The Natural Unconscious

Does Kanye West believe in free will? Yes and no...

John A. Bargh, Ph.D. — Fri, 18 Sep 2009 16:25:38 +0000

One important reason people believe in their own free will is the compelling feeling we have of personal agency, of having chosen and caused our behaviors. There are many reasons to be suspicious of this feeling as a valid indicator of actual free will -- as Dan Wegner has documented so well in The Illusion of Conscious Will (MIT Press, 2002), we do not have direct access to internal causal information. (This echoes the point made several centures ago by the British philosopher David Hume: that we never directly experience any cause but instead must infer that A caused B based on available evidence.) Because we don't have direct access, we must instead make a causal inference or attribution of what caused our behavior based on a few key factors. Wegner and colleagues have experimentally manipulated those factors and have shown that doing so moves around the participant's belief that they themselves caused something to happen. (If instead the person did have direct access to the causes of his or her behavior, Wegner would not have been able to move around these causal beliefs with his experimental manipulations.) What Wegner and others have shown, in brief, is that we can't use just our feelings of having caused our behavior as some kind of prima facie evidence that we did indeed cause it.

But there is another reason to be suspicious of our feelings of free will: somehow we don't tend to have them when those around us express their disapproval of our actions. The recent behavior of Kanye West at the MTV Video Music Awards last Monday evening, followed by his responses to Jay Leno on the latter's prime time television show the next day, illustrate the selective and strategic nature of the belief in one's own free will. When directly asked by Leno to explain why he upstaged Taylor Swift, who had just won the Best Female Video award, by taking the microphone away from her to very publicly question why she had won and not Beyonce, West suddenly did not appear to believe in his own free will in the matter. Instead, he explained his behavior by reference to his own mother's recent death (during plastic surgery) and the pain he feels over her loss. He did not take personal responsibility for what he had done, but instead apologized that "his own pain caused someone else's pain". He was saying, in other words, that he did not intend his behavior, he did not freely choose it, but instead it was caused by the understandable pain he still felt over the loss of his mother.

Now, Mr West has been roundly criticized in the media for not taking personal responsibility for his own actions and trying to focus the public's attention on his own suffering and not the embarrassment and humiliation he caused Ms Swift. But we should realize that his sudden, and quite convenient abandonment of his own presumed belief in free will, is something we all tend to do, although usually not on the Jay Leno show in front of 30 million viewers. The self-serving nature of how people ascribe the causes of their behavior have long been known, going back at least as far as the pioneering theory and research of Prof. Gifford Weary of Ohio State University, published in the Journal of Personality and Social Psychology in 1978. Weary (then writing as Gifford Bradley) showed how we all tend to believe in free will for positive outcomes and behavior, but not for failures and other negative outcomes. For some reason when we fail we quickly turn into social psychologists and start to see the external causes for our behavior.

In the Tampa SPSP debate on free will with Baumeister, chronicled in our respective Psychology Today blogs over the summer, my colleague Brian Earp and I presented new data showing that a person's self-concept and social identity includes many features of which we are proud but of which we had nothing to do with, causally. Examples include one's eye color, country of origin, gender, successful ancestors, and so on. Those data concerned the case of our positive outcomes that we did not ourselves cause. The present case of Kanye West's sudden conversion to social psychologist (most likely temporary in nature) on the Jay Leno show illustrates the complementary situation, in which we tend not to believe in our personal causation of negative outcomes.

So it all boils down to wanting to take credit for positive outcomes and avoiding blame for negative outcomes. We take personal pride in things about us that others like and admire, whether or not we personally caused those things, and we disown those things that others do not like and disapprove of, even if we did cause them. The Janus-like nature of a person's belief in free will suggests to us that the belief does not come from first principles, logic, or any actual evidence -- instead it is self-serving and selectively, strategically applied.

The political psychologist Philip Tetlock of the University of California-Berkeley has proposed a model accounting for such seeming inconsistency (Psychological Review, 2002), in which we automatically adopt different mind-sets and values depending on the situation. Because our social standing and relations with others in our group are so important to us, when faced with their disapproval we quickly adopt a kind of "defense attorney" mindset that focuses on the excuses and external causes of what happened, in order to deflect blame and maintain good relations with others. This is quite different from the "prosecutor" mindset we normally adopt when faced with disapproved, anti-social behavior of others in our group. That mindset is all about holding the miscreant's feet to the fire, holding them responsible for what they did, and only reluctantly acknowledging any mitigating circumstances. The importance of Tetlock's model is that it holds that both of these mindsets exist within each of us, and that we flip back and forth depending on the circumstances.

