Habits allow us to perform actions without attending to every detail; we can do complex things and more than one action at a time without overloading our cognitive and motor systems. They are goal-directed macro actions made up of a sequence of simple primitive actions. A habit allows a complex action to be launched as a unit and efficiently reach the goal of the habit without each step needing its own specific goal.
In forming a habit, a sequence of actions is consolidated by passing from a closed reward loop to an open reward loop. In other words the whole sequence comes to be evaluated rather than each step. Passing from step to step becomes much faster when it is automatic. “To explain how these sequences are consolidated, Dezfouli and Balleine distinguish between closed-loop and open- loop execution. At the beginning of learning, feedback is crucial. The organism needs a reward or some clues in the environment to identify and perform the proper behavior (closed-loop execution). In advanced stages of training, a step in the sequence is conditioned by the previous step, regardless of feedback stimuli or reward (open-loop execution). This independence accounts for the insensitivity to the outcome shown in experiments of reward devaluation and contingency degradation that are standard measures to determine if a habit has been acquired .” It takes persistent failure of the expected reward to disrupt the habit.
Learning is adaptation of an individual to the environment by changes in behavior resulting from regularities in the environment. Learning is adaptive because it is a response to regularity. As habits present regularities in the environment because one step automatically follows another, they can be the basis of learning.
The author, Balderas, (see citation below) uses the fast-mapping that dogs do in learning to associate a name with an object, to illustrate the intertwining of habit and learning. Only some dogs do fast-mapping: learning that a new word applies to the only new object available using exclusion logic. Other dogs stand about looking lost. She explains the learning of a particular dog, Rico, that uses two habits (automatic sequences): one is playing fetch and the other is associating a name with an object. The fetch sequence has three main actions (a) go for (b) select (c) deliver. Select however can be seen as a sub-sequence (1) look-for (2) match (3) take. If there is no new object/name then abc can be executed without interruption. But during fast-mapping it becomes more complex. “In this case, take can not start because match was not executed. Since Rico does not dispose of a name-object association that enables it to complete the task, it is in a situation where it has to make a decision in the middle of the selection task, so the goal-directed system regains control. After solving the problem, the fetching-game sequence follows its tendency to completion and Rico returns to the sequence: it goes to take and to c (deliver). This description also follows the hierarchical view because at the starting point the behavior begins as a habit, when a decision is required it becomes goal-directed and ends again as a habit after overcoming the difficulty.” The dog uses the exclusion principle, and that involves the matching of previously learned pairs to eliminate them. When the dog finds the only possible answer is the unmatched object, he must select this object in order to deliver and reach the end-point, the habit’s goal. This sequence results in learning a new name/object matching. Habits modulate behavior and guide the animal to detect and solve a problem and thus learn.
I have to admit that part of the reason for this post is my love of a former dog (a much missed border collie – husky cross) who could learn vocabulary, including by the exclusion principle. We were building a house and the internal walls were only the studs. I had shown people around and the dog had followed. I would stand in a space and say this is the kitchen and then go on to the next room. After a few times the dog preceded the group. Then I would stay in the middle of the house and say, “Badger, show them the kitchen”. She did the tour with me only naming the rooms. Then one day I said, “show them the basement”. The dog looked at me and around the space, a couple of times and then trotted to the top of the stairs to the basement. I don’t think she picked up the word ‘basement’ from conversations or she would not have been puzzled at first, but she did recognize that it was the only space left that she could possibly show them. From then on she could be told to go to the basement and she understood. When the walls were finished she could still be told to go to a particular room, although now she had to use the doors.
Balderas, G. (2014). Habits as learning enhancers Frontiers in Human Neuroscience, 8 DOI: 10.3389/fnhum.2014.00918