Cheating done right

Most of us grow up learning that cheating is bad. We should not steal. We should not lie. And above all: we should not exploit the groups we are part of, whether this is our family, our circle of friends, or even humanity itself. The implication is that cooperation and trust would be driven to extinction if such behavior were to spread – leading the downfall of not only the group, but also the individual cheats themselves [Kant sends his regards].

As many economists have pointed out, there are measures, largely through institutional design, that human populations can take to prevent a cultural spread of cheating. But what can non-human groups of organisms do against the spread of cheaters: that is, individuals that —unlike cooperators within the group— do not contribute to the collective good? Indeed, how is it possible at all that multicellular organisms such as ourselves evolved, if we are nothing but a highly integrated and cooperative group of individual cells? It turns out that the answer may force us to see cheats not as the doom, but rather as the savior of cooperation!

While economists and philosophers have often wondered how cooperation among humans is possible (see Veit 2019), biologists were primarily concerned with a different question: if cheaters benefit from their actions at the expense of others and thus will always do better than cooperators, then how is it possible that cooperation can not only evolve but also persist for an extended period of time. Shouldn’t evolution eventually eliminate all cooperators in a population in favor of cheats? The traditional answer here has been similar to that of economics – to find mechanisms or explore the conditions that aid the persistence of cooperators (see for instance Michod 1999). In much of traditional work in biology, cheats were seen as an unquestionable evil. Economic tools such as evolutionary game theory were used, often successfully, in biology (see Veit 2019). But through this process biologists inherited the agential explanations common in economics – making them blind to the potentially positive roles of ‘cheats’.

Evolutionary biologist Paul Rainey, in a number of influential publications with colleagues (2003, 2007, 2010, 2014), argued that this view is too narrow – that indeed, it misrepresents the importance of conflict within evolution. Through an exciting experiment on a possible mechanism for the evolution of multicellularity, that was eventually published in Nature (see Hammerschmidt et al. 2014), they argued that cheats —while initially undermining cooperation— could eventually restore cooperation and, indeed, even strengthen it within the group.

They argue that the problem for the evolution of multicellularity is that cooperating cells are often bound together. They cannot disperse when cheats arrive. Cheats, therefore, take over the population (due to their higher replication rate) and eventually come to dominate the population —leading to the eventual downfall of all.

Genuine multicellular organisms, however, are more than just a group of cooperating cells. They are integrated in special ways, such that they depend on each other, show division of labor, and have special cells (germ cells) that are responsible for the reproduction of the individual.

Unlike cooperators, cheats in ancestral populations, due to their lack of cell-glue production (i.e. the common good) can plausibly detach themselves from others. If they are able to produce daughter cells that stick together and cooperate, either by a genetic switch or by a mutation in the daughter cells, then a new group organism can be formed. This is a possible key that solves multiple problems for biologists at the same time: cooperation of individual cells and a mechanism of reproduction of the entire group. We get a life cycle and a new level of selection at which selection can occur. Prior to Paul Rainey, however, no one had come up with the simple hypothesis that ‘cheats’ could play the role of propagules, i.e. function as the reproductive system of the group. Why is that?

The answer lies in what some consider a pernicious way of thinking, common in biology. As already mentioned, biologists were eager to apply the tools of economics to problems in biology. Was this a mistake that eventually led to the introduction of agential thinking (Godfrey-Smith 2009) into biology? I would argue that this gets the causation backwards. Humans love agential explanations and have become addicted to narratives involving agents, goals, and purposes (see Rosenberg 2011 & Veit 2018). We constantly describe the behavior of our fellow humans as goal-oriented, and speculate about others’ hidden beliefs and desires. This theory of mind, we are all equipped with and use in our daily lives, has been called ‘folk psychology’ and criticized by many as scientifically misguided. Despite being used by humans for thousands of years, there seems to be little if any improvement in how we predict and explain the behavior of others within the last hundreds of years (in ordinary life).

“How is it possible at all that multicellular organisms such as ourselves evolved, if we are nothing but a highly integrated and cooperative group of individual cells?”

What justification do we have in biology for treating organisms, genes, and nature as agents? Biologists often argue that this agential way of thinking is merely metaphorical. It helps us to come up with new hypothesis and guide our thinking. Indeed, I have defended this way of thinking myself in a joint publication with Daniel Dennett (famous for his defense of what he calls ‘the intentional stance’), and other ‘Dennettian’ philosophers (Veit et al. forthcoming). But recognition of a useful role for such a mode of thinking cannot be an unlimited ‘get out of jail’ card. As Rainey’s work shows, this way of thinking can not only go wrong, but it can also stop us from considering alternative hypotheses, something that the heuristic defense of agential thinking is all about. Instead, we should embrace a model pluralism (see Veit 2019) that helps us to generate new hypotheses, without taking the model or heuristic ALL too seriously. By embracing a variety of different models, we can leave it to the empirical results from experiments to decide where our theoretical investigations should take us next.

In the end, however, we may have to choose between accepting that ‘cheats’ need not always be bad, or give up on the intentional language altogether.

If this raised your interest, you might be interested in the paper this brief entry in Biomusings is based on: “Evolution of multicellularity: cheating done right” (Veit 2019).