Skip to content

Emergence’s devil haunts the moral enhancer’s kingdom come

It is 2025. Society has increasingly realised the importance of breaking evolution’s chains and enhancing the human condition. Large grants are awarded for building sci-fi-like laboratories to search for and create the ultimate moral enhancer. After just a few years, humanity believes it has made one of its most major breakthroughs: a pill which will rid our morality of all its faults. Without any side-effects, it vastly increases our ability to cooperate and to think rationally on moral issues, while also enhancing our empathy and our compassion for the whole of humanity. By shifting individuals’ socio-value orientation towards cooperation, this pill will allow us to build safe, efficient and peaceful societies. It will cast a pro-social paradise on earth, the moral enhancer kingdom come.

I believe we better think twice before endeavouring ourselves into this pro-social paradise on the cheap. Not because we will lose “the X factor”, not because it will violate autonomy, and not because such a drug would cause us to exit our own species. Even if all those objections are refuted, even if the drug has no side-effects, even if each and every human being, by miracle, willingly takes the drug without any coercion whatsoever, even then, I contend we could still have trouble.

Surprisingly, the scenario imagined in the first paragraph is not that far-fetched. The field of cognitive moral neuroscience and the study of moral cognition have been flourishing; we have already found many neurochemical manipulations which seem to alter our social and moral preferences. Dr. Molly Crocket and Professor Ernst Fehr’s recent review published this August addresses many of the methodological considerations involved with the study of neuromodulation of social and moral behaviour. Such overview was deeply needed as there has been a surge of studies in that area in the last 5 years. The Oxford Martin School recently hosted an event on the psychological and neurological basis of pro-social and cooperative behaviour, which was broadcasted on this blog. Anyone interested in moral or social enhancement should be particularly interested on these areas. However, missing even from the excellent methodological review conducted by Dr. Crockett was an account of how to address between-groups cooperation. Yet this is arguably the ultimate problem the moral enhancement hopes to fix.

However attractive the cooperative paradise may seem, it is not clear that increasing individual agents’ levels of cooperation will increase overall cooperation on society. Although one might expect an increase in our individual tendency towards cooperation between individuals would increased between groups cooperation on higher levels, it should be made clear that what we want is the latter. That is, we want to increase higher level, between groups cooperation. We already know how to cooperate within small isolated groups fairly well. As Joshua Greene puts it, the problem is when groups collide and their cooperation solutions are incompatible. Then we have the meta-cooperation problem, as he states it.

Hence, the question is: does individual, lower level between individuals cooperation entail between group, higher level cooperation? The answer is no. But not only there is no necessary connection, there are in fact plausible mechanisms whereby lower level cooperation actually decreases higher level cooperation. If we bear in mind many well understood examples of higher level properties, this shall come as no surprise. Here I will understand high level properties (or emergent properties) as patterns or organizations which emerge out of simpler lower properties or interactions. A classical example would be snowflakes’ symmetrical patterns arising out of tiny supercooled cloud droplets.

High level organization can often possess unexpected features. This is true even for the simplest physical processes. Take a paten with water being heated from below. A natural heating convection flux will occur and the process governing heat conduction in the lower microscopic level will follow a disorderly random movement. However, under certain relatively simple settings, macroscopic, orderly and stable hexagonal structures will visibly surface. Flatten the paten, the hexagons turn into spirals; increase temperature too much, the patterns shatter into chaos. Could increasing cooperation too much shatter social institutions?

Individual ants present almost a random and chaotic behaviour. If one were to get acquainted only with individual ants, it would seem such feather-brained creatures were incapable of any complex organization. But, place many of those ants together, and patterns will emerge, synchronizing the ant colony as if it were a single macroscopic creature. Appositely, if we meddle the seemingly chaotic ant’s individual behaviour, the macroscopic pattern can change in dramatic ways. Could meddling with individual cooperation completely reshaped social structures?

Here are some examples I could gather on a Thursday evening:

(1) Parochialism: The most classic examples are cases where increased cooperation inside a group leads to decreased cooperation and even aggressiveness between groups. Each individual values his own group to such an extent that, sometimes he may harm himself and others if under the belief he will benefit his own group doing so. This is particularly worrisome since one aspirant for moral enhancement, oxytocin, is known to produce such effects, leading to ethnocentrism and parochialism.

