The idea I want to share today goes by a range of different names, connecting the school of Radical Behaviorism in psychology, to the theories of predictive processing, reinforcement learning, and active inference in the fields of artificial intelligence and neuroscience. Besides being particularly elegant, it’s also a very difficult “pill to swallow,” emotionally and philosophically, but at least in my personal experience, it’s well worth the sore esophagus.
The idea, which I’m going to call the “functional view of behavior,” is that our behavior is determined by its function – by what it does, or more accurately by its past outcomes. This is essentially an alternative to the more familiar idea of intention… tossing aside the issue of “free will” like a snotty tissue—which makes this idea easy to dismiss if – like me – you happen to be attached to the idea of self-determination. But, what makes this idea easy to discard is also what makes it so useful—namely that it does not typically match with our subjective experience of our own mental process, and therefore offers different insights than our intuition might suggest. I think for probably the same reason, many people seem to be much more comfortable with so-called “cognitive” theories of human psychology – which emphasize the explicit verbal and conceptual processes that we most readily identify as human – rather than the mechanisms of behavior we share with much of the rest of the animal kingdom. But, if you can accept that “consciously deciding” isn’t all that’s going on with your behavior, I think you might find that especially in relationships, we might be more animal than we’d like to think.
Backing up, at their most fundamental, relationships are made out of behavior—specifically the behavior of the participants in a relationship toward – and about – one another. When I say “behavior” here, I’m talking about everything you do, including things like speaking, thinking, imagining, and remembering. Everything you do has an outcome, (even if it’s only in the environment of your own mind) and to the extent that you unconsciously learn from past outcomes, your experiential past becomes the cause of your behavioral present, or that’s the idea, anyway.
The functional lens asks me: “What is your behavior doing?” rather than “what are you trying to do right now?” If we’re fighting, I might be intending to “express my feelings” (nothing wrong with that, right?) and you might be intending to “correct my misunderstanding” (I think we can all agree that accurate communication is important, can’t we?) We might notice, though, that my expression of feelings and your correction of misunderstanding are functioning to escalate the conflict… not to mention our respective blood cortisol levels. By understanding how these kinds of negative patterns emerge and persist, we may be better able to drive a wedge somewhere into their mechanism.
Reinforcement
Let me back up. The “functional view” as applied to human behavior is most famously represented in the work of B. F. Skinner, who conceptualized “learning” (technically: change in behavior over time) as a process analogous to that proposed by Darwin—namely one of selection (Skinner called it “selection by consequence”) akin to what Darwin called “natural selection.” Both describe a process where the environment “selects” features of the organism—be they behavioral or physical—encoded in the organism’s genome and nervous system, respectively. In natural selection, organisms are selected based on their success in surviving and reproducing, with the genes carried by organisms that succeed becoming more numerous in the larger environment. In the functional view, behavior is selected based on its outcome in the environment, with certain outcomes (generally those beneficial to the organism) making the behavior more likely in the future. Looking at it this way, both genes and behaviors are selected by the environment over time based on how well they work—the basic mechanism is: try stuff, see how it goes, and then do more of the kind of stuff that works well. Tying that together, both evolution and “learning” are processes of selecting by function.
So, the basic idea is that once you understand a behavior’s function, you understand the “reason” the organism is behaving that way: because that kind of behavior has been “reinforced” in the past. When we think of organisms adapting to their environment, it’s not the organism shaping itself based on what kind of body it thinks will work well in the future, it’s the environment shaping the organism based on what has worked so far.
Both evolution and learning are also fundamentally incremental processes. In evolution, when there’s an ecological niche to be filled, it’s not as if the environment simply fabricates a species to fill it from whole cloth—the new organism must adapt from existing species. Wings are reworked legs or fins, not an entirely new invention. Behavior shaped by function works the same way—each new layer adapted from previously learned behaviors. Selection and adaptation are not linear calculations of the one correct “solution” nor brute-force calculations of the optimum among all possible options, they are gradual processes of iterated adaptation.
