Columbia/NYU Philosophy Graduate Conference

Call for Papers: 2011 COLUMBIA/NYU Graduate Conference in Philosophy to be held Saturday APRIL 9th, 2011 at NYU

Keynote Speaker: Stephen Yablo (MIT)

The graduate students and faculty of the Philosophy Departments of Columbia and New York Universities invite papers by all graduate students in any area of philosophy.

Deadline: January 20th, 2011

Papers must meet the following requirements:

1. All papers must be between 3,000 and 5,000 words in length and suitable for a presentation of 30-40 minutes.

2. Papers must be submitted with an abstract no longer than 300 words.

3. Papers must be submitted electronically in blind-review format to:

www.philcolumbia.com/gradconf


PLEASE note that NO submissions by mail or email will be accepted. For more information please visit our website at www.philcolumbia.com/gradconf or email us at: gradconf@philcolumbia.com

Professor John Searle's 50 Years at UC Berkeley - A Celebration

Anti-Humeanism, Anti-Emotivism, and the Desire-as-Belief Thesis

Let us refer to the claim that the good is the object of desire in a sense analogous to how the true is the object of belief as the Broadly Anscombean View. One way of unpacking the Broadly Anscombean view is in terms of the Desire-as-Belief Thesis. In his paper, “Belief, Desire, and Revision”, John Collins defines the Desire-as-Belief Thesis as “the thesis that desire is a particular kind of belief—that to desire A is simply to believe that A would be good."

We may distinguish between weak and strong versions of the Desire-as-Belief Thesis; the claim that some desires are beliefs and the claim that all desires are beliefs, respectively. But this distinction, while significant in its own right, will have little bearing on the present discussion. This is because we are only concerned with the Desire-as-Belief Thesis as a means of unpacking the Broadly Anscombean View; namely, the claim that the good is the object of desire. When we unpack the Broadly Anscombean View in terms of the Desire-as-Belief Thesis, we arrive at the claim that desiring to φ is identical to the belief that φ is good. Now, according to the weak version of the Desire-as-Belief Thesis, only some desires to φ are identical to the belief that φ is good. However, on those occasions in which the desire to φ is not identical to the belief that φ is good, the desire in question does not have the good as its object. Therefore, the Broadly Anscombean View does not apply to such desires. Hence, insofar as we are interested in the Desire-as-Belief Thesis as a means of unpacking the Broadly Anscombean View, we are only concerned with those cases in which a desire may be said to be identical to a belief. Consequently, the distinction between the weak and strong versions of the Desire-as-Belief Thesis is superflous for our present purposes.

In his paper, “Defending Desire-as-Belief”, Huw Price highlights two motivations for the desire-as-belief thesis; a rejection of the Humean theory of desire and a defence of anti-emotivism in moral discourse. He writes:

In modern terminology the Humean view is thus that action is a joint product of an agent's beliefs and desires; and that these are distinct kinds of mental states, desires being distinguished from beliefs in virtue of their motivational role .... In recent years, however, several philosophers have questioned the Humean orthodoxy. They suggest that certain beliefs might be intrisically motivational—in effect, in other words, that some or all desires might themselves be beliefs .... An attractive feature of the suggestion that (some) desires might be beliefs has been its evident potential to undermine emotivist accounts of moral discourse (and so to enable ethical statements to be brought within the scope of a truth-conditional general semantics).

According to Price, the Humean view amounts to the thesis that desires are intrinsically motivational while beliefs are not. Moreover, Price construes emotivist accounts of moral discourse as entailing the denial of the claim that ethical statements display a truth-conditional semantics. I will refer to the rejection of the Humean view as Anti-Humeanism, and the rejection of the emotivism as Anti-Emotivism.

Although, in the passage just cited, Price associates the Desire-as-Belief Thesis with Anti-Humeanism and Anti-Emotivism, it is worth emphasising that the denial of the Desire-as-Belief Thesis is consistent with both. It is common ground between the Humean and the Anti-Humean that desires are intrinsically motivating. However, the Humean denies, and the Anti-Humean affirms, that some beliefs are also intrinsically motivating. But one may consitently subscribe to the claim that some beliefs are intrinsically motivating and yet deny that desires are beliefs. For example, suppose one held that being intrinsically motivational is a necessary but insufficient condition for a psychological state to be a desire. Then one may also consistently hold that desires are distinct psychological states from beliefs (thereby denying the Desire-as-Belief Thesis) and that some beliefs are intrinsically motivational (thereby affirming non-Humeanism).

Moreover, if we identify the content of moral discourse with the content of an intrinsically motivating belief (rather than with the content of a desire), then we may consistently hold that moral discourse displays a truth-conditional semantics and that the content of desire does not. One only needs to add the further assumption that the content of an intrinsically motivating belief displays a truth-conditional semantics. Hence, if we reject the Desire-as-Belief Thesis, it does not follow from this rejection that either Anti-Humeanism or Anti-Emotivism is false.

The preceding observations draw attention to an important feature of Price's account. Price associates the claim that “certain beliefs might be intrinsically motivational” with the claim that “some or all desires might themesleves be beliefs”; either equating the two claims or taking the second to be entailed by the first. However, the equivalence or entailment only holds if we assume that being motivational is a sufficient condition for a psychological state to be a desire. In brief, Price assumes that only desires are intrinsically motivational. However, it seems like both the Humean and the Anti-Humean alike is well within her rights in rejecting this assumption. For example, both the Humean and the Anti-Humean may hold that certain emotions, like love and hatred, are intrinsically motivational. However, it is not immediately clear, nor is it a fundamental assumption of either Humeanism or Anti-Humeanism, that love and hate are desires. In fact, both the Humean and Anti-Humean is free to deny that emotions, like love or hate, are (typically) propositional attitudes.

While I do not wish to either endorse or impugn such a position, I do wish to stress that it is not inconsistent with either Humeanism or Anti-Humeanism. The upshot is that both the Humean and Anti-Humean may reject the claim that only desires are intrinsically motivating. If this suggestion is right, then it is best to see Humeanism as being committed, not to the thesis that desires are the only psychological states that are intrinsically motivating, but rather to the claim that beliefs are not among those psychological states that are. The Anti-Humean, by contrast, wants to insist that at least some beliefs (perhaps, beliefs about the good or about what one should do) are included among the set of psychological states that are intrinsically motivationg. But once it is acknowledge that the set of intrinsically motivating psychological states is not limited to desires, then (as far as Anti-Humeanism is concerned) there is no need to insist that desires are beliefs.

Timothy Williamson Interview: The Armchair

Timothy Williamson on "The Armchair" from Pablo Carnino on Vimeo.

Propositional Content vs. Representational Content

In my last two posts, "Introducing Felicity Conditions" and "Anscombe and Felicity Conditions", I have tried to unpack a particular conception of the correctness conditions of the attitudinal component of the propositional attitude (in contradistinction to the correctness conditions of propositional content of a propositional attitude). Moreover, I hold that the correctness conditions of the attitudinal component of a propositional attitude (or what I refer to as "felicity-conditions" is determined by the attitude's representational component. When this claim is conjoined with the claim that the felicity-conditions of a propositional attitude may differ from the correctness conditions of its propositional content, the upshot is that the representational content of a propositional attitude may also differ from its propositional content. In this post, I want to say why I think this claim should be accepted.

According to the account of desires presently on offer—one that takes its inspiration from Anscombe—desiring that I will leave work at 2pm is felicitous just in case my leaving work at 2pm is an instance of the good. Hence, the felicity-conditions of desiring that I will leave work at 2pm coincide with the felicity-conditions of believing that my leaving work at 2pm is good. Recall, the belief that my leaving work at 2pm is good is felicitous just in case it is true that my leaving work at 2pm is good. This follows from the fact that the felicity conditions of a belief correspond with the truth-conditions of its propositional content. Similarly, my desire that I will leave work at 2pm is felicitous just in case it is true that my leaving work at 2pm is good. This observation makes it tempting to revise the propositional content of the desire to leave work at 2pm so that it coincides with the propositional content of believing that my leaving work at 2pm is good. Hence, we may be tempted to say that the propositional content of desiring that I will leave work at 2pm is not the proposition, “I will leave work at 2pm” but rather, the proposition, “my leaving work at 2pm is good”.

Unfortunately, it is untenable that the propositional content of desiring that I will leave work at 2pm is the proposition “my leaving work at 2pm is good”. Revising the propositional content of the desire in this way simply changes the object of the desire in question. To see that this is so, we merely have to register that desiring that I will leave work at 2pm is different from desiring that my leaving work at 2pm is good. Hence, we need to leave room in our theorising for desires with either propositional content. There are at least three reasons why this is so.

First, the satisfaction-conditions of desiring that I will leave work at 2pm is identical to the truth-conditions of the proposition, “I will leave work at 2pm”, while the satisfaction-conditions of desiring that my leaving work at 2pm is good is identical to the truth-conditions of the proposition, “my leaving work at 2pm is good”. The fact that the desires in question have different satisfaction-conditions suggests that they are in fact distinct desires.

