Cimpian & Salomon (C&S) propose the “inherence heuristic” as an explanation of the cognitive process that leads people to explain observed patterns based on their inherent features. They propose a unified account for a diverse set of empirical findings, including “is–ought errors” in moral reasoning (sect. 3.4). We agree with the authors about the relevance of heuristic cognitive processes in illuminating moral judgment and decision making. However, we doubt the ability of the inherence heuristic alone to explain a range of phenomena, including complexity in ethical reasoning, for two major reasons: (1) the inherence heuristic offers modest insights into the complex nature of both the is–ought tension in moral reasoning and moral reasoning per se, and (2) the inherence heuristic does not reflect the complexity of domain-specific moral heuristics.
First, is–ought inferences, which Cimpian and Salomon try to explain with the inherence heuristic, require critical scrutiny. Interestingly, David Hume himself did not frame such leaps in reasoning as straightforward errors or fallacies. In fact, his Treatise of Human Nature (Reference Hume1739/1975) is an attempt to develop a science of the human nature and mind, which grants an obvious importance to the understanding of cognitive processes and emotions for ethics. G. E. Moore (Reference Moore1903/1971) deserves credit for having transformed Hume's observation into a “fallacy” in support of a doctrine of antinaturalism in ethics. Here lies their first potentially problematic assumption: They have interpreted the is–ought tension as both a dichotomy and a fallacy – that is, “is–ought errors.” Granted, crude is–ought transitions are dubious because they rely on debatable overgeneralizations and moral conservatism, but this was exactly Hume's point of caution with respect to such transitions. Hence, the redescription of Hume's point in the language of inherence heuristic may not add explanatory power but carries the problematic overtone of antinaturalism. Such antinaturalism could lead one to reify the is–ought tension into an essence, which is exactly the problem that the authors identify. In fact, the relationship between essentialism and the is–ought tension had already been identified, for example, by Dewey as the “philosophical fallacy” – that is, the supposition that whatever is found true under certain conditions may forthwith be asserted universally or without conditions (Dewey Reference Dewey1922). In contemporary ethics, pragmatic and coherentist theories that recommend wide reflective equilibrium between ethical principles and factual claims provide a more generous and applicable understanding of how statements of facts inform considered judgments and moral deliberation (Beauchamp & Childress Reference Beauchamp and Childress2009; Racine Reference Racine2008). Importantly, they have avoided creating essences out of normative and descriptive categories based on semantic analyses. The reason is clear: moral judgments and moral reasoning cannot avoid factual assumptions about the nature of moral problems as well as the experience of moral agents. Otherwise, one is pressed to explain where any “ought” could logically come from (Callahan Reference Callahan1996), unless one assumes nonnatural properties as Moore does, which the authors appear to have done.
Second, the inherence heuristic is not contextualized in a broader web of domain-specific heuristics that more explicitly relate to moral theory. This diminishes the relevance and explanatory power of the proposal because the appropriate level of explanation is not identified. The authors provide many examples where the inherence heuristic seems to be at work, but it is more plausible to assume that such a psychological process underlies a number of simple domain-specific heuristics. It could be analogous to the domain-general “attribute substitution” process (Kahneman & Frederick Reference Kahneman, Frederick, Gilovich, Griffin and Kahneman2002), which is potentially implicated in a range of simple heuristics (see, e.g., Gigerenzer Reference Gigerenzer and Sinnott-Armstrong2008; Sunstein Reference Sunstein2005). However, it is certainly not a single heuristic that explains everything from beliefs about orange juice to ethical reasoning. The authors need to dissociate between domain-specific heuristics and domain-general “effort reduction” processes (Shah & Oppenheimer Reference Shah and Oppenheimer2008). Consider the example of different types of moral intuitions and related candidate domain-specific heuristics. Deontologists (e.g., Kant Reference Kant and Ellington1785/1993) claim that certain actions are obviously wrong (e.g., torture) or right (e.g., saving lives). They also sometimes believe that it is an inherent feature of a certain action to be wrong. Most of the time, this strategy works well, but sometimes it misfires and/or leads to problematic inferences (e.g., it is a moral imperative not to lie even to a serial killer).
However, it would be false to assume that this is a feature of only one type of moral intuitions and heuristics. Consequentialists (e.g., Sidgwick Reference Sidgwick1874/1981) use similar reasoning when they focus on consequences (e.g., saving five people as opposed to just one), regardless of the means used or the character traits of the agents involved. They also sometimes believe that it is an inherent feature of a certain consequence to be good, which again can misfire as a strategy and/or lead to morally abhorrent conclusions (e.g., killing one healthy person to harvest organs to save five people is rational).
Even virtue theorists (e.g., Aristotle 1890) use similar reasoning to make their case. Namely, they focus on the character – the habits, intentions, and motives of the agent (e.g., is this something appropriate for a virtuous person?). They also sometimes believe that it is an inherent feature of a virtuous person to react morally appropriately in every situation. Again, this might be a sound strategy most of the time, but it can certainly misfire and/or create moral outrage (e.g., if Jesus/Mohammed/Moses/Buddha/Marx responded in such a way, everyone should).
The so-called inherence heuristic could actually be a domain-general process that might inform a range of domain-specific heuristics. However, this needs to be acknowledged to ensure its integration into the rich body of theoretical and empirical moral psychology. Furthermore, it is wrong to assume that the is–ought divide is insurmountable and that fallibility inherently means faultiness. If the authors maintain such assumptions, they are falling prey to a process they themselves identified.
