Bryan Caplan  

From Intuitionism to Contrarianism: A Case Study

From Poverty to Prosperity<... A Good Line...
As an undergraduate, I spent hundreds of hours pondering the foundations of morality, also known as "meta-ethics."  In the end, the young Michael Huemer converted me to ethical intuitionism, a view I've held ever since.  (BTW, a decade or so later, Huemer literally wrote the book on the subject). 

Here's how I explained my intuitionism in the Caplan-Hanson debate:
Sensible moral reasoning begins with concrete, specific cases.  For example: It would be wrong for me to walk over to Robin right now and punch him.  From there, we can start to generalize.  It would probably be wrong for me to walk over and punch any of the people in this room.  At the same time, we can note exceptions.  If Robin had consented to box me, then punching him would be OK.  In fact, it would probably be wrong not to try to punch him, because I'd be cheating you, the audience.
Based on this passage, you would conclude that my moral views are extremely mainstream.  But they're not.  Indeed, I often think that conventional "morality" is evil.  Question: How do you get from intuitionist meta-ethics to contrarian ethical conclusions?  That seems like quite a trick.

The answer: You have to show that seemingly plausible ethical views conflict with other, even more plausible views. 

Take for example the view that employers should not be allowed to discriminate.  I'll admit that this conventional view is somewhat plausible.  To rebut it, you would need to show that it conflicts with a more plausible principle.   Like what?  There are many possibilities, but Steve Landsburg provides one in his latest book.  He points out that almost everyone thinks that employees should be allowed to discriminate against employers.  The two cases seem very similar, and the second intuition is much stronger than the first.  So it looks like the first intuition has to go.

Of course, this isn't absolute proof - which frankly is so rarely obtained that it isn't even worth looking for it.  But is Landsburg's argument any good at all?   Justin Martyr, who largely agrees with my meta-ethics, isn't convinced:
The gist of the argument is that we grant whites the freedom not work for blacks (because whites do not have to apply to firms owned by blacks) but we don't grant whites the ability to not hire blacks.

I think the distinction is clear. Employers have bargaining power and employees do not. As long as the market is monopolistically competitive then employers will have at least some bargaining power. A racist with bargaining power has the ability to harm the welfare of others; a racist without bargaining power is merely entertaining his own private thoughts.
Justin's trying to bolster the plausibility of the conventional view.  I've no objection in principle; an honest intuitionist should consider this possibility.  Still, I think his effort fails.  Yes, in the real world of imperfect competition, employers have some bargaining power.  But so do employees!  A worker who holds out for a wage 1% above the usual level often manages to find a job anyway.  

You could admit that both have bargaining power, but insist that employers have more.  Empirically, however, standard tests of race and gender discrimination find little evidence that discrimination reduces wages.  Either employers don't have much bargaining power, or they rarely bother to use it for discriminatory ends.

In any case, even if one side has lots more bargaining power than the other, it is counter-intuitive to claim that a coercive response is justified.  Suppose A and B be are dating.  A has an equally good outside option.  B can't bear to live without A.  A therefore has some bargaining power - vastly more than most employers, in fact.  Yet almost everyone thinks it would be wrong to force A to stay with B. 

You could say that if two people start with the same meta-ethics yet disagree about a relatively simple issue, this shows that their meta-ethics are wrong.  The cheap and easy reply is to point out that the same problem afflicts all meta-ethical theories; intuitionism is no worse on this point than any other view. 

But that's selling my own view short.  Intuitionism is better than the competition at handling disagreement.  How so?  Unlike most other moral theories, intuitionism doesn't pretend to derive ordinary ethical judgments from some Overarching Indisputable Principle.  So when intuitionists disagree about X, they usually try to argue about X, instead of searching for flaws in an obscurantist seventeen-step "proof" that X is wrong.  So while I'm not confident that I can change Justin's mind about employment discrimination, I do think that my chance is a lot better than it would be if we shared some other meta-ethics.

Comments and Sharing

COMMENTS (21 to date)
dWj writes:

I think most people are ethical intuitionists, as I understand it from this post; they're just very bad at it.

Doc Merlin writes:

I go one step further and say that ethicists are ethical intuitionists they just don't know it. From listening to ethicists debate, and reading their papers, the typical debate goes, "Your theory is wrong because it leads to x, which is obviously wrong."

In addition Paul in the New Testament uses a form of ethical intuitionism (only coupled with partial subjectivism) to say that its not immoral to eat meat offered to idols, but if someone feels guilty about doing so, then it is still wrong for them to do so.

hacs writes:

That seems "coherence" for me.

If A and B are "incoherent" (in some sense) and B is more plausible (in some sense) than A, then A should be false.

I do even not know whether "coherence" should be a necessary condition in applications of ethical analysis.

Is that "coherence" neutral (content independent)?

Are A and B really comparable?

If B is more "plausible" than A, C is more "plausible" than B, A and B are "incoherent", B and C are "incoherent", but A and C are not "incoherent", then should B be false? Should A and C be true? And, if C is not reminded in the evaluation? Should B be considered momentarily true? If yes, so that is not really true, but consensus.

And so forth...

Kurbla writes:

Yes, majority of people are "meta-ethical intuitionist", at least everyone who is willing to check his ethics against reality on similar way we checks other theories. There are many people, however, who believe that suspect in existing theory is the work of Satan.

Your illustration of the racism control is good one. I think it is not about bargain power, but about object of the state control - in the case of employer, it is state control over capital, in case of worker, it is state control over worker's body. In the spirit of the post, I'll leave it here.

Andy writes:

The best evidence I've seen that discrimination matters empirically is the recent Guryan and Charles paper:

It's not a 'standard' test for discrimination, it instead tests Becker's model of discrimination and how the distribution of racism determines the market-clearing discriminatory wage.

johnleemk writes:

One issue I have here is that the empirical evidence in this debate (on the specific issue of discrimination) can only go so far, literally -- all the studies brought up so far seem relevant only in an American context. Discrimination in the workplace is an issue in many countries and regions beyond North America.

Kevin writes:

Bryan, you can engage in this process of intuition harmonizing without being an intuitionist. An intuitionist like Mike thinks that intuitions indicate external, mind-independent moral truths. Rawls thought they were information about the shared moral ideas of our culture. Yet both routinely use the process you describe to approach more justifiably principles.

pjsw writes:

I second Kevin's sentiments above.

Also, doesn't an intuitionist (like cultural relativists) have a difficult time explaining moral progress? If "Sensible moral reasoning begins with concrete, specific cases," and you think that the intuitions *constitute* moral rightness/wrongness (instead of being evidence for independent rightness/wrongness), and most people get the intuition that, e.g., white people are superior to black people, then it would seem you have no basis for explaining the moral rightness of civil disobedience and the like.

The mistake being made here is the assumption of simple causality. One intuition = one ethical action. In fact, what we may have here are two intuitions at work for one ethical action, and only one for the other. I'm not saying that that is in fact the case (and, in the end, I am probably with Caplan on the specific case he gives), but it is something we need to think about.

Marc Hauser in fact argues that we evolved a set of intuitions that underlie our ethics and what we end up doing is using moral reasoning after the fact (after the action in question). This does not mean that we cannot fine-tune our intuition with moral reasoning (a la Hayek), but it hardly comes first.

John writes:


I agree with Kevin, your ethical methodology may involve reasoning from concrete cases to principles through intuition and your normative ethics may be fairly consequentialist without being intuitionist in the Huemerian sense. Why should we think that there are external moral facts that have mind-independent truth values? We need not infer any metaphysical consequences from our preferred methodology.

Murali writes:

A few points

1. The problem with intuitionism is that neither common agreement, nor even the "considered judgements" of the wisest members of society are in anyway dispositive of the right

2. The problem of intuitionism leading to relativism is very real. 150 years ago, americans had very different intuitions about slavery.

3. You cannot simply assert that punching tyler cowen in the face is immoral. You have to demonstrate it.

3.1 What are the wrong/bad making features?

3.2 Why are those features wrong/bad making?

4. the answer to questions 3.1 and 3.2 are very important when it comes to asserting whether something else is right or wrong. They may or may not share the relevant right/wrong making features.

5. Even though we commonly use our intuitions in our daily life and only justify them after the fact, it does not follow that there can be no rigorous justifications for our moral judgements.

6. The 17 step obscure justification has the upshot that if it indeed works, the conclusion thereby reached cannot be denied by any rational being.

7. The mere fact that most 17 step arguments out there are not completely sound, does not preclude the possibility of there being a working one. At each step, when explicitly presented, you could poke holes if possible in the logic of it.

In fact, accessing my website would in fact give you a good example of what at least part of a moral theory that did this would look like. Each step is explicitly stated and numbered. Any particular point can therefore be addressed individually.

Thomas Sowell pointed out that southern streetcar companies before 1900 did not discriminate against blacks, but did separate smokers and non-smokers. Separation by smoking met a market need. Separation by color would have angered customers with no increase in profit.

Further, streetcar companies resisted state and city laws requiring discrimination and separated seating.

So, if companies have to pay for their discriminatory intent, they would rather have the money than be discriminatory. Competition opposes discrimination, other than for economic reasons.

Discrimination in the south on streetcars and buses only took hold after the government bought and monopolized those services. Those governments then imposed separate seating regardless of the cost.

Does this sound familiar? Government imposes a monopoly (rules and regulations), then enforces them regardless of the cost.

People forget that oppressive, effective discrimination was only possible through government power.

Reference (search for "When streetcars")

Urstoff writes:

It seems that an intuitionist using a sort of reflective equilibrium method will still end up with (at best) a sort of cultural relativism. A population that holds similar intuitions may be able to systematize them, but a population with differing intuitions (even if the weighting is merely different) may end up with a different system. Call it indeterminacy of moral reflection.

Second, of course, is the problem of the questionable metaphysical status of external reasons. Normativity without means-ends relations is downright spooky.

David J writes:

1) True "conflicts" are rare or impossible since there is almost always some difference between any two situations. Without addressing these differences any assertion of conflict comes across as incomplete. When we discriminate we often "intuit" these differences - the other half of "intuiting" the similarities between situations.

2) Employee vs Employer and Dating - we treat the employer as a legal entity that is restricted in its ability to discriminate in legal "contract" matters; whereas the employee is a human than can choose to act discriminately in personal matters and those counter-party to a legal entity.

3) Intuitionism -> Relativism is a matter of reality; and is either by design or a natural consequence of the lack of an absolute truth.

The main difference between an employer and an employee is that the employee is a true human entity while the other is a legal entity. Humans may do the hiring but the corporate structure, however formed, is not a person; employees agree to work for the benefit of the company and not for the individuals (not unlike the U.S. government where people agree to serve the Constitution and not whomever is currently holding office).

Thus, in the employment and dating scenarios, we allow human individuals to be discriminatory while we disallow legal entities (e.g, companies and governments) to discriminate.

Even in the cases of a sole-proprietorship we hold this to be the case (if to a lesser extent) since, in principle at least, the corporate and personal life of an entrepreneur are distinct - we allow them to chose who to date and marry but restrict their ability to discriminate within their company.

Of course it is simply to take a handful of cases and distinguish them and then evaluate whether or not what is being done unconsciously is really what people want; but saying that such analysis can justify any position is a stretch - the evaluation still needs to occur after differences are discovered.

Not taking the time to find the root difference and making an evaluation on its relevance is simply being lazy; if two situations are treated differently then discovering internal differences and addressing them is necessary for meaningful analysis. If you want to poke holes in my "legal entity vs. human entity please go ahead; but ignoring that those are/could be important differences and simply saying that in both cases a human being is making the decision is ignoring the fact that a difference must exist. Thus, a true conflict between identical situations is likely rare or non-existent - all situations differ from one another in some way.

As for Murali, Item #2, relativism being a consequence of intuitionism is simply a matter of reality; even if there IS some form of absolute truth the fact that it is not written into our consciousness means that any assertion thereof is just another relative guidepost; and it is this aspect of our reality that leads to intuitionism. The only real consequence of such a reality is that global harmony is impossible and change and conflict will be a permanent feature of the human species. If there is a supreme being that establishes absolute truth the logical conclusion is that this permanent conflict state is by design. If there is none then relativism ethics is as good and logical an explanation as to why we have what we do.

Jesse Prinz offers a moral intuitionist view that doesn't not claim that intuitions reveal universal or mind-independent truths. Since some people have intuitions in which cannibalism is abhorrent, and other people have intuitions in which cannibalism is an honored, ritualized practice, it seems plausible (to me at least) that intuitionism is incompatible with strong claims about finding universal truths.

Such a viewpoint does lead to a relativist outcome, obviously, but not necessarily to abandonment of the idea of moral progress. See Prinz's book, The Emotional Construction of Morals, which addresses moral progress in the final chapter.

CJ Smith writes:


Good post - particularly like the highlighting of Caplan's flawed base conceptions in point 3.


The problem your example demonstrates is that ethical intuitionism as you describe it demonstrably creates fallacious beliefs about "universal" moral truths and dictates. You state a factual case, then jump to a moral determination, without delineating or evaluating the factual, moral and ethical considerations that really underlie your "intuitive" determination. The "17-steppers" you denigrate would at least identify that cost-benefit and maintenance of status quo are probably two of the primary evaluative criteria in your "intuitive" judgment, and that others may exist.

Convert to pure situational relativism - it's generally simpler, easier to apply and less prone to generating fallacious "universal" truths and moral dictates. Your ethical intuitionism is almost there, it just lacks the understanding (or you misrepresent the understanding) that while situational evaluation provides the foundation for all moral determinations, there are no "universal" rules - just evolving rules of thumb and generalizations that allow you to make truncated (intuitive) moral judgments because the evaluator either cannot or chases not to identify and evaluate all the factors in a situation requiring a moral evaluation. Situational relativism would also allow you to understand that for certain people (such as Murali and myself), hitting Robin in the situation you laid out might be a moral toss-up or even a moral action. It would also allow you to consider that even though one or both of us might come to a different decision, that decision is only wrong when you impose your moral framework in preference to ours.

Critics of situational relativism dislike it, because it can be used to justify any action, inaction or moral position, by simply questioning whether the person making the moral determination had all the facts, was consciously or intuitively make the decision, or by denying/disputing the underlying and collateral moral judgments contained in the decision framework. That is exactly what happens in real life. Sometimes the overreaching moral principle "harming others is bad" is outweighed by other moral considerations, such as "immediate cost/benefit," and "status quo."

The reason I believe you might be avoiding espousing situational relativism is that, in order to make a normative comment or imperative, rather than a descriptive observation, the commentator must recognize and acknowledge that:
1. the commentator is engaged in a situationally relevant moral determination;
2. when making a normative evaluation, the commentator's ethical framework contains the better evaluation criteria;
3. even worse, if the commentator shifts from making to a normative evaluation to making a normative imperative, the commentator has to accept that their own framework of situational relativism allows for situation specific normative imperatives that might change, given the next situation or set of facts. This is the huge stumbling block for most borderline situational relativists - having the willingness to acknowledge that they consciously or unconsciously prefer their framework to any other, and that even the well-thought-out imperative they posit may not be applicable in all times to all situations.

Bryan, trust me on this one - no one here would be shocked to learn that you consider your moral framework and judgments superior to everyone else's. While we would be shocked to discover that your stated position on matters can change over time or based upon circumstances, that's probably due more to your style or blogging.

caveat bettor writes:

Bryan: In your framework, is there the construct of a conscience, even if you cannot assert the origination?

Justin Martyr writes:

Hi Bryan,

I'm happy to defer to the professional economist on the state of the literature!

My overriding moral intuition is this: I don't want people unable to live the Good life because of man-made structural barriers, whether from racist social norms, racist laws, or whatever.

Thus I do not think all cases of bargaining power are a problem. In the case of the two daters there is no reason to assume that B couldn't have an equally fulfilling relationship with someone else. If B can't, then it is a problem internal to B and not created externally to him by society.

azmyth writes:

It's not about binding constraints, it's about political signalling. A company who wants to fire someone for being a minority will find an excuse, but politicians want to send a signal about what they believe in and what they think is right. It would also be nearly impossible to enforce a law targeting individuals.

Mario Rizzo writes:

I find it odd that a person with economics training would say look at the data first and then make a theory. Inductivism? No better in ethics than in economics.

I do not deny, however, that many individuals acting in the world are "intuitionists." But that does not mean that the analyst must be one too.

Mario Rizzo. "The Problem of Moral Dirigisme: A New Argument Against Moralistic Legislation" NYU Journal of Law & Liberty 1.2 (2005): 790-844.
Available at:

Murali writes:

Situational relativism would also allow you to understand that for certain people (such as Murali and myself), hitting Robin in the situation you laid out might be a moral toss-up or even a moral action.

Egads, I thought I made it clear that I was not a relativist. And Mr Smith, while I kinda appreciate your support, I have even worse things to say about moral particularism (the kind of thing that you are endorsing) than I have of generalist intuition-ism.

As for Murali, Item #2, relativism being a consequence of intuitionism is simply a matter of reality; even if there IS some form of absolute truth the fact that it is not written into our consciousness means that any assertion thereof is just another relative guidepost

Sorry, but wrong. The mere fact that the value of pi is not written into our heads, or any other difficult mathematical proposition), or for that matter facts about how old the world is etc etc etc does not therefore mean that talking about these things are mere signposts. (Ok, as long as we're talking about empirical stuff there might in fact be a good argument as to the existence, or lack thereof of tables etc)

Ethics is like mathematics in the sense that both are a-priori (as opposed to a posteriori) synthetic (as opposed to analytic) enterprises. Given uncontroversial, self evident premises, the conclusions that can be worked out are logically necessary.

and it is this aspect of our reality that leads to intuitionism. The only real consequence of such a reality is that global harmony is impossible and change and conflict will be a permanent feature of the human species. If there is a supreme being that establishes absolute truth the logical conclusion is that this permanent conflict state is by design. If there is none then relativism ethics is as good and logical an explanation as to why we have what we do.

The question ethics explores is not why we do what we do, but What ought we to do, and why should we do it? The former question is one Robin Hanson answers all the time. The latter question is the one which we need to ask and answer if prescribing a libertarian society is to mean anything. Does the mere fact that most people like the welfare state justify the welfare state? If you're a libertarian, certainly not.

Descending into relativism and saying that everything is relative is the last move that libertarians should make. That intuitionism leads to relativism is therefore a serious charge which libertarians should not take.

Also, moral non-realists out there have a deeper problem in that they have argued too much. According to them, there is no fact of the matter as to what we morally ought to do. But, by the same argument, they would have denied all normativity, which puts them in a very strange position because they are trying to make at least an epistemic prescription.

To make it explicit, I'm some kind of Kantian. I see moral duties as rationally required. It would take a lot more space than I have here to fully defend my view, but you can catch the inklings of it on my site. (Just click my name)

Comments for this entry have been closed
Return to top