Friday, March 30, 2007

The Democratic Cannibals

In a hospital emergency room, five critically ill patients desperately need organ transplants. A healthy man walks in. Should the doctors remove his organs to save the sick five? Most people will respond in milliseconds with a resounding "No way". Now imagine an out-of-control train about to run down five workers standing on the track. There's a fork ahead, and throwing a switch could divert the train to another line on which there is only one worker. It's the same question - should we sacrifice the one to spare the other five? - yet most of us would say "yes" just as quickly. How do we make these lightning moral judgements?

If you can spot the fundamental flaw in that paragraph, chances are you're already a democratic anarchist like myself. Otherwise you're probably asking how can you be a democrat and an anarchist.

Marc Hauser whose interview with New Scientist that piece introduced (cached), went on from that flawed beginning to expand on his entirely spurious and unnecessary conviction that we all have some kind of moral sense biologically built into us like the homing instinct of a pigeon:

What I call the moral faculty has that same aspect: we unconsciously deliver a response to right and wrong - and I use "unconsciously" in the same sense the linguist Noam Chomsky does in his work about language. In other words, there's something about the biology of our brains that has orchestrated a set of tools to build a moral system.

Frankly it was so fatuous and wrongheaded that my initial response was to ignore it. Nobody else would take it seriously. Why should I? It was a letter (see below) about the interview, a couple of weeks later, which changed my mind and made me realise that most people cannot see the flaw and that exposing it might help reconcile the division between the libertarians of right and left. As the libertarians tend to represent the intelligent component of the political spectrum, this exposure might have far reaching consequences.

[Edit August 2007: Rereading this months later I realise it comes across as though I'm deeply hostile to the concept of an inbuilt moral sense. This is not what I meant to imply. I don't buy into his idea, but I'm not hostile to the possibility of a genetic algorithm. I'm agnostic. My hostility is aimed only at the argument that you could reach such a conclusion (that morality is biologically built in) based on Hauser's examples of ethical dilemmas. They fail to support such a model because, as I hope I've described below, it is easy to analyze the dilemmas without recourse to Hauser's biological explanation.]

Confusion caused by an unspecified condition

The flaw in his argument is simple and obvious once it has been pointed out. It is contained in:

It's the same question - should we sacrifice the one to spare the other five?

It is NOT the "same question" at all.

In the "railway game" there is an unspecified but absolutely vital condition. Neither of the potential targets (solo worker or group of 5) are aware of the impending disaster. ONLY the "signalman" has a) the facts and b) the means to act. ONLY the signalman is in a position to make a decision and in those circumstances it is not even a difficult decision. Clearly to do nothing and allow maximum casualties would actually render him ethically liable for the additional deaths he could have prevented.

You wanna make that a difficult decision? Replace the anonymous solo worker with the signalman's eldest child. Now it's still easy for us observers to make the objective judgement that the signalman's daughter must die. But how many of you - in the signalman's shoes - would make that choice? And how many would condemn him if he didn't? But that's by the by. Not part of Hauser's game.

In the "hospital game", there is no implicit equivalent restriction of information. ALL parties who could be affected by the decision are aware of the relevant data.

To make the two games and questions truly equivalent, you'd have to create a highly artificial condition in the railway game. Like having all parties chained to the track but, conveniently, all having accessible mobile phones so they can be informed of the problem and discuss potential solutions.

It is the absence - in the hospital game - of that unspecified condition from the railway game, which makes the questions dramatically different. The real question in the railway game is "what choice should the SIGNALMAN make?"

The real question in the hospital game is "what choice should SOCIETY make?" (where the sick people, the healthy potential donor and all the relevant medical staff constitute a reasonable analog for "society")

Hauser's mistake is to assume that the ethical issues are limited to the resolution of the problem itself. Should we sacrifice the one to spare the other five? But as soon as more than one person is party to the decision making process, a whole new set of ethical issues come into play. Most obviously, the one who might be sacrificed in the railway game has no idea of his impending doom; cannot (within the implied rules of the game) be made aware and, therefore, cannot be part of the decision making process.

In stark contrast, the one who might be sacrificed in the hospital game is fully aware of the situation and is thus entitled not only to participate in the decision but, as most people recognise, also to veto any decision which requires his demise.

Like I said, once it's pointed out, it's obvious. And it doesn't require a mythical genetic algorithm to explain our recognition of its obviousness.

Who takes part in the decision making process?

In general terms it ought also to be obvious that:

in the event of a problem or opportunity which requires a social (shared) choice to be made; anyone who is
a) aware of the existence and nature of the problem or opportunity - or can be made so if necessary
b) believes they have a reasonable probability of being affected by the decision or its outcome and
c) gives a shit
should be entitled to share in the decision making process.

