On Practical Empiricism

On Practical Empiricism Joseph Ulatowski Abstract Elijah Millgram has argued that experience, or backward-directed commitments, matters most for figuring out what to do and is a useful guide for action. His argument has assumed that experiential beliefs are rock-bottom, where we need go no further than rock-bottom beliefs to start along the practical inferential chain. In this paper, I want to challenge Millgram’s adoption of experiential beliefs as rock-bottom. What I will argue here is that experiential beliefs depend upon our commitment to them, and our commitment to them is not experiential. If our commitment is not something that bottoms out in experience, then it is an antecedent condition, and not the experiential belief itself, that is rock-bottom. So, experiential beliefs are not rockbottom; something antecedent to experiential beliefs are. Keywords: practical reasoning, theory of rationality, practical induction, Elijah Millgram ! Practical reasoning helps us to figure out what to do, whereas theoretical reasoning helps us to figure out what to believe.1 Elijah Millgram has claimed that practical reasoning and theoretical reasoning are analogous, so, according to his “practical induction,” experience, such as goals, priorities, evaluations, and other pieces of the agent’s cognitive equipment, is a useful guide to action. Agents pursue goals because it will permit them to reach some end where reaching that end is something they want to reach. The aim of this paper is to challenge Millgram’s view that experience matters most for figuring out what to do.2 According to Millgram, some experiential beliefs are non-inferentially acquired. He calls these beliefs “rock-bottom.” By this, he does not understand experiential beliefs to be undoubtedly true or justified epistemological building blocks upheld beyond doubt; instead, according to his view, Millgram has suggested that we can learn from experience about what is worthwhile, what matters, or what is desirable. My assault on this position begins and ends with questioning how it is that we come to know or learn from experience. I will argue that having rock-bottom beliefs depends upon commitment, or what Millgram has called ‘conviction’. If commitment is an antecedent practical requirement necessary for figuring out what we believe 1

It is interesting to note that many of Millgram’s essays (1996, 1997a, 1997b, 1997c, 2001, 2005) begin with a declaration of the analogy between theoretical and practical reasoning such as the one above. It seems he has inherited the analogy from Geach (1976, 96) and it seems to be symbolic of his argument favoring a kind of externalism in practical reason (cf. Wallace 2014). 2

Elijah Millgram’s practical induction is the focus of this paper, but his position can be categorized with other views under the heading of “practical empiricism” (Cf. Brandom [1994, 2001] and Stampe [1987] ). Despite the similarities between Millgram’s own position and these other positions, I do not in this paper intend to offer a critique of any of them.

through experience and if commitment is non-experiential, then antecedent non-experiential commitment is rock-bottom. If antecedent commitment is not experiential, there is something other than experience that matters most for figuring out what to do. 1. Millgram’s Practical Induction and “Rock-Bottom” Experiential Beliefs Millgram’s position is an endorsement of reason externalism. If it is implausible that reasons for belief entail or depend on facts about desire or motivation, then rational belief is responsive only to evidence. Beliefs formed on the basis of desires are irrational. So, not all normative reasons are internal reasons. Internalism about practical reasons therefore might seem arbitrary and unmotivated (cf. Vogler 2002, Thompson 2008). According to Millgram, that new experiences can reveal to us previously unknown reasons for belief shows that such new experiences can reveal to us reasons for action. There seems to be two strategies open to an internalist who wishes to criticize Millgram’s account.3 First, the internalist might extend their practical reason internalism to theoretical reasons for belief, i.e., people have a desire for knowledge or truth (cf. Kelly 2003; Velleman 2000). This strategy seems to fail outright because it begs the question against the reason externalist. A second, and perhaps more promising, strategy is to deny that practical and theoretical reasons are relevantly similar. For example, Markovits (2011) has argued that there is no practical analogue of foundational beliefs in epistemic cases. This argument seems to be susceptible to the something like the Goodmanian critique of similarity (1972). Given the notorious difficulties associated with clearly defining similarity and giving a principled way of delineating it from other forms of similarity, we ought to abandon it. In this paper, I would like to address the possibility that a third criticism of Millgram’s externalism is possible once we have uncovered an implicit position his view has overlooked. In “Pleasure in Practical Reasoning” (Millgram 1993) and Practical Induction (Millgram 1997, ch. 6), Millgram has argued that carrying out our actions follows from experience about what is desirable, and this is a kind of inductive practical inference that plays a legitimate role in our thinking. “Practical reasoning includes an analog of induction, through which one can acquire new ends by learning from experience” (Millgram 1993, 396). On Millgram’s view, we draw general evaluative conclusions from the experiences we have. Practical induction may sound much like an instrumentalist theory of practical reasoning, i.e., the world is important only insofar as it plays some role in a means to some specified, desirable end, but the two should not be confused. In Millgram’s words, “On the instrumentalist view, you do not want things because they matter; they matter because you want them” (Millgram 1997, 5). Since instrumentalist theories fail to rule out completely unjustified desires - desires had “at pill” according to Millgram (1997, passim) - they do not provide an adequate basis for drawing any conclusion about what we have reason to do. Practical induction has to be informed by the way

3

It is interesting to note that no one has challenged Millgram’s account until now. What I argue here should not be considered a decisive criticism of Millgram’s position; rather, I take it to be a first-step toward a more full-fledged criticism of practical empiricism.

-!2-

things actually are. So, practical induction begins in a practical analog of observation.4 “Attention to the cognitive role of pleasure shows how the world is given a further say in practical reasoning. Practical reasoning is informed by something that can be considered the practical analog of observation” (Millgram 1997, 120). It is pleasure that is practical reasoning’s analog of rockbottom experiential beliefs, and just as experience is useful for forming beliefs, pleasure is a guide for figuring out what to do. According to Millgram, commitment to the truth of some belief cannot in practice be manifested in anything beyond one’s inferential commitments, and some beliefs have a special status. Millgram has called these special beliefs “rock-bottom experiential beliefs,” defined as beliefs “not inferred from other beliefs” (Millgram 1997, 113). Rock-bottom experiential beliefs put us in a position to check whether our beliefs stand in inferential relations of conflict, compatibility, support, and so on, with other beliefs (Ibid.). For example, Alice may inform us that she believes there are eight eggs in the carton. If Alice asserts her belief without checking the carton, then her belief is not rock-bottom. Experience acts as a litmus test for rock-bottom beliefs. From here, Alice can infer whether she has enough eggs for a souffle, for an omelet, or for baking cookies. Millgram has said that a feeling of conviction accompanies a rock-bottom belief. When Alice looks in the refrigerator and finds eight eggs in the carton, the “a-ha” feeling that accompanies checking the carton in the refrigerator is the feeling of conviction. Suppose that I ask a person how Alice knows there are eight eggs. She may respond by saying, I just looked in the refrigerator. What do you mean, ‘How do I know?’ Of course I know’. If I stop myself in the middle of such a tirade, the feeling I find there is typically a feeling of conviction. (Millgram 1997, 114) If the person cannot deny that there are eight eggs in the carton, and if the person is convinced by what she saw in the carton, then the feeling that accompanies this experience is the feeling of conviction. We do not always have to be consciously aware of the feeling of conviction, precisely because it is a feeling. Millgram has claimed: Let me pause a moment to say what I mean by ‘feeling’. I am going to distinguish feelings from sensations. […] Feelings... may be had unawares. [...] Feelings can of course involve sensations. [...] We often identify feelings by the sensations they involve. (Millgram 1997, 109) Feelings, like beliefs, fade from conscious experience. But, when we are reminded of the conviction, the conscious feeling returns. If someone asks Alice about the number of eggs later in 4

