Tuesday, March 26, 2013

Thoughts on a Version of Consequentialist Ethics

Ethics is never far from the surface of discussions of economics, especially practical economic polices. Adam Smith, for example, was also an ethical theorist. We also have the spectacle of the Austrian Murray Rothbard who tried to defend and justify his anarcho-capitalist system mainly on natural-rights ethical grounds.

The limitations of early versions of hedonistic utilitarianism are well known, of course: the difficulty of reliably comparing interpersonal utility, and so on.

But hedonistic “act utilitarianism” is only one version of a myriad of ethical theories that can all be subsumed in the broader category of consequentialism.

And consequentialism as a broad group of theories can itself be categorised as one of the two species of teleological ethics, as follows:
Teleological (broad Consequentialist-type) Ethics:
(a) Virtue Ethics
(i) Eudaemonist theories
(ii) Plato’s eudaemonist ethics
(iii) Aristotle’s eudaemonist ethics
(iv) Stoic ethics (with natural law)
(v) non-eudaemonist virtue ethics
(vi) agent-based virtue ethics of Michael Slote
(vii) neo-Aristotelian ethical naturalism (P. Foot; R. Hursthouse)
(b) Consequentialist (minimal/moderate moral realism)
(i) Utilitarian-type theories
- Hedonism
- egoism (the 17th-century English philosopher Thomas Hobbes)
- universal/universalistic hedonism = utilitarianism (Bentham, Mill, and Sidgwick): act utilitarianism, rule utilitarianism

(ii) Consequentialist/Other teleological/utilitarian-type
- ideal-moral-code rule consequentialism (Hooker 2000; Richard Brandt)
- preference rule consequentialism (John Harsanyi)
- Motive consequentialism (Robert Adams)
- ideal consequentialism/utilitarianism (G.E. Moore; Hastings Rashdall)
- preference consequentialism/utilitarianism (R. M. Hare; Peter Singer)
- two-level preference consequentialism/utilitarianism (R. M. Hare)
- negative consequentialism (Popper, Christoph Fehige and Clark Wolf)
- evolutionary ethics, with ends of action, survival and growth (Herbert Spencer) power, as in despotism (Niccolò Machiavelli and Friedrich Nietzsche)
- satisfaction and adjustment, pragmatism (Ralph Barton Perry and John Dewey)
- freedom, existentialism (Jean-Paul Sartre)
- Buddhist ethics
- Mohism (Mo Tzu/Micius)
- Ayn Rand’s objectivism?
There is in fact a high degree of compatibility between a number of modern consequentialist theories.

While I don’t directly endorse any one particular theory above, I suspect that a workable version of consequentialism would require concern for a number of ends, and would draw on various of these theories.

As I think of ethics more and more, I am starting to see some merit in certain aspects of Virtue ethics and Rawl’s rights-based ethics.

Morality must be concerned not just with happiness, but clearly with fairness, justice and justifiable human rights as well.

But my position is in no sense an endorsement of any natural rights theory of ethics, a theory which, I think, remains nonsense. Rights are not natural; they are ethical constructs, requiring rational justification, and requiring human institutions and human beings to enforce them.

I suppose a serious ethical theory must pass three tests:
(1) it must not commit the “appeal to nature” fallacy;

(2) it must explain how it overcomes or is consistent with G. E. Moore’s “naturalistic fallacy,” and

(3) it must explain how it overcomes or is consistent with Hume’s “is–ought” problem (sometimes called Hume’s Law and Hume's Guillotine).
A complete answer to the question whether the “good” is really identifiable with natural properties (as naturalism contends), or is an indefinable, non-natural property (as G. E. Moore argued in Principia Ethica) I leave as an open question for further thought, although I do now lean towards the view that the “good” is at least explicable for humans in naturalistic terms.

According to Moore, all naturalist ethical theories commit the “naturalistic fallacy,” though there is some dispute on what exactly Moore meant by this. The “naturalistic fallacy” is not simply a crude fallacy called the “appeal to nature,” which is the assertion that what is natural is therefore inherently moral. Nor would some philosophers say that the “naturalistic fallacy” is the same as Hume’s “is–ought” problem.

Moore’s “naturalistic fallacy” consists in the (alleged) fallacy of identifying the property of goodness with some other natural property or thing.

For example, one could say:
(1) Pleasure is good.
This is a proposition, but it has two possible meanings, as follows:
(1) Pleasure is a thing that has the independent property of being good.
(2) Pleasure is identical with the concept “good.” That is, they are one and same thing.
In order to understand what Moore means, we must understand the somewhat confusing grammatical differences in the way the word “is” is used in English.

