What are habits?
Have you ever been driving and been so lost in thought that you aren’t even consciously aware of what you are doing? For a well-practiced skill like driving (and especially for a well-practiced commute) you might even find yourself arriving at your destination with little recollection of all the steps you took to make it there. Or maybe you’ve had the humbling experience of constantly trying to turn on lights and various electronics while your power is out? Most of us have likely experienced something like this, where we act automatically or “out of habit”. As strange as it may seem, habits are exceptionally powerful; indeed, it is partly because habits happen outside of some of our more deliberative, cognitive processes that habits are so powerful.
A History of Habits
We have long recognized that habits exist, from Aristotle to William James. James, viewed by many as the “Father of Psychology” was a particular lover of habits:
“The more of the details of our daily life we can hand over to the effortless custody of automatism, the more our higher powers of mind will be set free for their own proper work. There is no more miserable human being than one in whom nothing is habitual but indecision, and for whom the lighting of every cigar, the drinking of every cup, the time of rising and going to bed every day, and the beginning of every bit of work are subjects of express volitional deliberation .”
And that does sound pretty nightmarish, right? Imagine having to consciously think about every little thing you do. Automation and machine learning are all the rage nowadays, but really, habits have been automating the dull stuff for millenia.
But what are habits anyways? In the early 20th century, many early behaviorists fell into two camps that attempted to explain animal behavior. Some, like Clark Hull, took a mechanistic view, where everything that animals (including us) did was elicited. What this means is that there was no choice or evaluation but only stimuli that elicit a response (for instance, if you saw a red light while driving you would always automatically press the brake). This view, where stimuli directly elicit responses is basically analogous to habits. In contrast to this view, scientists such as Edward Tolman proposed that we form a sort of “cognitive map” of the environment that allows us to guide our behavior . This cognitive map would hold all sorts of useful information – if we return to our red light example, the cognitive map might also include the speed your car is traveling and if you are being tailgated or not. This contextual information then might affect your decision to slam on the brakes. As is often the case in science (and society), these two sides engaged in increasingly contentious and polemical debates, with those in Hull’s camp arguing that all behavior is elicited, while those in Tolman’s camp argued that animals can use information like the context they are in or how motivated they are to modify their behavior.
Why not both?
In the 1980’s Anthony Dickinson and other researchers tried to reconcile these two viewpoints in a series of experiments. Couldn’t animals use both systems? Both the automatic stimulus-response system as well as the more evaluative system ? To investigate this, they trained rats to press a lever to earn a food pellet. Once animals learned the task they introduced an interesting manipulation; they devalue the food pellet. They did this in one of two ways. In one experiment , they paired the food with a drug that made the rats sick (imagine getting food poisoning after eating sushi from your local supermarket…that might make you a bit reluctant to try that sushi again in the future). In another experiment , rats were allowed to gorge themselves on food pellets for an hour (imagine how much you would want more pizza if you just ate an entire pie). After devaluing the outcome, rats were placed back in the chamber with the lever where the scientists measured the number of times they pressed the lever. The mechanistic account of behavior and the “cognitive map” account make different predictions for how devaluing the food pellets should affect behavior.
If animals are using this evaluative “cognitive map”, they should reduce the number of times they press the lever, since how much they want the reward can directly affect their motivation to perform the action. In contrast, if animals are behaving mechanistically/habitually, they should keep pressing the lever anyways. Essentially, the evaluative system (which Dickinson called “goal-directed”) can use information like how much you want the reward to modify the action, while the habit system cannot. What Dickinson and others found is that rats can use either strategy, and which they used depending on a number of circumstances. Thus, the strategy or system that is controlling behavior is not only the habit system or only the goal-directed system, but instead both exert some continuum of control.
What causes habits?
A number of factors have been found that bias the use of either a goal-directed or a habitual strategy. One of the big determinants is the amount of training. If rats only had minimal training on the lever task they would generally use the evaluative or goal-directed system, while if they had extensive training they would tend to use the habit system . This makes some intuitive sense for the habits in our lives. When you first start doing something (learning to drive a car) you have to actively think about every little thing you are doing, but over time, this can become habitual (where you can eventually drive without even thinking about what you are doing). This is also why self-help books often claim that you need to do something for X number of weeks or days in row for it to become a habit.
Another factor affecting which system is controlling behavior is how rats were trained to press the lever. In particular, habits were increased when rats were trained on schedules with variable (intermittent) reinforcement, where sometimes lever presses were rewarded and sometimes they were not . Incidentally, intermittent reinforcement is actually the most effective way to increase a behavior, so the next time you’re training a dog, don’t give them a treat every time they do something you want, but instead give it some of the time.
It turns out that this fact is well known by people who are trying to sell you things. The most obvious example is gambling. In gambling, of course, sometimes you win and sometimes you lose (thank you Kenny Rodgers). This intermittent reinforcement is a powerful promoter of habits, and casinos take advantage of this fact to keep people gambling. But this is also true in less obvious cases – take something like Facebook or Twitter or even your e-mail. The majority of the things you see on your feed will be uninteresting, but every now and then you see something interesting/funny/aggravating. By intermittently giving you what you want, these companies are encouraging you to stay on their website longer and habitually search for that next interesting thing to click on.
