The quantum mechanics of the waggle dance.

May 7, 2010 12:35 PM

Mathematician Barbara Shipman speculates that a honey bee's sense of the quantum world could be as important to their perception of the world as sight, sound or smell: "the mathematics implies that bees are doing something with quarks."
posted by jardinier (45 comments total)

Charm and Strange honey are delightful. Bottom is musky and Up is sickly sweet.
posted by Babblesort at 12:39 PM on May 7, 2010 [3 favorites]

So am I.

I just don't know exactly what.
posted by Drasher at 12:41 PM on May 7, 2010 [2 favorites]

This sounds really crank-y to me.

Also, 1997?
posted by empath at 12:42 PM on May 7, 2010 [2 favorites]

"Published online November 1, 1997"

So I am guessing that this was not a breakthrough discovery.
posted by 517 at 12:42 PM on May 7, 2010 [2 favorites]

I don't get it. I'm a big Deep Space Nine fan, and I don't remember seeing a bee on there even once.
posted by Halloween Jack at 12:44 PM on May 7, 2010

re. bottom - yes, more of a Bees-Einstein condensate. I so want a copy of Von Frisch’s "Dance Language and Orientation of Bees" on my shelf.
posted by falcon at 12:45 PM on May 7, 2010 [1 favorite]

It only hurts when I exist
posted by anazgnos at 12:46 PM on May 7, 2010 [2 favorites]

A few million neurons?

A friend of mine did a class project with a 'walking man' 3D model that could move around and do stunts with just a few dozen neurons.
posted by delmoi at 12:46 PM on May 7, 2010

And btw, it's really not physically possible for bees to 'perceive' quarks, whatever that means. Quarks simply do not exist outside of protons and neutrons.
posted by empath at 12:46 PM on May 7, 2010 [1 favorite]

For anyone mystified by Babblesort's comment, those are flavors of quarks.

I've heard the 'bees can sense quantum fields' meme in popular culture everywhere, and when I read this, I wondered if this speculation might have been the source. I've noticed the idea is often suggested as a possible cause of colony collapse disorder usually in a black-helicopter kind of way.
posted by jardinier at 12:50 PM on May 7, 2010

The bees, it appears, are able to triangulate as well as a civil engineer.

Watching the Civil Engineering students try and survey, bees are considerably more adept at triangulation. Perhaps we can find some way to get the students to do "something" with quarks. Maybe they would do awesome dances while they are at it. That would brighten the campus up something fierce!
posted by GenjiandProust at 12:51 PM on May 7, 2010 [3 favorites]

If you draw a line connecting the beehive and the food source, and another line connecting the hive and the spot on the horizon just beneath the sun, the angle formed by the two lines is the same as the angle of the waggling run to the imaginary vertical line. The bees, it appears, are able to triangulate as well as a civil engineer.

That's pretty awesome. The manifolds-and-quantum-mechanics stuff is mind-blowing, though.

I remember hearing some time ago that the ancients used bees as a way of finding water while they were traveling. This article sort of turns that into "the ancients carried water-finding quantum computers."

Also: The Discover article will mess with your head if you are into punctuation.
posted by circular at 12:52 PM on May 7, 2010 [3 favorites]

I don't get it. I'm a big Deep Space Nine fan, and I don't remember seeing a bee on there even once.

What do you mean? Work Bees were common on DS9. And Quark once claimed his bar would be 'busier than an Alvanian bee hive,' so there's that.

Oh Memory Alpha, is there no piece of Star Trek minutiae too trivial for you?

Quarks simply do not exist outside of protons and neutrons.

(Possibly) not so!
posted by jedicus at 12:53 PM on May 7, 2010 [2 favorites]

This article, from 13 years ago, promises that the work will be published "probably sometime next year". As far as I can tell, the work appears nowhere in the literature, and is not included in her online bibliography.

So it didn't work out.
posted by mr_roboto at 12:55 PM on May 7, 2010 [3 favorites]

Here's an interesting take on her work from a few years ago.

I honestly can't tell if the overwhelming aroma of crank is due to Discover's breathlessly shitty writing or Shipman's deeply odd assumptions.
posted by ged at 12:58 PM on May 7, 2010 [1 favorite]

Neat, albeit speculative. I wonder if any progress has been made since 1997. If bees can perceive subatomic particles in six dimensions, that lends credence to the Doctor Who theory of colony collapse.
posted by lholladay at 12:59 PM on May 7, 2010 [2 favorites]

This sounds really crank-y to me.

