Joshua Knobe's novel theory of how moral judgments influence people's intuitions about certain apparently non-moral notions, including intentionality and causation, is a version of contrastivism (see Sinnott-Armstrong Reference Sinnott-Armstrong2008). As with many other topics (including knowledge, free will, explanation, and morality), it is illuminating to consider the range of possible contrasts or alternatives and ask when and why people limit their attention to a smaller contrast class. So I am very sympathetic. I would, however, like to press Knobe to develop three aspects of his theory.
First, Knobe formulates his “general approach” in his section 5.1 in terms of relevant alternatives. Moral judgments are said to affect which counterfactual alternatives are seen or treated as relevant. Next, Knobe discusses his “case study” in his section 5.2 in terms of defaults on a continuum. Moral judgments are said to affect the position of the default. These views are not equivalent, because alternatives need not always fall on a continuum, and relevant alternatives might fall on either side of a default. Knobe describes the default as “a particular sort of alternative possibility,” but it does not seem to be the only relevant alternative, so comparing a default and comparing a range of alternative possibilities seem quite different. My question for Knobe is then: What exactly is the relation between these two theories?
Second, it is also not clear what the relation is between either of Knobe's suggestions and the views against which he has argued in the earlier parts of his article. Motivation, blame, emotion, and pragmatic context would seem to be promising candidates for explaining why we treat certain alternatives rather than others as relevant, or why we place the default at one point instead of another on a continuum. If Knobe agrees, then his own theory, though a crucial part of the story, would need to be supplemented by central aspects of the views he criticizes. His theory then works together with his supposed opponents, rather than supplanting them. But if Knobe denies that these features explain why we adopt certain relevant alternatives and defaults, then we need an alternative explanation of relevance and default. It is not enough to refer to alternatives and defaults without explaining how the alternatives and defaults get set.
Addressing this issue, Knobe says, “all sorts of different factors can play a role here” (sect. 5.1, para. 5). This is surely right, and he cites supporting literature. However, it leaves open the possibility that motivation, blame, emotions, and pragmatic context do sometimes play roles in determining which alternatives we see as relevant and where we place the default. No theory that focuses on one single factor can or should be expected to cover all examples, even if each factor does explain some variance in some areas. Hence, I also want to ask Knobe whether his arguments are supposed to show that motivation, blame, emotions, and pragmatic context do not always play a role or do not ever play a role in determining relevant alternatives and defaults.
Third, after he criticizes his opponents for failing “to produce any positive evidence in favor of the hypothesis” (sect. 4.1.4, para. 1), I would like to see Knobe's positive evidence in favor of his own hypotheses. Consider first his claim about relevant alternatives, and focus on his example of what caused the dent in the car. He suggests that, when we ask whether a certain person, object, or event caused a certain effect, we “think about,” “consider,” “compare,” and “pick out just certain specific alternatives” (sect. 5.1, para. 4). As Knobe says, we do not “consider the possibility that the car could have been levitating in the air.” But what is the positive evidence that we do consider other alternatives?
I do not deny that we treat a range of alternatives as relevant. However, it is not clear whether we actually represent these alternatives, even unconsciously. Another possibility is that we have a disposition to dismiss certain alternatives as irrelevant, if raised, and to accept other alternatives as relevant, if raised; but we never explicitly “think about” or “consider” the relevant alternatives any more than the irrelevant alternatives unless prompted.
How can we decide between these views? Perhaps we could get evidence that subjects consider or think about certain alternatives by asking the subjects, but self-report would not be reliable. Another method would be to measure subjects' memory errors, word completion patterns, or reaction times when asked whether certain words were in the scenarios. If subjects really do consider an alternative that would naturally be formulated in certain terms that were not actually in the scenario, then we would expect them to be more likely to misremember those terms as being in the scenario, to complete letter strings so as to form those terms, and to unscramble the letters from those terms more quickly than if they never considered that alternative.
However, before we can apply these techniques, we would need to formulate specific hypotheses about which alternatives are and are not thought about or considered in which scenarios. It is not enough simply to say that moral judgments affect the range of alternatives that are taken to be relevant. We need to know which alternatives are supposed to be seen as relevant. Only then can we test whether those alternatives are actually considered or thought about, as Knobe claims.
The same basic issue arises for Knobe's theory that moral judgments affect which point on a continuum is seen as the default. Certain hypotheses might seem plausible and might have explanatory power, but it is not easy to figure out how to gather positive evidence for the hypothesis that people actually set different defaults depending on their moral judgments. So my last question for Knobe is: How will you get positive evidence for your claims that moral judgments affect relevant alternatives and defaults?