Further research by Dickinson and others revealed that habits are also hard to break. To demonstrate this, researchers trained rats to press levers for food like before only now once the rats had learned the task, they changed the relationship between lever pressing and reward. Whereas before lever presses produced food, now lever pressing had no effect on food delivery; instead, food was just occasionally delivered at random. Rats that were habitual took much longer to figure out that the relationship between lever pressing and pellet delivery had been erased . This is analogous to you continuing to try to turn the lights on in the middle of a power outage. In contrast, goal-directed rats quickly realized that this relationship changed and they were able to just sit around and wait for pellets.
The 7 habits of highly effective addicts
Though there is an entire self-help industry based on getting you to acquire “good habits”, habits can definitely have a dark side as well. Some scientists have proposed that addictions arise through the habit system. A prominent aspect of addiction is that people are generally aware of how harmful their addiction is, both for themselves and for others, but they are still unable to stop. In fact, many of the criteria for addiction in the DSM-V (the Diagnostic and Statistical Manual of Mental Disorders) boil down to continuing to use despite negative consequences (known as compulsivity). This inflexibility in the face of change sounds a lot like habits. While most people initially try a drug (or gamble, or steal, etc.) in a controlled, goal-directed way, over time they lose control over their use until they are taking a drug habitually. Remember: increased experience on its own biases the use of habits (as it did for the well-trained rats). More than just experience though, drugs of abuse often directly affect brain regions like the striatum and the prefrontal cortex that are specifically involved in goal-directed and/or habitual decision-making. Some drugs affect these brain regions much more than a “natural” reward like food or sex does, and this can strongly encourage habit development.
Good and Bad in Goals and Habits
Dickinson’s two system model makes a good deal of sense. It seems obvious that we do some things in a goal-directed way, while we do others out of habit. Both goal-directed actions and habits have their benefits. As William James noted, habits allow us to “automate” a variety of mundane tasks, or even get into the habit of doing something beneficial for ourselves like brushing our teeth. This saves us a lot of cognitive energy, and oftentimes habitual actions are faster and more fluid (as in playing an instrument) than their goal-directed counterparts. The inflexibility of habits can also be both beneficial or extremely harmful. While it is very good for you to establish a habit of going to the gym every day, if you can’t stop going even after a crippling back injury, your gym habit can quickly go from helpful to harmful.
Still, it’s clear that both goal-directed and habitual systems have their place. Both systems are evolutionarily ancient, having been observed in simple organisms like fruit flies . By combining flexible, goal-directed behavior with inflexible but faster and more automatic habits, organisms are able to get the best of both worlds – as long as they balance the two appropriately. So the next time you find yourself constantly flicking the lights in a power outage, don’t be too hard on yourself. It’s just a habit.
. James W., The Principles of Psychology (Henry Holt and Company, New York, 1890), vol. 1, Chap. 4.
. Tolman, E.C. (1948). Cognitive maps in rats and men. Psychol. Rev. 55, 189–208.
. Dickinson, A. (1985). Actions and Habits: The Development of Behavioural Autonomy. Philosophical Transactions of the Royal Society B: Biological Sciences, 308(1135), 67–78. https://doi.org/10.1098/rstb.1985.0010
. Adams, C. D., & Dickinson, A. (1981). Instrumental responding following reinforcer devaluation. The Quarterly Journal of Experimental Psychology Section B, 33(2), 109–121. https://doi.org/10.1080/14640748108400816
5]. Dickinson, A., Campos, J., Varga, Z. I., & Balleine, B. (1996). Bidirectional instrumental conditioning. The Quarterly Journal of Experimental Psychology. B, Comparative and Physiological Psychology, 49(4), 289–306. https://doi.org/10.1080/713932637
. Adams, C. D. (1982). Variations in the sensitivity of instrumental responding to reinforcer devaluation. The Quarterly Journal of Experimental Psychology Section B, 34(2), 77–98. https://doi.org/10.1080/14640748208400878
. Dickinson, A., Nicholas, D. J., & Adams, C. D. (1983). The effect of the instrumental training contingency on susceptibility to reinforcer devaluation. The Quarterly Journal of Experimental Psychology Section B, 35(1), 35–51. https://doi.org/10.1080/14640748308400912
[8. Dickinson, A., & Mulatero, C. W. (1989). Reinforcer specificity of the suppression of instrumental performance on a non-contingent schedule. Behavioural Processes, 19(1), 167–180. https://doi.org/10.1016/0376-6357(89)90039-9
. Gomez-Marin, A., Stephens, G. J., & Louis, M. (2011). Active sampling and decision making in Drosophila chemotaxis. Nature Communications, 2, 441. https://doi.org/10.1038/ncomms1455
Featured Image: http://animalearning.com/guiding-s/guiding-s-reinforcement/
Slot Machines: Photo by Stokpic on Pexels.com