Ants That Can Count

thisIsOdd writes "NPR had a recent report about scientists at the University of Ulm who suggest that ants in desert environments count to help them get to and from their homes. Because the desert's windiness and sandiness is not conducive the 'smell-trail' method, where ants squeeze certain glands that leave a chemical trail, scientists were puzzled by the fact that these desert ants were able to leave and successfully return to their nest. The theory is called the 'pedometer theory,' and the experiment used to test it involves manipulating the leg length of some of these ants. Ants with longer legs would pass the nest on the way home, and ones with shorter legs came up... well... short."

Comments

This doesn't prove ants can count

[ls671]

If the experience results are valid, there is still a difference between counting and remembering and reproducing a sequence of movements.

Ants might remember that they have to do "step step step step step step step step" to get back to their nest without actually counting. This would seem much more natural to me.

Here is an example applicable to humans: As a drummer, I can create and reproduce the same roll on the fly. But if you asked me how many times I hit the drum pad, only then I would have to count. I did not need to count in order to reproduce the roll nor did I know how many times I actually hit the drum pads.

This leads me to believe ants cannot count, why would they need to. Counting is good for humans in order to trade, so they have developed that capability. Same goes for female animals that could notice one of their puppy is missing. They don't have to "count" them, they only have to remember a picture of all the puppies and notice the picture they now see is different from the normal picture. There is many more examples you can think off where one can appear to count without actually doing so.

Re:This doesn't prove ants can count

[tonycheese]

While you may be right, the examples you gave don't necessarily extend well to the ants. An animal looking for its babies would see less than 10 or 20 and reproducing a roll on your drum would involve varied rhythms or beats or whatever, or probably less than 50 hits if you were to reproduce it after hearing it once. The ants, however, are probably taking hundreds or thousands of steps and remembering the exact distance in one go. I cannot imagine a person hearing a roll go for 750 hits and then reproduce it in the same ballpark without counting time or hits (but I'm no drummer). The article described it as a "pedometer" and I think describing it as counting is perfectly valid - being able to distinguish between 1200 and 1300 steps would involve some form of "counting" in my mind, whether in the brain or by some physical mechanism.

Re:This doesn't prove ants can count

[MartinSchou]

As a drummer, I can create and reproduce the same roll on the fly. But if you asked me how many times I hit the drum pad, only then I would have to count

Been quite a while since I played drums, but I still remember having to learn how to reproduce certain sequences, and that involved counting the number of repetitions.

I can't remember what it's called but the one where you emphasise every third hit (i.e. HIT, hit, hit, HIT, hit, hit etc) came quite fast, and I can still do that one without even trying (including alternating between 2nd, 3rd and 4th). The one that has every fifth hit (HIT, hit, hit, hit, hit, HIT, hit, hit, hit, hit) is one I never got the hang of, and I remember spending a LOT of time trying to burn it into my muscle memory.

But the 3rd one, while easy, still required learning by counting "ONE, two, three, ONE, two, three" for quite a while.

My point is, when you get really really good at something, like drumming, you don't count at a concious level, but in order to get that good, you did need to count.

Re:This doesn't prove ants can count

[discbrain]

If you would have read the article (I didn't read it, but I'm from uulm and am familiar with the corresponding research results) or the referenced research, you would know that all what you stated is obviously taken account for. For example your "they only have to remember a picture of all the puppies and notices the picture"-statement is wrong because of the following experiment: An ant is taken from its current location and moved to another location some meters away. So the ant has no way to tell where it is located at, but it still runs the straight way "home" (though ending not at the ant colony but somewhere else)...

Re:This doesn't prove ants can count

[ebuck]

Math isn't just about a bunch of numbers. Push down automata can count (I know, it's incredible! Even more so considering they have no fingers) The program as I heard it basically described a behaviour which could easily be simulated via a push down automata.

As a drummer, you might be so accustomed to a particular rhythm that you don't count it out in the literal sense of counting out loud, but you do put eight strikes into a measure. Whether you acknowledge that as counting or not, it is still counting, it is just counting that you have become so accustomed to that you don't consider it counting because you need to reframe it in a different context before you can acknowledge to yourself that you are counting.

Rather than using your own logic to falsify the scientist's hypothesis, perhaps you should have listened to the details of the experiment and observed the results. You might have found a superior but alternate explanation, in which case you would have expanded the realm of possibilities a bit. You might even be able to suggest a follow up experiment to differentiate between the counting hypothesis and your alternative to determine which is more correct.

I take it that you haven't done much with functional programming languages, as there are often certain types of problems that are more easily solved in functional languages by counting in the manner of "one one one one one" (as five) than by actually storing a five.

And while I'm at it, trade doesn't require counting. Bartering might involve counting, or it might simply be a swap of my fishtank for your LP collection. However, the idea that we knew counting would be good for trade so we developed counting is a cunning bit of mental gymnastics; it's the mental equivalent of putting the cart before the horse.

Re:This doesn't prove ants can count

[4D6963]

No, you misread, he said that when you move an ant, they act like they haven't been moved. They don't get home, they go where they think home is without taking into account the offset taken when they were moved, which means they're blindly walking back with no regard for environmental clues.

Re:This doesn't prove ants can count

[hazem]

It's still not counting, though it can reproduce the effects. A calculator doesn't actually count (it's just bit switching), but it reproduces the effect. Granted it means that whatever it's doing can simulate the effect of basic counting, but it in no way represents the understanding of numbers

Well, this is science. These researchers had a hypothesis that ants can count and devised an experiment to test the hypothesis. Based on their assumptions, the evidence from the experiments support their hypothesis.

