Positive reinforcement
The dog sits, the ball gets thrown. Sitting happens more. This is most of what a session with me looks like, and for most dogs the currency is a game rather than food.
the science behindHow Dogs Learn
Everything I do with your dog rests on about a hundred years of research. Here is what it says, in plain English.
Training is not a secret
People come to me expecting a trick. Some word, some gadget, some way of standing that makes a dog fall into line. There isn’t one.
What there is instead is a small set of principles that describe how any animal learns, worked out in laboratories between the 1890s and the 1950s and confirmed every day since. None of it is mysterious. Most of it is obvious once someone says it out loud.
I keep studying this because knowing why something works is what tells you what to do when it stops working. A trainer who only knows the recipe is stuck the moment your dog does something the recipe did not cover. That is most dogs, most weeks.
Three names do most of the heavy lifting. Here they are, and here is where each one shows up in your living room.
Ivan Pavlov, and what your dog feels
Pavlov showed that an automatic, biological reflex can be paired with a previously meaningless thing in the environment until that thing produces the reflex on its own. This is classical conditioning.
The important word is automatic. Nothing here is a decision. Your dog does not choose to feel excited when you pick up the leash, and does not choose to feel sick in the vet’s parking lot.
Something neutral got paired with something that already meant a great deal, often enough that the neutral thing took the meaning on.
You already have a house full of it: the cupboard that holds the food, the particular jingle of your keys that means a walk rather than work, the sound of a specific car in the driveway. Nobody trained any of that on purpose.
You cannot instruct a dog out of an emotion.
This is the layer I work on with a reactive or frightened dog, and it is the layer most owners are trying to fix with obedience by mistake. Asking a genuinely scared dog to sit does not make it less scared. It just gives you a scared dog that is sitting.
That work has its own program — see behaviour modification, which starts with an in-person evaluation rather than a booking.
What changes the feeling is changing what the trigger predicts — patiently, and far enough away that the dog can still think.
While you are building an association, pair it every single time. This is the one place where being completely predictable is the point — the dog is learning “this always means that,” and every exception you allow slows it down.
Edward Thorndike, and the consequence
Thorndike put cats in a puzzle box and timed how long they took to get out. They did not reason it through. They scrabbled at everything until something worked, and then got faster at the thing that worked.
Out of that came the Law of Effect, which is the sentence the whole trade is built on: actions followed by a satisfying result get repeated, and actions followed by an unpleasant one fade out. Your dog is not being stubborn or dominant. Your dog is doing whatever has been paying.
That is the one everybody quotes. Thorndike wrote several more, and the rest are where the practical answers live — they are the reasons a dog that “knows” a command does not do it.
Actions followed by a satisfying result get repeated. Actions followed by nothing, or by something unpleasant, fade.
The law of exercise is why every board-and-train program ends with a go-home lesson and follow-ups, rather than handing your dog back and wishing you luck.
What is Thorndike’s Law of Effect?
Actions followed by a satisfying result are more likely to happen again. Actions followed by discomfort weaken and drop away. Everything else is detail on top of this. If a behaviour keeps happening, something is paying for it — and it might not be you. Counter-surfing pays in sandwiches whether or not you were in the room.
What is the Law of Readiness?
Learning happens when the animal is physically, mentally and emotionally prepared to act. Force it when the dog is not ready and you get frustration instead of progress. This is why I do not start a session with a dog that is over threshold, exhausted, bursting for the toilet or full of unspent energy. Meeting the dog’s needs first is not indulgence, it is a precondition for anything sticking.
What is the Law of Exercise?
A connection between a cue and a response strengthens with practice and weakens without it. Training is not a course your dog completes. The dog you get back from me is the dog you keep if you keep using it, and the dog you slowly lose if you don’t. Five minutes most days beats an hour once a fortnight.
What is the Law of Multiple Response?
Faced with a new problem, an animal tries lots of different things until one succeeds. That flailing is not disobedience, it is the search. It is also why a dog offers you three wrong behaviours before the right one when learning something new, and why punishing the search teaches a dog to stop offering anything at all.
What is the Law of Set, or Attitude?
What the animal is disposed towards shapes both what it will try and what counts as satisfying to it. Two dogs in the same situation are not working on the same problem, because they do not want the same things. This is the formal version of what I say plainly elsewhere: work out what actually motivates the dog in front of you rather than assuming it is a biscuit.
What is the Law of Response by Analogy?
In a new situation, an animal does what it did in whichever old situation feels most similar. This is why a dog that is perfect in your kitchen falls apart in a car park. To the dog those are not the same problem, and it has nothing similar to draw on. Generalising a behaviour deliberately — new places, new distractions, new surfaces — is the work, and it is the part most people skip.
What is the Law of Associative Shifting?
A response can be moved from its original trigger to a completely new one by pairing them and shifting across in small steps. This is the mechanism behind almost every cue your dog has. The word means nothing on day one; it acquires meaning by riding alongside something the dog already understands until it can carry the behaviour by itself.
What is the Law of Partial Activity?
An animal can respond to one element of a situation and filter out the rest, rather than reacting to the whole scene at once. Useful when it works for you and maddening when it doesn’t — the dog that only sits when you raise your hand has locked on to your hand, not your word. Half of proofing a behaviour is finding out which part of the picture the dog is actually keying on.
B. F. Skinner, and the four quadrants
Skinner took Thorndike’s Law of Effect and built it out into operant conditioning: the study of how voluntary behaviour is shaped by what happens immediately afterwards. He gave us the vocabulary the whole industry argues in.
Before the grid, one piece of housekeeping that causes more bad arguments than anything else in dog training. Positive and negative here do not mean good and bad — they mean added and taken away, like on a calculator.
Reinforcement means a behaviour goes up. Punishment means a behaviour goes down. That is all these four words have ever meant.
All four are running whether or not a trainer names them. A dog whose recall is ignored is being taught that the word means nothing. You do not get to opt out of the quadrants. You only get to choose whether you are using them deliberately — which is what the methods question on the programs page is really answering.
Negative reinforcement
Steady leash pressure stops the moment the dog steps towards you, so stepping towards you happens more. Every leash in the world works this way, including the flat collar on a dog that has never met a trainer.
Positive punishment
The dog does something and gets an unwelcome consequence, so it does it less. This is the quadrant everyone is really arguing about, and the one that does the most damage when it is used without the other three in place.
Negative punishment
The dog jumps up, the game stops and you turn away. Jumping happens less. It costs nothing and owners underuse it badly, usually because it does not feel like you are doing anything.
How often you pay
Getting a behaviour is one job. Making it survive contact with the real world is a different one, and it comes down to how often you reward.
While you are teaching something new, pay every single time. The dog is still working out what earns the reward, and skipping some just muddies it. This is also true of the Pavlov side: while an association is being built, be completely predictable.
Once the behaviour is reliable, stop paying every time and start paying unpredictably. A behaviour on an unpredictable schedule is far harder to extinguish than one that has always been paid.
That sounds backwards until you think about a slot machine. If a vending machine takes your money once, you stop using it. People feed slot machines for hours, because the next one might pay.
Pay every time while it is being learned. Pay unpredictably once it is reliable.
That is the whole trick behind a dog that still comes when called on the hundredth recall of a walk, having been paid on maybe the ninth. It is also, incidentally, the reason bad habits are so durable. The counter that has food on it once a month is the most compelling counter in the house.
What this looks like at Big Nick K9
Theory is only worth anything if it changes what you actually do. These are the foundations I build every program on, in the order they matter.
-
Build rapport
Before anything else, your dog needs to want to be around me. Nothing else on this list works from a standing start.
-
Build trust
The dog needs to know I am predictable and that nothing happens out of nowhere. A dog that is bracing cannot learn.
-
Establish motivation
Find out what this particular dog will work for. Thorndike’s law of set, in practice: reinforcement only reinforces if the dog actually wants it.
-
Establish cooperation
Get the dog participating rather than complying. Those look similar for a week and completely different after a month.
-
Create clear communication
Markers, cues and tools all do one job: tell the dog precisely which thing it just did was the right one. Timing is most of it.
-
Have a shared goal
You and I need to want the same dog at the end of this. That is what the go-home lesson and the follow-ups are really for.
-
Work ethically
Use the least I need to use to be clear, explain everything before I do it, and stop when a dog has had enough for one day.
-
Respect genetics
Breed and individual temperament set the range you are working in. Training shapes a dog. It does not replace the one you have.
Where the tools fit
I use a prong collar, and on longer programs an e-collar. Given everything above, here is exactly what they are for.
They are communication devices. A tool does not train a dog any more than a leash does; it gives me a way to say something clearly at the moment it needs saying, at a distance, without shouting and without dragging. Used properly the pressure is light, informational, and released the instant the dog is right — which, in Skinner’s language, is mostly negative reinforcement rather than punishment.
A tool does not train a dog any more than a leash does.
My primary motivator is still play. I use some treats, but most dogs, once their other needs are met, would rather chase a ball or play tug than eat — and that builds a different kind of bond than food alone does.
The tools sit on top of that foundation, never in place of it, and they go on late rather than early.
I will walk you through how and why before anything goes on your dog. If you are not comfortable, say so and we will talk about it. A tool used by someone who does not understand it is worse than no tool at all, which is the actual reason this article exists.
Where this comes from
I am a dog trainer, not a researcher. Everything above rests on work published by people who were, and you should be able to go and read it rather than take my word for it.
- Ivan Pavlov — biographical, The Nobel Prize. Pavlov took the 1904 Nobel Prize in Physiology or Medicine for the digestion work that turned up conditioned reflexes by accident.
- Animal Intelligence (1911), Edward L. Thorndike, in Classics in the History of Psychology, York University. The puzzle-box experiments and the Law of Effect, in his own words.
- The B. F. Skinner Foundation. Skinner’s own writing on operant conditioning, reinforcement and schedules.
- Behaviorism, Stanford Encyclopedia of Philosophy. The wider argument these three sit inside, including the objections to it.
Nick Gallina Owner and head trainer at Big Nick K9 in Halethorpe, Maryland. Sixteen years around dogs, professional since 2022, and still reading the research. More about how I got here, or see the programs and pricing.
Want this applied to your dog?
Tell me what is going on and I’ll tell you straight whether I can help, and which program fits.