Cimpian & Salomon (C&S) propose the “inherence heuristic” as an explanation of the cognitive process that leads people to explain observed patterns based on their inherent features. They propose a unified account for a diverse set of empirical findings, including “is–ought errors” in moral reasoning (sect. 3.4). We agree with the authors about the relevance of heuristic cognitive processes in illuminating moral judgment and decision making. However, we doubt the ability of the inherence heuristic alone to explain a range of phenomena, including complexity in ethical reasoning, for two major reasons: (1) the inherence heuristic offers modest insights into the complex nature of both the is–ought tension in moral reasoning and moral reasoning per se, and (2) the inherence heuristic does not reflect the complexity of domain-specific moral heuristics.
First, is–ought inferences, which Cimpian and Salomon try to explain with the inherence heuristic, require critical scrutiny. Interestingly, David Hume himself did not frame such leaps in reasoning as straightforward errors or fallacies. In fact, his Treatise of Human Nature (Reference Hume1739/1975) is an attempt to develop a science of the human nature and mind, which grants an obvious importance to the understanding of cognitive processes and emotions for ethics. G. E. Moore (Reference Moore1903/1971) deserves credit for having transformed Hume's observation into a “fallacy” in support of a doctrine of antinaturalism in ethics. Here lies their first potentially problematic assumption: They have interpreted the is–ought tension as both a dichotomy and a fallacy – that is, “is–ought errors.” Granted, crude is–ought transitions are dubious because they rely on debatable overgeneralizations and moral conservatism, but this was exactly Hume's point of caution with respect to such transitions. Hence, the redescription of Hume's point in the language of inherence heuristic may not add explanatory power but carries the problematic overtone of antinaturalism. Such antinaturalism could lead one to reify the is–ought tension into an essence, which is exactly the problem that the authors identify. In fact, the relationship between essentialism and the is–ought tension had already been identified, for example, by Dewey as the “philosophical fallacy” – that is, the supposition that whatever is found true under certain conditions may forthwith be asserted universally or without conditions (Dewey Reference Dewey1922). In contemporary ethics, pragmatic and coherentist theories that recommend wide reflective equilibrium between ethical principles and factual claims provide a more generous and applicable understanding of how statements of facts inform considered judgments and moral deliberation (Beauchamp & Childress Reference Beauchamp and Childress2009; Racine Reference Racine2008). Importantly, they have avoided creating essences out of normative and descriptive categories based on semantic analyses. The reason is clear: moral judgments and moral reasoning cannot avoid factual assumptions about the nature of moral problems as well as the experience of moral agents. Otherwise, one is pressed to explain where any “ought” could logically come from (Callahan Reference Callahan1996), unless one assumes nonnatural properties as Moore does, which the authors appear to have done.
Second, the inherence heuristic is not contextualized in a broader web of domain-specific heuristics that more explicitly relate to moral theory. This diminishes the relevance and explanatory power of the proposal because the appropriate level of explanation is not identified. The authors provide many examples where the inherence heuristic seems to be at work, but it is more plausible to assume that such a psychological process underlies a number of simple domain-specific heuristics. It could be analogous to the domain-general “attribute substitution” process (Kahneman & Frederick Reference Kahneman, Frederick, Gilovich, Griffin and Kahneman2002), which is potentially implicated in a range of simple heuristics (see, e.g., Gigerenzer Reference Gigerenzer and Sinnott-Armstrong2008; Sunstein Reference Sunstein2005). However, it is certainly not a single heuristic that explains everything from beliefs about orange juice to ethical reasoning. The authors need to dissociate between domain-specific heuristics and domain-general “effort reduction” processes (Shah & Oppenheimer Reference Shah and Oppenheimer2008). Consider the example of different types of moral intuitions and related candidate domain-specific heuristics. Deontologists (e.g., Kant Reference Kant and Ellington1785/1993) claim that certain actions are obviously wrong (e.g., torture) or right (e.g., saving lives). They also sometimes believe that it is an inherent feature of a certain action to be wrong. Most of the time, this strategy works well, but sometimes it misfires and/or leads to problematic inferences (e.g., it is a moral imperative not to lie even to a serial killer).
However, it would be false to assume that this is a feature of only one type of moral intuitions and heuristics. Consequentialists (e.g., Sidgwick Reference Sidgwick1874/1981) use similar reasoning when they focus on consequences (e.g., saving five people as opposed to just one), regardless of the means used or the character traits of the agents involved. They also sometimes believe that it is an inherent feature of a certain consequence to be good, which again can misfire as a strategy and/or lead to morally abhorrent conclusions (e.g., killing one healthy person to harvest organs to save five people is rational).
Even virtue theorists (e.g., Aristotle 1890) use similar reasoning to make their case. Namely, they focus on the character – the habits, intentions, and motives of the agent (e.g., is this something appropriate for a virtuous person?). They also sometimes believe that it is an inherent feature of a virtuous person to react morally appropriately in every situation. Again, this might be a sound strategy most of the time, but it can certainly misfire and/or create moral outrage (e.g., if Jesus/Mohammed/Moses/Buddha/Marx responded in such a way, everyone should).
The so-called inherence heuristic could actually be a domain-general process that might inform a range of domain-specific heuristics. However, this needs to be acknowledged to ensure its integration into the rich body of theoretical and empirical moral psychology. Furthermore, it is wrong to assume that the is–ought divide is insurmountable and that fallibility inherently means faultiness. If the authors maintain such assumptions, they are falling prey to a process they themselves identified.