Consider Joshua Knobe and the so-called "Knobe effect." Knobe is the high-priest of x-phi (or "experimental philosophy") who, in 2003, somehow induced the editors of Analysis to sully themselves by publishing the results of an empirical study in their respected philosophy journal.* The study purported to show something about the ordinary moral judgment of ordinary people that might be suprising. In Knobe's telling, his study seemed to show that "there seems to be an asymmetry whereby people are considerably more willing to blame the agent for bad side effects than to praise the agent for good side effects."** But I think Knobe mis-interpreted his own results, and, in this post, I want to explain why.
First, I'm going to quote from a friend's description of the experiment:
Knobe presented his questions to 78 people spending time in a Manhattan public park. Each subject was randomly assigned to either the 'harm condition' or the 'help condition'. Subjects in the harm condition read the following vignette
The vice-president of a company went to the chairman of the board and said, 'We are thinking of starting a new program. It will help us increase profits, but it will also harm the environment.' The chairman of the board answered, 'I don't care at all about harming the environment. I just want to make as much profit as I can. Let's start the new program.' They started the new program. Sure enough, the environment was harmed.
These subjects were then asked to determine how much blame the chairman deserved for what he did (on a scale from 0 to 6) and to say whether they thought the chairman intentionally harmed the environment. In the harm condition, most subjects (82%) said that the agent brought about the side effect (harming the environment) intentionally.
Subjects in the help condition received a vignette that was almost exactly the same, except that the word 'harm' was replaced by 'help':
The vice-president of a company went to the chairman of the board and said, 'We are thinking of starting a new program. It will help us increase profits, and it will also help the environment.' The chairman of the board answered, 'I don't care at all about helping the environment. I just want to make as much profit as I can. Let's start the new program.' They started the new program. Sure enough, the environment was helped.
These subjects were then asked to determine how much praise the chairman deserved (on a scale from 0 to 6) and whether they thought the chairman intentionally helped the environment. As predicted, the two conditions elicited two radically different patterns of responses. In the harm condition, most subjects (82%) said that the agent brought about the side effect intentionally, whereas in the help condition, most subjects (77%) said that the agent did not bring about the side effect intentionally!
In drawing the conclusion he does, Knobe implicitly depends on the assumption that the helping and harming cases are, in fact, morally symmetrical. The interest of his "asymmetrical" result depends critically on its being the case that subjects are not simply being appropriately sensitive to real moral differences between the helping and harming conditions. If we were convinced that the subjects's overall pattern of judgment was morally appropriate or essentially rational, why would it be interesting to call it an "asymmetrical" pattern? The interest in Knobe's conclusion is in the (implicit) suggestion that ordinary moral intuition is deeply irrational and inconsistent-- in particular, that it is inappropriately sensitive to features which couldn't possibly amount to a real moral difference.
What I want to suggest, however, is that the two situations--the helping and harming conditions--in fact are not morally symmetrical from the viewpoint of the praise- or blame- worthiness of the chairman of the board. And because they are not, the experiment simply fails to show the sort of irrational or iconsistent "asymmetry" that Knobe supposes. Knobe, in other words, is not licensed to draw the inference he does.
This may seem like a very strong conclusion on my part. After all, the two conditions were deliberately constructed so that the only difference between them is that one helps and one harms the environment. Where we are assuming equal amounts of help or harm, and where we are assuming that the valuable thing in question (the environment) makes equal moral demands to be helped or harmed (per unit of help or harm, respectively), where is the room for a real moral difference? (Incidentally, I don't question either of these two assumptions. They seem both plausible and a part of the original though-experiment.)
Despite this deliberate attempt to create moral symmetry between the two conditions, however, a moral asymmetry nevertheless crept in. The chairman in the first hypo is revealed as a person who is willing to destroy something of value, and the chairman in the second hypo is revealed as a person who is indifferent to something of value. The moral difference between the characters revealed in the respective hypos is the difference between Nazi Germany and Switzerland; it is the difference between an agent who is unmistakably (given his actual actions) willing to harm versus an agent who is (given his actual actions) possibly merely indifferent to either help or harm. This is a Real Moral Difference in the agent's culpable mental state, or mens rea. (Mens rea -- it ain't just fer' criminal law.)
Despite this deliberate attempt to create moral symmetry between the two conditions, however, a moral asymmetry nevertheless crept in. The chairman in the first hypo is revealed as a person who is willing to destroy something of value, and the chairman in the second hypo is revealed as a person who is indifferent to something of value. The moral difference between the characters revealed in the respective hypos is the difference between Nazi Germany and Switzerland; it is the difference between an agent who is unmistakably (given his actual actions) willing to harm versus an agent who is (given his actual actions) possibly merely indifferent to either help or harm. This is a Real Moral Difference in the agent's culpable mental state, or mens rea. (Mens rea -- it ain't just fer' criminal law.)
