• Positive Reinforcement (“reward” such as food or play)
• Negative Reinforcement (“relief” such as provided by painkillers or getting to breathe again if I force your head under water)
• Positive Punishment (this is the scary or painful kind of “punishment” such as pain, extremes of temperature, having me start forcing your head under water)
• Negative punishment (“penalty” such as getting a ticket or losing a privilege).  

What this means is that there are 1) good things and 2) bad things and that both these can 1) start and 2) end. Dogs, like all living organisms, are constantly trying to start and keep the good stuff (attention, walks, food, access to dogs if they like dogs, play etc.) and avoid starting or terminate the bad stuff (aversive methods, like pain or anything else that activates an animal’s “Imminent bodily injury or death” alarms). In our training methods, there’s no pain, violence or scary stuff. In other words, the dog is trained using positive reinforcement – rewards – and negative punishment – time-outs or “penalties” such as snatching food away when the dog breaks his stay. 

Pet dog training based on the military tradition that started in the post WWII era relied heavily on nasty punishments such as leash corrections to motivate the dog. Over the decades, there’s been an explosion in dog-friendly methods that make greater and more sophisticated use of reward and reward-removal as principle motivation. The control is as good, and in some cases better than that achieved with traditional methods and avoids the side-effects of aversive methods. Having violence done to you is motivating for sure but it is not without fall-out.

A reward is anything that increases the future probability of the behaviour it immediately follows. Examples of likely rewards for dogs are: attention, food, play, walks and access to other dogs. The trainer’s timing is everything. Because it’s necessary to reward the desired behaviour so precisely and it’s not always possible to get a treat, pat or toy to the dog fast enough, a key word or sound (“conditioned reinforcer” or “bridging stimulus”) that means “that’s it! - you win!” is a useful tool. Another valuable aid is a signal that means “oops – you just lost!” and is called a “No-Reward-Marker (NRM), such as “too bad” or “uh-uh.”

NRM’s require split-second timing just as reward-markers do and are always followed by reward removal. 

When training a new behaviour, a reward is given for every correct response. We want strong, not weak, or middling, behaviour. After the first few sets, rewards should not be shown up front but hidden and offered immediately after the desired action is performed (or immediately after the bridging stimulus, like the conditioned reinforcer used in training). So you’re still giving it every time, you’re just keeping it hidden on you until you give it. When the behaviour is strong and as fancy as you like it, you can maintain it with intermittent rewards, especially if these are unpredictable to the dog. The craftiest use of intermittent reward is to reward the best examples of the behaviour. 

Emotional Learning 

The second way dogs learn is via tip-offs in their world to events of relevance to them. This is why dogs are so happy when you bring the leash out – the leash is the tip-off that a walk is about to happen. It’s also why they look all forlorn when you pick up your keys and brief case – this picture is the tip-off that they’re about to be left alone. Dogs learn the sound of the cookie jar, they learn that towels and hoses predict baths, that a certain look on your face means you’re going to pin them to the ground (or some other violent act) etc. Dogs can also use time as a tip-off. If you reliably feed your dog at 7 a.m. and 5 p.m., he will salivate and feel happy anticipation shortly before those times. 

These are all examples of classical conditioning, the form of learning whose details were worked out by Ivan Pavlov around the turn of the last century. Those same details have bored generations of psychology students in intro-psych classes, an obligatory rite of passage to sexier psych courses on fancy aspects of the mind. There’s no doubt that classical (and operant) conditioning is wildly out of fashion in psychology, but it’s a goldmine for dog owners. It illuminates many of our “why does he…” questions and is a royal road to influencing whether our dogs like or fear things.  

To exploit Pavlovian style conditioning, you don’t need to think about behaviour and consequences (that’s operant conditioning). You just have to make something you’re trying to get your dog to like a tip-off to something he already likes. For example, if you want your puppy to like being brushed, do a bit of brushing before every walk, if he likes walks. If you want your dog to like kids, every time the dog encounters a kid, give him a fabulous and rare treat right afterwards. The order matters. It’s just not good enough for one thing to happen “around the same time” of another thing. The kid has to come first and the fabulous treat after. Otherwise it’s not much of a “tip-off”. 

Another thing that’ll help get stronger emotional conditioning is having as close to a 1:1 ratio between the two things as possible. So, try to always give the treat after the dog encounters a kid and don’t give that kind of treat except after the dog has seen a kid. The tighter this relationship, the stronger the conditioning will become. 

If your dog already has a negative emotional response about something - let’s say he growls at kids or trembles in terror at the vet (because the vet has reliably predicted being put up on a slippery metal table, poked and prodded and given shots), it’s tougher, because now you’re not just conditioning him to like the vet you’re counter conditioning an existing problem. It’ll take longer, you’ll need to go very slowly and you’ll need ultra-heavy artillery, treat-wise.

And don’t worry about “rewarding” his vet phobia by comforting him in the waiting room. Emotional responses are involuntary. If you’re afraid of spiders, when you scream it won’t make you more afraid next time. Same thing with punishment. If you’re afraid of spiders, me twisting your arm to the point of pain will not make you less afraid of them next time. The way to alter emotions is using Pavlov along with desensitization.