Tag Archives: canine

Dog Behavior: Beneath the Veneer of “Man’s Best Friend.”

by Kenneth W. Krause.

Kenneth W. Krause is a contributing editor and “Science Watch” columnist for the Skeptical Inquirer.  Formerly a contributing editor and books columnist for the Humanist, Kenneth contributes frequently to Skeptic as well. He can be contracted at krausekc@msn.com.

We love our dogs—often more than many fellow humans. In Homer’s 8th century BCE Odyssey, Odysseus referred to the domestic dog as a “noble hound,” and it may have been Frederick II, King of Prussia, who in 1789 first characterized Canis lupus familiaris as “man’s best friend.”  More recently, Emily Dickenson judged that dogs are “better than human beings” because they “know but do not tell.”  Dogs are capable creatures, certainly, but are they as intelligent and considerate as most of humans apparently believe?  Regardless of breed, can their levels of consciousness truly support qualities like nobility, loyalty, and friendship?

Ethologists attempt to assess animal behavior objectively by emphasizing its biological foundations. Classic ethology was founded on the notion that animals are driven by intrinsic motor-patterns, or species-specific, stereotyped products of natural selection (Lorenz 1982).  Modern practitioners, however, often introduce additional factors into the ethological equation.  Many suggest, for example, that intrinsic motor-patterns can be accommodated to developmental and environmental influences.  Some argue as well that complex and otherwise confusing behaviors can emerge from interactions between two or more simpler behavioral rules.

Evolving Motor-Patterns.

Fellow ethologists, biologist Raymond Coppinger and cognitive scientist Mark Feinstein argue that much, if not all, dog behavior can be explained by reference to these three ideas—without resorting to more romantic notions of consciousness or sentience, let alone loyalty and friendship (2015). In the crucial context of foraging, for instance, intrinsic motor-patterns manifest similarly in all canids.

When born, both dogs and wolves spontaneously demonstrate a characteristic mammalian neonatal foraging sequence: ORIENT (toward mom) > LOCOMOTION (to mom) > ATTACHMENT (to her teat) > FOREFOOT-TREAD (stimulating lactation) > SUCK. Here, pups’ mouths and digestive systems are well-adapted to challenges imposed by the foraging environment—that is, mom.  But despite the cozy evolutionary relationship between dogs and wolves, puppyhood is the point after which precise foraging parallels end.

As adults, canid predators display the following generalized foraging sequence: ORIENT > EYE (body still, gaze fixed, head lowered) > STALK (slowly forward, head lowered) > CHASE (full speed) > GRAB-BITE (disabling the prey) > KILL-BITE > DISSECT > CONSUME. But some species, and some individual dogs and wolves, might substitute one motor-pattern for another.  Tending to hunt small prey, coyotes might occasionally swap the FOREFOOT-STAB pattern for the CHASE pattern, and HEADSHAKE for KILL-BITE. Big cats like the puma, by contrast, might substitute FOREFOOT-SLAP for GRAB-BITE to bring larger prey down from behind.

The coyote "forefoot-stab" rule

The coyote forefoot-stab rule

The precise form of GRAB- or KILL-BITE can vary between species as well, often based on the predator’s evolved anatomy. The puma usually kills with a bite to the neck, crushing its prey’s trachea, or to the muzzle, suffocating the victim.  But the wolf often GRAB-BITES the prey’s hind legs, shredding its arteries and slowly bleeding it to death. Puma and wolf anatomies—jaw structure, dentition, and musculature, in particular—apply different mechanical forces, and thus demand the evolution of at least slightly different foraging behaviors.

Domestic dogs, on the other hand, tend to be far less purposeful.   Having long-relied on humans for sustenance, they rarely demonstrate complete predatory foraging patterns.  Instead, different breeds have retained programs for different partial sequences.  Border collies, for instance, are famous for obeying the EYE pattern. I, in fact, once owned an Akita that employed FOREFOOT-STAB with astonishing expertise to capture mice foraging deep beneath the snow.

Border collie "eye" rule.

The Border collie eye rule.

