We are in broad agreement with much of what Elqayam & Evans (E&E) have to say in their article. First, normative theories, process theories, and competence theories are genuinely distinct kinds of theory and ought not to be confused with each other. Second, the psychology of human reasoning ought not to be merely concerned with explaining patterns of deviation from, and conformity to, putative normative standards. Third, we agree that it is extraordinarily hard to adjudicate between competing normative theories. Fourth, we agree that the central goals of the psychology of reasoning are descriptive ones: to identify the patterns of inference in which we engage and the psychological states, processes, and mechanisms that are causally responsible for these patterns. Finally, we strongly suspect that success in this descriptive project does not presuppose a prior identification of a correct normative theory of rationality. For all that, there are two points on which we take issue with E&E's position.
The prudential case against normativism is problematic
Central to E&E's case against normativism is that it invites various infelicitous inferences: for example, from is-to-ought, and from ought-to-is. E&E's suggestion is that excising normative theories will remove such temptations. Yet, they also suggest that precisely the kinds of formal theories that have traditionally been treated as normative standards may also be retained in order to play a host of other roles: for example, constraining and inspiring psychological theorizing, and providing a framework for the articulation of competence theories. In short: Use the (presumed normative) formal theories in psychology, just don't use them as normative theories. But we wonder: If researchers really are as prone to conflating “is” and “ought” as E&E suggest, then why should they be any more successful in avoiding a slide back towards normative interpretations of formal theories? Of course, whether or not this would occur is an empirical matter. But the fact that formal theories have a longstanding dual function – descriptive and normative – and the fact that reasoning is an inherently normative phenomenon – perhaps the paradigmatic object of rationality attributions – leads us to suspect that excising normative theories from descriptive psychology will not have the desired effect.
Even supposing researchers avoid a slide back into normative interpretations of formal theories, we doubt that banishing normative theories will yield the results that E&E seek. For it seems that infelicities and biases closely analogous to those that E&E trace to normativism may well still arise from the misapplication of formal theories, even were those theories not construed normatively. Consider, for example, E&E's complaint that dual process theorists sometimes make the ought-is fallacy of concluding that System 2 underlies a response just because the response is normatively correct. We agree that this is a mistake. But an exactly analogous mistake can result from misapplying formal theories in characterizing cognition, even if one denies their normative status. Suppose, for example, that one incorrectly assumes that the competence theory for System 2 is accurately characterized by some formal theory, say probability calculus. Under such circumstances one may (mistakenly) conclude that, because a response conforms to the formal theory, it was produced by System 2. But in this case the problem cannot be the normative status assigned to the formal theory. Rather, the problem is simply that one is inaccurately using a formal theory to characterize the system's competence. Moreover, we think that the point generalizes. As far as we can tell, much the same is true for many of the other biases and infelicities that E&E attribute to normativism.
Overselling the case against normativism
E&E appear to suppose that genuine normative standards of rationality satisfy the following conditions:
-
(a) They are unconditional in at least the sense that their normative status depends on neither the goals of agents nor the functions of the mechanisms involved in reasoning.
-
(b) They are deontological in the sense that they specify what it is to reason correctly – what is constitutive of good reasoning – in terms of conformity to some appropriate set of rules or principles.
Now perhaps E&E introduce these constraints merely to limit the class of rational norms to those associated with the Standard Picture (Stein Reference Stein1996). Nevertheless, it is worth noting that this is a very narrow – and quite contentious – characterization of normative standards of rationality. Indeed, according to some of the most intensively discussed philosophical theories of rationality, normative standards routinely violate one or both of these constraints. One especially prominent kind of view, which is often called consequentialism, maintains that what it is to reason correctly – what is constitutive of being a rational reasoning process – is being an effective means of achieving some goal or range of goals. So, for example, according to one well-known form of consequentialism – reliabilism – a good reasoning process is one that tends to lead to true beliefs and the avoidance of false ones (Goldman Reference Goldman1986; Nozick Reference Nozick1993). But on such a view, normative standards of rationality are neither deontological nor unconditional. Moreover, as far as we can tell, E&E's worries fail to generalize to such consequentialist theories of rationality.
With this in mind, E&E appear to be guilty of overselling their case against normativism. Contrary to what they appear to suggest, their arguments cannot plausibly be seen as militating in favor of the view that “theories of higher mental processing would be better off freed from normative considerations” (target article, Abstract). Rather, at most their arguments provide grounds for rejecting a specific kind of normativism: one that construes normative standards of rationality in a narrow and not especially plausible fashion.
