Eat Babies for Galactic Peace

Eliezer at Overcoming Bias has outdone himself, producing a fascinating serial novella that explores the limitations of moral thought.

Maybe it is my own nihilism talking, but to me it serves as a fine illustration of the inherent arbitrariness hidden in the foundation of all moral systems. Once you can imagine creatures with a moral system radically different from your own, it's hard not to retreat to an ad hoc, pragmatic view of morality.

Through part II, my sympathies are with the baby eaters. Is that unusual?

Edit:

After pondering this story for awhile, I am less impressed with it. I may have missed the point that the author intended for me to get the first time around. It is still an entertaining yarn that will send a charge through your brain-crystals, but it could end up as a simple parable on the virtue of Yudkowsky's preferred morality.

Consider the evidence: Eliezer is known to consider suffering and death to be a great evil. He thinks that people who reject even a small chance of cheating death, such as by use of cryogenics, are attempting to rationalize a flawed world-view. To Yudkowsky, people who think death is a "natural" part of human existence are the babyeaters.

However, there is some art in the story, since Yudkowsky gives us his morality in two perspectives. He next introduces the alien race of the super happy ultra fun fun people (or something like that) who have done away with pain and live in an eternal state of supreme pleasure. They view suffering as a great evil, and view humanities' choice not to use technology to do away with it as something akin to the babyeaters' choice not to use technology to do away with the need for baby eating. At the conclusion of part 4, they are threatening to forcibly take over and remake the human race to eliminate death and pain.

I think Yudkowsky wants us to be convinced by the arguments of the super happy fun fun people and commit ourselves to do away with whatever pain we develop the technological capability to alleviate. I missed this point too the first time I read it. Instead, I thought the super happy fun fun people were sort of grotesque.

I am at peace with the fact that I have a human-centered worldview. My entire personality, identity, and consciousness is tied up with a particular piece of wetware with its own innate moral biases. Yudkowsky won't convince me to give it up that easily.

If I were a babyeater, I'd eat babies.

Alternatively, Eliezer might put some warts on the happy fun people and use them to illustrate some of his ideas on failed utopias. After all, most stories are a function of their authors.

Share this

Excellent story

Your utility function may be arbitrary, but it's still your utility function - all of your values. I'm pretty sure you don't want babies to be eaten, all else equal. This is all to say that you shouldn't do whatever you're immediately inclined to do. You have a utility function, so optimize it. And if you're human, you probably value things other than yourself, like babies. I'm not accusing you of missing any of this, but most self-proclaimed nihilists don't seem to realize it, so it's worth keeping in mind anytime anyone brings nihilism up. Personally, I avoid using the term because it's connotations typically prevent my actual meaning from being conveyed.

Alien babies do throw an interesting wrench in all of this. I'm not so sure that I care all that much about alien babies - not enough to go on a galactic crusade, but Eliezer seems to think that we (humans) would. In fairness, I may not be fully internalizing the different context in the story.

He's also very optimistic about the psychological unity of mankind keeping our utility functions similar enough that morality basically is universal among humans. I'm uncertain about this proposition, but not long ago I would have agreed with Eliezer. Now it seems that sociopaths and weird cultures seem problematic for his view.

btw, Part 4 is up.

If you are a babyeater then you eat babies

"I'm pretty sure you don't want babies to be eaten, all else equal."

He said, "If I were a babyeater, I'd eat babies.", therefore all else is not equal.

To which I would have to respond "If any of us were baby eaters then we'd eat babies".

I didn't really find the baby eaters all that disturbing. After all they are eating babyeaters. I didn't have much empathy for them. Nor would I feel compelled to help the baby babyeaters.

Nor do I think humans in general feel this compulsion to help expessed in the story, or to impose their values. There are still starving Africans despite all those TV commercials so people must be flipping the channel. The US hasn't invaded the entire Muslim world to impose individual rights. Hell, the countries we have invaded Iraq and Afghanistan we didn't even set up constitutions that allowed for religious freedom, we set up Islamic Republics. Stupid Bush.

Even in the case of expansive violent ideologies like Islam, many individual members do not feel any compulsion to impose upon others. Even though when you read the text of the Qur'an it's egging on Muslims to participate in military jihad.

The main danger of the babyeaters was their apparent intolerance of non-babyeaters and need to eliminate those who disagree with the consensus.

