Even though all rational people know that determinism rules human behavior, and in that sense there is no possibility of “choosing otherwise” at a moment of decision—absent quantum effects, which don’t in any way give us “free will”—this conclusion disturbs some people. Our sense of agency is so strong that it’s impossible for many of us to accept determinism of our behavior, or, if we do, to fully grasp its implications. Others, while accepting determinism, nevertheless confect other forms of “free will” that are compatible with determinism. I won’t go into the arguments for “compatibilism”, which I uniformly reject as simply semantic arguments designed to fool people into thinking that they’re free agents. Finally, a lot of compatibilists tout alternative forms of non-dualistic free will by the “Little People” argument: that if regular folks truly grasped determinism, with no loopholes, they’d become nihilistic, apathetic, or even immoral. In this way, belief in free will resembles belief in God, as many atheists are nevertheless pro-religion because they think that without faith society will fall apart. (That of course is wrong, as we know from seeing the godless countries of Europe.)
Several previous studies have buttressed the Little People’s argument for free will, including an oft-cited paper by Vohs and Schooler published in 2008, which showed that subjects “primed” by reading a passage by Francis Crick promoting determinism tended to cheat more in subsequent tests than those primed by reading an innocuous passage on consciousness. This paper has been used to show that it’s important to let people think they have free will, for they’ll behave badly if they don’t.
Unfortunately, the Vohs and Schooler paper, which has its own flaws (e.g., not a whit of evidence that the readings affected real-life cheating or lasted more than a day), was not replicated in two subsequent studies (see here and here). Nevertheless, new papers continue to come out with mixed results: some show that belief in free will promotes good behavior, others that it promotes bad behavior, and still others give mixed results. They all have their problems, including the article discussed here, a new paper by Emilie Caspar et al. in Frontiers in Psychology (free link; reference below). And by “problems”, I mean that these are all lab studies that give no conclusion about how one’s belief in free will or lack thereof affects regular quotidian behavior.
Nevertheless, I’ll summarize it briefly.
First, subjects (40 of them, recruited and tested in pairs) were “primed” by reading one of two passages written by Francis Crick: one attacking free will and promoting determinism, the other a neutral passage on consciousness (this resembles Vohs and Schooler’s method). But the psychologists also assessed participants’ “core beliefs” in (dualistic) free will versus determinism in psychological tests administered before the experiment.
Then the experiments. There were two that were relevant.
1.) Empathy. The first involved one member of a pair giving an electric shock to the other, and under two conditions: “coercive,” in which an experimenter sat next to the “shocker” and told him/her whether or not to administer a shock (a real one!) to a “shockee”. There was also a “free will” condition in which participants could decide on their own whether or not to administer a shock, earning € 0.05 for each shock they gave (they were already paid € 12 for participating). At the end, all participants reread the priming text they’d read before starting. After one bout, the positions were reversed so the “shocker” became the “shockee” (see point 2).
Results: They’re shown in the figure below, giving the number of shocks given in the “free will” condition (i.e., no hectoring experimenter telling you what to do). The data are divided into those primed by an innocuous passage (left two bars) and the “deterministic” passage (right two bars). The graph shows that overall, those primed with determinism showed significantly fewer shocks administered, but that is due to a huge reduction in the number of shocks administered by women who were primed with determinism. Men primed with determinism showed a nonsignificant increase in number of shocks administered, but the effect of female “empathy” outweighed that so there was an overall effect of priming. In fact, one can conclude that there’s a sex-by-priming interaction effect here, and that priming with determinism made females—but not males—more empathic. The authors, however, concentrate on the overall effect, which I think is misguided given the effect of sex (which they do, in fairness, mention).
When participants were told to shock or not shock others (data not shown), the results were pretty much the same, although participants who were coerced estimated longer interval estimates between their actions and the outcome, which, say the authors, is more characteristic of an involuntary than a voluntary action.
Finally, the “core beliefs” of participants in free will vs. determinism assessed before the experiment had no effect on the results. That is, these results appear to be due solely to the effect of priming.
2. Vindictiveness. The authors also estimated the degree of “vindictiveness” of participants: that is, the correlation between the number of shocks you got from your partner and those that you then gave to your partner. Here are those data from the groups primed for “no free will” and the controls. In both control and experimental groups, there was a correlation, i.e., evidence of vindictiveness, but that correlation wasn’t significant for females in the “no free will” group. In other words, priming females, but not males, with determinism made them more empathic—less likely to be vindictive. In this case, core beliefs did affect behavior for males but not females, but the effect was small. The main effect again was that induced by the priming.
What’s the upshot? The results are a bit complicated, for while there’s evidence that being primed with determinism makes you more empathic and less vindictive, this effect appears to hold for females but not males. The authors conclude that this runs counter to previous studies:
Moreover, we observed that the core beliefs of participants did not differ in these groups, and neither did their scores on empathy. Taken together, this suggests that the reduction of immoral behavior in the no free will group for female participants stems from the induced beliefs. The observed prosocial benefits of disbelief in free will may appear to go against the mainstream, since the literature mainly converged toward the prosocial benefits of believing in free will (e.g., Vohs and Schooler, 2008; Baumeister et al., 2009). However, numerous factors differed in our study, notably the social aspects associated with the presence of two co-participants who were aware that roles would be reversed at the middle of the experiment. Future work is required to explore this question more thoroughly.
and about vindictiveness:
Additionally, we observed that vindictive behavior was reduced for female participants in the no free will group compared to other sub-groups, and that the higher female participants scored on free will, the more vindictive they were. Importantly, our paradigm made it possible to investigate whether disbelief in free will influences the occurrence of vindictive behavior without the need to mention the notion of punishment to our participant, such as in previous studies. This tendency to behave vindictively is consistent with previous studies that showed that people who believe in determinism are less punitive and have reduced retributive attitudes toward others (e.g., Westlake and Paulhus, 2007; Krueger et al., 2014; Shariff et al., 2014). When people have to express a judgment about the morality of someone else’s behavior, their beliefs about the cause of these behaviors may greatly influence how they judge the severity of the act. Reducing people’s beliefs in free will might make them consider that individual responsibility is reduced, thus making them less retributive toward others.