This runs counter to our intuition about human behavior—we generally expect ourselves to “realize” something, and then behave, perceive, and feel categorically differently. The upshot is that – to the extent that your behavior is controlled by “functional learning” – you can expect behavior change to take many repetitions even after you “know better.” Especially in relationships, this can be a critical insight, not least because we often expect our partners to feel things based on their beliefs about us and the situation, and get disappointed when what they actually do feel turns out to be more about the past than the present or anticipated future.
Just to geek out for a second, I find it intuitive to think of both bodies and behaviors as “constructs” resulting from coded information in the DNA molecule and the neural structure of the brain, respectively. Where the genome represents a sort of “recipe” for building a body, the “connectome” (state and structure of the brain and nervous system) represents a program, or more generally, “computational process” being run in the brain. Both kinds of code are shaped by their contexts, with the code itself being continually “re-written” based on its outcome in the environment. The implication is that the solution evolution “engineered” to accomplish learning is mechanically very similar to natural selection itself, though it occurs over the timescale of an individual organism’s lifespan, rather than the successive generations of evolutionary time.
The Relational Environment
The other big shift (besides shifting from the anticipated future to the past) is from looking at behavior as a product of the self to the environment. In talking about the function of a behavior, we’re referring to the function in the environment, or more generally in context. The same behavior – say, moving your arm in a particular way, or speaking a particular sentence – can have entirely different functions, depending on the context in which it happens. This is “survival of the fittest” in evolution, where “fitness” is determined not simply by the structural “shape” of the organism, but by its relationship to its environmental context. A shape that’s highly “fit” or adaptive in the desert is likely to be “unfit” or maladaptive in the rainforest.
In an interpersonal relationship, “the environment” is – for the most part – the other person, and their behavior. As a foundationally social species, this kind of environment is precisely what we’re most specialized and adapted for, in that much of what we have evolved for is to effectively predict and control other people’s behavior. So, if you and I are in a relationship, a primary function of your behavior is to prompt and shape my behavior, which in turn functions to prompt and shape yours. We simultaneously control and are controlled by one another. As such, most of the relational phenomena we find problematic enough to devote significant attention to, are co-constructed—meaning that both partners are involved (even when one or both of us are sure it’s all the other person’s fault).
When thinking about either evolution or learning, we have a tendency to conceptualize the environment as relatively static, while the organism is the thing doing the adapting—evolving or learning. But, in both cases, (at least outside of laboratory conditions) it’s generally more accurate to see the whole system as engaged in a process of mutual adaptation: coevolution in ecosystems, or “reciprocal learning” in relationships, where my behavior is inevitably shaped in the process of shaping yours. If nothing else, this might be a good justification to let go of the fantasy of finding the “right thing” you can do to get your partner to behave the way you want them to, and accept the reality that any desirable state of relation is going to be a moving target.
Natural ecosystems represent interconnected networks of functional relationships, in the sense that the function a given organism serves in an ecosystem is dependent on its relationship to other species. A tree functions as many different things to many different other species—a trellis for vines, a competitor for other trees, an oppressor of grass, a home for birds, a source of wood for humans. Analogously, a given behavior in a relationship—say, detailing the events of my day to my partner, could have a range of effects on other behaviors, and also be affected by a range of different behaviors. It might encourage reciprocal sharing from my partner, prompt emotional disengagement, and might be made more or less probable by either “irritable” or “warm” behavior from my partner beforehand. It could be a facilitator of emotional intimacy, or a way to conceal a more emotionally pressing and relevant conflict I don’t want to talk about.
Zooming out a bit, one could look at the primary function of an organisms’ behavior as being the extraction of whatever is necessary for their survival from their environment, be that nutrients, water, sunlight, etc, and keeping that environment within parameters compatible with their survival (often just meaning that they move to locations that keep them alive and away from locations that are likely to kill them). Given that relational partners represent the primary environment for humans, it’s your partner(s) from which you must ultimately extract those “necessary resources,” (whether literally or figuratively) and whose behavior you must modulate within appropriate parameters. To put that in slightly more intuitive terms, your behavior in a relationship can be thought of as simply a description of what it takes to get your needs met by your partner.
