Deliberation and Acting for Reason

7/25/2019 Deliberation and Acting for Reason

http://slidepdf.com/reader/full/deliberation-and-acting-for-reason 1/32

Deliberation and Acting for Reasons

Nomy Arpaly

Brown University

Timothy Schroeder

Ohio State University

Theoretical and practical deliberation are voluntary activities, and like all

voluntary activities, they are performed for reasons and can be carried out

wisely or unwisely, reasonably or unreasonably, rationally or irrationally.

As a consequence, there must be processes that are nondeliberative and

nonvoluntary but that nonetheless allow us to think and act for reasons,

and these processes generate the voluntary activities making up ordinary deliberation. These nondeliberative, nonvoluntary processes by means of

which we are able to deliberate for reasons must be fundamental, on pain

of regress. As a result, the usefulness of deliberation to rational belief and

action is intermittent, contingent, and modest.1

We would like to thank a number of people for helpful discussion of the ideas presented

here, including Nicolas Bommarito, John Broome, Sarah Buss, Ben Caplan, David Chris-

tiensen, Justin D’Arms, James Dreier, David Estlund, Nico Kolodny, Philip Pettit, PeterRailton, Richard Samuels, Valerie Tiberius, and our anonymous referees. This article also

benefited from versions of it being presented to the philosophers of Massachusetts Insti-

tute of Technology; the Moral Psychology Research Group; Ohio State University; the

Ohio Philosophical Association; the Society for the Theory of Ethics and Politics; State

University of New York, Albany; University of California, Davis; and the Workshop on

Moral Expertise.

1. The general topic of interest to us has been treated by a number of philosophers

(including one of the authors of the present work). See, for example, Arpaly 2000, in

which rationality without deliberation is discussed; Arpaly 2003, chapter 2, in which deli-

beration itself is given as an instance of acting for reason without deliberation and a basic

regress argument is made briefly; Dreier 2001, which deals in depth with a related

Philosophical Review , Vol. 121, No. 2, 2012

DOI 10.1215/00318108-1539089

q 2012 by Cornell University

209



1. The Nature of Deliberation

Allan sees that the tea shop has no pastries today, and he wonders—

deliberates—about where to go instead. He considers various optionsand settles on the cafe up the street. Mira is asked whether she really

believes in animal rights or only the moral importance of animal welfare,

and she starts pondering— deliberating about— the differences between

the two, ending up with the conclusion that she does believe in animal

rights. The frequency of deliberation in our lives could be exaggerated,

but it is entirely common and entirely familiar. In this first section, we

sketch out what deliberation is with sufficient precision to go on.

Deliberation is commonly divided into theoretical and practicaldeliberation. Theoretical deliberation is primarily concerned with what

to believe, while practical deliberation is primarily concerned with what

to do. The product of theoretical deliberation is perhaps a new belief (or

entrenchment of an existing belief), or a new credence in a possible state

of affairs, or a sincere thought about the truth of a proposition, or some-

thing similar. The product of practical deliberation is perhaps a new

intention (or entrenchment of an existing intention), or an action, or

something similar—or the product of practical reasoning is a new belief

that now taking a certain course of action would be best overall, or a belief

that it would be fitting, or something similar.2 For our purposes, it is not

important what, exactly, the product of either sort of deliberation is, so

long as the above proposals are not all radically misguided.

Both sorts of deliberation are kinds of actions. (See also Hookway

1999 on epistemic deliberation and Railton 2009 on practical deliber-

ation.)3 They are mental actions, or perhaps mental actions conducted

issue; and Railton 2004, 2009, in which the first steps of the present argument are taken

(in the jargon we deploy later in this article, Railton makes the regress argument against a

combined version of Present Deliberation and Present Recognition). We see the present

article as a particularly noteworthy addition to this growing literature in (1) having a well-

developed theory of deliberation; (2) considering the full range of published arguments

for the thesis that deliberation has a privileged role in our thinking and acting for reasons;

(3) deploying a wide variety of counterarguments to the aforementioned arguments; and

(4) providing a distinctive account of what role deliberation actually plays in making us

responsive to our reasons.

2. This latter possibility would, of course, render practical reasoning a subtype of theoretical reasoning.

3. We are committed to no specific theory of action and hope to be as uncommitted

as possible in the course of this work. But if deliberation is an action, then it does seem to

us that there will be instances of deliberation that are actions of a less-than-paradigmatic

N O M Y A R P A L Y A N D T I M O T H Y S C H R O E D E R

210



with supplementary nonmental aids (notepaper, calculators, interlocu-

tors, the sound of one’s own voice speaking one’s thoughts aloud, and so

forth). We will, however, focus on the mental aspect of deliberation, as

mixed cases do not alter the plausibility of the conclusions for which we

will press.4 Purely mental deliberation, then, is a purely mental activity,

carried out by means of performing various mental actions aimed at some

suitable end. Both the actions and the end are important to deliberation.

For a mental activity to be deliberation, it must be aimed at deter-

mining what to think or do.5 As we said a moment ago, we maintain an

ecumenical stance on just what this amounts to, but there are limits to

ecumenism. A mental activity aimed only at amusing the subject is not

deliberation. A mental activity aimed only at facilitating sleep is not delib-eration. Determining what to think or what to do is thus a constitutive

end of deliberation, in the same way that persuasion is a constitutive end

of arguing, or getting coffee is a constitutive end of going for coffee.

Deliberation can fail to achieve its end, but it is an activity that aims to

achieve that end (perhaps among others).

For a mental activity to be deliberation, it must also be carried out

by means of particular mental acts: bringing to mind—to conscious-

ness— various ideas (general or particular, abstract or concrete) or ima-ges (visual, auditory, gustatory, and so forth). There are other ways to

determine what to do or what to believe, but these are not deliberation. As

a person with the aim of determining what to do, Maria might be aware

that she will come to be resolved on what to do if she can just enter a state

of meditative tranquility in which she does not do any conscious thinking

sort: instances of deliberation from which one is alienated (“I could hardly believe I was

contemplating covering up my misbehavior”), deliberation conducted for an end that isnot transparent to one (“I found myself thinking again about the facts of the case, uncer-

tain as to why my mind was drawn to them”), and deliberation conducted absentmindedly

(“I suddenly realized that, in my mental wanderings, I had started going over all the

reasons I had given myself for ending my past romances, weighing up each one”), to

choose just three examples. If the reader wonders about such cases, we hold that these

examples of apparent deliberation are actions just in case the counterpart cases involving

overt bodily movement are actions—and if they are not true actions, then they are not

true instances of deliberation either, for deliberating is something we do .

4. Note the existence of arguments that there is no principled difference between

“purely” mental actions and “mixed cases.” See, for example, Clark and Chalmers 1998.5. Perhaps there is also deliberation about what to feel, or what to intend, that is

distinct from deliberation about what to think or to do. But we are skeptical—and in any

case, the existence of such deliberation would not substantially change the arguments

that follow.


211



but simply empties her mind. And it might well be true that entering this

state is a mental act that Maria can perform, aiming through it to deter-

mine what to do. But Maria would not be deliberating if she were to

proceed in this way.

Mental acts are not guaranteed to be acts of deliberation even if

they bring ideas or images to mind with the goal of determining what to

do through filling one’s consciousness with them. Perhaps Randa finds

that what to believe about a difficult matter comes quickly to her if she

visualizes her fourth-grade teacher sternly asking, “What is the answer,

Randa?” This would not make her use of such an image into an act of

deliberation. Deliberation requires that one bring to mind ideas or im-

ages that are meant to have some rational relation to the propositions that would be the conclusion of one’s deliberation, with the aim of reaching a

conclusion.6 Even this is probably not a sufficient condition, but the full

complexities can be set aside for our purposes. Deliberation, at the least,

requires bringing to mind ideas or images meant to have some rational

relation to the topic being considered, in the service of reaching a con-

clusion about what to think or do.

Talk of mental acts that bring ideas and images to mind might be

unhelpfully unspecific. Experiences of deliberation suggest a variety of perhaps helpful concrete examples. Consider Harold, who is deliberat-

ing about whether to promise to meet his son in Calgary on Tuesday.