Overall, the authors say that “we observed that a disbelief in free will had a positive impact on the morality of decisions toward others” and that this “challenges current thinking.” Well, that holds only for women and not men, and I won’t speculate why.
The study is of course flawed because it assesses only short-term behavior in the lab and not the long-term effect of belief in determinism on empathy and vindictiveness, which is what we really want to know. Of course, all such studies are flawed in this way. But this new one suggests that previous studies showing an increase in bad behaviors caused by priming with determinism must be taken with a grain of salt. Reviewing the literature, the authors note that among all studies there’s simply no consensus on this issue, and nearly all those studies are of this short-term nature.
But in the end, are such studies necessary? I’m not sure. The first thing we need to do is figure out what the truth is, and we already know that: human behavior is ruled by the laws of physics, and, save any effect of quantum indeterminacy, that leaves us no room for dualistic free will. Then we need to deal with that truth, just as we need to deal with the even more unpleasant truth of our own mortality. For this secondary program it might be useful to have studies such as these, but I don’t think they’re necessary, especially because of their flaws. What’s important about grasping determinism is, as I’ve always said, is to apply it to our system of reward and punishment, being mindful that nobody has a choice about whether to act good or badly. To claim, as some readers have, that determinism has no effect on such judgments is a claim I don’t accept (and neither do Caspar et al.). A full grasp of determinism would have a marked effect on how our legal system deals with criminals, or even how we deal with our own lives vis-a-vis empathy, forgiveness, and our attitude towards those who are at the bottom of society.
h/t: Dom
____________
Caspar, E. A., L. Vuillaume, P. A. Magalhães De Saldanha da Gama, and A. Cleeremans. 2017. The Influence of (Dis)belief in Free Will on Immoral Behavior. Frontiers in Psychology 8. https://doi.org/10.3389/fpsyg.2017.00020
The thing that always puzzled me was how faithiests at the same time were so confident in their belief that human beings had “free will” and at the same time believed that their gawd had everything planned-out for them.
What do they really believe, that free will somehow enables humans to compete w/ gawd’s grand plan?
I would guess that most faithiests are oblivious to possible complex analyses of the meaning of “free will”. When the priest of the text says people must chose Christ to be saved, that’s as far as it goes. Free will is assumed without additional thought. Only theologians recognize further issues and the effect determinism has on their beliefs, and we know well how they deal with such mental speed bumps.
Look up Molinism – a very sophisticated way of trying to reconcile libertarian free will with the notion that everything is planned (though not technically determined) by God.
Whether he attempt is successful or not is a moot point but the attempt has certainly been made!
Sub
What’s the upshot? The results are a bit complicated, for while there’s evidence that being primed with determinism makes you more empathic and less vindictive, this effect appears to hold for females but not males.
In other words it doesn’t hold for half the population.
The population which makes up the greater part of the law makers, the judges, the lawyers, the police, prison wardens, and armed vigilantes.
Every generation has a blind spot for errors that our descendants will be quick to point to in disbelief. They really thought/did that?! I always wonder what those blind spots are. Might popular belief in free will be one?
We all have selective memories and tell ourselves stories that bolster our beliefs.
I love history, but am well aware that it isn’t all true; that it changes in accordance with cultural changes in populations over time. (At least every 25 years, according to what one history professor told me). All the stories we tell ourselves, about how the world works, about history, mythologies and religions are prone to modification. Even the words we use change meaning over time.
I find it fascinating that siblings, ostensibly raised in the same environment and culture, can turn out so very differently. I know that their genes are different and events are different, and schooling and activities, etc. But one would think that being in the same nest with the same parents providing a common history and belief systems would make then more similar.
I’m afraid I don’t agree that “history changes in accordance with cultural changes in populations over time”. There are certainly trends in some historical interpretations, but most of the main events of history are well documented, and for some of them at least there is a consensus on their causes and consequences.
And as for “how the world works”…I hope you aren’t referring to the physical and evolutionary processes that have brought about the current state of the Earth. They are even better understood, agreed and documented.
Like the Hitchens brothers?
I don’t think that most people really think much about freewill one way or the other. Not in a purposeful way as in the freewill discussions Jerry hosts here, and similar.
Yes, I agree. What I’m really pondering is whether scientists such as Jerry may eventually cause a sea change in public opinion such that the *consequences* of belief in free will (death penalty, punishment for crimes, etc) will be thought by our descendants to be antiquated and horrific.
I’m not sure you can generalise from an increase in ’empathy’ – even if it were proven not to be a short-term phenomena – to more humane treatment for criminals.
Empathy can more easily be manipulated by the prosecutor to create identification with the victim than with the accused. The reason defence councils try to keep their clients off the stand is that the less the jury thinkks about them the less likely they are to convict.
I’m with Stephen Pinker and Paul Bloom on this. Empathy isn’t conducive to fair treatment of criminals because empathy is directed towards the ingroup, not the outgroup.
I think an acceptance of determinism leads to more humane treatment of criminals.
The way I read the empathy part of the equation was general empathy rather than empathy in a specific circumstance.
I think criminals should be treated a lot better than they are, and I would like to see a model like that of Norway applied worldwide. I think prison regimes like that in the US are counterproductive.
However, my general empathy for criminals might not be in control if I came face to face with someone who hurt someone I love. I’d have to face that situation to know whether I could maintain my ideals.
For a certain narrow definition of empathy. In fact, I think it would be fair to say that someone who indiscriminately favors their ingroup is just not very empathic.
The ability to create a mental model of what someone else is experiencing under a given circumstance would seem to me to be neutral in terms of ingroups/outgroups. It would, however, be an essential part of formulating an ethical system. The Golden Rule and its variants are essentially calls to be empathic.
I had a similar argument with another commenter on another thread, and Bloom was mentioned there, too.
I wonder if many people, even despite knowing the distinction, still conceive of something closer to “sympathy” when they write about “empathy”.
The study is far from conclusive. I do think studies like it, but not yet conceived, would be useful.
