Representation, Bee Dances, and Daniel Hutto

Dan Hutto, who does great philosophy of psychology work, recently gave an ENSO seminar titled Beyond Content: Explications, Motivations and Implications. 

This is the ENactive Seminars Online (ENSO), put together by Merek McGann and Mathew Egbert. Dan gives a solid overview of the players working today on whether cognition necessarily entails representation - and what on earth that might mean. For those who don't know, after championing a Wittgensteinian view of folk-psychology theories in the beginning of his career,  Dan has spent the last decade or so as one of the heavy hitters in philosophy arguing that cognition need not entail mental representation. Though he thinks representation has a place in the conversation - post language - he does not think "basic cognition" requires it at all.


Towards the end of the talk, bee dances came into the story, and Dan covered different philosophers attempts to explain what the dances represent to other bees. Because my background is in animal behavior, I could wrap my head around this material quite easily, and so my mind began racing even faster. The superfluous language that others were arguing for brought to mind the chapter Andrew, Sabrina, and I collaborated on, which discussed the need for a better language with which to talk about neuroscience. The bee dance seemed like a very obvious context for revisiting the points made there.

Bee Dance Overview:

Probably you are all at least roughly familiar with the bee dance story. Roughly speaking, when a bee returns from a far away food source that still has a lot left over, it does a "waggle dance". The direction of the waggle represents the angle a bee would need to fly to get to the food source, and the length of the waggle represents the distance to the food source.

"Represents"?!? Did Eric really just say that something represents?!? Yes, it re-presents the crucial variables. And that much is nigh universally agreed upon. The interesting question is where we go from there. The first question should be "represent to whom"? And at least one clear answer is: To a human observer, with Karl von Frisch's discoveries. But it is probably more than that, probably it re-presents those things to the observing bees as well. It re-presents the location of the food in the same way that your driver's license picture re-presents your face. The re-presentation is physical and out there in the world. What I am going to deny is that there is any need to posit "mental representations." There is no need to suppose that there are any cognitive processes going on the bee that warrant discussion of "representational content." This was Dan's point as well.

What bees do:

The better way to talk about the bees would be with a language that doesn't import more than is needed. The first thing we know is that a bee that comes back to the hive after finding food is a different bee than the bee which left. The bee that returns has been reconfigured into the type of bee that wants to dance in a certain way. We also know that some of the bees hanging out in the hive at any given time are the type of bees that want to watch dancing. However, once those bees have seen one or more dances of a particular type, they become the type of bees that want to leave the hive and fly in a certain direction. 

Again, I think pretty much everyone agrees on all of that. That is a reasonable description of the thing-to-be-explained. 

How to explain what bees do:

If we agree on the description, then the crucial question is how to explain the changes the bees undergo. Do we need to posit that after visiting the food, the original forager returns to the hive with a "mental representation" of the food's location? Do we need to posit that after seeing the dance, the subsequent forager leaves the hive with a "mental representation" of the food's location? I agree that it is tempting to anthropomorphize the bees, assuming that what is happening in the hive is akin to one person giving another person verbal directions to a restaurant. I think Dan would agree that "representation" is typically happening when people exchange verbal directions. As such, I speculate, that he might think about this as a question of whether bees are doing something akin to that, as is implied by those who call the dances "bee language." (He doesn't think they are, but he might well agree that is a crucial issue.)

Andrew and Sabrina and I, as well as most ecological psychologists and behaviorists, would likely take a harder line, I think. We would demand that those who claim mental-representation is at play explain how we would examine the system to determine if representations are indeed present.  What experiment would we do to determine if we have a bee-configured-to-fly-in-a-certain-direction vs. a bee that is mentally representing food at a certain location, and therefore takes off towards it? If there is no investigation difference being proposed we would assert, then there is we should drop the added implications of using representational language, because it can only add vacuous confusion.

To those ends, I asked Dan, at the end of his talk, what method we could use to distinguish his position that the bees were "contentless" from the contrasting argument of all-that-plus-representations. His answer was that he couldn't think of any implications that could be investigated. He also pointing out, correctly, that it wasn't his burden to come up with such tests.

Where does that leave us? It leaves two issues that are a recurring theme in my work, over at Notes from Two Scientific Psychologists, and in the work of Dan, Tony Chemero, and many others.

1. We need a better language to talk about the changes taking place within the bee. We need to better explain the bee's changing from one type of system to another. Presumably, in the case of the bees, primarily involve light-weight neural reorganization, but we need systems language, rather that relic dualisitc language to describe what is happening during neural reconfiguration. A bee that wants to fly with the sun just over its right shoulder for a while, need not be representing a flight path, nor a destination. It simply has to be the type of thing that is statistically more likely to go one way rather than another. How can we talk about that better?
2. I need to get a better handle on what Dan (and many others) mean when they say that "mental content" has a role as soon as we get to linguistic systems. Certainly, by a broad definition, the bees are "communicating" but also Dan clearly does not think they are doing "language." What would the bees need to do for Dan to think they were doing language? (Maybe he will give us some hints, or sink some of his illustrious graduate students on us, below.)

A more problematic question: 
The interpretation of bee waggle dances leaves us in a precarious situation regarding representations. As Chemero keenly point out almost a decade ago, there are many situations in which you could talk about representations, but it doesn't add anything to the discussion. That is, the representation-talk, to the extent that it is accurate, simply provides a poor description of a system, which is completely covered and improved upon by an actual illumination of the mechanisms involved. This connects back to Peirce's pragmatist maxim. One of the many ways Peirce tried to explain his maxim was to assert that: Any two ideas that are found to have all the same implications are, despite appearances, only one idea.

There was a time when those who asserted the existence of mental representations were making bold, potentially testable, claims. As the evidence has stacked up against the more adventurous of those claims, proponents of representational theories have back tracked, broadened, and generally softened their definitions of what gets to count as "representation". Is it possible that their assertions have been watered down so much that "representation" and "not-representation" are ultimately indistinguishable positions?