Harold engages in a mental action known colloquially as “searching his

memory” for potentially conflicting promises, a process that itself might

involve the mental act of holding in mind the idea of Tuesday, but might

not even involve that so much as an effortful attentiveness in Harold’s

present context. This mental act will produce as a product a notable

nothing in consciousness or some idea or image tied to a competing

promise. Or consider Albert: when he deliberates about the implications

of curved space-time for time travel, he visualizes a warped rubber sheet

with a ball rolling along it, and this image is relevant—in the context of

his thinking—to the truth of the claim that a straight line in a curved

space-time is one that could loop back on itself. Of course, in saying this

we assume that there is a way of making intelligible the idea that there is a

rational relationship between a visual mental image and a proposition.

6. Or perhaps propositions can express or represent the conclusions of one’s delib-

eration, rather than being identical to them. And perhaps propositions cannot express

plans, though plans can be the conclusion of practical deliberations (thanks to an anon-

ymous referee for this point), and so we need to make room for them as well.


212



One line of evidence in favor of this thought is found in the rich literature

on the content of sensory perception and imagery: see, for example,

Peacocke 1983 and Tye 1995. If imagery has content, then perhaps that

content can stand in rational relations to the content of beliefs. (But we

do not wish to commit to any particular claim about how imagery could be

rationally related to beliefs.)

In thinking about deliberation, it is important to focus on the most

basic voluntary acts that make up an episode of deliberation. While it can

be perfectly correct to say that in deliberating one considered X, Y, and Z

and concluded that P, it would be a mistake to think of this whole process

as an indivisible unit. If one is interested in deliberation as an active,

agential process, it is important to see that it is made up of agentialunits, and these units are, in each case, an act in which something with

apparent (theoretical or practical) relevance is brought into conscious-

ness as part of a larger plan to reach a theoretical or practical conclusion.

Seen on this fine-grained scale, we are reminded that deliberation is often

a far cruder process than the step-by-step valid deductions that philoso-

phers enjoy describing. It is common for the process of deliberation to

throw up a fragment—for example, Harold might suddenly say to him-

self, “Oh, the planning council!”—rather than a complete set of ideasentailing the conclusion—I can’t go to Calgary on Tuesday—that is,

nonetheless, reached on the basis of the fragment.

Thinking of these basic agential units highlights the fact that ex-

tended deliberative actions are made up not just of these basic exercises

of agency over thought but also of the occurrence in consciousness of

thoughts and feelings that serve as input to deliberation but that are not

brought to consciousness through exercises of agency. When Harold

searches his memory for possible conflicting commitments—a voluntary

mental act—it might result in his consciously remembering that he has

committed to going to a meeting of the planning council. This conscious

remembering itself is something that might well happen in Harold as a

result of something he does directly without it being something he does

directly ( just as sinking a free throw is something that happens as a result

of something done directly—shooting the ball—without itself being

done directly). The remembering happens in Harold, and having hap-

pened, prompts (as it might be) another action, such as Harold’s saying

to himself, “ Yes, I mustn’t forget the planning council meeting. Now, when does that get out?” This act, in turn, might lead to other nonvolun-

tary spontaneous thoughts or feelings, or it might lead to another volun-

tary mental act—and so it will go, until Harold settles what to do or gives


213



up (for the moment, at least) on settling what to do.7 In this way (as in so

many others), extended deliberative acts are like extended overt bodily

acts: they are made up of more elementary acts selected, in part, because

of the ongoing consequences of working through the extended action.

Deliberation is performed by performing a sequence of mental

actions. Thus, like any action, we would expect deliberation to be more or

less reasonable or unreasonable on any given occasion.8 Sometimes, de-

liberating is smart, reasonable, or rational, and sometimes it is stupid,

unreasonable, or irrational.9 Furthermore, even when it is rational to

deliberate in general, a person can deliberate in a specific way that is

rational or in a way that is downright foolish. That is, the specific mental

acts making up the process of deliberation will be reasonable or foolish,rational or irrational, in addition to their product being such. The word

‘rational’ in the expression ‘rational deliberation’ is thus not a redundant

one: irrational deliberation is possible.

To see the possibility of both sorts of irrational deliberation, con-

sider some cases. First consider cases in which it is irrational to deliberate,

and yet one deliberates. Rebecca deliberates on whether she should really

be getting off the Interstate at exit 33, ignoring the fact that she and her

car are already on the off ramp for exit 33 and situated in swift-movingtraffic—though she has no wish to get into an accident and knows that

distraction in swift-moving traffic promotes accidents. The rational thing

for Rebecca to do is surely to drive first and deliberate later, perhaps once

she has reached a calm surface street. Deliberating about the wisdom of

taking exit 33 at just that moment is irrational.

The second sort of case to be illustrated is the case in which the

agent is rational to deliberate in general, but in which the particular

deliberative acts taken by the agent are irrational. Suppose that Katie

7. Thanks to Peter Railton for pointing out to us the importance of the back-and-

forth in deliberation between basic mental acts and the spontaneous thoughts and feel-

ings such acts often engender.

8. Compare our arguments on this over the next few paragraphs to Joseph Raz, who

expresses the view that “reason is inherently normative” (Raz 1999, 68).

9. We take a rational person to be a person who acts as he or she does for very good

reasons and thinks as he or she does for very good reasons; we also take it that facts about

being smart or stupid, reasonable or unreasonable, arerelevant to claims about rationality and irrationality. We do not rely here on the idea that rationality is tied to what is ideal

from the perspective of an omniscient outside observer, or anything similar. Newton was

rational in holding his theory of universal gravitation, in spite of that theory’s falsity, in

our way of thinking about these things.


214



begins to consider the grammaticality of the phrase ‘I’m loving’. If she

searches her memory for different uses of the phrase, she is probably

proceeding reasonably. Calling to mind a pop song, an advertising cam-

paign, and sentences uttered by different people will bring to conscious-

ness evidence that the phrase has widespread use, and this is worthwhile

evidence to consider when determining what to think about a phrase’s

grammaticality: in conservative circles the verb ‘to love’ is not used in the

progressive, but there might be a linguistic change in progress. But sup-

pose that Katie begins to consider ways to express ongoing enjoyment of

something without using the phrase ‘I’m loving’. She calls to mind the

phrases ‘I’ve been enjoying,’ ‘I relish’, and others. She says to herself, “no

one needs to say he’s loving something.” Perhaps she finds the phrase“I’m loving” distasteful, and her emotions about it are influencing her

effort to determine what to believe about grammar. If she proceeds in this

way, it seems that Katie’s deliberation is now proceeding unreasonably—

there is a sort of wishful thinking at work in her. After all, the existence of

another way to say much the same thing as ‘I’m loving’ is no evidence one

way or the other regarding the grammaticality of ‘I’m loving’ itself. As

Katie knows, a phrase can be grammatical even though it is also optional

as a means of expressing some idea. Given that her goal is to determine what to think about the grammaticality of the phrase ‘I’m loving’, she has

taken a wrong turn. Each time she calls to mind another alternative ex-

pression, or declares to herself the inessentiality of the expression ‘I’m

loving’, she is acting unreasonably in her deliberation. And Katie’s un-

reasonable deliberation exemplifies just one way in which deliberation

can be improperly influenced: in addition to deliberation that is exces-

sively oriented toward what we wish for (that is, wishful thinking), there is

also fearful thinking, antagonistic thinking, and the like (see Lazar 1999).

There is a lot in the philosophical literature about rational delib-

eration, while irrational deliberation is rarely mentioned. There is talk

of people who deliberate correctly or who deliberate incorrectly, or of

people who deliberate under good conditions (“a cool hour”) or under

bad conditions (in emotionally fraught times), but there is little mention

of deliberating irrationally or unreasonably. Is there, then, some inde-

pendent standard for deliberation, failures of which are best called ‘in-

correctness in deliberation’ rather than ‘irrationality in deliberation’?

Consider what it means to deliberate correctly. It does not mean to delib-erate in a cool hour, because incorrect deliberation can be the product of

a cool hour (this sometimes seems to happen, for example, when major

economic policies are designed) and at least some correct deliberation is


215



the product of “hot” hours (for example, when righteous indignation

sharpens thoughts on the subject of human rights). Deliberating well or

correctly—the thing that cool hours make more likely—is deliberating

without fallacies and logical errors, as well as without such things as self-

deception or wishful thinking—in a way suited to determining what to

think or do. In other words, correct deliberation is in conformity to the

requirements of rationality. It thus appears that incorrect deliberation is

simply deliberation in violation of those requirements.

