« Useful Statistical Biases | Main | The Error of Crowds »

April 01, 2007


In my view, a powerful argument for moral coherency is a form of Occam's Razor. There are only so many moral principles, so many deep emotions; different judgments that arise from these same sources should be similar to each other. On the other hand, there's an unlimited number of situations in which our own selfish impulses may move us to create exceptions to our own rules - an unlimited temptation to special cases.

In the real world, incoherency is sometimes a force for good, but only where something else is already very wrong. Someone has a declared moral principle, or a declared allegiance, which would seem to move them to dreadful deeds, but their compassion opposes this and causes them to flinch away from the consequences, so they make exceptions to keep themselves from actually doing the dreadful deed. Jewish rabbis carefully defined away most of the Bible's many death penalties - for example, talking back to your parents is to be punished by being stoned to death before the gates of the city - by surrounding them with so many special legal rules that there is no realistic way of imposing them.

If you find yourself tempted to be "morally incoherent" because the alternative is hurting someone, I would say: question morality.

Do moral systems have to make sense?

This is the big question of the age as we come to accept a mechanistic, deterministic explanation of ourselves and the old timey notion of free-will fades. It stikes me though, that humans, we big brain creatures of the 21st century, have made ourselves too much of a slave to logic, too ready to cast off everything that doesn't please our brain's cold logic.

I accept that my desire for fat-rich, edible objects arose for evolutionary reasons, but the mere fact my desire for chicago style pizza can be explained from an evolutionary or scientific basis doesn't affect the pleasurable sensations from eating the stuff, though I may understand exactly why that pleasure should exist in me. Similarly, I accept that my moral emotional states exist in me because they served an evolutionary purpose, but the mere fact logic has pulled away the curtain to reveal no one has free will (and thus logically my anger at my trespassers makes no sense), the moral emotion against them rages on in me no less fiery hot.

The real problem seems to be a division in our emotional/moral intuitions, between the angry sensation at someone who purposefully does harm and our, also, emotional/moral intuition that we shouldn't be angry against someone who does harm but had no choice to do it. (What's the old saying? Even a dog can tell the difference between when its been stumbled on and when its been kicked? The problem is our big brains that say no free will now tell us everything is a stumble.) But because we've evolved to feel anger at harm-doers that 'Appear' to be acting purposefully, the mere fact we 'know' otherwise doesn't cause that sensation to go away.

Really, the question becomes should society be organized to make logical sense or to make its constituents happy, regardless whether that pleasure or displeaure "makes sense." I'm not so sure the latter isn't best.

There is a small set of very widely accepted moral rules
that are held, with caveats and peculiarities, in pretty
much all societies. This small sense can be seen to be
pretty sensible if one is thinking about maintaining a
reasonably functioning social order. They are basically
about five of the last Ten Commandments: Thou shalt not kill,
thou shalt not steal, thou shalt not bear false witness, thou
shalt not commit adultery, and honor thy father and mother.
The first of these three are actually in most legal codes,
and the latter two, although widely ignored, certainly underpin
maintaining family structures.

As for the rest of the Ten Commandments, they are mostly highly
sectarian and not at all universal. Anybody for refighting the
Iconoclastic War?


I thought long and hard about it, and I just can't make any sense of "Murder is quite likely to be wrong." What could this possibly mean?

- That it is quite likely that murder has overall negative utility? (All murder, or just particular instances?)
- That it is quite likely that Platonic values exist in metaphysical space? (Seems quite unlikely...)
(I'm sure there are other possibilities.)

What makes the statement to which you compare this less appealing statement different is that it basically says that there's a rule that has exceptions. Some might object to your phrasing and claim that you shouldn't have used the term murder - that any cases covered by the "exception clause" should not properly be called murder (but rather "justified killing" or something similar). What counts as murder is always, it seems, open to question; this is what some moral theorists mean by calling moral concepts "essentially contestable": the possibility of critique and revision is always open. But how do we decide what counts as murder? This doesn't seem to me an empirical question at all - i.e. one that admits of probabilistic assessment. To decide what counts as murder is to make a decision about the grammar of this concept in our moral discourse (or more plainly, to decide the term's meaning). But how is the meaning of a moral concept an empirical question? (This is like saying that a yard probably equals thirty-six inches: either it does or it doesn't, and whether it does depends upon what we mean by a "yard." Perhaps you mean to be asking how we can know which moral standard of measurement is going to turn out to be the best?)

Matthew, see my defense of impossible worlds. They allow probabilities over moral possibilities, even if the topic is not "empirical."

Robin, I am probably being dense here, but I don't see how this is going to help us adjudicate the question, "What is murder, really?" (Suppose we agree on the definition: Murder is the unjustified killing of an innocent person. Well, what's unjustified? What's innocent?) I see, roughly, how possible worlds might allow us to visualize the arrangement of the different possible answers to these questions, but how we arrange these possible worlds raises yet another normative question: for we aren't trying to figure out which world is closest to (or is) the actual world - we want to know which one is the ideal moral world. If we arrange the possibilities with reference to the actual world, then what? Do we place worlds that are closest to our way of answering these questions next to our world: what does that show about the likelihood of truth (and whose way of answering is at work)? (Again, what is it to say that a particular definition of murder is true???)