It should also be (but rarely is) bleedin obvious that our share in the decision making process is NOT necessarily EQUAL. That share is determined by where on the spectrum your interests in the decision/outcome lie between involvement and commitment.

The difference? As a friend of mine occasionally reminds me, "When you're having ham and eggs for breakfast, the chicken is involved, the pig is committed"

If you're involved, you get a vote. If you're committed, you need both a vote and a veto.

Why we need the veto as well as the vote

In the hospital game - the medical staff are involved; the patients and potential suicide donor are committed. Even if they all (presumably except the donor) vote in favour of his sacrifice, his veto trumps them all. Similarly, even if the medics were to recommend a somewhat more rational option - that one of the already terminal patients be sacrificed for the benefit of the others - that patient's veto would block any amount of votes in favour. A vote never beats a veto, so even the 5 terminal patients' votes cannot outweigh the donor's veto. Why not? (other than the trivial observation that this is what "veto" means)

Let's not try to avoid or evade the question. Of course we know that - in real life - no 5 terminal patients in such a situation would be selfish enough (despite the jaundiced views of some economists) to demand the sacrifice of the donor. But this is a game we're playing. And the rules of this game include the stipulation that they ARE that selfish.

Why are there some situations in which a veto MUST trump any number of votes?

According to the author of the letter I mentioned above (cached), the reason the decision would have to be in favour of the donor's veto is that there is an implicit contract between all members of the community (regardless of their health status) and all hospitals such that it is understood that we only cross the threshold of the hospital premises on the strict condition that we will NOT be made involuntary donors under any circumstances whatsoever. Breach of that contract, even on a single occasion, would simply ensure that no one would voluntarily visit a hospital ever again. This is the "contractarian" view of ethics which, he argues, is superior to the utilitarian approach

Since contractarian theories predict the correct result, but utilitarian theories do not, this is really an argument against utilitarian theories, not a puzzle about right and wrong.

His "contractarianism" does, on this occasion, indeed predict the correct result, but it still doesn't get close to the heart of the issue. Merely predicting people's expectations does not constitute an ethical argument. But he is also right to point out that mere "majoritarianism" (greatest good for the greatest number) doesn't cut the mustard either. The relevant question is why shouldn't the random healthy visitor expect to be plundered for his spare parts? Why are they entitled to veto their own dismemberment?

Reciprocity - the borderline between Democracy and Anarchism

The ethical answer is one of the simplest, longest established and most widely recognised of all ethical principles. Reciprocity. Do as you would be done by. The Golden Rule.

Would you accept being made an involuntary donor if you were the healthy one in the hospital game? No? Then - even if you swap places and become one of the terminal patients - you can't justify doing it to anyone else. Period.

What we are looking at here are the "buffers" at the end of the decision making process. This the point at which social decisions cannot be allowed to override personal autonomy.

The edge is well defined. The terminal patients, or their representatives, can go so far as to petition the donor, but no further. The donor's decision is final. The donor's autonomy is absolute.

This is the borderline between social decision making - collectivism, democracy - and autonomy or anarchism; the zone in which only the personal writ runs. There is no conflict between democracy and anarchism, provided no-one crosses that border in either direction.

In a real world situation, you might well find that the donor's actual decision was a compromise. S/he may be able and willing to make some contribution short of suicide which helps at least one of the patients. Perhaps s/he can spare some blood; maybe some bone marrow; perhaps even as much as a kidney, all without fatal consequences.

But nobody can "insist" that the donor offers even so much as a toenail clipping, because we would not accept such an imposition on ourselves. It really is that simple. It doesn't require sophisticated rationalisation.

There is a common cliche of which I have been unable to trace the source. It's one of those which people treat as a pearl of wisdom but is, in fact, ethical and tactical bollocks.

"Your right to throw a punch (or "swing a fist") ends " - the cliche begins - after which you can find a variety of formulations which basically boil down to a choice of two. Take your pick:

"where my nose begins" or
"one inch from my nose"

Both are, of course, wrong and tell us that the source led a sheltered life and was never exposed to any level of combat training (or forgot all about it).

I am entitled to take evasive measures from the moment it becomes clear that you have begun to throw the punch. Some would argue that the moment your intent to strike becomes clear is the point from which I can reasonably begin my self-defence - but given that your intent might be to hit me at the end of next month that is clearly too vague. In any event, I certainly do not have to wait until the blow is about to land. Such a ludicrous proposition would imply, for example, that Britain would not have been entitled to go to War against Hitler until the first bombers crossed the territorial boundary of the United Kingdom.