I will not have the space in this paper to contend with Millgram’s wholesale dismissal of instrumental theories of practical reasoning. Millgram has argued against instrumentalist theories of practical reasoning in several places, and he has argued that instrumentalist theories are not the same as what has been called ‘Humeanism’ (Cf. Millgram 1995).

-!3-

the day, she will respond that there are eight. How does she know? Earlier in the day she looked and they were there. Alice believes it because she is convinced of it. Millgram has claimed the experience test shows there is little reason to think there is a distinction between belief and conviction. If, by looking, I “know” that there is an egg carton in the refrigerator, then I must believe it. Conviction, thus, gloms onto beliefs in virtue of the proposition’s logical content. Now that I have given a brief overview of Millgram’s rock-bottom experiential beliefs we may turn our attention to criticisms of it. In the next section, I will argue that the feeling of conviction upon which experiential beliefs depend is an antecedent commitment or condition that has to be had for the cognitive phenomenon to be rock-bottom. What is rock-bottom, then, is not experiential beliefs but the feeling of conviction. If experiential beliefs are not rock-bottom, then we should expect that pleasure, practical reasoning’s analog of rock-bottom beliefs, is not rockbottom, either. 2. Are beliefs and convictions identical? Millgram’s observational account has presumed that a belief and a feeling are indistinguishable. If it is possible to show that beliefs and convictions are non-identical, then this distinguishes the two and creates problems for Millgram’s account. Let’s begin with discussing whether the feeling of conviction is also a belief. What distinguishes a feeling from a belief is that the latter is truth evaluable, i.e., beliefs can be either true or false and feelings cannot be. Either the feeling of conviction, whether we speak of it as having cognitive content or not, is truth evaluable or it is not. According to Millgram, the feeling of conviction is the “a-ha” moment that accompanies one’s experience. The “a-ha” moment does not necessarily have any propositional content. Even if the “a-ha” moment had propositional content, it would seem unusual for a bystander to claim that a person’s “a-ha” response is wrong or false. So, it must not be the case that the feeling of conviction is truth evaluable. Beliefs do have propositional content and are truth evaluable.5 That Ned believes that the Empire State Building is 250,000 feet tall is false and wrong. The content of Ned’s belief is inconsistent with the way the world is. Anyone present when Ned proclaims such a falsehood would be warranted in asserting, “Ned is mistaken, the Empire State Building is 1250 feet tall.” If beliefs have a property feelings do not, then feelings and beliefs are not identical. The feeling of conviction is not truth evaluable and, therefore, cannot be a belief. Call this the “simple argument” against the identity of the feeling of conviction and belief. The simple argument shows that conviction and belief have non-overlapping properties. Perhaps there is a way in which we think of beliefs and convictions where the two come out to be 5

One might contend that intuitive beliefs are less like beliefs than they are the content of a feeling of conviction. To my mind, this would be to misrepresent intuitions. They, like beliefs, are truth evaluable, even though philosophers often speak of having a feeling accompanying them that sounds very much like what Millgram has described as the feeling of conviction. Since they are truth evaluable, they function in the same way as beliefs do. Intuitions have propositional content in much the same way that ordinary beliefs do. I will not have the space in this paper to address any concerns one might have with intuitions.

-!4-

identical. One could argue, for instance, that if a conviction and a belief have all the same causes and same effects, then the two are identical (Cf. Davidson 1969). In order to undermine Millgram’s adoption of this response to the simple argument it will have to be shown that our beliefs and convictions do not have all the same causes and same effects. Interestingly, there is an example from Davidson’s understanding of what it is to be human that might speak against Millgram. Imagine we uphold a conviction that ‘humans think’, but we do not seem to deny that human-like creatures think. Would we say that a creature not composed of the same biological materials as us fails to think? If I were to discover that my best friend had been born by hatching from an egg, or had been conceived by a process that required the active collaboration of three creatures of different sexes, it would probably not influence my opinion that he or she could think... [I]f my friend turned out (after all these years) to be made of silicon, I’d change my mind about what materials a person might be made of, not my judgment that he was a person. (Davidson 2004, 79) Davidson is convinced that his friend who was spawned by three distinct beings and composed of silicon is a thinking thing despite that Davidson’s conviction, i.e.,’humans think’, would undermine his view that his friend is a thinking thing. The discovery that his friend is not human in the ordinary biological sense would give rise to the belief ‘my friend is not human’, but this would harm neither his belief that ‘his friend’s a thinking thing’ nor his feeling of conviction ‘only humans think’. All of this seems to show that commitment and belief can be treated separately. Therefore, just as the simple argument concluded, the identity of feeling and belief is mistaken. Despite the simple argument, Millgram has defended the view that the feeling of conviction and belief are identical. He has argued: It may sound peculiar to identify a feeling with a judgment. Possibly this is because one feels that there must be more to a judgment than a feeling: possibly because one can make judgments that do not feel like much at all. But recall that I am not using ‘feeling’ as a synonym for sensation: because it is not a sensation, a feeling is not precluded from having cognitive content in the way one might think sensations are; and, as I remarked earlier, one can have feelings of which one is hardly, or not at all, aware. I am avoiding distinguishing feelings from judgments because the distinction seems too forced here to be useful. But one can just as well think of these feelings as aspects or accompaniments of judgments. (Millgram 1997, 116) A rather crude rendering of Millgram’s argument is the following: “Since a feeling is not precluded from having cognitive content and since beliefs have cognitive content, feelings and beliefs are identical.” The reconstruction is crude because it fails to account for the aspectual nature of identity Millgram seems to endorse toward the end of the selection. Notice that Millgram seems to argue that the identity he has in mind is one where “feelings [are] aspects or