(2) Necessary evil: The way society’s economy is currently organized heavily relies on individual agents being at least moderately self-interested. Macro-economical models and policies often depend on such assumptions, and we aim at building our higher level cooperational structures over that lower-level individualistic foundation. Western capitalist societies are noted in particular for their reliance on individualism. It may well be that we could achieve higher level cooperation far more easily if everyone were completely cooperative. A society could even become vastly more cooperative that way. But so much for wonderful contractual possibilities if to get there we would have to decrease individualism in a society which relies on individualistic self interest.

(3) Leaders: The way politics is organized also relies in individualism. One common feature of cohesive cooperative societies is the presence of leaders. Often, the only feasible way of constraining so many desires, values, opinions and positions into one cooperative group is to delegate responsibility to fewer individuals. But if we nearly extinguish individualism, it might be the case that no one will ever want to stand out and become a leader. Again, it is plausible that such highly cooperative societies would not need a leader. However, if there is only one single iteration on the cooperative enhancement process which leads to a society cooperative enough to extinguish great leaders but still individualistic enough to make cooperation without leaders unfeasible, uncooperativeness would surface.

(4) Polarization: Advocates of moral enhancement trying to improve our crude primitive morality might try to erase black and white political thinking. Some see certain political debates as the most clear cut examples of irrationality. Many cognitive biases arise when arguing for inflamed positions. Contra Karl Popper, we selectively search for evidence favouring our position and neglect evidence which goes against it. We often are blind to the fact that likely each side has its faults and merits. As mentioned above, we might even take actions to harm other parties and ourselves when defending our party. Now the boorish and thirsty moral enhancer would say “If we ever had a chance to morally enhance, this would be it! By eliminating such blind political irrationality!”. Not so fast. Higher level organization between thousands or millions can only arise if they compromise with having only a few opposing choices or political parties. If individual’s positions could cut through opposing political positions – as reason would dictate -, chaos would emerge. Imagine if people could solely cheer good moves on football instead of supporting teams, shortly after, football would disappear as a social or profitable institution. A world full of politically unbiased individuals would fall apart.

Paradise on the cheap might cost us dearly. Moreover, if the neuroscience or psychology of morality and social behaviour solely focus the search for moral enhancements on the individual level, then, not only might we create social catastrophes but we will be ignoring very important aspects which pertain to the very nature of the problems moral enhancement desires to fix: between groups meta-cooperational issues. Although properly addressed in Savulescu & Persson’s Unfit for the Future and elsewhere[4], the consequences of moral enhancement to politics, international relations and conflict resolution can only be fittingly addressed by de facto scientific research pursuing the development of moral enhancements if the experimental settings focus on group’s social strategies towards other groups, rather than individual’s strategies towards other individuals.

More generally speaking about any enhancement, a strong case has been made that evolution – that morally blind, fickle, and tightly shackled tinkerer who should be in jail for child abuse and murder – often gets stuck on local optima, and that, bestowed with the powers of technology, we could rescue our species out of these immoral imprisonments. Nevertheless, getting out of evolutionary local optima might not be so easy; they are local optima for a reason: every nearby solution is even worse, even if far away solutions are vastly better[3]. On the impetuous of breaking evolution’s chains, one should be careful not to stumble and fall in the local optimum’s moat. Having opened the gates of evolution’s perverse captivity, enhancement advocates should pay attention when searching for easy money on design-space’s garden, for its low-hanging fruits, so close to our reach, might be guarded by abrupt moats one might be forever stuck in. Indeed, it seems plausible there would be a technological path out of the local optimum. But perform willy-nilly human enhancement, and we fall on the moat. As the four examples above indicate – particularly so 2 and 3 -, the safe path might as well be a very thin ridge. Should we fall the mountain pass, we could be left wishing for a slippery slope, instead of the painful rocky cliffs of a disrupted society.  Better get stuck at local optima than absolute minima.


[1] Strictly speaking, some local optima will also be a global optimum. However, here I’m relevantly concentrating in those cases where although neighbouring candidate solutions are worse, there is in fact a better solution further away. In those cases, any hill climbing optimization process will likely fail at finding the global optimum.