You have
Access
Joshua Knobe's novel theory of how moral judgments influence people's intuitions about certain apparently non-moral notions, including intentionality and causation, is a version of contrastivism (see Sinnott-Armstrong Reference Sinnott-Armstrong2008). As with many other topics (including knowledge, free will, explanation, and morality), it is illuminating to consider the range of possible contrasts or alternatives and ask when and why people limit their attention to a smaller contrast class. So I am very sympathetic. I would, however, like to press Knobe to develop three aspects of his theory.
First, Knobe formulates his “general approach” in his section 5.1 in terms of relevant alternatives. Moral judgments are said to affect which counterfactual alternatives are seen or treated as relevant. Next, Knobe discusses his “case study” in his section 5.2 in terms of defaults on a continuum. Moral judgments are said to affect the position of the default. These views are not equivalent, because alternatives need not always fall on a continuum, and relevant alternatives might fall on either side of a default. Knobe describes the default as “a particular sort of alternative possibility,” but it does not seem to be the only relevant alternative, so comparing a default and comparing a range of alternative possibilities seem quite different. My question for Knobe is then: What exactly is the relation between these two theories?
Second, it is also not clear what the relation is between either of Knobe's suggestions and the views against which he has argued in the earlier parts of his article. Motivation, blame, emotion, and pragmatic context would seem to be promising candidates for explaining why we treat certain alternatives rather than others as relevant, or why we place the default at one point instead of another on a continuum. If Knobe agrees, then his own theory, though a crucial part of the story, would need to be supplemented by central aspects of the views he criticizes. His theory then works together with his supposed opponents, rather than supplanting them. But if Knobe denies that these features explain why we adopt certain relevant alternatives and defaults, then we need an alternative explanation of relevance and default. It is not enough to refer to alternatives and defaults without explaining how the alternatives and defaults get set.
Addressing this issue, Knobe says, “all sorts of different factors can play a role here” (sect. 5.1, para. 5). This is surely right, and he cites supporting literature. However, it leaves open the possibility that motivation, blame, emotions, and pragmatic context do sometimes play roles in determining which alternatives we see as relevant and where we place the default. No theory that focuses on one single factor can or should be expected to cover all examples, even if each factor does explain some variance in some areas. Hence, I also want to ask Knobe whether his arguments are supposed to show that motivation, blame, emotions, and pragmatic context do not always play a role or do not ever play a role in determining relevant alternatives and defaults.
Third, after he criticizes his opponents for failing “to produce any positive evidence in favor of the hypothesis” (sect. 4.1.4, para. 1), I would like to see Knobe's positive evidence in favor of his own hypotheses. Consider first his claim about relevant alternatives, and focus on his example of what caused the dent in the car. He suggests that, when we ask whether a certain person, object, or event caused a certain effect, we “think about,” “consider,” “compare,” and “pick out just certain specific alternatives” (sect. 5.1, para. 4). As Knobe says, we do not “consider the possibility that the car could have been levitating in the air.” But what is the positive evidence that we do consider other alternatives?
I do not deny that we treat a range of alternatives as relevant. However, it is not clear whether we actually represent these alternatives, even unconsciously. Another possibility is that we have a disposition to dismiss certain alternatives as irrelevant, if raised, and to accept other alternatives as relevant, if raised; but we never explicitly “think about” or “consider” the relevant alternatives any more than the irrelevant alternatives unless prompted.
How can we decide between these views? Perhaps we could get evidence that subjects consider or think about certain alternatives by asking the subjects, but self-report would not be reliable. Another method would be to measure subjects' memory errors, word completion patterns, or reaction times when asked whether certain words were in the scenarios. If subjects really do consider an alternative that would naturally be formulated in certain terms that were not actually in the scenario, then we would expect them to be more likely to misremember those terms as being in the scenario, to complete letter strings so as to form those terms, and to unscramble the letters from those terms more quickly than if they never considered that alternative.
However, before we can apply these techniques, we would need to formulate specific hypotheses about which alternatives are and are not thought about or considered in which scenarios. It is not enough simply to say that moral judgments affect the range of alternatives that are taken to be relevant. We need to know which alternatives are supposed to be seen as relevant. Only then can we test whether those alternatives are actually considered or thought about, as Knobe claims.
The same basic issue arises for Knobe's theory that moral judgments affect which point on a continuum is seen as the default. Certain hypotheses might seem plausible and might have explanatory power, but it is not easy to figure out how to gather positive evidence for the hypothesis that people actually set different defaults depending on their moral judgments. So my last question for Knobe is: How will you get positive evidence for your claims that moral judgments affect relevant alternatives and defaults?