Nevertheless, say Coppinger and Feinstein, “certain commonalities have long persisted in the predatory motor-pattern sequences of all carnivores, reflecting their shared ancestry” and “an intrinsic ‘wired-in’ program of rules.” Learning is neither necessary nor optional.  Indeed, as soon as their foraging sequences are interrupted for whatever reasons, some wild-types are rendered incapable of continued pursuit.  Pumas can’t consume an animal that’s already dead, for example, and, although wolves generally can enter their foraging sequences at any point, they often can’t perform GRAB-BITE if interfered with following expression of the partial EYE > STALK > CHASE sequence.

Dogs have similar limitations. Coppinger and Feinstein recall two conversations with different sheep ranchers.  The first shepherd commended his livestock-guarding dog for independently standing watch over a sick ewe for days without consuming it.  The second, however, complained because his guarding dog ate a lamb that tore itself open on a barbed-wire fence.  To the ethologists, these seemingly inconsistent behaviors were anything but.  Livestock-guarding dogs generally do not express the DISSECT motor-pattern. In fact, their only foraging rule is CONSUME. The first dog wasn’t a “good” dog, necessarily—it was just an unlucky dog.  And the second animal wasn’t really a “bad” dog—it simply performed its intrinsic program when afforded the opportunity to do so.

Accommodating Environment.

However, that many motor-patterns are stereotyped and non-modifiable does not imply that dogs and other animals are mere automata driven solely by internal programs. Certain behaviors can arise as well from an accommodation of the intrinsic to the contingencies of external forces.  Under the “right” circumstances, in other words, animals commonly act in species- or breed-specific manners.  But when exposed to other environmental conditions, their behaviors can look very different (Twitty 1966).

Indeed, if animals encounter such conditions during a developmentally “critical” or “sensitive” period—that is, a species-specific, time-bound stage of growth—they might never display certain typical behaviors. For example, many prospective service dogs flunk out merely because they can’t negotiate stairs, curbs, or even sewer grates.  Why not?  According to Coppinger and Feinstein, their vision systems never fully developed because they were raised in kennels that were sterile and spacious, but nevertheless lacking in three-dimensional structures.

Research also suggests that canids have critical periods for social bonding, during which exposure to a given stimulus will reduce the animal’s fear of that stimulus in the future (Scott and Fuller 1998). Some argue further that certain conspicuous behavioral differences between canid species can be explained, at least in part, by distinct onsets and offsets of these periods (Lord 2013).  For instance, the sensitive bonding period for dogs begins at about four weeks and ends at about eight weeks, while the same period begins and ends two weeks earlier for wolves.

Crucially, dogs and wolves develop their sensory abilities at about the same time—sight and hearing at six weeks, smell much earlier. As such, wolves have only their sense of smell to rely on during their sensitive bonding period.  One general result is that more stimuli, including humans, will remain unfamiliar and thus frightening to them as adults.  But dogs can suffer similar consequences when raised in the absence of direct human contact.

With bonding periods in mind, Coppinger and Feinstein invite us to guess why the Maremma guarding dogs they studied in Italy would abandon their human shepherds to trail their flocks. Were they merely obeying their evolved, gene-based intrinsic motor-patterns—or perhaps their training?  Did they actually understand the importance of their job?  None of the above, say the authors: guarding dog behavior can be “explained by accommodation to particular environmental factors during a critical period in the development of socialization.”

Maremma guarding-dogs.

Maremma guarding dog pups.

During his famous, Nobel Prize-winning experiments in 1935, Konrad Lorenz was able to transfer the social allegiance of newly-hatched greylag geese from their mothers to not only Lorenz himself, but to inanimate objects including a water faucet as well. Coppinger and Feinstein produced similar effects with their Maremmas.  When raised with sheep instead of humans, the dogs usually stuck with the sheep.  Interestingly, however, a few Maremmas preferred to remain at home when both the sheep and their shepherds left for the fields.  These dogs had actually bonded with milk cans that no doubt smelled very much like the sheep.

Emerging Complexities.