In fact, as that case adequately demonstrates, I may even feel free to react to you throwing a punch at a friend of mine which never threatens me directly at all. Do I have the "right" to step in to help my friends? I don't give a damn, I'm going to do it anyway. You'd better factor that into your calculations before you attack my friends.

The Rules of the Democratic Game

Now, when it comes to making decisions which concern us all, I'm only prepared to play that game if it is accepted by ALL the other players that this boundary exists; beyond which the group cannot tread into my personal domain. The game cannot include a rule which demands my compliance with a decision which directly threatens my interests or wellbeing. No rule, if you like, can oblige me to accept the punch on the nose and forbid me to evade it, to defend myself or to retaliate. If you don't agree to those rules, I'm not going to take part in the game.

This is the essential consensus which must be present BEFORE we can begin any kind of democratic debate. Once we understand that, it becomes reasonably obvious how and why Democracy doesn't actually exist in the modern world. When were you ever given the opportunity to provide informed consent for or even to consider the "rules of the game" before being asked to consider the substantive issue? Come to that, when were you last given the opportunity - regardless of the rules - to decide ANY substantive issue?

Anyway, once we've reached consensus on the rules, we can hold the debate. How would this work in practice?

There are ten of us left in the lifeboat, and we've been adrift for 37 days. To make this game more interesting, we're OK for water but we're out of food and unable to catch any, the problem is obvious: If we don't eat something, we're all going to starve to death.

A proposal is made: If we were to begin eating each other, there is a chance that one or more of us will still be alive when and if salvation is finally at hand. It is proposed, therefore, that we decide, democratically, two issues:

1 In the circumstances, Cannibalism is a necessary and acceptable step to improve the chances that one or more of us will survive this ordeal.

2 If there is a consensus in favour of that proposition, the only fair way to select who is going to be sacrificed is by drawing random lots.

That first question is vital. It establishes the necessary consensus (or not) without which the drawing of lots cannot proceed. If even one person rejects the first proposition - that cannibalism is necessary and acceptable - then we cannot proceed to the second step. At least - we cannot proceed democratically.

But what if nine voted in favour of cannibalism and only one objected. Couldn't the nine over-rule or overpower the tenth? Probably, but the simple fact of their majority wouldn't justify their actions. Their murder of the dissenter wouldn't in any sense be ethical. And, if he was capable of defending himself against such odds, then anything he did in such self-defence WOULD be entirely justified.

Having said that, it does not mean that the dissenter's veto prevents all further action. For example, one thing the nine could still do, while remaining consistent with the golden rule, would be to conduct a further debate amongst themselves as to whether - given that the dissenter has refused to accept the basic premise (the current necessity of cannibalism) - they can or should proceed to select a sacrifice from among the nine who have accepted that premise.

If there is a new consensus in favour of that revised proposition, then not only can they now proceed to the method of selection and make the sacrifice, but they would also be entitled - again, purely on the basis of reciprocity - to withhold the "benefits" from the dissenter (not that s/he's likely to be inclined to accept a slice of the action in this case anyway)

In all the "survival at sea" or similar stories I've come across, I've never read of that initial question being put. They always seem to take that question for granted (that such ends do justify even those means). Or perhaps they have always reached an informal consensus where everybody clearly accepts the necessity and acceptability of cannibalism. It is, of course, difficult to predict your own reaction in such circumstances unless you've been in a similar situation.

Personally, if I ever find myself in that situation, I hope that I would be a dissenter. I would not wish to spend the rest of my life aware that I was only still here because I helped kill and eat a fellow traveller. And I certainly wouldn't permit the other nine to elect or select me. I might, of course, be unable to avoid compliance because I am outnumbered. But I would at least go down fighting.

On a "good" day, I might, perhaps, be inclined, to volunteer for the pot myself (however, given my innate physical cowardice, it would probably be because I misunderstood what sort of pot they were talking about) In fact, I'd have a hard time accepting even a volunteer. At least not one who publicly made such an offer. I could not condone their suicide any more than their murder, for my benefit.

If, on the other hand, we woke up on the 38th morning and found one true altruist had committed suicide overnight and left a note begging us to eat him or her, I would be utterly appalled but almost conscience driven to try to digest a piece of the corpse if only to try to provide some meaning for their noble self sacrifice.

Others may draw their personal line at different points. That's the point of personal autonomy. I can't judge what you will tolerate or where your boundary lies. You have to make that decision for yourself. Similarly I won't let you tell me where my boundary must be.

There is nothing ethically wrong with ANY method of arriving at a decision providing ALL parties give their free and informed consent to the decision making process and ALL indicate similar free and informed consent to be bound by the outcome.

In case I didn't make it clear enough, the key word in that paragraph is ALL. The dissenter cannot be bound by rules s/he did not accept in the first place. The fact that a majority has accepted a) the rules and b) the decision - doesn't provide them with any ethical justification for imposing their decision on the dissenter who didn't agree the rules.