Second, the first desire is the type that typically yields an intention while the second is not. For example, desiring that I will leave work at 2pm may yield the intention to leave work at 2pm. However, it is not typically the case that I can intend that my leaving work at 2pm be a realisation of the good. Moreover, even if I could intend that my leaving work at 2pm be a realisation of the good, it is entirely conceivable that I may successfully carry out that intention without actually bringing it about that I leave work at 2pm. For example, I may intend that my leaving work at 2pm be a realisation of the good by completing some work related project three hours early. But I may complete my project early, and thereby make my leaving work at 2pm a realisation of the good, without actually leaving work at 2pm. Thus, if I did form the intention that my leaving work at 2pm be good, it would still be a different intention to the one I would form if I intended to leave work at 2pm.

Third, desiring that my leaving work at 2pm is good requires that I possess and deploy the concept “good”, while desiring that I will leave work at 2pm does not. Hence, it is important that we preserve the distinction between a desire with the propositional content, “I will leave work at 2pm”, and a desire with the propositional content, “my leaving work at 2pm is good”.

The preceding considerations impugn the claim that desiring that I will leave work at 2pm has the same propositional content as believing that my leaving work at 2pm is good. However, there may still be a way to salvage the idea that the desire that I will leave work at 2pm and the belief that my leaving work at 2pm is good share something in common, content-wise. Towards this end, we may distinguish between the propositional content and representational content of a propositional attitude. Desiring that I will leave work at 2pm has the same propositional content as believing that I will leave work at 2pm; namely, the proposition, “I will leave work at 2pm”. However, while believing that I will leave work at 2pm represents the proposition, “I will leave work at 2pm”, to be true, desiring that I will leave work at 2pm does not represent the proposition, “I will leave work at 2pm” to be true. Rather, desiring that I will leave work at 2pm represents the truth of the proposition, “I will leave work at 2pm”, to be a realisation of the good. Given this difference, it seems natural to say that believing that I will leave work at 2pm has a different representational content to desiring that I will leave work at 2pm. After all, how could two propositional attitudes have the same representational content if one represents something to be true and the other represents something to be good. The upshot is that two propositional attitudes that share the same propositional content may nevertheless exhibit different representational contents.

Language and Thought

Anscombe and Felicity Conditions

In my previous post, I introduced the notion of felicity-conditions; defined as the correctness-conditions of the attitudinal component of a propositional attitude (in contradistinction to the correctness-conditions of the propositional content of a propositional attitude). In the present post, I want say a little about how the notion of a felicity-condition relates to Anscombe's claim that the good is the object of wanting.

Let us refer to the claim that desires do not have felicity-conditions that are identical to the truth-conditions of their propositional content as the negative thesis. The negative thesis is neutral on the question of whether or not desires have felicity-conditions at all; to wit, it takes no stand on whether or not the attitude of desire has correctness-conditions beyond the correctness-conditions of its propositional content (i.e., the truth-conditions of the propositional attitude). The primary motivation for the negative thesis comes from our ordinary linguistic practice; namely, the fact that we do not ordinarily conceive of desires as true or false. We may unpack this intuition by registering that while beliefs represent a certain proposition as true, desires do not. For example, the belief that I will have a slice of cheesecake represents the proposition “I will have a slice of cheesecake” to be true. However, the desire that I will have a slice of cheesecake does not represent the proposition “I will have a slice of cheesecake” to be true. If it did, then it would make sense to say that the desire that I will have a slice of cheesecake is true when the proposition “I will have a slice of cheesecake” is true, and false when the proposition is false. However, as we already noted, we do not ordinarily speak or think this way.

In addition to the negative thesis, I wish to argue that, like the attitude of belief, the attitude of desire has felicity-conditions. However, while the felicity-conditions of a belief are identical with its truth-conditions, I hold that the same is not true of the felicity-conditions of desire. Let us refer to this claim as the positive thesis. Something along the lines of the positive thesis must be accepted if we wish to buy into Anscombe's theory of desires. Recall, according to Anscombe, desire (or wanting) stands in a roughly analogous relation to the good as belief (or judgement) does to the true. However, the negative thesis—namely, the claim that the attitude of desire lacks a truth-value—offers little support for this claim. To wit, the mere fact that the attitude of desire lacks a truth-value does not show that it stands in a certain relation to the good. Moreover, preserving Anscombe's thesis requires that we see the attitude of desire as having something along the lines of what I have been calling 'felicity-conditions'. On this score, the following passage from Anscombe is instructive:

The conceptual connexion between 'wanting'. . . and 'good' can be compared to the conceptual connexion between 'judgment' and 'truth'. Truth is the object of judgement, and good the object of wanting; it does not follow from this either that everything judged must be true, or that everything wanted must be good.

If the fact that the good is the object of desire does not entail that everything desired must be good, then it follows that it is at least possible that a particular desire may get things wrong. Moreover, it seems safe to assume that Anscombe is also committed to saying that it is possible for desire to sometimes get things right. Hence, by Anscombe's lights, desires may be described as two-valued, such that there are cases in which a desire may be said to get things wrong and cases in which a desire may be said to get things right. This suggests that desires have some kind of correctness-conditions, which determine when a desire can be said to get things right. But since the cases in which a desire gets things wrong or right do not correspond with those cases in which the propositional content of a desire is true or false, the correctness-conditions of the attitude of desire (whatever they happen to be) cannot be identical to the truth-conditions of its propositional content.

One very natural way of understanding the claim that the good is the object of desire is to say that a desire is felicitous just in case its object is an instance of the good. Rephrased in the language of propositional attitudes, we may say that a desire is felicitous just in case the truth of its propositional content is a realisation of the good. For example, desiring that I will leave work at 2pm is felicitous just in case the truth of the proposition, “I will leave work at 2pm”, is a realisation of the good. However, some caution is required here. The present claim is not that desiring that I will leave work at 2pm is felicitous only if I do leave work at 2pm; to wit, it is not necessary that the propositional content of the desire be true in order for the desire to be felicitous. Rather, it is only necessary that, were the propositional content true, it would be a realisation of the good. This comports with Anscombe's observation that “goodness is ascribed to wanting in virtue of the goodness (not the actualisation) of what is wanted.”

In other words, the felicity-conditions of a desire should not be confused with its satisfaction-conditions. The satisfaction-conditions of desiring that I will leave work at 2pm just are the conditions under which the proposition, “I will leave work at 2pm” is true. However, if my leaving work at 2pm is not an instance of the good—because, let us suppose, it would involve my leaving an important project uncompleted—then desiring that I will leave work at 2pm is not felicitous even if I do leave work at 2pm. Thus, desiring that I leave work at 2pm may be infelicitous even if the satisfaction-conditions of the desire have been met. Moreover, desiring that I will leave work at 2pm may be felicitous even if the satisfaction-conditions of the desire have not been met; to wit, even if the proposition “I will leave work at 2pm” is false. For example, if my leaving work at 2pm is a realisation of the good—because, let us suppose, it would allow me to spend some quality time with my family—then desiring that I leave work at 2pm is felicitous even if I do not leave work at 2pm. The upshot is that the felicity-conditions of desiring that I will leave work at 2pm, when construed along Anscombean lines, are different from the satisfaction-conditions of desiring that I will leave work at 2pm.

Let us take stock of what we have seen thus far. First, we noted that although the propositional content of desire has truth-conditions, just like the propositional content of belief, the attitude of desire is ordinarily assumed to lack a truth-value. Using the label 'felicity-conditions' to refer to the correctness-conditions of an attitude (in contradistinction from the correctness-conditions of the propositional content of an attitude), we may say that the felicity-conditions of a desire, assuming that it has felicity-conditions, do not correspond with the truth-conditions of its propositional content. Moreover, I have argued that preserving the Anscombean thesis that the good is the object of desire entails that desires have some kind of correctness-conditions.

Introducing: Felicity Conditions

Beliefs, and desires are widely regarded as examples of propositional attitudes; relational mental states connecting a person to a proposition. Examples of propositions include: “It is raining outside”, and “I will leave work at 2pm”. Examples of propositional attitudes include: believing that it is raining outside, and desiring that I will leave work at 2pm. As their name suggests, propositional attitudes are typically conceived of as being made up of two components: a proposition, which may either be true or false, and an attitude or mode of entertaining that proposition. I have reservations about the claim that beliefs and desires always involve relations between persons and propositions. I hold that on at least some occasions, they may involve relations between persons and actual objects or states of affairs. However, I will not argue for that conclusion here. Instead, I will simply adopt the standard philosophical practice of describing beliefs and desires as propositional attitudes.

Let us say that the proposition, “I will leave work at 2pm” is true just in case I do leave work at 2pm, and false otherwise. According to the standard view, I may both believe that I will leave work at 2pm and desire that I will leave work at 2pm. Thus, both beliefs and desires are attitudes that may be taken towards a proposition—i.e., a truth-value bearing item. However, while the belief that I will leave work at 2pm is ordinarily deemed true when the proposition, “I will leave work at 2pm”, is true, and ordinarily deemed false when the aforementioned proposition is false; the desire that I will leave work at 2pm is not ordinarily deemed true when the proposition “I will leave work at 2pm” is true, nor is it ordinarily deemed false when the aforementioned proposition is false. In fact, the categories of truth and falsity are not ordinarily taken to have application to desires.