Ad hoc morality

Once you can imagine creatures with a moral system radically different from your own, it's hard not to retreat to an ad hoc, pragmatic view of morality.

compare:

Once you can imagine creatures with a blood pressure radically different from your own, it's hard not to retreat to an ad hoc, pragmatic view of blood pressure.

Whatever you want to call it, "pragmatic" or "ad hoc" or whatever, there is a healthy human range for blood pressure, it's not arbitrary, it's not a matter of whim, it's not a matter of personal preference. Whether you live or die might, of course, be a matter of personal preference, but the list of things that will kill you is not. This remains true even if you can imagine creatures that have a very different list of things that can kill them.

And similarly for human morality.

Constant, I used to make the

Constant,

I used to make the same arguments. To Jacob, in particular. I've come over to his side.

But I am so glad to have someone where I used to be who defends himself with such panache.

And yet, you are so

And yet, you are so libertarian that you would let the world blow up in the "murder a child or a madman blows up the world" moral dilemma. Some nihilist.

I have a few reasons for being a moral nihilist. First of all, let's define "positive morality" to be the moral intuitions that people naturally have. This could be measured with psychological experiments, surveys, and statistics.

Positive morality is a challenge to philosophical "abstract morality". Positive morality is lumpy, it is contradictory, it is ad hoc. On the other hand, abstract moralities are smooth. No overriding set of consistent principles can likely be constructed to fit our species' positive morality without glaring misses. The topology is wrong.

I'll flesh this out in a post at some point, when my work load hits a lull again.

Positive and abstract medicine

Positive morality is a challenge to philosophical "abstract morality". Positive morality is lumpy, it is contradictory, it is ad hoc. On the other hand, abstract moralities are smooth. No overriding set of consistent principles can likely be constructed to fit our species' positive morality without glaring misses. The topology is wrong.

Let's substitute "medicine" for "morality" as follows:

Positive medicine is a challenge to philosophical "abstract medicine" (a hypothesized approach to medicine which attempts to deduce it from a small set of medical axioms). Positive medicine is lumpy, it is contradictory, it is ad hoc. On the other hand, abstract medicines are smooth. No overriding set of consistent principles of medicine can likely be constructed to fit our species' positive medicine without glaring misses. The topology is wrong.

The human being is a sufficiently complex organism that one cannot derive medicine from five or ten or a thousand "first principles". Rather, medicine is a vast collection of facts which cannot be deduced from a small set of axioms discoverable a priori from the comfort of one's armchair but must be learned through long experience. Positive morality as you define it is learned empirically from "psychological experiments, surveys, and statistics". Similarly, positive medicine is learned empirically from practice and from controlled studies done on many patients.

Are you a medical nihilist? I surely hope not.

Why should the failure of a certain, admittedly popular, approach to the study of morality lead you to be a moral nihilist? Does the failure of alchemy lead you to disbelieve in the existence of metals? Does the failure of lamarckism lead you to disbelieve in the reality of evolution?

More nihilist than thou

And yet, you are so libertarian that you would let the world blow up in the "murder a child or a madman blows up the world" moral dilemma. Some nihilist.

I don't see the tension. What does the objectivity (or lack thereof) of morality have to do with my particular brand of morality? It seems you are confounding two concepts: principle and objectivity. Just because I rely on a few principles to derive my moral views doesn't imply that those moral views are objective. I also consistently like sushi -- you could say my gustatory likes are based on the sushi principle -- but surely this doesn't imply sushi is objectively better than other food.

On a superficial level, certainly allowing the world to be destroyed for something as silly as a trespass seems the more nihilistic stance.

Positive morality is a challenge to philosophical "abstract morality". Positive morality is lumpy, it is contradictory, it is ad hoc. On the other hand, abstract moralities are smooth. No overriding set of consistent principles can likely be constructed to fit our species' positive morality without glaring misses. The topology is wrong.

That's one interpretation. Another is that, so far as people have moralities that fail to be smooth, those people are in error. The failure of the entirety of human morality -- positive morality, as you've designated it -- to be smooth no more implies that morality doesn't exist than the fact that the entirety of human views on, e.g., physics fails to be smooth -- positive physicality -- implies that the domain of physics doesn't exist.