When two people link up in some kind of relationship, and begin to get to know each other, their behavior towards one another immediately begins “co-evolving” mutually adapting continuously in a particular and unique “way to be together.” So, partners can be expected to shape one another’s behavior into a custom-designed system for getting interpersonal needs met by one another. It’s not necessarily the best or optimal way to be together—selection just finds something that works, and there’s no telling what kind of unwanted or sub-optimal effects might come along with the solution at which you arrive together.
Further contributing to that sub-optimality is the fact that as our behavior is shaped by our relational environments over the course of a lifetime, we naturally apply behavior learned and shaped in one relationship to subsequent relationships. Like an organism transplanted from one environment to another, it’s common for us to find ourselves behaving in ways adapted to past environments and past partners that no longer “fit” our present relationships. A child who grew up in a home where all their wishes were indulged with minimal effort on their part may find their behavioral strategies suddenly ineffective once they try to live independently. Similarly, someone who learned how to survive in a psychologically dangerous environment might be expected to struggle to “let their guard down,” even if they find themselves in a safe social context. Much of what we see as problematic or “maladaptive” behavior, in other words, might be best understood as behavior adapted to a different relationship than the one the person is in currently. Knowing something about the environment to which a given behavior was originally adapted in comparison to the kind of relationship you’re in (or want to be in) now, may offer insights about necessary environmental changes to meaningfully affect the behavior.
Action and Perception
Behavior occurs in a sort of cycle where what a behaving organism does is based on what it senses about its present environment—perceiving, then acting, then perceiving, and so on. Functionally speaking, the effectiveness of any action is entirely dependent on its context—the situation the organism is in when it performs that action. If you want to keep your body within a temperature range conducive to remaining alive, for instance, you naturally need to do opposite sorts of things depending on whether the environment is currently too hot as opposed to too cold, which depends on your ability to perceive your current temperature.
Pragmatically, we go around alternating the questions “What situation am I in?” and “What should I do about it?” with each answer informing the next question. Perception categorizes the current environment into a “kind of situation” which leads to doing the “kind of thing” that has been effective in that situation in the past. It might be a bit counterintuitive to think of perception as a behavior that’s learnable and shapeable, because our subjective experience of our perceptions is that they directly represent reality. On a functional level, though, what we experience as the world, is actually a very personal, idiosyncratic guess about what’s most likely out there and what is most likely relevant to our actions—a guess that is being arrived at on a subconscious level, which largely determines what we do. In other words, perception is creative and subjective, and it’s a particular way of piecing together the sensory input data you’re working with at the moment.
In considering relationships, it’s important to take into account the functional implications of perceiving things a particular way, especially given how subjective, ambiguous, and “up for interpretation” human interactions (especially conflicts) tend to be. Simply, though, expect yourself and your partners to tend to see things in a way that serves a necessary function, or otherwise confers some functional advantage. This should not be surprising, given how much weight our social and interpersonal structures tend to place on the definitions of situations—how a given interaction is perceived. This is perhaps most starkly illustrated in our justice systems, where questions of perception like “was this murder or self-defense?” have enormous practical consequences.
Even in relatively cooperative everyday interaction, though, there is often significant conflict between how each participant perceives what’s going on: one person thinks they’re just being friendly, while the other perceives flirtation; or one sees an attack while the other is “just joking.” In discussing our perceptions in relationships, we’re often preoccupied with the idea of accuracy, asking “which perception is right?” when it is often more important to consider the function of the perception: “what is the effect of seeing it that way?” or “What am I getting out of seeing it this way?” Much unproductive conflict in relationships is waged over which view of the situation is to be accepted as truth, when each participant has material incentives to insist on their own definition. As Upton Sinclair famously put it “It is difficult to get a man to understand something when his salary depends on his not understanding it.”