Like Wegner's related work on the illusory nature of the feelings of conscious will, the shifting and strategic manner in which people believe versus do not believe in personal control over their actions is a compelling reason to distrust the phenomenal feeling of personal causation. As Baumeister highlighted in his half of the Tampa debate and on his blog over the summer months, the belief in free will serves many important motivational functions for us as individuals (as opposed, say, to feelings of helplessness, which are maladaptive and highly demotivating). But it seems that people don't really believe in it very strongly or deeply, as a bedrock principle -- rather they believe in free will when it suits their purposes, and not when it doesn't.

New study: TV food ads provoke automatic eating in adults as well as children

John A. Bargh, Ph.D. — Tue, 21 Jul 2009 18:09:30 +0000

In a new study published this month in the journal Health Psychology, TV food ads were found to significantly increase eating while viewing, in adults as well as children. To go directly to a description of that study and a discussion of its implications for the obesity crisis, just jump down to the photo of the hamburger ad.

The following is another installment in an ongoing Psychology Today blog debate with Roy Baumeister concerning the existence of free will, for which the new study on automatic effects of TV ads is highly relevant.

In his reply to my blog entry The Will is Caused, not Free, Roy Baumeister echoed recent statements by Kathleen Vohs and Jonathan Schooler that "scientists should not tell the general public that determinism is a proven fact, because to do so may have socially undesirable effects." Baumeister argued further that "tell[ing] the public that free will is an illusion... [is an] irresponsible misrepresentation of opinion as fact [which] has additional and damaging consequences, insofar as the unfounded opinion they advocate will promote antisocial behavior."

As I pointed out in Part 1 of my reply to Baumeister, it is not at all clear at present what, if any, consequences there might be in communicating to the general public scientific results that suggest limits to the scope of their free will. Initial studies by Vohs and Schooler, and by Baumeister himself, have found increases in cheating and aggression in experimental participants, after they were told by the experimenters that science had proven that free will was an illusion. In my prior post I noted that it is quite premature to draw strong conclusions from these few initial studies, as Baumeister and Schooler have, about any such negative social consequences, for several reasons: First, these initial studies only examined antisocial forms of behavior, and so only negative and no positive consequences could have been observed. Secondly, I noted that we already have the historical precedent of evolutionary science and religion, in which substantial amounts of evidence in favor of the principle of evolution over the past 150 years has not shaken people's belief in supernatural causes such as the act of Creation -- so we probably don't need to be so afraid of informing the public about studies indicating their lack of free will.

Now, Baumeister has been careful in his recent posts to distinguish determinism from causality, and he and I seem to be converging, via this battle of the blogs, on the notion of relative freedom. I see the body of social psychological research on unconscious or automatic causation of social judgment (e.g., stereotyping, forming impressions of others), social behavior, and social goal pursuits (such as cooperation, competition, achievement, and affiliation) as showing that the glass of free will is mostly empty, and Roy sees it as mostly full -- or at least fuller than I do. But I cannot and do not conclude on the basis of that evidence alone that free will does not exist, just as Roy cannot and does not conclude from that evidence that the will is entirely free.

SHOULD WE TELL THE PUBLIC THAT THEIR BEHAVIOR IS BEING CONTROLLED?

Given this, I'm not sure what Baumeister, Schooler, and Vohs' position is concerning informing the public about experimental evidence about limitations in the scope of free will, or of situations and domains in which people believe they are exerting free will when in actuality they are not. This is not the same as telling the public that determinism is an established scientific fact, or that free will is an illusion (in an absolute sense). Rather, it consists of reporting to the public evidence of external and internal causes of their choices and behavior of which they are not aware and did not consciously intend. In that spirit I want to alert you to a new study, just published this month in the journal Health Psychology, that illustrates perhaps the most important reason why it is irresponsible for us as scientists not to tell the general public about evidence of limits to the free scope of their will.

As Baumeister emphasized in his Tampa SPSP debate presentation, the belief in free will (as distinguished from free will per se) serves important motivational and subjective functions for the individual. But it also can cause problems when it leads us to ignore or dismiss the possibility that there may be powerful influences on our behavior that we don't know about -- even for important behaviors such as who and what we vote for, or the kinds and amount of food we eat. If we believe that we are the absolute captain of our soul, then we don't worry too much about these potential influences -- and thereby leave ourselves wide open and vulnerable to them. (See also the recent commentary on this point by Tom Clark of the Center for Naturalism.)