In general, I think it's important not to get distracted by the form of words in the hypos. The asymmetry I'm suggesting is not about the amounts of help or harm to the environment, or about any inequality in the strengths of the demands to help or not harm. Attend instead to the complete picture evoked by the words in each case, and especially, to the sort of agent revealed in that picture. Do that, I think, and you'll see that there is an asymmetry in kind of agent pictured, and that it's not about help or harm per se. The asymmetry is about the sort of mens rea we attribute to the agent, and those mens reas are not symmetrical in any sense.
So Knobe's subjects, in fact, are arguably simply being sensitive to the real moral differences in the respective situations. If so, then there is not much mystery, and we need not think that this is a case of "asymmetry" in willingness to ascribe praise or blame dependent on whether the side-effects are good or bad. At least, thinking this would require sustaining a further non-trivial moral argument that the ordinary way of connecting praise- or blame- worthiness to mens rea is improper.
So Knobe's subjects, in fact, are arguably simply being sensitive to the real moral differences in the respective situations. If so, then there is not much mystery, and we need not think that this is a case of "asymmetry" in willingness to ascribe praise or blame dependent on whether the side-effects are good or bad. At least, thinking this would require sustaining a further non-trivial moral argument that the ordinary way of connecting praise- or blame- worthiness to mens rea is improper.
Incidentally, none of what I'm suggesting will make sense if you insist on understanding praise- and blame- worthiness exclusively in terms of causing harm or causing help. In that case, you would be unable to make sense of the difference in mens rea I'm suggesting, since, by hypothesis, there is no difference in the amount of harm or help. But, as far as I can tell, that's simply a very good reason to abandon the consequentialist account of praise- and blame- worthiness as an account which can't properly encompass the relevant moral sense of things we actually have.
One final thought. The results of an "x-phi" experiment could only reveal interesting philosophical results if the results are correctly interpreted. But the question of correct interpretation is itself a conceptual question, at least in part. So, it would seem that you really can't escape the armchair in philosophy.
----
*- Experimental philosophy, or x-phi, bugs me. Understood as part of a tradition of clever empirical psychology or sociology, I have no objection. Understood as part of an essentially philosophical tradition, I view it, in my gentler moods, as a vicious betrayal. Where's the conceptual insight? Where's the concern for (in Aristotle's excellent formulation) being qua being? Or (in my preferred formulation) for thought qua thought? Where's the armchair attempt to tell how possibly some x might be the case? X-phi trades in all of that classically philosophical stuff for a mere empirical science. It's 'philosophy' for weaklings who don't know the difference between philosophy and psychology, or who anyway don't care to keep the distinction clean, even in thought. Treated as philosophy, it's a rank confusion.
** - Knobe offered a second sub-conclusion to the effect that people are more willing to ascribe intention in the case of bad effects than in the case of good effects. This result is less interesting to me. If I were going to hazard a comment here, however, it would be to suggest that in the particular context of the experiment -- particularly, in a context in which the subject is being called upon to form some kind of moral opinion about a person in a hypo -- the ordinary, non-philosophical person may well have treated the concept of "intention" as an essentially moral concept, and not as (as the researcher supposed) a concept to do with a certain kind of (essentially morally nuetral) mental state. Anyway, it seems plausible to me to suppose that many of the subjects may have thought that they were supposed to use the word 'intentionally' especially to mark out the degree of praise- or blame- worthiness of the hypothetical chairman, and not to mark out the intentionality of action strictly speaking. After all, subjects were being called upon to make some moral judgment, and many may have assumed that the question about "intentionality" was simply meant to be a further question about how praise- or blame- worthy they thought the hypothetical agent was. In other words, many of the subjects may have understood the researcher to mean, by "intentionality," "especially praise- or blame- worthy." And because most thought (for reasons explained above) that the harming chairman was especially blameworthy but that the helping chairman was not especially praiseworthy, they would have diverged in their ascription of intentionality, if they did understand 'intentional' in this way.
