About the SA Blog Network

The Thoughtful Animal

The Thoughtful Animal

Exploring the evolution and architecture of the mind
The Thoughtful Animal Home

Dogs, But Not Wolves, Use Humans As Tools

The views expressed are those of the author and are not necessarily those of Scientific American.

Email   PrintPrint

Sometime between fifteen and thirty thousand years ago, probably in the Middle East, the long, protracted process of domestication began to alter the genetic code of the wolf, eventually leaving us with the animals we know and love as domestic dogs. While there are several different theories as to exactly how dog domestication began, what is clear is that there were some wolves who were less fearful of humans than others. Over time, those wolves were incorporated into early human settlements. Perhaps humans and early dogs learned to hunt cooperatively – both species hunt primarily by outrunning their prey – or perhaps early dogs instead learned that they could avoid hunting by scavenging on the leftovers of human hunting parties. Whatever the initial reason for the incorporation of wolves into human society, there their descendents still remain.

By sharing an environment with humans, dogs left behind their ancestral environment and found a place in a new one. No longer would they have to hunt to eat; humans would come to provide for their care and feeding. It is probably no accident that the relationship between dogs and their owners mirrors the attachment relationship between parents and their children, behaviorally and physiologically. Indeed, humans who have strong bonds with their dogs have higher levels of oxytocin in their urine than those with weaker bonds.

But it isn’t only the source of their food that changed as wolves became dogs; their entire social ecology changed. Instead of sharing social space primarily with other wolves, dogs came to treat humans as social partners. This is one of the critical differences between a domesticate and a wild animal that is simply habituated to the presence of humans. Domestication is a genetic process; habituation is an experiential one. Domestication alters nature, habituation is nurture.

Several years ago, scientists at Eotvos University in Budapest wanted to determine whether the social-cognitive differences among dogs and wolves was primarily genetic or experiential. To do this, they hand-raised a group of dog puppies and a group of wolf pups from birth, resulting in roughly equivalent experiences. Any differences between the two groups’ social cognitive skills, then, would be attributable to genetics.

Wolf and dog pups were raised by humans starting four to six days after birth, before their eyes had fully opened. For the first months of their lives, the wolf and dog pups were in close contact with their human foster parents nearly twenty-four hours per day. They lived in the homes of their caregivers and slept with them at night. They were bottle-fed, and starting on the fourth or fifth week of life, hand fed with solid food. Their human caregivers carried them in a pouch so that the wolf pups and dog puppies could participate in as much of their daily activities as possible: traveling on public transportation, attending classes, visiting friends, and so on. Each of the pups had extensive experience meeting unfamiliar humans, and at least twice a week, they were socialized with each other as well as with unfamiliar adult dogs. The guiding principle for the hand-rearing paradigm, according to the researchers, was based not upon competition or aggressive interactions, but “to behave rather like a mother than a dominant conspecific.”

Would wolves, having been raised by humans, demonstrate social-cognitive skills that approached the sophistication of dogs? Or is social-cognitive aptitude encoded in dogs’ genes, a direct result of domestication?

In one simple task, a plate of food was presented to the wolf pups (at 9 weeks) or to the dog puppies (both at 5 weeks and at 9 weeks). However, the food was inaccessible to the animals; human help would be required to access it. The trick to getting the food was simple: all the animals had to do was make eye contact with the experimenter, and he or she would reward the dog with the food from the plate. Initially, all the animals attempted in vain to reach the food. However, by the second minute of testing, dogs began to look towards the humans. This increased over time and by the fourth minute there was a statistical difference. Dogs were more likely to initiate eye contact with the human experimenter than the wolves were. This is no small feat; initiating eye contact with the experimenter requires that the animal refocus its attention from the food to the human. Not only did the wolf pups not spontaneously initiate eye contact with the human experimenter, but they also failed to learn that eye contact was the key to solving their problem.

A second experiment, conducted when the wolves and puppies were between four and eleven months old, found similar results. Each animal was presented, in different testing sessions, with two different types of tasks. First, each of the wolves and dogs was trained to retrieve a food reward by opening a bin (in one task) or pulling a rope (in the second task). Then, after they had mastered the task, they were presented with an impossible version of the same problem. After attempting to retrieve the food, the dogs looked back towards the human caregivers. The wolves did no such thing. Dogs spontaneously initiated a communicative interaction with the humans earlier, and maintained it for longer periods of time, than did the human-reared wolves, who all but ignored their human caregivers.