Acts of deliberation can be not only closer to or further from

what the reasonable or rational person would have done but also actions

performed for good or bad reasons. When one acts for a reason one does

not simply act or believe in accordance with the reason: one’s accordance with the reason is a result of the fact that one is acting for the reason. Thus,

consider Kiyoshi. It might be rational for Kiyoshi to turn on his television

because there is a program he enjoys that has just started, and this is a

good moment for a pleasant diversion. But Kiyoshi could have done what

is reasonable entirely by accident: he has a bad habit of turning on his

television when he walks in the door, and it was the habit he was acting on

rather than his good reasons. He could also have turned on the television

because of the good reasons to turn it on. There is acting in accordance with one’s reasons, and then there is the further step of acting for one’s

reasons. A crucial claim for our article is that what is true of turning on the

television is also true of deliberation. It could be reasonable for Katie to

call to mind the advertising campaign based on the phrase ‘I’m loving it’

when deliberating about whether the phrase is grammatical. But she

could call it to mind not because of whatever it is that makes it reasonable

to do so (that is, not for the right reasons) but because she has been

thinking about the advertising campaign off and on for a week, for a

course on marketing, and so would have called to mind the campaign

regardless of its relevance—“just out of stupid habit,” as she might have

said. If this had been the truth of the matter, then Katie would have done a

reasonable or rational thing but not have done it for the reasons there are

to do so. Likewise, Katie could also call the advertising campaign to mind

because her friend insistently urges her to consider the advertising cam-

paign. Perhaps the friend is inarticulate about why the advertising cam-

paign is relevant evidence (or even says something false about why it is

worth considering), and yet the friend will not relent until Katie accedesand deliberates about what relevance the advertising campaign might

have as evidence regarding the grammaticality of using ‘love’ in the pro-

gressive. Under this condition, again, Katie calls to mind a relevant idea in


216



deliberating about what to believe, but again she does not do so for the

reasons she has to do so. Cases like these, displaying a contrast between

deliberating about something for the right reason and deliberating about

something for no good reason or for no reason that bears on what to

think or do, are strong evidence that there is such a thing as acting for the

right reasons (and for better or worse reasons) even when the action is

deliberation itself.

Nothing so far should be very controversial. Though most philos-

ophers interested in deliberation have not provided us with theories of

deliberation or particularly naturalistic examples (Seidman 2008 pro-

vides one noteworthy exception), most of them have said nothing in-

compatible with the foregoing—which in any case we take to be largely familiar and conservative. And yet it seems to us that following the con-

sequences of this account of deliberation leads to some very controversial

theses indeed.

2. Deliberation and Responding to Reasons

The idea that deliberation (sometimes also known as reflection, reason-

ing, practical reasoning, theoretical reasoning, and so forth) is crucial to

thinking and acting for reasons is widespread in philosophy. Consider thefollowing claims:

When acting rationally, an agent undertakes to act in light of her belief

about what she has reason to do. She chooses her action because it is

supported by reasons. In this sense, rational action seems to embody a

distinctly rational form of motivation in which the agent guides herself by

the thought that an action is recommended by reason. This guiding

thought need not always be explicitly articulated. For rational action to

be possible, however, the agent must, at some level of awareness, concep-tualize the features to which she is responding as reason-giving. (Barry

2007, 232)

To make a reason for doing something the agent’s reason for doing it,

the reason must enter into a process of practical reasoning. . . . I do not

insist that every stage of the agent’s reasoning be consciously carried out.

The agent may have developed short-cuts in reasoning, and some of these

steps can become programmed into his brain and be carried out auto-

matically. (Chan 1995, 140 and n. 10)

There are reasons why animals act as they do, but . . . only in the most

tenuous sense can we say that they have reasons for acting as they do. Only

a language-using creature can reason and deliberate, weigh the conflict-


217



ing claims of the facts it knows in the light of its desires, goals and values,

and come to a decision to make a choice in the light of reasons. (Hacker

2007, 239)

For the person [who wants not to have a toothache, and who believes

that going to the dentist will cure the toothache] to act rationally, she must

be motivated by her own recognition of the appropriate conceptual

connection between the belief and the desire. (Korsgaard 1997, 222)

Once we are aware that we are inclined to believe or to act in a certain

way on the ground of a certain representation, we find ourselves faced

with a decision, namely, whether we should do that—whether we should

believe or act in the way that the representation calls for or not. Once the

space of reflective awareness—reflective distance, as I like to call it—opens up between the potential ground of a belief or action and the belief

or action itself, we must step across that distance, and so must be able to

endorse the operation of that ground, before we can act or believe. What

would have been the cause of our belief or action, had we still been oper-

ating under the control of instinctive or learned responses, now becomes

something experienced as a consideration in favor of a certain belief or

action instead, one we can endorse or reject. And when we can endorse

the operation of a ground of belief or action on us as a ground, then we

take that consideration for a reason. . .

. [W]e now both can have, andabsolutely require, reasons to believe and act as we do. (Korsgaard 2009,

31–32)

An agent [must] take herself to have good reasons for the option she

chooses. . . . [And her taking herself to have these good reasons] must not

change under appropriate reflection. (Tiberius 2002, 343)

This selection of quotations strikes us as representative of a

substantial portion of the literature on thinking and acting for reasons

(if the reader finds some of them unclear as expressions of ideas about deliberation, the significance of particular paragraphs will be made

clearer in the next section). There is, thus, some reason to explore the

weaknesses inherent in the idea that deliberation—somehow—makes

thinking and acting for reasons possible. To this we now turn, focusing on

the special case of the act of deliberation itself.

3. Regress Objections

It might be suggested that an act of deliberation can be an action

taken for reasons because it is an action with an appropriate relationship

to a previous occasion of deliberation. Call this approach “Previous


218



Deliberation.” In the past, perhaps, Harold engaged in some delibera-

tion, at the end of which he embraced the principle that he should not

make a promise without considering (that is, deliberating upon) possible

conflicting prior promises. And that is why, when he searches his memory

for such promises, he not only does what he has reason to do but does it

for a reason. His present deliberating for reasons is constituted by his

acting in accordance with the conclusion of previous deliberation about

how to deliberate in such cases. Or, in a slightly more sophisticated vein,

his present acting (that is, his present searching of his memory) for

reasons is constituted by his acting out of a habit, policy, or virtue that

was engendered by his previous deliberative conclusion about how to

deliberate.10 We assume, along with Anscombe (2000) and philosophers as

different as Davidson (1980) and Korsgaard (1996), that every action is,

by necessity, done for a reason (or reasons; we will stick to saying “a

reason”). It need not be done from overwhelming reasons or even

good reasons. The action need not be completely justified, but it needs

to be at least somewhat rationalized: done for a very bad, foolish reason,

perhaps, but still for that reason, and not for no reason at all.11 Since

deliberation is an action, deliberation cannot be deliberation without

10. The role of habit in making deliberation responsive to reasons appears to be

invoked by Chan 1995, as quoted above. It is emphasized by, for example, Hookway (1999)

and Herman (1993). Hookway does not clearly endorse Previous Deliberation, how-

ever—perhaps as a result of having a slightly different framing of the issues from the

one that concerns us. Railton (2009) also emphasizes the role of habit, but Railton is very

clear that he rejects anything like Previous Deliberation.

11. What about actions, like choosing one of two equally appealing bales of hay, that

appear to be taken for no reason? An anonymous referee helpfully suggests that, insteadof holding that every action is taken for a reason, we might hold the weaker thesis that

every action coincides in the actual world with an action performed for a reason. On this

suggestion, choosing the left-hand bale of hay is not something one does for a reason

(because one has no reason to prefer it over the right-hand bale), but it is an action that

coincides with choosing either of the bales, and one does have a reason for this coinciding

action (perhaps, to feed a hungry donkey). If this is the best way to sort out these action-

theoretic issues, we can accept this suggestion: it does not change our arguments. But

notice that when a person acts irrationally, there is no rationalizing explanation of why he

or she does what is irrational rather than what is rational. Yet irrational actions still seem to

be actions, performed for (inadequate) reasons. So perhaps there is a theory of action within which it is possible to say that irrational actions are performed for reasons that do

not justify taking the action rather than some alternatives, and perhaps the same theory

will allow that one bale of hay can be chosen for a reason that nonetheless does not justify

choosing that bale rather than the alternative bale.