Television and other forms of advertising is expressly directed at getting us to do something that is in the best interests of the advertiser, but not necessarily our own. We have already recognized this in the case of cigarette (tobacco smoking) advertising and as a consequence it has been banned now for many years. In the new study, Jennifer Harris and Kelly Brownell of the Rudd Center for Food Policy and Obesity at Yale and I showed that passive exposure to food advertising on television may contribute to the ongoing obesity epidemic by automatically triggering eating behavior, right then and there while watching TV. Experiment 1 focused on elementary school children because the Federal Trade Commission has reported that they see an average of 15 TV food ads per day and that fully 98% of these ads promote products high in fat and sugar. We simulated the natural television viewing situation for our young participants by having them watch a 5 minute cartoon that contained a few 30 second food ads -- or, in the control condition, non-food ads. While they watched the cartoon a bowl of goldfish crackers was made available to them. As we had suspected, those children exposed to food ads during the cartoon ate significantly more of the snack food than did the children in the control condition. Unexpected, at least to me, was the size of this effect: children consumed 45% more of the snack food when exposed to food advertising.

Now, we all know that children are not as able as adults to defend themselves against ads for toys, cereals, clothes, DVDs, etc., so perhaps this finding is not that surprising. It certainly suggests that there is a direct and automatic effect of food ads on consumption behavior in children right then and there while they are watching television, not only on their preferences for certain brands or products for their parents to buy them. And the sheer size of the effect strongly suggests that this automatic effect on consumption is indeed a contributor to the public health problem of obesity in children.

However, nearly all of the social psychological research on automatic and unconscious causes of human judgment and behavior over the years has been conducted on college-age or older adults, suggesting that adult television viewers might be just as vulnerable to the deleterious effects of food advertising as are children. In our Study 2, we showed a group of adults a short television documentary that incidentally included either snack-food ads, nutritious-food ads, or no food ads. After the program, they took part in what they thought was a separate study in which they taste-tested a range of healthy (e.g., fruits) and unhealthy snack foods. We found that the adults who had been exposed to the snack food ads ate more of all types of food during the taste test compared to the other conditions. Thus the automatic effect of snack food ads to increase the amount eaten while watching television holds for adult as well as child viewers, suggesting that TV snack and fast-food ads are a contributor to adult obesity as well.

In neither study was the amount eaten related to the participants' reported levels of hunger, and in careful questioning after the experiment was over, no one showed any awareness or appreciation that the amount of food they ate while watching the show was influenced by the ads they saw. (In studies such as these, participants typically strongly resist such suggestions.) These are unconscious effects, and so by definition one is not aware of them while they are happening. Because people do not experience these influences on their behavior, they have no chance at correcting or controlling them. The only way then for the general public to know that there are interested agents out there (e.g., advertisers, government) exerting control over their behavior in these ways is for us as scientists to do these kind of studies and to report the findings publicly and as widely as possible.

Bottom line: we should be telling the public the truth about limits to their free will as revealed in experimental studies, and not decide for them what is good or bad for them to know. We may think we are doing them a favor by permitting them a positive illusion, but there are also negative consequences to such naivete -- namely, leaving oneself wide open to being controlled by others who are not so naive.

John Bargh and ACME Lab at Yale University conduct research on the unconscious causes of our preferences, motivations, and social behavior. ACME research publications are available here

Roy Baumeister and Some Misunderstandings about John Bargh

John A. Bargh, Ph.D. — Fri, 10 Jul 2009 18:36:38 +0000

In this very interesting post by Roy Baumeister, he notes several apparent misunderstandings I have about free will and about the underlying message of research showing some negative consequences of telling people that free will does not exist. Naturally, I don't think these misunderstandings are entirely my doing, a point some others have already made in replies over at Roy's "Cultural Animal" blog. I have already posted a reply there concerning the shifting definitions of free will in Roy's post and also the possibility that he and I are both right but are operating at different levels of analysis. Here I respond specifically to Baumeister's charge that researchers are being irresponsible when they draw conclusions against the existence of free will, and that expression of such conclusions to the public will promulgate anarchy and antisocial behavior.