But first people need to learn bit more about what can be determined. For example, it might be (probably is) the law that specific streets must have a sidewalk. The vast majority of people would not question this yet it constrains their behavior. Additionally, the slope of sidewalk cannot exceed some value based on wheel chair use. Again, a universally accepted constraint determining our behavior.
Humans have eagerly adopted engineering controls that embed physical constraints to our behavior that are manufactured and justified by ethical concerns for the greater good, e.g. a small baby falls from an open window at a hotel. Next month, hotel has windows babies can’t fall from. Next problem….
Make the universe our cradle: it’s already determined, why not take advantage of it instead of it taking advantage of us.
Lady MacBeth was onto something.
I do not see why one needs determinism – and I speak of absolute determinism here – as an argument against free will, also in the sense of absolute, supernatural free will. Isn’t it sufficient to allude to naturalism alone?
And is it not far more interesting to understand the emergence of our undeniable sense of agency than to denounce it as a mere illusion?
Strict determinism would mean that the future is not open. Is this actually common understanding of all natural sciences? I doubt it.
What is interesting and what is useful are two different things. And yes, naturalism encompasses determinism.
Finally, yes, strict determinism means the future is not open except for the vagaries caused by quantum indeterminacy. In everyday life, no, the future is not open, and read Sean Carroll’s book “The Big Picture” to see why.
This is exactly what also bothers me a bit: that determinism is useful for what you want to achieve with it. To believe something because it is useful for something else – is this a scientific stance?
I have read Sean Carroll, and do understand the physical basics of determinism. But this might not be the last word on the issue, after all, no?
Besides, I am a great admirer of your intellect. I am just wondering about the issue of determinism, but I am certainly not defending free will.
Umm. . . . this is verging on a Roolz violation. Do you seriously think I accept determinism because it’s USEFUL, regardless of the scientific evidence? No, I believe it because that’s where the evidence points, and yes, pending anything new, it is the best word on the issue.
Once you accept the truth, then you can deal with its implications. I can conclude from your post only that you haven’t read anything I’ve written about or spoken about on this topic.
That’s a bit harsh, I think. Rosmarie was being perfectly polite, so far as I can see.
cr
Jerry, I think perhaps you’ve misunderstood Carroll’s brand of determinism. I don’t think he believes the future is closed in the sense you mean. Rather, as a staunch Everettian he believes that the set of possible futures is fully determined by quantum mechanics, that the number of futures in that set is far greater than one, and that all of those possible futures are physically realized.
This view of quantum determinism obviously does not license belief in unphysical libertarian free will, but neither does it justify the sort of single-outcome, can’t-do-otherwise determinism you advocate. In my opinion Rosmarie is right to doubt that claim.
Indeed, Carroll has affirmed compatibilism on many occasions.
“Strict determinism would mean that the future is not open. Is this actually common understanding of all natural sciences?”
But see chaos theory. Complex systems are absolutely deterministic BUT not predictable in detail without running through all the steps in sequence (at least as I understand it). So they appear random.
For example consider the waves on the beach – how far up is the next wave going to come? That depends on the backwash from the last wave, and the size of the next arriving wave, and the timing – sometimes they combine to slosh a long way up as a ‘super-wave’, usually not. This is accentuated if the shoreline is shaped into a little bay, for example. It’s entirely deterministic but completely unpredictable. For predictive purposes you might as well call it ‘random’ but in fact it isn’t.
cr
In he science fiction series, Hyperion, the AI is working on creating a being that can determine all variables and in this way become god.
Hmmm. Yes, an interesting concept, and I agree that would make it ‘god’, but ‘all variables’ is the killer. The power required to compute (in, presumably, real time) all the variables of, say, my example of waves surging up the beach would be astronomical. It might not require every single molecule of water to have its position and momentum individually calculated, but certainly every cc would have to be so treated. In fact the simplest way to do it could be to physically build the beach and run the sequence.
Now imagine that replicated for every beach and stream and wave and gust of wind in the world. And that’s without bringing living things into it. Living things always screw things up. No wonder the Auditors of Reality hated life. Or that the mice found it necessary to make an entire planet to compute the question for 42. (Of course their program hit a fatal bug).
cr
Yes; Laplace’s Demon is precluded by physics.
But if chaos theory tells us that macroscopic outcomes are sensitive to microscopic perturbations, and quantum theory tells us that microscopic events are not deterministic (in Jerry’s sense of yielding unique outcomes), then that seems to undermine rather than support the claim that “complex systems are absolutely deterministic”.
You can’t build a completely deterministic system on a nondeterministic foundation. You can come close; there are ways of insulating high-level events from low-level noise (such as error correction in digital electronics). But it can never be perfect, and you can’t then claim that “the laws of physics” mandate whatever degree of determinism you’ve engineered in.
Hmmm. Suppose for the moment that the microscopic events are not so small as to be subject to quantum uncertainty. I would say that, in the complex (chaotic) system, the effects of small changes are so magnified and become rapidly so complex that the results are completely unpredictable (except maybe statistically). That would still be completely deterministic but also not predictable.
Quantum uncertainty is almost the icing on the cake. I personally think that would, genuinely, make the whole system non-deterministic but I hasten to say I’m no sort of expert in this field.
cr
In the most cases, the “outcomes” which arise are close to being equivalent. The way I think about it (physicists are welcome to correct me) is like how one understands the kinetic theory of gases: a distribution of the relevant properties is massively likely under most circumstances. So there are “many ways” to have pressure P (or P +/- some amount, or whatever).
Gas dynamics aren’t chaotic at human or laboratory scale. On a planetary scale it’s a different story; gas laws are of little use in predicting the weather.
The point of chaos theory is that small perturbations don’t always average out; in some systems they’re amplified to produce wildly divergent outcomes.
I suggest that natural selection frequently has reason to favor systems of the divergent sort. In meiosis, for instance, the goal isn’t to produce a million sperm that are mostly identical; it’s to maximize genetic diversity through random recombination.
No, they aren’t chaotic, but they are “coarse grained” (I believe that’s what the jargon is here). (I was mentioning one way you can more or less the same results from divergent starting points.)