You might say that we have to stipulate a definition of murder, and then attach different value-claims to the concept in different worlds: but that assumes that you can strip the term murder of all its in-built evaluative content. (Some would claim that the concept itself implies a negative value judgment: that murder is not an evaluatively neutral term.)

If free will is an illusion that came about through evolution then what adaptive function does it serve? Does the illusion itself influence behavior? It seems to me that the illusion of free will can have no influence if there is not something that it is free to choose or change. Or is the illusion of free will impotent and just an evolutionary side affect of some other development?

It seems to me that these questions need to be answered before we change our moral systems. I personally believe in free will just because I have more data (personal experience) in regard to free will than I have data in regards to evolution and it seems to me that if my data is bad in regards to free will then the data for evolution will also be tainted. I also do not deny that evolution has some basis in truth.

Hmm, I do not find this discussion making much sense. Part of the problem is that "making sense" has not been defined. Are we talking about some kind of logically coherent axiomatic system with well and universally applicable defined terms or concepts? This may "make sense" but strikes me as being pretty stupid.

There are other criteria, e.g. Kant's categorical imperative, which can be defended one evolutionary social stability grounds. Thus, my defense of the five basic principles from the Ten Commandments, although they are subject to all kinds of definitional variations across cultures (e.g. "what is murder?" and "if someone can have more than one spouse, when is there adultery?'). But, it is not hard to see that societies may have problems preserving and reproducing themselves if they allow murder, theft, official lying, and the breakdown of families, however defined. Such a criterion makes sense, and may indeed explain why those particular strictures have evolved to be so widely regarded as valid "moral" injunctions.

Barkley, for the record, I would never use the word "axiomatic." My point was that the statement, "Murder is quite likely to be wrong," would be a bizarre claim to make within a moral system. (Would such a statement mean that I can use some fancy-pants statistical formulas to estimate how likely it is that shooting my mother in the face for taking away my pudding would be wrong?)

At any rate, what you say about the 5 commandments seems sane. I suppose all the debate is about what exactly the exceptions are...(Or were we to take, "Thou shalt not kill" literally? Oops.)

The point is that no-one ever claims "murder is quite likely to be wrong", even if someone who did claim that would behave identically to someone who thought "murder is wrong" but added practical caveats. Why can we conceive "murder is wrong" or even "murder isn't wrong" but not the in between position? And we do this only with moral values, not with other types of statements.

My post was going through an example of when I should change my moral values based on incomplete information, and exploring the reasons for my reluctance to do so. More eloquently, Eliezer expresses a practical reason to hew to absolute statements. But all these reasons feel like rationalizations - surely, some philosophers, not exactly the most practical of people, must have come up with partial moral truths at some point, if only practicalities stood in the way.

That they nearly always go for the "truth absolute - reality contingent" approach tells me that there is something very strong pushing our morality this way. I've met quite a few fundamentalist Christians who told me that they found intolerable the idea of there being different correct behaviours in the same circumstances. We seem to have the weaker version of finding it intolerable that there may be different correct moralities for ourselves.

Are we wired to need axioms in morality?

Why can we conceive "murder is wrong" or even "murder isn't wrong" but not the in between position?

For many people, the answer to this question is that murder is a morally loaded term - a "thick" concept.

For my part, I am growing increasingly skeptical about "general moral principles" as (your) philosophers traditionally conceive of them. This is probably because unlike those Fundamentalists, I am not bothered by the possibility of situations in which what the "right thing to do" is depends upon the person in the situation.

(I have a post on roughly this issue queued; waiting for Robin to push the button.)

I don't know if we're wired to need axioms, but we don't like feeling like what we're doing is purely arbitrary. Axioms would seem to solve that problem (if only they could actually tell us what to do in concrete, real-life situations where there is a REAL conflict between moral possibilities!).

And we get into definitional problems, that also get far afield. Thus, in the original Hebrew, the Commandment forbids "murder," not "killing," but in the usual English (and Latin and Greek) translations it has ended up as "killing," which may lead to pacificism, but that is not implied in the original. Thus, most legal codes do not label killing in self-defense to be "murder" in any degree, and legal codes take seriously matters of intent, hence the key distinction between first and second degree murder in the common law traditions.

I also do not deny that evolution has some basis in truth.
That's awfully generous of you.

But all these reasons feel like rationalizations - surely, some philosophers, not exactly the most practical of people, must have come up with partial moral truths at some point, if only practicalities stood in the way.
What constraints are philosophers under that will cause them to act in a certain way? People who study other subjects tend to reach something approaching truth because being wrong is more costly. In philosophy you never know if you're wrong and there are no costs.

In philosophy you never know if you're wrong and there are no costs.
But that's precisely why it's strange they never explore these particular alternatives. And they're under constraints to be known, and a beginning like "murder is quite likely to be wrong" would attract attention.

It can certainly be built up into a coherent system, too, so that can't be the problem. So what is?

It all a load of Rubbish!! what makes human's think they can find answers to any questions at all is beyond thought.We are but frogs in a pond and haven't seen the ocean have we? All the talk of morality is not something we can summarise as this chapter should be left to the God Almighty. he will come and kick everyones arse one day and put justice to real action. Morality will be understood. For know lets just stick to the basic Bible as a guide to what our behaviour should be.

The comments to this entry are closed.

Less Wrong (sister site)

May 2009

Sun Mon Tue Wed Thu Fri Sat
          1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30