Yet other canine behaviors cannot be easily explained by reference to either intrinsic motor-patterns or their accommodations to environmental influences. Consider the collaborative hunting of large prey in wolves, for example.  Here, individuals within the pack appear to work closely together according to preconceived plan, synchronizing their movements and relative positions to prevent prey escape.  At first glance, the spectacle tempts us to infer not only extraordinary intelligence, but insight as well.

Coppinger and Feinstein, however, suggest an explanation relying not on naïve anthropomorphisms, but rather on our knowledge of canine behavior plus the intriguing concept of emergence.  Nothing new under the intellectual sun, emergence proposes that complex and novel phenomena can arise from the accidental, “self-organizing” interaction of far simpler rules and processes.

Two classic examples illustrate the principle. Known for their towering, complicated, yet surprisingly well-ventilated structures, termite mounds obviously are not designed and erected by hyper-intelligent insects.  Similarly, many species of migrating birds, Canada geese in particular, tend to fly in a conspicuous V-pattern, but not because individual geese possess advanced senses of aesthetics.  The more likely explanation, say the authors, is that members of each species act only according to very simple rules.  Termites transport sand grains to a central location.  Perhaps their movements are sensitive to humidity levels and the relative concentration of gasses.  Geese evolved to fly long distances and to draft behind others to lessen their burdens.  The impressive results were never planned; they simply emerged from the interaction of much humbler, species-specific rules.

The practice of collaborative hunting among wolves is no different, according to Coppinger. He and his colleagues recently created a computational model including digital agents representing both pack and prey (Muro et al. 2011).  When they imposed three basic rules on individual predators—move toward the prey, maintain a safe distance, and move away from other wolves—the model produced a successful pattern of prey capture appearing remarkably similar to the real thing.  But in no way were Coppinger’s results dependent on agent purpose, intelligence, or cooperation.

Collaborative hunting among wolves.

Collaborative hunting among wolves.

Consider dog “play” as well. One popular explanation of play generally is that it evolved as a “practice” motor-pattern to prepare animals for escape.  But play only partially resembles any given adult motor-pattern sequence.  And to be even minimally effective, escape has to be performed correctly the first time, which is precisely why motor-patterns are intrinsic, stereotyped, and automatic—as Coppinger and Feinstein observe, “no practice is ever required.”

As any dog owner will attest, canine play is commonly manifested in the “play bow”—a posture in which the animal halts, lowers its head, raises its rear-end, and stretches its front legs forward. Many have interpreted the bow as a purposeful invitation to engage in play.  But more careful observations and experiments suggest otherwise.  Play bows often result when dogs and wolves enter into an EYE > STALK motor-pattern sequence only to be interrupted by their subjects’ failures to react—that is, to run. As such, the bow might actually reveal a combination of two conflicting rules: stalk and retreat.  If so, the posture itself is neither an adaptive motor-pattern nor a signal of intent.  Rather, say Coppinger and Feinstein, “it is an emergent effect of a dog (or wolf) simultaneously displaying two motor-pattern components when it is in multiple or conflicting states.”

None of which should diminish the love and allegiance we typically bestow upon our dogs. That they have no desire to please us—indeed, that their conscious goals are severely limited in general—is no reason to deny ourselves the great pleasure we so often derive from their company.  Even so, our failure to pursue a more objective understanding of dog behavior frequently results in disaster, for both ourselves and our pets.  For some, ignorance might be bliss.  But it’s never a solution for anyone or any thing.


Coppinger, R. and M. Feinstein. 2015. How Dogs Work. Chicago: University of Chicago Press.

Lord, K.A. 2013. A comparison of the sensory development of wolves and dogs. Ethology 119:110-120.

Lorenz, K. 1982. The Foundations of Ethology: The Principle Ideas and Discoveries in Animal Behavior. NY: Simon and Schuster.

Muro, C., R. Escobedo, L. Spector, et al. 2011. Wolf-pack hunting strategies emerge from simple rules in computational simulations. Behavioral Processes 88: 192-197.

Scott, J.P. and J.L. Fuller. 1998. Genetics and the Social Behavior of the Dog. Chicago: University of Chicago Press.

Twitty, V. 1966. Of Scientists and Salamanders. San Fransisco: W.H. Freeman.