How much time passed before the animals would look back towards their human caregivers?

Both dogs and wolves were equally adept at learning the two tasks, indicating that there were no group differences in terms of motivation or physical abilities, but large differences emerged when given impossible problems to solve. In both impossible tasks, as well as in the earlier eye contact experiment, dogs instinctively shifted their attention away from the food and towards the humans. Despite the fact that they had been fully socialized, the wolves treated each of the situations as physical problems rather than social ones. Only rarely did they ever attempt to engage in a communicative problem-solving interaction with a human. It’s not that wolves are unintelligent; it’s quite the opposite, in fact. Wolves are cooperative hunters, skilled at negotiating within their own social networks. It’s just that even after being raised by humans, wolves simply do not see humans as potential social partners. The dogs, however, quite rapidly took a social approach to solving each problem they were given. In one sense, this is a remarkable example of tool use. Only in this case, the humans were the tools, and the dogs the tool-users.

ResearchBlogging.orgGácsi M, Gyori B, Miklósi A, Virányi Z, Kubinyi E, Topál J, & Csányi V (2005). Species-specific differences and similarities in the behavior of hand-raised dog and wolf pups in social situations with humans. Developmental psychobiology, 47 (2), 111-22 PMID: 16136572

Miklósi A, Kubinyi E, Topál J, Gácsi M, Virányi Z, & Csányi V (2003). A simple reason for a big difference: wolves do not look back at humans, but dogs do. Current biology : CB, 13 (9), 763-6 PMID: 12725735

Might Pleistocene Fido Have Been A Fox?
Biological Evidence That Dog is Man’s Best Friend
Did Dogs Gain Their Social Intelligence By Accident?
Dingoes Ate My Nametag: Tool Use in a Dingo
Man’s new best friend? A forgotten Russian experiment in fox domestication

Photos: Grey wolf via Wikimedia Commons/Domestic dog copyright the author.

Jason G. Goldman About the Author: Dr. Jason G. Goldman received his Ph.D. in Developmental Psychology at the University of Southern California, where he studied the evolutionary and developmental origins of the mind in humans and non-human animals. Jason is also an editor at ScienceSeeker and Editor of Open Lab 2010. He lives in Los Angeles, CA. Follow on . Follow on Twitter @jgold85.

The views expressed are those of the author and are not necessarily those of Scientific American.

Rights & Permissions

Comments 29 Comments

Add Comment
  1. 1. TTLG 10:37 am 04/30/2012

    I think that what this shows is not that wolves will not cooperate with humans, but that they do not look to humans for help: perhaps they do not consider humans their superior in abilities or intellect so do not think that humans can solve any problem that they themselves cannot.

    Link to this
  2. 2. bertrand_ducharme 11:00 am 04/30/2012

    The study is not really convincing: «Wolf and dog pups were raised by humans starting four to six days after birth». Ask a human mother and she will tell you that four to six days after birth, there is already a profound relation between her and her child. The child have been exposed to all sorts of stimuli coming from her and from the environment. This study doesn’t really separate genetic factors from environmental ones. It says nothing about the environmental factors coming into effect before birth.

    Link to this
  3. 3. K9TrainerTina 11:04 am 04/30/2012

    I’m not at all surprised by the results, but there does appear to be a fundamental flaw in the methodology: “In one simple task, a plate of food was presented to the wolf pups (at 9 weeks) or to the dog puppies (both at 5 weeks and at 9 weeks). However, the food was inaccessible to the animals; human help would be required to access it. The trick to getting the food was simple: all the animals had to do was make eye contact with the experimenter, and he or she would reward the dog with the food from the plate…”.

    Why was only the “dog puppy” group presented with the learning opportunity at the age of 5 weeks?

    Link to this
  4. 4. Jason G. Goldman in reply to Jason G. Goldman 12:21 pm 04/30/2012

    @K9TrainerTina: they were, actually. however, the 5 month old wolf pups mostly just went to sleep after initially failing to get the food, so they were excluded from the analysis.

    Link to this
  5. 5. Jason G. Goldman in reply to Jason G. Goldman 12:23 pm 04/30/2012

    @bertrand_ducharme: you’re right, no study can possibly control for everything. however, there is converging evidence from other research that, at least as far as the effects of domestication are concerned, prenatal environment doesn’t impact on domestication – it’s all about the genetics.