The verb “is” has 3 possible grammatical uses in English:
(1) to covey identity, e.g., “He is John” (They are one and same thing);
(2) predication (as the so-called copula or linking verb), e.g., “This desk is white” (that is, this desk has the property of being “white,” but obviously the “desk” per se is simply not the same thing as “white”);
(3) existence or being, e.g., “There is a high mountain two miles from here” (that is, there exists a mountain that is two miles from this location).
Just because we can use the word “good” with “is” in sense (2), Moore says, it does not follow that anything supposedly having the property goodness can be identified with the “good” in sense (1) (that is, identity). To assert that anything natural x is actually identical with the “good” in sense (1) is the “naturalistic fallacy.” Moore therefore contends that the “good” is an ineffable and non-natural property.

Whatever one thinks here, it is still possible in principle to be a moral consequentialist and to accept Moore’s non-naturalist views on the nature of the “good” (Levine 2002: 136), and then proceed to construct a consequentialist ethics. Moore himself did so, and was a moral consequentialist.

As to Hume’s “is–ought” problem, this seems rather more difficult. Now many philosophers contend that it is very easy to derive an “ought” from an “is” in a non-moral sense. For example, a working clock keeps time, so logically a clock ought to keep time in a sense that is non-ethical, that is, merely teleological. But that does not really help the moral theorist.

For example, one can observe the empirical (descriptive) fact that leaves naturally fall under the influence of gravity to the ground. Does it follow from this in a clear logical way that a leaf has a moral right to fall to the ground? No, it does not. My stopping the leaf falling to the ground does not appear to be immoral in a way logically derivable from the descriptive fact. Just because the nature of a leaf is to fall to the ground under the influence of gravity, it does not follow that the leaf has any moral right whatsoever to fall to the ground.

You can complain that morality should only involve sentient beings. But the difficulty is still apparent. If a person takes possession of unoccupied property never claimed by anyone else, this can also be expressed as a descriptive fact. But does it follow from the descriptive fact that such a person now has a moral right to absolute ownership and use of his homesteaded property? Is the moral statement “John ought to have the right of absolute ownership and use of his homesteaded property” a proposition that is true and justified (a moral fact) derivable from the previous mere descriptive fact? It does not logically follow.

If we say that “theft is the cause of serious problems in human society,” how is the moral injunction that follows, that “you ought not to steal,” a morally justified, true ethical fact? How do you derive that “ought” from the “is”?

If one wants to argue that all moral injunctions are really teleological imperatives to achieve some aim or end, then I suppose one has solved Hume’s “is–ought” problem, and endorsed some form of non-cognitivism, the view that there are in fact no “moral facts” as such and no “moral knowledge.” One version of non-cognitivism is the prescriptivism of R. M. Hare: what we think are ethical statements are just imperative statements. “You ought not to steal” becomes “do not steal!” but in a sense that is not a moral fact, but an imperative. Strictly speaking, the prescriptivist argues that any moral judgement is not a descriptive statement and is never entailed by any other descriptive fact.

That is, if one wants to admit that a moral “ought” statement does not really possess the property of truth or falsity, and in fact is never true or false, and can only be obeyed or disobeyed (since it only urges a course of action as an imperative does), then one has evaded Hume’s “is–ought” problem (though the prescriptivist, however, asserts that one can still reason about moral imperatives, not just about descriptive facts).

But many might complain that this a high price to pay for having solved (or perhaps evaded) Hume’s “is–ought” problem.

Other moral naturalists contend that one can derive the moral “ought” from “is,” and see the answer in the nature and logic of goal-directed behavior. I am not sure they have really solved the problem, and at best one should subject all claims to deriving a moral prescription from some descriptive fact to careful scrutiny.

Perhaps the solution really is to say that assumed moral injunctions are imperatives. Whatever justification is given consists in asserting that actions are urged to achieve certain ends to allow human society to function in a way that is conducive to order, known rules, predictability in life and social harmony. A direct moral injunction does not constitute a true moral statement at all, but is an imperative to urge action in a teleological sense to make society work.

That is the basic utilitarian function of what we call morality, and once one has proposed ends or aims, one can
(1) defend them by showing how other moral theories are flawed and unjustifiable;

(2) defend one’s proposed ends or aims in discourse to see how far one can obtain agreement from other human beings, and

(3) logically and empirically examine how to achieve the teleological ends: that is to say, what consequences follow about what should be done and should not be done to achieve the stated aims.
Returning to the main argument, crude versions of utilitarianism focusing merely on happiness/pleasure/utility as the sole good are inadequate, but not more sophisticated forms of consequentialism that argue that multiple moral aims (or consequences) are what must be aimed at.