For example. Let's say it's a fact that the world is round. Not everyone agrees on this: the Flat Earth Society is still about. Given that disagreement, this lumpiness, I very much doubt you could construct a consistent physical theory that allows for both views. But of course, this is silly: we don't need to take account of the flat-earthers' views, because they're wrong.

By the same token, let's say it's a fact that utilitarianism is right. Not everyone agrees on this. Given the disagreement, the lumpiness, I very much doubt you can construct a moral theory that allows for both views. Does this imply morality doesn't exist? Or mightn't the utiltarian simply argue the non-utilitarians are, as the flat-earthers, simply wrong?

Nothing beats that metaphor

Nothing beats that metaphor to explain these ethical paradoxes.

I hadn't seen that. It's

I hadn't seen that. It's pretty good. I suppose Jacob's response would be denying that there are any easy cases, i.e. cases far away from the fractal fence.

I saw it

I was not impressed. Quoted stuff doesn't make any sense. Invisible fences? What is that supposed to mean? Do invisible fences make invisible neighbors?

In order to fight

In order to fight procrastination, here's an imperfect and never perfected draft that I did more than a year ago when I read this entry, it's the ethical yin & yang

I'm not happy with it, it needs to be improved but I never got around to do it

We used to have a collection

We used to have a collection of three dozen emoticons available at the old site, but everyone thought they were corny. I've never wanted to use one until now.

*applauds*

Arthur, this is a family-friendly board

Please refrain from posting pics of hairy vulvas.

What made you pick "that"

What made you pick "that" picture? heh

I've been dorking around with "Math Art" for awhile and I've seen that design a few times while rummaging about on the information super-highway.

Uh, yeah

I've seen that design a few times

If you mean the Mandelbrot set, well, yeah, I suppose you would, but the way you say it makes it sound like it's not as famous as it was ten or twenty years ago.

I suppose I can get serious

I suppose I can get serious here for a moment.

Yes, the Mandelbrot set. It's just a strange coincidence that it appeared here in a comment, when I've been messing around with a few "mathy" things these past few days (and weeks) that have used that image, and many others.

Fractal Moral Boundaries

Because it's a picture of a fractal, and some, including me, have used fractals and an analogy to moral boundaries.

A woman who shaves her pubis

A woman who shaves her pubis in the shape of the Mandelbrot set is a a keeper!

Strange Attractor

That would certainly be a strange attractor.

Pity they call it the

Pity they call it the "Mandelbrot Set". Not only is the label in error it can't compare to a "Vulva Sundae".

Defeating illogic is not enough

I used to make the same arguments.

It's not enough to defeat the illogic of moral anti-realists. I've seen you arguing against the illogic of moral anti-realists but apart from this I haven't seen you argue for the reality of morality. If you have no compelling positive reasons for believing in the reality of morality, it is not surprising that you would eventually stop.

Moral arguments - arguments which derive one moral claim from another moral claim - can't establish a system of moral claims, they can at best demonstrate internal consistency. So they are no defense against nihilism. Almost everything that I see from moral realists is moral arguments. Moreover, if I've understood their responses, moral realists have asked me to present moral arguments for the correctness of my own theories of morality. Of course any such moral argument would be equally vulnerable to nihilism.

There is an objectively real phenomenon, which I have described before on various occasions, which correlates well with what we typically think of as morality. I consider this reason enough to identify it with morality. This is no different from the empirical reason for accepting any scientific explanation of anything at all. It is not, however, a moral argument, and thus unlikely to satisfy either moral realists or nihilists. Too bad for the lot of them.

Moral intuitions are objectively real, but they are not the heart of morality. They are akin to other intuitions, such as physical intuitions. Intuitions are windows on reality, but are not themselves the reality that they intuit. Physical reality teaches us physical intuitions, which allow us better to navigate physical reality, and similarly, moral reality teaches us moral intuitions, which allow us better to navigate moral reality. Moral intuitions can be "measured with psychological experiments, surveys, and statistics," but in the end this is like measuring physical intuitions. Physical intuitions can also be measured with psychological experiments, surveys, and statistics. But this isn't the only or necessarily the best way to learn about physical reality, and similarly, doing surveys of moral intuitions isn't necessarily the best way to learn about moral reality, though it may be better than a priori moral philosophy.

