A debunker's dilemma
So far the skeptical threat looks to be motivated by a genealogical debunking argument, in which some beliefs or other attitudes are allegedly influenced by illicit processes. Nietzsche and Freud, for example, famously attacked ordinary moral and religious beliefs as being influenced by wishful thinking, egoism, and rationalization. More recently, some philosophers argue that ordinary moral beliefs are unjustified because they have been too heavily shaped by extraneous evolutionary forces (e.g., Joyce Reference Joyce2006). Whatever the targeted attitudes, such debunking arguments intend to reveal that the attitudes are problematic because they're in fact substantially influenced by unseemly forces (see Nichols Reference Nichols2014).
Genealogical debunking arguments can be made to work, particularly when informed by the relevant empirical research, provided they aren't too wide-ranging (Kumar & May, Reference Kumar and Mayunder review). Empirical evidence can reveal that some of our decisions are influenced by arbitrary factors, but it's more difficult to establish that most of our behavior is so influenced. Indeed, there is plenty of evidence that many of our choices are influenced by good reasons (see e.g., discussion in Batson Reference Batson2011; Miller Reference Miller2013; Seligman et al. Reference Seligman, Railton, Baumeister and Sripada2016).
With evidence on both sides, it looks like we ultimately have to do the hard work of determining what does drive most of our decision-making. Yet Doris can't just show that our choices are slightly influenced by arbitrary factors. That would leave room for reflective direction or for being influenced by unreflective yet appropriate factors. We must ask: Are (a) most of our choices (b) substantially driven by (c) genuinely arbitrary factors?
My own view is that a close examination of the empirical literature suggests otherwise (May, Reference Mayforthcoming). Certainly, some of the influences on our choices are truly unwelcome and arbitrary. For example, our choices certainly shouldn't be determined by racial or gender bias, irrelevant feelings of disgust, or the mere order in which information is presented. But a comprehensive look at the literature, including meta-analyses, suggests such arbitrary influences are often rather small (see e.g., Oswald et al. Reference Oswald, Mitchell, Blanton, Jaccard and Tetlock2013; Landy & Goodwin Reference Landy and Goodwin2015; Demaree-Cotton Reference Demaree-Cotton2016). This leaves plenty of room for being motivated primarily by the right reasons. For example, feeling queasy from food poisoning might make one think stealing is slightly worse than one would judge otherwise (May Reference May2014). But the reason most people don't embezzle from their employer is because they think it's unfair, harmful, disrespectful, or just plain immoral.
Other choices are substantially influenced by various factors, such as group size, ambient smells, being in a hurry, similarity to a victim, and honor codes (see e.g., Latané & Nida Reference Latané and Nida1981; Carlson et al. Reference Carlson, Charlin and Miller1988; Batson Reference Batson2011; Ariely Reference Ariely2012). But these forces, while often powerful, aren't necessarily something we'd reject as non-reasons once we examine the effects in more detail. Consider, for example, being in a hurry, being in a good or bad mood, or being reminded of one's moral commitments. We may be happy to cite these as genuine reasons for either helping or not helping a stranger in minor need. Imagine: “Why didn't I stop to help that man pick up his dropped papers? I was in a hurry and I'm just not in the mood to talk to anyone right now.” Even when a stranger's situation appears to be dire, one has good reason not to help if one infers that no real help is needed because someone else will do it or everyone else who hears what's going on isn't helping. Even if one recognizes another is in serious need, feeling compassion may be a good reason for help. I may empathize more because the victim and I share a similar background and gender, but that may just draw my attention to a relevant reason to help (e.g., he's in serious need).
This isn't just ad hoc whack-a-mole. There may well be a general dilemma here for wide-ranging debunkers like Doris: Influences on many choices tend to be either substantial or arbitrary but not commonly both. Indeed, some influences may turn out to be neither substantial nor arbitrary. Eye spots, for example, aren't necessarily arbitrary, for they may serve as a moral reminder that draws one's attention to reasons for being fair and honest (same goes for honor codes and the like). And one meta-analysis of 25 studies suggests the eyes-effect is small and quickly diminishes (Sparks & Barclay Reference Sparks and Barclay2013). Either way, at least one of the conditions for a debunking argument isn't met.