-!5-

accompaniments of judgments” (ibid; my emphasis). The identity that Millgram has in mind here does not seem to be the traditional strict form of identity but compositional identity. Undermining Millgram’s compositional identity might require engaging in some conceptual gymnastics because this form of identity rejects the central tenets of strict identity. To claim that feelings are “aspects of judgments” is to claim that one is part of a greater whole. Suppose that Diana’s watering her petunias is made up of three activities, Diana’s filling the watering can, Diana’s walking over to her petunias, and Diana’s tipping the watering can. There is some sense in which we might claim that Diana’s watering her petunias is nothing more than these three activities. According to the composition as identity thesis, Diana’s three activities are identical with Diana’s watering her petunias (cf. Baxter 1988, Lewis 1991, Sider 2007).6 If compositional identity is what Millgram seeks to endorse, then it suffers from myriad problems. First, and perhaps most obviously, if we believe in the commonsense definition of identity, i.e., for x and y to be identical, x and y must share all the same properties and Diana’s activities have properties Diana’s watering her petunias does not, then parts of Diana’s watering her petunias are not identical to her activities. Second, since Diana’s filling the watering can occurs prior to her tipping the watering can, the two cannot be identical with Diana’s watering her petunias. The act of watering her petunias would extend through time where we might not want to say at one time she’s watering her plants. If Diana were to die but her watering can were to tip in just the way we might think it should had she lived to see it through, we would not want to say of dead Diana that she’s watering anything. After all, she is dead and dead people cannot act. Finally, given that watering the petunias continues to take place after Diana has completed her activities, we might have to argue that there are aspects of Diana’s watering her petunias which are beyond her control. We would not want to say that something beyond the control of a person is an aspect of an action. Otherwise, we would be forced to say that unintended side effects of some action are the direct result of an agent and they should be held morally accountable for those side effects. This discussion leads us to reject the compositional view that conviction accompanies or is an aspect of belief, as we might have gleaned from Millgram’s discussion above. Overlooking the compositional nature of Millgram’s concept of identity would lead us to believe that the only thing we have to show is that convictions and beliefs are non-identical in the more traditional Leibnizian understanding of identity. The argument might go something like this. If feelings and beliefs differ on any property, they cannot be identical according to Leibniz’s Law. Any difference in the properties of X and Y indicates that they are not strictly numerically the same thing. X and Y are not identical. The question, then, for Millgram’s account is whether the property of “not being precluded from having cognitive content” is the same as the property of “having cognitive content.” Using a practical example, I believe we can show that there is a distinction between the two ways of conceiving the property of cognitive content. “Nothing prevents Kevin from asking Winnie for a date” is not the same as “Kevin asking Winnie for a 6

My discussion of composition as identity cannot do justice to the different ways in which we might conceive of the composition thesis. For a good overview of the different conceptions of the view, e.g., the strong and weak compositionally thesis, and arguments against it, see Wallace 2011a,b.

-!6-

date.” Acknowledging that nothing precludes Kevin from asking Winnie for a date means that no impediment, pardoning his own embarrassment or his reluctance to ask a woman for a date, prevents Kevin from asking, while asking Winnie for a date means not only that there was nothing preventing Kevin from asking Winnie but that he actually did ask her on a date. When Millgram claims that beliefs have cognitive content and convictions are not precluded from having them, he seemingly admits that differences exist between the two; any differences between them, thus, shows that they are not identical. What makes Millgram’s identity of the feeling of conviction and beliefs critical for his overall position is summarized in fn16, 114f: When a belief is rock-bottom, we fall back on conviction; so it is not surprising that we invoke conviction to signal the non-inferential origins of a belief. (Millgram 1997, 114f, fn16) Millgram supposes that a feeling of conviction is critical for the formation of belief, i.e., what we are to believe. If one has a belief without the feeling of conviction, then it appears that the belief could be easily dismissed by the agent. If one has a feeling of conviction in a belief the preponderance of evidence shows that we ought to reject, the person might be reluctant to give it up. Millgram argues that: The feeling of conviction (or, I will just say, conviction) plays an important role in one’s epistemic economy—particularly when it is of the rock-bottom, experiential variety. If, on considering a proposition, I find it unconvincing—if I lack the feeling of conviction— I may decide that I am on the wrong track (or that someone else is). If I do not feel conviction in situations in which I’m face to face with the object of my would-be belief (that is, when experiential rock-bottom belief is at issue), then it’s going to be very difficult to convince me. And if in such situations I do feel conviction, my views on the subject will be difficult to dislodge. (Millgram 1997, 115) At least one way in which the feeling of conviction plays an important role in the formation of beliefs is for the two to be identical. Just as the feeling of conviction is critical in the formation of belief, it must have a practical analogue, namely desire, which is critical for the formation of action. So, we might believe that the “rock-bottom” conviction is identical with the belief it accompanies. Millgram’s account rests on a certain feeling of conviction. I agree with him that the feeling of conviction is important for rock-bottom beliefs, but I disagree with him when he thinks that the feeling of conviction bottoms out in experiential beliefs. There is some antecedent condition that must be satisfied for conviction to take hold of an agent. In this section, I have shown how belief and the feeling of conviction may be distinguished. If the two components critical for the success of practical induction are not identical, then we have reason to ask whether Millgram’s view that experiential beliefs is more fundamental or rock-bottom than the feeling of conviction. According

-!7-

to Millgram, the former and not the latter is rock-bottom. In the next section, I will argue that the feeling of conviction should be more fundamental than, and antecedent to, experiential belief. 3. Are convictions more fundamental than rock-bottom beliefs? Millgram’s practical induction unjustifiably attaches commitment to belief, but, from the argument above, we are able to see why the two could be distinct. On my view, they are not only distinct but convictions are more fundamental than experiential beliefs, thus making the feeling of conviction rock-bottom. In this section, I will argue for this claim so that in the next section I can argue that convictions are logically prior to experiential rock-bottom beliefs. When we look more carefully at Alice’s belief about the number of eggs in the carton, we may ask whether her feeling of commitment is derived from the belief’s propositional content or from the principle that her perceptual faculties do not usually fail her under ordinary circumstances. If Alice’s conviction depends on the propositional content of her belief, then it is experience that leads her to having the conviction. If Alice’s conviction is the result of some antecedent condition, i.e., the principle that perceptual faculties in fine working order do not most of the time fail the perceiver, that frames the disposition to believe what is part of the propositional content of experience, then it is not experience that leads her to have the conviction she does. Something preceding or closely accompanying her experience guides her to be committed to what she experiences. One might contend in opposition to my argument that a more general experience with certain beliefs might lead to certain convictions.7 No one should deny that a more general experience, say having the same particular experience over and again, is not an experience in the way Millgram contends, so his view might not be mistaken after all. While I agree that there is no denying that we have a more general conception of experience as a whole, which would include the more particular direct perceptual experiences we have through our sensory modalities, I believe that my argument applies equally well to the whole of experience, which might pertain not only to me as a cognizer but to all cognizers, as it does to the particular beliefs. Think of it in the following way. If some further and more general experience recommends that our perceptual faculties usually work properly, then our belief that experience on the whole is more trustworthy because particular experiences have given rise to correct information. What lies behind experience generally is the conviction that we ought to trust any faculty that has repeatedly given us good information, upon which we can base our beliefs (in theoretical matters) or our actions (in practical matters). Behind a thin veil of experience lies the true nature of one’s practical and theoretical commitments. The conviction which gives rise to the trustworthiness of general experience has to do with our notion of trust.

