This is a repost of something I wrote in 2011, with a new data point added.
⁂
Original 2011 post:
In “Transposition,” a sermon delivered during World War II and published in 1949 in Transposition and Other Addresses, C. S. Lewis refers to dogs’ inability to understand pointing.
You will have noticed that dogs cannot understand pointing. You point to a bit of food on the floor; the dog, instead of looking at the floor, sniffs at your finger. A finger is a finger to him, and that is all.
If you’ve ever owned a dog, you will no doubt find this a rather extraordinary thing to say. Dogs obviously understand pointing, even without any training, and it’s quite common to train dogs to respond to pointing as a command (for example, pointing to a doorway to tell the dog to go into the room indicated). No dog I’ve ever met would waste time sniffing my finger when I’d just pointed out a bit of food it could eat. Cats, yes, but certainly not dogs.
However, Lewis had already had no fewer than six dogs by the time “Transposition” was published (details here), so it’s hard to dismiss what he says about them. This isn’t Pliny the Elder we’re dealing with, reporting hearsay about animals he’d had no personal contact with. Lewis knew dogs well and must surely have known from direct experience how they respond to pointing.
Is it possible that Lewis was right, and that dogs have changed in the half-century since he wrote?
We know that dogs’ ability to understand pointing is a relatively recent evolutionary development. According to dog expert Stanley Coren (as quoted in a 2009 Bloomberg article), domestic dogs understand pointing but their wild conspecifics do not.
“Suppose I point at something — the dog recognizes that I’m indicating something in that direction and looks,” Coren said, referring to a 2004 experiment carried out by Harvard anthropologist Brian Hare, which focused on the increase in dog IQ from domestication. “They do this even if they’re eight to ten weeks old, whereas a wolf, reared since puppyhood in a human environment, would look at my hand,” explained Coren.
Is it possible that the change Coren alludes to could have happened within living memory, sometime after the Second World War? It would be interesting to comb old books for references to dogs’ understanding or not understanding pointing and try to infer when the change took place.
I suppose it’s also possible that geography is a factor. Perhaps the North American dogs studied by Hare and Coren have abilities which English dogs do not. (Iain McGilchrist, a Scot, also refers to dogs’ ability to understand pointing, but he seems to be drawing on the same American research as Coren, not on his own experience.) Most of my own experience with dogs has been in America, but I often see stray Taiwan Tugous (a local breed far removed from anything in Europe or America) and should be able to test their responsiveness to pointing.
If you have any direct experience with dogs and pointing, or if you know of any references to it in books, please leave a comment.
⁂
2023 addendum:
This is from p. 36 of The Hidden Springs: An Enquiry into Extra-sensory Perception (1961) by Renée Haynes:
Humans . . . may also observe that no domestic animal can understand the human gesture of the pointing finger; cats and dogs alike may sniff or lick that finger, but will never follow the line it indicates towards a bone or dish of milk.
"Cats and dogs alike"! As if this were not one of the most conspicuous cognitive differences between the two species!
I have not been able to discover how much first-hand experience Renée Haynes may have had with dogs, so it is not clear whether she speaks from her own knowledge or merely passes on received opinion. At any rate, to say that dogs do not understand pointing was apparently considered uncontroversial as recently as 1961 -- just 43 years before it was experimentally demonstrated that they do. It seems fantastic that canine nature could have changed in such a short time. But, supposing it did not change, it also seems fantastic that such a basic misconception about such a very familiar animal could have persisted for so long -- and in England, of all places, a country Haynes calls a "Dog's Paradise where Cerberus himself would be fed with vitaminized biscuits."
15 comments:
Maybe some breeds understand pointing more readily than others.
Interesting. I'd always assumed it was breed and intelligence related. I've had shepherd mixes who understood pointing, they were littermates and one got it before the other did, so the second may have learned it from the first rather than understood it innately.
I now have a pit/pointer mix and he doesn't understand pointing, I guess that he is more pit dominated, but I don't have experience with either breed.
As for books, I wonder what Where the Red Fern Grows (1961) has to say on this subject, or Old Yeller (1956).