I’m not sure about the label ‘vindictiveness’.
From the first of those graphs it appears that men in the control group who received not shocks gave no shocks and those who received 60 gave 60 in turn. That’s not vindictive, that’s tit-for-tat; an entirely rational strategy under game theory.
On the other hand, some members of the ‘female – no free will’ group appear to have administered shocks despite not receiving any themselves.
I maintain that comaptibilism seeks to solve a problem that does not exist. Namely that people who accept determinism lose their moral compass. They clearly do not. In fact I believe the opposite occurs. In my case aceptance of determinism made me far less hateful, vengeful spiteful and shameful.
From the conversations I’ve seen compatibilists seem more concerned with losing the right to moralize rather than genuinely concerned that people will lose their morality. The latter is just not evident in any way I can see. Incompatibilists seem to be among the most moral people I have ever met.
Have you got some evidence of, say, Daniel Dennett, being more concerned with moralizing than with people losing their morality?
Has Dan Dennett produced any evidence whatsoever that people need compatibilism to maintain morality or for any other reason? Has he or any other compatibilist shown that acceptance of determinism messes people up and therefore we need the semantics of compatibilism to continue with normal life and morality?
For what other reason are compatibilist semantics put forward?
Compatibilism is *not* put forward out of concern that, without it, people will not be moral.
Rather, compatibilism is put forward as an attempt to understand what is meant by concepts such as “choice” and “moral responsibility” in a deterministic world.
Not always. Some people promote compatibilism because of the supposed invidious effects of determinism. I’m sorry,but I have quotes to back that up, and I disagree with you.
No, not really, it’s rather asking a different question: why do we find the concept “choice” useful in understanding the world, and what do we mean by it?
Really? Compatibilists don’t always lend their arguments to the subject of morality? They don’t go from “What does choice mean” to “in what sense are people responsible agents for their choices and actions?”
What compatibilists are you reading? Let me know and I’ll find passages where they concern themselves greatly with “moral responsibility” in the face of determinism?
Sure yes, compatibilists discuss morality. After all, “moral responsibility” is another thing that we need to understand, given that the world is deterministic and that people do not have God-given souls.
And “compatibilism” helps us understand moral responsibility how? Determinism helps us understand morality better than we ever have. Compatibilism messes with the helpful breakthrough that determinism gives us. Namely the revelation of determinism makes feelings of hated revenge and shame irrational. Compatibilism brings them back. I don’t understand how any intelligent person can not see that determinism is helpful with understanding morality not harmful or confusing. Compatibilism confuses the issue and is harmful in my opinion. Still looking for a compatibilist to show it’s usefulness or how it helps us “understand” anything without making the “little people” argument.
I don’t see “responsibility” as achieving much beyond a primitive attempt to find someone to blame, in defiance of the complexities of real-world causality. In that regard, it’s the problem of retribution writ large because they both stem from a narrow and human-exceptional approach to morality, without even a token regard to moral luck or historical contingency. In short, it would be like trying to find and punish viruses when a much more sensible measure might be pre-emptive vaccination or a course of various customized therapies.
To that extent, I much prefer an approach akin to medical testing: find likely causes and factors, and use this knowledge of cause-and-effect to either prevent future diseases or to cure or alleviate existing ones. Moral science would become more accurate and more effective with that mindset than with the semi-mythical ones of “responsibility” and “choice”.
Hear hear hear hear.
Responsibility isn’t just about blame; it’s part of the taxonomy of behavior that helps us distinguish curable bad actors from those that must simply be quarantined.
Gregory, I agree.
As a compatibilist, I favor the “quarantine”
attitude toward bad actors. It has nothing to do with hatred, vindictiveness, retribution etc. It’s simply the acknowledgement that bad actors have to be separated in society insofar as they pose a danger to society.
“Responsibility isn’t just about blame”
The fact that it’s about blame at all is the problem, because blame is one of those concepts that ties into the muddle of free-will thinking. We don’t blame a virus for infecting people; we try to work out the factors (such as prior infection, exposure to other people, general prevalence of the virus, etc.) that caused the disease, and even then chiefly as part of a strategy for reducing its incidence in the future, not as a way to find something to punish. Applied to diseases, the notion of “blame” is revealed as transparently anthropomorphic at best, stupid at worst. But that’s mainly because of our primitive ideas of what’s “anthro” anyway; it’s no less problematic when applied to humans because the nature of causality is no different there.
“it’s part of the taxonomy of behavior that helps us distinguish curable bad actors from those that must simply be quarantined.”
Then cut out the middle man and describe one set as curable (for rehabilitation) and the other set as incurable (for isolation, or potentially deterrence). You can’t pretend the concept of responsibility doesn’t have problematic free-will baggage, any more than you can pretend punishment and deterrence are the same thing. By contrast, embracing a more explicitly medical-like approach makes it clearer that you’re not swallowing the free-will terminology of old while claiming to be determinist.
Once you make it clear causality is the aspect of interest, you can think more clearly about both the goals of morality (reducing suffering and premature killings in the future) and the methods (empirically, whatever is demonstrated to achieve said reductions versus whatever we traditionally assume is a good idea). It also more obviously suggests looking at the big picture and trying to factor in as many causes as possible, not just finding people to treat but thinking about the whole society and even the non-living aspects of the world. That’s why you need cold hard reason to get the job done, not encouraging words about how we’ll still be able to make perpetrators responsible in an age of determinism.
I’m still waiting for some incompatibilist to propose an empirically sound methodology for rehabilitation that gets the job done without the use of encouraging words about taking responsibility and making better choices.
Interesting monomania, there. Are you arguing for a position, or arguing for a tactic? Which do you want to do: tell the truth about the world, or manipulate people into behaving better? You seem to be unable to make the distinction.
If lying to people about “choice” and “responsibility”, in order to get them to behave, works empirically, then it works empirically. That says nothing about whether “choice” or “responsibility” are actually true. Altering future outcomes by saying stuff: that is a totally different issue from whether what you’re saying is correct or not. You can’t have it both ways.
Besides, you seem very confident in the belief that talk about “choice” and “responsibility” is the be-all and end-all of rehabilitation, absent any actual techniques or programs. To speak metaphorically, you don’t improve chess player performances by wooing them with words like “responsibility” and “choice”; you teach them actual strategy, the conditions for using said strategy, and specific details about case studies and notable games. Wholly within the bounds of cause-and-effect, and without the metaphysical mythology.
I don’t have time today to pursue this further, and we’re pushing the limits of Rool 9 in any case. So I’ll just say that as a compatibilist, I don’t see a conflict between telling the truth about the nature of human agency, and helping people understand how to apply that truth to improve their own behavior. Deceptive manipulation of the sort you’re describing is neither necessary nor (in my opinion) likely to be successful in the long run.
“Vaccination” for human actors includes the use of concepts like responsibility. We try to neutralize a bad actor’s virulence by either quarantining him or getting him to see how his actions are unacceptable.
That’s a bad analogy. Quite apart from the mechanics of vaccination having little to nothing to do with persuading someone not to commit crimes again, the concept of responsibility can’t be honestly used if its basis is free will. And yet I’d bet you’d blanche if we described rehabilitation more mechanistically: an attempt to re-engineer a recidivist into a non-threat, using his or her psychological processes to achieve the goal of a reduced future crime rate. Because when you stop buying into the narrow focus of blame and responsibility, harder but more relevant questions open up that call the whole enterprise into question.
For instance, is this actually, empirically, the most cost-effective way of reducing the future crime rate? Is everybody who commits a crime a potential recidivist, or is their crime a statistical one-off that is unlikely to be repeated? Are more people killed by easily curable diseases than by murderers or even by suicide? How do we compare the suffering of an individual crime victim with the suffering of, say, a natural disaster or a war? Is property violation really such a big deal that we have to punish responsible, blameworthy actors for it? Where are we going to procure empirical evidence in favour of or against any particular measure, and how will we quantify both the suffering and the resources needed to deal with it?
They’re not nice questions, and some of them may lead to deeply unintuitive answers which our pseudo-consciences look upon with horror. So be it. Since it’s obvious cause-and-effect is the main subject of interest when we’re both trying to make sense of the world and when we’re trying to achieve a more moral future, I think it’s far more valuable and honest to make that explicit than to blur the boundary by maintaining old ideas with dubious connotations.
Research like this drives me crazy. I’ve been using the Vohs & Schooler 2008 study for a couple years now in my classes to illustrate the dangers of underpowered research, but unfortunately so much of experimental social psychology falls into this trap.
This Caspar et al. paper is woefully underpowered, in the statistical sense. Only 5 pairs per blocking variable (group x gender) with human subjects in a social experiment. This is not laboratory control: sample units are not interchangeable save for the blocking conditions. Not even close. This is always the problem with social psychology research, and yet these papers treat the statistics like they are in a botany lab testing different heat and light conditions on interchangeable units.
With such small sample sizes and large variability, observed effects are almost completely driven by chance. As Jerry pointed out, the observed decrease in shock delivery among females was not recovered in the male subjects; in fact, there was an observed increase from control to experimental condition. This is a huge red flag that these observed effects are just noise. Furthermore, the vindictiveness graphics explain some of the initial differences between the very small treatment samples: the female-experimental treatment group was far less vindictive, driving the observed effect the authors base their entire paper on (note also that these groups are in fact unbalanced – 6 and 4 pairs of participants across gender – contrary to what the authors state earlier in their paper).
Even ignoring the methodological problems, if we take the observed effects at face-value, there is no evidence of anything in this paper. They calculate an 11 point difference between control and experimental response, pooled across groups, and although they do not give the standard error of this difference, we can calculate it from the statistics they do report: 8.97 points. In confidence interval terms, that means the 95% CI for treatment difference is roughly [-7.5, 28.5].
Now perhaps some nonzero treatment effect really does exist. If it does, it is likely to be very small, i.e. near zero, maybe about 1 point. This means that the author’s study has only about 5% power, which is essentially useless to detect anything meaningful. There’s roughly a 40% chance that their effect estimates have the wrong sign, and their observed effect sizes are likely to be 20 times too large (see Gelman & Carlin 2014, “Beyond Power Calculations: Assessing Type S and Type M Errors” to see how I got these numbers). These are exactly the type of “pure noise” results that the Reproducibility Project exposed so effectively last year.
Perhaps cognizant of these criticisms, the authors here have a strange line in their Methods section that: “The sample size was based on estimated power size considering our experimental design”. I would like to see this calculation, as I am skeptical much was actually done. If it was, why not report the target effect size the study was designed to detect? This had to be specified beforehand if an actual power calculation was done. And if they really used this to decide on their sample sizes, then necessarily their target effect size had to be huge. How could they possibly justify expecting such a large effect, given that all previous research produced either no effects, or noisy effects in the opposite direction?
Sorry for the long rant, but this type of statistical malpractice is rampant in the psychology literature. Until it’s fixed, I think these authors are doing a disservice to themselves and to the rest of us who try to learn something from their work.
Thankyou for that. Throughout reading Jerry’s summary and looking at the graphs, I was incredulous that so few subjects and experiment iterations could possibly reveal anything but noise. I appreciate your wisdom and the time you took write this.
Indeed, thank you from me as well.
You keep repeating this over and over. It is misleading. Even if dualism was true, you still cannot have libertarian free will. Libertarian free will contradicts itself before you get to any physics.
There is always the hidden assumption that empathy is good and vindictiveness is bad.
In his book “Against Empathy” Paul Bloom reminds us that empathy also has its downsides.
Further I would argue that revenge is not always a bad idea. A believable threat of retaliation can be a good thing; world peace depends on it.
Am I alone in interpreting “empathy” as an emotionally neutral cognitive ability? Like the ability to recognize faces. As I wrote above, I think Bloom is really critiquing sympathy.
What is the downside to being able to mentally model the experience another person is having, or their perspective on something?