7

I am grateful for an anonymous referee pointing out this potential response to my argument.

-!8-

Convictions, on Millgram’s view, can be drawn from experience or testimony of others’ experience.8 The feeling of conviction is not limited to direct perceptual experience. So, there should be cases where we can decide to give up some of our convictions or refuse to surrender our commitment. Even if there are formidable countervailing considerations that should overwhelmingly persuade us to give up our conviction, we might not do so. I am thinking here of situations where we uphold a belief with such a feeling of conviction that we see no reason to reject it, even though there are plenty of reasons why we should not be committed to that belief. Suppose that Ted believes that ‘aliens pilot unidentified flying objects (“UFOs”)’, and his sense of conviction accompanying this belief is quite high. Ted may observe a UFO in the night sky, and, based upon his conviction which he may have formed by listening to the testimony of others calling in to AM radio late-night talk shows or after having been “abducted” as a child, he may believe that ‘this UFO is piloted by aliens’. Ted may discover later that what he observed was a military test aircraft piloted by a human. His belief turned out false in this particular instance. Ted may abandon his belief that the UFO he observed was piloted by an alien, but he need not abandon his belief based upon a strong feeling of conviction that ‘aliens pilot UFOs’. Ted might not be so easily persuaded to give up his belief that aliens pilot UFOs by learning that the UFO he observed this one night was piloted by military personnel. Of course, given his feeling of conviction, Ted might not be so quick to judge that the UFO he observed in the night sky was piloted by a military test pilot. Nothing proscribes Ted from upholding his belief that that particular unidentified aircraft he observed that particular evening was piloted by an alien, and this would be in consonance with Millgram’s view of rock-bottom beliefs. Ted may be a conspiracy theorist, one who believes that the government attempts to cover-up anything that is contrary to what they want ordinary citizens to believe. His conspiratorial nature may have developed during the course of the Vietnam War era, where the massacre of women and children by American GIs was common and the military’s cover-up of these heinous wartime “operations” was even more common. When Ted hears the news that the UFO was a military test aircraft, he sloughs it off and dismisses the report as yet another instance of a government cover-up.9 On account of his own experience with previous government cover-ups, Ted hangs on to his 8

One might be tempted, albeit wrongly, to suppose that my view hoists upon Millgram’s notion of rock-bottom beliefs that we must conceive of them in one way and only one way for my argument to work, namely that experiential rock-bottom beliefs are absolutely foundational and drawn from one’s direct perceptual experience. This is a gross misinterpretation of my view; rather, I take it that rock-bottom beliefs are multiply realizable and that they can be drawn from numerous sources. I have tried to be clear that I do not take up this mistaken view of Millgram’s notion of rock-bottom beliefs, and, instead, I have chosen to focus upon what ultimately seems to hold rock-bottom beliefs together, i.e., what makes them “rock-bottom,” which Millgram seems to overlook time-and-again in his discussion. 9

Compare this belief with what climate change deniers have said in response to the general scientific community’s claim that climate change is real, occurring now, and much of it is human induced.

-!9-

conviction that the UFO he observed in the night sky was piloted by aliens. Ted uses the rockbottom experiential belief, “the government deliberately hides information from its citizens,” to uphold his view that aliens piloted the UFO, and this is consistent with Millgram’s view. Millgram might complain that the Ted example fails to account for the defeasibility of beliefs. When we claim that a belief, experiential or not, is rock-bottom, the designation implicitly suggests that it should not be abandoned unless there is good reason to do so. A rock-bottom belief is a strong, immediate belief that is not derived from other beliefs. It could be, though it need not be, foundational, in that, the rock-bottom belief anchors the other beliefs that can be derived from it. Under circumstances where the strength of a belief has been compromised by a more powerful belief, we should consider rejecting the less powerful belief. Our overturning a belief, however, has nothing to do with whether the belief is experienced or not; rather, we derive some belief being compelling from the conviction we have associated with it. In Ted’s case, the strength of the commitment outweighs the belief’s propositional content. The experience either calls into question the commitment or it does not. Ultimately, the feeling of conviction could possibly win over the experiential belief. The argument that an experiential belief might not be rock-bottom suggests that there is something other than it that makes a belief rock-bottom. If Alice’s commitment to the belief is rock-bottom, then it is the feeling of conviction and not the experiential belief that makes it so. Alice’s conviction may be a principle she would not give up unless there was a stronger conviction that promised a greater practical pay-off. The principle Alice seems to assume is (something like) whatever I observe is an accurate reflection of the way the world is. Whatever Alice sees, she will trust corresponds with the world because her visual perception under ordinary circumstances has not deceived her in the past. Since this is an instance where ordinary perceptual experience has not failed her, Alice believes that there are eight eggs in the carton. Hitting rock-bottom in Alice’s account is to uncover the conviction upon which her belief rests. We can imagine that if Alice was not convinced by experience, then she would check the contents of the egg carton over and over and over again, ad infinitum. Alice may not be satisfied by checking the carton once, twice, or three times. No matter how many times Alice checks the carton, unless she has some form of conviction to the reliability of her perceptual faculties, she will continue to check. So, what is rock-bottom might not be a belief but a commitment or conviction that accompanies a belief. There is nothing about the propositional content of Alice’s belief or how Alice acquires the belief that makes it suitable to be rock-bottom. Conviction that our perceptual faculties are trustworthy yields the acceptance of beliefs gleaned through experience. Alice can believe that her experiential beliefs are correct if and only if her conviction is intact. Alice will believe what she sees because the accuracy of her perception is a backward inferential commitment of hers that she will not repudiate under ordinary circumstances. 4. Are Commitments Logically Prior to Beliefs? What I have argued so far is that conviction is a more fundamental component of our cognitive architecture than the experiential beliefs we derive from observation, so experiential