I found pdf's online and did a quick word search for "point".
No reference in Old Yeller based on that.
In WTRFG the boy points and the dogs respond, the judge and his dad remark on it being odd that redbone hounds understand him point, the judge says it's like the dogs can read the boy's mind. He does attribute it to intelligence, says that if they were collies or some other breed he would expect that behavior.
Rhere's a synch for you, 1961 shows up twice.
A few more dates, Rawls was born in 1913, and collies have been around since the 1700's according to the internet.
If one pointed to a vitaminized biscuit on the floor, would all three of Cerberus' heads look, or only one?
I was under the impression that dogs couldn't understand pointing. Would disturb me a bit if a dog understood my pointing.
ben, have you ever owned dogs?
Yep. I tried it just now with a border collie / kelpie cross. Didn't really seem to work but hard to tell.
My English Springer Spaniel does not understand human pointing, but then I've never tried to train her for it. I was under the impression that dogs just can't understand it. My childhood dogs didn't, and my husband says the same about his childhood dog. My ESS is a fast learner, so I'll see if I can teach her. We also have an obedience class tomorrow, so I'll ask the instructor about it.
On the other hand, she has always had a strong pointing instinct with zero training in it. Ever since we brought her home as a pup she has pointed at birds in flight, and toward the sound of gunshots (we live in the country and hike in the nearby wilderness, so we hear gunshots all the time). So on some level she understands the concept...just not when people do it.
Ended up with quite a few cats during the lockdowns. Some cats understand pointing and some don't. One took a long time to train to stop looking at my hand and look at the food I was pointing to. Pretty much had to train it by touching the food and then it figured out what pointing was.
I should clarify that I don't have a Border Collie crossed with a mythological water kelpie, but with an Australian Kelpie dog.
So the data point stands.
My bernedoodle sometimes gets it, but it's not consistent.
For my sibling's Japanese Spitz, it apparently depends on the subject of the pointing.
If memory serves, my previous cat usually just licked my finger at most. The new cat, meanwhile, is liable to attack my entire hand, though that's possible almost regardless of what I'm doing. (During his adoption, I was told his previous home was... chaotic, so he likely wasn't treated well. I'm hoping to train him to stop seeing my hands as chew toys.)
Thanks for the feedback, everyone. Apparently the ability to understand pointing varies a lot among different breeds and individuals.
Ben Pratt, this was the first I'd ever heard of that particular sort of dog. "Labradoodle" I get, but "Bernedoodle"? Where does the first "d" come from? Shouldn't it be "Bernesoodle" or "Bernoodle" or something? Or possible Bamdoodle, reading BMD as a Hebrew-style acronym?
One confound in this question of dogs responding to human pointing is their response to eye contact, and tendency to follow the human gaze. This 2017 paper [0] reports on the latter (incidentally beginning with "Dogs are renowned for being skilful at using human-given communicative cues such as pointing").
This 2013 paper [1] (paywalled) includes the excerpt:
> More than a decade ago, initial studies revealed that dogs are capable of using various forms of human communication to a degree that many other non-human species cannot (e.g., Hare et al., 1998, Kaminski et al., 2004a, Miklósi et al., 1998, Soproni et al., 2001). The vast majority of these studies used the so-called object choice paradigm (see Anderson, Sallaberry, & Barbier, 1995). In this task, a human experimenter hides food under one of several containers out of the dog's view
Hare et al. (1998) [2] is a small n technical report, and addresses the gaze confound:
> In the first study both dogs were able to follow human pointing reliably to one of several locations where food was hidden, both in front of them and behind them. They also showed some skills at following human gaze direction in this same task, when both head and eyes indicated the same location. They did not follow eye direction when it conflicted with head direction.
This 2001 paper [3] further investigates pointing, gazing, head orientation, and mixtures thereof:
> On the basis of a study by D. J. Povinelli, D. T. Bierschwale, and C. G. Cech (1999), the performance of family dogs (Canis familiaris) was examined in a 2-way food choice task in which 4 types of directional cues were given by the experimenter: pointing and gazing, head-nodding ("at target"), head turning above the correct container ("above target"), and glancing only ("eyes only"). The results showed that the performance of the dogs resembled more closely that of the children in D. J. Povinelli et al.'s study, in contrast to the chimpanzees' performance in the same study. It seems that dogs, like children, interpret the test situation as being a form of communication
This 1999 paper [4] finds above-chance significance primarily in the multiple cue case:
> Ten domestic dogs (Canis familiaris) of different breeds and ages were exposed to 2 different social cues indicating the location of hidden food, each provided by both a human informant and a conspecific informant (for a total of 4 different social cues). For the local enhancement cue, the informant approached the location where food was hidden and then stayed beside it. For the gaze and point cue, the informant stood equidistant between 2 hiding locations and bodily oriented and gazed toward the 1 in which food was hidden (the human informant also pointed). Eight of the 10 subjects, including the one 6-month-old juvenile, were above chance with 2 or more cues.
This same paper discusses mid-90s research on non-human primates, which is mixed with regards to gaze versus pointing fluency:
> For example, Itakura (1996) had a human approach nonhuman primate individuals and either look or look and manually point to the right or left. Many of the 11 different nonhuman primate species tested visually tracked the pointing gesture, but only 1 (an orangutan) visually followed gaze direction by itself.
Continued:
I would encourage testing pointing in the context of gaze and head orientation, and see if the results differ. Also, in general, I would strongly expect social communication to be breed and training dependent, and would not be surprised if the urban Lewis had simply never encountered dogs attuned to non-verbal communication, which is currently, and presumably has been throughout the life of the breed, a staple of herding dog training. See this 1981 BBC report [5] especially, in which a deaf sheepdog is taught an ad-hoc command sign language based on what is described as "waving", but which in the short video appears to include an element of pointing.
Miscellaneous references to modern canine hand gesture training:
[6] http://www.bordercollierescue.org/advice/content/unicommands.html (specifically mentions pointing based commands)
> In BCR all verbal commands used are accompanied by hand signals which serve to emphasise and compliment the instruction. This is of great benefit when training deaf dogs, but is also useful if a hearing dog is some distance away or conditions are such that the sound of verbal commands is drowned out by other background noises. There is no mystery or weird science involved in training hand signals. Most dog owners will probably have unknowingly trained their dog to respond to sign simply because they use particular gestures or facial expressions with certain commands.
Even over some distance, a dog is able to differentiate small variations in signals, such as if the finger is extended or the palm of the hand facing up or down, the arm fully extended or the elbow bent.
Border Collies are known to be very intelligent, so it is important to be very consistent in the use of any command, verbal or sign. Even a small deviation may cause the dog to think it is getting a different command from the one intended.
[7] https://www.psychologytoday.com/us/blog/canine-corner/201606/are-voice-commands-or-hand-signals-more-effective-dogs
> "It is important for you to remember that in actual competition you can only give one command for each action. That can be either a voice command or a hand signal, but you can't give both."
[8] https://www.akc.org/expert-advice/training/understanding-your-dog-can-make-you-a-better-trainer/
> Hand signals are easy to teach, particularly if you use lure-and-reward training. For example, the common hand signal for sit, raising your hand from your side to be parallel with the floor, comes from holding a treat to your dog’s nose and lifting it over his head to lure a sit position. Although it is best to train using the hand signal first, even if you have already taught a verbal cue, you can add a hand signal later by giving the hand signal before you say the verbal cue.
[0] https://royalsocietypublishing.org/doi/10.1098/rsos.170349
[1] https://www.sciencedirect.com/science/article/abs/pii/S0023969013000325
[2] https://typeset.io/papers/communication-of-food-location-between-human-and-dog-canis-t0nkjmmzgn
[3] https://typeset.io/papers/comprehension-of-human-communicative-signs-in-pet-dogs-canis-4tv9bktddk
[4] https://typeset.io/papers/domestic-dogs-canis-familiaris-use-human-and-conspecific-1kijba9ewv
[5] https://www.facebook.com/watch/?v=725548181874458
Post a Comment