My Favorite Things: BF Skinner and the Invention of the Algorithm

Posted inCreative Voices

We know that we like our favorite things more than other things.

What does it mean to โ€œlikeโ€ something?

Weโ€™re used to thinking of liking something as a private, personal matterโ€ฆan attraction or preference of one thing over others. Last time, we saw that some of the things we prefer are the result of associating those things with basic human needs, like survival or reproduction.

Those two fundamental sets of human needs express themselves in varying ways in modern life. For example, โ€œsurvivalโ€ today isnโ€™t merely about having enough food and shelter (although threats to those basics call all other bets off), itโ€™s also about maintaining a way of life.

Could I โ€œsurviveโ€ without a smartphone? Of course I could. My life would be significantly affected by its absence, but my physiological functions would (probably!) keep doing their job.

To โ€œsurviveโ€ as the kind of member of modern society that I choose to be, however, would be much more difficult. The role that my smartphone plays in my life is to bring me the bits and pieces of the world that I have โ€œselectedโ€ as important to being โ€œme.โ€

That selection process is pivotal for us to understand how our favorite things become favorites.

Itโ€™s obvious that, for example, I didnโ€™t just randomly choose a smartphone from the vast array that the marketplace offers. I selected one that was somehow consistent with a way-of-being-in-the-world, that reflected some qualities that resonated with me-as-a-person. Maybe even more importantly, the phone did more than just reflect those qualitiesโ€ฆit enhanced them, strengthened them, made them clearer.

In other words, my smartphone reinforced my identity and I chose it because it did.

Not consciously, of course. Consciously I tell myself, and others, that I chose it because it (now, storytelling starts): โ€œuses the latest chipโ€; โ€œallows side-loadingโ€; โ€œonly cost $400โ€; โ€œjust works with my other tech stuff.โ€

Digging into these stories a bit we find that we choose objects mostly because they go with other objects we already possess. In other words, we โ€œcurateโ€ the world according to โ€œstandardsโ€ that lead to a relatively coherent โ€œcollection.โ€ And, we mostly do this without thinking about it!

How?

Thereโ€™s a great French word that captures the process: bricolage. A bricoleur (one who practices bricolage) is an improvising tinkerer, putting together a diverse range of things that are available and fit together to form a new whole.

In moving through the world as bricoleur, we are picking and choosing things that both fit and strengthen a way-of-living in todayโ€™s culture. What is most interesting is how clearly we know what the โ€œstandardsโ€ are that weโ€™ll use in our bricolage without ever having to articulate them.

Need a watch? What will it be? Rolex? Apple Watch? Omega? You already know the general answer because your identity is like a searchlight shining on the available choices in the watch category, already clear about which of the options will likely be the best-for-me for the current momentโ€™s purchase. Your bricoleur will pick up from here.

One reason that we can move through the commercial world in this way is that society has structured the marketplace to signal which products and services go with which lifestyles. From the very earliest moments of 20th century retail, merchants have taught customers that some things are for them and some things are not. Sears Roebuck & Co. made the idea of โ€œGood,โ€ โ€œBetter,โ€ and โ€œBestโ€ merchandise an integral part of shopping that still persists today. Each of us learned which levels of products were โ€œfor meโ€ in just about every category of goods and services.

And, we have rats and the guy whose picture heads up this issue to thank for that.

Thatโ€™s Burrhus Frederic (BF) Skinnerโ€™s picture up there. Fred Skinner was an experimental psychologist and devised what is probably the most powerful method for shaping behavior in human history. Itโ€™s called โ€œoperant conditioning.โ€

You might remember Pavlovโ€™s experiments in classical conditioning. He demonstrated that a dog could be taught to perform an instinctual behavior (salivating in the presence of food, for instance) even in the absence of the original stimulus (food) if another stimulus (a bell, for example) could be presented simultaneously with food. Eventually, the bell itself would elicit salivation.

Skinner went further. He wanted to teach animals new behaviors, not just to perform instinctual ones associated with various stimuli. So, he set out to train a rat to press a lever and receive a food pellet reward if and only if two conditions were met: letโ€™s say, if a red light was on in a box and there was no tone playing on a speaker. He then varied the conditions (including a slight electric shock on the grid in the bottom of the box) to determine which set of conditions led to the rat learning to correctly press the lever in the fewest number of trials, and to keep pressing it even after the food pellets stopped coming. Red light, toneโ€ฆdonโ€™t press. Blue light, toneโ€ฆshockโ€ฆdonโ€™t press. Think of the combinations!

