A Delta Society publication aspiring to set professional standards for dog trainers lists three primary ethical criteria that the authors believe qualify a procedure as humane: “Humane dog trainers use and advocate methods that rely on: eliciting and reinforcing desired behaviors, inhibiting and discouraging unwanted or potentially dangerous behaviors, minimizing the use of aversives while doing either of the above” (). Defining ethical behavior exclusively in terms of technical means irrespective of aims is inherently circular and limited with respect to ethical practices.
Without defining the aims of training, such rather tinny behavioristic criteria as set by the Delta Society are virtually meaningless with respect to humane-practice criteria. Numerous questionable training activities might be construed as ethical and humane practices merely because they are performed in adherence to these sorts of standards. The second criterion violates the dead-dog rule and the third criterion neglects to consider the adverse impact of overly intrusive methods (see Hydran-Protean Side Effects, the Dead-dog Rule, and the LIMA Principle). According to these recommendations, putting a dog in a crate and training by means of a tedious autoshaping program would be considered ethical and humane, even though a human is never involved in the training process. Evaluating the ethical and humane use of means in the absence of appropriate consideration of the ends to which they are applied is logically flawed and produces many absurd implications. For example, dogs trained by the Russians during World War II were desensitized and conditioned with food rewards until they either ran between the treads of an approaching tank or ran alongside it, whereupon an ordinance strapped to their bodies was detonated. Since antitank dogs were trained by “eliciting and reinforcing desired behaviors,” applying the aforementioned criteria suggest that training of antitank dogs was ethical and humane.
The criteria listed for identifying behaviors to be reinforced or “inhibited” are strictly limited to anthropic interests; that is, those canine activities that the trainer finds desirable or undesirable without reference to the dog’s needs and quality-of-life (QOL). Canine adjustment problems cannot be effectively treated without normalizing the social and environmental circumstances producing the objectionable behavior. Merely reinforcing or punishing what one likes and dislikes is not likely to prove very beneficial in the long run.
These concerns emphasize the profound influence that training objectives have on matters pertaining to ethical standards. A ruthless person may use kindness and gifts in order to deceive and gain the confidence of others, with the goal of eventually harming or taking advantage of them in some way. Such manipulation is obviously unethical and can hardly be considered humane. In a poignant way, Bertrand Russell () has stressed the potential risks associated with the formation of faulty expectations based on an assumed uniformity between means and ends:
Domestic animals expect food when they see the person who feeds them. We know that all these rather crude expectations of uniformity are liable to be misleading. The man who has fed the chicken every day throughout its life at last wrings its neck instead, showing that more refined views as to the uniformity of nature would have been useful to the chicken. ()
Russell’s chicken, like the Russian antitank dogs, was duped by a faulty extrapolation from means to ends.
Under natural conditions, animals change their behavior in order to improve their ability to control environmental resources and events that contribute to their survival and well-being. Accordingly, instrumental behavior is strengthened when it succeeds in enhancing an animal’s ability to predict and control the occurrence of some significant event and weakened when it fails. Dog training is based on a contrivance that exploits canine learning to shape and control an artificial repertoire of behaviors patterned toward some training objective that may or may not be in a dog’s best interest. In the context of training, reward and punishment are arbitrarily arranged to occur in accord with the trainer’s plans to render the dog’s behavior more predictable and controllable. The dog’s ability to predict and control these events is preempted by rules established by the trainer in advance; that is, the game is rigged to allow only those changes in the dog’s behavior that suit the trainers purposes.
Just as a dog’s behavior is shaped by its ability to control attractive and aversive events arranged by a trainer, the trainer’s behavior is subject to modification by the relative success or failure of training exchanges to meet objectives. Efforts that successfully increase control over a dog’s behavior are strengthened, whereas efforts that fail to increase control are weakened. As such, practical dog training is inherently reward based insofar as contingencies are arranged to enable dogs to gain control over motivational events by means of signaled actions consistent with the training objective. Only an irrational trainer would set up contingencies that were deliberately arranged to make a dog fail. The most unethical and punishment-based training does not stem from the use or nonuse of aversive motivational stimuli but rather from incompetence. Incompetent trainers may lack the basic know-how and skill to use basic procedures humanely or to organize the contingencies and steps of a training plan. As a result of an inconsistent or incoherent organization of training events, dogs may be unable to control significant attractive and aversive motivational stimuli, thus resulting in persistent punishment, distress, and an adverse overall impact on their ability to cope.