"NOBODY WANTS RESEARCH TO STOP"

This is good to hear. Here are the reasons I had thought otherwise: At Tampa I noted the Vohs & Schooler paper and the more recent Baumeister paper in which telling research participants about purported scientific findings that free will was an illusion and did not exist increased the participants' subsequent negative social behavior -- cheating, stealing, and aggression. I noted the particular choice of dependent measures in these studies -- all negative social behaviors -- and wondered aloud what the point of these studies really was. I also noted that at least one of these studies was funded by a philanthropic foundation promoting spirituality and the reconciliation of religion and science. This was in the context of a discussion of the roots of the free will concept in early Christian theology, in particular St Augustine's use of the concept of free-will to solve the 'problem of evil' -- how an all-good and all-powerful God could permit the existence of evil in the world. (Answer: because people must be free to do good or evil as they choose for there to be a basis for the Final Judgment upon each of us.) Again, in this context, I noted how people in the 19th century worried that a widespread belief that there is no God would result in anarchy and negative social behavior (mainly because of the removal of the threat of eternal punishment in the afterlife), and how similar this worry was to that of Vohs-Schooler-Baumeister that telling people there is no free will would produce the same outcomes. And I noted other research showing that conservatives have the same worries about what would happen if it were proven that there is no God as they do about what would happen if it were proven there is no free will -- that the social order would break down -- and that liberals too have the same worries in both cases, though different than that of conservatives -- that there would then be no meaning to life. I raised all of this to the Tampa audience because of the remarkable conflations I was discovering between religion and free will. I brought all of this up to provide a broader context to the recent conduct and publication of research specifically designed to show that without a belief in free will, social order will break down. As an example, I noted that in designing these studies, the researchers chose only to look for negative social behaviors as consequences of being told free will did not exist.

Now, you can and should draw your own conclusions as to why these studies were (and continue to be) conducted -- at the Tampa debate I left matters hanging at this point and did not explicitly attribute any such motivations. However, I was of course suggesting that there might be deeper motivations, even historical ones, behind these studies, of which perhaps even the researchers themselves are unaware. For example, philosophers of science long have noted the operation of the Zeitgeist, or the 'spirit of the times' as a kind of unseen hand that helps drive ideas for studies as well as what topics are seen as important for publication by reviewers and editors. (Psychology researchers, myself included, are of course not immune from the operation of such unconscious forces and motives.) But this implication or suggestion of an unstated, underlying motive for the research is what I think brought some heat to the debate, especially during the question period following the debate. No one likes to have their motivations questioned. And although I had meant to suggest that there might be deeper, implicit or unconscious motivations driving the design and conduct of these studies, and not necessarily explicit ones on the part of the investigators, in retrospect I don't think I made that distinction as clearly as I could and should have to the Tampa audience.

Now, I mentioned above that in the published studies on the consequences of being informed that science had shown free will did not exist, the researchers chose to look only for negative social behaviors. Why not positive ones as well? After all, among the last words of Jesus Christ on the cross were: "Father, forgive them, for they know not what they do." In other words, perhaps knowledge that free will did not exist would cause us to be more understanding and forgiving towards the faults and misbehaviors of others. And there may be other benefits as well. As I learned after the debate, Jonathan Schooler and colleagues had already thought of and conducted such a study. They indeed found that telling experimental participants that free will did not exist caused those participants to be more forgiving towards the transgressions of others.

This is a very important finding. After all, one key premise behind Baumeister's admonition to researchers to not to tell the public that free will is an illusion, is that this expression of opinion (as if it is merely a matter of personal opinion, regardless of how informed this opinion might be of the theory and empirical evidence bearing on the issue) "has additional and damaging consequences, insofar as the unfounded opinion they advocate will promote antisocial behavior." I believe Roy's conclusion is premature, because this general line of research has only just started, and as further data come in we will have a more complete and balanced picture of the social and behavioral consequences of a disbelief in free will.