219



It might be suggested, in the light of the foregoing, that an act of

deliberation is performed for reasons because it is an action with an

appropriate relationship to a merely possible act of deliberation. Call

this approach “Possible Deliberation.” Perhaps it is true that, were Harold

to deliberate about it, Harold would conclude that he is being reasonable

(or unreasonable) in trying to call to mind promises that might conflict

with meeting his son in Calgary. And perhaps it is this fact about what

conclusion he would have reached that makes it true that Harold’s act of

trying to call these promises to mind is indeed a reasonable (or unrea-

sonable) act.14

Possible Deliberation, like its counterparts, faces a serious regress

problem. Suppose it is true that had Harold deliberated about thereasonableness of calling to mind promises that might conflict with meet-

ing his son in Calgary, he would have reached the (theoretical) con-

clusion that it would be reasonable of him to do so. Now, if Harold had

deliberated, he would have done so for a reason; and so (by Possible

Deliberation), it would have to be true that had he deliberated about

whether he had reason to reach the theoretical conclusion that it would

have been reasonable to call to mind possible conflicts, he would have

reached the theoretical conclusion that, yes, it would have been reason-able to reach the theoretical conclusion that it would be reasonable to call

to mind possible conflicts. And from here it just gets worse, as Harold

must support infinitely many ever-more baroque counterfactuals about

what he would have reasoned about his reasoning about his reasoning

about his reasoning . . . about his reasoning about his calling to mind

possible promises that would conflict with meeting his son in Calgary.

At a certain point, Harold’s ability to support these counterfactuals about

reasoning about reasoning . . . will give out, because his intellectual abil-

ities in this regard are (like all of ours) modestly finite, but the need for

counterfactuals about how he would reason about reasoning about

reasoning about . . . is infinite. That is, there is no problem with there

are undertaken for reasons just in case the act of deliberation judges itself to be reason-

able, in addition to judging whatever else it judges. The content of Katie’s deliberation

(for example) might then be as follows: there is an advertising campaign using the phrase

‘I’m loving’ and this is a reasonable thing to be considering. But Harman’s proposal doesnot solve other problems for Present Deliberation, found in the next section.

14. Possible Deliberation appears to be defended by Tiberius (2002), quoted above.

It would also seem to be defended in McGeer and Pettit 2002, 294, on the assumption that

believing for a (better or worse) reason is what makes one responsible for one’s beliefs.


221



being infinitely many possible Harolds, but there is a serious problem

with it being the case that Harold could have had an opinion about what

it would be reasonable for him to believe about his reasoning about his

reasoning about . . .

(and so on for another fifty iterations). . .

about his

reasoning about calling to mind possible promises, because this is just too

complex a thing for Harold to have a genuine opinion about. This is the

viciousness of the regress. It will not help to appeal to how Harold would

ideally reason, notice, because how he would ideally reason is via delib-

erative acts that would themselves be acts performed for reasons—acts

that would themselves be reasonable—and so no appeal to ideal reason-

ing can be made in explicating what it is to deliberate for a reason on pain

of ontological circularity. A view that combines elements of Present Deliberation and Pos-

sible Deliberation is the view that Harold deliberates as he does for rea-

sons if his deliberation stems from an actual mechanism that ensures that

he does whatever he would have done had he deliberated. But it is hard to

see how this view makes progress over Possible Deliberation. Presumably

the combined view holds that Harold deliberates as he does for reasons if

his deliberation stems from an actual mechanism that ensures he delib-

erates as he would have done had he deliberated on how to deliberate, which again would have had to be for a reason. This sets off the regress

problem again.15 We have no objection to the idea that thinking and

acting for reasons relies on some mechanism—one that could be vali-

dated by an ideal observer who deliberated about the operations of the

mechanism. But there is a related idea that is unsupportable and that we

reject: the idea that what makes the mechanism one suited to ensure that

we think and act for reasons is reducible to facts about counterfactual

reasoning about the mechanism.

Possible Deliberation is no more tenable than Previous or Present

Deliberation. And with it we exhaust our imaginations. There are surely

other ways to defend the claim that a deliberative act is performed for a

reason just in case it bears the right relation to some other deliberative

act. But we have covered the main options that seem worth covering and

15. The combined view is also subject to the problems raised in the next section: if

Katie has foolish views about how she ought to deliberate, views that do not ordinarily

come up when she deliberates about this or that but do come up when she deliberatesabout how to deliberate about this or that, then this combined view must hold that to

deliberate for good reasons in the present, she must now deliberate in the manner that

she would, foolishly, validate upon considering her deliberative process. More on this

below.


222



have found no defense that succeeds. It seems that acts of deliberation are

performed for reasons (good or bad) in virtue of something other than

their relation to deliberation.

4. Other Objections

Apart from regresses, there are various other problems for Previous, Pres-

ent, and Possible Deliberation. Quite salient is the fact that the experi-

ence of deliberating as to how to deliberate is rare and the experience of

deliberating about how to deliberate about how to deliberate almost

nonexistent. Many people seem to get by in life with a bare minimum

of reflection upon their own thought processes. Must we believe that they have, nonetheless, managed at some time or other to have deliberated

about how to deliberate in sufficiently broad terms that now their acts of

deliberation are all capable of being taken for good (or bad) reasons?

This seems required by Previous Deliberation but unlikely to be true. And

then, Present Deliberation requires an act of deliberation that is not

phenomenologically familiar yet that should be as easy to experience as

ordinary deliberation itself. At the very moment at which Harold con-

sciously tries to call to mind competing promises that would prevent him

from promising to meet his son in Calgary, he is not (if he is anything like

us) consciously evaluating the reasonableness of this conscious search of

his memory. Yet Present Deliberation requires that there be such an act.

So the second deliberative act required by Previous Deliberation would

seem not to be a conscious act at all—which makes it an act not of

deliberation but of some other kind.

More problems appear when we turn the discussion from acting

for reasons per se to acting for good reasons, or acting rationally or reason-

ably. Consider the view that acting for good reasons depends on a previous, present, or possible act of deliberation. Problems for these views

arise from the possibility of people deliberating about how to deliberate

but reaching foolish conclusions.

For Past Deliberation, consider people who, perhaps as a result of

adolescent philosophical discussions, have embraced foolish principles

constraining deliberation: to believe nothing in the future without seeing

a proof ; to do nothing without first asking how it will benefit me . People

who manage to constrain their deliberations in these ways seem lessreasonable, not more, in virtue of their conformity to their chosen prin-

ciples. If they neglect their principles (as they often do, common sense

reasserting itself unnoticed), they deliberate in ways that are more


223



reasonable, it would seem, while in violation of their most recent settled

policies for deliberation. (One of us has discussed such cases at length in

Arpaly 2000 and Arpaly 2003, chap. 2.)

Present Deliberation entails, similarly, that an utterly unreason-

able deliberation about how to deliberate has the power to make the

deliberated-upon deliberation itself reasonable. Thus the view makes it

quite unclear how an act of deliberation could be unreasonable so long as

an agent is, at the moment of deliberation, subjectively content with the

act. Yet, as we have seen, such unreasonable acts of deliberation seem

common enough: if Katie is deliberating about the grammaticality of ‘I’m

loving’ and her deliberation has gone down an unreasonable path in

which she is asking herself about all the alternatives that exist to that phrase, her subjective contentment with the course of her deliberation

does not make it reasonable. She is failing to act on the reasons she has to

think about other things if her goal is to determine the grammaticality of

the phrase ‘I’m loving’, and no contemporaneous endorsement of her

thinking can change this fact.

A variant on Present Deliberation holds that an act of deliberation

is one performed for good reasons just in case it is an act that is believed

reasonable, consciously believed reasonable, thought to be justified by the reasons before the agent, or otherwise taken (cognitively) to be ap-

propriately licensed. Call this variant “Present Recognition.”16 Strictly

speaking, Present Recognition might fall outside our purview, because

the recognition it requires is not an action (unless it is a judgment, and a

judgment is an action). But Present Recognition is worth addressing; it

also falls prey to the sort of objections being raised in this section to Past,

Present, and Possible Deliberation.

