Not Today and Not for My Sheepdogs
In our desire to make lives more comfortable and fulfilling for the dogs who live among us, and for ourselves as we live alongside them, we often turn to scientific inquiry to inform our understanding of them. Here, we lean on paradigms as though they provide definitive answers that we couldn’t find through observation, enquiry, and study of our own dogs. While science indubidably improves our knowledge and understanding of the world, we must also recognise that when it comes to the dynamism and individuality of the dogs with whom we live, it will necessarily be limited in terms of what it can offer: the foundational studies on principles of extinction, counterconditioning, etc., took place in highly controlled contexts with animals who have been bred into those settings and therefore who have no experience of rewards. We may find that these principles and effects that are set out in these studies don’t transfer so easily into the real world. (Article: Science Doesn’t Have All The Answers)
Further, these foundational studies took place in a culture in which a dog’s compliance and obedience was valued, with an emphasis on the behaviours produced rather than the emotions underlying them. This can skew our perception towards viewing our dogs’ learning as a matter of input-output, which further diverts our focus away from how our dogs learn as opposed to what they’re learning. And to further complicate matters, many of these protocols and procedures are antithetical to the way our sheepdogs have been selected to understand the world, and they therefore can often provoke confusion or even frustration in them. In this blog post, I’ll take a brief look at three common protocols (or sets of protocols) employed putatively to “resolve” common issues that arise for sheepdogs in a domestic environment, and explain why they may not be the optimal solution for these dogs.
Extinction is a protocol that seeks to diminish a previously reinforced behaviour by withholding the reinforcer. For example, if a dog jumps up, and if that jumping up has previously been reinforced by social approval (touch, talking, etc.), then the recommendation is that withholding that social approval will cause the behaviour of jumping up to extinguish. Extinction is commonly used with sheepdogs whose attempts to control the movement of their person has appeared “endearing” when they were younger, but is now becoming a nuisance at best and dangerous at worst; in this case, the withholding of the reward is the frustration of a need for that dog, and would be much better managed by the careful construction of alternative and compatible rewards.
Let’s think about this from the dog’s perspective: the dog has learned that a stimulus that has previously predicted a reward no longer predicts that reward. This certainly happens in nature, but when we are part of that picture—when we’ve consistently been the source of the reward—this may result in distrust of the learning process that they undertake alongside us. For sheepdogs who have been selected for cooperative work, and who are acutely responsive to human attentional cues (Kaminski and Nitzschner, 2013), who must be sensitive to patterns in the environment and observe and respond to changes, it is not beyond the realms of probability that such protocols are aversive.
Extinction procedures are incredibly difficult to apply and have lots of nasty side-effects – so much so that they are often described with ample cautions for both behaviorual and emotional consequences. In fact, there are even some studies to suggest that expecting a reward and not receiving it (or wanting to escape from something aversive and not being able to) can lead to what the researchers term or at least analogise to depression, and also an inhibition of learning (cf. Huston et al., 2013).
Michael Domjan even notes that “a true reversal of acquisition is rarely achieved and may not be possible. The phenomena of spontaneous recovery, renewal, reinstatement, and resurgence all attest to the fact that extinction does not erase what was learned previously.” He goes on to emphasise that extinction does not erase learning: “extinction does not erase what was originally learned and the extinguished behaviour can reappear with the passage of time (spontaneous recovery), a change in context (renewal), reexposure to the US (reinstatement), or the extinction of another response (resurgence).”
“Impulse Control” is a popular phrase in the dog business, and often regarded as a core consideration in living with sheepdogs. In human terms, the idea of “impulse control” refers to our acquired ability to make rational choices in the face of particular stimuli. In living with dogs, though, it’s often framed as the ability for a dog to pause between the presentation of a stimulus (e.g. a reward) and their response to it, in order to make a rational choice; the argument goes that this will reduce the dog’s tendency to experience frustration when they can’t have what they want, and it will slow down their predatory responses (ironically, this desire for a pause never seems to be a consideration when the stimulus is a human-offered cue for a behaviour…in that case, trainers generally want their dogs to respond as fast as possible).
Yet, while we accept that humans can learn to control their urges just by virtue of existing in an environment where not all desires or urges can be satisfied, we assume that dogs must be taught, and in a way that we deem appropriate. Further, the idea of impulse control in dogs assumes that we must make the decision on their behalf–they must comply with our choice in the face of that stimulus, and that they should wait for our permission to engage.
Unfortunately, because the idea of impulse control in dogs is a human construct, it hits our sheepdogs particularly hard: these are dogs who we have selected to be incredibly alert to the environment, to make fast decisions in response to stimuli, and to be tenacious in their work.
The idea of impulse control in our dogs is another manifestation of the focus on compliance: that dogs are irrational beings whose impulses can and should be curbed by us; that we can achieve this by showing our dogs that we control access to the objects that satisfy their basic biological needs; and that we are fundamentally at odds with them, and they with their ‘best selves’, but this can be resolved by their obedience to us.