    Link to this
  6. 6. David Ropeik 1:21 pm 04/30/2012

    Two things. First, the dogs may be looking to the leader of the pack to solve a problem they can’t solve by themselves. That is more cautious and less anthropomorphic than your “Using humans for help” language. And your suggestion about the long evolution of wolf to dog ignores the work of Belyaev with silver foxes, in which he selectively bred the most timid/passive foxes and in just a handful of generations the foxes changed in many of the same ways wolves changed to dogs…body shape and proportion, fur color, behavior. this is, as I understand it, the growing belief about wolf to dog evolution, not the much longer story to which you refer.

    Link to this
  7. 7. Jason G. Goldman in reply to Jason G. Goldman 1:32 pm 04/30/2012

    David: if you look carefully, you’ll see that I link to my own posts about Belyaev’s study in this post, way up there in the first paragraph. That dogs were domesticated on the basis of reduced fear and aggression is indeed likely, as I write above. However, the fact that Belyaev saw results so quickly does not mean that the wolf-to-dog transition occurred so quickly too. Belyaev allowed roughly 1% of his foxes each year to mate – the tamest one percent – allowing the process of domestication to occur quite rapidly. Out in the real world, where the selection pressures on those early wolf-dogs weren’t nearly as precise or controlled, it likely took much longer before wolves became the dogs that they are today. Also, regarding your “leader of the pack” hypothesis, I don’t see a reason why this would apply to dogs but not to wolves. And, for what its worth, the “wolf pack” theory of the 1970s that’s been pushed so hard by people like Cesar Milan is generally considered by those in the scientific community to be incorrect.

    Link to this
  8. 8. RedJohn 1:53 pm 04/30/2012

    I’m far from an expert in canine behavioral patterns, so I’ll accept the article as is, with two exceptions.
    First, I do not understand the idea that wolves or dogs and humans hunted by outrunning their prey. From my admittedly limited understanding of the hunting techniques of both species, a human is rather incapable of matching the speed of say, a fleeing deer, and neither is a wolf.
    Both species used pack hunting, and cut off the prey, obviating the need to engage in a extended chase.
    And second, why and how can the author even make any connection between what seems to be a form of symbiotic relationship and any form of tool use?

    Link to this
  9. 9. Jason G. Goldman in reply to Jason G. Goldman 2:04 pm 04/30/2012

    RedJohn: To your first point, you’re exactly right. Humans can’t out-sprint most prey (such as deer), but we excel at distance running. Which means we just have to keep on running until the deer gets tired. Wolves use a roughly similar strategy. (But that doesn’t mean other strategies as you describe aren’t also used).

    To your second point, for every tool out there there is a different definition of what constitutes tool use. That said, I’m using the idea of tool use lightly here, in the sense that dogs attempt to manipulate humans in order to alter the local environment (which is, more or less, one definition of tool use: manipulating an object in order to alter the environment)

    Link to this
  10. 10. RedJohn 2:35 pm 04/30/2012

    Jason: I will accept that, as again, my understanding of the hunting tactics of both species is purely based upon second hand information and I have no expertise in the matter, and I’m unwilling to proclaim myself an instant Google expert.
    I can’t help however but feel that both the title of the article, while attention getting, (obviously,) and the last sentence is rather disingenuous, as it rather extends the idea of tool use rather outside the commonly understood norms.
    I was fascinated with the experiment, and it clarified many questions that I had in regards to the domestication of the canine, from wolf to dog, even as it raised further questions of how this could have possibly come about.
    I am familiar with the experiments of Belyaev mentioned by David, and I agree with your response entirely.
    It wasn’t until I reviewed the article after your response and re-read the replies of others that I realized the you were the author of this piece, and if I may be both so bold and so crass as to be a computer chair critic…Dogs still do not use tools, in anyway shape or form. Ever.
    On that note I have to get out of my chair, my cat requires a can opener.

    Link to this
  11. 11. singing flea 3:21 pm 04/30/2012

    “Dogs still do not use tools, in anyway shape or form. Ever.”

    Not necessarily true. I have neighbor that takes her red setter gardening with her. It will carry a basket to put the fruit vegetables in and carry it back to the house for her. That dog will even put the veggies in the basket if told to do so. That basket is a tool which allows the dog to carry many more potatoes then it could without the basket. The basket is after all a tool invented by humans to carry things in.