To my mind, one potential benefit to getting people to not believe so strongly in the power of their own personal agency or free will is that they might then be more concerned about external influences or even explicit attempts by advertisers, government, etc. to control what they do (eat, drink, buy, vote). Research by Tim Wilson and Nancy Brekke (Psychological Bulletin, 1994) has shown that people do not worry very much about these influence attempts because they believe they are the captains of their minds and in near-complete control over their judgments and behaviors. For example, people do not believe negative campaign advertising affects them, and so do not attempt to counteract or defend themselves from the effects of such ads, yet that variety of campaign advertising is in actuality so effective that it became nearly the exclusive form of campaign ads during the recent 2008 US presidential election. And Jennifer Harris and colleagues in our ACME lab have recently shown unconscious effects of television ads on snack food and cigarette consumption, such that these ads contribute to societal health problems of obesity and smoking (see www.yale.edu/acmelab/publications.html). Thus I can see significant positive benefits in informing people of their (at least relative) lack of free will in the behavioral impulses triggered by the ads, both in their own health outcomes and in their ability to counteract presumed unwanted influences on their important decisions, such as who they want to lead their country. Indeed, given that Baumeister has expressed his belief that telling people that free will may not exist is 'irresponsible', I can make the case that not telling them is perhaps even more irresponsible, because it leaves them at the mercy of corporations and governments who are not quite so naive.

Another reason I'm not as concerned as Vohs, Schooler, and Baumeister are about the potential negative social consequences of telling peole the scientific reasons why free will may be an illusion, is that we have a historical comparison case already. Coincidentally the debate in Tampa was held the day before Darwin's 200th birthday, and in the 150th year since publication of Origin of Species. In the course of those 150 years a massive amount of research evidence has been generated, across several scientific discipline, in favor of the theory of evolution, and against the prior dominant notion of an act of Creation by a supreme being. Have people taken this overwhelming evidence to heart and therefore abandoned their belief in Creation and in a supreme Creator? Hardly. Most of the people of the world hold religious beliefs in some form, and even in highly industrialized countries such as the U.S., recent surveys show 90% believe in God, 60% believe in a literal Final Judgment at the end of time, and even 1 out of 6 high school science teachers believe in Creationism! Now, the evidence that already exists in favor of evolution is far, far greater than that out there right now against the existence of free will -- and it still hasn't done much to change the cosmology of the average person. So need we really be this worried about the mass effects of telling people the results of psychological experiments bearing on the existence, or at least relative scope, of free will in their lives?

SO SHOULD WE BAN THE STUDY OF ECONOMICS?

By the way, other forms of knowledge have been shown to induce negative social behaviors such as greed, cheating, and lack of concern for others -- economics, for one. Research has shown that economics faculty donate the least amount to charity among all academic departments in universities, and that training in economics is the culprit: in essence, giving things away to others is considered irrational behavior in economic theory. Applying the same logic as does Baumeister and colleagues, perhaps we should ban the study of economics -- or at least the teaching of it to others, because doing so is an irresponsible advocacy of opinions (i.e., that altruistic giving to others is irrational) that undermines community and promotes antisocial behavior.

In short, I am against any suggestion that we should withhold or not tell the public about research findings bearing on the existence or illusory nature of free will; to the contrary I believe it is irresponsible and a disservice to not communicate this research to the general public. Baumeister, Vohs, and Schooler want to argue that (as Roy does in the post to which I'm replying here) the "assertion [that free will is an illusion] is unproven and unprovable" so that conclusions about the existence of free will based on scientific evidence are unfounded and can only reflect the personal opinion of the author. But at the same time, Roy discusses in his post the idea of "partial and relative freedom... some actions are freer than others" and I want to point out that it was originally research back in the 1970s and 1980s that discovered this. Back then, there was a near-absolute faith in free will and the centrality of conscious intention and choice in producing all of the higher mental processes in humans. It was research on automatic, implicit, unconscious influences that discovered that (at present) this freedom is in fact partial and relative. Without this research we would never have known even about the partial and relative nature of our freedom to act. In my opinion, the trajectory of this research, with more and more domains of complex human functioning being discovered each successive year to operate outside of conscious intention and awareness, does give us some clues and basis for speculation about the absolute question, but I agree with Roy and colleagues that we should be careful to present such conclusions as theory and not as established fact. We can and should, however, draw -- and express to the public -- informed conclusions about the relative scope of free will and personal agency in daily life based on the available empirical evidence, just as Roy has done in his post.

There is more for me to respond to in Baumeister's post, but it will have to wait for next week -- this is certainly enough for now!