Your hypothesis is that it's not counting but something else. It seems the next step is for you to devise a way to isolate counting from doing a counting-like behavior in ants and do an experiment to test your hypothesis.

However in a way, you're just playing with the definition. What does "understanding of numbers" mean? And is it really integral to counting? If you use pacecounter beads (Ranger beads: http://www.instructables.com/id/Army-Ranger-Beads/), you are "counting" on a piece of string but not actually keeping numbers in your head. In fact, the whole point of those is that you don't have to keep track of numbers because it's hard to do when you're exhausted and have all the other soldier-things to keep track of. You could use these beads to go out some distance, turn around and come back the same distance. You wouldn't have to use numbers in your head, but counting is still being done.

Re:This doesn't prove ants can count

[TropicalCoder]

I don't the ants are consciously thinking one... two ... three in their minds as they walk

These six-legged insects count in base 6, of course.

Re:This doesn't prove ants can count

[mdwh2]

I admit I'm not a drummer, but I have played other instruments - surely if we're talking about an entire song rather than one bar, the person still has to count lines/bars (e.g., this bit happens 4 times, before going onto the next bit, and these two sections alternate two times)? This would be especially true if the drummer was playing on their own, without being able to rely on listening to the music.

Ants can count. The reason it sounds uncomfortable is because it might imply a comparison to how humans count - we do it using our sentient mind. I doubt that this is the case for ants. But even if it's done by some automatic mechanism, I don't think "counting" is unreasonable (I mean, we say that computers can count, even if it's just following an automatic process that a human set up).

I felt a pang...

[crioca]

couldn't help imagining what it would be like for one of the ants that had it's legs cut off, was made to walk home across the desert on it's stumps and then was totally bewildered as to where it's home had gone. I know they're just ants, but damn that's sad.

Re:I felt a pang...

[MarkRose]

I'll, too, admit that the thought of having my limbs chopped makes me a bit antsy...

Re:I felt a pang...

[DreamsAreOkToo]

Ditto.. something about humans and the way they treat life with utter contempt...

I could make the same argument about humans. "...something about humans and the way they treat all life as sacred..."

There are very few creatures on the planet that actually react with sadness when something is killed (other than their very close kin)

that accounts for distance...

[gandhi_2]

but not direction. Like doing "dead reckoning" with pace but no azimuth.

Just had to do it.

[Psaakyrn]

So, it's like changing the tires of a car to a larger or smaller one then miscounting the distance traveled based on rotations?

How many steps before register overflow?

[dirkdodgers]

I can think of a number of follow-on experiments to tell us more about this mechanic.

First I think you'd want to establish more conclusively that it is counting or memory of steps or actions, and not something in the environment:
- Replace the sand behind them on their path and see whether they can still get back.
- Put them on a treadmill to get to their location and back so that their aren't actually moving relative to the earth and see whether they still get back.
- Once this get to the food, rotate the artificial section of ground it is on 180 degrees and see whether they still get back.
- Change the wind direction in an artificial environment and see whether they can still get back.
- Reverse the location of the primary light source in an artificial environment and see whether they can still get back.

Then explore the limits of the counting or action memory mechanism:
- Keep extending the number of steps to get to food until they can't remember how many steps to get back.
- Keep extending the number of steps in a path with a turn in it, on each side of the turn, and compare to the path with no turn.

This is oooold news

[TubeSteak]

Science 30 June 2006:
The Ant Odometer: Stepping on Stilts and Stumps
http://www.sciencemag.org/cgi/content/abstract/312/5782/1965

And here's the original /. story from 2006
http://science.slashdot.org/article.pl?sid=06/06/30/006245

Re:This is oooold news

[thePig]

Thank you.
The comments in the old /. article are worth a read.
Esp.
http://science.slashdot.org/comments.pl?sid=189951&cid=15635693
http://science.slashdot.org/comments.pl?sid=189951&cid=15634139
http://science.slashdot.org/comments.pl?sid=189951&cid=15634257

Makes an interesting read. Also, good to have a comparison between the average quality of comments from 06 and 09 in /.

Re:I heard one of the ants in the experiment speak

[MarkRose]

That's got to be awful confounding... I bet the poor thing is stumped!

Sounds like frogs ...

[dltaylor]

To determine how what proportion each leg contributed to a frog's jumping distance, a scientist trained a frog to jump on command. He then measured the distance with all legs, and remeasured after successively removing one leg at a time.

His conclusion: that since the frog, with all legs removed, did not jump after hearing the command, that the frog was now deaf.

Re:god, that name!

[Antarius]

You're just being PedAntic

Re:Experimental set-up raises a few questions

[ebuck]

Well, that might be a good explanation for the reason the short legged ants failed to arrive home. However, it doesn't explain why the artificially leg lengthened ants overshot their nest. I mean, if it were you or me, we would have seen our home and stopped, so the ants must really heavily rely on step counting.

True masters of the Earth

[criptic08]

If ants have mastered abstract thinking we're all in deep trouble.

avoid the "pom" word on Net

[Anonymous Coward]

> He could as well go "Pom, pom, pom"

Everyone who saw "P o r n, P o r n, P o r n", please raise your hands?.... Thought so!

Re:god, that name!

[smoker2]

That's why we spell it paedophile, and not fall into the lazy practice of spelling things the way they sound, which often results in conflicting definitions for similar words. And it's not even the same sounding word. Pedometer is pronounced ped-(rhymes with bed)-oh-meter. When you ride a bicycle, do you pedal or do you peedal ? And I think that pedes is latin, hence stampede, impede, millipede etc, whereas pod is greek, giving us arthropod, bipod, tripod, podiatrics. Paedo is greek for child, not feet, and is not confusing even if you switch ancient classical languages.