If the preceding observation regarding desires is right, then it does not follow from the fact that the content of a propositional attitude (which we are, for the purposes of the present blog post, assuming to be a proposition) has a truth-value that the attitude itself has a truth-value. We may accommodate this claim by distinguishing between the correctness-conditions of the propositional content of a propositional attitude and the correctness-conditions of the attitude itself. Hence, we may say that while the propositional content of a desire has truth-conditions, the desire itself does not. In this regard, desires differ from beliefs, since both the propositional content of a belief and the belief itself may be true or false.

In order to avoid confusion, it would be helpful to have different terms to refer to the correctness-conditions of the content of a propositional attitude and the correctness-conditions of the attitude itself. Since we are assuming that the content of all propositional attitudes are propositions, and since all propositions have truth-conditions, let us refer to the correctness-conditions of the propositional content of a propositional attitude as the truth-conditions of the propositional attitude. Moreover, let us refer to the correctness-conditions of the attitude itself as the felicity-conditions of the propositional attitude. The distinction between truth-conditions and felicity-conditions will provide us with the theoretical machinery we need to characterise the aforementioned contrast between beliefs and desires; namely, that although both beliefs and desires have truth-value bearing propositional content, only the former is ordinarily conceived of as being either true or false. We may say that in the case of beliefs, the felicity-conditions of the attitude are identical to the truth-conditions of the attitude's propositional content. Thus, not only is the propositional content of the belief true or false, but so is the belief itself.

It is not part of our ordinary linguistic practice to talk about the propositional content of a propositional attitude. The notion of propositional content is a theoretical one. Hence, it seems plausible that when we ordinarily talk about beliefs being true or false, we have their felicity-conditions (i.e., the correctness-conditions of the attitude) in mind. This assumption would also go some way towards explaining why we do not ordinarily conceive of desires as being true or false. While the propositional content of a desire has a truth-value, the desire itself does not. Thus, the felicity-conditions of a desire—if it has felicity-conditions at all—are not identical to the truth-conditions of its propositional content. Since our ordinary intuitions about which propositional attitudes have a truth-value track the propositional attitude's felicity-conditions (rather than the truth-conditions of its propositional content), the fact that the felicity-conditions of desire are not identical to the truth-conditions of its propositional content would explain why we do not ordinarily conceive of desires as true or false.

To recap, we have distinguished between the truth-conditions of a propositional attitude and the felicity-conditions of a propositional attitude. The former, we have identified with the correctness-conditions of the propositional content of a propositional attitude and the latter we have identified with the correctness-conditions of the attitude itself. Moreover, we noted that our ordinary intuitions about which propositional attitudes are true or false seem to correspond with the felicity-conditions of a propositional attitude, rather than with the truth-conditions of its propositional content.

Empirical Psychology on Free Will

Theoretical and Practical Transitions

Let us refer to all belief-yielding psychological transitions (e.g., the transition from a perceptual experience to a belief) as theoretical transitions, and all intention-yielding psychological transitions (e.g., the transition from a desire to an intention) as practical transitions. A transition qualifies as psychological, on the present view, only if it has representational content. The distinction between theoretical and practical transitions hinges on the recognition of two distinct kinds of representational content. The first is associated with psychological states that represent a certain state of affairs as being the case and are legitimately deemed faulty if the state of affairs in question is not the case. The second is associated with psychological states that do not represent a certain state of affairs as being the case and which are therefore not deemed faulty if the state of affairs in question is not the case.

I maintain that a psychological state is good just in case it is not legitimately deemed faulty, and bad otherwise. The conditions under which a specific psychological state is legitimately deemed good or bad constitutes that state’s correctness conditions. Thus, I hold that goodness and badness (i.e., goodness conditions) are the primary and most basic form of correctness conditions for a psychological state. Moreover, I hold that truth is simply one instance of the good, relative to psychological states that represent a certain state of affairs as being the case (i.e., those that fall under the umbrella of theoretical transitions). A psychological state represents truly just in case it represents a particular state of affairs as being the case and the state of affairs in question really is the case.

Because of the important role played by the notion of truth in our ordinary discourse, we often find it necessary to speak about truth in ways that set it apart from other forms of goodness. Truth is primus inter pares. In this regard, the concept of truth is analogous to the concept of the human; for while human beings are considered a type of animal, the important role played by the notion of the human in our ordinary discourse often requires that we speak about humans in ways that set them apart from all other animals. Hence, just as we may juxtapose humans and animals, although strictly speaking, humans are a type of animal), so too we may juxtapose the true and the good (although, strictly speaking, truth is a type of goodness).

The juxtaposition of humans and animals is legitimated by the fact that humans have a special characteristic (i.e., the ability to respond to reasons as such) that all other animals lack, and this characteristic plays a special role in our ordinary discourse. Similarly, the juxtaposition of truth and goodness is legitimated by the fact that truth has a special characteristic (i.e., its identification with states of affairs that are the case) that all other forms of goodness lack, and this characteristic plays an important role in our ordinary discourse. Hence, we may refer to the correctness conditions of a psychological state that represents a particular state of affairs as being the case as truth conditions, and I will describe all other psychological states as having goodness conditions.

It is often observed that the specification of the content of a perceptual experience or perceptual belief is worded in terms of a that-clause. One may, for instance, see that there is an apple on the table, or believe that there is an apple on the table. By contrast, the specification of the content of a desire or intention typically takes the form of a to-clause. One may desire to eat an apple or intend to eat an apple. However, while such grammatical details are suggestive, they are far from conclusive. For example, Myles Brand observes that the content of an intention may often be specified in terms of a that-clause; a claim he illustrates with the example: “Richard intends that he vote in the next election.”(Italics mine)

I believe that a more philosophically interesting way of unpacking the difference between theoretical and practical transitions is to note that an ideal rational agent that holds a belief (i.e., a psychological state resulting from a theoretical transition) to the effect that S obtains, when confronted with conclusive evidence that S does not obtain, will revise her belief, whereas an ideal rational agent with an intention (i.e., a psychological state resulting from a practical transition) to bring S about, may retain this intention even in the face of conclusive evidence that S does not obtain.

One upshot of the claim that theoretical and practical transitions have different kinds of content is that the content of a theoretical transition could not feature in a practical transition, and vice versa. Dennis Stampe commits himself to such a view when he notes that it is impossible to transition from the propositional content of a belief to the content of an intention. He writes:

Starting from It would be good if p, and perhaps Only my doing A will make it the case that p, by what logic do we pass to I will do A? All that seems to follow is that It would be good to do A—and this neither denotes an action nor the content of an intention; further premises would yield I ought to do A, but to believe that is not to do A. We confront a logical gap. And it cannot, it seems, be bridged by the addition of further beliefs.

Here, Stampe observes that practical and theoretical transitions are distinct types of transitions with their own respective types of content. Any attempt to reduce the content of an intention (or desire) to that of a belief or judgement is therefore moribund. In the discussion that follows I take Stampe’s observation, along with the definition of practical transitions as intention-yielding, seriously. A transition that is belief-yielding, but which happens to have practical content (i.e., happens to be about some action the agent is performing) does not qualify as a practical transition on this view.

On Types of Explanation: Ben’s Reply to Me

The following is Ben's reply to my post, On Types of Explanation: My Reply to Ben.

Thanks for the reply. Let me clarify what I meant to say in my own post about two points germane to the issue between us.

First, I did not mean to suggest that there cannot be different kinds of explanation, where a difference in kind marks, say, a difference in the sorts of considerations that fill in explanatory gaps (e.g., law-governed particles vs. psychological states). Nor did I mean to rule out differences in kinds of explanation where this marks the fact that there are different standards involved (e.g., due to differences in the background understanding of the recipient of the explanation). This last point is relevant to the issue of whether justification is a kind of explanation, which I will return to below. What I did want to stress is that these differences need not mark different concepts of explanation (e.g., folk-psychological vs. scientific). I did not mean to say, on the one hand, that there is only one type of explanation, and then, on the other, speak of different kinds of explanation. What I wanted to say is that there is one concept and different ways of falling under it.

I think this is all compatible with saying that our scientific practices and folk-psychological practices of explaining phenomena in the world appeal to different ‘storehouses’ of explanans. What I want to insist on is that the activity of explaining is in all cases formally similar (the explanation is offered by one who aims to thereby plug a gap in someone’s understanding). The appeal to different sorts of information does not make different kinds of explanations different in this regard. And neither does a difference in what it takes to plug the gap in understanding. One reason why I think this picture is helpful is that it allows us to explain differences in types of agent (e.g., non-rational vs. rational; reasons-responsive vs. responsive to reasons as such) according to differences in parameters on adequate explanation.

The second point I would like to clarify has to do with justification. By a 'justified agent' you take me to refer to the agent the explanation is about--as you put it, the explanandum. But--and this is crucial here--in some cases this agent may also be the recipient of the explanation. So we may speak of a 'justified agent' in the following two senses, call them third personal and first personal.

From the third personal point of view, we speak of an agent as justified when we can articulate an explanation of her behavior that appeals to her take on the situation and that plugs a gap in our normatively-loaded background understanding—i.e., a background understanding that includes things like norms governing formations of beliefs and intentions. When we speak of a 'justified agent' from the third personal point of view, this does not entail (i) that the agent can recognize reasons as such nor (ii) that the agent has a justifying reason for her behavior. Presumably, the agent has at least a motivating reason (a consideration that that motivates her behavior) and our explanation is in terms of this (e.g., what plugs the gap is the agent’s supposed representation of an approaching poacher). Moreover, as we are the sorts of agents who can receive explanations that fill gaps in a normatively-loaded background understanding--i.e., justifications--the agent's motivating reason may justify her behavior to us. But the agent herself need not have a justifying reason because she need not be the sort of agent who can have a normatively-loaded background understanding.

In the first personal case, a 'justified agent' is both the object and recipient of the explanation of her behavior. So the distinction between the one performing the behavior and the one normatively situating the explanation of this performance collapses. An explanation of one's own behavior that satisfies a gap in a normatively-loaded understanding of the situation entails that the agent (i) can recognize reasons as such and (ii) has a justifying reason for the behavior.

So in the first personal case the object and recipient both are relevant to the standards in terms of which a proper justifying explanation is given. In the third personal case this is not so. And this difference between the two cases explains why the capacity for reflection is necessary for having justifying reasons. The 'justified agent' who first personally explains her behavior to herself by plugging a gap in her normatively-loaded background understanding--i.e., articulates a justifying reason for her behavior--manifests her capacity for reflection. We can say that she has a justifying reason. The 'justified agent' whose behavior is third personally explained does not necessarily manifest a capacity for reflection. And this is why this second agent might not have a justifying reason. Nevertheless, we can speak of the agent as justified because we can articulate a consideration that plugs a gap in our normatively-loaded background understanding when we look at things from the agent’s point of view. And we can assume that agents other than ourselves have justifying reasons because we can speak of them as justified in this third personal way and assume that they, like us, can receive justifying explanations. So my view does not commit me to a suspect solipsism.

Again, as we have discussed in earlier posts, this does not mean that we illicitly import desires or beliefs about the situation when considering the situation from the agent’s point of view. Rather, we take up what we assume to be the agent’s own psychological representations of the situation (though we can be wrong about what psychological states we suppose the agent to have). But given that we are the sorts of creatures who understand the world in terms of norms, we can situate the supposed representations of the agent in a normative context. Perhaps we can also abstract away from the normatively-loaded understanding through which we usually view the world. We might then articulate a mere explanatory reason, or mere motivating reason. I think this is plausible. But the point I want to insist on is that there seems to me to be a clear way in which we can articulate justifying reasons for other agents and not thereby commit ourselves to the claim that those agents have justifying reasons at all. The reason we can do these two things is that we can take up their perspective in the context of our own normatively-loaded understanding. Since the adopted perspective is theirs (including their supposed representation of the situation) it is apt to talk of the reasons being their reasons. But since the normative background is ours, it is at the same time apt to talk of them not having these reasons.

I hope this clarifies some of what I meant. Thinking about your remarks has helped me to get clearer on things in my own mind.

On Types of Explanation: My Reply to Ben

The following is my reply to Ben's post, On Justification and Explanation: Ben's Reply to Me:

In your post, you express some reservations about my distinction between two types of explanation. I suspect that there is no canonical answer to the question: is there a single type of explanation? Usually, whether or not we decide to distinguish between two or more subclasses that fall under a general heading will depend on our goals and the level of specificity necessary for achieving them. You prefer to keep things general; restricting yourself to a single conception of explanation; one that applies unequivocally to all replies to why-questions.

However, I believe the question presently at hand demands a greater level of specificity than that offered by your broad account of explanation. Presumably, even you would concede that there are important differences between the two examples of explanation you mention; namely, an explanation of why water boils and an explanation of why a monkey climbs a tree. The explanation of why water boils only appeals to considerations that display the regularity characteristic of the type of laws described by the physical sciences. However, the explanation of why a monkey climbs a tree includes considerations that do not display the same kind of regularity – namely, psychological processes or events. Our folk psychology may be seen (inter alia) as a storehouse of criteria for the attribution of psychological states; it encodes norms governing psychological processes and events. No such storehouse is necessary for the explanation of water boiling; the special function served by our folk psychology has no bearing on such explanations. This, no doubt, constitutes an important difference between the two explanations; one that I believe is salient to the question of agency and our reason-giving practises.

Now, it seems to me that when you say that justification is a “special sort of explanation”, you thereby concede that there is more than one type of explanation. Consequently, there seems to be some tension between your claim that there is only one type of explanation and your claim that justification is a special kind of explanation. You take as your point of departure the claim that explanation is a matter of filling in gaps in the understanding. But this raises the question: gaps in whose understanding? Initially, your answer is gaps in the understanding of the person to whom the explanation is being given (as opposed to the person the explanation is about - i.e., the explanandum). This suggests one possible diagnosis of why it was tempting for you to conclude that the explanations of a monkey’s behaviour and the behaviour of water molecules are, in all salient respects, the same. After all, in both cases, there is a gap being filled in the understanding of the person to whom the explanation is given. Consequently, if the understanding of the person to whom the explanation is being given is the only salient factor vis-a-vis the kind of explanation given, then it would follow (from the fact that the person to whom an explanation is given is always an agent of a single broad type – namely, a rational agent) that there is only one type of explanation.

However, later in your post you seem to reject (albeit implicitly) the assumption that the person to whom an explanation is given is the only relevant factor vis-a-vis the type of explanation given. For example, you claim that in order to be justified, an agent must have the capacity to recognise reasons as such. Moreover, you take justification to be a type of explanation. But notice, if explanation is simply about filling in gaps in the understanding of the person to whom the explanation is being given, then there would be no need for a justified agent (i.e., the person that a particular explanation is about) to recognise reasons as such. Thus, your claim that justification requires that the justified agent (i.e., the explanandum) be able to recognise reasons as such entails that the person to whom a particular explanation is given is not the only salient factor vis-a-vis the type of explanation given. The upshot is that it does not follow from the fact that the explanation of a monkey’s behaviour and the explanation of the behaviour of boiling water both fill in gaps in someone’s understanding (namely, the person to whom the explanation is given), that the respective explanations are of the same kind.

Once we dispense with the offending assumption – namely, the assumption that the person to whom an explanation is given is the only determinant of the nature of an explanation – we are now in a position to recognise that the object of an explanation may also determine the nature of the explanation given. This brings us back to the point with which I began this post; namely, that since the object of explanation in the boiling water example lacks psychological states and the object of explanation in the monkey example possesses psychological states, the types of explanations at play in the two cases are different. I unpack this claim by noting that the first kind of explanation only appeals to considerations that display the regularity indicative of the laws described by the physical sciences, while the second kind of explanation - i.e., folk psychological explanations - appeals to considerations that fail to display the same kind of regularity - namely, psychological processes and events.

On Justification and Explanation: Ben's Reply to Me

The following is Ben's response to my previous post, On Justification and Explanation: My Reply to Ben.

Thanks for the reply. It clears things up for me a bit. Let me offer the following remarks in response.

I intend to be using the terms 'justification' and 'explanation' with respect to reasons for action in a sense that is nicely captured by Raz in 'Reasons: Explanatory and Normative'. I'll quote from that paper:

[Reasons] are both normative and explanatory. They are normative in as much as they guide decision and action, and form a basis for their evaluation. They are explanatory in that when an action for a purpose occurs the purpose for which it is performed, the reason for the action as the agent sees things, explains its performance.

Raz goes on to argue that 'reasons' has two meanings. I agree, but I put the distinction in terms of explanation and justification instead of explanatory and normative.

By the way, I do not claim that these different meanings pick out different considerations. The same fact may both justify and explain--that is, one consideration can be both the purpose that explains an agent's action from that agent's perspective and it can guide performance of the action and serve as a basis for evaluation of that action (and the agent who guided her behavior on its basis). So the distinction between justifying reasons and explanatory reasons is not a distinction between different sorts of considerations.

I don't quite know what you mean to signal by a distinction between the folk psychological concept of explanation and the concept of explanation used by the sciences. But I do not want to make any sort of distinction like this. On my favorite view (cf., e.g., van Fraassen; Wright; Scriven) explanation is a matter of filling gaps in understanding. So a consideration explains when it answers a why-question. I am taken by the fact that explanations are given. And this presupposes that there is someone to give the explanation to. So what counts as a good explanation can depend on the needs of the one it is given to.

Why did the monkey jump into the tree? Because he saw the poacher coming through the bushes. Or: In order to grab the bananas that were in the tree. Why does water turn to steam when it reaches 100 degrees Celsius? Because the movement of the particles … The form of explanation is the same in both cases. I think this is the correct, general account of explanation.

As for justification, perhaps I am thinking of it as a special sort of explanation. It plugs a special sort of gap in the understanding. And in order to have this sort of gap in one’s understanding, one must have certain capacities—e.g., to recognize reasons as such. So where we have someone with these reflective capacities asking why an agent did something we can not only offer a consideration that explains why the agent did what he did but also why the cited consideration shows the behavior to be appropriately responsive to reasons. The why-question that signals a request for justification is situated in a normative context.