The obvious objection to Present Recognition is the same that

was just made to Present Deliberation: that foolish approving attitudes

would seem unable to make otherwise foolish deliberative acts into ones

performed for good reasons.17 As before, it would seem that if Katie is

getting caught up with thinking of alternatives to the phrase ‘I’m loving’

16. Present Recognition appears to be endorsed by Barry (2007) and Korsgaard

(1997), quotedabove. Niko Kolodny (2005, 520) seems to endorse something like Present

Recognition in writing that for a mental transition to happen for a reason, it must result

from awareness of the justification for the transition. But in a footnote, Kolodny wants toallow that unconscious, automatic awareness is possible. Since there is no such thing (in

our understanding) as unconscious awareness, we find this a bit puzzling but suspect that

the problems do not go away by burying the putative awareness in the unconscious.

17. Wedgwood (2006, 675) raises similar objections to a similar move.


224



while deliberating about its grammaticality, then Katie is deliberating

irrationally—even if she consciously believes that she is proceeding

very reasonably, or otherwise has a positive cognitive attitude toward

her responsiveness to reasons. Her conscious belief is simply mistaken.

A related objection asks: is an agent’s conscious belief that he or

she is proceeding reasonably (or whatever other variant of Present Rec-

ognition is preferred) itself believed for good reasons? If it is not, then it

seems not to be the sort of thing that could render an act one that is

performed for good reasons. Why would mere faith in one’s reasonable-

ness, held for no good reason, be something that could actually render

one’s acts reasonable? (If the action happens to be in conformity to the

requirements of rationality, then one is merely lucky that one’s act is asrationally acceptable as one guessed it to be. How does this lucky coinci-

dence justify the claim that the act was not merely in accordance with

rationality but in fact performed for good reasons?) But if the belief in

one’s own reasonableness is believed for good reasons, then we can look

back to when the belief was formed and ask how it was that the belief was

formed for good reasons. It cannot be because the belief was the product

of appropriate deliberation about what to believe, on pain of circularity

orregress.Soitmustbesomethingelse.Andthuswefindthatifwestartby holding Present Recognition, we end by recognizing a process outside of

deliberation and voluntary action such that beliefs can be formed for

good reasons via that process. So long as this process is equally capable

of generating actions for good reasons (and at this point there is no

reason to deny it), the conclusion is exactly that toward which we are

pressing. As a result, Present Recognition fares no better than Present

Deliberation.

Finally, Possible Deliberation faces these same problems stem-

ming from the possibility of foolish views about deliberation. Katie is,

suppose, acting for good reasons when she says to herself, “Lots of people

seem to say ‘I’m loving’ without scare-quoting it or anything.” She is trying

to determine the grammaticality of the phrase, and saying this to herself is

rehearsing relevant information, and she has said it to herself for the

reason that it is relevant information. But suppose Katie has some unrea-

sonable beliefs about good reasoning. In particular, suppose Katie

believes that no piece of evidence is worth considering if it is not conclu-

sive. This belief of hers is unreasonable, of course, and normally she doesnot revise her beliefs in accordance with it (though occasionally she does,

when this belief spontaneously occurs to her while she is reasoning). But

in contexts in which she theorizes about how her reasoning should be


225



conducted, this belief comes to the fore all the time. It seems to us that

this scenario is possible. And yet, Possible Deliberation must say it is

impossible, because the fact that Katie would reject her own deliberative

act as unreasonable makes it true, according to Possible Deliberation,

that her own deliberative act actually was unreasonable. The mere pos-

session of a foolish belief about what counts as evidence worthy of con-

sideration in deliberation deprives Katie of the capacity to deliberate for

good reasons. We think this is false. It goes against what seems the most

natural interpretation of the case, which is that Katie can be reasonable

in actuality while being counterfactually disposed to reject her own

reasonableness, and also it seems that believing the contrary requires

attributing too much power to single metatheoretical beliefs. We donot think that believing (in the grips of bad psychology) that every act

is selfish can make every act by that agent selfish. Why should beliefs

about the nature of deliberation be so much more powerful? So we

take it that Katie can deliberate for reasons even though her deliberation

does not conform to her belief about how such deliberation should

proceed.

Tiberius (2002) defends her form of Possible Deliberation against

this objection. Her defense focuses on making it plausible that agents who force themselves to conform to foolish practical principles act ratio-

nally in so doing. Agents who believe that they should consult an astrol-

oger as part of ideal deliberation and who, as a result, do not get proper

medical treatment for something act rationally in not getting the medical

treatment (if doing so follows from the astrological advice), according to

Tiberius. But Tiberius does not consider the question of whether these

agents would be acting irrationally if they had never deliberated (and so

never concluded that they ought to start by consulting an astrologer) but

had simply acted on their belief that doctors know what to do for people

with mysterious pains, plus their belief that they have a mysterious pain,

plus (if needed) their desire not to be in pain. It seems to us that these

agents would be acting for a good reason if they were to just spon-

taneously visit a walk-in clinic as they were walking past one. And if so,

there is something wrong with Possible Deliberation. The fact—if it is

one—that agents could undermine their justification for taking this ac-

tion via reasoning on the basis of a foolish deliberative principle does not

change what the agents can do for good reasons in the absence of suchmisguided deliberation.


226



5. Is There an Ambiguity?

We have shown that there must be a way of thinking and acting for reasons

independent of deliberation and that this way of thinking and actingfor reasons must be what makes acts of deliberation more or less reason-

able as actions. But it is worth asking: is there another way—a different

way—of thinking and acting for reasons, one that does rely upon delib-

eration? That is, are there perhaps two senses of ‘thinking for reasons’ and

‘acting for reasons’, a weaker sense in which deliberation is not necessary

and a stronger sense in which deliberation is necessary after all?18

There is rarely anything wrong with making distinctions in philos-

ophy, and we see nothing to object to in the present proposal. Let there beone sense of what it is to think or act for a reason, the sense under dis-

cussion in this work so far, and let there be another, called ‘thinking and

acting for reasons par excellence’, that is instantiated only when the

thought or act in question has been considered, and validated, by the

thinker or actor’s own deliberative processes.

With this distinction made, however, it is worth asking what the

status is of beliefs formed and actions performed for reasons par excel-

lence. Are these actions the only ones that are properly said to be actions

as opposed to mere activities? (Are these beliefs the only ones that are

properly said to be beliefs as opposed to mere cognitions?) Are these

thoughts and actions the only ones that are properly said to be rational,

reasonable, sensible, intelligent (or irrational, not very reasonable, fool-

ish, unintelligent)? Are these thoughts and actions ones that are certain

to be more rational than they would have been otherwise? Our answer to

all of these questions is ‘no’, for reasons found in what has come before.

And if the answer to all of these questions is ‘no’, then while we are happy

to grant the existence of thinking and acting for reasons par excellence, we see no special role for this sense of ‘acting for reasons’ in the philo-

sophical literatures with which we are most familiar. It is the sense of

thinking and acting for reasons that is not “par excellence”— plain think-

ing and acting for reasons—that has all of the properties we are most

interested in when we are interested in thinking and acting for reasons.

Begin with the question of whether “thinking” and “acting” for

reasons in a sense wholly independent of deliberation (not “par excel-

lence”) really counts as thinking or acting. There are philosophers whodistinguish the “mere activity” of some animals from the full-blooded

18. We thank an anonymous referee for urging us to consider this possibility.


227



actions that only people can perform.19 Perhaps the philosopher inter-

ested in thinking and acting for reasons par excellence would want to

hold, in a similar vein, that being moved by reasons par excellence is a

necessary condition for being a belief or action. So consider the case of

the apparent act of deliberating. Deliberation itself is very rarely some-

thing on which one has deliberated, but deliberation is a paradigmatic

mental action, an action one can perform— like any action—voluntarily.

(The reader can experience the voluntary aspect of deliberation by now

electing to take a minute to deliberate on, say, the reasons for the dwin-

dling memberships of unions in the United States, should the reader wish

to do so.) Deliberation is also subject to the common disorders of volun-

tary action, which is further evidence that it is an ordinary voluntary action in general (regardless of whether one has just deliberated about

how to deliberate). For instance, one can decide that it is best to cease to

deliberate about how best to flirt with someone unsuitable only to then be

akratic and unable to stop thinking about it.

