March 27, 2008

Interstellar trade, brains in vats, superspace, life extension and the varieties of silliness

BusinessbeingsI just read Paul Krugman's very entertaining The Theory of Interstellar Trade. That led a colleague to show me the paper Stuperspace, and I had to respond by showing her the Brain in a Vat essay. Laughter all around!

Krugman writes: "It should be noted that while the subject of this paper is silly, the analysis actually does make sense. This paper, than, is a serious analysis of a ridiculous subject, which is of course the opposite of what is usual in economics". While funny, there is something problematic with this outlook. Why is it silly to consider interstellar economics?

That we haven't got interstellar business and that it will not arrive for a long time doesn't preclude serious study. Looking through the economic literature I am pretty certain we can find an enormous number of papers about situations that have not happened or will never happen (and that leaves out all the unrealistic Gedankenexperiments).

What does make interstellar trade silly, while the economics of world government seems to be non-silly? Both papers use real economic theory, empirical data and known physics to make statements, but the later seems far less likely to be regarded as silly. It can't just be that it is in a peer reviewed journal and nicely typeset. Maybe we can imagine world government more easily than interstellar travel, but that seems unlikely to make much of a difference - most people can just imagine interstellar trade as very long sea journeys, while globalisation is far harder to visualize. Similarly, it is easy to guess that the engineering issues involved in interstellar trade can be enormous, but most people have little knowledge of the actual details. Is a low prior probability enough to move it into the silly category while world government (with presumably big political difficulties) is assumed to have a much higher probability? Again, this seems unlikely. Most people would lump them into the category "really unlikely but possible".

Graziella Caselli, Jacques Vallin, Demographic Trends: Beyond the Limits?, Population: An English Selection, Vol. 13, No. 1, Biodemographic Perspectives on Human Longevity (2001), pp. 41-71 is a silly paper about a serious matter - or a serious paper about something silly. They look at the demographic effects of various scenarios of life extension and gender bias, finding various results. Again, the methods are conservative and the assumptions are not that extreme (no immortals etc). Yet the authors are constantly making fun of the whole issue. In the discussion they reject their own conclusions by regarding them as equally far-fetched as Martian invaders kidnaping Earth women! Why is a 150 year life expectancy silly? As the paper mentions, current trends show no sign of slowing and we have historically seen doublings of life expectancy - even when leaving out radical technological possibilities 150 year life expectancies are within the range of the thinkable.

One reason they make fun of their own work is that they realise the arbitrariness of their assumptions (Krugman does this too). Caselli and Vallin have to make some assumptions to feed their model and they know these can be varied endlessly: even a valid prediction will be undermined if we are very uncertain about the value of its parameters. But this does not mean the predictions are useless: the population oscillations caused by delayed fertility or the asymmetry between male and female gender selections are obviously relevant.

Meanwhile the brain in a vat paper makes fun of discussions philosophers tend to take somewhat seriously by lumping them all together into absurdity. The trolley problem has produced a large body of writing, and it is by no means trivial - especially the fat man version has produced an intense side-discussion about neuroethics (see Greene, Singer and Kahane). The example may be silly, but most people recognize that it is serious in meaning. But when combining other serious thought experiments they suddenly become silly - the brain in a vat is non-silly (despite being completely science fictional), the trolley is non-silly, but a brain in a vat in a trolley is silly. Partially it may be because the suspension of disbelief engendered by stating that it is a thought experiment intended to explore something important (the nature of reality, utilitarian morality) is broken when it is conflated. But presumably there might be serious matters that deal with both perception of reality and utilitarianism? That they are rare in normal life is not an issue, we do not suffer ethical dilemmas most of the time either.

Stuperspace is just plain silly, parodying the style and hyperbole of string theory. There doesn't seem to be any serious conclusion, not even that string theorists are pompous, just enjoyable wordplay and parody of academic style.

Maybe one could apply a nonsense approach to distinguishing these kinds of silliness. Nonsense is an utterance within a language that does not have any identifiable meaning. There may be levels of nonsense, since some meaning may be hidden or the absurdity lies on different levels. Stuperspace is just barely grammatical (and the math is mostly random). The trolley story makes sense as a fictional setting, but is an exceedingly unlikely situation - the link to reality is extremely tenuous, making the meaning (if any) equally tenuous. But both the interstellar trade paper and the population paper take place in hypothetical worlds which are much more closely linked to plausible future worlds. They have meaning, they could guide actions in a rational way (should I invest in a space business venture? Is a particular population policy likely to produce good results?), but they are largely hypothetical. This does not make them nonsense, just as the extrapolations of the world government paper does not make it nonsensical.

Another approach to silliness is to argue that certain papers are foolish in the sense that they waste time and effort that could be used for something more useful. But this evaluation is both author/reader dependent (distracting from some pointless work would make the paper non-silly) and depends on a risk/benefit estimate. The impact of life extension is likely to be big and is often used as an argument about whether funding such research is a good thing. Hence a paper looking at it is at least useful for guiding those funding decisions. Some risks are exceedingly unlikely but if they would happen they would be disastrous (e.g. existential risks); hence it may be rational to not just explore them and come up with protections, but also to explore whether they have any appreciable probability in the first place. But clearly we need to use our priors to draw the line somewhere, since there are an infinite number of improbable and unthinkable risks we could consider (planet-eating space goats!)

But again, regarding some things as silly does not seem to result from an estimation that the probability is extremely low, it seems to be a direct rejection of it as unthinkable and irrelevant - not the same thing, although the rejector will quickly argue that the chances of the things happening are minuscule. The rejection has many similarities to the yuck reaction we see in ethics, where certain possibilities are rapidly rejected as immoral with little reflection (c.f. the work of Haidt). So maybe the best explanation of what makes a paper silly is just that it goes against the social intuitions we have built up about thinkable, serious subjects. Space travel is science fiction and science fiction has low status, so hence papers about the economics of space travel must be silly. Life extension is silly, so papers looking at its consequences must be silly. Framing world government in terms of non-silly globalisation makes it non-silly.

If this is right, we should expect that many silly things are actually worth taking seriously on their own terms. As long as they are not nonsense.

[ Overcoming bias has a post on the subject with some relevant comments. ]

Posted by Anders3 at March 27, 2008 07:31 PM