Competent dog training is intrinsically rewarding and bond enhancing, whereas incompetent training is intrinsically punishing and bond degrading. The enhanced control resulting from successful training yields an enhanced capacity for the partnership to produce mutually rewarding exchanges that enable the partners to enjoy each other while adjusting and learning about each other. Conversely, social exchanges that result in a loss of mutual control tend to promote increased punitive interaction and conflict. The interactive conflict stemming from incompetent exchanges contributes to an increased reliance on excessive confinement and isolation, emphasis on passive head and jaw restraint, and needless distress.
The practical ends of dog training have significant welfare implications, especially for dogs trained to perform services that are inherently aversive and require significant compulsion to achieve. One cannot disclose the nature of one’s training objectives to a dog in advance, nor can it appreciate intuitively the significance of its training. To some extent, the dog’s innocence and ignorance with respect to the end task renders the ethical dilemma somewhat less consequential than if the dog was deliberately kept in the dark by deception and omission regarding ultimate aims and purposes. However, to the extent that the dog is an object of care, its innocence about such matters obligates the trainer, as a humane being, to consider thoughtfully the ethical implications of training procedures and the practical purposes for which the dog is trained. Training goals that are ultimately enslaving or harmful to a dog are inherently unethical and cruel, regardless of the means used to attain them.
One hypothetical test that can serve as a general guideline for making decisions about practical use is to ask oneself whether the dog, given a choice, would likely select the career being chosen for it. The notion that a dog might truly love to work is not far-fetched, since selective breeding has produced a staggering array of canine skills integrated with ready-made motivational systems that make their performance intrinsically rewarding. Breeding biologically prepares dogs with drives and functional capacities that are uniquely compatible with certain practical tasks while being inimical to others. Dogs are born with a set of innate threshold values predisposing them toward certain traits and action modes. These predispositions point toward activities that a dog seeks and is gratified to perform. These various preferred activities represent a core focus of competency that a dog is naturally inclined to integrate skills around. Discovering a dog’s core competency and building on those interests and incentives is an important aspect of both cynopraxic and practical dog training. Integrating skills relevant to a dog’s core competency serves to activate power and freedom incentives, which are typically expressed in playful activities of various kinds. The availability of these playful activities can be budgeted and made to overlap with the performance of related practical activities of significant value to the trainer but of no particular significance to the dog, except insofar as resonating with core interests and affording an opportunity to play or the possibility to obtain other rewards. When training working dogs, the competent trainer discovers, stimulates, refines, and sublimates various innate appetites and propensities (drives) conducive to the performance of useful services in accordance with the accepted cultural and functional expectations of the breed.
Just as a competent craftsperson would not use a screwdriver in place of chisel or a wood saw to cut metal pipe, a competent trainer would not train a beagle to hunt grouse or train a Brittany spaniel to hunt rabbit. In essence, the practical trainer’s work is to actualize an innate potential present in the dog for some useful activity. As a result of such training, a convergence human and canine interests dovetails, with the dog obtaining significant gratification and playful stimulation as the result of performing some valuable practical activity, thereby satisfying both the trainers objectives and the dog’s interest in play and companionship. Dogs successful in such work have an increased likelihood of being bred and thereby perpetuating the genes responsible for their responsiveness to training.
