Philosophy Phriday: Ruining Picnics with Epistemology

The Daily Ant hosts a weekly series, Philosophy Phridays, in which real philosophers share their thoughts at the intersection of ants and philosophy. This is the forty-fourth contribution in the series, submitted by Will Fleisher.

Ruining Picnics with Epistemology

Suppose you know that there is a picnic going on somewhere in a nearby park, but you aren’t sure where. You want nothing more than to ruin this picnic, and you have a bunch of friends with you who share your desire. You know that the most efficient way to find the picnic is to spread out and search through the park. So, while you head toward the lake, Anton searches in the woods, Antonia looks by the hill, and Brant heads to the disc golf course. If one of you finds signs of a picnic, you will signal to the others, and some of them will come join the one who found the signs. Once the picnic is discovered, you will all rush in to steal the food.

I will be informing no one reading this blog by telling you that ants are famous for their division of labor. The kind of division of labor I’m interested in here isn’t the kind facilitated by the different castes of ant (queen, worker, soldier, etc). Instead, I’m interested in the kind with the goal of picnic ruining. That is, the kind where ants divide up their exploratory labor when seeking out food, shelter, building materials, and opportunities for picnic sadism.

Ants-Big Final scene-build-FINAL
Ants blessing a picnic. Photo: Greg Hardin

Ants achieve a surprising efficiency at locating their needs using the general strategy of dividing their labor in exploring the local terrain. Discovering the details of how this works, I understand, is an ongoing subject of research. The effectiveness of ants has not gone unnoticed by computer scientists, however, who have devoted a whole sub-field of research to trying to emulate them. That these researchers into artificial intelligence are emulating ant behavior is suggestive of the epistemic power of the social. Individual ants, after all, aren’t particularly intelligent. They aren’t deciding what to do, or where to go, based on carefully reasoning about their evidence. Rather, they are responding to local interactions with their fellows to determine what kind of behavior they should be engaging in, and where they should be going. For instance, some ants track the frequency with which they bump into their fellows to determine whether they should leave the colony to go foraging, or whether they should leave a particular search area for another (for readers who are also myrmecology newbies, I read about this here).

Recently (I mean, relative to the history of philosophy), epistemologists have also become interested in the social aspect of our epistemic lives. “Epistemic” here just means “stuff that has to do with knowledge and inquiry.” Social epistemologists are interested, in particular, in what kinds of epistemic benefits and pitfalls we find when research takes place in the context of a community (and in what kind of benefit there is to our philosophizing when we recognize this fact).

One benefit of communal research is that inquirers can emulate ants in their division of labor. In our case, this is a cognitive division of the labor working toward epistemic goals. For instance, different researchers can focus on exploring diverse questions, and developing the skills necessary to do so. A myrmecologist can focus on learning about insects, and need not be versed in the ins-and-outs of Martian seismology, and this specialization improves our efforts to learn about both ants and marsquakes. This specialization works much like the ant caste system. But a community attempting to answer a single question can also benefit from a cognitive division of labor.

Many castes, many tasks. Photo: Alex Wild

Recall the example I led with, where we supposed that you have become perhaps too enamored with ants, and so have taken up their interest in picnic-wrecking. This toy example illustrates our ant-like division of cognitive labor. Each member of the party desires to learn the same thing: the location of the picnic. They improve their odds of finding it by improving their coverage of the terrain. Similarly, researchers pursue different research strategies, programs, paradigms, and theories, even while trying to answer a common question.

Sometimes, division of labor results from disagreement. You and Antonia might disagree about the best place to look for the picnic, and this explains your different search choices. Similarly, researchers disagree about which theory is the best one to pursue. This leads to a division of their efforts into different research projects, in an attempt to vindicate their own theories.

Consider the case of Homo floresiensis. Paleoanthropologists were divided over whether the fossilized remains of hominins on the Indonesian island of Flores are the remains of a newly discovered species, or whether they are the remains of a population of Homo sapiens which happened to be particularly small (perhaps due to one of several heritable conditions shared by the population, or to insular dwarfism). The debate over the question of whether there really was a Homo floresiensis, while now apparently on the path to being settled in the affirmative, provides a rich case study (De Cruz and De Smedt 2013, Currie 2017) for the benefits of the division of labor that can develop due to scientific disagreement.

Many scientists, pursuing the small Homo Sapiens hypothesis, spent a great deal of time and effort on determining whether there was any heritable condition that occurs in modern humans which could explain the microcephaly (small head size) of the fossils on Flores. Others attempted to take DNA samples from the fossils. Meanwhile, the proponents of the new species hypothesis researched ways that insular dwarfism arises in other animals, in an effort to show that the Flores specimens could not have evolved in this fashion, lending credence to the idea of an earlier speciation event. All of these efforts improved the scientific community’s knowledge, even those that ultimately failed to support their preferred conclusion. Sometimes the failed attempt was itself instructive because it helped rule out versions of a theory. Thus, the community was able to “cover more ground,” to learn more and to settle the question faster, in virtue of the division of labor resulting from the disagreement.