In humans, impulse control disorders are regarded as clinical psychiatric disorders stemming from altered neurochemistry (e.g. serotonin levels, the balance of serotonin and dopamine, opioids). These are not common in humans, the aetiology (cause) is still disputed, and there is currently no one known treatment modality. That we attempt to apply the term to our dogs is another example of the cultural norm is of pathologizing their difference from us, thinking of training as a remedy rather than a learning opportunity.
Furthermore, the protocols for “teaching” our dogs impulse control often involve withholding things that they want, or setting them up to make an error so that we can take things away from them: these are bound to lead to disappointment and even frustration for a vast majority of dogs, and don’t go very far towards helping the dog make choices that we find favourable. In fact, what we often see is a hesitation that looks like control; this is purely the suppression of behaviour.
What we are trying to help the sheepdog learn in SheepBalls®, in part, then, is predatory control: how to control themselves around the prey (ball) so that they have the best success at attaining the rewards that enacting their predatory behaviour can bring. Dopamine bathing – the stillness, the crouch, the dropped tail, the lowered ears, the raised paw, or whatever posture your dog adopts in stillness as they await the capture without startling the prey – is one demonstration of predatory control.
Counterconditioning involves the modification of a response by association with a stimulus of opposite valence: we often see this attempted where a sheepdog lunges at cars and is offered food in an attempt to transform what is perceived to be an aversive stimulus into an appetitive one.
Just as the brain alters behaviour, so too does the behaviour alter the brain. What this means is that the sheepdog’s control of movement is incredibly rewarding, and as a result of that reward, the set of actions that led to it cause the brain to select that option in the future. As a result, if a sheepdog has had a history of working the wrong sheep, then the sight of that sheep in the future will stimulate the electrical activity in the brain that led to the behaviour involved in seeking to control it.
But while the wrong sheep may represent a potential reward (a sheepdog may successfully control the movement of a cat or their person), it may also represent a frustrated reward. The sight of a car or a motorbike, for instance, may predict a desired reward (capture) that has frequently been unattainable. The dog may respond, then, by spinning, barking, yelping, and so on: behaviours that humans like to term signs of “frustration.”
A major failing of the attempt to apply counterconditioning protocols to our dog, though, is the human perception of what is aversive: we assume that Sheepdogs lunge at cars because of a negative emotional response to them, rather than considering the dog’s predisposition to being stimulated by and attempting to control movement. We also misinterpret the valence of rewards to a dog in a particular moment in time: approach-seeking predatory behaviours shut down peristalsis (digestion of food) to prepare the body for action (Löw et al., 2008), meaning that food is unlikely to be rewarding in this context, and if that is the case, then counterconditioning cannot be achieved, since the presented stimulus is not of opposite valence to the undesirable one.
Counterconditioning has also been shown to be functionally similar to extinction, in that renewal of the original response is likely; in fact any fear response that returns may, in fact, be greater after renewal (Holmes et al., 2016).
Instead of seeking to suppress through extinction, the solution is to help our sheepdogs channel their learning through the availability and reliability of rewards, so that the activities that we wish them to select for themselves become rewarding in themselves; this is the foundational method of Learning About Dogs.
Our priority should be helping these dogs to develop skills through a rich reward process, and in this facilitating their ability to choose this, not that. The desirable side-effect of this is that in setting them up for success in their learning, and in making the reward process so rich and pleasurable for them, the stream is going to be more likely to flow in the direction of this. What others regard as failures of impulse control, behaviours that need to be extinguished, or responses that require counterconditioning, then, may be a gap in learning: a lack of experience of the rewards of this and a strong history of self-rewarding with that.
When it comes to our dogs’ learning and the gaps we observe, it serves us well to think in terms of “not yet,” and to build the this pathway on a strong reward process and, where appropriate, management that they are more inclined to select it over that. A selection and provision of appropriate rewards will ensure that we’re not pursuing the futile endeavour of trying to suppress the heritage of our sheepdogs, but instead offering them a life of ample and appropriate rewards and channelling their needs towards appropriate outcomes.
Azrin N.H., Hutchinson R.R., Hake D.F. (1966) “Extinction-Induced Aggression.”Journal of Experimental Animal Behavior 9.3:191-204. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1338179/
Bouton ME (2004). “Context and Behavioral Processes in Extinction.” Learning & Memory, 11.5: 485-494. https://pubmed.ncbi.nlm.nih.gov/15466298/
Domjan, Michael (2015) The Principles of Learning and Behavior. Seventh Edition. Cengage.
Holmes N.M., Leung H.T., & Westbrook R.F. (2016). “Counterconditioned Fear Responses Exhibit Greater Renewal than Extinguished Fear Responses.” Learning & Memory, 23.4: 141–150. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4793199/
Huston, J.P., et al. (2013) “Animal Models of Extinction-Induced Depression: Loss of Reward and its Consequences.” Neuroscience and Biobehavioural Reviews 37.9: 2059-2070. https://doi.org/10.1016/j.neubiorev.2013.02.016
Kaminski, J. and Nitzschner, M. (2013) “Do Dogs Get the Point? A Review of Dog-Human Communication Ability.” Learning and Motivation 44.4, 294-302.
Löw, Andreas, et al. (2008) “Both Predator and Prey: Emotional Arousal in Threat and Reward.” Psychological Science 19.9: 865-873. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3612950/