The foregoing represents the ideal circumstances surrounding the training of working dogs, whether they are used for military scouting and reconnaissance, drug and explosive detection, search and rescue, hunting, herding, and so forth — dogs work to play. Some training objectives, however, require that dogs perform work for which they are neither biologically prepared nor from which they can obtain play gratification. Training dogs for such purposes requires the implementation of means other than play to shape and render reliable the requisite repertoire of trained behavior needed to perform the service. Typically, such dogs are systematically manipulated by social encouragement together with the presentation and withdrawal of attractive and aversive motivational stimuli in the process of compelling adjustments compatible with the training objective that they would otherwise not perform. These issues raise difficult ethical questions, requiring that one weigh the cost of such training and service to the dog versus the potential social benefits that the service provides for the end client. If the social value of the service is minimal or harmful to society, it is easy to conclude that such training is inappropriate. It is sobering, though, to consider the previously mentioned Russian antitank dogs in the context of a social cost-benefit assessment. From the perspective of the Russian military and a great many Russian people faced with a mortal threat, such an end use of dogs was probably considered a tremendous social good. Indeed, the use of such dogs may have saved Russian lives, but the notion of training dogs to become living bombs is abhorrent, just as making prostitutes of adolescent children for the sake of espionage is abhorrent. In both instances, the violations of innocence and trust make such uses intrinsically inhumane. This example underscores the danger of ethical judgments weighted in the direction of social cost-benefit assessments and social consensus without appropriate consideration having been given to the cost and harm to the dog.
The line becomes even more blurred in the case of some canine services that provide a significant amount of social benefit, such as those performed by assistance and guide dogs, but involve the training of unprepared action sequences. As noted previously, in addition to reward-based training efforts, such practical training activities often require a significant amount of inhibitory training and compulsion. Unlike working dogs in which the end use is compatible with breed-typical appetites and drives, some service applications are intrinsically foreign to dogs (see Serpell et al., 2000). Dogs that perform such services do so as the result of intensive training that gradually shapes a repertoire of behaviors having little intrinsic reward value for the dogs themselves but possessing immense comfort and reward value for persons needing such assistance. Although the close interaction and companionship provided by the client-care-giver provide the dog some compensation, nevertheless such dogs would also benefit from special compensatory stimulation consistent with their breed-typical play and work interests when not performing services, including appropriate opportunities to play and exercise — activities that the client-care-giver may not always be able to provide the dog. Perhaps what is needed is a national volunteer-type organization that would provide assistance to these dogs and their owners in the form of play and other outdoor activities not otherwise available to them.
There is nothing intrinsically incompatible between cynopraxic priorities and work, provided that the training process is competent and consists of bond-enhancing exchanges that improve a dog’s quality-of-life (QOL). These cynopraxic priorities stress that the working dog team is the result of a mutual process of adaptation conducive to belongingness, playfulness, and trust. Even relatively unnatural skills, if trained in a competent manner and gradually brought under the motivational influence of power and freedom incentives, can become the source of significant pleasure for a dog to perform.
Although practical dog training can involve behavioral objectives that are exploitive or inimical to a dog’s well-being, many practical uses of dogs can be achieved in a manner that does not run afoul of cynopraxic bond and life-experience imperatives. Some authors have argued that working dogs, in principle, cannot be effective workers while living in a home as companions, suggesting that selective breeding for traits that make them good working dogs bars them from making good household companions. These sorts of general hypotheses conflict with numerous counterexamples indicating that working dogs can, and do, enjoy the quality-of-life benefits of companionship and still remain effective workers. Finck (), who studied the affect attunement between children and dogs, found that working and hunting dogs were generally highly receptive to a child’s attunement efforts, which is consistent with the enhanced abilities of such dogs to follow human instruction.
Breeding that would so alter behavioral thresholds in ways that preemptively excluded a dog from household companionship would likely also significantly impair the dog’s capacity to perform most types of cooperative work. The notion that canine traits conducive to work are incompatible with domestic life is unfounded and contradicted by common experience, since thousands of police dogs have demonstrated that they can be both good workers and welcome household companions. Many police departments in the United States use dogs that work and then go home with the officer/handler as companions. Police dogs share the risk of a dangerous job and are entrusted to make safe contact with the general public. Why, then, should they be denied the familiar surroundings of the home after their duty shift is over? How might the isolation and confinement of overnight and weekend stays in a kennel help such dogs in their work? Dogs are first and foremost companions who are also helpers, and specialized uses of canine labor that preemptively exclude them from living in a home or deny them the benefits of human companionship so that they can be made into cheap and useful “tools” raise a number of significant welfare concerns.