Further complicating the interplay of action and perception, is the issue that actions often function to generate sensory information to inform further perceptive behavior. Interpersonally, we often take actions in order to learn things about our partners and their current state, the most obvious example of which would be asking questions. However, verbal questions are far from the only behavior in relationships that function primarily to generate information about your partner’s state, and unsurprisingly, how you go about information-gathering can have substantial impacts on the state of the relationship overall. Humans seem to be particularly fond of experiments—behaviors emitted specifically to see what will happen—specifically to test how the environment will respond. If I do something annoying to check whether my partner’s in a bad mood, there’s a good chance I will wind up with a partner who’s in a bad mood—and if I exhibit problematic behavior to test the strength of the relationship, especially over time, the “stress-testing” is likely to weaken the relationship.
Returning to the evolutionary analogy for a moment, both action and perception can function to select environments to which they are already adapted, like an animal migrating to a location to which it is better adapted than the one where it started. Perceptively speaking, we try to find environments we are good at understanding, as well as good at acting in. Among other things, this can lead us to unconsciously prefer familiar relationships and situations, even if they are more painful or less valuable than available alternatives.
Virtual and Verbal Behavior
When considering the action-perception cycle in relationships in particular, a lot can be summarized with the quintessential human behaviors of speaking and listening, or in other words, communication. What I mean is that a lot of our “active” behavior in relationships comes down to saying things to partners, and our “perceptive” behavior has a lot to do with our listening to – and interpreting – what they say. The human capacity for communication goes far beyond verbal language, though, to the extent that some theorists would assert that all behavior is communication.
Related to the use of language is the human capacity for what Skinner called “private behavior,” covering things like thinking, imagining, and remembering. He conceptualized these sorts of behavior as simply “private” versions of “public” behaviors like speaking, moving, hearing, or seeing—and similarly governed by reinforcement. This view sees thought as literally “saying to yourself,” or “hearing to yourself,” and imagining as something like “seeing to yourself.” The private versions of active and perceptive behaviors collectively represent something like a simulated world, complete with a simulated point of view that we identify as a self, and where we behave and experience virtually without “really” behaving. Ultimately, the behaviors of symbolically representing, communicating, and simulating things are all part of the human organism’s functional relationship to the relevant environment, rather than being some disconnected mental universe—they are an integral part of how we arrive at effective behavior in the physical world, and based on input from that physical world.
In a relational context, symbolic language and simulated behavior together enable us to do what functionally amounts to sending and receiving virtual experiences with other minds—representing features of the world symbolically, and then communicating those representations to one another. With language, we conjure simulated experiences for one another—which may then prompt and shape behavior like any other environment. When I listen to you talk, I’m constructing a simulated world of what I think you “mean”—how you see me, yourself, and the environment. When I speak to you, I’m trying to condense or represent some version of my own experienced or invented private world that you will be able to “reconstruct” based on our shared history and meanings associated with the words and symbols we use. As an evolutionary adaptation, this provides our species with the huge advantage of being able to effectively “share” reinforcement history from experiences and environments with which we’ve had no direct contact, benefitting from costly lessons learned by others, as well as potentially valuable perspectives outside our own—but also ushers in its own range of problems and complications.
Verbal behavior allows us to construct and utilize mechanisms like rules, instructions, and beliefs—which can effectively function as an efficient distillation of an otherwise long and painful reinforcement learning process, with some drawbacks. A rule like “don’t touch the stove when the red light is on” alleviates the pain associated with “learning the hard way.” A well-written recipe can stand in for years of trial-and-error culinary disasters. One issue with these kinds of verbal constructs, though, is that they can prevent one from learning from experience. One downside of beliefs, rules, and instructional/procedural “programs” is that they have a tendency to persist well past their usefulness because they are often reinforced even when not directly or optimally functional. When you’re following a rule, you often stop “experimenting” and you might not realize there’s a better way to do it. How big a disadvantage this is, depends on the circumstance. In relationships, implementing rules like “never talk about your feelings again” or “only apologize if they do first” might be a useful shortcut through painful experience, but may be just as likely to prevent you from enjoying potential benefits of the relationship.