    I realized it was trained to do this, but never the less, it is at least one example of a dog using a tool.

    Link to this
  12. 12. RCWhitmyer 4:08 pm 04/30/2012

    Redman have you read ? They didn’t use a can opener but it’s a start. Often I felt like I’ve been used a tool by my GSD when she wants an itch scratched.

    Link to this
  13. 13. JamesDavis 5:48 pm 04/30/2012

    Here is something else the author did not mention: The wolves that realized that the closer they stayed to humans, the more food they would get, and the wolves that the humans trusted more got more food. The wolves realized that if they wanted more food, they would have to make some change so the humans would trust them more and not be afraid of them or hurt them. In less than 50 years, the wolves who wanted the humans to trust them more, lowered their ears and rounded their eyes, plus changed their personalities to be more passive like that of a human child. Now, you cannot help from loving a baby puppy. A dog and an elephant is the only two animals in nature that has droopy ears. To look at just a dog’s eye ball beside a human eye ball, it is difficult to tell the difference and most humans can’t. These more trusting wolfs, more or less, turned themselves into human children, and they have had humans wrapped around their little paw ever since, and I love them.

    Link to this
  14. 14. Kestrel 6:22 pm 04/30/2012

    I was under the impression that if a human made eye
    contact with a wolf, the wolf would view the contact
    as a hostile act. I’ve read accounts by those who
    are supposed to know about these things that if you
    see a wolf, coyote or even a pit bull that looks
    threatening,don’t make eye contact with it.

    Maybe this had a part in the wolf pups not making
    eye contact with the humans in the trial.

    Link to this
  15. 15. David_H 7:35 pm 04/30/2012

    Thank you for the very interesting article and also for your interaction here. “Humans can’t out-sprint most prey (such as deer), but we excel at distance running. Which means we just have to keep on running until the deer gets tired.” I’m not sure how my greater endurance helps me if within seconds the deer is long gone and out of sight :-) . Or perhaps I should say “my greater endurance if I didn’t like pizza and root beer so much” :D .

    Link to this
  16. 16. RedJohn 7:50 pm 04/30/2012

    Mr. Davis:
    You really have no understanding of the very article you are commenting upon do you?
    I really did appreciate the effort to equate dog’s to elephants,via the droopy ears idea, in fact I was just considering how often elephants ran with human hunters.
    It really is a shame how most humans cannot differentiate between a dog’s eye and a human eye, but I bet there are really smart humans with machines who can figure out how to do so without the shame.
    The last paragraph demonstrates the most abysmal ability to comprehend the written word that I’ve ever, ever seen.
    However, I’m glad you love ‘them’.

    Link to this
  17. 17. RedJohn 8:02 pm 04/30/2012

    singing flea:
    You utterly failed to see the humor in my post didn’t you?

    Link to this
  18. 18. K9TrainerTina 1:03 am 05/1/2012

    Jason, Thanks for your response to my question. Which validates my opening line. “I’m not at all surprised by the results…”

    Link to this
  19. 19. bertrand_ducharme 1:09 am 05/1/2012

    Once again about the environment: If the mother dog is attentive to humans and reacts to them(via human voice or otherwise) during pregnancy and during the first 6 days after birth and not so much the mother wolf, how can someone pretend that prenatal environment doesn’t impact on domestication? I am curious to see your reference on that.

    Link to this
  20. 20. Riverain 10:44 pm 05/1/2012

    @redman: the thing about attempts at online humor is they often go unrecognized due to lack of eye contact.

    Link to this
  21. 21. Riverain 10:45 pm 05/1/2012

    Duh, i meant redjohn. Not redman.

    Link to this
  22. 22. LMac4 12:22 am 05/2/2012

    I am a bit confused by your comment that “dogs came to treat humans as social partners”. Humans have bred dogs intensively since the practice of line breeding and show dogs became commonplace in the 19th century with the express purpose of increasing the animal’s dependence on humans. Selecting specifically for neoteny means that dogs lack the neural development and independence that wolves have. Dogs did not decide accept human – animal bonds; humans have molded dogs for the express purpose of companionship and dependence.

    Link to this
  23. 23. denke42 3:47 am 05/3/2012

    There is no doubt a profound relation between a human mother and her child by 4 to 6 days after birth, but puppies at that age cannot see, hear, or probably smell. Mainly they move away from cold toward warmth and, if hungry, root when they bump into something nipple-like and suck if they manage to get their mouth around it. They don’t seem to be in a position to draw many inferences from their mother’s relationship to the humans in the area.

