Articles Covered:*

“Vindicating the Normativity of Rationality” Nicholas Southwood, pages 9-30.
“Reasons: Explanations or Evidence?” Stephen Kearns and Daniel Star, pages 31-56.
“Reply to Southwood, Kearns and Star, and Cullity” John Broome, pages 96-108.

*The Symposium on Broome’s work also contains an article by Garrett Cullity: “Decisions, Reasons, and Rationality.” A review of that article by J.J. Swindell has already appeared in the Philosopher’s Digest, April 7, 2009.

Main authors discussed (by Southwood): John Broome, N. Hussain, D. Velleman, M. Bratman, N. Kolodny, T. Scanlon.
Main authors discussed (by Kearns and Star): John Broome.

In his work on rationality Broome distinguishes between requirements of rationality and reasons. Rationality, he argues, cannot be reduced to responding correctly to reasons; it is also a matter of fulfilling requirements like belief consistency, belief closure, instrumental reasoning, and enkrasia (that is, the requirement to intend to X if you believe you ought to X). For Broome ‘ought’ (and not ‘reason’) is the fundamental normative concept. A ‘reason’ is defined as an explanation, or part of an explanation, of normative facts, i.e., an explanation of why N ought to F. One important question is whether rationality is normative. In his more recent work Broome is sceptical: He thinks rationality is normative in its own right, but he admits that he has no decisive argument to show why this is so.

In his highly interesting paper Southwood aims to provide the missing argument why rationality is normative. He argues against the strategy of tackling the question by offering independent reasons each agent has for compliance with the requirements of rationality (e.g. prudential reasons, or reasons deriving from our valuable status as believers and agents). Just as it is a mistake to answer the “Why be moral?” question by trying to offer a justification outside morality, so it is a mistake to try to answer the question “Why be rational?” by searching “for a justification for rational compliance outside of rationality, a source of normativity that is external to rationality” (18). Whether rationality is normative can, Southwood claims, be answered only by providing a theory of rationality that explains what kind of claims rational requirements are and why they are “by their very nature, normative” (19).

Southwood discusses several accounts of rationality recently offered and then comes up with his own proposal. A first position he rejects is the “distinctive object account” of rationality, which attempts to explain the requirements of rationality by specifying the objects to which they apply. As an example he cites N. Hussain’s claim that rational requirements describe the norms of how we should reason. Southwood rejects Hussain’s interpretation as being too narrow: rational requirements (e.g. demands of coherence and consistency) apply to us independently of whether we are engaged in reasoning (21).

Another account Southwood considers is the “proper functioning account”: rational requirements describe the conditions attitudes have to meet to serve agential functions, either cognitive aims (Velleman) or the practical function to coordinate action (Bratman). According to Southwood, the functioning account fails to show why rational requirements have normative force for particular agents (23).

The “subjective reasons account” (Kolodny, Scanlon) interprets rational requirements as reasons we have to form or revise our beliefs or intentions, given the reasons we think we have because of our other beliefs or intentions. However, Southwood objects that requirements of rationality have real normative force, not only “(seeming) normative force” (Kolodny) relative to the reasons subjects think they have (25).

Southwood considers his own interpretation of rational requirements, the “first-personal authority account,” to be an improved version of the subjective reasons account. Rational requirements must have a reference to “our point of view” since they are claims that we ought to have or form certain attitudes. To avoid the troubling subjectivism of Kolodny’s and Scanlon’s account, Southwood distinguishes between ‘subjective demands’ and ‘standpoint-relative demands,’ the latter being demands that are “relative to the particular first-personal standpoint of the agent who is subject to it” (27). Rational requirements are demands we have to fulfill in order to be agents with first-personal authority. We have first-personal authority if we are standing in a specific relation of accountability to ourselves. “The normativity of rationalty,” Southwood claims, “is a matter of honoring our first-personal authority” (30).

Southwood’s article gives a most valuable exposition of current explanations of rationality. His positive solution raises some questions, e.g., what exactly “honoring our first-personal authority” means and how Southwood’s account is related to an account like Korsgaard’s, where first-personal authority involves commitment to a principle of autonomy in terms of which reasons are defined. In his reply, Broome objects that a reference to standpoint-relativity and first-personal authority does not explain why rationality is normative. The rational requirement not to believe a contradiction is, as Broome points out, not standpoint relative and not dependent on first-personal authority. Broome’s objection raises the question whether the senses of “normativity” at stake here are different: one question is whether rational requirements are normative given their specific content; the other question is why we should comply with them. Since Broome himself emphasizes that one cannot be a person if one violates too many requirements of rationality (99), Southwood’s accounting for the normativity of rationality by explicating the presuppositions of agency does not seem implausible.