Metafilter is so predictable sometimes. I actually RTFA, and in the article Ms. Shipman is quick to make explicit that she does not have empirical evidence for what amounts to educated speculation about bee communication. That is, she is upfront about what in her theorizing remains unsubstantiated. But she's also clearly not just some crank.
posted by HP LaserJet P10006 at 12:59 PM on May 7, 2010

The thing about generating bee dances from the flag manifold is interesting; the thing about the quarks is really setting off my bullshit alarm. (Or at least my fanciful-speculation alarm.)

Really, showing that the apparent variety of the bees' dances can in fact be described by a single mathematical entity makes it much more plausible that a little bee brain could produce them. The wild-and-crazy symbol-recombining world of real language would be much more difficult to achieve with a relatively simple physical system than something straightforward like projecting a six-dimensional surface into 2D (after tweaking some variables to account for distance and direction.)

A friend of mine did a class project with a 'walking man' 3D model that could move around and do stunts with just a few dozen neurons.

Ehhh. I'd be careful about drawing conclusions from that. "Neurons" in artificial neuron networks vary significantly in their behavior from real biological neurons. Despite some superficial similarities, they aren't really the same thing except in name -- and small artificial neural networks seem to be much more malleable to modification through evolutionary algorithms than biological neural networks to modification through real evolution. (Large artificial neural networks, on the other hand, get comparatively useless fast. There was one guy who thought he could simulate a cat brain using an artificial network with as many neurons as a cat. Failed spectacularly, of course.)

Unless you mean that he used mouse neurons in a petri dish or something, in which case that is so cool
posted by magnificent frigatebird at 1:05 PM on May 7, 2010 [2 favorites]

Fears for crops as shock figures from America show scale of bee catastrophe
posted by homunculus at 1:05 PM on May 7, 2010 [2 favorites]

however, wasn't it maths that definitely proved the bumblebee could not fly?
posted by infini at 1:18 PM on May 7, 2010

however, wasn't it maths that definitely proved the bumblebee could not fly?

No, it proved that they couldn't fly if they were fixed-wing aircraft. Or possibly even if they were little birds. Being neither, a little further research showed how they flew, since their flying was never in much doubt.

Although bees with tiny propellers is kind of a cool idea.
posted by GenjiandProust at 1:20 PM on May 7, 2010 [2 favorites]

It looks like at one point she produced a manuscript entitled "A mathematical foundation for the dance language of the honeybee". This title is cited here in preprint form. There, a paper titled "Investigating bee behavior from the standpoint of fundamental physical principles" is cited as appearing in something called Am Bee J, which is presumably the American Bee Journal. This, however, is a mis-cite; that paper was written by Shipman's father.

It seems like this work was never published.

It's getting a ton of play in the crackpot corners of the web, however, as one can see by subjecting oneself to a few searches....
posted by mr_roboto at 1:21 PM on May 7, 2010

Wow I think I just stepped in a pile woo.
posted by ExitPursuedByBear at 1:26 PM on May 7, 2010

This sounds really crank-y to me.

Yeah, but science is full of odd speculation, some fruitful some not. Perhaps, someone will look at her speculation one day (or even today) and find something useful or inspiring for further research. She does call it speculation and implication and so does the post, but that kind of imagination plays a vital role in the history of science (in a Haraway-an sense).

can't tell if the overwhelming aroma of crank is due to Discover's breathlessly shitty writing

The article was written by a working PhD in Physics, Adam Frank who also writes for popular lit, he may be a poor writer, but he's a theoretical astrophysicist.
posted by jardinier at 1:44 PM on May 7, 2010 [1 favorite]

I just want to know which flag manifold Shipman is talking about... does anyone know?
posted by ennui.bz at 1:52 PM on May 7, 2010

I'm sure she'd tell you if you contacted her - though she might have a filter on her email the trashes anything about bees... but a real math question like that should get through!
posted by jardinier at 1:57 PM on May 7, 2010

Ms. Shipman is quick to make explicit that she does not have empirical evidence

Should be easy enough to collect. Lots of honeybees dancing in lots of places.
posted by Blazecock Pileon at 2:25 PM on May 7, 2010

But she's also clearly not just some crank.