    Dogs have evolved for millennia and been bred for centuries to be attentive to and even dependent on humans. Common sense suggests that this is by now at least partly genetic. Of course, common sense is what tells us that the world is flat, so it’s good to have some scientific evidence on the topic. In this case, the evidence supports my common-sense hypothesis. Fascinating article.

    Link to this
  24. 24. bucketofsquid 5:13 pm 05/3/2012

    Humans have used animals as tools for millennia so it isn’t surprising they have learned to do the same back to us. I know I feel like a tool every time my teacup chihuahua demands elevator service up to the top of the recliner.

    Link to this
  25. 25. bertrand_ducharme 1:19 pm 05/4/2012

    In reply to Denke42: You say «puppies at that age cannot see, hear, or probably smell». Even though the majority of scientists used to think that the embryos of vertebrates are not aware of their environment, it has been shown recently by Pr Evan Balaban of McGill University (in “Current Biology”)and his colleagues that unborn chicks can hear loud twittering by measuring their cerebral activity . It would not be surprising if other animal embryos can hear too.

    Link to this
  26. 26. doug 1 1:52 am 05/7/2012

    So, I’m curious as to why they didn’t try the wolves for eye contact at 5 weeks…I don’t so much dispute the findings as wonder the researchers failed to consider that the kinds of social instincts at 5 weeks, particularly in dogs which we already know have been selected for social connectivity, might not linger in wolves since as they mature those kinds of social interactions become less important and pack dominance becomes more important for wolves. Dogs of course are selected to accept their human companions as being higher standing in their social structure, or so it seems. Anyhow, a good look at current thinking in a subject that I find endlessly interesting.

    Link to this
  27. 27. Linhiril 9:46 am 05/8/2012

    Here’s a thought suggested by another comment. In the animal kingdom, eye contact just means two things: I want to fight you or I want to eat you. This vestigial memory is given as the reason we humans are so often nervous when giving speeches. My question would be, How often did the wolves and dogs establish eye contact with each other?

    Link to this
  28. 28. Jason G. Goldman in reply to Jason G. Goldman 1:12 pm 05/8/2012

    @doug: they did, actually, test the wolves at 5 weeks. most of them just went to sleep, so they were excluded from the analysis at that timepoint

    Link to this
  29. 29. Janice Koler-Matznick 6:45 pm 07/6/2012

    I replicated the rope-pull test with New Guinea singing dogs (dingoes) and even extending the time of the trials none of the 5 ever looked to me at all. They would stop trying to pull the rope, walk around and try again without looking toward me (their owner/friend) or my assistant. Singing dogs have no qualms about looking directly into a human face/eyes, and in fact some people are made uncomfortable by their unwavering gaze, which is not in any way connected to aggression/dominance. The Singers are merely curious and learning about the human.

    BTW: wolves that are less fearful of humans are, like most large wild animals, especially predators, the most dangerous ones to humans. The first sign of wolves habituating to humans as potential prey is they stick around to observe the people, over time coming closer and closer. This watchful observing can last weeks or even months. Eventually this escalates into testing by attempting to touch the potential prey (human), and the next stage is biting to see if the prey has any serious defences. Unarmed humans alone – especially children – are basically defenseless, especially easy prey, and even armed with spear or club a lone adult human could never defend itself from more than a single wolf. Only wolves that have learned to fear humans – by being hunted with long distance weapons or by learning by observing older pack members – are not a threat. So, I doubt that Middle Stone Age humans allowed wolves to hang around their camps, let alone invited them to cohabit. The wolf/human facts are from The Wolves of Russia: living with anxiety by Will Graves, reports by Dr. Valerius Giest (wildlife biologist) on Vancounver Island wolves habituated to humans, reports of the child-lifting wolves of India, and other sources. I admire wolves greatly and want them to always be out there in their natural habitat. But pollyanna mythology about wild wolves will only end in tragedy for wolf and human alike. Wolves and humans do not mix well, and I think this was true in the Stone Age too. So where did the dog, the closest genetic relative of the gray wolf, come from? Ahhh – that is the question that needs to be answered.

    Link to this

Add a Comment
You must sign in or register as a member to submit a comment.

More from Scientific American

Email this Article