September 29th, 2009

A partial retraction on opium

In my last rant on Afghan opium, still pushing what seems to me the sound ploy of buying the crop to expand pain relief in poor countries, I accused the drug warriors in the policy establishment of ignoring the idea out of “pathological tunnel vision.” Mark, who is unlike me a card-carrying drugs policy expert, tells me I’m wrong. The plan, in the form presented by ICOS, has caused a stir, is being looked at in Washington, and there’s even a draft paper submitted to a leading journal criticising it. So my accusation is factually wrong and I’m glad to withdraw it. (I wasn’t to know; the public sphere is still silent.)

How about the tunnel vision? It’s sans objet as the French say. But my language was I hear resented. “Pathological” went too far so I also retract that. I didn’t say and and didn’t mean to suggest that the policymakers are personally indifferent to African cancer sufferers, just professionally so, because they don’t see it as their job. Recognizing that most of the world’s problems can’t be your business is the only way for any of us to stay sane and get anything done. I’m not for instance doing anything just now to help the local cancer charity shop or foreign prisoners in Spanish jails: there are plenty of volunteers for the former and I don’t have the skills for the latter, so public-policy blogging looks a better use of my time.

I’m not retracting anything more than this. Accusing experts of tunnel vision, aka cognitive bias, is a cheap shot because it’s almost always true. Part of what it means to be an expert is to belong to a community with broadly shared assumptions about what problems are important, what analytical approaches are legitimate, which past work is a fundamental reference point, and where the debate takes place. If you have a hammer,¬†problems tend to look like nails. As Thomas Kuhn pointed out in his great The Structure of Scientific Revolutions, scientific progress in normal, non-revolutionary times depends on a mesh of such assumptions. Scholars cannot do real work if everything always has to be argued from first principles. You do not have to accept Kuhn’s more extreme claims about the cognitive incommensurability of different paradigms to accept the plain truth of this weaker, sociological version.

If you like, each scientific community is a work gang of miners chipping away inside its own tunnel. Nothing stops the miners from exploring other coalfaces. But as a matter of fact, and reflecting their professional incentives,  most do not.

Francis Bacon started the analysis of cognitive bias with his picturesque fourfold scheme of idols that our minds falsely bow down to: of the tribe (innate human tendencies), the cave (individual temperament and social roles), the market-place (from semantic confusion and the tyranny of words), and the theatre (from sophistry and false learning). Great stuff, but his belief that we could at one bound escape the four idols with his terrific invention of the scientific method was hopelessly optimistic, as Kuhn showed. Each tribe of scientists still carries its handy portable altars. The crucial point is not an impossible blank slate, but the possibility of self-correction and objective learning imposed by the covenant of all the tribes to accept the ultimate sovereignty of facts; the principle honoured in the title of this blog.

Where am I going with this? Evaluating ICOS’ and my proposal, if you are going to do it properly, requires three sets of expertise: public health, on the delivery of pain relief in poor countries; counter-insurgency, on running programmes in villages to help Afghan farmers and cut the flow of funds to the Taliban; and drugs policy, on the risks of diversion and the impact on the supply of illegal opiates. Since it’s uncertain whether any interventions in a single country can significantly affect the global illegal supply and reduce heroin abuse at home, I don’t see how the drugs policy perspective can be particularly useful here, and it certainly shouldn’t dominate.

I also suspect that the cognitive bias of the drugs policy field includes pessimism. Drug abuse, like child abuse, the cycle of poverty, suicide, murder and other social ills is highly intractable: the aim of policy is to contain and if possible reduce the evil, knowing that many attempts to make things better have actually made them worse. Public health and counter-insurgency types tend to be optimistic: quite reasonably in the former case, as they can look back on a long roster of defeated diseases; far less so in the latter, where the list of successes is short and their human cost often appalling. Anyway they’d be more likely to see new stuff as worth trying.

Since we can’t escape cognitive bias simply by knowing more, problems of framing are also inescapable. You can only apply a systematic utilitarianism from a vantage-point, like mine, of superficial pauciscience. Still, I think we pauciscients can still be useful, just by asking new and annoying questions that cross traditional boundaries of expertise and bureaucratic competence.

You can see the way this works out very well in my idea. If you frame the problem as illegal opium in Afghanistan, it looks like a choice between three strategies at village level :

A. Eradicating the crop. Pros: none. Cons: ineffective, alienates villagers.

B. Substituting other crops like apples and pomegranates. Pros: genuine long-term solution. Cons: hard to set up and sustain, expensive initially.

C. Legalising opium production. Pros: genuine long-term solution; potentially benefits African cancer sufferers; longer-term income guarantee; partially pre-funded by postulated cancer scheme. Cons: hard to set up (=B); risk of diversion; no cancer scheme in place for Africans etc.

In this analysis, the legalisation is a bad bet absent the currently nonexistent scheme for the Africans, and at best no better than substitution.

But if like me you frame the problem as pain relief for the world’s poor, there are two strategies:

X. Buying lots more legal opium from existing suppliers (India, Australia, Turkey). Pros: sure thing using existing legal and farming infrastructure, no additional risks. Cons: none really.

Y. Buying lots more legal opium from Afghanistan. Pros: helps win war in Afghanistan. Cons: much harder to set up (but the alternative in Afghanistan is plan B, so you already have a policy and village-level infrastructure).

In this perspective, the Afghan option offers the potential of a double payoff and is worth at least trying an actual experiment, as Bacon would have suggested.

But let’s be clear: the failure to provide morphine to millions of poor sufferers is a terrible scandal; it’s easily and cheaply fixed as evils on that scale go; and if the morphine can’t be sourced from Afghanistan, then Australia will do just fine.


Share this post:
  • Twitter
  • StumbleUpon
  • Digg
  • Reddit
  • Facebook

2 Responses to “Idols of the tribes”

  1. Avocet says:

    The real problem, it seems to me, is to identify crops other than opium as to which various regions of Afghanistan have a competitive advantage. Taking transport costs into account, surely there is something, but I have seen little discussion of this issue in the press.

  2. Mark Kleiman says:

    There’s an extended comment inthe next post. Short version: since the opium to make a dose of morphine costs a penny, under-treatment of pain – which is certainly an important problem – does not result from the cost of opium, and therefore trying to fix it by increasing opium supplies won’t work.