A little H. sapiens creature… but what is an ape doing on this blog?

Ants, presumably, don’t divide their labor in virtue of explicit disputes over the best place to look for picnickers. As previously mentioned, one thing they actually do is pay attention to what their fellows are up to. This is another way that researchers can divide their labor: they can watch what others are doing, and either follow them, or go off in a different direction. Social epistemologists have been using economic models, and computer models, to try to find the best way for researchers to divide up their research in this way (for an overview of this literature, see Muldoon 2013).

One way of doing this closely follows the example of ants: it involves computer models of agents searching an epistemic terrain (e.g., Weisberg and Muldoon 2009, and Thoma 2015). The agents look for the highest point in the terrain, which represents the truth, by paying attention to their changes in height as they move. Sometimes they are allowed to pay attention to the behavior of other agents. Another kind of computer model involves networked agents sharing information about their own research (e.g., Zollman 2010 and Grim et. al. 2015). Moreover, there are economic models arguing that scientific researchers, when provided with the right reward scheme, can be motivated to divide up their efforts, and work on under-explored theories, based on a sensitivity to how many researchers are working on any given project (e.g., Kitcher 1990, Strevens 2003).

What I find most interesting about each of these models is that they involve sensitivity to considerations about what one’s fellows are up to (just like the ants). These considerations, or reasons, don’t seem to have to do with the truth of any particular belief or theory. This is despite the fact that they are all reasons concerning inquiry, a truth-aimed activity. This can be illustrated by appeal to our picnic-wrecking example.

Brant, if you will recall, goes looking for the picnic at the park’s disc golf course. This isn’t a particularly promising place to look. Indeed, you might think it is the least probable place of all the options (who wants to picnic while constantly watching out for flying discs?). Brant, we can suppose, is well aware of this. However, he knows that his picnic-hating co-conspirators will all be heading to the other areas. Because of this, the best way for him to contribute to the search is by looking in the less likely place. Brant going to the disc golf course might not make it much more likely that the conspirators find the picnic in time to ruin it, but it does make it more likely than if Brant just went along with Antonia to the most likely place, on the hill. It would certainly be a bad strategy if the whole group simply went to the hill, rather than checking out the other options.

Brant, much like the ants that inspired him, and like the epistemic agents in our models, is being sensitive to considerations about collective inquiry. Brant is unlikely, by himself, to discover the picnic, and would be more likely to find the picnic by himself if he went to the hill. But the community does better in virtue of his searching another area. Brant’s reasons to search the disc golf course are not reasons to think that he is very likely to find the picnic there. Rather, they are reasons to change his search behavior because of his fellow searchers actions, so that the group can do better.

Ants on a hill. Photo: Alex Wild

Researchers in a variety of fields make similar judgments: they choose to work on new, and sometimes quite unlikely, theories. They might do so because few other people are working on the new theory, or they have figured out a new way of running experiments to investigate the new theory, or perhaps they happen to work where the best equipment for testing the theory resides. Either way, they are being sensitive to what will benefit collective inquiry. And it’s good that they are, since otherwise our inquiry might be stifled, and arrive at premature consensus on a theory which isn’t true. We want our researchers to divide up their labor just like our picnic-searchers, with sensitivity to what others are up to, in order to benefit the group as a whole.

Many epistemologists have thought we can distinguish epistemic reasons from practical (or moral ones). Sometimes we believe things because they make our lives easier or better. For instance (to borrow from L.J. Cohen), a lawyer might accept that her client is innocent, because defending her client is her duty (and because that’s how she pays the bills). This might even happen when the lawyer has excellent evidence to believe her client is guilty. Our lawyer is acting on practical (and ethical) reasons. Epistemic reasons, by contrast, are considerations we have to believe something because we care about getting to the truth. Many philosophers have thought that the only kind of epistemic reason is evidence to think something is true (e.g., the eye-witness testimony and DNA evidence implicating the lawyer’s client).

However, we have seen that sometimes researchers can be motivated like Brant: they want the group to find the truth, but this means personally pursuing the less likely option. They have reason to pursue a new theory based on the fact that it is under-explored. But this lack of exploration isn’t evidence to think the theory is true. I like to call such reasons extrinsic epistemic reasons (if you are interested, more on that here). They are reasons concerning the health of collective inquiry, rather than reasons to think a theory is true.

So, we can learn a great deal about successful social inquiry from ants, and social epistemologists have recently been doing just that. Good inquiry involves being sensitive to what your fellows are up to. Of course, that’s only one way to wreck a picnic with philosophy. You can always fall back on Socrates’ strategy of showing up to the picnic and demanding that the picnic-goers define justice.


Will Fleisher is a Ph.D. candidate at Rutgers University. His interests lie in epistemology and philosophy of mind, as well as decision theory, cognitive science, and philosophy of science. For his paper on “Virtuous Distinctions”, see here. Or if “Rational Endorsement” is your thing, go here. If ants are your thing, stay right here.