Hereโ€™s a diagram of the experimental setup thatโ€™s come to be called a โ€œSkinner Box.โ€

What Skinner learned was that the fastest way to teach the rat to press the bar under the right circumstances was to dispense a food pellet when the red light/no tone combination came upโ€ฆbut not every time it came up! He started with giving the rat a pellet each time the rat pressed the lever when red light/no tone was presented. As time went on, he discovered that the fastest and most persistent way to keep the rat pressing the level was to change the ratio of correct presses and food dispensing. For the next five correct presses, the rat will get one pellet on the third try, but none on the other four correct tries. Then, three correct tries in succession will yield three pellets. None for the next six tries. And so on. The rat continues to press the lever, essentially โ€œhopingโ€ that the next correct press will yield a pellet. This method is called โ€œvariable ratio positive reinforcement learning.โ€

Weโ€™ve all experienced it. It is the fundamental behavior shaping technique that makes gambling such an enormous moneymaker for casinos. Slot machines are computerized Skinner Boxes programmed to keep the player playing by dispensing exciting positive reinforcements (money, sounds, blinking lights) in just the right amounts, after just the right number of spins, to keep the player playing as long as possible. The longer the player plays, the higher the likelihood that the casino will win all the playerโ€™s money.

Photo by Kvnga on Unsplash

Even if you donโ€™t gamble, a technique this powerful couldnโ€™t stay locked up in the desert forever. Social media engagement techniques are designed according to similar principles. As we scroll through our timelines, pausing to read some stories, clicking Like on others, sharing some, writing comments on a fewโ€ฆthe algorithms that monitor and control the items we see in our newsfeeds take every behavior into account and start using variable ratio positive reinforcement to increase time spent on the site.

Remember, keeping your attention is the siteโ€™s top objective. As long as we are on a site the chances of seeing an ad for something that might be interesting increases. Click on one of those and you have strengthened the algorithmโ€™s capacity to predict and control your behavior by showing you more of the sameโ€ฆor more of what others who have clicked on that ad also clicked on. As in the Skinner box, one of the key tenets of operant conditioning is that, โ€œpast behavior predicts future behavior.โ€

OK, but, where do bricolage and favorite things come in?

Well, what โ€œcatches your eyeโ€ in a store? Chances are certain colors, shapes, aromas, textures or brands just โ€œjump out at meโ€ as youโ€™re walking through the aisles. Think of those moments as ones in which your bricoleur is on low-level alert, not โ€œshoppingโ€ exactly, but simply scanning the world for possibilities. When an item jumps from background to becomes figureโ€ฆgrabs your attentionโ€ฆthereโ€™s a very strong likelihood that the item is similar to others that you have received positive reinforcement for in the past.

When people say, โ€œoh, you look great in that color,โ€ itโ€™s the kind of survival/reproduction-related bit of positive reinforcement that may very well lead to a color becoming a lifelong favorite. Collecting items (weโ€™ll talk about collecting people in a later essay!) that cohere with our pre-existing, reinforced, ways of being-in-the-world is the curatorial process by which we bricolage our way through life.

If you look at your favorite things in light of both classical and operant conditioning youโ€™ll probably discover that many have been chosen for their positive impact on the sense of self that they reinforce.

Next, weโ€™ll take a look at the many ways in which even the simplest objects can play an outsized role in constructing and maintaining a coherent sense of ourselves as purposeful ways of being-in-the-world. Itโ€™s why even the tiniest detail on a piece of clothing can make it one of My Favorite Things.


Tom Guarriello is a psychologist, consultant, and founding faculty member of the Masters in Branding program at New Yorkโ€™s School of Visual Arts. Heโ€™s spent over a decade teaching psychology-based courses like The Meaning of Branded Objects, as well as leading Honors and Thesis projects. Heโ€™s spearheaded two podcasts, BrandBox and RoboPsych, the accompanying podcast for his eponymous website on the psychology of human-robot interaction. This essay was originally posted on Guarrielloโ€™s Substack, My Favorite Things.

Header image via Wikimedia Commons.