-!10-

beliefs cannot be rock-bottom. A question we may have is whether commitments are logically prior to beliefs, especially the content of one’s beliefs? Commitment arises from either (i) experience or (ii) an antecedent condition distally present in experience. For (i), Alice has a commitment to the experiential belief that “there are eight eggs in the carton,” since she has checked the egg carton and there were eight eggs in it. For (ii), an antecedent condition distally present in experience exerts influence upon Alice to be committed to the belief that there are eight eggs in the carton. A condition “distally present in experience” is a state-of-the-mind or principle that an agent need not be consciously aware of unless the agent is reflecting upon it. If Alice’s belief that “there are eight eggs in the carton” arises from observing them as well as a condition distally present in experience, the conviction that the perceptual faculties under ordinary circumstances yield correct information, then Alice is committed to whatever it is she sees. In the latter case, conviction is logically prior to experiential beliefs. I have argued that there is a distinction between deriving a commitment from experience and deriving commitment from an antecedent condition present in experience. The propositional content of belief concerns whatever a belief is about. Conviction is a feeling or disposition a person takes toward the propositional content of her belief. The distinction between the two leads us to a second effect. It seems that we cannot have (i) without (ii). If the correct antecedent condition was not present, then a person would not necessarily be committed to the propositional content of experiential belief. The content of an experiential belief depends upon the commitment to an antecedent condition that frames our perceptual beliefs. For example, if Alice was a skeptic and she did not trust her visual perception, then she would not be committed to the propositional content of the belief. Alice’s commitment to the propositional content of her belief arises from her feeling that perceptual faculties are reliable, for example. Thus, (ii) is a necessary condition of (i). A commitment to the trustworthiness of perceptual faculties is logically prior to the allegedly rock-bottom beliefs obtained via sensation. When we discuss commitment, we need to be able to distinguish between (i) and (ii) and to show that (i) necessarily requires (ii). If an account fails to distinguish between the two types of commitment or does not argue for the presence of certain antecedent conditions, then it is inadequate. Millgram’s account fails to distinguish the two. So, in the remainder of this section, I will argue that his account is inadequate. Let me begin with an example. Suppose that Elaine is a member of a town’s public school administration. The school board is a governing body that determines the length of the school year. She trusts whatever the school board claims about the length of the school year. On Millgram’s view, it is an experiential belief underlying Elaine’s conviction. Experience, however, does not seem to inform Elaine of why she believes the school board; rather, it is the trust she has in school boards. The school board is the primary administrative unit endowed with the power to dictate the length of the school year. When the school board says that the school year ends June 22nd, regardless of what her experience has been with a school board or its members, Elaine believes that the last day of school is June 22nd. Elaine’s feeling of conviction is not determined by the propositional content but by her feeling that school boards are trustworthy groups of

-!11-

people. With her trust in the school board, she can begin planning her summer vacation to Macau. Commitment determines which beliefs are rock-bottom. The rock-bottomness in Alice’s case is that she is convinced of the accuracy of her senses. Elaine’s belief is rock-bottom because of her trust in school boards, not because she has heard the school board declare the last day of school is June 22nd. The feeling of conviction in Elaine’s case and Alice’s case depend on the trustworthiness of the sources of experiential beliefs. A critic might claim that we obtain trust through experience since knowing whether an agent or faculty is trustworthy comes from the experience we have of the faculty or agent. The trouble seems that this need not be the case. For example, Elaine would trust the school board from the time before the first day on the job.10 In fact, if Elaine asks the school board for evidence guaranteeing that the school year ends on June 22nd, the school board may believe that she is questioning their authority. If trust is not something we obtain through experience, then it is an antecedent condition. Therefore, commitment is an antecedent condition more fundamental and logically prior to beliefs, particularly those beliefs we learn through experience. Practical induction fails to adequately account for this sort of commitment. It places the content of belief before commitment. As we have seen, an adequate account of the rockbottomness of belief has to make room for commitment. In fact, commitment determines that the beliefs are rock-bottom. Since commitment depends on nothing other than itself, it seems that practical empiricism would have to accommodate this strong, immediate belief independent of all experience. Practical induction, however, cannot argue that commitment is an antecedent condition logically prior to the content of belief. Millgram has argued that one’s belief is convincing only insofar as it is drawn from the appropriate circumstances. Millgram writes: Suppose I inform you that I believe that there is milk in the refrigerator. I am committed to the milk’s in fact being in the refrigerator: if it is not there, I am wrong. I expect that if I go and look, it will be there. Now what does going and looking amount to? I put myself in appropriate circumstances (by walking up to the refrigerator and opening the door): in these circumstances, I come to have a belief that there is a carton of milk on the top shelf. (Millgram 1997, 113)

10

It could be argued that Elaine’s trust of school boards decreases as time passes. Elaine could lose her trust in school boards because an experience exposes how untrustworthy some members of the school board are or how the school board intentionally deceives employees. In cases like this one, it is the experience that drives her to surrender the original conviction. Millgram might argue that this shows how rock-bottom beliefs are experiential. The trouble with this response is that should Elaine move to another school district or that all of the school board members be fired overnight, her original commitment, trusting the school board, would be restored and would be unaffected. She would have no problem upholding her trust of the new board members. It is up to them whether Elaine’s trust waxes or wanes during the course of her tenure.

-!12-

By walking up to the refrigerator and opening the door, we find the milk carton on the top shelf. “Seeing is, often enough, believing” (Millgram 1997, 115). So, it seems that experience - visual perception, that is - tells us what we believe. There seems to be quite a bit loaded into the “appropriate circumstances.” Millgram’s analysis has presupposed that my sensory modalities are reliable belief forming faculties, and sensory data are adequate evidence for belief. If Millgram is correct, then his position is: M1: S is justified in believing p if and only if S has adequate experiential evidence supporting p. The adequate experiential evidence involves checking that some state-of-affairs obtains in the world. The background commitment Millgram seems not to acknowledge is: M2: S always ought to follow his/her evidence. This is the distally present non-experiential antecedent condition I spoke of above. If it is an antecedent condition, as I want to claim, and if it is a part of M1, then commitment to M1 is not rock-bottom because it depends upon M2. M2 is a principle one has independent of the experience of M1. So, rock-bottom belief is not a matter of experience but an antecedent condition present in experiential beliefs. Without M2, M1 is unwarranted. If we adhere just to M1, we are left with adopting a form of practical skepticisms. Skeptical scenarios question our commitment to experience as a justification condition for knowledge. But skeptical scenarios seem to undermine something even more robust than experience. The purpose of skeptical scenarios is to raise questions about whether we ought to be committed to experience as justifying some knowledge claim. For instance, Russell’s skeptical scenario about the fiveminute old universe shakes our commitment to memorial beliefs. So, any belief we may uphold about what happened ten minutes ago or ten years ago is incorrect, despite the apparent vividness of our memories. Just as skeptical scenarios apply to epistemic cases, they may hold for practical cases, too. What practical skepticisms might look like is quite disturbing indeed. Returning to the Alice case I mentioned above, if she was a practical skeptic, she might constantly check for the eggs. No matter how many times she confirms her belief, her desire for checking that she has eight eggs might never be fully satiated. For Elaine, she would be constantly questioning her allegiance to the school board and what they declared in their most recent meeting. The practical skeptic would keep returning to the refrigerator or keep questioning the authority of the school board, despite having checked multiple times. Practical skeptics, interestingly enough, are not figments of our imagination since people with obsessive-compulsive disorder seem to behave in ways consistent with the practical skeptic.11 The behavior might be labeled ‘neurotic’, but, ultimately, this sort of neurotic behavior would be no different than the epistemic skeptic unsure of the existence of an external world. Of course, 11

I am grateful for a referee pointing out that the practical skeptic I have in mind is not as outlandish as one might believe.