Why did the monkey jump into the tree? Because he saw the poacher coming through the bushes. This can both explain and justify. Given background assumptions about the monkey’s motivation to stay alive, the poacher’s goal of killing monkey, etc., it is intelligible in a special sense why the monkey jumped into the tree. This behavior is appropriate given the facts. What I have been insisting on is that the fact that he saw the poacher coming can explain the monkey’s behavior, both to us and to the monkey. But this fact cannot justify the monkey’s behavior to the monkey because the monkey cannot have the relevant sort of gap in the understanding. His perspective is not relevantly normative in that he does not guide his behavior on the basis of reasons as such. No fact can provide a basis for evaluation of behavior for him of the sort that justification involves. But it can for us because we do guide our behavior on the basis of reasons as such.

This brings me to perspective. When I say that we adopt the monkey’s perspective in judging that the fact that the poacher was approaching justified his jumping into the tree, I do not mean either (i) to deny that the monkey has a perspective of his own (this is what we adopt if we do things right) nor (ii) that we import our desires into the monkey’s perspective (again, insofar as we do things right). I mean that we adopt the monkey’s perspective, as we take it to be, which we can be wrong about, and this involves assuming his motivations, which, again, we can be wrong about. Think of this as the relevant background understanding, the gap in which is filled by the fact that is purported to be a reason. The answer to the question why the monkey jumped into the tree is only explanatory (and so also only justificatory) insofar as it fills a gap in understanding. The fact that the poacher was approaching can do this, on the assumption that the monkey represented this fact to himself and also that the monkey had the relevant motivations (e.g., not to be killed by the poacher, which we are assuming he recognized as a threat). If we are assuming that the monkey sees bananas but no poacher, then the fact that the poacher was approaching cannot explain (and so also cannot justify) his jumping into the tree. So my view allows that we can get it wrong and has something to say about what is wrong when things do go wrong.

Notice that my view can agree with you that a transition (e.g., from the representation of the poacher to an intention to jump into the tree) can be rationally assessable even if the agent making the transition cannot assess it. We seem to disagree about whether the transition is reason-providing for the agent. You say it is in that it provides the monkey with an entitlement. I say it depends on what sort of reason you have in mind. It provides the agent with an explanatory reason. But since the monkey cannot receive justifications, it cannot provide him with a justifying reason. Some of the disagreement may be only apparent, depending on how your view of entitlements vs. justifications lines up with my view of explanation vs. justification.

One last point about to-be-doneness and justification. I think it is correct that if there were no reflective creatures than there would be no justifications. There would be no one to receive the justifications if there were no one with the relevant gaps in the understanding. But this does not mean that the same considerations that do justify given the presence of reflective creatures fall out of the picture entirely. They could still explain, supposing that there were creatures that could be given the relevant sorts of explanation. Barring that, they could still motivate behavior in the sense necessary even for them to explain it. The lion can still recognize that eating his paw is not to be done, even if there is no one capable of posing the question why he did not eat his paw. If this amounts to the consideration making the behavior intelligible, in some sense of ‘intelligible’, then I am good with that. But on the assumption that there are no creatures intelligent enough to be given explanations (and so also justifications) then this consideration does not explain (or justify) anything because there is no gap in understanding to be filled by it.

On Justification and Explanation: My Reply to Ben

What follows is my response to the questions posed by Ben in the comments following his post.

I’m not sure how you’re using the words ‘justification’ or ‘explanation’ here; these are terms of art and philosophers tend to define them in different ways. Moreover, on the way I’m inclined to use the word ‘explanation’, your question as to whether entitlement falls on the justification or explanation side of the divide does not lend itself to a straightforward answer. In fact, given my way of carving things up, the question would be somewhat ill-formed; the distinction between justification and explanation is simply not salient. Instead of talking about explaining a belief or action, I prefer to speak about making a belief or action intelligible. (The 'making intelligible' locution is meant to highlight the difference between the types of explanation implicated in our folk psychology, and that offered by the natural/physical sciences). The difference is not simply terminological. The philosophically interesting respects in which my preferred way of carving things up differs from your own may be seen in the following summary of the view I find most attractive.

When we ask why an agent believes such and such (at least in the sense of ‘why’ I presently have in mind), the question presupposes that the transitions that give rise to the belief are subject to certain norms relating to truth. Analogously, when we ask why an agent performs some action, the question presupposes that the transitions that give rise to the agent’s intentions are subject to certain norms relating to what ought to be done (what I refer to as ‘goodness-conditions’). I refer to both the norms relating to truth and goodness as ‘rational norms’. This way of putting things involves broadening the applicability of reason-talk to include more than the norms relating to truth. Moreover, I hold that a transition may be subject to a rational norm even if the agent engaged in the transition lacks the conceptual resources necessary to understand that norm. The upshot is that even the psychological transitions of animals may be subject to rational norms.

On the present view, a transition’s being subject to a rational norm does not entail or require that the agent performing the transition be able to assess whether or not that transition is rational. (Assessment is, of course, reserved for agents with the required conceptual capacities; namely, rational agents.) Thus, I distinguish between the conditions necessary for performing a transition that is rationally evaluable and the conditions necessary for evaluating a transition’s rationality. The first set of conditions are met by all agents with beliefs and intentions, while the second is met only by those agents that can reflect on (i.e., rationally evaluate) their beliefs or intentions. Moreover, it is only agents capable of evaluating their own beliefs and intentions that may be correctly described as ‘rationally responsible’. To say that an agent is rationally responsible is to say, not only that the agent’s beliefs or intentions, but also the agent herself is rationally evaluable. Thus, I also distinguish between the necessary conditions for an agent’s beliefs and intentions being rationally evaluable and the necessary conditions for an agent being rationally evaluable in the light of her beliefs and intentions. The latter are only met by rational agents.

To say that a transition is rational is, on the present view, to say that it is reason-providing. This reason provides the agent with an entitlement to beliefs or intentions based on that transition. When a suitably equipped agent comes to recognise her reasons as such, her entitlement (eo ipso) becomes a justification. Thus, a justification is what an entitlement becomes when an agent comes to recognise her reasons as reasons. In sum, I hold that the reasons that provide an agent with an entitlement to a certain belief or intention also constitute an agent’s justification just in case the agent recognises those reasons as reasons. The upshot is that the very reasons that constitute an agent’s entitlement may also constitute her justification; this leaves no room for a distinction between the kind of reasons that justify and kind that explain.

The preceding upshot – namely, that the account on offer leaves no room for the distinction between explanatory and justificatory reasons – may seem like a glaring omission. However, I believe that the intuitions that motivate the standard distinction between explanatory and justificatory reasons may be accommodated (without violence) by the distinction between theoretical and practical transitions. Time won’t allow me to fully unpack the view here. But the gist of it is that the content of theoretical transitions roughly corresponds with the kind of reasons that fall on the justificatory side of the traditional divide, and the content of practical transitions roughly corresponds with the kind of reasons that fall on the explanatory side of the traditional divide. Another way the point may be put (though it does not get things quite right) is to say that while theoretical transitions justify, practical transitions explain. This does not get things quite right since, as I noted above, I prefer to say that theoretical and practical transitions are salient to our respective attempts to render an agent’s beliefs and intentions intelligible. Both represent modes of folk-psychological explanation (which, again, must be contrasted with the types of explanations in the natural/physical sciences). In sum, I hold that the concepts of entitlement and justification (whether theoretical or practical in form) are explanatory concepts; they represent a fundamental part of our folk psychological attempts to render beliefs and intentions intelligible.

With regards to your second question, I believe that the notion of entitlement, like all normative notions, can only be applied by an agent equipped with the necessary concepts. But the question that presently concerns us is not who may apply a normative concept, but rather what a normative concept may have application to. I believe a certain state of affairs can be described as being good for a cat (in the most literal sense imaginable), even if the cat lacks the conceptual resources necessary to understand this fact or apply the concept of goodness in its own case or anyone else’s. Moreover, when we say that a certain state of affairs is good for a cat, we do not mean that it would be good for us if we were in the cat’s place. We mean it is good for the cat. For example, if I were in the corner of the room chewing on a freshly killed mouse, this state of affairs would not be good for me. But given the desires and needs of the cat, it may certainly be good for it. In short, when I say that “chewing on a freshly killed mouse” is good for the cat, I mean from the cat’s perspective (given the cat’s needs, beliefs and desires), not my own.

The need to distinguish between our perspective and that of an animal, whose beliefs and intentions we happen to be evaluating, becomes important when one considers that the evaluations in each case may actually be at odds. For example, a die-hard vegan may believe that it would be bad for her (or any other human being) to ingest the flesh of another animal and yet hold that it is good for a lion to do so. When the vegan says it is good for the lion to do so, she is not putting herself (with all her human capacities and desires) in the lion’s place, for if she were to do so, then whatever moral qualms she has about humans consuming animal flesh would continue to apply. Thus, her assessment of the lion cannot simply be a result of her projecting human capacities to the lion. Rather, she reflects on the lion’s needs, beliefs and desires (not her own) and concludes that the act of ingesting animal flesh is good (from the lion’s perspective. Moreover, reflecting on the lion’s perspective presupposes that it has one, and that this perspective belongs to the lion; not to the human being evaluating the lion’s actions.

