process, Broome believes it need not
involve any second-order belief.
In this restricted form of reasoning, rationality is equivalent to responding
correctly to reasons. Of course, conditional on the assumption that attitudinal reasons
exist. However, unfortunately, Broome argues, attitudinal reasons do not exist.
In this point, I partly agree with Broome. Although I am not sure whether
attitudinal reasons exist or not, I am ready to accept that they do not exist, at least as far
as it concerns my argument. I have been tackling the possibility abovementioned because
I will further explore Broome's description of the way in which it connects reasoning
with reasons. In this description, the attitude whose content is "it is not raining" is a
reason for a thinker not to believe it is raining. Since Broome understands attitudes as
dispositions, the belief that it is not raining can unconsciously be (causally or in some
11
other way) connected with its attitudinal reasons. The relation in question we can count
as responding correctly to reasons.
Having in mind the above Broome's example, in my objection to Broome's theory
I will try to show
a) that there is a particular aspect of reasoning that must be recognized as the requirement
of rationality,
b) that this particular aspect of rationality consists in responding correctly to reasons, and
c) that those reasons are not attitudinal reasons.
The first step of my argument is to show that there is a requirement of rationality
that Broome does not mention explicitly.
Let us start with one of Broome's rationality requirements, the modus ponens
requirement:
(MPR) Rationality requires of N that, if N believes that p and N believes that if p
then q, N believes that q.
MPR (or some of its modifications) is, Broome holds, a genuine requirement of
rationality. I would like to remark here that, instead of modus ponens, which is the
content of this requirement, we certainly could take a more simple logically valid form (I
will use conjunction elimination).
If MPR is the genuine requirement of rationality so it should inevitably be a kind
of reasoning that is fairly neglected in Broome's work. What I mean here is the piece of
reasoning that leads me from the fact that a particular inference form seems to me correct,
to the belief that this inference form is correct. If I am required to believe q based on my
believing that p and my believing that if p than q (as MPR requires), it seems doubtless
12
that I am also required to believe that the inference form on which I have based my
reasoning is correct.
Let me exemplify it in this way:
Suppose that a friend told me that his son had obtained a scholarship and enrolled the
university. From this conjunction, (p&q), I can easily conclude that his son obtained a
scholarship, (p). Rationality required of me to believe p (or q) if I believe (p&q).
However, it cannot be the case that rationality does not require of me to believe that,
when faced with the instance of the conjunction elimination transition, I believe that this
transition is correct. In addition, I will try to show that I have a reason to believe that
transition T is correct.
Hence, I am not arguing here that conditional on my believing that (p&q) I have a reason
to believe p. My argument, inspired with C. Peacocke account for understanding logical
constants is this:
Facing this simple inference form (transition)
(p&q)
p
I am able to form the belief that this transition is correct5. This impression that T is
correct is a reason for me to believe that the transition is correct. Even more, this reason
is a sufficient reason. Note that the transition in question is a valid transition. This kind of
a valid transition has a feature that some other transitions may not have. It has a feature
that it is obviously valid, or correct, looking from the naïve subject's perspective. The
5
The correctness I am talking about is of course the logical property of validity.
However, from the standpoint of a logically naive subject, it is correctness; in the
argument, I will continue to use this term.
13
obviousness of this impression is not based on any further inference. It is direct, and as
such, primitive. The class of transitions that has the feature of primitive obviousness is a
subclass of the class of valid inferences.
I have to argue now that the reasons in my example are not an attitudinal reasons.
If I find that a transition is correct, the content of my attitude is about correctness. The
correctness of a transition consists in the fact that the instance of the transition always
preserves the truth. Again, looking from the naïve subject's perspective, this can be
formulated as: it cannot be the case that, if premises are true, the conclusion is false. It is
important to notice two things. First, validity, as a logical property is a fact independent
of my attitudes. Secondly, the impression of correctness, corresponding to validity, is an
attitudinal fact, but this fact is primitively obvious, and, furthermore, having it, one is
infallible.
The analogy with pain would be helpful here. The majority of theoreticians hold
that if it seems to one that he is in pain, he is in pain. Nevertheless, pain is a mindindependent fact. In the same way there is a class of simple transitions that have the
following feature. Namely, having the impression that T is correct is partly determined by
understanding the connective. Understanding the connective is a necessary condition for
mastering it. If I understand the connective, I cannot but use it in a definite transition, in
our example, in the instance of conjunction elimination. If this claim is correct, I have
shown the plausibility of the second step in my argument, namely, that the requirement of
rationality one is satisfying here consists in responding correctly to reasons.
The third step is to argue that the reason in question, the fact that the transition is
valid, is not an attitudinal reason.
14
Perhaps in the above sentence I have presupposed too quickly the fact that
validity is mind-independent. We should start from a kind of Euthyphro dilemma. The
question could be: whether my impression that T is correct determines the T's validity, or
the T's validity determines my impression? The realist about logical entities will accept
the first horn of the dilemma. The proponent of anti-realism or psychologism will
naturally choose the second horn of the dilemma. My theoretical sympathies are strongly
on the realist side, or, better to say, on one of the possible moderate realist positions.
However, the decision on the ontological status of the property of validity presupposes
the choice among the above theories. As a realist in this area, I am strongly inclined to
say that validity is a mind-independent fact.
15
16
Download