Larry Barsalou is brilliant. In my view, he made one of the most important scientific discoveries in the modern era of psychological science: ad hoc concepts. It is not hyperbole to say that my own scientific efforts, in part, owe their existence to Larry’s discovery of ad hoc concepts.
Before Larry’s seminal work, a category was simply any group of instances that are similar enough to be treated as interchangeable, yet different enough from the instances of other categories so as not to be confused with them. A category simplified variation by finding similarities among differences that render them equivalent, and a concept was a summary of those similarities. Categories were assumed to exist in the world and to endure in time and space, whereas concepts were assumed to exist in an animal’s head (or inside its body if it doesn’t have a head) and to change with time and experience. To categorize, an animal was assumed to retrieve a concept to assign a newly encountered “anything” to an existing category, giving it meaning as “something.”
Beginning with his dissertation research in the 1980s, Larry discovered that categories are momentary, contextualized and situated — extemporaneous groupings whose features of equivalence are related to the functions that the instances serve in a particular spatio-temporal moment. The probability distribution of features for instances of a momentary category is highly variable, and the the prototype corresponds to the ideal instance of the category for meeting a goal in a particular situation, even if the instance is a simplification that has never existed in reality and must therefore be inferred.
This insight, that categories are dynamic events, rather than stable cognitive structures, is the centerpiece of my research on brain function as it relates the predictive regulation of the body and the lived experience that emerges in the process. In my own work, categorization is not something the brain does at the end of processing sensory signals from the world and from the body, after sensation, perception, and other forms of cognition have been completed. Instead, my work suggests that categorization is how a brain processes signals from the very start. It’s a basic operating principle of the brain, a natural consequence of its architecture. Categorization begins as the construction of an ad hoc category that is tailored to the functional requirements of the specific situation and ends as a specific category instance that guides action and creates lived experience in that situation.
Categorization, when viewed in this way, is a whole brain event with sensation, perception, memory, action, and other psychological phenomena as different facets. That the brain is a continuous category constructor dissolves the old boundaries between category, concept, and exemplar. It also has implications for the philosophy of science and the relational nature of reality.
Larry Barsalou is also a mensch. He is one of the most generous, humble people I have ever met. He regularly astounded me with his encyclopedic knowledge of cognitive science as he patiently and selflessly exposed me to a smorgasbord of topics and indulged my sometimes exuberant scientific speculations. I will be forever grateful for the years I collaborated with him.
2 Comments
Lisa, I never understood emotion or emotion research until you explained to me how emotions are constructed, along with how deeply problematic the innate emotion view is. It was amazing to see the central roles that conceptualization, situations, and embodiment play in constructed emotions. Learning to see your perspective was so interesting, inspiring, and important that I started moving away from the ‘pure’ study of concepts to a variety of applications in more social and clinical areas. It’s really all your fault how far off the deep end I’ve gone, focusing now on psychometrics and behavior change with respect to behaviors like trichotillomania. David will be talking to you…
All mental states as ad hoc categories… Interesting, it’s indeed amazing how flexible and dynamic the brain is at constructing each of its states, no one ever the same, each one contextualized by everything in the brain, body, and world at the current moment, reflecting previous learning up to that point. Perhaps a quintessential example of quantum cognition?
Yet, at the same time, I’m equally impressed by how stable the brain is. You, too, have probably noticed how hard it is to get people to change. Behavior change interventions have become notorious for how ineffective they typically are. The problem is the existing (often bad) habits that people have, along with all the conditioning that has gone into them for decades. These habits (implemented by ineffaceable dopamine circuits in the brain) certainly appear to produce a LOT of stability, perhaps too much. The only way to compete with them is to condition new habits that are more rewarding, with the potential to take over, at least for a while, more optimistically, forever. Within this context, stability and flexibility comfortably co-exist. Strong attractors with infinite adaptability? Both have high utility for survival. These days, I’m finding myself more attracted to habits and stabilities in cognition, emotion, and behavior than to their variability. Where does all this stability come from, and how do we work with it effectively, not just at the individual level but at the systems level? A lot of destruction around the world results from stable human tendencies. A lot of good results as well.
Hmmmm. That behavior is difficult to change is not necessarily evidence that the brain is stable. The brain might appear stable for many reasons. Off the top of my head: (1) the brain shapes the environment it responds to, constructing its niche in such a way as to make certain predictions (i.e., ad hoc categories) more likely than others, which has the effect of stabilizing behavior (as in the case of habits). Something can be very metabolically efficient in an immediate sense while being maladaptive in the longer-term. (2) Most experiments are designed to constrain variation and therefore construct stability; it is not discovered. The mistake is in believing that the stability is real in the brain as it functions in the wild (as opposed to in the scientific workshop). (3) All of us experience ourselves in the world through the lens of categories. All categories are simplifications . The world (including brains) are immensely intricate ensembles of signals that are statistically non-stationary. Categories are momentary reductions of complexity and variation in the service of some goal or for the purpose of achieving some function. Perhaps there is more variation than you think (even in dopamine circuits).