How Dogs Learn
- Greg Roder
- Jan 9
- 9 min read
Cornerstones of canine behaviours and learning.
Foundations
Understanding canine behaviours that “come naturally” - and distinguishing those from learned behaviours - is an important aspect of successful dog guardianship, management and training. Further, appreciating how dogs actually learn, that is, by observation, imitation and the outcome/consequences of their actions, are key to not just training canines, but appreciating the depths of their understanding of the physical world. There are even some surprising aspects of dogs in their observations of humans and events that we might not intuitively appreciate. So, here are some basic elements of canine behaviours and learning to aid in the dog companionship and training arenas.
· Innate behaviours: Those expressed with no specific training or experience needed, and have a strong genetic basis, as they are encoded in the animal’s genes and do not require any learning to occur. An example is the suckle reflex of newborn puppies.
· Instinctive behaviours: Inherited, but occur in response to particular stimuli and are species-specific, with little variation in pattern – can be refined with experience. An example would be the hunting or herding instincts, stronger in some breeds than others.
· Encoded learning: The ability of an animal to learn is encoded in its genes, but it is also important to remember that all learning is influenced by the animal’s physical health, emotional state, environment and previous learning experiences. Just like humans, learning engenders learning.
· “Do as I do”: Trainer demonstrates an action for the dog to follow. Once a dog has learned the cue to copy a behaviour, they will copy new behaviours. Examples are jumping obstacles, catching objects and encouraging the dog’s interest in guardian behaviours, or focussing on a target using pointing gestures, e.g., for going to a place, touching an object or searching a location. Dogs watch humans, observe, and follow pointing and nodding gestures, follow the human’s gaze and look to them to determine what to do next.
· Fairness: Range et al[1] and Greenfieldboyce[2] found that when two dogs were asked to perform identical tasks in view of each other and one dog got a reward and the other didn’t, the second dog was less likely to perform the behaviour again. Whereas if neither dog received a reward after performing the task, or if both dogs received rewards, they were more likely to perform the behaviour again – because (it seems) of their sense of the “ethics of equality”.
· Dogs watch human interactions: Marshall-Pescini[3] found that dogs were more likely to solicit attention from someone they have observed being kind to another person, than someone who was less kind (see also Hoi-Lam Jim et al[4]).
· People worth watching: Takaoka, et al [5] allowed dogs to observe people opening boxes containing food, while others opened empty boxes. The two groups of humans later used a pointing gesture to indicate to the dogs which box to investigate. Dogs did not follow the pointing indication from those who they had seen opening empty boxes, but did for experimenters who had opened boxes of food. Miklosi[6] and Marshall-Pescini[7] came to related conclusions regarding dogs watching humans.
· “Learn to learn”: Ottosson[8] found that trained dogs were more often successful at accessing the hidden food in food puzzle toys, showed more focus on a toy and were faster to manipulate it. Various researchers have attributed this success to the fact that trained dogs acquire the skill to “learn to learn” – as well as being more motivated to work for food rewards, so the toy becomes more attractive to explore.
· Imitation: Dogs are seen to observe each other and follow suit, copying the other dog’s behaviour. Farmers may use this in training young dogs, running through the work alongside an experienced dog (this is generally a complimentary method, not the complete foundation and not the refinements required).
The Guiding Principles of teaching dogs.
1. There are 4 stages of dog training (or “learning” – rather like a human learning a new language):
a. Acquisition; b. Fluency; c. Generalisation; d. Maintenance
Acquisition is learning the task or skill – and performing it when cued to do so.
Fluency is performing the action (almost) every time it is cued.
Generalisation is performing the cued action no matter the environment and distractions (i.e., in the yard, in the street, in the park, with an audience, etc.).
Maintenance is maintaining/being periodically reminded of the cue and requested action (dogs may lose responsiveness to a cue if not reminded for about three months – but obviously a great deal of variation in this).
2. “Doggie Zen” is the foundation - comes first, even before the Acquisition phase – build the bond and trust (see article on this website “Zen and the Art of Companion Dog Training”).
3. As a guardian/dog trainer, you get out what you put in – start out the way you wish to continue and base the learning on the bond with your dog.
4. Dog training is simple – because your dog already knows how to do most things.
But that doesn’t mean it is easy – we teach the dog what we want them to do and how exactly we want them to do it, when we want them to do it – using positive reinforcement. So, teaching a dog to sit on cue is probably the easiest action – ever met a dog which doesn’t sit? Compare this to training a Border Collie or Kelpie to herd sheep. These breeds have a strong natural herding instinct, they already know how to herd sheep – getting them to move the sheep where the farmer wants them to go, often working in twos or threes, in a quiet, orderly manner, is the trainer’s challenge.
5. Consistency, Perseverance and Patience are key – “5 minutes a day” for a young dog (longer for a mature dog working on complex actions and behaviors). Remember - dogs (like humans) learn in different ways at different rates.
6. In the end – it’s all down to you, the guardian/trainer – build relationship, bond and trust with your dog and enjoy their company and loyalty – dogs love to please you.
Always have fun with your dog – accept constructive advice (rise above criticism); laugh off errors and mistakes and rejoice in successes; never get stressed about your dog’s (or your own) learning, prowess or capabilities – your dog is the best dog in the world.
Dog learning and positive reinforcement: essential background understanding.
a) Impact of Stress on Learning in dogs.
· “Stress” results in physiological and/or psychological responses to internal or external threats, pressures or pain (stressors). Stress involves changes affecting nearly every system of the body, influencing how animals feel and behave. Stress has important functions in that it is involved in:
• motivating animals to act e.g. hunger results in stress and motivates the animal to seek food.
• prepares the body and brain to respond to threats e.g. run out of the way of a falling tree.
• redirects energy resources and blood flow away from non-essential processes towards essential ones for the survival of the animal.
The “SEVEN F’s” in dog reactions to Stress[9]
1. Fight (defend/attack)
2. Flight (run away)
3. Freeze (locked in place)
4. Fidget (fiddle, wriggle, look around)
5. Fawn (obsequious/lie on back)
6. Flock (go to family for protection)
7. Faint (fall down unconscious)
Hopefully it is obvious that the “stress” referenced here is not the mind and body exercise type of stress accompanying, for humans, playing sport or chess or Scrabble - and for canines, learning/participating in Obedience, Agility, Rally-O, Dock Diving, Hoops, etc. Rather the subject stress (which might cause one of the “Seven F” reactions) is that creating concern or fear, caused, for example, by a perceived threat or environmental predicament.
· “Distress”, on the other hand, is when the intensity, frequency or duration of stress is prolonged and prevents the animal from returning to a calm relaxed state readily. If an animal is frequently, intensely distressed, or in a state of stress for prolonged periods, this is maladaptive, that it, the abnormal emotional state interferes with an individual’s activities of daily living or ability to adjust to or participate in certain situations. The distress may also lead to unpleasant, or dangerous behaviour.
b) Before applying the tools of operant conditioning[10] in dog training, here are some foundation principles to remember:
· Behaviour is repeated if it is rewarded.
· Positive reinforcement of behaviours/responses engenders a desire/willingness to learn – seemingly finding other ways/responses to cues to achieve success and a reward[11].
· Behaviour, once learned, will be remembered longer if rewarded intermittently (often referred to as the “gambler’s logic – I might win next time”).
· You cannot teach an animal that is not ready to learn (because of stress, distractions, environment, poor past experiences relevant to the teaching objective, etc.).
· If behaviour ceases to be rewarded in any way, it will not be repeated, or phrased another way, if the animal finds no reinforcing value in the behaviour, that behaviour will diminish. In detail, behaviour increases in duration, frequency and intensity when it is reinforced. This occurs whether the trainer reinforces the behaviour, or the animal finds the behaviour innately reinforcing[12]. When using positive reinforcement techniques to train a behaviour, we reinforce the animal for a “correct” response to increase the likelihood of the behaviour recurring. If you do not reinforce within a second of a behaviour occurring, you may have “missed the window of opportunity” and are even possibly reinforcing another behaviour.
· To use positive reinforcement, we first need to identify something the animal finds reinforcing. It is easiest (and hence, most common) to use a “primary reinforcer” (usually a food treat, but also play and for certain breeds it might be a structured exercise of “seek and find” – the successful “find” being the reward in itself, probably reinforced by a happy trainer).
Acknowledgements and further References.
The information summarised in this article combines understandings from documentation (books and research articles), personal experience in training dogs and instructing dog guardians, as well as learnings from course work undertaken by the author.
Articles of particular relevance have been referenced in the text (footnotes below) – these can lead to many other research papers on those topics, for those interested in delving further.
In addition, certain books have proved invaluable in building an understanding (and appreciation) of canine behaviour. Of particular note (although not the only ones of relevance – have fun finding others) are:
Serpell James (Ed) (2017) The Domestic Dog; its Evolution, Behavior and Interactions with People; 2nd Edition Cambridge University Press; 416pp.
Horwitz, D. F. & Ciribassi, J. [with Steve Dale] Eds. (2015) Decoding Your Dog - Explaining Common Dog Behaviors and How to Prevent or Change Unwanted Ones; American College of Veterinary Behaviorists; Mariner Books 360pp.
Horowitz, A (2009) Inside of a Dog - What Dogs See, Smell and Know; Simon & Schuster, UK; 353pp.
Burch, M. R. and Bailey, J. S. (1999) How Dogs Learn; Howell/Wiley, NJ; 188pp.
Vette, M. (2017) Dog Zen: Everything You Need to Know to Transform Your Dog; Random House NZ; 400pp: (2019) Puppy Zen: Eight Weeks to Train Your Pup and Create a Lifelong Bond; Random House NZ; 432pp: (2020) Zen Heart: What I've Learned From Animals and Life; Random House NZ; 304pp.
----------------------------------------------------------------------
[1] Range, F, et al (2009) The absence of reward induces inequity aversion in dogs; PNAS;
[2] Greenfieldboyce, N. (2008) Dogs Understand Fairness, Get Jealous, Study Finds; npr;
[3] Marshall-Pescini S, Passalacqua C, Ferrario A, Valsecchi P, Prato-Previde E. Social eavesdropping in the domestic dog. Anim Behav. 2011;81(6):1177–83; URL https://www.sciencedirect.com/science/article/abs/pii/S0003347211000959
[4] Hoi-Lam Jim et al (2020) Do dogs eavesdrop on human interactions in a helping situation? PLoS One
26;15(8); Nat. Libr. Med.; URL https://pmc.ncbi.nlm.nih.gov/articles/PMC7449479/
[5] Takaoka, A. et al (2014) Do dogs follow behavioral cues from an unreliable human? Animal cognition; 18 (2); URL https://www.researchgate.net/publication/267743102
[6] Miklosi, A. et al (2003) A simple reason for a big difference: wolves do not look back at humans, but dogs do; Curr Biol; 29;13(9):763; at National Library of Medicine; URL https://pubmed.ncbi.nlm.nih.gov/12725735/; Note – references many studies with similar conclusions.
[7] Marshall-Pescini, S. et al (2017) The role of domestication and experience in ‘looking back’ towards humans in an unsolvable task; Scientific Reports; 7(1):46636; URL https://www.researchgate.net/publication/316251885; also (2014) Do Domestic Dogs Understand Human Actions as Goal-Directed?
[8] Nina Ottoson (1990+); creator of puzzle food toys for dogs and cats; URL https://www.nina-ottosson.com/about-us/nina-ottosson-story/
also (2014) Do Domestic Dogs Understand Human Actions as Goal-Directed?
[9] And see article on this website on “Dog Body Language”.
[10] Operant conditioning (or instrumental conditioning) is a theory of learning where behavior is influenced by its consequences. Behavior that is reinforced (rewarded) will likely be repeated. E.g. McLeod, S. (2024) URL https://www.simplypsychology.org/operant-conditioning.html
[11] Thorndike, E. L. (1911). Animal intelligence. New York: Macmillan: Skinner, B. F. (1938). The behaviour of organisms: an experimental analysis; New York, Appleton-Century: (1951) How to Train Animals; Scientific American, 185(6), 26–29. (1976) About Behaviourism; Vintage books (Random House) NY; 291pp.
[12] Think about the dog barking at the Postman/woman – they deliver the mail and leave – mission accomplished, the dog barked, the Postie ran away, the dog won, the behavior was reinforced. Also consider the “counter-surfing” dog – if it finds food, success and reinforcement – will definitely try that again.