Moving beyond deliberation, there are all of the apparent actions

people perform and the apparent beliefs they form wholly independently

of deliberation. If we have any grasp of action, then passing the salt when

asked to “please pass the salt” is an action, and is so whether or not onedeliberates (as rarely happens, in the case of salt passing) about whether

to pass the salt. Likewise, if we have any grasp of belief, then what

typically happens when one opens the fridge and sees everything that

one expects but butter is that one comes to believe that there is no more

butter, and this counts as a belief even in the absence of deliberation on

butter’s availability. To deny these theses in the cases of ordinary adult

language-using human beings (at least) is, it seems to us, to lose any grasp

one might have had on what actions and beliefs are.20

19. See, for example, many of the essays in Velleman 2000. We note that although

Velleman’s views have a strongly Kantian cast, he need not fall prey to any of the arguments

in this article. The desire to make sense to oneself, from which all genuine action must

stem on Velleman’s account, is one that might move one for good reasons in the absence

of deliberation upon those reasons.

20. Someone wishing to nonetheless diminish the status of these putative actions and

beliefs by denying them these names could always reserve the words ‘action’ and ‘belief’for things upon which there has been deliberation—but then the philosophical interest

of true beliefs, as opposed to schmeliefs (things just like beliefs except for not having been

deliberated upon), and true actions, as opposed to shmactions, becomes open to

question.


228



So thinking and acting for reasons par excellence is not required

for performing actions or forming beliefs. It is likewise not required for

these actions or beliefs to be more or less rational, performed for better or

worse reasons. As we held earlier in the article, nothing can be an action

without being performed for some reason or other, whether a better or a

worse one. So if we have shown that there can exist actions that are not

performed for reasons par excellence, then we have shown that there are

actions that are performed for better or worse reasons without being

performed for reasons par excellence. And the same reasoning, it

seems, should hold for belief as well as action.

Now, even if this much is agreed to, it might still be held that

thinking and acting that does not happen for reasons par excellencemust always be somewhat deficient in its rationality, intelligence, or

reasonableness. But this too seems wrong. Imagine, for example, Oscar

Wilde, who according to one story bragged that he could make a pun

on any subject. Presented with the challenge “the queen,” he quickly

answered, “the queen is no subject.” A conversation like this can happen

much too fast for any deliberation to take place and surprise even the

pun-maker; this is a big part of what makes a great conversationalist. Yet

Wilde’s answering is a paradigmatic action, and a paradigmatically witty,smart, well-chosen, intelligent action as well. It would have been less witty

and a lesser display of intelligence had Wilde deliberated before his quip;

this strongly suggests that Wilde’s quip strikes us as performed for excel-

lent practical and aesthetic reasons, in a way that warrants this kind of

praise, without any presumption that deliberation was involved, and

so without any presumption that it was performed for reasons par

excellence.21 Similar thoughts apply in the epistemic domain. It is a

true dullard who concludes that he or she is out of butter only after

deliberating upon the fact that no butter is in evidence in the refrigerator

and after concluding that the absence of visual evidence is a reason to

believe in the absence of butter. A more intelligent person would be

shown to be more intelligent—would have his or her rational faculties

on full display —in drawing the same conclusion without engaging in any

deliberation at all. There are cases in which human beings cannot draw

the correct epistemic conclusions for good reasons without relying on

deliberation (for example, about complex philosophical matters; see

below)—but that such cases exist does not show that every case is one

21. Arpaly (2003, chap. 2) gives witty conversation as an example of possibly rational

action without deliberation.


229



in which the only way to believe something for a good reason is to have

first thought it over.

Finally, there is the evidence (drawn on in the previous section)

that theoretical and practical conclusions reached on the basis of delib-

eration can be less reasonable, less rational, than the theoretical and

practical conclusions that would have been reached in the absence of

deliberation. People with a foolish theory of how to deliberate can make

themselves less reasonable than they would otherwise be by deliberating.

People whose deliberations about whether or not there is any butter do

not rest until a proof is at hand are people whose deliberations are likely

to harm their epistemic status, not help it; had they only not delib-

erated (and so not drawn upon their thesis that all beliefs need proofsfor their justification), they would likely have believed what was reason-

able on the grounds that made it reasonable. Similarly for their practical

counterparts.

In short, while there is room for the idea that there is such a thing

as responding to reasons par excellence, there is no particularly interest-

ing property that attaches to all and only instances of responding to

reasons in this way, in the light of deliberation about one’s reasons.

There is an important role for deliberation to play in our thinking andacting for reasons, as we will now turn to demonstrating. But this role is

one that is intermittent, contingent, and modest.

6. The Role of Deliberation

We concluded that acts of deliberation are in no way required for other

acts of deliberation to be performed for reasons. We take it that, if acts of

deliberation are not required, no other acts are required: acts of delib-

eration were the only promising candidates from the realm of actions. So we conclude that it is possible to act for reasons in virtue of things—

processes—that are themselves nondeliberative and nonvoluntary. And

while we have only argued that this is true in the realm of deliberation, we

think it natural to generalize: it is possible to believe for reasons and to act

for reasons in all sorts of domains via nondeliberative, nonvoluntary

processes.

We now face an unusual question. What is the value of delibera-

tion? Within the theory of rationality, the value of deliberation is not normally questioned, but if people are endowed with the capacity to

think and act for reasons without deliberation, then the question of delib-

eration’s value naturally arises.


230



Thinking and acting for reasons via nondeliberative, nonvolun-

tary (‘ND’) processes has an impressive role in our lives. But it is also

evident that there are obstacles to thinking and acting for the very best

reasons using only our ND capacities. Other causal processes are at work

within our minds in addition to those that mediate our ND abilities to

think and act for reasons. Quick wit and even ordinary grammaticality can

suffer from low blood sugar, from background emotional distress, from

sitting in view of a distracting television display, and more. There is, as a

result, room for various interventions to enhance our ND abilities to

think and act for reasons.

Caffeine and a good night’s sleep and a bite to eat all enhance our

ND abilities to think and act for reasons, but these are purely nonrationalinterventions. What room is there for deliberation to contribute to think-

ing and acting for reasons?

Consider first some more carefully selected ways in which we can

enhance our ND abilities to think and act for good reasons. These abil-

ities can be diminished by distraction: a television in the background,

noisy children running around, the presence of a very attractive per-

son— all can make it hard to maintain witty banter, generate philosoph-

ical insight, or do other things displaying ND reason-responding at itsbest. (They can even make it difficult to act on the reasons one has to use a

verb conjugated for the third person singular or to move the fork in one’s

hand to one’s mouth.) One thing people do to enhance their ND abilities

to think and act for reasons is to dismiss the distractions: turn off the

television, send the children to play outside, and gaze at the ceiling in

preference to the attractive person.

Another familiar obstacle to our ND abilities to think and act for

good reasons is lack of inspiration. Wit without a set of topics to organize

it is challenging, and new philosophical insights are often promoted by

being asked new questions. To enhance their ND reason-responding,

people will take advantage of available sources of inspiration: glance

about for topics to comment on or listen to the very particular cadences

of a conversational partner’s speech, if one hopes to find something witty

to say, or encourage a conversational partner to “never mind how other

people think about the problem, just tell me whatever you think,” when

seeking to jump-start philosophical insight.

Another barrier to the optimal use of our ND abilities to think andact for reasons is generated when beliefs (and perhaps also plans, desires,

and so forth) are not at the foreground of our awareness. If Harold’s son

asks him to meet in Calgary on Tuesday, Harold’s thinking will tend to be


231



most responsive to the contents of his mind that were just under discus-

sion and to those that are most linked (by association?) to the ideas of

meeting, Calgary, and Tuesday. Harold’s thinking will also tend to reflect

ideas that he has dwelt on recently, especially ones he has dwelt on re-

peatedly. But Harold is in some danger of not taking into account facts he

knows but that he has not much thought about recently in any guise. If,

for instance, Harold had agreed five weeks ago to lunch with someone on

the twentieth, and Tuesday falls on the twentieth, and going to Calgary

requires departing well before lunch, this is much less likely to come to his

mind than a regular Tuesday meeting of the city planning council. Things

one believes (plans, desires, and so forth) but that are not being thought

about at the moment seem less efficacious in influencing one’s thoughts,even when relevant, than things equally believed (planned, desired, and

so forth) but currently or recently thought about, all else being equal.

Automatic retrieval is clearly possible (we spontaneously recall relevant

facts all the time), but it is also clearly not implemented in human beings

with optimal efficiency. Fortunately, there are a number of overt acts one

can take to combat the problem: Harold can keep a day planner and

consult it, for instance.