Verbal agreements, in particular, represent one of the most powerful tools for reshaping relational environments that have evolved in directions their participants are maybe not so happy about. By making—and keeping—an agreement such as “we don’t have sex with people besides one another” partners are able to intentionally alter the effective environment of their relationship and potentially “steer” the evolution of a behavioral system that might otherwise arrive at solutions that aren’t particularly good for anyone. Verbal behavior enables humans to agree to behave as if the world were different than it would be otherwise, and radically reshape the development of behavior in the system. There’s more discussion of the concept of story to come in future posts, but for now I just want to note that the stories we tell ourselves and our partners about what’s going on between us, why it’s going on, and what it means for us all serve functions for us and our relationships, and therefore may be selected based on that function like any other behavior.
Application
The “functional” lens I’ve been texplaining sees both the cause and the effect of behavior as a relationship between the behavior and the state of the environment in which it occurs; or in other words, seeing behavior in terms of what it does. This means looking at your behavior in a relationship the way an ecologist would understand an organism in its natural environment, i.e. in context, and in terms of its survival needs and its functional relationships in the larger ecosystem.
To restate a bit, what a behavior does is a different (if often related) question from the questions we’re more familiar with answering about behavior, such as what it is intended to do, or its moral justification, or why it was the only rational thing to do at that moment, or how one consciously perceived their environment at the time they behaved that way, or how the behavior is a product of the “personality” or “character” of the person doing the behaving. This means asking questions leveraging our understanding of functional relationships, perhaps the most basic of which is: What happens next? or more specifically: What has happened after this kind of behavior in the past?
While these might be simple questions to ask, they’re more complicated to try to answer for a range of reasons, one of which is that an infinite number of events happen after any given behavior. I could be looking at what happens within myself—what I do next—or what my partner does next, or what happens in our physical environment, or the overall effect on the relationship. This is where some experimentation, creativity, and intuition are called for as you try to figure out what the relevant function outcome might be.
For example, if you’re confused about why you feel “helpless” when your partner is around, we might notice that your partner typically treats you with additional care and consideration after your expressions of overwhelm, which might function in your relational context to bring you and your partner together, or to reassure you that you are loved and cared about, or give your partner a validating sense of being strong and necessary. You can make these kind of guesses by observing what happens after the behavior in question, and asking yourself what you, your partner, or the relationship might be “getting out of” these outcomes. In contrast, the intuitive answer to “why am I crying?” would almost certainly be some version of “because I was sad” or “because a sad thing happened.” If, for whatever reason, you were interested in crying less, the intuitive answers might turn out to be less helpful in suggesting potential courses of action for changing the behavior.
Then there’s the question of what about that outcome is reinforcing. Even when the overall outcome is obviously undesirable, there may be some part of it that serves a necessary function—that’s needed by one or both of you—or at least is preferable to the known alternatives. In other words, a behavior’s outcome only needs one “necessary” function in order to persist. This functional outcome can also be removed from the behavior in question by some amount of time, and any number of intervening behaviors, as when a behavior with decidedly problematic immediate outcome is embedded in an overall pattern with a desirable outcome. An example might be a couple whose most fulfilling and meaningful cooperation only occurs after long, painful, and otherwise-unproductive fights, and functions to reinforce the fighting.
We often have well-thought-out ideas about what sorts of behavior we want from ourselves and our partners, but these desirable behaviors end up being reinforced less effectively than more undesirable behaviors. Thinking ecologically, for “good” behavior to predominate, the relational environment must be one where good behavior is capable of out-competing bad, which is far from a foregone conclusion. Since the “fitness” of an organism or a behavior is contextual—it’s about fitness in relation to environment—a principal concern when you notice that you and your partner are behaving “badly” a lot of the time, is work to understand what is making your environment more hospitable for bad behavior than good. In other words, what is it about your relationship as a whole that is making it a better environment in which to behave uncooperatively than to work together in good faith? A relationship with a long history of mean-spirited arguments is likely an inhospitable environment for “vulnerability,” for example.
Alternatively, if you’re confused about why your partner is behaving the way they are, you might ask yourself what sort of situation that their behavior would be the optimal response to, and try to understand how you might be contributing to their perception of your behavior as constituting that sort of situation.