We learn more about physical reality, and get better intuitions of physical reality, when we conduct experiments. The moral analog of experiments in physics is conflict resolution. Conflicts arise and are resolved, and over time we develop rules which allow us to more quickly and painlessly resolve conflicts and, best of all, avoid them altogether by following the rules in our everyday behavior. These rules are deeply impressed in our minds as moral intuitions. Some of these rules are arbitrary - they can vary from culture to culture - but certain key rules are not arbitrary and so are invariant across cultures. They are invariant because they occupy a stable equilibrium in the space of possible rules. Any talk of an equilibrium implies an underlying dynamic and this is no different. The social dynamic that underlies morality is akin to the economic dynamic that underlies money. Key moral rules are stable and universal in something like the way that money is stable and universal - that is, all economies eventually develop a medium of exchange (money), and once they develop it, they keep it. They don't to back to barter. And similarly for key rules of morality.

I might be classified as a moral nihilist by those who believe that there must be something more to morality than the sort of thing I have been describing - those who see what I have been describing as mere physical reality and not moral reality. These are the people who ask: "granted, societies have moral customs, but why am I morally obligated to follow these customs?" This I take as asking for a moral argument - I talked about this above. But I don't consider myself to be a moral nihilist. Among other difference from nihilists (as I've encountered them), I think that moral intuitions are fallible windows on moral reality, whereas nihilists think moral intuitions (or even personal preference) are the heart of morality.

In that case, our views

In that case, our views don't seem to diverge at all. As I understand the landscape, you're an ethical subjectivist. You believe moral statements can be true or false, but that those truths and falsehoods are judged by reference to human preferences. Whereas a moral nihilist doesn't believe that moral statements can be true or false at all. And a moral realist believes moral statements can be true or false without reference to human preferences.

It's not enough to defeat the illogic of moral anti-realists. I've seen you arguing against the illogic of moral anti-realists but apart from this I haven't seen you argue for the reality of morality. If you have no compelling positive reasons for believing in the reality of morality, it is not surprising that you would eventually stop.

Perhaps. But I, by my lights, don't have any compelling reason for believing in the reality of anything. I rest my belief of what exists and what is merely socially constructed on nothing more than a gut level intuition. For a while, that gut intuition told me morality was objective, but the feeling ebbed. The gut intuition that the material world exists has yet to ebb.

Murdering you isn't

Murdering you isn't unhealthy for me under some counterfactual, yet I still shouldn't do it. This seems to be purely a matter of personal preference - I don't like it when people are murdered, and neither do you... I hope...

You are being too literal

Murdering you isn't unhealthy for me

I was addressing a specific argument for moral nihilism by substituting terms while retaining the internal logic of the argument. A logically sound argument will survive a logic-preserving substitution of terms and this one did not. It is not necessary that blood pressure and murder have any particular features in common, be it "being unhealthy" or "being measurable in millimeters of mercury."

To recap: the conceivability of another species with a different XXX from the human XXX does not mean that XXX is arbitrary, subjective, a matter of personal preference, or even nonexistent. There is an unspoken assumption in such an argument that XXX, if it existed objectively, would necessarily be universal across all possible species - and thus, if there is a conceivable species where XXX differs, then XXX is a matter of personal preference etc. In other words, the assumption is that the only alternatives are cross-species universality, or utter subjectivity and objective nonexistence. I dispute the assumption. Some moral realists do, apparently, claim that morality is necessarily universal across all possible species, but I don't subscribe to that view. My view is that morality is biological, and so it may differ among different social species, since those species will have different biology.

This seems to be purely a matter of personal preference

I argue that there is more going on than personal preference in another comment. The argument in a nutshell: the intuition that murder is wrong derives from the social custom that murder is wrong. (The social custom is, being a custom, necessarily shared, which rules out the personal preference view of morality.) The social custom that murder is wrong derives from the resolutions of countless conflicts in which a person was killed. These resolutions were not themselves arbitrary but tended as a result of the underlying dynamics to be resolved a certain way. (A poor resolution might, for example, spark further conflict, which might discredit the judge and cause his opinions to be forgotten - subjecting the evolution of law to a kind of selective pressure.) The dynamic that governs human conflict is an outcome of human biology. Aliens have a different biology and so may have a different morality.

With preferences, there is a

With preferences, there is a simple way to derive normativity. With social custom, where do you get the normativity? Why should any individual care about social custom outside of the consequences of ignoring it?