The problem I have with your claim that talk of a lion’s reasons is simply metaphorical is that it threatens to obscure the above observations; it gives the impression that the lion really does not have a perspective, independent of the perspective of the human being evaluating the lion's beliefs or actions. This is why I emphasised that the human evaluator could get things wrong when she attempted to attribute certain beliefs and desires to an animal. The possibility that the human evaluator may be mistaken entails that there is something (independent of the human evaluator) that she could be mistaken about. Moreover, I maintain that the lion would continue to have a perspective even if there were no humans around to make evaluations based on it. But your view seems to entail that if all human beings died tomorrow, lions would no longer have a perspective from which certain states of affairs could be evaluated as good or bad. Of course, there would be no one around to carry out the evaluations, but that would not mean that ingesting wood chips and animal flesh would suddenly be seen as on a par by lions. Lions would continue to see ingesting wood chips as something not to be done (which, on my account, constitutes a certain state of affairs being represented as bad) and ingesting animal flesh as something to be done (which, on my account, constitutes a certain state of affairs being represented as good). In brief, when I say that a lion has a reason to stalk and kill a gazelle (for example, in order to feed its cubs), I am saying something about the lion, not about myself.

On Reasons and Animals: Ben's Reply to Me

What follows is Ben's response to my previous post:

Thank you for paying such close attention to my comments and for such a thorough and thoughtful reply. Let me mention that I really do not want to deny that animals are agents (that would be absurd) or that we can give explanations of animal behavior using ‘reason’ in (one of) its ordinary meanings. That is, animals are properly and non-metaphorically spoken of as agents and as behaving intelligibly (i.e., in ways that make sense given certain considerations).

That said, I still have a complaint about your conclusion that the monkey’s transition of thought in Hurley’s case (from A is dominant over B and B is dominant over C to A is dominant over C) gives the monkey a reason to believe that A is dominant over C—or, as you take them to be equivalent, justifies the monkey’s belief that A is dominant over C.

You’ve brought out two nice distinctions in your last two posts: first, the distinction between responding to reasons and responding to reasons as such; second, the distinction between reasons-for-which and reasons-with-which. I assume that these are not identical distinctions, but they share something important in common. In both cases, it takes an extra capacity to be the sort of creature to which the latter terminology is applicable. Human beings but not all agents act for reasons as such, and the same is true with respect to reasons-with-which. And this is the case because human beings have the capacity for reflection. I think this is all important and true.

But I also think that it does not show that a transition of thought like the one in the monkey case is reason-giving for the monkey. This is the claim at issue. You say that the transition of thought is reason-giving, but you deny that the monkey acts on the recognition of a reason as such (and so also, I assume, you deny that the monkey acts with a reason). I agree that the monkey does not act on the recognition of a reason as such (or with a reason). But I deny that the transition of thought gives the monkey a reason. The reason that I deny this is that I do not think that the rationality of the transition (i.e., its intelligibility in response to the relevant considerations) gives the monkey anything. From the monkey’s perspective, nothing is more or less justified before or after the transition of thought. Equivalently: nothing is justified for the monkey.

I think there is a similarity between the inanimate object and the monkey. Neither the monkey nor the knife, say, can gain a reason on the basis of a transition of thought. The explanation of why this is so in each case reveals a difference between the monkey and knife. The knife cannot gain a reason based on a transition of thought because the knife cannot have thoughts. The monkey cannot gain a reason based on a transition of thought, I claim, because the monkey cannot have reasons in the relevant sense. The relevant sense is, by your own lights, that the monkey’s belief is justified.

Now I can see two ways of making sense of the claim that the monkey’s belief is justified. First, the belief may be justified for the monkey. But this, I claim, requires that the monkey be able to countenance a reason as such. And we both deny that. Second, the belief may be justified for us when we consider things from the monkey’s perspective. This is the view I tried to advocate for in my comment to your previous post. You quote a sentence from those comments to the effect that we project our deliberative perspective on the monkey when thinking about the intelligibility of the monkey’s behavior. I stand by this claim. And I don’t think that it ignores the distinction between the monkey and the knife. We might project on both, even if the one can have thoughts and the other cannot.

To press you further, I think that there is an important distinction missing from this discussion. We might cite a reason for the monkey’s behavior in the sense of a consideration that explains why the monkey did such-and-such. But this is not yet to say that the consideration justifies anything. A justifying explanation requires that the one to whom the explanation is given has the capacity to take up and evaluate the explanatory considerations. And the monkey cannot do this, I think. We can, I think, because we have the capacity to reflect on possible determining grounds of our behavior and choose between them—we have the capacity for reflection.

My view allows us to still account for the ordinary practice of saying things like the reason for which the monkey leapt into the tree was that there was a predator approaching, where ‘reason’ has justificatory force here. This explains the monkey’s behavior in terms that engage our rational, reflective capacities. When we think about the circumstances from the monkey’s point of view, we take the approaching predator to be a reason to jump into the tree, in the sense of a consideration that justifies doing this. We talk as though the monkey took this as a justifying consideration as well. But this talk is metaphorical. (The same goes for the transition of thought regarding dominance.)

What might not be metaphorical is talk that attributes an explanatory role to the consideration that the predator is coming. I think that is the case. But then this consideration does not give the monkey a reason in the sense you are after. You claim that reason-givingness involves justification. And explanation of the sort we have here does not justify anything to the monkey. The monkey is not a proper recipient of justifying explanations because the monkey cannot take them up and assess them in the right way. The monkey does not have the relevant capacity for reflection. We do have this capacity, and we also have the capacity to take up the monkey’s perspective from within our own deliberative point of view. And when we do this, the same considerations that explain things about the monkey’s behavior also justify this behavior to us. But the move from explanation to justification requires the capacity for reflection. And I want to insist that if the monkey does not have this capacity, then the monkey cannot be given a reason, where this involves justification. I also want to insist that this does not undermine the propriety of our talking about the monkey having a justification so long as we recognize that our language is in this case metaphorical.

The upshot is this: we can talk of the monkey having a reason for doing something in both a metaphorical and non-metaphorical sense. In the first case, it is metaphorical because we take reasons to be justifying and the monkey cannot receive justifications. In the second case it is not metaphorical because we take reasons to be explanatory and the monkey’s behavior is explained by certain considerations engaging the monkey’s motivational states and issuing in the relevant behavior. (In case you are worried that the two collapse because, say, a good explanation justifies, I should note that I don’t think the monkey is given an explanation of his behavior here, and for similar reasons—the monkey does not have the requisite capacities to receive this sort of explanation. So I suppose I would still take issue with the (perhaps weaker) claim that the monkey receives a reason, where this involves the monkey receiving an explanation. The sense in which I think the monkey receives a reason is that he is made aware of a consideration that engages his motivational states and issues in the relevant behavior. This reason both explains and justifies his behavior, but the explanation and justification are given to us, not to the monkey.)

One last point. You are right that we may be incorrect about why the monkey leapt into the tree. Perhaps it missed the predator but saw some food in the tree. So we might mistakenly attribute mental states to the monkey, and our ordinary linguistic practices are sensitive to this. But I don’t see how this bears on the issue of whether or not our talking as if the monkey gains a justification is metaphorical or not. If we misattribute a mental state to the monkey, this impugns both the claim that the relevant consideration is justifying and the claim that it is explanatory. If there is no representation of a predator to justify the leap into the tree, then there is equally no representation of a predator to engage the motivational states of the monkey. I don’t see a special problem for my view here.

On Reasons and Animals: My Reply to Ben

Special thanks to Ben for his thoughtful feedback on my previous post (see comments). I concede that our ordinary linguistic practice often includes the metaphorical attribution of agency. But of course, that does not mean that such attributions are always metaphorical. For example, no one worth taking seriously (at least for the purposes of the present discussion) would say it is metaphorical in the case of fully competent human beings. So the question is, on which side of the divide does our agency-attributions to animals fall? Is it like our agency-attribution to inanimate objects, or is it more like our attributions to human beings?

Now there are clearly numerous respects in which animals are more like human beings than they are like inanimate objects. Hurley’s monkey, for example, has a heart, brain and haemoglobin bearing blood cells; as do we. But these points of similarity are not salient to the question at hand. Thus, in this post, my task will not be merely to show that animals are more like humans than they are like inanimate objects. That’s a given. Rather, it will be to show that they are similar to humans in ways that are salient to the question of agency. Consequently, when assessing my claims, one cannot simply ask if they are true; one must ask if they’re relevantly true. This is the question I wish to take up.

One striking difference between animals and inanimate objects is that the former are ordinarily assumed to possess motivational states while the latter are not. Moreover, the claim that animals possess motivational states is meant to be taken literally. This is a point made quite emphatically by Mary Midgley:

There is nothing anthropomorphic in speaking of the motivation of animals. It is anthropomorphic to call the lion the King of Beasts, but not to talk of him as moved, now by fear, now by curiosity, now by territorial anger. These are not the names of hypothetical inner states, but of major patterns in anyone’s life, the signs of which are regular and visible. Anyone who has to deal with lions learns to read such signs, and survives by doing so. Both with animals and with men, we respond to the feelings and intentions we read in an action, not to the action itself. (Midgley (1978), Beast and Man., pp. 105-6.)