These sorts of voluntary overt actions have corresponding part-ners in deliberation. Consider first dealing with distraction. In delibera-

tion, one can take mental actions to deal with distraction: one can refocus

one’s attention onto one’s question or the evidence already called to

mind, one can use verbal imagery to say to oneself “I need to say some-

thing about causation, something about causation” and otherwise use

mental behavior to block the effects of potentially distracting stimuli.

When these actions involve contents that are relevant to theoretical or

practical conclusions one is pursuing, these actions are acts of delibera-

tion. In this way, deliberation can be a covert action that improves ND

reason-responding, taking the place of an overt action that would im-

prove ND reason-responding. Of course, sometimes acts of deliberation

are a poor substitute for more vigorous action: if the kids are making a

ruckus, saying “Is causation always lawful?” repeatedly to oneself is prob-

ably a less effective way of dealing with distraction than sending them

out to the yard. But sometimes acts of deliberation are more effective

than most overt acts available: if emotional distress is getting in the way of

philosophical contemplation, it might be that mentally refocusing one’sattention is going to be more effective in reaching a theoretical con-

clusion than giving oneself a slap to the face and muttering “Get a grip!”


232



Note that the deliberative acts just described would themselves be

actions performed for reasons as the result of ND mechanisms. When

deliberation is brought to bear to enhance the efficacy of ND mechan-

isms for thinking and acting for reasons, it does not do so as an alien

intrusion into an ND process. Rather, deliberation is brought to bear as its

own sort of action, performed through ND mechanisms that make the

deliberative acts themselves more or less reasonable acts: reasonable at

promoting the end of enhancing our ability to think about other things

and act in other ways, for good reasons.

Consider now dealing with lack of inspiration. There are acts of

deliberation that can cope with a dearth of inspiration in the external

world and, in this way, enhance ND reason-responding, just as overt ac-tion might. If Heather is planning to make a funny speech at a wedding

and finds herself lacking inspiration, she could take overt actions that

might help, such as browsing through a photo album with pictures of the

bride- and groom-to-be. But she might also engage in deliberative acts,

such as calling to mind things that comedians often talk about, in the

hope of stumbling across a topic of humor that particularly suits the

people involved or the ceremonial occasion. And it might be that on a

given occasion Heather would be better off turning to the world outsideher head for inspiration, while on another occasion her best bet would be

to turn inward and rely on deliberative acts to gain insights into how to

improve her speech.

Next, consider dealing with the problems generated by infor-

mation that has not recently come to conscious attention. Acts of delib-

eration by their nature bring ideas and images into consciousness and so

are excellent vehicles for the promotion of neglected facts—so long as

something (some ND process that guides deliberative acts) causes the act

of deliberation to focus on the buried information in the first place. To

make it more likely that we will recall such buried information, we use a

number of techniques in deliberation: going over a conclusion or an

argument again, “looking” (as we say) for potential problems. (If we

are skilled deliberators, we tend to use techniques such as putting the

same ideas into different “words,” or approaching them from a different

organizational principle or perspective, in order to enhance the prob-

ability that we will call to mind buried information.) Sometimes we need

to check an encyclopedia and so forth for such purposes, but often delib-eration takes us far enough.

One other circumstance in which deliberation helps (the last we

will mention, though we make no claim to being exhaustive) is when we


233



are faced with puzzles that are, for some reason or another, too “big” for

us. Deliberation can be used to sequence together a set of steps, each one

of which can be performed without deliberation, in such a way that

the sequence results in solving the “big” problem with which we began.

Consider complex sums: faced with the task of adding 3,545 to 869 we

find ourselves unable to see the result directly. But if we focus our atten-

tion first on the sum of five and nine, and keep in mind that the result

ends in four (a conclusion that, itself, can be reached without the aid of

deliberation), and then sum one, four, and six . . . and so on, then we

can perform a sequence of mental actions with the final result being

that we can draw the correct conclusion about the sum of 3,545 and

869. Whether or not a particular task calls for this treatment dependson the individual’s capacity for ND reason-responding: young children

require deliberation even for simple arithmetic, most adults require it

only for more complex calculations, and a few arithmetical wizards im-

mediately see—for the right reason—sums and products that the rest of

us must work through with painstaking effort. It also depends on the

circumstances: deliberation can suddenly be required for a task not nor-

mally requiring it if one is overwhelmed by fear or another strong

emotion, for instance. And presumably it depends on the nature of thetask itself: arithmetic lends itself to algorithms requiring only small steps

(relative to our ND reason-responding capacities), and so does shopping

for groceries, but some complex tasks seem not to helpfully decompose

(perhaps generating major new philosophical insights falls into this cat-

egory). Yet, even with these tasks, deliberation often seems helpful: it is

hard to imagine how a human being could develop a good philosophical

theory, devise a reasonable scientific experiment, or write a structurally

complex short story without it.

The moral that is lurking in these examples is that what delibera-

tion does, when it enhances responding to reasons, is to act to remove

barriers to the ND processes through which we normally think and act for

reasons. Perhaps it is true that, as children, we first learn the value of

talking to our parents, then learn the value of talking to ourselves aloud,

and finally learn the value of voluntarily generating auditory images of

our voices, as a means of thinking more reasonable thoughts about what

to believe and do. But even if this is not true of ontogeny, it is a suitable

myth for the sort of view we are defending. Deliberation is not the foun-dation of our ability to think and act for reasons but a tactic we have for

enhancing our preexisting, and foundational, abilities to think and act

for reasons: our ND reason-responding.


234



Because this is the role of deliberation, the value of deliberation is

intermittent, contingent, and—in some ways—modest.

It is intermittent because sometimes we think and act for reasons

without recourse to deliberation, and at other times we are better able to

enhance our abilities to think and act for good reasons through turning

off the noise in the background than through deliberation.

The value of deliberation is contingent because, so far as we can

see, there is no reason in principle that we have to be designed such that

blood sugar, ambient noise, lack of recent conscious attention, or even

computational tractability are factors that are able to disrupt our ability to

think and act for good reasons. There could exist creatures whose ND

abilities to think and act for good reasons are flawless, or close to it. Forsuch superhuman creatures, all deliberation would be a waste of time, in

the way it would be a waste of time on our parts to deliberate on trivial,

obvious topics (what is five plus nine? should I brush my teeth at exactly

thirteen minutes to the hour, every hour?). There are animals incapable

of seeing at a glance that a reaching device must be used in order to

secure a certain banana, but we are capable of this feat. There are

human beings among us who are capable of seeing at a glance that 39

goes into 351 nine times, while others of us are not capable of this feat.(There are also reports of savants capable of seeing at a glance that certain

seven-digit numbers are prime. Though the power is mysterious, we see

no reason to deny that these savants grasp what they do for excellent

reasons, unaided by deliberation.) Why then should there not be possible

creatures who can see at a glance what to do to bring the world to a just

global peace, or what to think about the correspondence theory of truth?

It seems to us that these creatures are possible, though they would be very

unlike us indeed.

There could also be creatures who have all of our difficulties with

ND reason-responding but who also have trouble controlling their behav-

ior if they start thinking about a question. These creatures would be ones

who, once started on the question of whether or not to have a precise time

for brushing their teeth, would have trouble doing anything other than

brushing their teeth for the rest of the afternoon. They would be like

obsessive-compulsives who would be prone to developing topics of obses-

sion and compulsion from dwelling on them in thought. These creatures

are possible, if less well adapted to a world like ours than we ourselves are.It is only a contingent fact that the Earth is populated by neither

sort of creature: neither creatures who are rationally perfect without de-

liberation nor creatures who are grossly impaired by it. We have great


235



need of our powers of deliberation and no doubt owe to them much of

the technological and social complexity that we have and that hunter-

gatherers lack (and they, in turn, no doubt benefit greatly from delib-

eration as well). But things need not have been this way; what deliberation

does for us is not expressive of its nature so much as it is expressive of

our nature: for others deliberation would be otiose or systematically

harmful, but given the kinds of creatures we are, deliberation is by and

large a boon.

Finally, the role of deliberation is modest in that deliberation can

fail to respond correctly to reasons just as much as other actions can.

Deliberation can fail to call to mind the neglected but vital fact, can

look in worthless places for inspiration, can fail to shut out the distrac-tions that are making it so hard to draw any conclusions at all, can take us

to the wrong step in an attempt to solve a problem sequentially. It can

focus our attention on certain facts that support one conclusion while

diverting our attention from other facts that support another (better

supported) conclusion. Deliberation is a valuable but imperfect tool

for improving our more valuable but even more imperfect abilities to

think and act for reasons without it.