** - Knobe offered a second sub-conclusion to the effect that people are more willing to ascribe intention in the case of bad effects than in the case of good effects. This result is less interesting to me. If I were going to hazard a comment here, however, it would be to suggest that in the particular context of the experiment -- particularly, in a context in which the subject is being called upon to form some kind of moral opinion about a person in a hypo -- the ordinary, non-philosophical person may well have treated the concept of "intention" as an essentially moral concept, and not as (as the researcher supposed) a concept to do with a certain kind of (essentially morally nuetral) mental state. Anyway, it seems plausible to me to suppose that many of the subjects may have thought that they were supposed to use the word 'intentionally' especially to mark out the degree of praise- or blame- worthiness of the hypothetical chairman, and not to mark out the intentionality of action strictly speaking. After all, subjects were being called upon to make some moral judgment, and many may have assumed that the question about "intentionality" was simply meant to be a further question about how praise- or blame- worthy they thought the hypothetical agent was. In other words, many of the subjects may have understood the researcher to mean, by "intentionality," "especially praise- or blame- worthy." And because most thought (for reasons explained above) that the harming chairman was especially blameworthy but that the helping chairman was not especially praiseworthy, they would have diverged in their ascription of intentionality, if they did understand 'intentional' in this way.
Very good, Michael. I think the asymmetry you suggest is there and that it is what the subjects are sensitive to. I would just remind you, with respect to the anti-consequentialist bit at the end, that not all consequentialists are "actual consequence" consequentialists. (Actually, consequentialist's ought not be actual consequence consequentialists since the view is so grossly counterintuitive). I think a "value adjusted possible consequence" consequentialist way of assessing praise and blame worthiness could fit with your point about mens rea.
ReplyDeleteOkay, you have to explain "value adjusted possible consequence" consequentialist a bit, I'm afraid. Is the thought that, actually, the consequences in the two cases are assymetrical, for one reason or another? That could save the rationality of the "assymetry" in ascribing praise or blame even on the consequentialist picture, except that I was trying to assume that the consequences were -- all relevant adjustments considerered -- equivalent between the helping and harming condition.
ReplyDeleteIncidentally, though, I'm not so sure I can assume this equivalence in accounting for the difference in mens rea. I'm not completely sure that the sense of difference in mens rea is independent of the sense we (might?) have that the duty to not harm is somehow stronger than the duty to help. I said above that I was assuming that the demand to not harm was equal to the demand to help. On reflection though, I'm not so sure about this.
(I remain sure, though, that there really is a moral assymetry in the two situations that is reflected in an idea of different mens reas. The question is how we might account then for that different in mens rea and whether, despite my first assumption, we actually need the idea of asymmetrical duties or some such.)
In assessing the rightness or wrongness of an action and, relatedly, the praiseworthiness or blameworthiness of the agent (as consequentialists) we don't let the ACTUAL consequences do the work. If we did, we'd have to say that the guy shooting randomly into a crowd of innocents who just so happens to miss everyone but one individual (not an innocent) who was seconds from detonating a nuclear bomb did the right and praiseworthy thing. He maximized goodness! Instead, the consequences that do the work for us are the value adjusted possible consequences: We expect agents to have taken (to the extent reasonably possible) a survey of the possible outcomes (consequences) of the different courses of action that are available at some given time and act appropriately. But, we also want agents to be rationally open to pursuing an action that has some very, very low probability of resulting in a somewhat significant negative outcome if it also has a very high probability of resulting in something really, really good. (We also want agents to avoid actions that have even a low probability of severe, devastating, really bad consequences - assuming they are bad enough). Thus, we want agents to adjust the value assigned (positive or negative) to each of the possible consequences in accordance with the probability that the outcome will result. In this way, we can reserve room for facts about the agent's intention (or motive or will) in acting in our overall moral assessment. Its not just Kantians that get to be concerned about agent's will. My point was that I thought a more sophisticated consequentialism would allow us to be appropriately concerned with the agent's culpable mental state. (Did I miss your point in saying the view doesn't make sense from within a consequentialist framework?)
ReplyDeleteGood. But I did mean to suggest, originally, that the value-adjusted possibilities in question (harming or helping the environment) were equivalent. If that is assumed, then even on this view of consequentialism (which seems exactly the right version to me), it can't make sense of a relevant difference in mens rea in the case in question because there is (on its account of blame) no relevant difference in intention. (And the thought-experiment is set up, of course, so that the actual results are completely predictable and known in advance.)
ReplyDeleteSo, I guess my thought is that, in this case, even the value-adjusted possible consequence view doesn't account for our sense of assymetry in blameworthiness, given that the value-adjusted possible consequences are symmetrical in either case, and predictably known to be so in advance. I suspect that the mens rea difference might have to do with an assymetry in duties toward the environment (the duty to not harm is stronger than the duty to help)-- but whether this thought can (on current assumptions) be maintained depends on whether an account can be given of the duties which is independent of the consequences.
Still, notice in any case that determining what the 'knobe effect' amounts to requires doing lots of old-fashioned arm-chair work!