Brains, Persons, and Society *** ABSTRACTS    Cervelli, Persone e Società ***ABSTRACTS

Stuart Yasgur
London School of Economics

Accounting for Rationality

“Several philosophers have argued that rational requirements must be somehow inherent in the nature of the mental states they are concerned with. I am sure they are right in some way… However, I do not know how this general idea can be worked out in detail, to provide a criterion for determining what rationality requires.”

- “Requirements of Rationality” John Broome

This paper explores the possibility that rationality requires a connection between recognizing a fact to be a reason and taking a fact to be a reason.  Further, it is argued that this connection can accommodate some of Broome’s recent contributions to the bootstrapping literature.

Some of Broome’s recent contributions:

Claim 1: If one intends an end, and believes the means are necessary to achieve it, but does not intend the means, then one is not as one ought to be.

Claim 2: Intending an end does not give one reason to intend the means one believes necessary. 

While many seek to hold both of these claims, there is an obvious tension between them.  For, if intending an end does not give one reason to intend the means one believes necessary, in what is the agent at fault if he does not intend to take the means?

Broome’s recent works have made a substantial contribution to resolving this tension by arguing that the normative requirement referenced in (1) is not the reason relation mentioned in (2).  Broome has argued that we would be well served to recognize a requiring relation that is distinct from the reason relation.  Indeed, Broome has identified the four following possible relations.[1] 

In this current paper we are only concerned with the Reason and Requires relations.  From this we can see that while the Reason relation has a narrow scope, it is slack.  Whereas, the Requires relation has a broad scope, but is strict.  We can illustrate these differences as follows.

Narrow vs. Broad:

If an agent has a reason to intend an end, and a means is necessary to achieving that end, then the agent has reason to take the means.[2]  The Reason relation is narrow.

However, if an agent intends an end, and believes that a means is necessary to achieving the end, he is not required to take the means.  Rather, what is required of him is not to be in a state in which he intends the end, and believes the means necessary, but does not intend the means.  He can avoid being in such a state by coming to intend the means, or by dropping the end, or by coming to believe that the means is not necessary.  Any of the three will fulfill the Requires relation.  The Requires relation is broad.

Strict vs. Slack:

While there are three different ways in which the agent can comply with the requiring relation in the current example, the strictness of the requiring relation demands that one of the three.  The Requires relation is strict.

The Reason relation is weaker in this regard.  That the agent has a reason to take the means is consistent with his not intending to take the means, because he may have more reason not to take the means.  The Reason relation is slack.

Suggestion

While Broome’s discussion of the different characteristics of the Requires relations makes a considerable contribution, it also has a major drawback.  As described, it seems troublingly opaque in two ways.  First, there is no basis to enumerate what it requires.  We might be able to come to a consensus based on intuition about what it requires, but then it seems talk of the Requires relation is little more than a convenient way to talk about a number of strict and broad requirements.  In other words, the real work seems to be done by Broome’s arguments that certain requirements are strict and broad.  The stipulation of a Requires relation seems to do little work.

Second, though the characteristics of the Requires relation, strictness and broadness, do seem to carve things at the joints, there is no basis for understanding why this is the case. 

My suggestion is that we recognize broadness and strictness as features of a more basic account of rationality.    Rationality is the ability to be ruled by one’s own reason.  One cannot be ruled by one’s own reason if one does not take to be a reason that which one recognizes to be a reason.  Therefore, rationality requires that one take to be a reason that which one recognizes to be a reason.  Further, if one fails to do so, one is thereby irrational.

This requirement of rationality is strict.  Each case in which an agent fails to take to be a reason that which he recognizes to be one, is a case of irrationality.  To see what this means, let’s return to the earlier example.  The agent intends an end, but does not intend to take the means he thinks necessary to achieve it.

If the agent intends the end, then he recognizes himself as having a reason to bring the end about.[3]  If he also believes a means to be necessary, then he recognizes himself as having a reason to take the means.  For to recognize that a means is necessary to achieving an end which one recognizes oneself as having reason to bring about is to recognizes oneself as having a reason to take the means.  The belief that the means is necessary is a belief that there is an identifying reason to take the means.

Having said this, there are two relevant scenarios[4] in which the agent may fail to have the intention to take the means.  First, the agent may fail to take himself to have a reason to take the means.  This is straightforwardly a case in which the agent fails to take as a reason that which he recognizes is a reason, and is therefore irrational.

Second, the agent may not form the intention to take the means because he realizes he has more reason not to take the means than he does to take the means.  But if this is the case, and the agent recognizes that the means is necessary to achieving the end, then the agent recognizes himself as having reason not to achieve then end.  As in the first scenario, the belief that the means is necessary is the belief that there is an identifying reason against taking the ends.  Given this belief, if the agent takes himself to have more reason not to take the means than to take it, then he recognizes himself as having more reason not to take the end than to take it, and rationality requires that he take this to be the case.

From this discussion, we can also see that the scope of rationality is broad.  The agent’s irrationality does not consist in his failing to take the means, but in his intending to take the end while not intending to take the means that he believes is necessary.  Or, to say the same thing more explicitly, the agent’s irrationality does not consist in failing to take himself to have a reason to achieve the means, but in failing to take himself as having such a reason when he recognizes that he does.  The agent would no longer be irrational if he took himself to have a reason to take the means or if he no longer recognized himself as having a reason to do so.  This could be the case if he no longer believed the means to be necessary, or if he no longer took himself to have a reason, all things considered, to achieve the end.  These are the same possibilities that we discussed above.

Moving Forward

This paper identifies a specific requirement of a narrow conception of rationality and argues that it is both strict and broad in Broome’s sense. If, as I will argue in future papers, the failure of this specific rational requirement can be located in each of the canonical versions of irrationality (e.g., akrasia, believing a contradiction, intention inconsistency, and failure of means end rationality), then we may think that the distinction between recognizing a fact to be a reason and taking a fact to be a reason goes a long way towards offering a general account of the requirements of rationality.