randombio.com | science commentary
Sunday, June 10, 2018

Bees do not really understand zero

Scientists in Australia have claimed that honey bees understand the concept of zero.

A nother study has turned up[1] purporting to show that animals have a concept of number. This time it's bees.

The authors say bees have been shown to be able to count up to four. So they trained bees to fly to whichever image had the lower number of objects. When the researchers gave the bees a choice between an image of one object vs. a blank image, they chose the blank one 63% of the time. The authors claim this means the bees understand the “concept of zero.”

A bee, concept­ualizing flowers
FFT of Circles
FFT of Circles

But this is not necessarily so. Animals don't make decisions based on raw images; they process them. We can't draw any conclusions about how they're deciding until we understand how the bees were processing the image. For example, the bees might have been extracting the edges, identifying the center point of the image, or doing any number of other things. Or they could have responded to the busyness, or should I say bzzziness, of the image. A bee's task is to navigate in a three-dimensional space. Maybe it sees the patterns in terms of contiguous navigable space. We have no way of knowing.

At right are four images similar to the ones in the paper. The total brightness of the image is the same except for the blank one. A simple count of the total pixels on feature edges would support the conclusion that the bees are deciding simply on the image complexity. They could also be processing the images by spatial frequency (as shown in the Fourier transforms in the lower image), which is to say they looked at the pattern as a whole and flew toward the simplest one.

Here's the thing: counting is a function of language. Just because bees can tell the difference between patterns of different degrees of complexity doesn't mean they're counting them. To conclude that bees have a “concept of zero” just because they flew toward the less complex pattern when trained to do so is an extravagant overinterpretation of the data. The fact is that we cannot decide from a behavioral test whether they are using concepts at all.

1. S.R. Howard et al. Numerical ordering of zero in honey bees. Science. Vol. 360, June 8, 2018, p. 1124. doi:10.1126/science.aar4975.

jun 10 2018, 3:22 am

Related Articles

Weird science news: Can chickens really do arithmetic?
Maybe, but they can't do calculus. Reason: they are chicken.

Birds Discover Fire
A new article reports that birds deliberately set fires to trap prey

Fake Science
Can bread do associative Pavlovian learning? How's about trace conditioning?

On the Internet, no one can tell whether you're a dolphin or a porpoise

book reviews