You have
Access
We are in broad agreement with much of what Elqayam & Evans (E&E) have to say in their article. First, normative theories, process theories, and competence theories are genuinely distinct kinds of theory and ought not to be confused with each other. Second, the psychology of human reasoning ought not to be merely concerned with explaining patterns of deviation from, and conformity to, putative normative standards. Third, we agree that it is extraordinarily hard to adjudicate between competing normative theories. Fourth, we agree that the central goals of the psychology of reasoning are descriptive ones: to identify the patterns of inference in which we engage and the psychological states, processes, and mechanisms that are causally responsible for these patterns. Finally, we strongly suspect that success in this descriptive project does not presuppose a prior identification of a correct normative theory of rationality. For all that, there are two points on which we take issue with E&E's position.
The prudential case against normativism is problematic
Central to E&E's case against normativism is that it invites various infelicitous inferences: for example, from is-to-ought, and from ought-to-is. E&E's suggestion is that excising normative theories will remove such temptations. Yet, they also suggest that precisely the kinds of formal theories that have traditionally been treated as normative standards may also be retained in order to play a host of other roles: for example, constraining and inspiring psychological theorizing, and providing a framework for the articulation of competence theories. In short: Use the (presumed normative) formal theories in psychology, just don't use them as normative theories. But we wonder: If researchers really are as prone to conflating “is” and “ought” as E&E suggest, then why should they be any more successful in avoiding a slide back towards normative interpretations of formal theories? Of course, whether or not this would occur is an empirical matter. But the fact that formal theories have a longstanding dual function – descriptive and normative – and the fact that reasoning is an inherently normative phenomenon – perhaps the paradigmatic object of rationality attributions – leads us to suspect that excising normative theories from descriptive psychology will not have the desired effect.
Even supposing researchers avoid a slide back into normative interpretations of formal theories, we doubt that banishing normative theories will yield the results that E&E seek. For it seems that infelicities and biases closely analogous to those that E&E trace to normativism may well still arise from the misapplication of formal theories, even were those theories not construed normatively. Consider, for example, E&E's complaint that dual process theorists sometimes make the ought-is fallacy of concluding that System 2 underlies a response just because the response is normatively correct. We agree that this is a mistake. But an exactly analogous mistake can result from misapplying formal theories in characterizing cognition, even if one denies their normative status. Suppose, for example, that one incorrectly assumes that the competence theory for System 2 is accurately characterized by some formal theory, say probability calculus. Under such circumstances one may (mistakenly) conclude that, because a response conforms to the formal theory, it was produced by System 2. But in this case the problem cannot be the normative status assigned to the formal theory. Rather, the problem is simply that one is inaccurately using a formal theory to characterize the system's competence. Moreover, we think that the point generalizes. As far as we can tell, much the same is true for many of the other biases and infelicities that E&E attribute to normativism.
Overselling the case against normativism
E&E appear to suppose that genuine normative standards of rationality satisfy the following conditions:
(a) They are unconditional in at least the sense that their normative status depends on neither the goals of agents nor the functions of the mechanisms involved in reasoning.
(b) They are deontological in the sense that they specify what it is to reason correctly – what is constitutive of good reasoning – in terms of conformity to some appropriate set of rules or principles.
Now perhaps E&E introduce these constraints merely to limit the class of rational norms to those associated with the Standard Picture (Stein Reference Stein1996). Nevertheless, it is worth noting that this is a very narrow – and quite contentious – characterization of normative standards of rationality. Indeed, according to some of the most intensively discussed philosophical theories of rationality, normative standards routinely violate one or both of these constraints. One especially prominent kind of view, which is often called consequentialism, maintains that what it is to reason correctly – what is constitutive of being a rational reasoning process – is being an effective means of achieving some goal or range of goals. So, for example, according to one well-known form of consequentialism – reliabilism – a good reasoning process is one that tends to lead to true beliefs and the avoidance of false ones (Goldman Reference Goldman1986; Nozick Reference Nozick1993). But on such a view, normative standards of rationality are neither deontological nor unconditional. Moreover, as far as we can tell, E&E's worries fail to generalize to such consequentialist theories of rationality.
With this in mind, E&E appear to be guilty of overselling their case against normativism. Contrary to what they appear to suggest, their arguments cannot plausibly be seen as militating in favor of the view that “theories of higher mental processing would be better off freed from normative considerations” (target article, Abstract). Rather, at most their arguments provide grounds for rejecting a specific kind of normativism: one that construes normative standards of rationality in a narrow and not especially plausible fashion.