Elqayam & Evans (E&E) have served up an impressive collection of thought-provoking ideas. Most importantly, in my view, they raise challenging questions about what the goals of research on thinking should be, and about the types of theories of thinking that researchers should strive to produce. They argue that the primary goal should be to discover the rules that describe how thinking is actually done, as distinct from how it ought to be done – that the theories we strive to develop should be descriptive, as distinct from normative. They contend that criteria for evaluating the quality of thinking are essential to the practical objective of improving thinking, but not to the objective of investigating and understanding thinking.
E&E's argument regarding the importance of description as a goal of research strikes me as compelling, but their dismissal of normativism, defined as “the belief that people ought to conform to a normative standard” (sect. 1, para. 10), does not. I find it easy to agree with the emphasis on description, but difficult to accept the idea that appeal to norms for the purpose of investigating and understanding thinking is inapt.
E&E acknowledge that “formal systems such as logic and Bayesianism have provided major incentives and inspiration to countless research paradigms” (sect. 7, para. 2), but they see these systems' pitfalls as outweighing their positives by enough to warrant consideration of proscribing normativism. “Can we,” they ask, “make do in reasoning and JDM [judgment and decision making] without normative theories altogether?” (sect. 7, para. 2).
In my view, the answer is no. Or, better, that we could make do without them, but only at unacceptable cost. I take a normative theory of thinking to be a theory that specifies how we should think, taking our capabilities and limitations into account. Some refer to this type of theory as prescriptive, and reserve normative to connote what would be prescriptive for a creature without human limitations. I find this distinction unhelpful; what may be normative for a creature without human limitations would not only not be prescriptive for us, it would not be normative for us either.
We want to know not only how we reason, but how we should reason. Maybe the way we reason is how we should, and maybe not. My sense is that the truth is somewhere in between – that our reasoning is not entirely consistent with what any reasonable normative model that takes account of human capabilities and limitations would prescribe, but also not quite as bankrupt as some presentations of the biases to which we appear to be prone seem to suggest.
The study of thinking is motivated by a variety of goals, description, and evaluation – the latter of which requires norms – among them. The results of countless studies show that neither traditional logic nor probability theory is descriptive of human reasoning as it is generally done. And in the opinions of many scholars, neither logic nor probability theory constitutes a plausible normative theory of how reasoning by mere mortals should be done. Disputes abound among philosophers and logicians regarding what should be considered normative in both deductive and inductive reasoning. This is illustrated, for example, by the long-running debates about the rules of conditional inference: Are conditional assertions truth functional? Should If A, then C be interpreted as the material conditional? Is the probability of the conditional the conditional probability? What does it mean for a counterfactual world to be maximally (or sufficiently) similar to the actual world? The literature is rife with arguments and counterarguments about such questions, most of which are never settled decisively.
That the challenge of developing a plausible normative theory of thinking is daunting has long been recognized by philosophers. MacIntyre (Reference MacIntyre1988), who discusses it at length, takes the position that it is not possible to identify principles of rationality that will be universally recognized as valid. Still, we must have norms, even if any norms that we adopt are unlikely to be acceptable to everyone. We need norms, not only to live by, but in terms of which to understand our thinking – how it is good and how it is not. Even to say that normativism is wrong is to imply the existence of some norm that justifies that assessment.
There are approaches to the study of reasoning and decision making that attempt to bridge the gap between descriptive and normative theories. One example is that of Fox and colleagues (Fox Reference Fox2003; Fox & Parsons Reference Fox, Parsons, Hunter and Parsons1998; Fox et al. Reference Fox, Beveridge and Glasspool2003), whose approach aims to deal realistically with situations in which people have to operate – in which problems often are poorly formed, relevant knowledge may be lacking, the circumstances may be fluid and time limited, but action is required. Perhaps the best chance of developing truly useful normative theories for such cases is via (descriptive) study of how the more effective reasoners and decision makers deal with them. This is the rationale, I believe, that motivates the naturalistic decision-making approach of Klein and colleagues (Klein Reference Klein1998; Pliske & Klein, Reference Pliske, Klein, Schneider and Shanteau2003; Zsambok & Klein Reference Zsambok and Klein1997). In both cases, the goal is not only accurate description of how reasoning and decision making are actually done in real-world situations, but an understanding of the processes that can be used for prescriptive purposes as well.
E&E note several ways in which formal systems can be useful for psychological theorizing, and vice versa. This strikes me as an especially important part of their discussion. However, I wonder about the impression conveyed by their Figure 2 that the flow is primarily from formal systems to psychological theory and data rather than the reverse, and about E&E's exclusion of validation from the ways in which psychological theory and data influence formal theories (see target article, sect. 7). Excluding validation from the ways in which psychology can influence formal theories raises the question of what the basis of the authority of formal theories could be, if not the warrant given by human judgment. One accepts the laws of (some) logic as binding, if one does, because one finds them psychologically compelling; to what else could one turn?
