Why you shouldn't be rewarding your dog! Punished by Rewards with dogs.

The book (Punished by Rewards by Alfie Kohn) argued that dangling carrots in front of peoples (or dogs) faces, ‘rewarding’ them for desirable behaviours, will in many cases remove the intrinsically rewarding aspects of that behaviour. Simply by the presence of a reward.

There are many examples and studies of this effect presented by Alfie Kohn. One example would be rewarding children for reading books. It pointed to common schemes that are used by schools, libraries and even parents to encourage the child to read more books. All of these schemes included the use of rewards or brides to manipulate the child into reading more.

What was found by these schemes is that children would pick shorter books with more pictures and graphics. In other words, easier books. Once the schemes ended there was no improvement in the children's want to read, in fact the opposite was the case. The children resented reading, if there was no bribe to coerce them to read.

Although it is important to note that as B F Skinner says, it's quite likely that without reinforcement, reading would probably never form on its own. Once the child knows how to read, rewarding them for reading a story book when they actually want to read it, will reduce the intrinsic motivation to read them without the presence of reinforcement.

Reading is not a natural process to begin with, unlike talking which can be learnt purely through observation and is quite intrinsically/extrinsically rewarding as soon as it is learnt.

The trouble with using rewards, is that as soon as you reward a behaviour that may have already been intrinsically rewarding to the organism, you reduce or even eliminate the intrinsic reward mechanism associated with that behaviour. Because not only do you get the organism focussed on the extrinsic value of performing the behaviour, but you blatantly say to the organism “I am manipulating you to perform this behaviour”. How are you doing that? By rewarding the behaviour!

I'll give another example, this time with dogs. Let's say you're teaching a dog to play fetch/play retrieve/chase and catch. And you decide to shape the behaviour by successive approximation with food. What this does, is it immediately turns something that would have been intrinsically motivating to the dog (grasping imaginary ‘prey’ in his/her mouth and interacting in a playful non-competitive way with you) into just another obedience command.

This would be fine if you were teaching a formal obedience retrieval (dumbbell retrieve etc), in which case you could reinforce the obedience retrieval, with the use of a primary reinforcer like food or play that has been allowed to be intrinsically motivating to the dog. But you will not be able to use that type of dumbbell retrieve as a reward, as there is no intrinsic reinforcement in that type of retrieve, as we have blocked it from occurring with the primary reinforcer. This is fine when teaching a form of an obedience retrieve, however when we are trying to get a dog to find an innate biological satisfaction in chasing an imaginary prey (our toy) and then the dog knowing how to trigger that dopamine hit again (by bringing it back to us, to throw again for the dog) is something entirely different. I want the dog to enjoy the game and play it for its own sake. This is not possible when we are rewarding the dog, for playing something that in itself should be a deeply enjoyable experience for the dog. This includes using food to teach an ‘out’ command. *I do not generally teach the dumbell retrieve in the way that I listed above.

I recently saw someone try and use food to ‘reinforce’ a dutch shepherd puppy for biting a decoy. This dog you could see was already extremely motivated to bite, but for some strange reason the handler felt it necessary to shove a bowl of food in front of the dog when he commanded the dog to ‘out’. By the end of the utterly stupid training session, the dog barely knew how to bite! The grips were terrible, the dog no longer bit the decoy on command instead just barked in order to fulfil the ‘obedience protection’ (it had to be agitated by the decoy before it decided to bite). Whereas before all of this nonsense, the pup was biting with completely perfect motivation! This is what can happen when you reward behaviours (like biting) that are already intrinsically rewarding to the dog, especially in the dog's initial stages of acquisition.

The conclusion:

Using extrinsically motivating rewards can ‘ruin your relationship’ with your dog when they are used incorrectly. If you are going to use rewards, if possible it is best to surprise the dog with the reward rather than bribe the dog to do a certain behaviour.

You should avoid rewarding a behaviour that the dog already enjoys, with a primary reinforcer (food etc).

I hope you have benefited from reading this! If you would like to read more of my insightful articles in the future you can subscribe to my mailing list.