7. Thinking and Acting for Reasons without Deliberation

All of this could lead one to the conclusion that a new theory of thinking

and acting for reasons is needed to explain nondeliberative, nonvolun-

tary processes of reason-responding. We do not propose to offer a theory

here because such a theory has been proposed, elaborated, and defended

at length by one of us in Arpaly 2006, and is defended in Wedgwood

2006. But to make things clearer, we will say a few words about what we

take a suitable theory of ND reason-responding to look like. A suitable theory of ND reason-responding considers transitions

between mental states: from some beliefs to others (when believing for

reasons), from beliefs (and perhaps desires and plans) to an intention or

willed action (when acting for reasons), and perhaps other transitions as

well. Now, some such transitions happen as a matter of blind material

causation: Katie’s belief that grammar depends on social practices might

physically rest on a blood vessel that, because of the activity of the neurons

making up the belief, ruptures, causing an end to the neurons making upher belief that ‘Object’ is an amusing name for a cat. This sort of tran-

sition is obviously not one that happens for reasons. Other transitions

happen because of causal relations that are sensitive to the contents of the


236



mental states involved but not because these contents have any rational

relation to one another. Harold might think of the planning council, and

this might—because he associates the idea of planning with Michael

Bratman—cause him to think that Michael Bratman, author of Intentions,

Plans, and Practical Reasoning , is very tall. Though physical connections

certainly mediate this causal relation, the causal relation is in some way

sensitive to content as well: it is no accident that planning reminds Harold

of Bratman and not, say, of mice. Just how to understand this sort of

causation-in-virtue-of-content (or, perhaps better, this explanatory rel-

evance of content) is debated by philosophers of mind, but it seems

reasonable enough to us that some true story exists to be told. Still,

even once told, this is not yet a story of mental transitions in virtue of reasons. Free associative links are not reasons (or do not represent

reasons) to shift from one thought to another. Finally, some transitions

seem to happen because the contents in question have certain logical

relations to one another. The transition from the beliefs that all owls are

strigine and the belief that Bubo is an owl to the belief that Bubo is

strigine is one that could be caused not just by free association but by a

process that is sensitive to the fact that the first content entails the

second.22

Just how a brain might be organized to make this true weleave to scientists. And just what theory of causation, or causal relevance,

might permit this sort of claimed causal relation to be validated? It seems

to us that it is a special case of the general problem of how to understand

causal relations in virtue of properties (or how to understand causal rel-

evance in terms of properties), and so the problem is one to be left to

philosophers working on causation and causal relevance. It is not a prob-

lem specific to action theory or the theory of reasons. The claim that a

belief that Bubo is an owl gave birth to the belief that Bubo is a strigine by

virtue of the logical connection between it and the belief that all owls are

strigines is no more mysterious than the claim that the rock broke the

window by virtue of its mass and rigidity and the fragility of the glass. We

assume that the metaphysical problem has a solution.

22. It is perhaps worth reminding thereader that we do notrequire thoughts brought

to mind to actually have a logical relation of some sort to anything of note in order to be

part of one’s deliberation, much less a logical relation as strong as entailment. All that is

required is that the thought brought to mind be one that is meant by the agent to bearsome logical relation to what to think or what to do— which might amount to simply being

a consideration that is hoped to be one that inductively undermines one of several con-

siderations counting in favor of taking a certain action or thinking a given thought, for

example.


237



We take reason-responding to require nothing more than mental

transitions that occur because of the logical relations (theoretical entail-

ment, practical entailment, statistical relevance, or the like) between the

different attitudes involved. And we take this to be nothing fancier than

causal relations in virtue of certain properties, something that neither

deliberation nor voluntary action more generally is required to mediate

in general.23

We have argued that deliberation is not essential to rationality.

This is not to deny that deliberation is wonderful. We can think of no

realistic way in which human beings could have achieved the insights in

philosophy, the arts, and the sciences that we have achieved without deli-

beration. So we celebrate deliberation. But we celebrate it as a powerfultool well suited to ameliorating our human weaknesses as imperfect ND

reason-responding agents. We are unfortunate in needing deliberation

but fortunate in having the tool that we need. Deliberation is a beautiful

tool for its purpose. Philosophers go wrong only when they misconstrue

deliberation as, not a tool in, but the foundation of, our power to think

and act for reasons.

References Anscombe, Elizabeth. 2000. Intention . Cambridge, MA: Harvard University

Press.

Arpaly, Nomy. 2000. “On Being Rational Against One’s Best Judgment.” Ethics

110: 488– 513.

———. 2003. Unprincipled Virtue: An Inquiry into Moral Agency . New York: Oxford

University Press.

———. 2006. Merit, Meaning, and Human Bondage . Princeton: Princeton Uni-

versity Press.

Barry, Melissa. 2007. “Realism, Rational Action, and the Humean Theory of

Motivation.” Ethical Theory and Moral Practice 10: 231–42.

Chan, David. 1995. “Non-Intentional Actions.” American Philosophical Quarterly

32: 139–51.

Clark, Andy, and David Chalmers. 1998. “The Extended Mind.” Analysis 58:

10–23.

Davidson, Donald. 1980. Actions, Reasons, and Causes . Oxford: Oxford University

Press.

23. Though sketchy, our approach has the virtue of being precise about what sort of

ND mental state transitions mediate thinking or acting for reasons and distinguishing

these from the sorts of ND mental state transitions that do not. In this respect, our work is

an advance over Railton 2009.


238



Dreier, James. 2001. “Humean Doubts about Categorical Imperatives.” In Varieties

of Practical Reasoning , ed. Elijah Millgram, 27– 49. Cambridge, MA:MIT Press.

Hacker, Peter. 2007. Human Nature: The Categorical Framework . Oxford: Blackwell.

Herman, Barbara. 1993. The Practice of Moral Judgment . Cambridge, MA: HarvardUniversity Press.

Hookway, Christopher. 1999. “Epistemic Norms and Theoretical Deliberation.”

Ratio 12: 380–97.

Kolodny, Nico. 2005. “Why Be Rational?” Mind 114: 509–63.

Korsgaard, Christine. 1996. Sources of Normativity . New York: Cambridge Univer-

sity Press.

———. 1997. “The Normativity of Instrumental Reason.” In Ethics and Practical

Reason , ed. Garrett Cullity and Berys Gaut, 215–54. Oxford: Oxford Univer-

sity Press.———. 2009. “The Activity of Reason.” Proceedings and Addresses of the American

Philosophical Association 83: 27– 47.

Lazar, Ariela. 1999. “Deceiving Oneself or Self-Deceived? On the Formation of

Beliefs under the Influence.” Mind 108: 265–90.

McGeer, Victoria, and Philip Pettit. 2002. “The Self-Regulating Mind.” Language

and Communication 22: 281–99.

Peacocke, Christopher. 1983. Sense and Content: Experience, Thought, and Their

Relations . Oxford: Oxford University Press.

Railton, Peter. 2004. “How to Engage Reason: The Problem of Regress.” InReason and Value: Themes from the Moral Philosophy of Joseph Raz ,ed.JayWallace,

Philip Pettit, Samuel Scheffler, and Michael Smith, 176– 201. Oxford:

Clarendon.

———. 2009. “Practical Competence and Fluent Agency.” In Reasons for Action ,

ed. David Sobel and Steven Wall, 81–115. New York: Cambridge University

Press.

Raz, Joseph. 1999. Engaging Reason: On the Theory of Value and Action . Oxford:

Oxford University Press.

Seidman, Jeffrey. 2008. “Caring and the Boundary-Driven Structure of PracticalDeliberation.” Journal of Ethics and Social Philosophy 3: 1–36.

Tiberius, Valerie. 2002. “Practical Reason and the Stability Standard.” Ethical

Theory and Moral Practice 5: 339 –53.

Tye, Michael. 1995. Ten Problems of Consciousness . Cambridge, MA: MIT Press.

Velleman, David. 2000. The Possibility of Practical Reason . New York: Oxford

University Press.

Wedgwood, Ralph. 2006. “The Normative Force of Reasoning.” Nou ˆ s 40:

660–86.


239

Deliberation and Acting for Reason

Documents

Transcript of Deliberation and Acting for Reason