-!13-

it might not be neurotic, in the clinical sense of the term. Sometimes when I leave the house I wonder whether I turned off the oven, even though I know that I checked it twice before leaving. Or, if I think that I left the garage door open and if I remember pushing the remote control button closing the door as I left the driveway, nevertheless, I return home to double-check whether the garage door is actually closed. “I might be a brain-in-a-vat,” says the epistemological skeptic. So too for the practical skeptic, “I might not have eggs.” Millgram’s practical inducer does not seem to fare any better than the practical skeptic. There is one final criticism I would like to argue against the view that it is not conviction that is rock-bottom. It has to do with whether experiential beliefs are contextually insensitive. If Millgram’s rock-bottom experiential beliefs are context insensitive, then the rock-bottom belief one has could not morph for alternative arenas of experience. Our backward looking inferential commitments might have potentially disastrous practical outcomes. Suppose that Alice’s friend, Carol, is playing golf one day and approaches a young lad, who’s selling gently-used golf balls in egg carton containers for $12.00. Since all of her previous experience tells her that these kinds of containers hold eggs, she is at first surprised by the price of these particular eggs, $12.00 is exorbitantly expensive. When she checks the containers, she finds golf balls in place of the eggs, undermining her previously held conviction that egg cartons contain eggs. That is the “a-ha” moment of which Millgram has spoken. After her round of golf, when Carol goes to the grocery store, she happens upon the dairy aisle where there are numerous cartons of eggs. Do these cartons contain eggs or golf balls? Her experience has suggested cartons sometimes contain eggs, while at other times contain golf balls. Given that Millgram’s account does not seem to vary by context, it would not be unwise for Carol to check every carton in the dairy aisle for the one containing golf balls. If conviction depends upon experiential beliefs, then a situation like Carol’s would not be uncommon. Every time she had an experiential belief altering her conviction, she would uphold it until shown that it ought to be abandoned. On Millgram’s account, we are left to wonder whether contextual factors play any role in determining whether we uphold or abandon some rock-bottom experiential belief. Millgram’s account seems to leave no room for the contextual features of an experience to determine an agent’s experiential belief. 5. Conviction is an all or nothing affair I have shown that Millgram’s observational account depends upon rock-bottom antecedent proximal conditions accompanying experiential beliefs and that insofar as his account has to do so we must admit that it is not proffering the most fundamental element by suggesting that experiential beliefs are rock-bottom. There is at least one more argument I would like to level against Millgram’s observational account. In this section, I will argue that conviction is an underlying feature of experiential beliefs, which we either have or do not have. Some might believe that this all-or-nothing component exemplifies a failure of an account of reasoning, but I show why this is not the case. When our experiential rock-bottom belief is at issue and we do not feel any conviction toward it, convincing us otherwise is virtually impossible. If we do feel conviction, then our views on

-!14-

the subject will be difficult to dislodge. The feeling of conviction is something that either a person has or does not have. It seems there is no median state. Perhaps one might argue that this is a false dilemma because conviction comes in degrees, conviction can fluctuate, or fall somewhere in the middle where a person is neither convicted nor not convicted. First, when a person is neither convicted nor not convicted to x, we cannot claim that the person upholds any kind of conviction whatsoever. The person may hold the middle ground between two extremes. This middle ground, however, is outside the parameters of conviction. The person may be undecided or teetering between two choices. Think here of the political independent that composes a large part of the U.S. electorate. Uncommitted voters have not decided whether to support one candidate or another. Their convictions are absent. Because no convictions are present, we say they are undecided. This does not stop candidates from pandering to them for votes. It is precisely because these constituents do not have convictions that candidates feverishly pursue them. The candidate wants their conviction, i.e., their vote. So, whatever being uncommitted or undecided is, it does not exemplify any kind of conviction. Second, if conviction fluctuates for a single individual, then a person might be convicted to x at one moment and not convicted to x at another. According to my interlocutor, if this is correct, then the view that the feeling of conviction is present or it is not is incorrect. I fail to see how this is a viable response to my claim that either a person is convicted to something or not. For this response to work, conviction must be atemporal. By that, I mean that it must be said of an agent that the conviction upheld at time t1 but suppressed at time t2 can be simultaneously ascribed to the agent, from time t1 to t2. Conviction C believed by agent S at time t1 and Conviction C* (whether it is the contradictory of C or not) believed by the same agent, S, at time t2 must be acknowledged and endorsed simultaneously for us to believe that conviction fluctuates. But our convictions do not behave like that. For example, Brown has a conviction at time t1 that the 1961 New York Yankees had the greatest lineup in baseball history. If Brown later rethinks his conviction and revises his opinion of the greatest baseball lineup to the 1975 Boston Red Sox (at time t2), then we cannot say that Brown is convicted to both the 1961 New York Yankees and the 1975 Boston Red Sox as being the greatest baseball lineup. Convictions are time- or contextsensitive, but this sensitivity does not mean that they are in a constant state of flux. When we uphold one commitment, we reject all others. So, it is still the case that we either have one conviction or we do not have it. Finally, a critic might argue that convictions come in degrees.12 If a person sorta has a conviction that Hunter Thompson is an amusing American journalist, then there might be times when the person’s conviction to the claim is greater (perhaps when she is thinking of Fear and Loathing on the Campaign Trail ’72) than at other times (say when she is thinking of The Rum Diary). “Sorta” is a practical modal operator that strengthens or weakens according to relevant contextual factors. Should this be an accurate rendering, so the critic will argue, convictions need not be an all-or-nothing affair. Relevant contextual factors may determine the strength of conviction, the degree to which an agent may be committed to some claim, but contextual 12

I believe that Millgram (2009) has done yeoman’s work of partial modifiers on practical matters of truth. Perhaps my rendition of this criticism is not compatible with what he says there, but I will not address that work in this paper.

-!15-

elements do not eliminate or constitute the convictions one has. We are convinced that Hunter Thompson is an amusing American journalist, whether we acknowledge it more so when we are thinking of one book than we would when thinking of another book of Thompson’s. An agent had to have had the conviction in the first place for contextual elements to adjust the strength of one’s convictions. Therefore, again, the potential counterargument fails to undermine my view that either one has a conviction or not. I have argued in §§2, 3, and 4 of this paper that the cognitive origin of conviction is a distally present antecedent commitment an agent either possesses or not, and we should contrast this position with Millgram’s observational account where we are committed to those beliefs we draw directly from experience. Experience seems less likely to give us something we would call rockbottom; instead, we may call our experiential beliefs noninferential. A noninferential belief amounts to a belief which cannot be inferred from other beliefs. Non-inferential beliefs should be contrasted with rock-bottom convictions, which are those commitments that do not require experience at all. Non-inferential beliefs accompanying experience are strong, immediate beliefs of ours.13 This section has shown that convictions are an all-or-nothing affair. If certain experiential beliefs may be rock-bottom and accompanied by a feeling of conviction, then it is possible to give an account of convictions independent of our account of experiential beliefs. This is exemplified by the point that we either have the feeling of conviction or we do not. In the next section, I will argue that pleasure as practical reasoning’s analog of rock-bottom experiential beliefs turns out to be something non-experiential, as well. 6. Pleasure: Practical Reasoning’s Analog of Experiential Rock-Bottom Conviction According to Millgram, the pleasure that we have in response to an object of experience amounts to an estimate of the object’s desirability. I argued in previous sections that, for the theoretical analog of Millgram’s observational account, experiential beliefs depend on some antecedent condition held independent of experience. In this section, I will show how our practical inferences cannot bottom-out in our experience of pleasure. Some antecedent condition 13