Skeptical hypotheses
Doris might avoid this dilemma by reframing the ambitions of his argument and thus its explanatory burdens. At one point he does explicitly model his approach on skeptical hypotheses in epistemology meant to undermine knowledge of the external world (Doris Reference Doris2015b, p. 65). Such perceptual skeptics argue that your evidence would be the same if you were hallucinating or being fed fake experiences by a Cartesian evil demon. Because you can't rule out the possibility that your experiences are systematically deceiving, you don't know there is a physical world beyond your senses (Brueckner Reference Brueckner1994). The idea is not at all that this grand skeptical scenario is actual, only that it's possible.
Doris accordingly thinks he has only to raise the mere possibility of a skeptical scenario. There is a “large, odorous, and ill-tempered animal under the awning of agency,” he writes, and thus “for all one knows, any decision may be infested by any number of rationally and ethically arbitrary influences” (2015b, p. 64). For Doris, the “critical question concerns not how often defeaters should be thought to obtain, but how their presence can be ruled out” (2015b, p. 68).
However, if this is the form of argument, then we didn't need all of the empirical evidence. Imagination alone can generate hypothetical scenarios in which it systematically seems we're motivated by good reasons though we're not. Moreover, while philosophers have long been fascinated with this form of argument, it's not necessarily because they find it compelling.
Perhaps Doris's idea is that his skeptical hypothesis argument should be more persuasive because there is some positive scientific evidence that the skeptical scenario is actual (compare Sinnott-Armstrong Reference Sinnott-Armstrong2006). Nevertheless, skeptical hypothesis arguments make no claims about the actual genealogy of the relevant mental states. The only empirical claim in such arguments is about the actual character of one's evidence – namely, that it can't rule out the skeptical scenario – but this isn't a claim about the source of one's attitudes (May Reference May2013). So, given Doris's explicit and extensive appeal to evidence that our choices and decisions are in fact influenced by arbitrary factors, he seems to be offering a wide-ranging debunking argument.
Doris presumably requires a debunking argument anyway for his purposes. Skeptical hypothesis arguments lead to a sweeping denial of knowledge. Perceptual skeptics conclude that we don't know there's an external world; Doris concludes we don't know our behavior is defeater-free. Such negative conclusions cut both ways: We neither know that we are, nor that we aren't, perceiving an external world or acting for good reasons. Doris, however, aims to establish the positive claim that much of our behavior is in fact influenced by defeaters. His rejection of reflectivism, for example, relies on knowledge of the influences on our actions. As he says, the argument “gets its bite from a family of empirical observations indicating that reflection does not [in fact] play the sort of role in self-direction that reflectivism supposes” (2015b, p. 33). Moreover, Doris's own dialogic theory of agency is motivated by such claims about the actual springs of human action. So it seems Doris needs to do more than raise the specter of malodorous influences on our choices.
You have
Access
John Doris (Reference Doris2015b) argues forcefully and eloquently that human thought and action aren't quite what they seem. He deftly points to empirical research that suggests that our actions are commonly influenced by a wealth of unconscious and, importantly, unseemly factors: “Many studies identify causes of behavior that are not plausibly taken as reasons for behavior” (p. 43). One of his favorite examples is the finding that people appear to cheat less when there is a depiction of eyes watching them (e.g., Bateson et al. Reference Bateson, Nettle and Roberts2006). Few people would happily say “I did it because of the eye spots” (Doris Reference Doris2015b, p. 43).
Human actions, Doris concludes, are often driven by unconscious and unreflective processes that amount to “defeaters.” These are influences “the actor is unaware of, and would not recognize as a reason justifying the behavior, were she so aware” (Doris Reference Doris2015b, p. 52). Doris doesn't quite give this view a label. He just associates it with a “skeptical threat,” which he thinks we can avoid by adopting his own preferred theory of agency.