It simply doesn’t work as well – or at all – once you start thinking in terms of populations rather than individuals. That’s why most charity adverts focus on pictures of one or two (sometimes named) victims rather than on statistics. Past that point, you need to use reason to figure out the appropriate social policy and strategy, or else you might in fact be behaving inconsistently or hypocritically without realizing it. Relying on empathy OR sympathy is likely to make one short-sighted and one’s actions correspondingly limited.
See, I think you’re really only talking about sympathy here. See my comments to peepuk below.
No. I mean what I say. Even the empath’s ability to understand another’s mental state diminishes the larger the population you wish to understand. Given the limited resources of the brain, it would frankly be surprising if there wasn’t some limiting factor involved.
Pinker makes the case in more detail in Better Angels, but the upshot is that the only feasible means of getting a clear idea of how best to benefit society involves applying cold reason to moral problems. The most obvious consequence of this is that it opens your mind to counterintuitive but effective strategies that even an empath would be reluctant or unable to understand. It also means recognizing a situation where you’re going to lose either way, and therefore acknowledging that you’re forced to compromise even if the empathic side of you sees the pain involved for the loser.
I’m not claiming empathy can be applied to a population. I’m also not claiming there aren’t other strategies for solving moral problems. I’m claiming empathy is an important one, and that it’s not actually emotionally charged, as many seem to think. When you write “…even if the empathic side of you sees the pain…” you’re alluding to sympathy. The more accurate way to put it would be “…even if your empathic ability caused you to feel sympathy for the loser.” Empathy frequently but doesn’t always lead to sympathy. I can understand what’s going on in a narcissist’s mind but I don’t feel sympathetic when he reaps negative consequences from his behaviors.
“I’m not claiming empathy can be applied to a population.”
But what else is morality except a concern for the social interactions of populations of people? Especially since most moral issues are society-wide, a lone thinker needs to grasp, however abstractly, the wider ramifications of whatever moral code they want to act upon.
“When you write “…even if the empathic side of you sees the pain…” you’re alluding to sympathy.”
You’re reading words I haven’t printed. A torture technician sees the pain of their victim: to be any good at their job, they have to figure out the best way to draw it out at all. But seeing something and feeling for it are two different things, and I’m sure I can tell the difference. While I appreciate “to see an emotion” is potentially ambiguous, I don’t believe anything in my comment was weighted towards “synonym for empathy”.
In any case, I stand by what I said previously. In principle, a moral system could be worked out by an amoral strategist, so long as they knew which rules to apply. It just happens to be a subset of rules we imbue with a lot of emotional gravitas, but they are ultimately cold hard rules that we can apply without thinking particularly keenly about their effect on minds. Indeed, since our ability to conceive of another’s mental state is intrinsically limited and parochial – and I’m not talking about sympathy, but about “yes I see you are experiencing an emotion” empathy here – to some extent it has to be worked out like that.
Well, I’ve had my say. I don’t really have anything new to add, so I’ll see myself out of this thread by just reiterating that I’m essentially making a semantic argument. I think people are importing the emotional connotations from “sympathy” into “empathy”. I don’t see why understanding another’s experience or perspective should be *inherently* emotional.
I’ll also add that individual-level phenomena are not precluded from having important effects at the population level. “Don’t murder” can only be applied to individuals but it’s still an important part of our general moral system. And, since this is WEIT, think about evolution and selection. Selection occurs only at the individual level but it is what the evolution of populations is founded on.
While Reasonshark makes a valid point in his response, I think, it isn’t really on topic with respect to the point you made.
On your point, I agree entirely. Empathy and sympathy are not the same thing and I agree that many people confuse them. Empathy often engenders sympathy, but they are very different things.
It is interesting to me how often that people noted for being among the best at beating opponents in some sort of conflict, ranging from a chess championship to commanders in war, often describe something like “coming to understand their opponent so well that they feel a connection to them” as being a major step in their process of preparing for the engagement.
Wikipedia:
Sympathy is the perception, understanding, and reaction to the distress or need of another life form. This empathic concern is driven by a switch in viewpoint, from a personal perspective to the perspective of another group or individual who is in need.
Oops but shows how closely related empathy and sympathy are 🙂
Are you maybe referring to cognitive empathy?
Paul Bloom is referring to emotional empathy: putting yourself in the shoes of others; feeling their pain and suffering.
This is what I mean. Putting yourself in someone else’s shoes is not an emotion. It is a cognitive ability. It can lead to emotions, and of course basing judgments on emotions is not always a good idea. But putting yourself in someone else’s shoes, ie, creating an accurate mental model of what they are experiencing or what their perspective is can equally well be a part of reaching a reasoned conclusion.
Imagine you spot an old flame while you’re out and about. She doesn’t see you. You are filled with different emotions. The fact that your brain was capable of recognizing her is not an emotion, but it led to emotions.
Empathy is an ability. Sympathy is an emotion.
Choices can be considered ‘free’ even if they are deterministic (obey the laws of physics) insofar as they are done in accordance with the desires of the agent.
When a magician asks someone to pick a card from the deck, sometimes the choice is ‘free’ (the person can pick any card they like), whereas other times the choice is ‘forced’ by the magician such that the person is made to pick the specific card desired by the magician.
In my view, the ‘free will’ debate is really (or should be) about dualism vs materialism, not about determinism or indeterminism.
I think these studies are basically apologetics, regardless of the conclusions a particular study may draw.
If you believe in free will, then you probably believe that believing in free will makes people more moral, and that believing in determinism makes them less moral. If you believe in determinism, by contrast, you probably *don’t* believe that believing in determinism makes people any less moral.
No one actually changes their mind based on these sorts of studies. With very few exceptions, they accept the studies that yield the conclusions consistent with the view they came to the table with, and reject the ones that don’t. For example, Dr. Coyne has probably never made a blog post touting a study indicating that determinism makes people less moral. Like everyone else, he draws attention to the studies that support his own preconceptions.
So why does anyone do these studies? I think it’s due to scientism. Many people are reluctant to admit that their beliefs about this follow from their philosophical commitments, so they feel the need to find a “study” that says that they are right.
