Bryan Caplan  

Reduction to Banality

Robin Hanson on Health Care... Better than Health Insurance?...

Last week, Tyler Cowen blogged on a story about a special ed program that uses electric shocks to make students to behave. His post included the somewhat cryptic remark that "I view this as a reductio ad absurdum on Bryan Caplan's view that mental illness does not exist."

What is Tyler getting at? As best as I can tell, he is alluding to my argument that if you can change someone's behavior by changing incentives, then he could have done differently all along. And if a person could do X but decides not to, it is silly to blame his decision - however weird - on a "disease." If an alcoholic stops drinking because his wife threatens to leave him, we can infer that heavy drinking was a choice he made, not a disease that happened to him.

So does this article reduce me to absurdity? I don't see why it would. The article shows that you can make the mentally retarded sit still by shocking them when they fail to sit still. Does this "prove" that mental retardation is a choice? Hardly. To do that, you would have to show that shocking them could raise their IQ up to normal levels - which it clearly can't.

So what does the fact that you can make the retarded sit still by threatening to shock them prove? It proves the obvious point that, besides being low in intelligence, the retarded also have unusual preferences. For example, they dislike sitting still more than the average adult does. They also probably prefer cartoons to opera. And if you raise the cost of satisfying their unusual preferences, the mentally retarded change their behavior, just like the rest of us. Big surprise.

Comments and Sharing

COMMENTS (10 to date)
Martin Kelly writes:

Och, Bryan, away and take a loan of yourself!

Firstly, don't you realise that this post seems incredibly self-important and conceited?

Secondly, of course the retarded have preferences - they don't want to suffer pain!

Thirdly, have collegiate relations within the economics dept at GMU broken down to the extent that yourself and Tyler Cowen communicate via blog postings?

And does Boudreaux know?

What surprises me is that people consider this position untenable. Isn't it... obvious? Like Heinlein said, nobody does anything they don't want to do - some people are just more honest about it than others.

I believe that most people are unwilling to control their own behavior beyond a certain point, and once they get to that point they prefer not to justify it... so we culturally enable them to blame an illness or a disease over which they can *have* no control, and they are absolved of the responsibility to even attempt any behavioral control.

Of course, if you happen to subscribe to any theory of personal responsibility, that concept is just retarded.

Matt McIntosh writes:

Well, how far do you want to take this? You can apply the same kind of conditioned response learning to mice. Would you say that mice make choices? Just trying to tease out what implications of your position you're willing to accept...

Constant writes:

Well, it seems to me that the following two claims are quite different from each other:

1) Mental illness does not exist. (or, to put it more softly, some things commonly considered to be mental illnesses are not in fact mental illnesses)

2) Mental illness is a matter of having unusual preferences rather than some sort of mental incapacity.

I can readily accept that some mental illnesses are in fact nothing other than unusual, possibly extreme preferences. But it does not follow that the people who have these extreme preferences are not mentally ill. Surely there can conceivably be such a thing as a healthy appetite and an unhealthy appetite - a healthy preference and an unhealthy preference.

Surely the preferences of some of the criminally insane - e.g., a preference for destroying other humans - can arguably be called "sick", and in need of curing.

I think Constant hits on an interesting point: sometimes an unusual preference is the result of a mental dysfunction.

What traditionally drives the distinction is an inability or unwillingness to alter one's preference when it is inappropriate. At that time, we declare the preference a mental dysfunction.

The flaw in this is that propriety is itself an unusual preference which should be occasionally altered for the comfort of others. An inability or unwillingness to do so is its own mental dysfunction.

Eric Crampton writes:

I'd tend to see mental illness more as the combination of extreme preferences and inadequate preference-repression technologies. Specify that not acting on impulses consistent with innate preferences is increasing in the distance between ideally preferred behaviour and socially acceptable behaviour, and put some alpha multiplier on the cost function such that alpha can vary across individuals. The special ed program in the article would then be working in the short term to make the cost of exhibiting the behaviour higher than the costs of not exhibiting the behaviour, and in the long run to improve the repression technology embodied in the alpha.

Mightn't a fair bit of the action be in heterogeneous technologies for repressing extreme preferences? Specifying a model where preferences are homogeneous across people, with every person having preferences on one dimension that are deemed antisocial, variance in behavior then is exclusively a function of heterogeneity in repression technologies. More realistically, preferences and technology are normally distributed but the combination of extreme-tail preference and low-tail repression technology gets called mental illness.

Ragerz writes:

I think that Caplan has a simplistic and I would say crude model of "choice." And that simplistic and crude model may account for his libertarian tendencies.

Lets take the alcoholic. If you had any experience with people, you would realize that some are more disciplined than others. That is, some people are more able to align their ideals with their actions than others. Getting drunk is pleasurable - for many, the most pleasurable thing they could do in a given period of time. For some, to maximize utility, alcohol consumption is the way to go. Now, this may be true for two people with identical utility functions from alcohol consumption and identical ideals with respect to the normative question of whether they ought to drink alcohol. But then one, the more disciplined one, does not drink, while the other, the less disciplined one, does. One maximizes utility while the other ensures that their actions are aligned with their ideals. Of course, the person who aligns their actions with their ideals, is disciplined, and is more admirable.

If an alcoholic quits using alcohol because his wife threatens to leave him, it is nonsense to suppose that he had a "choice" all along. This person is not disciplined. Losing his wife reduces his utility more than losing alcohol. He has no "choice" but to maximize utility, given his lack of discipline. But for the threat by the wife, he would have been unable to quit drinking. It is not the case that he would have been able to quit, absent that threat. He is nothing more than a short-term utility maximizing program.

If No Threat
--Then Drink Alcohol;
Else If Threat
--Then Don't Drink Alcohol;
(Sounds like a younger, less disciplined version of George W. Bush)

To call this "choice" is to fail to recognize the difference between Pavlov's dogs and a disciplined human who sacrifices pleasure for ideals. Which is something much more noble and great than the so-called "rational" man that merely maximizes utility. (I would argue that the utility maximizer is the opposite of rational.)

Fortunately for those who sometimes lack discipline, discipline is something that can be increased and is something that varies with context. A person might be very disciplined about going to the gym, but still have bad eating habits. A person might be disciplined about writing academic papers, but then fail to get enough exercise because they are obsessed with some video game. However, a person does not simply CHOOSE the level of discipline they have in a given period of time. If only it were so simple. (Otherwise we would have less people spending too much time playing video games, eating unhealthy food, and avoiding the gym.) It is something that is cultivated; a parent teaches their children discipline. If that doesn't occur, maybe the child learns it elsewhere. Or maybe they don't learn it. But for those who don't learn discipline, it is nonsense to say that they "choose" to be undisciplined to the extent that level of discipline is dependent on parental upbringing or other unchosen experiences in life. Did you "choose" your parents??

Discipline is also something that one must continue to exert. What obese person has not refrained from eating a Ding-Dong for lunch?? But then, that person cancels out the benefits of that "choice" with a subsequent "choice" to have one for dinner. I submit, that the difference in Ding-Dong eating is not likely to be the amount of pleasure that eating the Ding-Dong will bring at dinner rather than lunch. Rather, it is a matter of how focused that person is on their goals and ideals at lunch versus dinner. Which is in terms a function of a person's ability to focus on what matters and not be distracted; that is, a function of discipline. The undiscplined obese person who does not eat a Ding Dong for lunch just got lucky; they just happened to be focused on losing weight during lunch. (Or it did not occur to them to eat a Ding Dong at all during lunch - perhaps before dinner, they saw an advertisement for Ding Dongs that they didn't see during lunch.) With a more disciplined person, it wasn't a matter of luck. It was a matter of having cultivated an ability to focus on their ideals and be true to those ideals.

Overall, your concept of "choice" is not descriptive of reality.

You write: "It proves the obvious point that, besides being low in intelligence, the retarded also have unusual preferences. For example, they dislike sitting still more than the average adult does."

Here is a better explanation. It is not that the retarded have unusual preferences as much as it is that they lack discipline. They are truly an example of Pavlov's dog; only through shocks can their behavior be controlled. People who are not retarded are not merely blessed with a different utility function; rather, it it is a matter of being more disciplined. Or having the ability to disregard their utility function.

Has it ever occured to you that people who "choose" to do nothing but maximize utility are not free, but are rather slaves? Only through discipline does one become free. In turn, whether one is disciplined or not is itself not entirely a matter of choice, but also a matter of upbringing and unchosen experiences and even personal epiphany.

An interesting question is to what extent your flawed conception of "choice" feeds into your political ideology. For example, of course we should leave vending machines full of junk food in high schools. Because of course, whether students eat the food is simply a matter of "choice." And of course, "coercion" doesn't exist in the private sector, because all one need to is "choose" to get a different job. And on and on.

- Ragerz, the ex-libertarian

Constant writes:

Ragerz -

"Lets take the alcoholic. If you had any experience with people, you would realize that some are more disciplined than others."

This distinction could be captured by time preference (preferring momentary gains over long term gains). Thus, it can be explained as a difference in preferences, rather than a sameness of preference combined with a difference in discipline.

If you subscribe to the notion that "preference" is just a structured way of talking about action, then a person simply by definition cannot set aside his own preferences: whatever he does, he reveals that he prefers to do it.

To put it more straightforwardly: your discussion of alcoholism and discipline is deeply nonsensical to someone who has learned and internalized the concept of preference as an economic concept. Where you write, "the ability to disregard their utility function," - that is nonsensical. It's like saying, "the ability to choose what one does not choose."

Finally, there is nothing particularly libertarian about it - it is a matter of employing economic concepts.

Where you write, "Has it ever occurred to you that people who "choose" to do nothing but maximize utility are not free, but are rather slaves?" - you are just abusing the concept of utility, to mean something other than what it means in economic discussion. It's as though you had said, "has it ever occurred to you that people who choose what they choose are not free."

BT writes:

Alcoholism is a disease not a choice. Likewise mental retardation is a disorder with a biologcal cause not a choice. Economics assumes the ability to utilize one's mind. Mental retardation and any addiction means that a person has lost the ability to think rationally.

Bryan do get out of the ivory tower at GMU. Not everythiing in life can be explained by economics and role playing sessions.

Ragerz writes:


I thought I would respond to a few points.

"This distinction could be captured by time preference (preferring momentary gains over long term gains)."


It cannot be explained that way. A disciplined person would abstain over alcohol both in the long-run and the short-run, even if they get the same enjoyment from alcohol consumption as someone else, and even if they value the health benefits equally.

Discipline is not simply about a more sophisticated hedonism, where one restrains ones inclinations now so one can splurge later. It is about aligning ones actions with ideals. To some people, who are greater and more noble than hedonists, it isn't about "gain," either in the short-run or the long-run.

You write:
"If you subscribe to the notion that "preference" is just a structured way of talking about action, then a person simply by definition cannot set aside his own preferences: whatever he does, he reveals that he prefers to do it."

I suppose one could define preference that way, but then it is completely meaningless. If by definition all of our actions are aligned with our preferences, we might as well not talk about preferences at all. If someone points a gun at my head, by your definition of preference, I might "prefer" to give them my money. Your definition of preference is not very useful, and it does not capture a very important distinction that is captured by a more ordinary, and I would say superior, definition of preference. Only through a more narrow definition, can the concept have any meaning.

"Your discussion of alcoholism and discipline is deeply nonsensical."

If you think that a discussion of discipline and alcohol is "nonsensical" that may be an indicator that you lack discipline and thus fail to understand it. Alternatively and more likely, that you think that any discussion that goes "out of bounds" of your meaningless (you say "economic") definition of "preference" is nonsensical.

As far as your definition of "preference" as an "economic concept," it is not the case that every economist would agree with you concerning the best meaning of "preference" as an "economic concept." The terms utility and preference are elastic enough to perhaps cover, say, the noble action of living in alignment with your ideals, if you adopt an overly broad definition. Unfortunately, they would still also cover actions that are vile. And further, since defined so broadly all actions are said to be utility maximizing actions, these definitions are useless, except as a means to obscure and obliterate important distinctions. Obviously, one should not put the sublime with the vile together as if they are the same thing.

And this is my criticism. If one says that people are "rational" because they "maximize utility" and every action maximizes utility (or is a reflection of preference), that means that everyone, by definition, is rational. Which makes the whole concept of rationality meaningless. There are better, more discerning definitions of rationality. Just as if all actions reflect preferences, the concept of preference is meaningless. One must narrow one's definitions for them to have any meaning. If Caplan says that people "choose" to do whatever they have done, because by definition all actions are "chosen" then I would have to say this is a meaningless and useless definition of choice. Do I "choose" to give the guy with the gun to my head my money? Using the most broad and definition of choice, yes. Using a more sensible and discerning definition of choice, no. Maybe some would say this is merely a case of revealed preferences. I revealed my preference to give my money to the guy with the gun. But wouldn't this be missing something?

Ultimately, what is missing here is any distinction between mere inclination and those actions that are the result of disciplined choice. And disciplined choice is the only kind of choice that is truly free. If you think that a discussion of alcoholism and discipline is "nonsensical" I can only say that these unfortunate definitions chosen by some (but hardly all) economists have managed to make you deeply confused. Because some economists don't talk about discipline, you apparently think it is irrelevant as an "economic concept." If you are right, I think that would render economics fairly useless. Why should we listen to the policy positions of any economist who aggregates the utility of watching porn, engaging in prostitution, drinking alcohol and weighs it against going to the gym, getting medical care to those who need it, preventing murder, and ensuring that people have access to disease-free drinking water?? If you were right, then economics would be a rather useless subject for making policy choices. At least to any person interested in more than amoral and immoral hedonism.

Happily, nothing in economics prevents us from making a distinction between "maximizing utility" which can only be rendered meaningful by narrowing it in some way, and "disciplined action." And I would put under the rubric of "maximizing utility" short-term sacrifices for long-term indulgence. This distinction does not save you, because more sophisticated long-term hedonism is really no better than short-term hedonism both of which must be distinguished from disciplined action. (Which itself might occur over either the long or short term.)

I prefer a more robust sort of economics. It is not normatively neutral to aggregate the pleasure people derive from porn and weigh it against the "utility" that people derive from drinking clean disease-free water. To the extent that your definitions fail to enable you to distinguish the two, your definitions should be amended or, alternatively, your policy conclusion discarded.

With respect to your assertion that there is "nothing libertarian about it," where "it" refers to the adoption of overly broad definitions of "preference" and "utility" I would have to agree. Libertarians are often concerned with rights and many libertarians think such rights may not be compromised in the name of maximizing aggregate societal utility. That is, they distinguish between higher goods and crass maximization in at least some contexts.

However, I wasn't criticizing Caplan for being a utilitarian or for having an overly broad definition of utility. I was criticizing him for having an overly broad definition of choice. I am simply pointing out that Caplan's definition fails "to recognize the difference between Pavlov's dogs and a disciplined human who sacrifices pleasure for ideals." And as such is a bad definition. His adoption of an overly broad definition has led him to miss important distinctions and overlook important data. If our definitions lead us to group things and pragmatically assume they have like properties, (i.e. this is an X, and this is a Y. All things that are X can be considered alike for the purpose of this analysis...), I am simply pointing out that all the data you put in the "choice" category doesn't really belong there. It belongs somewhere else. To adopt overly broad categories (definitions) is to obscure reality. In this case, I think that Caplan's overly broad definition of choice might be one of the key contributing factors to his libertarianism. (Or maybe not. I can't claim to read his mind.) Whatever the case, Tyler Cowen was perfectly correct in criticizing Caplan's conception of choice and Caplan's defense here fell flat.

-Ragerz, the ex-libertarian

Comments for this entry have been closed
Return to top