Scientists Discover That Horses Can Use Symbols To Talk To Us (sciencemag.org)

sciencehabit writes from a report via Science Magazine: Scientists have discovered that horses can learn to use another human tool for communicating: pointing to symbols. They join a short list of other species, including some primates, dolphins, and pigeons, with this talent. Scientists taught 23 riding horses of various breeds to look at a display board with three icons, representing wearing or not wearing a blanket. Horses could choose between a "no change" symbol or symbols for "blanket on" or "blanket off." The horses did not touch the symbols randomly, but made their choices based on the weather. If it was wet, cold, and windy, they touched the blanket-on icon; horses that were already wearing a blanket nosed the "no change" image. But when the weather was sunny, the animals touched the blanket-off symbol; those that weren't blanketed pressed the "no change" icon. The study's strong results show that the horses understood the consequences of their choices, say the scientists, who hope that other researchers will use their method to ask horses more questions. The report has been published in Applied Animal Behaviour Science.

Comments

Cart before the horse?

[SpiralBound]

Hmmm... While it has been shown in a variety of ways going back at least as far as the original Mr. Ed that horses are smart and capable of performing a large repetoire of tricks, I do wonder in this particular case if the horses are touching the symbols at the appropriate times because they understand what the symbols mean, or because they were trained to touch the symbols at the appropriate times. It is a vitally important distinction. Just how did they teach the meaning of each symbol to them without instead accidentally training them to perform without any true understanding of the symbols themselves? Humans have instinctual behaviours towards pattern recognition, anthromorphism, self-delusion, and rationalisation, thus experimential methods must be very carefully designed to remove these influences.

Re:Cart before the horse?

[Rei]

Reading over the study, in case you're curious: they discuss the Clever Hans effect, how it manifests (the horse observes the trainer or audience and does what it needs to get a reward based on the audience's reaction). In this study, the trainer was off to the side, out of the field of view of the horse, and they monitored where the horse was looking to ensure it was only looking at the symbols when making its choice. Then the horse was given a reward regardless of what symbol it chose; it simply had to choose one to get a treat. The symbol touches were easy to record and unambiguous, so there was no "interpretation" of the horses behavior in question. The researchers also observed side behaviors. For example, once the horses had learned to use the symbols to control their blanketing status they often became very eager to go into the testing facilitity (before the phase where treats were on offer). It was observed that horses that sought to have their blanket removed in this manner tended to be sweaty underneath it, while those that didn't seek it out weren't.

To be fair, there's even more that could be done. I really liked the controls that Pepperberg did in her studies of Alex (the African Grey parrot). She had it set up so that the person asking a question didn't know the answer, and neither did the person scoring the result. In this case here, clearly the trainer knew the weather and thus what would be the "optimal" blanketing status. But by paying attention to where the horse's focus is, whether it's hesitating, etc, and giving it a reward either way, I agree that that's some pretty good controlling for the Clever Hans effect.

Racist Scientists

[Anonymous Coward]

As a dog, I am highly offended by this article. When I grab my leash and whine, my human knows to let me show him around our territory. How is that not using a human tool to communicate? Or is seeing a leash an instinctive reaction in humans that causes them to need to be led by someone smaller than them? That must be it because the little humans never let me walk them when they see the leash.

Never mind, I guess they're not racist. Humans are just dumber than I realized. I thought they understood me, but apparently they're simply too primitive.

Re:19th century smart vehicles

[kria]

My great grandfather drove a milk wagon. His horses knew all his stops, apparently, and would start off for the next one when he got back in the wagon without any signs from him. I assume that's not as intelligent behavior as what they're testing, but it's still pretty cool. :)