I would propose a version of consequentialism that holds we should aim at more ends than just utility/happiness such as the following:
(1) preservation of human life where the people in question wish to continue to live and we can clearly ascertain this* (e.g., the man dying of disease in a highly privatised system of health care, the starving human being who is unemployed and who finds no private charity). Such an end has precedents in 20th century forms of consequentialism.

(2) minimising suffering, especially where this can be done with redistribution of resources that are relatively abundant (on this point, I wish to take Karl Popper’s negative consequentialism and make it part of a larger consequentialist system).

(3) aiming at the end of respecting certain individual rights where not doing so would harm the functioning of society significantly or would violate the sense of justice or fairness that an individual must feel to have respect for the law and confidence in others, in so far as this does not lead to gross violations of (1) or (2) (on this, I think notions of justice and fairness are legitimate ends, as in Rawls’s critiques of some crude utilitarian theories).

* Although this is quite compatible with the idea of voluntary euthanasia when expressed clearly by those of sound mind suffering from painful terminal disease, I will leave this question open.
These additions to consequentialist ends aimed at overcome most of the traditional objections to utilitarianism.

I suspect the alleged incompatibility of consequentialism with other theories is exaggerated.

This is an ethics that strongly supports the right to be free from unjustifiable coercion, unjustifiable suffering, arbitrary killing, and theft (non-justifiable appropriation of external property), but at the same time the limited right to control over external property.

But also one’s right not to die simply because one cannot find a job in a laissez faire society or successfully beg for private charity; and the right not to live in poverty or without adequate health care, and so on.

Here is a puzzle for grammarians and philosophers.

What is the grammatical status of this sentence?:
“When playing a normal* game of chess, one ought to play to win the game.”

* Normal in the sense of not playing to deliberately lose or draw.
I suppose the obvious answer is that it is a sentence in the subjunctive mood.

Some might complain that the grammatical form of the sentence is misleading. Perhaps it is really just an imperative: “Play a normal game of chess to win!”

But can it be understood as a true proposition? That is, it is simply a true statement that a person playing a normal game of chess plays to win in a teleological sense. This is the purpose of chess.

Can a subjunctive even be true or false?

In addition, it is clearly not an ethical proposition. If you do not play a normal game of chess to win, then you are not immoral or evil, but simply acting contrary to the convention of behaviour required by the game.


Levine, A. 2002. Engaging Political Philosophy: From Hobbes to Rawls. Blackwell, Malden, Mass.


  1. Wow! I'm bookmarking this post ASAP for future reference. You have outdone yourself, Lord Keynes. You should write more of these philosophy related posts, as you come across as someone who has done extensive academic work on the subject and possibly has some type of degree in the field.

    What is your opinion on logical positivism?

  2. regarding the addendum,

    The key to the sentence is the qualifier "normal." It is fine to classify the sentence as an imperative. Imperatives don't signify ethical good/evil; one deals with imperatives all the time such as "don't enter" or "stay off the grass." Breaking these imperatives doesn't constitute an evil act, because the imperative itself is a ethically subjective proposition, it is the command of one individual. An imperative that is not merely advanced by an individual but accepted by the majority of an entire community becomes a norm. If one plays a game to chess to lose, one may face negative consequences in terms of losing one's reputation. If one plays against any of the rules of society, one may face a number of negative consequences, including id-imposed guilt. But people in one society may call an act evil and feel guilt when committing it, whereas people in another society may call the same act acceptable and not feel guilt when committing it. Despite the commonplace usage, we cannot conflate norms with ethics.

    Whether there is an ethics independent of norms is a vital question...but a theory of ethics should not be confused with a theory of norms.

    Maimonides of the Andalusian era wrote that norms/morals evolve throughout history. The anthropological record makes this fact fairly undeniable. Specifically, he thought there was a true ethics, that is an objective value system of good/evil, that is determined by the will of God. However, humanity's ability to come to terms with this true ethics has developed throughout history. Each successive revelation on moral behavior passed on from God to man has brought man closer to God, and closer to this true ethics.

    One can accept this theory without sharing the theology; suppose that the evolution of morals approaches an objective ethical ideal as history develops, through non-divine processes.

  3. Interesting to see this here - I have been deeply concerned with consequentialist ethics in the social sciences more generally for some time (e.g., Determinism and the Antiquated Deontology of the Social Sciences, Clint Ballinger )
    Haven't had time to read this yet but look forward to it.

  4. Excellent post, and an excellent blog.

    My response: Pragmatism, a naturalistic, consequentialist, and loosely utilitarian meta-ethical philosophy.

  5. All that talk about "God wants something because it is good" versus "it is good because God wants it" and not one mention of Euthyphro...