I should say here that “rock-bottom” does not mean “incorrigible;” one should not imply from what I have argued in this paper assumes that “rock-bottom” convictions are indefeasible. I would like to remind readers of what I said above regarding the revision of some conviction in light of certain contextual or temporal elements. A certain conviction may be altered in response to new evidence. If I was suggesting that a rock-bottom conviction was indefeasible, then I would be begging the question against Millgram’s observational account. Millgram has acknowledged the corrigibility of rock-bottom beliefs: A belief’s being rock-bottom carries no implication of indefeasibility: no matter how “observational” my belief, I may later retract it. Also a belief’s being rockbottom carries no denial that there are necessary conditions of its acquisition that must be described in terms of further beliefs. A good deal of background is typically required for coming to have rock-bottom belief. (Millgram 1993, 398)

-!16-

exists, such that there is some distally present principle holding together the pleasure we experience. If pleasure is not the rock-bottom judgment of desirability, then, as I will claim, it must be some principle that does not originate in pleasurable experience. Therefore, Millgram’s account of rock-bottom judgments of desirability is mistaken. Millgram has argued: Often, to the expectation that when one puts oneself in the appropriate situations, e.g., that when one actually gets what one wants, one will not be horribly disappointed and wish that one had never heard of it. Rather, one expects that when one gets it, it will turn out to be desirable. Now a primary indicator of whether something is desirable or not is pleasure. Pleasure is the rock-bottom judgment of desirability of an object of present experience (and likewise, displeasure is the rock-bottom judgment of undesirability, also directed toward present experience). (Millgram 1997, 115) Wanting to do something and actually doing it, according to Millgram, will likely result in a desirable experience. Whether we find something desirable depends on the pleasure the experience elicits.14 Our power to anticipate that a particular experience will be pleasurable does not always work correctly, though inferential commitments are a source of an explanation of or reason for action. Millgram also has claimed that, besides pleasure, there are other indicators of desirability. We may infer from our own experience of activities that have been pleasurable to other experiences we may not have had the pleasure of experiencing yet. Millgram has identified that these kinds of inferences are not experiential. He has written:

14

Millgram is careful to distinguish his observational account from hedonism. Pleasure is a way of perceiving reasons for action, not acquiring them (Millgram 1993, 394-415, Cf. Korsgaard 1996, 145-155). E.g., Practical reasoning tends to take one from a position of lesser pleasure to a position of greater pleasure. ... Some philosophers have noticed this tendency, and concluded that pleasure is one’s sole and necessary goal. In this they could not be more mistaken. Hedonists err in roughly the way that someone who thinks that the goal of enquiry is to maximize conviction might err. Normally, one’s enquiries tend to take one from a position of lesser conviction to a position of greater conviction. ... However, ... one’s goal is not conviction: one’s goal is truth. Conviction is epistemically important as a guide to truth, but conviction per se is not the object of my efforts. (Millgram 1993, 397-401) Since I am not opposing his view by arguing that the two are not nearly as distinct as Millgram might believe, I set aside the worry that his view is nothing other than hedonism.

-!17-

Other indicators may be inferences, or rock-bottom judgments of desirability that, like hunches, are not experiential. It is therefore not the case that taking something to be desirable entails expecting it to be pleasurable, any more than believing something entails expecting an encounter with the subject of your belief to produce conviction. (Millgram 1997, 117) Our believing some activity will be desirable inferred from past pleasurable experiences might turn out not to be pleasurable at all. Just because we may believe that some pursuit will be desirable does not mean that we expect such an endeavor will be pleasurable. Pleasure and its inferential correlates form judgments of desirability for Millgram’s observational account. From here, Millgram has to distinguish between his view and the view that has it we pursue pleasure for its own sake because one might contend that nothing other than theoretical induction is at play in practical inductions. Arguing in this way would equate Millgram’s observational account with what has been called “instrumentalism.” But Millgram has argued that the two are distinct. He has shown how we are not to think about practical inductions; I re-create his argument here (Millgram 1997, 118f). Remember Alice. If she pursues only those endeavors that she supposes will produce the most amount of pleasure (pursuing pleasure for pleasure’s sake), Alice might be tempted to try her hand at making a soufflé. Perhaps Alice has had quite a few soufflés in the past. They have all been delicious. So, the amount of pleasure a soufflé would produce for her is exponentially greater than the amount of pleasure an omelet would produce. But Alice fails to anticipate the difficulty making deliciously poofy soufflés. So, the pleasure (for pleasure’s sake) Alice seeks seems ill conceived. Millgram has written: “It is evidence of the strategy’s incoherence that knowingly putting oneself in the way of acquiring convictions in this manner will impede one’s ability to acquire them” (Millgram 1997, 119). Experiential judgments of desirability of objects, rather than pleasure (itself), dictate what we desire and what we find pleasurable. The question I want to pose for Millgram’s practical analog of experiential rock-bottom beliefs is whether experiential judgments of desirability are rock-bottom. I will argue that, like theoretical belief, some antecedent condition governs our feelings of desire. If this is correct, then experiential judgments of desirability are not rock-bottom. Pleasure, according to Millgram’s account, is an intentional state directed toward a feature of oneself or the world and is restricted to experiential judgments of desirability. If we derive pleasure from experience, then it is nothing other than the experience itself that is pleasurable. If experiences are pleasurable, then based upon our experiential judgment we should be able to clearly distinguish between different instances of some pleasure. The pleasure Bob experiences in watching a particularly colorful sunset should be clearly distinguishable from the pleasure he experiences in consuming an In-’n-Out double-double burger. The problem is that pleasures, though for an agent they may be more or less intense, are not as distinct from one another as at first it might seem. Failing to distinguish between experiential judgments of desirability implies that some pleasures are not derived from experience. Our experiential judgments of desirability at least partly depend upon a more deeply held backward-directed inferential conviction that whatever my sensations find pleasurable is desirable. This is a non-experiential background