I believe that the difference between animals and inanimate objects highlighted by Midgley is salient to the question of agency. In the knife case, the focus is on the event (the cutting of the hand). But in the animal case, the focus is on the animal’s inner (read: psychological) states. If one sees a lion walking towards you, it is important to be able to distinguish between whether it is being motivated by a desire to eat you or a desire to reunite with its pride that is gathered under a tree fifty feet behind you. Our folk psychology is what provides us with the tools we need to interpret the lion’s actions. But notice, we are not simply paying attention to the bodily movements of the lion, but also to the inner psychological states (e.g., the lion’s intentions), which we hope to opine by observing the bodily movements. Our concern is primarily with the lion’s psychological states since they will determine what the lion will do next. Moreover, while we may attempt to put ourselves in the lion’s shoes in order to help us opine what its intentions are, we nevertheless assume that these psychological states are real and that they belong to the lion.

The preceding claims are supported by a key assumption of any folk-psychological attempt to make sense of animal behaviour—namely, that when we attribute motives and intentions to animals, we may get things wrong. For example, we assume that it is possible for us to mistakenly conclude that the lion intends to reunite with its pride when it actually has its sights set on us; and this assumption entails that the lion’s inner states exist quite independently of our beliefs about them; that they are not simply hypothetical states we attribute to the lion. This does not seem to be true of inanimate objects, which occupy a very different place in our folk psychology. We do not, for example, think we can be mistaken about the motivations of an inanimate object. There is simply nothing to be mistaken about. I believe this difference suggests that while the motivations we attribute to inanimate objects are metaphorical, the same is not true in the case of animals. When we attribute a motive to an animal we typically take ourselves to be attributing something objectively real, something we may possibly be mistaken about.

I believe the above observations may reveal a difficulty with Ben’s alternative proposal. He writes:

We attribute rational agency to non-rational animals (and even inanimate objects) because we think about the relevant behavior/events (in these, but not necessarily all cases) by adopting the animal's (or object's) perspective from within our own first-personal point of view.

The problem with this proposal is that it overlooks what we ordinarily take to be an important difference between animals and inanimate objects. In the case of inanimate objects, like a knife, there is simply no perspective to adopt. The knife does not enjoy any representational states; it does not see, taste our experience the world in any way. It does not have desires, wants, or appetites. This is why talk of adopting a knife’s perspective can only be metaphorical. But animals, like Hurley’s monkey, are ordinarily assumed to have a perspective in the most literal sense imaginable. They enjoy representations of the world, and are plausibly assumed to be motivated by desires and appetites, and so on.

I hold that having motivational psychological states and being moved to act by them is sufficient for agency. Since the term agent, at least as it features in philosophical discussions, is a term of art, this definition may be seen as stipulative. Given the account of agency I favour (and one is of course free to suggest a better philosophical account), it follows from the conclusion of the preceding paragraph that animals may be agents (in a non-metaphorical sense). But terminological issues aside, the claim I actually set out to defend in my previous post has little to do with whether or not animals are agents and more to do with the concept of a reason and with what it means to be responsive to reasons. My contention is that there is an ordinary usage of the word ‘reasons’ that is consistent with the claim that animals may be responsive to reasons, and it is this usage of the word that I wish to subject to philosophical analysis.

In attempting to identify the ordinary usage of the word ‘reason’ that I have in mind, it may be helpful to distinguish between reasons-for-which an agent acts and the reasons-with-which an agent acts. The latter refers to considerations that an agent takes to be a justification for her carrying out a certain action. By contrast, let us say that the former refers to whatever motivates an agent to act. (Note: I may not be drawing this distinction in the same way others have.) When we specify the reasons-for-which an agent acts, there are typically two forms that our explanations take: We may say that a lion is walking towards me in order to devour me (an mode of explanation that emphasises the lion’s goals; think pull rather than push), or we may say the lion is walking towards me because it wants to devour me (a mode of explanation that cites specific desires that the lion possesses; think push rather than pull).

I believe that there is an ordinary and intuitive use of the word ‘reason’ according to which the following is true: to say that the lion approached the man because it desired to devour him is to give the lion’s reason for approaching the man. Likewise, to say that the lion approached the man in order to devour him is, again, to give the reason the lion approached the man. It is this usage of the word ‘reason’ that I am interested in preserving: reasons as reasons-for-which. I do not claim that it is the only usage of the word; I do not even claim that it is the only usage of the word that has a legitimate claim to being part of our ordinary linguistic practice. There may be multiple ordinary usages (i.e., meanings) of a single word. I only claim that it is a legitimate part of our ordinary linguistic practice (rather than my personal concoction), and it also happens to be a part of our ordinary linguistic practice that I wish my account of reasons and agency to preserve.

McDowell on Rational Animals

In my previous post on this topic, I described Burge's alternative to M-rationalism. Following Burge, I hold that justification is not the only type of warrant. Moreover, consistent with Burge, I hold that a psychological transition need not be motivated by a higher order attitude that takes the transition as its object in order for that transition to provide that agent with a justification for beliefs based on that transition. However, I part ways with Burge by advocating a broader and more inclusive conception of reasons; one that has application not only to justification but also to entitlement. In a slogan: only reasons warrant. I wish to say that the transitive inference performed by Hurley’s monkey provides it with a reason that warrants the monkey’s inference-based belief. On this view, a transition may count as reason-conferring even if the agent lacks the conceptual resources necessary to recognise that it accords with a rational principle. This less restrictive conception of reasons has application to all potentially belief-yielding psychological transitions, not only the psychological transitions of rational animals (in the sense reserved for mature humans). In short, one does not have to be a rational animal to have a reason.

The broader and more inclusive conception of reasons presently on offer is consistent with that of John McDowell, who draws a distinction between “responsiveness to reasons” and “responsiveness to reasons as such”:

The notion of rationality I mean to invoke here is the notion exploited in a traditional line of thought to make a special place in the animal kingdom for rational animals. It is a notion of responsiveness to reasons as such. That wording leaves room for responsiveness to reasons, though not to reasons as such, on the other side of the division drawn by this notion of rationality between rational animals and animals that are not rational. Animals of many kinds are capable of, for instance, fleeing. And fleeing is a response to something that is in an obvious sense a reason for it; danger, or at least what is taken to be danger. If we describe a bit of behaviour as fleeing, we represent the behaviour as intelligible in the light of a reason for it. But fleeing is not in general responding to a reason as such. (Italics his)

Following McDowell, I reserve the title “rational animal” for those agents that have the capacity to conceive of reasons as such. What makes one a rational animal (in the sense reserved for mature human beings) is not the ability to possess reasons—for this is an ability shared by some non-rational animals—but the ability to recognise one’s reasons as reasons. However, as McDowell’s “fleeing” example illustrates, an agent does not have to be a rational animal in order to possess or be responsive to reasons simpliciter. Hence, a distinction is drawn between the necessary conditions for possessing reasons; conditions that do not require the possession of the relevant concepts; conditions that some non-rational animals can meet; and the conditions for having beliefs about one’s reasons; conditions that do require the possession of the relevant concepts; conditions that only rational animals can meet. One does not have to be a rational animal to have reasons; but one must be a rational animal to reflect on them.

This manner of speaking strikes me as more natural, from a terminological point of view, than that of Wallace and Burge. To say that a psychological transition is rational is, intuitively, to say something about that psychological transition. To say that an agent is a rational animal is, intuitively, to say something about that agent. But on the Wallace-Burge way of putting things, to say that a psychological transition is rational is actually to say something about the agent enjoying the transition. This is because both Wallace and Burge make the status of a psychological transition dependent on whether or not the agent enjoying the transition possesses certain conceptual capacities, a fact that I take to be independent of the nature of the transition itself. By contrast, the conception presently on offer makes a transition’s status as rational contingent on the transition’s non-arbitrary truth-conduciveness, and an agent’s status as rational contingent on the agent’s conceptual capacities. This preserves a clear distinction between the rational status of a transition and the rational status of an agent; providing us with at least the option of talking about the two independently, should we need to do so.

The need to speak of the rational status of a transition independently of the status of the agent performing it arises in the case of Hurley’s monkey. We can imagine circumstances in which inferences performed by Hurley’s monkey comply or fail to comply with the norms of rationality. It would therefore be useful to be able to talk about the rational status of such an inference even if we recognise that Hurley’s monkey is not a rational animal. Moreover, there may be instances in which there is simply insufficient empirical evidence available to determine whether or not a particular agent has the conceptual resources necessary to recognise reasons as such. It would therefore be useful to be able to talk about the rational status of a transition performed by an agent without having to take a stand on whether or not the agent is rational. As it happens, our ordinary linguistic practice already provides us with tools needed for such a task, for it is part of said practice to talk about the reasons that some nonhuman animal has to behave in such and such a manner or hold such and such a belief, while leaving fixed the question of whether or not that animal can respond to reasons as such. This brings us to the most straightforward motivation for adopting a conception of reason that has application to agents that cannot recognise their reasons as such, agents that may be entitled but not justified; to wit, such a conception comports with our ordinary linguistic practice.

Burge's Alternative to M-rationalism

In my previous post, I presented a number of objections to the M-rationalist account suggested by R. J. Wallace. An alternative to Wallace's account is provided by Tyler Burge, who maintains that “a norm need not be understood or intentionally adhered to by the individual that it applies to or governs.” By Burge’s lights, the inference performed by Hurley’s monkey may be described as governed by the principle of transitivity even though the monkey performing the inference lacks the concept of transitivity and is not guided in its cognitive activity by its recognition of the concept. Hence, Burge’s slogan: “norms need not guide.” If we assumed that being governed by a rational norm is sufficient for a psychological transition to be rational (i.e., reason-conferring), then Burge’s account of what it means for a psychological process to be governed by a rational principle would entail that the transitive inference of Hurley’s monkey is a rational transition. However, as we shall soon see, things are not so simple on the Burgean account.

