Here’s a scenario that’s not too difficult to imagine: a drug company has recently released a paper claiming that a product they produce is both safe and effective. It would be foolish of any company with such a product to release a report saying their drugs were in any way harmful or defective, as it would likely lead to a reduction in sales and, potentially, a banning or withdrawal of the drugs from the wider market. Now, one day, an outside researcher claims to have some data suggesting that drug company’s interpretation of their data isn’t quite right; once a few other data points are considered, it becomes clear that the drug is only contextually effective and, in other cases, not really effective at all. Naturally, were some representatives of the drug company asked about the quality of this new data, one might expect them to engage in a bit of motivated reasoning: some concerns might be raised about the quality of the new research that otherwise would not be, were its conclusions different. In fact, the drug company would likely wish to see the new research written up to be more supportive of their initial conclusion that the drug works. Because of their conflict of interests, however, expecting an unbiased appraisal of the research suggesting the drug is actually much less effective than previously stated from those representatives would be unrealistic. For this reason, you probably shouldn’t ask representatives from the drug company to serve as reviewers for the new research, as they’d be assessing both the quality of their own work and the quality of the work of others with factors like ‘money’ and ‘prestige’ on the table.
“It must work, as it’s been successfully making us money; a lot of money”
On an entirely unrelated note, I was the lucky recipient of a few comments about some work of mine concerning inequality aversion: the idea that people dislike inequality per se (or at least when they get the short end of the inequality stick) and are willing to actually punish it. Specifically, I happen to have some data that suggests people do not punish inequality per se: they are much more interested in punishing losses, with inequality only playing a secondary role in – occasionally – increasing the frequency of that punishment. To place this in an easy example, let’s consider TVs. If someone broke into your house and destroyed your TV, you would likely want to see the perpetrator punished, regardless of whether they were richer or poorer than you. Similarly, if someone went out and bought themselves a TV (without having any effect on yours), you wouldn’t really have any urge to punish them at all, whether they were poorer or richer than you. If, however, someone broke into your house and took your TV for themselves, you would likely want to see them punished for their actions. However, if they were actually poorer than you, this might incline you to go after the thief a bit less. This example isn’t perfect, but it basically describes what I found.
Inequality aversion would posit that people show a different pattern of punitive sentiments: that you would want to punish people who end up better off than you, regardless of how they got that way. This means that you’d want to punish the guy who bought the TV for himself if it meant he ended up better off than you, even though he had no effect on your well-being. Alternatively, you wouldn’t be particularly inclined to punish the person who stole/broke your TV either unless they subsequently ended up better off than you. If they were poorer than you to begin with and were still poorer than you after stealing/destroying the TV, you ought not to be particularly interested in seeing them punished.
In case that wasn’t clear, the argument being put forth is that how well you are doing, relative to others ought to be used as an input for punishment decisions to a greater extent – a far greater one – than absolute losses or gains are.
Now there’s a lot to say about that argument. The first thing to say is that, empirically, it is not supported by the data I just mentioned: if people were interested in punishing inequality itself, they ought to be willing to punish that inequality regardless of how it came about: stealing a TV, buying a TV, or breaking a TV should be expected to prompt very similar punishment responses; it’s just that they don’t: punishment is almost entirely absent when people create inequality by benefiting themselves at no cost to others. By contrast, punishment is rather common when costs are inflicted on someone, whether those costs involve taking (where one party benefits while the other suffers) or destruction (where one party suffers a loss at no benefit to anyone else). On those grounds alone we can conclude that something is off about the inequality aversion argument: the theory does not match the data. Thankfully – for me, anyway – there are also many good theoretical justifications for rejecting inequality aversion.
“It’s a great home in a good neighborhood; pay no mind to the foundation”
The next thing to say about the inequality argument is that, in one regard, it is true: relative reproduction rates determine how quickly the genes underlying an adaptation spread – or fail to spread – throughout the population. As resources are not unlimited, a gene that reproduces itself 1.1 times for each time an alternative variant reproduces itself once will eventually replace the other in the population entirely, assuming that the reproductive rates stay constant. It’s not enough for genes to reproduce themselves, then, but for them to reproduce themselves more frequently than competitors if they metaphorically hope to stick around in the population over time. That this much is true might lure people into accepting the rest of the line of reasoning, though to do so would be a mistake for a few reasons.
Notable among this reasons is that “relative reproductive advantage” does not have three modes of “equal”, “better”, or “worse”. Instead, relative advantage is a matter of degree: a gene that reproduces itself twice as frequently as other variants is doing better than a gene that does so with 1.5 times the frequency; a gene that reproduces itself three times as frequently will do better still, and so on. As relative reproductive advantages can be large or small, we ought to expect mechanisms that generate larger relative reproductive advantages to be favored over those which generate smaller ones. On that point, it’s worth bearing in bearing in mind that the degree of relative reproductive advantage is an abstract quantity compromised of absolute differences between variants. This is the same point as noting that, even if the average woman in the US has 2.2 children, no woman actually has two-tenths of a child laying around; they only come in whole numbers. That means, of course, that evolution (metaphorically) must care about absolute advantages to precisely the same degree it cares about relative ones, as maximizing a relative reproductive rate is the same thing as maximizing an absolute reproductive rate.
The question remains, however, as to what kind of cognitive adaptations would arise from that state of affairs. On the one hand, we might expect adaptations that primarily monitor one’s own state of affairs and makes decisions based on those calculations. For instance, if a male with two mates has an option to pursue a third and the expected fitness benefits of doing so outweigh the expected costs, then the male in question would likely pursue the opportunity. On the other hand, we might follow the inequality aversion line of thought and say that the primary driver of the decision to pursue this additional mate should be how well the male in question is doing, relative to his competitors. If most (or should it be all?) of his competitors currently have fewer than two mates, then the cognitive mechanisms underlying his decision should generate a “don’t pursue” output, even if the expected fitness costs are smaller than the benefits. It’s hard to imagine how this latter strategy is expected to do better (much less far better) than the former, especially in light of the fact that calculating how everyone else is doing is more costly and prone to errors than calculating how you are doing. It’s similarly hard to imagine how the latter strategy would do better if the state of the world changes: after all, just because someone is not currently doing as well as you, it does not mean they won’t eventually be. If you miss an opportunity to be doing better today, you may end up being relatively disadvantaged in the long run.
“I do see her more than the guy she’s cheating on me with, so I’ll let it slide…”
I’m having a hard time seeing how a mechanism that operates on an expected fitness cost/benefit analysis would get out-competed by a more cognitively-demanding strategy that either ignores such a cost/benefit strategy or takes it and adds something irrelevant into the calculations (e.g.,” get that extra benefit, but only so long as other people are currently doing better than you)”. As I mentioned initially, the data shows the absolute cost/benefit pattern predominates: people do not punish others primarily on the basis of whether they’re doing better than them or not; they primarily punish on the basis of whether they experienced losses. Nevertheless inequality does play a secondary role – sometimes – in the decision regarding whether to punish someone for taking from you. I happen to think I have an explanation as to why that’s the case but, as I’ve also been informed by another helpful comment (which might or might not be related to the first one), speculating about such things is a bit on the taboo side and should be avoided. Unless one is speculating that inequality, and not losses, primarily drives punishment, that is.