Development of Behavior: Evolutionary Background - Normal Dog Behavior

Karen L. Overall, MA, VMD, PhD, DACVB, ABS Certified Applied Animal Behaviorist

To understand canine behaviors and breeds it is critical to remember that, fundamentally and ancestrally, the story of canine behavior is that of working with humans at many tasks. Dogs are not just altered wolves: as part of the co-evolution of our co-domestication they became their own, and a very different species. Listen to them and look at them to know this is true.

Outline of canine social systems:

All behavior has environmental and genetic components. The variation in the genetic component is sufficient to produce a wide array of individual behavioral phenotypes in the absence of any specific breed. Hence, not all domestic long-haired cats or mixed breed dogs look alike or demonstrate identical behaviors in response to like situations. This is true within and between litters. Evolutionary biologists since and including Darwin have recognized these differences. The extent to which behavioral plasticity is a function of genetics is a hotly debated issue in the fields of behavioral ecology and evolution. Some recent experimental evidence on desert toads living in highly variable environments indicates that selection may be operating to maintain developmental and concomitant behavioral plasticity, rather than selecting one or a few modes, each of which would persevere under alternate conditions. These issues are important and relevant when one discusses the issue of breeds.

One function of establishing and maintaining a breed is to canalize some of this overall genetic variation. Although domestic canine breeds have a relative body size that spans two orders of magnitude, such variation in size is absent in wild canids. The process of domestication, alone, relieved many of the pressures for which wild canid body size was a response, allowing the underlying genetic variation to respond to artificial selection. In the process of selecting for certain physical and behavioral traits within any breed, one has also selected for some variation in that trait. Accordingly, when one discusses what have been considered breed specific behaviors, caution is urged. Selection, natural or artificial, cannot act if there is no underlying genetic variation. Some variation (termed additive genetic variance by quantitative geneticists) must be present for a trait to be developed.

This is extremely easy to visualize if one is considering a physical trait such as coat or feather color. It is less easy to see that selecting for a behavioral trait, such as protectiveness, which is really a constellation of behaviors, is going to produce a continuum of protective behaviors, some of which will not be what the selector desired. In fact, some of these behaviors are going to be inappropriate, because they are not complete or forceful enough, and some will be unacceptable because they are too forceful and out of context. Under natural, instead of artificial, selection these behaviors would have been selected against at their extremes; however, it would be an error to regard the wild environment as producing absolute phenotypes.

The demographic and local climatic environments act in concert to determine what scope of the continuum of variation will survive. In a very good year even the most inept hunter might live to reproduce and contribute genes to the next generation. This is the source of the additive genetic variance. That such genetic variance exists is demonstrated by the extent to which artificial selection has developed so many and such varied breeds in a few hundred years, while thousands of years of natural selection has not developed that degree of canine variability, although the initial stock should have been similar.

Hence, if one has developed a breed for certain specific behaviors (rather than overall survivability, as in the wild situation), one should expect that there will be variation around that behavior and that some of this variation will result in inappropriate, out of context behavior. It is in this light that charges made about breed predilections should be viewed.  This means that if one has selected a breed for protectiveness or guarding, some of the individuals in that breed may inappropriately protect or guard against objects which pose no threat. Some believe that herding behavior was developed from the first phases of predatory behavior. It is conceivable that unless selection were extremely discrete, sufficient variation should exist so that the occasional herding animal exhibits inappropriate predatory behavior. This is commonly known by sheep ranchers. Such concepts of genetic variability in the development of behaviors are difficult, but given the amount of misinformation regarding some breeds and prejudice towards others, the issue should be addressed.

Attempts have been made to group breeds of dogs according to certain constellations of behavioral attributes. Although their classifications are based on subjective opinions (which can invoke prejudice and folklore) rather than objective classifications of individual behaviors, and their categories confound discrete behaviors and behavioral diagnoses, some overall patterns of behaviors correlated with breeds are apparent. Cluster analysis grouped animals according to reactivity, "trainability", and aggression. It is no surprise that high trainability characterized most working and guard dogs; prior to selecting for any specific other behavior the ability to work with and be trained by, people would have to be elaborated. Less reliable are the characterizations of aggression and reactivity, since these are both diagnoses and descriptions of amalgam behaviors.

More recent studies have attempted to focus on specific behaviors (growling when disturbed while sleeping; stalking small animals, barking at approaching strange people, et cetera). This is important because there is scant documentation of the frequency, duration, intensity, and pattern of occurrence of the actual behaviors that are involved in behavioral problems. It is only in this context that fair evaluations of breed related behaviors should be made. Furthermore, because so little is known about normal behavior and behavioral precursors of serious problems, early signs are not recognized. A survey of the faculty and senior students at VHUP, and local practitioners revealed that virtually all individuals thought that there were more and less aggressive breeds and could rank these, the ranks of the three survey groups were different within and between groups with no rank being statistically significant, and, with the exception of the majority of the students, few individuals in the other two survey groups could, when provided with a list of discrete behaviors, accurately identify those that were outright aggressive or precursors of future aggressive behavior.

In summary, caution is urged regarding any generalizations about breed based behaviors. It is best to view selection for specific behaviors as a risk assessment analysis: breeds that have been selected for one or a few particular and specific behaviors may be more at risk for developing unsavory variation for those behaviors. This does not mean that dogs selected for protective behaviors are more aggressive than dogs for which this selective pressure was absent. It does mean that that breed may be more at risk for developing a disproportionate number of dogs who exhibit inappropriate, out of context protective aggression. Inherent in this concept is that any dog, regardless of breed, can also exhibit the inappropriate behavior. A further corollary is that dogs who are selected for tenacity and jaw strength in their in-context work (Bull Terriers, Rottweilers, Rhodesian Ridgebacks), will, when they respond inappropriately or out-of-context in another behavioral setting, exhibit this same tenacity. Coupled with the physical traits attendant with such selection (large jaws, heavy musculature), they can and will do large amounts of damage on a first strike. These factors, rather than increased breed-specific aggression, are the cause of the severity of wounds, when inflicted.