The purposes and ultimate goals of training obviously affect the ethical assessment of the means. With the broader picture of ends in place, a more sensible discussion of means is possible. First and foremost, procedural means are employed to produce some series of immediate effects toward the realization of an ultimate goal or objective. In dog training, these means consist of a variety of manipulations involving the differential application and removal of attractive and aversive stimuli with the intent of altering a dog’s behavior in some way. The ethical implications of means depend on a number of competing considerations. There is a widely held belief that attractive stimuli are of a more humane and ethical nature than aversive ones. Of course, both people and dogs share a preference for pleasure over pain; however, the ultimate significance of attractive and aversive stimuli depends not only on their momentary hedonic and emotional effects but also to a significant extent on ultimate ends and the aggregate effects that such events have on the human-dog relationship and the dog’s quality-of-life (QOL). The experience of brief discomfort in exchange for the acquisition of long-term safety from pain or injury, together with increased opportunities for pleasure and liberty, unquestionably warrants the limited use of such stimulation, especially if other means are not available for producing a similar effect in a timely manner. However, using only mildly aversive or annoying procedures might be considered inhumane in situations where nothing of lasting benefit is achieved.
Motivational stimuli consist of hedonic (likes and dislikes) and incentive (needs) values. Seeking pleasure, comfort, and safety, on the one hand, and avoiding pain, loss, and risk, on the other, are strong motivational incentives regulating both human and dog behavior. Such events also exert influential excitatory (arousal) and inhibitory (calming) autonomic effects. Training consists of the response-contingent application of both attractive and aversive events in order to change a dog’s behavior in some specific or general way. Applying aversive stimulation excessively, incompetently, or unnecessarily represents prima facie abusive treatment. Although the arbitrary and incompetent use of attractive and pleasurable events does not invite similar criticism, it too can be highly abusive and destructive. Since attractive incentives can be used to deceive or to obtain ends harmful to a dog or fail to obtain beneficial ends, a reliance on nonaversive procedures may be unethical in cases where the training objectives are contrary to the dog’s best interests or when they result in harm to the dog that could have been averted by the use of a more effective but aversive procedure. Conversely, aversive procedures that cause significant discomfort but act as a means to some beneficial end (e.g., prevent the dog from running into the street) may confer a positive moral value to the procedure insofar as it is efficacious and achieves, on the whole, an ultimate good that could not have been achieved by nonaversive efforts alone.
When considered independently of ends, attractive events and aversive events are morally indeterminate, although the former is hedonically preferable to the latter. Consequently, on balance, if some behavioral objective is equally obtainable by attractive or aversive means, then the trainer, with due respect for the dog’s preference for pleasure, is ethically bound to use attractive means rather than aversive ones. On the other hand, if some behavioral objective promising benefit to the dog is obtainable only by aversive means, then the trainer, with due consideration for the dog’s well-being, is ethically obligated to use the least aversive procedure necessary. However, on the whole, the competence and confidence needed to integrate an adaptive coping style depend on a balance of reward derived from the successful control of both attractive and aversive motivational stimuli. To conceive of attractive stimuli as being intrinsically good and beneficial to a dog’s welfare and aversive stimuli as being intrinsically bad and inimical to a dog’s welfare is the brew of fanatics, not the fruit of science or a sincere concern for the dog’s welfare. From the perspective of cynopraxic theory, punishment, especially those efforts that thwart or attempt to suppress a dog’s control efforts by incompetent interference or excessive confinement or restraint, is a far more significant and serious threat to canine welfare than is the balanced use of aversive motivational stimuli in the context of reward-based training.
In practice, the decisions regarding the use of valenced motivational stimuli involve a variety of practical and ethical considerations. At minimum, for attractive and aversive events to be used humanely, the trainer must possess sufficient skill to apply them effectively for the purpose of attaining well-defined target objectives (competency factor) and possess a reasonable idea of the immediate and remote consequences of the events used, including a realistic appreciation of potential side effects. In addition to general training and behavioral knowledge, the trainer must also possess significant experience with dogs of different breeds and temperament types at various ages — factors that strongly influence the selection and appropriateness of training procedures.