Defeaters are supposed to be particularly damaging to a “reflectivist” tradition, which holds that human thought and action are normally guided by “accurate reflection” on one's own mental states. Now, various other philosophers and scientists, including myself, likewise believe that the human mind isn't so reflective (e.g., Watson Reference Watson1975; Arpaly Reference Arpaly2003; Seligman et al. Reference Seligman, Railton, Baumeister and Sripada2016; May, Reference Mayforthcoming). However, Doris also doubts the “accurate” bit, suggesting that we often aren't motivated by what we'd regard as genuine reasons. Here I want to suggest that the skeptical threat has been overstated, and yet a weakened version isn't enough for Doris's ambitious purposes.
A debunker's dilemma
So far the skeptical threat looks to be motivated by a genealogical debunking argument, in which some beliefs or other attitudes are allegedly influenced by illicit processes. Nietzsche and Freud, for example, famously attacked ordinary moral and religious beliefs as being influenced by wishful thinking, egoism, and rationalization. More recently, some philosophers argue that ordinary moral beliefs are unjustified because they have been too heavily shaped by extraneous evolutionary forces (e.g., Joyce Reference Joyce2006). Whatever the targeted attitudes, such debunking arguments intend to reveal that the attitudes are problematic because they're in fact substantially influenced by unseemly forces (see Nichols Reference Nichols2014).
Genealogical debunking arguments can be made to work, particularly when informed by the relevant empirical research, provided they aren't too wide-ranging (Kumar & May, Reference Kumar and Mayunder review). Empirical evidence can reveal that some of our decisions are influenced by arbitrary factors, but it's more difficult to establish that most of our behavior is so influenced. Indeed, there is plenty of evidence that many of our choices are influenced by good reasons (see e.g., discussion in Batson Reference Batson2011; Miller Reference Miller2013; Seligman et al. Reference Seligman, Railton, Baumeister and Sripada2016).
With evidence on both sides, it looks like we ultimately have to do the hard work of determining what does drive most of our decision-making. Yet Doris can't just show that our choices are slightly influenced by arbitrary factors. That would leave room for reflective direction or for being influenced by unreflective yet appropriate factors. We must ask: Are (a) most of our choices (b) substantially driven by (c) genuinely arbitrary factors?
My own view is that a close examination of the empirical literature suggests otherwise (May, Reference Mayforthcoming). Certainly, some of the influences on our choices are truly unwelcome and arbitrary. For example, our choices certainly shouldn't be determined by racial or gender bias, irrelevant feelings of disgust, or the mere order in which information is presented. But a comprehensive look at the literature, including meta-analyses, suggests such arbitrary influences are often rather small (see e.g., Oswald et al. Reference Oswald, Mitchell, Blanton, Jaccard and Tetlock2013; Landy & Goodwin Reference Landy and Goodwin2015; Demaree-Cotton Reference Demaree-Cotton2016). This leaves plenty of room for being motivated primarily by the right reasons. For example, feeling queasy from food poisoning might make one think stealing is slightly worse than one would judge otherwise (May Reference May2014). But the reason most people don't embezzle from their employer is because they think it's unfair, harmful, disrespectful, or just plain immoral.
Other choices are substantially influenced by various factors, such as group size, ambient smells, being in a hurry, similarity to a victim, and honor codes (see e.g., Latané & Nida Reference Latané and Nida1981; Carlson et al. Reference Carlson, Charlin and Miller1988; Batson Reference Batson2011; Ariely Reference Ariely2012). But these forces, while often powerful, aren't necessarily something we'd reject as non-reasons once we examine the effects in more detail. Consider, for example, being in a hurry, being in a good or bad mood, or being reminded of one's moral commitments. We may be happy to cite these as genuine reasons for either helping or not helping a stranger in minor need. Imagine: “Why didn't I stop to help that man pick up his dropped papers? I was in a hurry and I'm just not in the mood to talk to anyone right now.” Even when a stranger's situation appears to be dire, one has good reason not to help if one infers that no real help is needed because someone else will do it or everyone else who hears what's going on isn't helping. Even if one recognizes another is in serious need, feeling compassion may be a good reason for help. I may empathize more because the victim and I share a similar background and gender, but that may just draw my attention to a relevant reason to help (e.g., he's in serious need).