Well no, because you’re confusing two different issues here:
1. Is free will/determinism/other true?
2. What happens when people believe/are convinced that free will etc. are true?
The studies, including the one in the OP, are largely concerned with the latter. As Jerry himself points out, using the studies as support for the case of the former is nonsense. Determinism is true or it is not true: how people response when they believe either way is an appeal to consequences, a fallacy he accuses some compatibilists of making.
In any case, it’s not a philosophical commitment to point out reasons why one case is stronger than another.
Not that I don’t agree with the cynical “seek evidence to support predetermined case” idea, but it doesn’t apply here. Besides, that ain’t scientism: that’s just straightforward bad science and bad reasoning. Most of the time, “scientism” is little more than an anti-science boo word used by people who value academic PR over honesty.
Compatibilists do accept determinism.
The most important thing(s) to grasp about determinism is what it doesn’t imply. It doesn’t imply that the future isn’t open, i.e., dependent on our thoughts and attitudes. It doesn’t imply that people aren’t responsible for their actions. It doesn’t even imply that there’s no such thing as probability (read up on how the Everett interpretation can provide for quantum probabilities). The key to all these realizations is to use modern scientific understandings of causality and time. If you use your intuitive, quasi-Aristotelian physics to understand “causality”, then that really is incompatible with free will. But that isn’t the kind of causality science delivers. Causal laws describe patterns, not the coercion of events by earlier events.
The “Little People argument” is a bit of a straw man. Doubtless there are a few “compatibilists” who make it, just as there are a few adherents of any viewpoint who will make almost any dumb argument you care to name. But the serious thinkers, no. Dennett in particular is not guilty of making the argument.
(I apologize if this becomes a double post. My browser apparently ate my first try.)
Dennett in fact has made this argument; it’s in there twice in his essay, “Erasmus: Sometimes a Spin Doctor is right.” Here’s one of the two statements:
If you take “illusion” as I do, meaning “something that’s other than what it seems”, and the view that most people have dualistic free will, then this is the Little People argument perfectly expressed. Now you can construe it otherwise, but I’d see that as an example of confirmation bias.
Thanks for responding and for bringing new evidence to the table. New to me, anyway. But I don’t think the Erasmus paper establishes your criticism of Dennett. Quite the opposite, when the whole paper takes shape.
The quoted passage looks kinda like a Little People argument. But it also looks much like a “why free will is worth writing about” argument. That’s what I thought at first. But reading on, I think Gregory Kusnick has it right: Dennett is warning against hasty inference that goes beyond the scientific evidence into philosophical interpretation.
Dennett accuses some free-will skeptics of making the Little People argument:
I think it’s pretty clear: Dennett is against the Little People argument.
Dan Dennett is one of the top two or three philosophers that I have the most respect and liking for. I want to start with that to try and get you to accept that when I say that I have no bias against Dan, that I actually am biased in favor of Dan, that you might believe me.
With that in mind, I don’t understand how you can claim that Dan has not made the little people argument while relating his views on freewill. He clearly has, more than once. The clearest example, in my opinion, is during the Moving Naturalism Forward conference hosted by Sean Carroll. I’d recommend watching the whole thing for context, if you haven’t already.
Let me also point out that I don’t think Dan having made the little people argument is a big deal. For one, no ones perfect. For two, little people arguments aren’t always wrong, merely wrong often enough to not be considered good arguments. For three, though the little people argument is often considered unethical I have no doubts about Dan’s ethics. From what I know of him he is one of the more ethical people I know of. But I do disagree with him on this topic, while taking him very seriously.
Dennett’s argument is not that people should be shielded from the truth, but that they should be shielded from reckless misinterpretations of the truth. “You don’t have free will” is not a true statement in Dennett’s opinion, and should not be promoted as such. I think he’s pretty clear about this both in the Erasmus lecture and in Moving Naturalism Forward (see the bits I quoted in the comment linked by Paul above).
I don’t see Dennett’s position on this as that much different from Jerry’s position on epigenetics and “Darwin was wrong!” Both Jerry and Dennett are trying to combat bad science with good science. That’s not a Little People argument.
I respect your argument and recall you asserting it whenever this topic comes up, but I disagree.
Dan does agree that that is an accurate statement with respect to dualistic and other magical conceptions of freewill that Incompatiblists mean when they make that statement. I don’t see any interpretation of Dan’s misgivings about the average person hearing an Incompatiblist account of freewill that doesn’t include concern that the average person won’t be able to understand it. Or understand Dan’s Compatiblist account for that matter. After all, the average person has just as easy access to Compatiblist accounts of freewill as to Incompatiblist ones.
Dan knows what people like Jerry mean when they say freewill, he agrees it doesn’t exist, he knows that the freewill that he is a proponent of is different than the one Jerry says doesn’t exist, and he knows that Jerry knows that. So if Jerry and Dan can understand all that, why be concerned that the average Joe can’t? Maybe concern about that isn’t misplaced, but it’s still the little people argument.
Here’s the difference between your interpretation and mine:
“You don’t have free will”, without qualification, is not just a claim about dualist or magical free will. It makes the stronger claim that there is nothing you have that could be reasonably called free will.
It’s that stronger claim that Dennett objects to — not because the Little People can’t understand it (he thinks they can), but because in his opinion (and mine) it’s not an accurate statement of the facts, and scientists shouldn’t go around authoritatively asserting claims that aren’t accurate.
A problem with that interpretation is “without qualification.” People on both sides of this issue qualify their terms and what they mean all the time. Telling Incompatiblists that they shouldn’t push their ideas because if ordinary people hear an unqualified “You don’t have freewill” is a bit of a strawman argument.
Or do you mean that people can’t be expected to ask some questions or do some minor internet research if they hear a scientist say “Freewill doesn’t exist”? I’d say that is also a little people argument. And, again, concern about that might not be misplaced, but still LPA.
If that’s a Little People argument, then (again) so is the fear that people will get the wrong idea about evolution from sensationalized popular accounts of epigenetics.
You are claiming that the Compatiblist account of free will is as evidentially validated as Evolution and that the Incompatiblist account is as evidentially invalidated as sensationalized popular accounts of epigenetics? That would be our largest area of disagreement. They are not even in the same category let alone of comparable magnitudes.