-!18-

principle distally present in experience and one required for us to possess an judgment of desirability. What I am not suggesting when I argue that some antecedent conditions must be present for us to have rock-bottom experiential beliefs is that we pursue pleasure for pleasure’s sake. Like Millgram, I believe that accounting for our observations in determining what is desirable and what is not desirable is an integral part of practical evaluation. There is nothing in my argument that suggests the non-experiential principle distally present in experience is the target or focus of my desire. Our goal is not the pleasure we achieve through myriad pleasurable experiences. The underlying principle is accessible when we consider responses of Anscombe’s ‘Why?’question. Asking a person, “why did you do that?” or “what do you want to do that for?” should not go on infinitely. It is, according to Anscombe, reasonable to expect that explanations of one’s actions will terminate. Millgram has proposed that his account terminates with ‘It’s pleasant.’ He has argued: On the observationalist account, however, while ‘It’s pleasant’ does indeed terminate explanation, it does so in much the way ‘That’s just how it looks to me’ terminates explanations in the theoretical realm. ‘It’s pleasant’ more or less amounts to: ‘In experiencing it, I find it desirable’. One is not adducing a further goal, but affirming that the goal one has just mentioned is desirable. (Millgram 1997, 120f) According to Millgram’s observational account, bottoming out in the ‘It’s pleasant’ response to Anscombe’s ‘Why?’-question is the practical analog of the “a-ha” moment of experience bottoming out theoretical beliefs. The practical analog, like its theoretical kin, depends upon our experience of something desirable for us to judge it desirable. The explanation of action terminates with the judgement that the experience is desirable, i.e., “pleasant.” If we look at each response to Anscombe’s “Why?”-question, what we discover is that at each interval we should assume that the agent finds what he is doing pleasant. The judgment of desirability accompanies each response to Anscombe’s “Why?”-question. Suppose that we ask Alice, “why are you checking the refigerator?” She might respond, “To find eggs.” Of course, there is a close-by possible world where a more sarcastic resides, call her “Alice*”. Alice* responds, “because going to a farm several hundred miles from my home would be far too tedious and take too long for me to do.” Contrasting the response of Alice with Alice* we see that underlying Alice’s response is a presumption that going to the refrigerator is far more pleasant for her than traveling a great distance to a farm for the eggs she needs for her breakfast omelet. When we continue to ask Alice for an explanation of her action and it bottoms out in ‘It’s pleasant’, what she is really attesting to is not that she is seeking what she judges to be desirable but that each time we ask her why she is doing what she is doing the answer could have been ‘It’s pleasant’. Alice’s aim is not pleasure or the satisfaction of her desires; rather, she finds the object of her experience the most pleasurable way of navigating the world. Each of the responses she offers to Anscombe’s “Why?”-question includes the assumption that whatever the response is will be pleasant for me to do. The so-called final terminating response, ‘It’s pleasant’, is employed by

-!19-

agents to cut-off the questioning and alert the inquisitor of the background principle guiding the responses in the series of questions. What I have argued in this section is that underlying Millgram’s account of pleasure is a principle of conviction distally present in experience. Like with theoretical reasoning and with what I argued in response to rock-bottom experiential beliefs, pleasure’s experience cannot be the practical analog of rock-bottom beliefs. 6. Conclusion We have seen that if theoretical reasoning and practical reasoning are analogous in the way that Millgram has recommended and if we can explain that experiential beliefs depend on distally present antecedent conditions for us to accept them, then there is reason to deny that the experience is rock-bottom practical for theoretical reasoning and pleasure is rock-bottom for practical reasoning. This paper has called into question some of the main characteristics of Millgram’s observational account. If this assessment is correct, then we ought to reconsider the fundamental elements of practical empiricism before we wholeheartedly embrace it.

-!20-

References Anscombe, G.E.M. 1957. Intention. Cambridge, MA: Harvard University Press. Baxter, D. 1988. Many-one identity. Philosophical Papers 17: 193-216. Bonjour, L. 1997. In Defense of Pure Reason. Cambridge: Cambridge University Press. Brandom, R. 1994. Making It Explicit. Cambridge, MA: Harvard University Press. Brandom, R. 2000. Articulating Reasons. Cambridge, MA: Harvard University Press. Davidson, D. 2004. Turing’s Test reprinted in Donald Davidson, Problems of Rationality. Oxford: Clarendon Press. Geach, P. 1976. Reason and Argument. Berkeley, CA: University of California Press. Goodman, N. 1972. Seven strictures on similarity. In Problems and Projects, pp. 437-447. Indianapolis, IN: Hackett Publishing. Kelly, T. 2003. Epistemic rationality as instrumental rationality: a critique. Philosophy and Phenomenological Research 66: 612-640. Korsgaard, C. 1996. The Sources of Normativity. Cambridge: Cambridge University Press. Lewis, D. 1991. Parts of Classes. London: Wiley-Blackwell. Markovits, J. 2011. Why be an internalist about practical reasons? In Oxford Studies in Metaethics, volume 6, edited by R. Shafer-Landau. Oxford: Oxford University Press. Millgram, E. 1993. Pleasure in practical reasoning. The Monist 76.3: 394-415. Millgram, E. 1995. Was Hume a Humean? Hume Studies 21.1: 75-93. Millgram, E. 1996. Williams’ argument against external reasons. Nous 30: 197-220. Millgram, E. 1997a. Practical Induction. Cambridge, MA: Harvard University Press. Millgram, E. 1997b. Incommensurability and practical reasoning. In Ruth Chang (ed.). Incommensurability, Incomparability, and Practical Reason, pp. 151-169. Cambridge, MA: Harvard University Press. Millgram, E. 1997c. Varieties of practical reasoning. In Georg Meggle (ed.). Analyomen 2: Proceedings of the 2nd conference ‘Perspectives in analytical philosophy’, pp. 280-294. Berlin: deGruyter. Millgram, E. 2001. Introduction. In Elijah Millgram (ed.). Varieties of Practical Reasoning, pp. 1-26. Cambridge, MA: MIT Press. Millgram, E. 2005. The Method of Practical Reasoning. In Elijah Millgram. Ethics Done Right: Practical Reasoning as a Foundation for Moral Theory, pp. 1-32. Cambridge, UK: Cambridge University Press. Millgram, E. 2009. Hard Truths. London: Wiley-Blackwell. Sider, T. 2007. Parthood. Philosophical Review 116: 51-91. Stampe, D. 1987. The authority of desire. Philosophical Review 96: 335-381. Thompson, M. 2008. Life and Action. Cambridge, MA: Harvard University Press. Velleman, J.D. 2000. On the aim of belief. In The Possibility of Practical Reason. Oxford, UK: Oxford University Press. Vogler, C. 2002. Reasonably Vicious. Cambridge, MA: Harvard University Press. Wallace, M. 2011a. Composition as identity: part I. Philosophy Compass 6.11: 804-816. Wallace, M. 2011b. Composition as identity: part II. Philosophy Compass 6.11: 817-827.

-!21-

Wallace. R.J. 2014. Practical reason. In Edward N. Zalta (ed.). The Stanford Encyclopedia of Philosophy (Summer 2014 Edition). URL= .

-!22-