Kearns and Star develop a highly challenging critique of Broome’s account of reasons in terms of explanations of ought-facts. Their central claim is that “reasons are evidence, rather than explanations,” they are “evidence of the obtaining of ought facts” (32). Kearns and Star draw the distinction between explanation and evidence in the following way: if fact F explains another fact G, then F makes it the case that G. Evidence, however, merely indicates or makes it probable that G is the case. If F is evidence for G, G may not be true. So their analysis of a reason amounts to the claim: “Necessarily, a fact X is a reason for an agent N to F if and only if (iff) X is evidence that N ought to F” (37).

Kearns and Star consider it an important advantage of their analysis of reasons that it takes into account the being/having distinction. It is crucial to distinguish between reasons there are and reasons an agent has, and talk of evidence makes sense of that distinction: a reason I have is evidence I have; ‘there being a reason’ means ‘there is evidence’ (38).

They raise three objections against Broome’s analysis of reasons: first, Broome cannot explain the practical role of reasons in deliberation; second, Broome does not provide a “full-blooded” analysis of reasons; third, Broome offers no unified account of reasons, because he introduces two different kinds of normative reasons, namely perfect reasons and pro tanto reasons, which cannot be analyzed in similar terms. The property of being an explanation of a normative fact (perfect reason) and weighing in favor of a certain action (pro tanto reason) are for Kearns and Star completely different properties.

The first two objections deserve more detail. In regard to the first objection: that a fact is an explanation of what an agent ought to do is according to Kearns and Star not crucial in guiding an agent in her deliberations. Much more relevant for them is that a fact is evidence for what an agent should do. Moreover, given Broome’s analysis of reasons in terms of explanations, persons would not need reasons in order to figure out what they ought to do (which would eliminate reasons from practical life). Their example: A person might come to the conclusion that she ought to eat cabbage by the fact that a reliable book advises her to do so, yet this fact is not an explanation why she ought to eat cabbage (40, 41).

In regard to the second objection: Broome’s analysis would, Kearns and Star argue, be full-blooded if he would analyze the concept ‘reason’ generally in terms of another concept, namely ‘explanation.’ This analysis, however, holds only for ‘perfect reasons.’ Because Broome introduces ‘pro tanto reasons’ (which play a favoring role in a weighing explanation of what one ought to do) he faces a problem: the idea of facts that “count in favor of a certain action” cannot according to the authors be explained otherwise than by the idea of a reason. To say that a fact counts in favor of a certain action is to say that “the fact is a reason to perform this action” (43). Moreover, Broome’s analysis cannot specify the strength of reasons; something that can be easily achieved by referring to the strength of evidence (44).

The authors then provide further arguments supporting their account. For example, they argue that their analysis of reasons avoids the assumption that ‘ought’ depends on unknowable ought-facts, an assumption which would violate the reasonable condition that ‘ought’ implies the “possibility of our knowing what to do” (53). What sense would it make to criticize agents for ignoring reasons that they could not know? Kearns and Star point out that their interpretation merely presupposes that “if one ought to F, there is some (knowable) evidence that one ought to F,” which seems plausible since a fact is evidence of itself. The main difficulty of their account seems to be that it does not amount to all-things-considered oughts and normative truths since, as Kearns and Star concede, “evidence doesn’t generally play an active role with respect to fixing the truth” (55). However, as the authors point out (referring to Parfit’s example of the trapped miners), evidence makes a contribution to figuring out what we ought to do.

In his reply, Broome objects that the “evidential property” is not the decisive criterion for a normative reason (101). Broome accuses Kearns and Star of confusing “epistemology with the determination of facts” (102). Weighing the evidence for (or against) the truth of a proposition is a matter of epistemology. We can weigh evidence for the proposition that one ought to X against the evidence that the proposition is false (it is not the case that one ought to X). But whether one ought to X is dependent on the weight of the reason that you ought to X against the weight of the reason that you ought not to X (which is different from ‘it is not the case that one ought to X’). The weight of a reason is dependent on the facts, not the evidence you have.

Reviewed by Herlinde Pauer-Studer
University of Vienna

No Comments

No comments yet.

RSS feed for comments on this post. TrackBack URI

Leave a comment