This isn't just ad hoc whack-a-mole. There may well be a general dilemma here for wide-ranging debunkers like Doris: Influences on many choices tend to be either substantial or arbitrary but not commonly both. Indeed, some influences may turn out to be neither substantial nor arbitrary. Eye spots, for example, aren't necessarily arbitrary, for they may serve as a moral reminder that draws one's attention to reasons for being fair and honest (same goes for honor codes and the like). And one meta-analysis of 25 studies suggests the eyes-effect is small and quickly diminishes (Sparks & Barclay Reference Sparks and Barclay2013). Either way, at least one of the conditions for a debunking argument isn't met.
Skeptical hypotheses
Doris might avoid this dilemma by reframing the ambitions of his argument and thus its explanatory burdens. At one point he does explicitly model his approach on skeptical hypotheses in epistemology meant to undermine knowledge of the external world (Doris Reference Doris2015b, p. 65). Such perceptual skeptics argue that your evidence would be the same if you were hallucinating or being fed fake experiences by a Cartesian evil demon. Because you can't rule out the possibility that your experiences are systematically deceiving, you don't know there is a physical world beyond your senses (Brueckner Reference Brueckner1994). The idea is not at all that this grand skeptical scenario is actual, only that it's possible.
Doris accordingly thinks he has only to raise the mere possibility of a skeptical scenario. There is a “large, odorous, and ill-tempered animal under the awning of agency,” he writes, and thus “for all one knows, any decision may be infested by any number of rationally and ethically arbitrary influences” (2015b, p. 64). For Doris, the “critical question concerns not how often defeaters should be thought to obtain, but how their presence can be ruled out” (2015b, p. 68).
However, if this is the form of argument, then we didn't need all of the empirical evidence. Imagination alone can generate hypothetical scenarios in which it systematically seems we're motivated by good reasons though we're not. Moreover, while philosophers have long been fascinated with this form of argument, it's not necessarily because they find it compelling.
Perhaps Doris's idea is that his skeptical hypothesis argument should be more persuasive because there is some positive scientific evidence that the skeptical scenario is actual (compare Sinnott-Armstrong Reference Sinnott-Armstrong2006). Nevertheless, skeptical hypothesis arguments make no claims about the actual genealogy of the relevant mental states. The only empirical claim in such arguments is about the actual character of one's evidence – namely, that it can't rule out the skeptical scenario – but this isn't a claim about the source of one's attitudes (May Reference May2013). So, given Doris's explicit and extensive appeal to evidence that our choices and decisions are in fact influenced by arbitrary factors, he seems to be offering a wide-ranging debunking argument.
Doris presumably requires a debunking argument anyway for his purposes. Skeptical hypothesis arguments lead to a sweeping denial of knowledge. Perceptual skeptics conclude that we don't know there's an external world; Doris concludes we don't know our behavior is defeater-free. Such negative conclusions cut both ways: We neither know that we are, nor that we aren't, perceiving an external world or acting for good reasons. Doris, however, aims to establish the positive claim that much of our behavior is in fact influenced by defeaters. His rejection of reflectivism, for example, relies on knowledge of the influences on our actions. As he says, the argument “gets its bite from a family of empirical observations indicating that reflection does not [in fact] play the sort of role in self-direction that reflectivism supposes” (2015b, p. 33). Moreover, Doris's own dialogic theory of agency is motivated by such claims about the actual springs of human action. So it seems Doris needs to do more than raise the specter of malodorous influences on our choices.
Conclusion
In the end, I think Doris is quite right that much of our behavior is determined by unreflective processes – certainly more than commonsense suggests. But only some of our choices are substantially determined by unsavory causes. Of course, we should still pay close attention to these. Even small biases can add up, generating large social problems. What's much less clear is whether small or rare biases warrant overhauls in our conception of human agency and moral responsibility.