Chimps Seem to Know What Others Know—So Do Dogs at Play

Let's revisit play behavior.

Posted Nov 17, 2017

Are chimpanzees and humans the only animals who take into account what others know?

A recent essay in The New York Times by James Gorman called "Chimps Tailor Alarms to What Other Chimps Know" caught my eye not only because it reports on a very interesting and important study, but also because it contains some statements that are few too primatocentric—primate centered—for my liking. This focus on primates ignores what we know about non-primates and can be misleading to people who don't keep up with comparative research on the cognitive lives of animals other than primates. Reading about the research on chimpanzees also reminded me of some research my students and I conducted years ago about the ways in which young dogs, coyotes, wolves, and adult free-running dogs communicate their intentions to play. I'm glad to have been motivated to revisit these data. 

The research essay on which the Times piece focuses was published in Science Advances by Catherine Crockford and her colleagues and is titled "Vocalizing in chimpanzees is influenced by social-cognitive processes." It is available online. Basically, using a set of very creative experiments the researchers learned that wild Ugandan chimpanzees seemed to know what others know about the presence of a model of a dangerous snake and change their alarm calls based on this information. They write, "We conclude that chimpanzees are aware enough of another’s perspective to target information toward ignorant group members, suggesting that the integration of signaling and social cognition systems was already emerging in early hominoid lineages before the advent of more language-specific features, such as syntax." 

In the Times essay we read, "When a chimp saw a realistic model of a snake, the animal would make more sounds—called hoos—and make a greater effort to show where the snake was if it seemed that other chimps in the area were unaware of the danger. If it seemed other chimps already knew about the snake, it would make fewer calls and stay a shorter time at the danger." All well and good, and the researchers' conclusions surely fit in with their data. They carefully consider alternative explanations that invoke less rich cognitive capacities, and I agree with their overall conclusions that chimpanzees surely seem to know what others know—whether there is a dangerous predator around and if they already know this—and change their alarm calls based on this knowledge.

In the Times essay, Dr. Crockford notes that this study "challenges the view that only humans keep track of what others know and change their communication to match." She correctly states, “This experiment shows they are monitoring their audience."  Along these lines, Mr. Gorman writes, "Three scientists testing wild chimpanzees in Uganda reported Wednesday in the journal Science Advances that chimpanzees can do something that previously had only been known in human beings. They change the way they are communicating to take into account what their audience knows."

Fair play and what dogs and other animals likely know about what others know: Fine-tuning behavior on the run

Dr. Crockford and her colleagues' study sets the standard for future research centering on questions of what individuals know about what others know and if and how they change their behavior based on this information. It is admirably detailed and the database is sound and extensive.

When animals display the ability to mind read others, to use what others know and what they feel, they're said to have a Theory of Mind (ToM). Having a ToM refers to "the ability to attribute mental states—beliefs, intents, desires, pretending, knowledge, etc.—to oneself and others and to understand that others have beliefs, desires, intentions, and perspectives that are different from one's own."

Marc Bekoff
Source: Marc Bekoff

In a previous post ("Theory of Mind and Play: Ape Exceptionalism Is Too Narrow,") I argued that there are data that show that nonhumans other than chimpanzees likely have a theory of mind and that studies of social play are a good place to investigate whether this is the case. When I wrote my essay I focused on a piece by Rachel Feldman in The Washington Post called "Can all great apes ‘read minds’ like humans do?" about a research paper titled "Great apes anticipate that other individuals will act according to false beliefs." Revisiting both of these pieces made me reflect once again on the taxonomic distribution of ToM among nonhumans.

In my essay, I argued for the possibility that when animals engage in social play there is every indication that to be able to play fairly and cooperatively they too display a ToM (for more on this please see "How and Why Dogs Play Revisited: Who’s Confused?"). Ample data for a number of different species show there are predictable rules of play, namely, ask first, be honest, follow the rules, and admit when you’re wrong.

As an example, I considered social play between two dogs named Harry and Mary. During play there is a good deal of rapid thinking, feeling, and fine-tuning on the run. So, for example, let's consider the possibility that Harry has thoughts and feelings based on what he thinks and feels Mary is likely to do during an on-going interaction (and vice versa). These sorts of interactions make it clear that play is also a good place to observe and to study ToM because Harry and Mary need to pay very close attention to what each has done and is doing, and how that will influence what she or he is likely to do in the future. There are a number data that show that dogs who are familiar with one another play differently from those who don't know one another.

Play bows punctuate on-going play sequences

One set of data about which I failed to write in the above essay is how young captive dogs, coyotes, and wolves and adult free-running dogs change the ways in which they ask other animals to play using an action called the bow. When dogs and other animals bow they crouch on their forelimbs, raise their hind end, and occasionally wag their tail and bark (please see the image above). This move is used to initiate play and to maintain the play mood. Dogs and other canids punctuate play sequences with bows to reinforce the play mood. All in all, bows essentially are contracts to play that change the meaning of the actions that follow, such as biting and mounting. They also serve to reinitiate play after there’s a pause in the action. Bows also allow dogs to perform a wide variety of different actions as they spring up after bowing.

Play bows are extremely stereotyped (they are what ethologists call a modal action pattern) and they vary in shape and duration depending on where they are performed in a play bout. For coyotes, beagles, and adult dogs, while there were no statistically significant differences in form between bows performed at the beginning of and during play sequences, in all cases bows performed during sequences were slightly more variable. It's surely possible that the animals themselves read these subtle variations. These differences can be due to the need to communicate more clearly that individuals want to play before play begins and it's less important to do so after they begin playing and when they know others want to play. Of course, it also could be that bows performed during a sequence are more variable because the animals choose to use them from a wide variety of different positions. These data are reported in an essay called "Social communication in canids: Evidence for the evolution of a stereotyped mammalian display" published in Science in 1977. 

Revisiting these data made me reflect on how It's the ways in which play bows change in form and duration that indicate that they know what others know—that this is play rather than aggression or an attempt to mate, for example—and I offer that this is another example of animals knowing what others know and then changing their behavior based on this knowledge. The overall conclusion of this study of play bows was that "canid play signals seem to be more important in the initial soliciting of social play and less so for the maintenance of the 'play mood.'" This would indicate that each player knows what others know and want to do and that they change their behavior accordingly. 

When dogs and other animals play, and for each individual to know that their playmate wants to play rather than fight or mate, it seems very reasonable to conclude that they need to know what others are thinking and what their intentions are. Each individual needs to pay close attention to what the other individual has done and is doing and each player likely uses this information to predict what the other is likely to do next.

This form of mind reading might be one reason why play only very rarely escalates into real fighting, although many people vastly overestimate how often this actually occurs. For example, Shyan, Fortune, and King (2003) reported that fewer than 0.5 percent of play fights in dogs developed into conflict, and only half of these were clearly aggressive encounters. Their data agree with observations on wild coyotes and free-running dogs at play.

When these and other data are considered together, evidence is increasingly showing that different animals likely have a theory of mind, and one of the main ways we’ve discerned this is through research on social play behavior. Play sequences in the canids mentioned above are extremely variable and statistically more variable than sequences observed in other contexts.1 Thus, it seems unlikely that play sequences and the order in which individual actions are performed during play are hard-wired in any way that I can imagine.

It seems highly reasonable to postulate that the players are thinking about what is happening when they are playing and alter their behavior "on the run." It's also possible that they're using creative combinations of different actions to maintain fair play based on what they know about what the other players know and want to do, even when they're engaged in frenetic "zoomies" (for more discussion in this topic please see "It's OK For Dogs to Engage in Zoomies and Enjoy FRAPs"). Interestingly, play in larger groups of dogs sometimes breaks down, not due to play escalating into fighting, but rather because it seems that when there are too many dogs they can't read one another accurately.

I look forward to more detailed comparative research on the dynamics of social play in a wide variety of species. 

Humans and other non-primates are not alone in the ToM arena

So, are humans and chimpanzees and possible other nonprimates the only animals who possess a theory of mind and know what others know? It's too early to say, but concluding they are alone in the ToM arena is a bit too fast for me.

Observations are data

While studies of the ways in which play is communicated are based more on detailed analyses of film rather than on experiments in which different variables are manipulated, these sorts of observations are data that need to be taken seriously. I only mention this because when I do interviews or talk with people about my and others' research they'll sometimes ask something like "Are there data to support these conclusions?" I simply say that observations are data and most come to agree that they are. Ethologists know there's no substitute for taking the time to observe what other animals do in a variety of contexts. 

Of course, we can always use more than solely this type of information, but simply watching other animals has generated extremely valuable data and has motivated further detailed studies that bear on the questions at hand. The study about which I'm writing on chimpanzees and how they alter their alarm calls is an excellent example of what needs to be done—it's relevant to how chimpanzees live and is about as non-invasive as it can be so researcher interference isn't a factor in interpreting the data. 

Please stay tuned for more comparative research and new discoveries about what animals know about what other individuals know and possibly feel. It's incredibly important and exciting to learn more about what's happening in the minds of other animals and to learn more about mind reading in different species.

Comparative research on animal minds is helping to put aside long-standing speciesistic myths about human or other animal exceptionalism and replace them with facts about the fascinating and previously unknown to cognitive and emotional capacities of other animals. It wouldn't surprise me to discover that mind reading has evolved in diverse species as a very useful way for facilitating social communication in many different social contexts.


1 Marc Bekoff and John Byers, 1981, "A critical reanalysis of the ontogeny of mammalian social and locomotor play: An ethological hornet's nest." in K. Immelmann, G. W. Barlow, L. Petrinovich, and M. Main, Eds., Behavioral Development, The Bielefeld Interdisciplinary Project. New York, Cambridge University Press, pp. 296-337. I thank Robert Fagen for doing the statistical analyses.