[2] Araujo, Marcelo (forthcoming) “Moral enhancement and political realism: Can bioengineering make the world safe”

Share on

7 Comment on this post

  1. Very polemical.

    I was thinking of : “Contra Karl Popper, we selectively search for evidence favouring our position and neglect evidence which goes against it. We often are blind to the fact that likely each side has its faults and merits. As mentioned above, we might even take actions to harm other parties and ourselves when defending our party.”
    I wonder to what extend this is actually the case, to what extend we cannot de-polarize political opinion and emerge unharmed. There is smaller variation within the political opinions of Fins than Colombians I guess. Which doesn’t seem to be causing devastating troubles to the political climate in Finland. Or Costa Rica, for that matter.
    There are several paradoxes impeding the possibility of politics: The cognitive bias, the Condorcet voting paradoxes, the unilateralist curse, and the fact that if your actual position is 0.7 but you know most people are voting 0.4, then you are fully justified in voting 1.0, though you don’t actually believe 1.0. (Not to mention the fact that voting in itself is counterfactually irrelevant).
    There are too many problems in politics for us to be able to enhance ourselves out of the conundrum on the fly, but I wouldn’t expect that the reasons why we can’t do it is because being less polarized would harm more than help.

    1. My main point is rather that de-polarizing might harm, not that it will harm. This is sufficient to prove that this prima facie straightforward way of enhancing could fail in unexpected ways.
      Nevertheless, I do think de-polarizing is likely to bring disruptive instability. In your example of Colombia and Fins the correct case would be to make it so both Colombians and Fins had opinions which would make them hard to be identifiable as either Colombians or Fins. I think if we generalize this case such that every individual would be unidentifiable according to his country or political party, then the amount of noise would be enough to make democracy impossible. There’s a need to compromise. This is not to say such compromise wouldn’t be rational on a meta-level, only that focusing solely in the individual level would decrease this meta-level rationality (and also that many frameworks we might use to develop moral enhancements – say cognitive biases – are blind to this meta-level).

  2. This is a special case of the more general principle: Societies are complex. To predict the overall social consequences of some particular change, you need to know a lot of social science. There may be some simple changes that one can reasonably expect to have positive overall consequences, but making individuals more “moral” is not one of them.

    1. I don’t think there’s any simple change with expected value as big as moral enhancement (done properly). We are a kludge of good-enough-floppy-solutions to-problems-that-do-not-exist-any-more build upon another. It is unreasonable to expect this will be sufficient to deal with modern problems. The fact that enhancing social traits will be extremely complex does not dismiss the question. We need to have proper moral enhancement, complex as this might be. More fatalistic, such types of enhancements are already being pursued, so we better get them right.
      More generally, enhancement technologies will not be confined to individual traits such as cognition, and the social impacts of these new types of tech need to be addressed (as you have done for emulations, so I gather we agree on that).

  3. The degree of co-operation required to undertake any such transhumanist project on a global scale would be well beyond anything we can achieve for the foreseeable future, or perhaps ever. Although it may one day be possible for humans to take control of their own evolutionary destiny through genetic engineering and other biotech techniques, it’s probable that different groups will seek to “evolve” in quite different directions, in line with differing ideologies, cultural traditions, “identity politics” and so on. Any group that seeks to become more peaceful, rational and co-operative will probably need to contend with groups trying to achieve much the opposite. It may be that the only way the more peaceful and co-operatives humans can continue their project would be to leave Earth and find somewhere else to live, as far away as possible from those choosing nastier evolutionary fates.

    1. Even leaving Earth won’t save you from competing with other groups. In order to move the distribution of features away from a competitive equilibrium, we’d have to collectively take control of reproduction everywhere in the universe.

    2. I agree with your first claim, my scenario in the beginning was merely argumentative. I intended to show that even if we do everything else right but focus on the wrong between individuals’ cooperation, things would go wrong.
      Regarding having to leave Earth, I disagree. Evolution has been maximizing for fitness and this has lead to an increase in cooperative levels in the long-run. Hence, it does not seem like non-cooperative groups would survive for long.

Comments are closed.