After reading the article, I would agree, but all kinds of natural phenomena have more than three dimensions to them, and that doesn't make them manifestations of quantum physics, it just means there are many parameters that can generate a certain result. Evolution itself, for example, can be considered a slice through a really large hyperspace of possible outcomes, but that doesn't turn evolution into quantum physics.
posted by Blazecock Pileon at 2:35 PM on May 7, 2010 [2 favorites]

can't tell if the overwhelming aroma of crank is due to Discover's breathlessly shitty writing

If that article constitutes shitty writing, then 99.9% of writing is shitty.
posted by ekroh at 2:39 PM on May 7, 2010

The assumptions beehind this theory go beeyond the empirical constraints of beehavior-based approaches to an extent that is frankly unbeelievable.
posted by dephlogisticated at 2:43 PM on May 7, 2010

What do you mean? Work Bees were common on DS9. And Quark once claimed his bar would be 'busier than an Alvanian bee hive,' so there's that.

NERRRRRRRRRRRRRRRRRRRRRDS!

well-played, sir
posted by Halloween Jack at 2:46 PM on May 7, 2010

The wild-and-crazy symbol-recombining world of real language would be much more difficult to achieve with a relatively simple physical system than something straightforward like projecting a six-dimensional surface into 2D (after tweaking some variables to account for distance and direction.)

Can... not... resist
posted by hal9k at 2:56 PM on May 7, 2010

This does not pass the smell test.
posted by fantabulous timewaster at 3:11 PM on May 7, 2010

so do bees smoke the shit or eat it in brownies?
posted by kitchenrat at 3:14 PM on May 7, 2010

And here comes the science:

http://www.nature.com/nature/journal/v435/n7039/pdf/nature03526.pdf

(sans quanta)
posted by pjenks at 3:53 PM on May 7, 2010

There's more than one kind of crank. One kind lives in a shack in the middle of nowhere and mails 200 page books to academics explaining how intelligence results from the interaction of love particles and energy particles (sadly, not exaggerating). However, there's another kind where an academic who is qualified to do top-notch research in one field chooses to wildly speculate in another completely unrelated field. Shipman is (or was) playing around in two fields she doesn't understand.

Quarks don't couple to large material objects. Period. End of story. This is in part because the nature of quarks is that they ALWAYS are bound to other quarks. But it's mainly because they are tiny, tiny, objects that one only interacts with at crazy high energies.

The other thing is that you'll find no shortage of people describing the computational power of nervous systems. Generally, those people don't study them. This whole line of thinking goes along the lines, "we know bees can't possibly have enough computational power in their tiny little brains, so therefore my crazy theory is worth thinking about". Except that first part is utter bullshit. No one's ever done any calculation deriving the computational power of a real biological nervous system, in part because there's no clear definition of "computational power" when describing something that isn't even really a computer, isn't built like a computer, doesn't act like a computer. It's not even clear what the fundamental units of "information" or "processing" is in real nervous systems. People often SAY the processing units are neurons for simplicity. But it could be subbranches of neurons, glia could be involved (okay, they ARE, but they may do independent processing).
posted by Humanzee at 3:57 PM on May 7, 2010 [6 favorites]

This whole line of thinking goes along the lines, "we know bees can't possibly have enough computational power in their tiny little brains, so therefore my crazy theory is worth thinking about". Except that first part is utter bullshit.

It's utter bullshit even before you get to the first part. Just because the brain is able to direct a certain action doesn't mean that the brain has to "compute" how to perform that action. The brain of a basketball player does not compute force vectors and parabolic trajectories as he's standing at the free throw line; there's something more along the lines of a heuristic going on there, involving muscle memory and vision.
posted by mr_roboto at 4:09 PM on May 7, 2010 [1 favorite]

And unsurprisingly it doesn't pass the read test either. Everything up to

At this point Shipman departs from safely grounded scholarship and enters instead the airy realm of speculation.

is more or less reasonable. Manifolds are interesting as complicated mathematical objects that arise, in general, from fairly simple rules. It's a general feature of mathematics that two problems which are not obviously related may, in fact, be isomorphic to one another. If the primary purpose of bees' communication systems is to locate and collect pollen and nectar, and the internal representation of this information has six variables, then maybe these two six-dimensional problems might overlap. (Why would there be six dimensions? Maybe there are three geometrical variables, fixing the source in three-space, and three variables describing the species, quality, and quantity of the find.) That much of the idea is plausible.