Coercion and Compliance

The only advantage that arises from having a majority in this scenario is a military one. The majority outnumbers the dissenter. They can choose to force the dissenter to comply. But in that case, there is no ethical barrier to the dissenter taking whatever steps s/he can to defend themselves against majority attack. Of course, in practice, the balance of forces might be enough to persuade the dissenter to comply, but that constitutes only a pragmatic reason for compliance, not an ethical one.

Which is not to deny that there will be situations where the majority will, nevertheless, need to exercise that military advantage and force compliance. For example, let's imagine that the dissenter's primary reason for dissent is that he holds private property in the form of a significant hoard of food he has managed to stash away and is refusing to share. In the circumstances, of course, his refusal to share is grossly unethical. More importantly he is breaking the golden rule, or at least exercising it in a way the others are unlikely to accept.

It is unlikely that even the most ardent defenders of "private property" (among the nine who voted for cannibalism) would prefer to starve or to kill and eat a fellow cannibal rather than deprive the dissenter of any part of his stash without his consent. Of course, they would prefer he shared it willingly and they might even agree that, after rescue, they would compensate him for his losses at a fair market rate.

But if he stubbornly refuses to share life giving resources despite all attempts at persuasion and negotiation, then the only rational conclusion the other nine can come to is that the dissenter is now an obstacle to their continued survival. At which point the military option becomes the obvious choice and the dissenter, and his stash, will all be sacrificed to the common good. The wise dissenter will concede well before matters reach that ugly stage, albeit for pragmatic rather than ethical reasons.


So what conclusions can we draw from all this?

That the veto is at least as important as the vote and is required to prevent "Tyranny of the Majority".

That "greatest good for the greatest number" - simple "majoritarianism" might be a good general guide but cannot be universally applicable. It only works where the issue being decided doesn't cross anyone's personal boundary. It isn't ethically valid in any situation where a potential dissenter feels strongly enough to exercise a veto. It implies that a majority can justify harming a minority if such deliberate harm is perpetrated for their own (majority) benefit.

We all implicitly reject such a social decision making structure (if we're awake) because we all recognise that on some issues we may each, one day, be part of the minority rather than the majority and we want to ensure that if we're ever in that position, we too will be treated fairly and reasonably, so we agree to do as we would be done by.

This is how simple reciprocity scales up to the level of democratic society. It establishes the borderline of majority rule. Democracy, as a decision making process, has to begin from the position of universal consensus that it is the way to decide the particular issue under discussion.

Where it achieves that prior consensus, it is probably the best way to make the decision because it will ensure minimum dissent regarding the outcome. Where, however, we cannot achieve the prior consensus that a given issue is appropriate for decision by the democratic process, the imposition of that process and its outcome is no better, no more ethical, than any other form of tyranny.

What also becomes obvious is that when an issue arises, resolution is never a matter of a single question. With the democratic cannibals, there were at least two questions; was the cannibalism necessary and how should we select the victims. We also saw that further questions could arise if one or more of the stranded group exercised veto over the cannibalism.

In more generic terms the process can be outlined in the following questions:

1 Do we agree there's a problem or opportunity?

2 Do we agree on a potential solution (or exploit)?

3 Does the potential solution/exploit cross anyone's personal boundary and, if so, does that person wish to exercise veto?

4 If veto is exercised, can a limited version of the solution/exploit be implemented which achieves the ambitions of those who support it but doesn't breach the boundaries of the veto wielders?

5 If the answer to (4) is "No", and there is a substantial majority in favour of the solution, is there consensus among that majority that the issue is so serious that they need to persuade or even coerce the veto wielders into compliance - and if so how?

6 In the light of a "Yes" answer to (5), does the veto wielder still wish to maintain their veto. Will they compromise or resist? Is the matter negotiable or do we have to go to war?

Those questions summarise almost any potential democratic debate. The first 3 should always be asked. The last 3 only become necessary once vetos are exercised. Whenever we need to address question 5, then what we need to try to find is a "pareto efficient" solution - which is one in which no one loses and at least one party gains. That's a whole new topic which can wait for another day.

Meanwhile, what the anti-democratic Libertarian Right have to acknowledge is that it is equally bleedin obvious that some decisions MUST be collective and require total compliance. (Obvious example: Which side of the public highway should we drive on?) What the Libertarian Left have to acknowledge is existence of the boundary condition which must be protected by the veto, without which Democracy could indeed become tyranny.

So now we've sorted that out, can we please get together and set about removing the real tyrants?

(made front page of K5 same day. See that page for some of the discussion which has already taken place)

No comments: