Learning by Trial and Error Part 4 – Negative Punishment

Negative Punishment – the removal of a stimulus following a behaviour which reduces the likelihood of that behaviour re-occurring.

Before I move forward, punishment in terms of Operant Conditioning means anything which reduces the previous behaviour. Punishment is neither good or bad in this context, but it is the term we are stuck with unfortunately. In terms or Operant Conditioning, we punish behaviours, we do not punish the learner.

Examples of negative punishment

  1. A dog barks excitedly on seeing another dog when he is on the lead. He is friendly and wants to go and see the other dog. I immediately move him away from the other dog. As a result, the dog learns that if he lunges and barks excitedly, the other dog will be moved further away, thus reducing the lunging and barking.
  2. I toddler is playing with a toy and another child wants to play with the toy too. Toddler one screams at todler 2 resulting in Mum removing the toy and giving it it toddler 2. Toddler 1’s screaming has been punished (if this works) by the removal of the toy.
  3. You call your dog to you and put the lead on and off lead play time ends. You have negatively punished the recall by removing play (not a great strategy, which is why we need to train recall with lots of fun stuff as well, so the dog learns recalling is a good thing). If this is the only time you recall your dog, your dog will qucikly learn not to come back to you.
  4. Your teenager stays out beyond curfew. You ground him. You have (hopefully) negatively punished his breaking curfew by removal of social activities.

In my own experience, negative punishment has limited uses in dog training. When used effectively, it can be a very useful tool however. When we first got our dog, she would lie down in the “stalking” position when she saw another dog. She did this in anticipation of the chase. As soon as she did this, I immediately gave her a verbal prompt and walked off with her in the opposite direction. As a result, the lying down part of the stalking behaviour reduced.

Lastly, withholding a treat for an incorrect response is not negative punishment, it’s non-reinforcement. Non- reinforcement is doing nothing. With negative punishment, as in any of the other 3 quadrants of Operant Conditioning, we (or the environment) has to act to remove or add something after the dog’s behaviour. If I am teaching my dog to lie down before I throw the ball and she sits instead, I am not negatively reinforcing the “sit” by not htrowing the ball, I am just not reinforcing it.

In the next blog of this series, I shall discuss positive punishment and discuss why it’s not the best tool to rely on, but is so often used by trainers.

Happy training.

Advertisements