The Science of Function

Discussing function as an entity can lead to some confusion. As with physical things, we can name functional things, such as 3D vision or justice. The names are not the things themselves, but if we understand the reference then we know the function being discussed. As with physical things, sometimes instead of explicitly naming a function we will describe it, and the description then acts as a reference to it, still without being the thing itself. What gets confusing is that a name or a description is functional itself. Naming something is a way of declaring it a concept, and describing it is a way of attaching traits to the concept, whether named or not. Concepts and traits are intrinsically functional, where the function of concepts is to help us group similar entities under one name in a general way. So while the name or description of a function, like 3D vision or justice, is not, as I said, the function itself, it serves its own function, namely to name or describe another function. So Boyle’s Law and descriptions of it are not the function themselves, but let us discuss it.

I mention this because our descriptions of things are always sketchy. First, they always presuppose a great deal of context which is presumably understood and agreeable. Then, they focus only on the most salient aspects in the hopes that the rest will be implied. For physical things, one can presumably see the object and make other physical observations that provide deeper understanding far beyond the verbal description. Most functional things also have instinctive and subconceptual functional support, such a 3D vision and justice. So while we can’t see them in the physical world, we can experience them as functions in our mental world to understand them beyond our verbal descriptions. Physical objects can be objectively described through measurement by instruments (which at least push subjective factors back one level), but it is harder to measure functional things objectively. Still, techniques are possible, such as the observation of the effects of function, e.g. of behavior.

Herein I am only describing the mind, that is, its overall and composite functions. I will, therefore, be sketchy, but I am hoping that a great deal of context is understood and agreeable, which should bring what it is into much better focus than before. Especially considering viewing the mind as a functional entity is a somewhat new perspective, I will start from the top down with the most salient aspects and fill in more detail as I go. My explanations will appeal to and depend on our experience of mental function, which is to say how we subjectively experience our own minds. This approach is introspective, which a challenge to objectivity. I will address that challenge in more detail later, but in short I will look to introspection to stimulate hypotheses, not to test them. The resulting descriptions of mind I develop will constitute a theory to be tested. Like all theories, it is not intended to have the same function as the mind or to be complete, only to be internally consistent and supported by the evidence. I intend to show that it is consistent with prevailing scientific perspectives once those perspectives are interpreted in the framework of form and function dualism.

All scientific theories are descriptions, and hence sketchy representations of reality. Boyle’s Law describes the relationship between volume and pressure of a gas at constant temperature. It is “sketchy” even though it seems to work perfectly because volume and pressure are approximate measures of reality based on instruments and not fundamental properties of spacetime. But we can measure volume and pressure almost perfectly most of the time, and in these cases the law has always worked to our knowledge and so we can feel pretty comfortable that it always will. Though unprovable and sketchy, we can label it “true” and depend on it without fear. Of course, to use it we have to match our logical model which models gases as collections of independently-moving particles with volume and pressure to a real-world circumstance involving gas, which depends on observations and measurements to provide a high probability of a good fit of theory to practice. This modeling, matching, observing, and measuring involves some subjective elements which have some uncertainty and vagueness themselves, so all objectivity has limits and caveats. But we can accurately estimate these uncertainties and establish a probability of success that is very close to 100%.

In principle, theories of function can also be almost certainly true. We have to start with what we are most certain about and be careful about introducing less supportable hypotheses. The existence of function itself is the first thing about which we need certainty. We have more direct access to function than to form (“I think therefore I am”), but our scientific explanations of form have so much coherence, i.e. corroborating evidence, logical consistency and explanatory power, that they set a very high bar. We probably can’t achieve a level of certainty or objectivity about function that is as high as we can for form, but we can still do pretty well because we have a number of objective and scientific sources of information about function. It is not all opinion.

So in what sciences is the subject matter functional? The physical sciences strictly study form because nonliving systems don’t have function. The formal sciences strictly study function, because they are not physical and function is all that is left. Though they don’t study physical form, they are named after formal systems, where “form” means well-defined. The formal sciences are indirectly functional; they specify and follow rules, and following rules is a good way to achieve functions, though not the only way. The biological and social sciences study the functions of living systems. So three out the four branches of science study function. The physical and formal sciences have achieved significantly more definitive results than the biological and social sciences. Physical laws can be tested in many ways with great precision. It doesn’t mean physics is solved; general relativity improved on Newton’s law of universal gravitation, and MOG (MOdified Gravity) may improve on general relativity. But we can predict physical phenomena with great accuracy and reliability, which has led to very empowering technologies. The formal sciences have been instrumental to that success, as physical sciences and technology depend heavily on mathematics and computer science. Significantly, our objectivity in the formal sciences is in many ways perfect. Knowledge within a well-defined formal system can be known for certain and provably so. While some define objectivity as being based only on facts and evidence, which is an empirical standard, for functional matters the broader definition of being without of bias, judgment, or prejudice is more appropriate. Any formal system in and of itself is necessarily objective by this standard, but the motivations for creating the formal system that way in the first place are necessarily subject to some degree of subjectivity, as are the methods used to apply a formal system to a given problem. So we can conclude that the study of formal function has met with great success.

Living systems, however, quickly introduce complexities that make the levels of certainty seen in the formal and physical sciences seem very distant indeed. Function in living systems derives from iterative mechanisms that refine knowledge and strategies over time. By construction, then, it is dynamic and adaptable rather than fixed and repeatable. However, many creatures in Earth’s history remained nearly unchanged for millions of years, suggesting that their function was accurately and reliably tuned to the needs of their niche. Human tools, too, sometimes remain unchanged for generations, like the dagger and paper clip. So it is not that elements of living functional systems can’t be fixed and repeatable, it is just that the playing field is much larger. The universe has just one physical system, one set of rules, but every lifeform is designed differently to its own set of rules, on a playing field where the design specifications often change. Much more than that, most physical systems work with no or little feedback, but most functional systems adapt their reactions continuously from feedback. So nothing about the design and not too much about the behavior of functional systems can be explained or predicted by equations. We consequently call the physical sciences hard science can call biology and the social sciences soft science, which makes them sound inherently wishy-washy. This is unfair; they are trying to explain much more complex phenomena. But they do need a firmer foundation: they need to acknowledge that they are built on function, not form. They also need to develop stronger ideas about what function means in living systems. We can never achieve the same precision with the life sciences that we can with formal and physical sciences, but precision is not really the goal. Function is all about capacity, not precision. If we understand what is functionally possible and why, we will be able to say what we can do and should do.

This raises the question, what is the best scientific method for studying the functional systems of living things, or, more broadly, all non-formal functional sciences? Is it definitely not the same method that has been refined for the physical sciences. The physical sciences produce fixed laws of action while functional sciences produce models of capacity. Physical laws can be tested with precision, but models of capacity are less tractable. Yes, we know exactly many of the capacities that math and computer science bring to the table, but living systems were not designed to solve idealized problems but to survive in a complex, interconnected environment. Unlike physical laws, which remain unchanged across billions of years, the functions of living things are being interpreted within biological niches which are potentially changing all the time. Positive and negative feedback in living systems has spurred the evolution of systems almost infinitely more complex than nonfunctional physical systems. Rather than hard and soft, the division in the sciences should be characterized as fixed and variable. The fixed sciences are necessarily more tractable and knowable than the variable sciences, but the variable sciences can yield almost infinitely more designs capable of almost infinitely more functions.

Does this mean we need to revamp the scientific methods used in the soft sciences? Not exactly. The methods have evolved to right about where they should be despite the lack of a firm philosophical footing. But yes, I will shore them up and it will highlight some shortcomings. Lacking a clear vision, the soft sciences have more or less adopted the same scientific method used by the hard sciences. (Note that I am not using the line between social and natural sciences because biology is a functional science and hence soft, despite the information captured by genes being more fixed than the information of minds and societies). The basic elements of the method are: observe, hypothesize, predict and test. Peer review and anti-biasing techniques (like preregistering research) have been added to control human factors. The steps are iterated as often as needed to refine the match of model (hypothesis) to reality (as observed). Feedback loops like this are the source of all function and information, but the iterations of the scientific method put a premium on truth and not just utility. