The reversal test, status quo bias, and opposition to human cognitive enhancement

Abstract

Bostrom and Ord’s reversal test has been appealed to by many philosophers to substantiate the charge that preferences for status quo options are motivated by status quo bias. I argue that their characterization of the reversal test needs to be modified, and that their description of the burden of proof it imposes needs to be clarified. I then argue that there is a way to meet that burden of proof which Bostrom and Ord fail to recognize. I also argue that the range of circumstances in which the reversal test can be usefully applied is narrower than they recognize.

Keywords:

• Argument from risk
• backfiring
• burden of proof
• human cognitive enhancement
• reversal test
• status quo bias

Acknowledgment

Thanks to Matthew Kopec, Neil Levy, Toby Ord, Russell Powell, and two anonymous referees for helpful comments on earlier drafts of this paper as well as audiences at the University of Oxford, Charles Sturt University, the University of Helsinki, and the Australasian Association of Philosophy annual conference in 2015.

Notes

1. Bostrom and Ord (2006) has been cited over 144 times, as of 3 April 2016 (Googlescholar).

2. Bostrom and Ord (2006) also introduce a second test for status quo bias, the ‘double reversal test.’ This rather complicated test does not appear to have been applied by any philosophers, at least not in published articles and books. Criticisms of it have been made by Weidemann (2009, 125–7) and Nordmann (2007, 39, n. 22).

3. There is some critical discussion of reversal tests in a recent paper by Sparrow (2015) as well as in a set of published commentaries on that paper. However, Sparrow does not apply Bostrom and Ord’s (2006) reversal test. Instead he applies his own test, which is loosely modeled on Bostrom and Ord’s test. The differences between Sparrow’s test and Bostrom and Ord’s reversal test are discussed in two of the commentaries: Campbell and Wahlert (2015) and Powell (2015).

4. Nordmann also accuses Bostrom and Ord (2006) of failing to comprehend that they need to treat approaches to ethics that are neither consequentialist nor deontological seriously. See (Nordmann 2007, 39, n. 22).

5. Nordmann compares Bostrom and Ord’s use of the reversal test, to reverse the burden of proof against the opponent of human cognitive enhancement, to contexts in which creationists illegitimately claim that creationism is on an evidential par with the theory of evolution, because biologists cannot offer absolute proofs of the truth of evolutionary theory (2007, 39). This comparison misses its mark. Creationists seek to evade the burden of proof that mainstream science imposes on those who promulgate theories that are inconsistent with accepted science. However, creationists do not seek to impose a burden of proof on evolutionary theorists. They merely seek equal treatment for creationism and evolutionary theory. So they seek to shift, rather than reverse, the burden of proof. Also, by treating evolution and creationism as just two unproven theories, the creationist evades consideration of the overwhelming evidential basis for acceptance of evolutionary theory. But there does not appear to be any analogously strong evidential basis for the rightness of opposition to human cognitive enhancement that Bostrom and Ord (2006) attempt to evade.

6. It is not clear that Nebel appreciates that he is redescribing Bostrom and Ord’s reversal test when he describes it as a test for irrational status quo bias. See Nebel (2015, 453–455).

7. The endowment effect is nicely illustrated by a simple experiment due to Jack Knetsch. See Kahneman (2011, 296–297).

8. For discussion of a simple study, which vividly illustrates the influence of loss aversion on ordinary thinking, see Kahneman (2011, 283–284).

9. For a recent survey of effects contributing to status quo bias, see Eidelman and Crandall (2012).

10. Humans already enslave and exterminate other humans. Presumably, Annas, Andrews, and Isasi’s (2002) underlying concern is that post-humans would be more effective enslavers and exterminators of humans than are humans.

11. References for some more suggested risks are collected by Bostrom and Ord (2006, 663, n. 14).

12. However, there is some discussion of this topic in Huxley ([1958] 2007).

13. There are many risk-averse approaches to the management of risk and many involve applying variants of the ‘precautionary principle.’ The precautionary principle is not often advocated explicitly in discussion of the right and wrongs of human enhancement, but there are commonalities between the forms of risk aversion advocated by many of the opponents of human enhancement and those advocated by proponents of the precautionary principle (Briggle 2014).

14. On the origins of urban diseases, see Diamond (2012, 295).

15. Modern lifespans are, on average, significantly longer than the lifespans of Australopithecus (Sacher 1975). If Bostrom and Ord’s Australopithecus were to raise their intelligence to the level of modern humans then they would, presumably be able to figure out how to extend their lifespans, at least to some extent. This seems like a clear point in favor of enhancing their intelligence. It is hard to know how happy Australopithecus were. But available evidence suggests that members of pre-industrial cultural groups who lead uncomplicated lives are at least as happy as those living in modern industrialized societies (Biswas-Diener, Vittersø, and Diener 2005). A cognitively enhanced Australopithecus would, all things being equal, live a longer life than an unenhanced Australopithecus, but there seems to be no good reason to think that it would be a happier life.

16. Bostrom and Ord (2006) might concede that the various cognitive biases that infect ordinary reasoning would make it hard for the Australopithecus to appreciate the full extent of the dis-benefits of modern levels of intelligence, but go on to argue that cognitive biases might lead them to fail to weigh benefits and dis-benefits even-handedly; and so fail to appreciate the overall benefits of modern levels of intelligence. Given the large number of cognitive biases that infect human cognition, in different and conflicting ways, and which can combine to have unexpected effects, it seems that such a line of argument would be speculative at best.

17. A potential response to this line of objection would be to invoke the possibility of moral enhancement, as advocated by Persson and Savulescu (2012). If cognitively enhanced people were also morally enhanced, then their selfish and parochial tendencies would be replaced by pro-social tendencies. I don’t think this is a credible response. I am persuaded by Powell and Buchanan (forthcoming) that an appreciation of the evolutionary history of human morality should lead us to conclude that proposals to use biological and technological interventions to morally enhance people are very unlikely to succeed.

18. For an extended discussion of this theme in conservative thought, see Kekes (1998, 68–90).

19. Some of the concerns raised by opponents of human cognitive enhancement are about the potential deleterious side effects of successful human cognitive enhancement. Others are about the potential deleterious effects of attempts to cognitively enhance humans backfiring. As far as I know, no opponent of human cognitive enhancement has explicitly distinguished between these two different forms of potential harm.