In practical application, one insight is that in order for unwanted behavior to go away, it’s often necessary to replace the necessary function that the unwanted behavior was serving—for which an imprecise but practical shorthand might be “get the same need met in a less costly way,” or “get the same thing done with less collateral damage.” A commonly-cited example of this general psychological process is that the chemical addiction to nicotine is often secondary to other functions of the unwanted behavior of smoking, such as serving as a reward for accomplishing a task, as a marker of time periods in the day, or a reason to go outside or socialize with other smokers. Replacing these functions, for example, by finding other “rewards” or “time markers” or “reasons to go outside”—is instrumental in the task of quitting smoking. To give a relational example, if you want your partner to trust you, you’ll need to find other ways to assert your freedom besides lying.
The inverse insight is the idea of trying to disrupt the reinforcement of an unwanted behavior, by intervening in the behavioral process so that the same behavior has a different, generally “non-reinforcing” outcome. There’s no set formula for how you might go about this, but one example might be asking my partner to stop responding to my self-deprecating comments by telling me how amazing I am, if I suspect my self-hatred might be being reinforced by the praise.
On the other hand, if there’s a particular way you want to be behaving, or you want to behave in a particular way more often, making that change sustainable tends to depend on your ability to make that behavior enjoyable or necessary besides just being “right” or “healthy” or “what your partner wants.” Behaviors and patterns which are not reinforced will not persist—for healthy behavior to take root, it has to work. Again, there’s no formula for this, but if you and your partner can make and keep agreements to—for example—start “fighting fair” at the same time, you’re much more likely to sustainably change conflict patterns than if you each just occasionally drop an I-feel statement into the middle of an insult-battle. Similarly, a shared, concerted effort to bring more honesty to the relationship is more likely to take root than if one or the other of you just occasionally tries out the tender seedling of “emotional intimacy” in an otherwise barren landscape of polite, withdrawn distance.
It would be absurd to completely ignore the role of (what we at least subjectively experience as) conscious decision making in human relational behavior. If you had no power to affect your own behavior, there would be no use in trying to change anything. Adding the functional lens to your repertoire, though, can be instrumental in identifying where to focus your finite willpower, understanding what you would actually have to do to make a sustainable difference in the ongoing, unintentional patterns of relational interaction. Rather than giving up on self-determination, in other words, I see an understanding of the functional account of behavior as potentially giving you more leverage to steer the course of your relationships with intention.
Viewing relational interaction functionally moves us from the idea of a rational dialogue, or even a process of explicit communication of information, to a cycle of mutual cause and effect, where the stream of behavior emitted between two people – a relationship – is a living, evolving thing, organizing itself into a shape and structure apart from the participant’s conscious, subjective experience, though intimately interrelated with that experience. Turning my attention to the effect I’m having on my partner, and vice versa, how we’re shaping one another continually, how our actions, reactions, and perceptions work in the network of behavioral influence—often gets at something fundamental and useful about what’s going on between us.
In any case, I hope I’ve at least given you a feel for how to use the functional lens to understand behavior. Perhaps the most useful thing is just to have this in your back pocket, to “try on” for a given problem or pattern you observe in your relationships—sometimes it doesn’t tell you anything beyond what an intuitive analysis does, but sometimes it reveals a whole different story than what is going on on the surface. Being able to see the behavioral ecology of your relationships might just start to illuminate some ways forward if you and your partner find yourselves scratching your heads about why you just keep doing the same stupid thing.
Sources
Gregory Bateson (1979) Mind and Nature: a necessary unity
Mecca Chiesa (1994) Radical Behaviorism: the philosophy and the science
Richard Dawkins (1976) The Selfish Gene
Yuval Noah Harari (2014) Sapiens: a brief history of humankind
Douglas Hofstadter (2007) I Am a Strange Loop
B. F. Skinner (1953) Science and Human Behavior
Mark Solms (2021) The Hidden Spring: a journey to the source of consciousness