You have
Access
Elqayam & Evans (E&E) have served up an impressive collection of thought-provoking ideas. Most importantly, in my view, they raise challenging questions about what the goals of research on thinking should be, and about the types of theories of thinking that researchers should strive to produce. They argue that the primary goal should be to discover the rules that describe how thinking is actually done, as distinct from how it ought to be done – that the theories we strive to develop should be descriptive, as distinct from normative. They contend that criteria for evaluating the quality of thinking are essential to the practical objective of improving thinking, but not to the objective of investigating and understanding thinking.
E&E's argument regarding the importance of description as a goal of research strikes me as compelling, but their dismissal of normativism, defined as “the belief that people ought to conform to a normative standard” (sect. 1, para. 10), does not. I find it easy to agree with the emphasis on description, but difficult to accept the idea that appeal to norms for the purpose of investigating and understanding thinking is inapt.
E&E acknowledge that “formal systems such as logic and Bayesianism have provided major incentives and inspiration to countless research paradigms” (sect. 7, para. 2), but they see these systems' pitfalls as outweighing their positives by enough to warrant consideration of proscribing normativism. “Can we,” they ask, “make do in reasoning and JDM [judgment and decision making] without normative theories altogether?” (sect. 7, para. 2).
In my view, the answer is no. Or, better, that we could make do without them, but only at unacceptable cost. I take a normative theory of thinking to be a theory that specifies how we should think, taking our capabilities and limitations into account. Some refer to this type of theory as prescriptive, and reserve normative to connote what would be prescriptive for a creature without human limitations. I find this distinction unhelpful; what may be normative for a creature without human limitations would not only not be prescriptive for us, it would not be normative for us either.
We want to know not only how we reason, but how we should reason. Maybe the way we reason is how we should, and maybe not. My sense is that the truth is somewhere in between – that our reasoning is not entirely consistent with what any reasonable normative model that takes account of human capabilities and limitations would prescribe, but also not quite as bankrupt as some presentations of the biases to which we appear to be prone seem to suggest.
The study of thinking is motivated by a variety of goals, description, and evaluation – the latter of which requires norms – among them. The results of countless studies show that neither traditional logic nor probability theory is descriptive of human reasoning as it is generally done. And in the opinions of many scholars, neither logic nor probability theory constitutes a plausible normative theory of how reasoning by mere mortals should be done. Disputes abound among philosophers and logicians regarding what should be considered normative in both deductive and inductive reasoning. This is illustrated, for example, by the long-running debates about the rules of conditional inference: Are conditional assertions truth functional? Should If A, then C be interpreted as the material conditional? Is the probability of the conditional the conditional probability? What does it mean for a counterfactual world to be maximally (or sufficiently) similar to the actual world? The literature is rife with arguments and counterarguments about such questions, most of which are never settled decisively.
That the challenge of developing a plausible normative theory of thinking is daunting has long been recognized by philosophers. MacIntyre (Reference MacIntyre1988), who discusses it at length, takes the position that it is not possible to identify principles of rationality that will be universally recognized as valid. Still, we must have norms, even if any norms that we adopt are unlikely to be acceptable to everyone. We need norms, not only to live by, but in terms of which to understand our thinking – how it is good and how it is not. Even to say that normativism is wrong is to imply the existence of some norm that justifies that assessment.
There are approaches to the study of reasoning and decision making that attempt to bridge the gap between descriptive and normative theories. One example is that of Fox and colleagues (Fox Reference Fox2003; Fox & Parsons Reference Fox, Parsons, Hunter and Parsons1998; Fox et al. Reference Fox, Beveridge and Glasspool2003), whose approach aims to deal realistically with situations in which people have to operate – in which problems often are poorly formed, relevant knowledge may be lacking, the circumstances may be fluid and time limited, but action is required. Perhaps the best chance of developing truly useful normative theories for such cases is via (descriptive) study of how the more effective reasoners and decision makers deal with them. This is the rationale, I believe, that motivates the naturalistic decision-making approach of Klein and colleagues (Klein Reference Klein1998; Pliske & Klein, Reference Pliske, Klein, Schneider and Shanteau2003; Zsambok & Klein Reference Zsambok and Klein1997). In both cases, the goal is not only accurate description of how reasoning and decision making are actually done in real-world situations, but an understanding of the processes that can be used for prescriptive purposes as well.
E&E note several ways in which formal systems can be useful for psychological theorizing, and vice versa. This strikes me as an especially important part of their discussion. However, I wonder about the impression conveyed by their Figure 2 that the flow is primarily from formal systems to psychological theory and data rather than the reverse, and about E&E's exclusion of validation from the ways in which psychological theory and data influence formal theories (see target article, sect. 7). Excluding validation from the ways in which psychology can influence formal theories raises the question of what the basis of the authority of formal theories could be, if not the warrant given by human judgment. One accepts the laws of (some) logic as binding, if one does, because one finds them psychologically compelling; to what else could one turn?