Behind the Scenes at the Tampa Free Will Debate

John A. Bargh, Ph.D. — Mon, 29 Jun 2009 20:08:23 +0000

One of the highlights of the past year for me professionally was the opportunity to 'debate' my long-time friend Roy Baumeister on the topic of free will, at the annual Society for Personality and Social Psychology convention held in Tampa last February. This was actually the third time Roy and I have debated the free will issue at international social psychology conventions -- first at SPSP in 2000, and then at the Society for Experimental Social Psychology circa 2005. Each of these occasions has been an invaluable learning experience for me -- Roy is a world-class expert on the self generally, and self-control more specifically, and his ideas and research findings have posed the greatest challenge to my own over the years. It's kind of like what they say about New York City -- if you can make it there, you can make it anywhere -- so too if you can find successful counter-arguments to Roy's and his colleagues concerning personal agency. Hopefully we will have the opportunity to do this debate thing again in a few years. But back to Tampa: there appear to have been some confusions about was actually said at the debate itself, which I think are largely due to the multiple outlets and formats into which this content was subsequently poured. And so it is in an attempt to clarify the situation that I provide below my personal recollection of what happened, when it happened, and where we seem to be going from here:

1. The debate itself. Roy and I were introduced by moderator Constantine Sedikides (see photo), who did a wonderful job keeping the lid on the proceedings. (Coincidentally, we were debating at the same time as the annual reenactment of the pirate-ship invasion of Tampa Bay, right outside the convention center, which may help explain the particular hat Constantine is wearing.)
I presented first for 30 minutes, followed by Roy. By prior agreement, neither of us knew in advance what the other would be saying -- although we certainly knew about each other's published work and general positions on the debate topic. But still, as a consequence of how the session was organized and designed, our talking points did not easily or obviously match up: Roy had his own presentation prepared and so did not respond to the points I made during the session, and my own talk did not anticipate many of the points Roy made.

2. The SPSP Newsletter. The July 2009 issue has just been published and contains short articles about the debate, one by Roy and Kathleen Vohs, and one by myself and Brian Earp, a Yale undergraduate cognitive science major who was our ACME lab manager this past year. However, the two articles do not correspond directly to the two sides of the debate. Rather, the Newsletter articles came about as Roy and Kathleen submitted their piece on determinism and causality to the newsletter editor, who then asked me for a companion article on my side of the issue. Brian and I chose to use the opportunity to summarize the arguments made during the actual debate, but again, this was not the intent of the Baumeister-Vohs article.

3. The Blogs (a.k.a. You Are Here). Accordingly, today Roy and I are using our blogs to provide a more complete discussion of what we and our co-authors believe to be the important issues and points to be made about the free will issue, based on relevant empirical research findings as well as logical and philosophical considerations. We only had 30 minutes each at Tampa, and only 1000 words or so each in the Newsletter articles, so that in the Debate itself and the Newsletter we were not able to present all of the relevant arguments in any complete fashion. That's what we are seeking to do now, here in the blogosphere. Stay tuned!

The Will is Caused, not "Free"

John A. Bargh, Ph.D. — Tue, 23 Jun 2009 16:06:46 +0000

Note: The following is a summary of our side of a recent debate with Roy Baumeister on free will, held at the annual convention of the Society of Personality and Social Psychology last February in Tampa, Florida. It appears in the current issue of Dialogue, the SPSP newsletter, along with a companion piece by Roy and Kathleen Vohs on determinism and causality. My co-author is Brian Earp, erstwhile ACME lab manager.

We welcome the opportunity to summarize our main points from the SPSP debate; first though we will respond to the additional arguments by Baumeister and Vohs in this issue concerning determinism and causality. We see no problem with the assertions that psychologists need not be strict determinists to practice their science, and that determinism and causality are not the same thing. However, neither of these points is relevant to the basic question of free will. The ‘free' in free will means freedom from causation, either by external forces (in the political sense of the term) or internal ones (in the psychological sense); and in our view it is just as problematic to claim that the will is uncaused as it is to argue it is not determined.

Free will may be defined as an agent's ability to act on the world by its own volition, independently of purely physical (as opposed to metaphysical) causes and prior states of the world. The folk notion of free will is laden with the concept of a soul, a non-physical, unfettered, internal source of choice-making-in other words, an uncaused causer. "The soul" may have gone out of fashion, and "the mind" taken over many of its functions and connotations, but the intuitive notion of free will has stayed much the same: there is something inside each of us that allows us to make "real" choices-choices that even an omnipotent being, one who knew every environmental influence, and every physical fact leading up to the choice-making event, could not foretell with perfect confidence and accuracy. Determinism, if it were true, would indeed rule out this sort of free will, or shunt it into the realm of total redundancy. But indeterminism (of whatever flavor) isn't any kinder to the notion. Just because some event is not strictly determined by prior physical data doesn't mean it is caused by a free will. It may be simply indeterminately, probabilistically, or (to whatever degree) "randomly" caused by prior physical data. (If one wishes nonetheless to use the existence of error variance as evidence for the existence of free will, we can only point out that our business as scientists is to strive to reduce this unexplained variance by replacing it with explanation. Calling it ‘free will' and walking away satisfied rather misses the point.)