I'm not familiar with the "flag manifold" in the context of quarks. I know that there are some models for the quark-quark interaction that use the special unitary group of dimension six, or SU(6), but those are approximate models that don't describe the behavior or ordinary protons and neutrons. And at any rate there is the problem that ... that ... this is the kind of idea that's not even wrong. I thought maybe I could say something illuminating about just how it's wrong, but I would have to know something about how the author misunderstood quantum coherence to start. Quarks interact in a quantum-mechanically coherent way inside the nucleus, but getting nuclei to interact with each other coherently is hard, and it's just silly to claim that it happens only in bees.

Anyway I think I've also heard this claim before (that "bees sense quantum fields") and I'm glad that I'll be able to rebut it if it comes up again.
posted by fantabulous timewaster at 4:10 PM on May 7, 2010 [1 favorite]

fantabulous timewaster: I thought maybe I could say something illuminating about just how it's wrong, but I would have to know something about how the author misunderstood quantum coherence to start.

Well, again the author is a theoretical astrophysicist, so if you email him, he can probably tell you how he misunderstood quantum coherence using language only the two of you would understand.
posted by jardinier at 4:23 PM on May 7, 2010

this sounds a bit quarky.
posted by rudster at 5:04 PM on May 7, 2010

jardinier: Which author? I thought that Frank was probably representing Shipman's ideas accurately, bogons and all. Anyway I'm sure neither of these folks would be interested in a correspondence where I explain why their idea was wrong, a dozen years after that idea was apparently abandoned.

mr_roboto: is a heuristic not a type of computation?
posted by fantabulous timewaster at 6:46 PM on May 7, 2010

To my way of thinking, it's a type of computation, since I use "computation" to mean a useful transformation of information or stimulus. But many people feel strongly that computation is something computers do: something representable by sequences of logic gates, generally understandable as straightforward equations, i.e. solving basketball trajectories using calculus and physics. The brain doesn't do this, nor do most smaller nervous systems (there are a few that are truly simple, such as some reflex responses to stimuli). I think there's merit to both ways of talking, so I generally don't dispute people who prefer one to the other.

In any case it should be understood that since the brain doesn't act like a computer (even insect brains), it's very difficult to quantify performance or capabilities.
posted by Humanzee at 7:24 PM on May 7, 2010

Well, I don't think anyone thinks that bees are doing calculus, but there are ways to solve complicated geometry problems using relatively simple machinery. It might be analog computing, but its still computing of a kind.
posted by empath at 8:00 PM on May 7, 2010

I don't know a thing about quarks, but I do know a couple things about flag manifolds.

A 'complete' flag manifold is where you start with an n-dimensional space, and then choose an n-1 dimensional space inside of it, and then an n-2 dim. space inside the n-1 dimensional space, and so on until you run out of dimensions. If n is 2, this means choosing a plane, and then a line inside the plane, and then a point inside the line, which looks a little like a flag. The choice is constrained; you should be choosing subspaces, so all should go through the point labelled 0: this means that there is no choice for the starting n-dimensional subspace or the final zero-dimensional space.

In any case, each 'point' in the flag manifold is such a choice of subspaces: it is the manifold of flags in an n-dimensional vector space. This is equivalent to choosing a _basis_ for the vector space, which is in turn equivalent to choosing an n-by-n invertible matrix. The dimension of the n-by-n invertible matrices is n(n-1)/2. Taking n=4, you then get a six-dimensional space. Four is, perhaps unsurprisingly, the number of dimensions in Lorentzian space-time, so if there's going to be a connection to the physical world, it seems like n=4 is a good place to work.

Ok, that's all for now...
posted by kaibutsu at 2:28 AM on May 9, 2010

First: someone (possibly Adam Frank, but possibly the web editor) is so goddamned stupid that they can't even use quotation marks. Stupid, stupid person.

Second: "nmr" is correctly spelled "NMR." Stupid, stupid person (whomever). The fact that this acronyym is correctly capitallized when quoted in ged's link leads me to believe the webstoid is to blame for both errors.

Third: Barbara Shipman wouldn't be the first insightful genius in history to make stunning breakthroughs (bee dances can be described with 2-D projections of flag manifolds, acceleration is indistinguishable from gravity's effects, and orbital hybridization are responsible for chemical bonds, for three examples), but then went on, giddy with their success, to postulate the nuttiest of theories (bees can detect quarks, God wouldn't allow the quantum mechanics, Vitamin C cures cancer and a host of other boogeymen). (Shipman, Einstein, Pauling).
posted by IAmBroom at 6:02 AM on May 11, 2010