What exactly constitutes a "poor resolution" of a conflict? This sounds like a normative evaluation, yet that would be circular reasoning.

Just to make sure we aren't arguing past each other, I'm a moral realist in the same sense Eliezer Yudkowsky is - morality is based on our preferences, but determining these preferences isn't a simple matter, and nether is actually making decisions based on them rather than immediate inclinations. I think morality is more or less universal among humans with some possible exceptions, but this is only because we tend to have similar preferences sets - we don't like it when people die or are injured, etc. Note that preference =/= pyschological gratification.

Social custom and norms

With social custom, where do you get the normativity?

You seem to be asking for a moral argument. I explained what I consider to be a weakness of all moral arguments with respect to nihilism here. I offer an alternative to the circularity and defenselessness against nihilism of moral arguments.

However, I can take your question a different way. A social custom is a social norm. The normativity is already there, in a certain sense of "normativity". Feel free to elaborate a different sense of "normativity". I will feel free to critique it.

If you want to step outside a particular social custom and judge it on some external basis, you can do that by asking whether the custom is inherently stable and universal within our species. I wrote:

Some of these rules are arbitrary - they can vary from culture to culture - but certain key rules are not arbitrary and so are invariant across cultures. They are invariant because they occupy a stable equilibrium in the space of possible rules. Any talk of an equilibrium implies an underlying dynamic and this is no different. The social dynamic that underlies morality is akin to the economic dynamic that underlies money. Key moral rules are stable and universal in something like the way that money is stable and universal - that is, all economies eventually develop a medium of exchange (money), and once they develop it, they keep it. They don't to back to barter. And similarly for key rules of morality.

..

Why should any individual care about social custom outside of the consequences of ignoring it?

This question can also be interpreted different ways. You suggest one answer (the consequences of ignoring it), and this suggests that you mean your question as the following request:

"Please offer a self-interested individual reason to follow custom."

An alternative interpretation is:

"Please explain why it is moral for an individual to follow custom."

Two very different questions. The first question, you already answered. The second question is a request for a moral argument, which I decline to provide, in part because moral arguments are pointless as arguments for moral realism and against nihilism (as I explain in the linked comment), and in part because I am not offering a theory of morality from within morality, and therefore it misconstrues what I am doing to attempt to answer that question. These are two sides of the same reason. If you argue from within morality, you're still completely vulnerable to moral nihilism.

To put it another way, I am not trying to justify morality. That would be to justify justification. It's circular. And in particular it doesn't gain you any ground against nihilism.

In my linked comment, I mention

These are the people who ask: "granted, societies have moral customs, but why am I morally obligated to follow these customs?"

which looks like a reasonably close anticipation of what you just now asked, and there I discuss what I think of such questions.

What exactly constitutes a "poor resolution" of a conflict? This sounds like a normative evaluation, yet that would be circular reasoning.

It is one which leads to the extinction of influence of the resolution as a precedent for future resolutions. Our moral intuitions are an end result of a process in which certain judgments have gone extinct and certain other judgments have survived, so as a result of our heritage, we now intuit the poorness of the poor resolutions. Our moral intuitions are thus a product of evolution (both biological - which I have not gone into here - and social - which I have been discussing). The intuitions are a reflection, a kind of memory, of the moral lessons that we have learned - lessons that were drummed into us by this long evolution. Moral reality came first. Our moral intuitions came next. Similarly: physical reality came first. Our physical intuitions came next.

I'm a moral realist in the same sense Eliezer Yudkowsky is - morality is based on our preferences

I am not a moral realist in that sense - I do not believe that morality is based on our preferences - at least, not in the way that I think you mean, based on your other comments (you write, "This seems to be purely a matter of personal preference - I don't like it when people are murdered" - you are here simply identifying morality with preferences, and thus are talking about a personal morality). I believe that our preferences are based on our preferences. Eliezer can call that "morality" if he wants, but if he says he's talking about preferences then I will take him at his word. There's no shame in talking about preferences, no need to dress it up as morality.

determining these preferences isn't a simple matter, and nether is actually making decisions based on them rather than immediate inclinations.

I agree. Preferences are not that simple.

Our morals are not shaped by

Our morals are not shaped by our preferences. Morals are an outcome of our human nature (and other mingling factors), which is an outcome of evolution. This was stated earlier, but not in so few words.