Burge distinguishes between two types of epistemic warrant; entitlement and justification. Regarding the former, he writes:

Individuals can be epistemically entitled to a belief without having reasons warranting the belief, without having the conceptual repertoire necessary to have relevant reasons for the belief, and without having the concepts needed to understand or even think the entitlement.

By contrast, Burge defines justification as “warrant by reason that is conceptually accessible on reflection to the warranted individual.” By his lights, Hurley’s monkey may be entitled to its inference-based belief. But insofar as Hurley’s monkey lacks the conceptual resources necessary to recognise that the inference is generalisable, it does not have a justified inference-based belief. And herein lays the rub. Burge only applies the label “reasons” to warrants that are conceptually accessible to the warranted agent—i.e., justifications. The upshot is that while Hurley’s monkey is entitled to its inference-based belief, it does not have reasons for this belief.

In this respect, Burge appears to be on the same page (albeit not the identical paragraph) as the Wallace. When Wallace conceives of a rational norm “governing” a psychological transition, he specifically has something akin to Burge’s notion of justification in mind. By contrast, when Burge conceives of a rational norm “governing” a psychological transition, he has not only justification but also entitlement in mind (the latter failing to even enter Wallace’s consideration). Thus, we have two different conceptions of what it means to be “governed” by a rational principle, the first (due to the Wallace) confined to transitions that take place within the space of justifications and the second (due to Burge) that includes transitions that take place within the space of entitlements.

Apart from Burge’s introduction of the notion of entitlements, the differences highlighted thus far between the two thinkers strike me as primarily terminological. Insofar as Hurley’s monkey lacks the appropriate concepts, both thinkers are committed to saying that it lacks reasons for its inference-based belief and both thinkers are ultimately committed to denying that the monkey’s inference-based belief is justified. However, Burge goes beyond Wallace in identifying a category of epistemic warrant that the latter simply fails to consider; namely, entitlement. Consequently, Burge has the resources to say that the inference-based belief of Hurley’s monkey is warranted, while Wallace does not.

There is also a substantive difference in the conception of justification adhered to by both thinkers. According to Wallace, a psychological transition is reason-conferring only if the agent engages in the transition because she recognises that it accords with a rational norm. By contrast, Burge only requires that the agent possess the conceptual resources necessary to recognise that the psychological transition is generalisable; it is not necessary that the agent engage in the transition because she recognises that the transition is generalisable or accords with a rational norm. Thus, Burge's account lacks the motivational implications of Wallace's account.

Criticisms of M-rationalism

In my previous post, I adumbrated the Motivational Rationalist (henceforth, M-rationalist) objection to my claim that one may instantiate a rational transition even if one lacked the concept of a rational transition. In the present post, I will limn a few criticisms of M-rationalism. Specifically, I will argue that M-rationalism can preserve our quotidian intuitions—e.g., entail that most of the transitions we would ordinarily take to be rational are rational—only if it is able to accommodate the possibility that many (if not most) of the inferences we would putatively regard as rational are not motivated by a belief to that end. But this does not seem to be a condition that M-rationalism can fulfil. To see why this is so, consider the following example:

The Unreflective Inference Example:
Mark is talking on the phone and is informed by his girlfriend, Jessica, that she will be arriving at 7:30pm that evening, and that his best friend, Rick, will be arriving a few hours after her. After hanging up the phone, Mark’s roommate asks him if Rick will be getting in before 7:00pm that evening. Without much hesitation Mark says no. I maintain that in the typical case, Mark does not form the belief that the inference he performs accords with the principle of transitivity and, a fortiori, no such belief motivates him to make the inference. In fact, Mark may not have formed any belief that takes as its object the particular (token) inference he performs. When his roommate poses the question the answer may simply come to Mark, without any accompanying beliefs about the inference itself. Presumably, Mark is aware of the content that forms the premises of the inference: the fact that his girlfriend will be arriving after 7:30pm, the fact that his best friend will be arriving after his girlfriend, and the fact that 7:00pm is earlier than 7:30pm. However, I wish to allow that he may have no beliefs about the inferential pattern in which this content features; for example, that the inferential pattern accords with the principle of transitivity or some unspecified rational norm. On this score, I hold that an inference-based belief may be very much like a perception-based belief; neither the inference nor the perceptual experience need feature as the object of some higher-order belief in order for the agent to have a belief based on that inference or perceptual experience.

Whether motivated by the relevant belief or not, I wish to say that Mark’s inference constitutes a rational transition; it provides Mark with a reason for his inference-based belief that Rick will be arriving after 7:00pm. But according to the Token M-rationalist, Mark’s inference is a rational transition only if it is motivated by his belief that it is generalisable. However, it is not immediately clear that Mark has the belief that the particular inferential token accords with a rational principle or, a fortiori, that such a belief motivates Mark to perform the inference. In fact, if one were to ask Mark if his inference was motivated by the belief that it is generalisable, he may very well respond by saying that he had not given the matter any thought. If we were to take such a reply from Mark at face value, we would have to conclude that Mark had not formed the relevant belief and that it was therefore unavailable to motivate his inference. Consequently, the Token M-rationalist would be forced to say that Mark’s inference fails to constitute a rational transition. But this strikes me as highly counterintuitive.

It should be clear that the preceding argument has some force against Token M-rationalism. However, it seems to have less force against Type M-rationalism. If we assume that having the concept of a particular inferential transition entails having the belief that transitions of that type are generalisable, then the fact that Mark has the concept of a transitive inference entails that he has the relevant belief.

However, Type M-rationalism is still saddled with the following two difficulties. First, it significantly narrows the scope of application of Wallace’s distinction between inferences that are motivated by a rational principle and inferences that merely correspond with a rational principle. Presumably, Wallace intended that the distinction be applied on a case by case basis, so that a single agent, Mark, could perform a transitive inference that was rationally motivated on one occasion and then perform a transitive inference that was not rationally motivated on another (future) occasion. However, Type M-rationalism runs into difficulties when it tries to accomodate this aspect of Wallace’s distinction. If we assume that the motivating belief is a standing belief, then it should be available to motivate all of the inferences that Mark performs once he has acquired the concept of a transitive inference.

Moreover, Type M-rationalism simply fails to provide us with the theoretical resources we would need to draw any possible distinction between those cases in which a particular inference is motivated by a standing belief and those occasions when it is not. The upshot is that an agent either has all transitions of a certain type motivated by a standing belief (assuming that the relevant standing belief is present) or she has no transitions of that type motivated by a standing belief (assuming that the relevant standing belief is absent). While this allows for the intelligibility of Wallace’s distinction when it is construed in terms of a contrast between a rational agent (equipped with the relevant concepts) and a non-rational agent, it does not allow for such a distinction with respect to a single rational agent at different times. However, this implication seems to fly in the face of Wallace’s aims in introducing the distinction.

Second, there remains a lacuna in the Type M-rationalist argument. Recall, the Type M-rationalist maintains that any agent may be ascribed a standing belief that a particular transition is generalisable simply in virtue of having the concept of that transition. I am largely sympathetic to this claim. However, the Type M-rationalist wants to make the further claim that the standing belief is what motivates the agent to complete the relevant inference. However, it is not immediately clear that this is so. The Type M-rationalist may have shown that there is a standing belief available, but she has not shown that it motivates the agent to complete the inference. Moreover, it seems that the question of whether or not it is the standing belief that does the motivating is an empirical rather than conceptual question. Once we recognise that it is possible (not only conceptually but also empirically) for a particular inferential pattern to be instantited without the belief that the inferential pattern in generalisable, it becomes an empirical question whether or not that inferential pattern is motivated by such a belief on a given occasion. This represents a significant challenge to Type M-rationalism since it makes the plausibility of the thesis hostage to future scientific investigation. Should future investigation reveal that it is generally the case that our unreflective inferences are not motivated by a standing belief, then Type M-rationalism would entail that such inferences fail to constitute rational transitions.

To recap, I have argued that M-rationalism (in both of its forms) faces a challenge when it comes preserving our ordinary intuitions about which transitions are rational. In the case of Token M-rationalism, the objection is much more straightforward. Although we would ordinarily regard unreflective inferences, like that performed by Mark, to be reason-conferring, Token M-rationalism seems to entail that such inferences are not rational transitions. In the case of Type M-rationalism, whether or not it is able to accommodate cases like that of Mark will depend on future scientific investigation. Should we discover that the inferences of mature humans are motivated by a standing belief that they are generalisable, all is well. But if we were to discover that this is not so, then Type M-rationalism would have the highly counterintuitive consequence that such inferences are not rational transitions.

I believe that it is a weakness for a philosophical theory to have its plausibility be contingent in this way. At best, we are forced to wait until all the facts are in before we can endorse the theory, and at worse, one begins to suspect that this contingency is merely a symptom of a deeper ailment—namely, that the theory fails to capture what is most fundamental about the concepts it sets out to explain.