No hypotheses or accounts of freewill are comparable, as of yet, to any of the major well established, well validated theories of modern science. The C vs IC debate is, currently, a philosophical debate and we don’t know enough yet to take it beyond that.
No, I’m saying that sensationalized accounts of neuroscience that proclaim “Free will is dead!” are like sensationalized accounts of epigenetics that proclaim “Darwin was wrong!”
If it’s true, as you say, that we don’t yet know enough to take the free will debate beyond philosophy, then you presumably agree with Dennett that such accounts ought not to exist, and that neuroscientists ought to exercise restraint in interpreting their findings, just as Jerry urges epigeneticists to do.
So we aren’t talking about Incompatiblist accounts of freewill but merely sensationalized accounts of neuroscience that proclaim “Free will is dead!” without qualifying what they mean? I agree with that, on nearly any subject. But that is not all that Dan has argued against.
Exactly. We’ve gone over this before here (and I remember posting various links/quote showing Dennett is explicitly not making the little people argument – I think he was even explicit about it in his talk with Sam Harris on free will, as well).
It’s disappointing to still see Dan pegged with this mischaracterisation.
Great first paragraph.
“Even though all rational people know that determinism rules human behavior,”
As a pessimistic incompatibilist, I disagree on a technicality. It is, at best, an approximation – and not a foolproof one – of what happens at our everyday level of existence, but ultimately the causality supporting it is a special case of the mathematical functions of quantum mechanics, which I’m given to understand are probabilistic rather than, as determinism would require, absolute. I certainly wouldn’t say we “know” determinism is true: not without qualification.
“Our sense of agency is so strong that it’s impossible for many of us to accept determinism of our behavior,”
I suspect this is more cultural conditioning than anything else. Myself, I don’t feel like I’m a free agent because I’m repeatedly aware of how much I’m blown about by the random winds of perception and cogitation: my senses, discoveries, urges, desires, interests, thoughts, reasons, beliefs, and emotions. All of these, as Sam Harris points out, emerge seemingly out of nowhere. Even when I gain control of the system – say, during concentrated activities, when I’m in a heightened emotional state, or when I’m in the zone – I’m aware that that’s entirely contingent on me feeling like gaining control. If I didn’t feel like it, it doesn’t happen until, for whatever reason, I do feel like it. And whether that feeling’s there or not is a matter of historical contingency, which in my limited perception is little different from a historical lottery. In short, even when I’m unambiguously a cause, I still see myself as an effect too.
As for the current experiment, as you say it’s nothing particularly impressive, but in my case I also frown at the low sample size. Forty people is too small a number from which one can make many generalizations. But even if the evidence was clear-cut and robust, that would be an empirical bonus, not strictly speaking necessary to the case for determinism. Otherwise, that argument would amount to a fallacious appeal to consequences.
Interesting question, though. I’d love to see more studies into how our beliefs on causality and human cognition affect behaviours.
Conversely, even when we’re an effect, it can be an effect of our own prior causes. We can reprogram our brains with new habits and new preferences that change what we feel like doing in the future.
“Conversely, even when we’re an effect, it can be an effect of our own prior causes.”
Which is immaterial to the present effect. My own prior behaviour is simply another contingency like any other, no more special than the weather or the social climate of yesterday.
This is not a way out of the mindset in any case; the prior causes are themselves effects of prior causes that have nothing to do with us. A canoe doesn’t gain any magical properties just because it’s floating on another canoe floating on the water. It still responds lawfully to the overall ebb and flow.
Unless, of course, you think there’s something causally magical about whether or not a prior cause is “our own”?
This is why folks like Kane try to figure out some sort of “self origination”.
Not magical, no. But a powerboat with an engine and onboard control system can go places that an empty canoe drifting passively with the current can’t go. The recognition that we are causes as well as effects — more like the powerboat than the canoe — is itself causally relevant.
I think you misunderstand the metaphor. It’s not a comparison of autonomous device versus non-autonomous device. The sea represents the field of cause-and-effect, the canoe represents one’s position within that field (if it helps, a dot on a traditional Cartesian graph will suffice). You can assign as many canoes to the spot as you like: the things which you had done causing the things you did, which in turn cause the thing you do, which in turn will cause the thing you are about to do. But you’re still floating on the causal sea, still a function on a Cartesian graph. On that metaphor, powerboats are magical free will.
Or if the distinction between canoe and sea is too confusing because it suggests a fundamental difference between oneself and causality, try links in a series of criss-crossing, many-to-many chains.
What you’re doing is highlighting the links that correspond to an individual human’s lifespan, and saying “these links are special” because… they have internal cause-and-effect.
Fundamentally, they’re not special. They’re made of the same material. They follow the rules that apply to all the chains. There is no escape clause.
That’s because indicating human uniqueness by pointing out how one’s prior chains link to one’s later chains in the human web… is no more causally spectacular than indicating a star’s uniqueness by pointing out how its prior chains link to its later chains in the star’s web. But no one sane is arguing that star evolution proves stars have rudimentary free will.
Whether you’re an effect, an effect-cause-effect, or an effect-cause-effect-cause- … -effect makes no difference to the working. I wouldn’t make much of internal causality when it’s ultimately dependent on external causality.
You’re barking up the wrong “free”. Internal cause-and-effect loops aren’t special because they’re supposed (by some) to be un-physical. They’re special because that’s what self-governance is. Compare erotic love. It isn’t because it comes from a god named Eros, that it’s special. It’s special because it’s fricking *erotic love*. The god-story is impressive sounding, but ultimately the phenomenon speaks for itself.
Reading through the comments reminded me of experiences I had in college with highly religious people. For a couple of years I lived in a private dorm that also housed a sizable contingent of Lebanese students, both Christian and Muslim. Tensions were often high. I became friends with some Christians and some Muslims.
One thing that really struck me, even back then, was how fatalistic all of them were. Regardless of how they may have answered questions about freewill, they behaved as if they fully believed in determinism. To an extent that sometimes seemed to be more like a caricature that might have been made by freewill proponents to ridicule determinism. It was alien to my experiences so far, 18-19 years old, in the US and Western Europe.