But let us assume that there is a free, internal source of control that guides our behavior and is ultimately responsible for ‘real' choices. To attribute human behavior to this mystical source is to place one's bets on an increasingly shrinking sphere. The project of social psychology, after all, has been to identify (a) external-to-the-individual causes of judgment, motivation, and behavior, such as situational influences, and (b) internal-to-the-individual causes, which research has shown increasingly to operate outside of awareness and conscious intention-not "freely chosen" in any sense of the term. Are there some human behaviors that are possible only if free will exists and is a true causal source of action? There may be. But let's not give up on the search for non-mystical causes just yet.

This brings us to an area of agreement revealed in the debate: that a belief in free will is important for human strivings. People cherish their sense of control over the world and their own behavior. In the debate, we noted recent empirical articles by Vohs and by Baumeister showing negative consequences (cheating, aggression) of informing participants that free will does not exist. Our response to these ‘new' articles is that our field revealed the existence of such positive illusions decades ago, and we already know how essential they are to normal functioning. Clearly it is motivating for each of us to believe we are better than average, that bad things happen to other people, not ourselves, and that we have free-agentic control over our own judgments and behavior -- just as it is comforting to believe in a benevolent God and justice for all in an afterlife. But the benefits of believing in free will are irrelevant to the actual existence of free will. A positive illusion, no matter how functional and comforting, is still an illusion.

And we must caution against drawing conclusions from such research findings (implicitly or explicitly) that we should either (a) not make findings against the existence of free will known to the public or (b) stop doing such research altogether. The belief in personal free will is a deeply rooted aspect of human phenomenal experience, and is so powerful that even those who do not subscribe to it intellectually still feel it in their personal lives as much as everyone else. It is not uncommon for one's first-person experience to be at odds with physical reality: 500 years after Copernicus we still see a morning sunrise, not the earth (and ourselves) tilting towards the sun, even though we know better scientifically. As Dan Wegner, Paul Bloom, Dan Dennett, and others have argued, there are strong natural supports for the belief in supernatural entities, just as there are for free will -- and sunrises too, for that matter. And if, as countless recent surveys show, the prodigious evidence in favor of evolutionary theory accumulated over the past 150 years has done little to erode the popular belief in a creator-god, then we can rest assured that the relatively nascent research on unconscious causes of motivation, judgment, and behavior will not result in anarchy or the collapse of social norms and moral behavior.

We should also not forget past social psychological research demonstrating that the belief in personal free will is selective: people routinely make self-serving attributions about the causes of their behavior. We take credit for the positive things we do (free will), but not for our misdeeds and failures ( "I had no choice", "I was abused as a child", "I was angry"). This suggests to us that much of the emotion surrounding the issue of free will is not about freedom per se but about self-esteem maintenance. We take personal pride in our ancestors, our blue eyes or rich brown skin, our height or birthday or name (as in the name-letter effect)-none of which we chose or had any control over. Accordingly, we analyzed hundreds of individuals' spontaneous self-descriptions, and indeed 34% of their first-to-mind completions to the stem "I am _____" were such non-chosen aspects of self. It seems that people do not possess a consistent belief in free will so much as they strongly wish to take credit for the good things they are and do (regardless of whether they caused them), and to distance themselves from the bad things (even if they caused them). Evidently, the belief in free will is not principled, but socially strategic in nature.

So what, then, if one's will is not ‘free' of internal causation? It is still your will and my will and each is unique: a confluence of genetic heritage, early absorption of local cultural norms and values, and particular individual life experiences. After all, one can claim personal ownership of one's will just as much as one claims ownership of one's name, eye color, and birthday, and be as proud of one's will and its products as one is proud of the exploits of great-great-Grandma the pioneer, even though one's ‘free will' played no role in any of these.

John Bargh and ACME Lab at Yale University conduct research on the unconscious causes of our preferences, motivations, and social behavior. ACME publications are freely available at www.yale.edu/acmelab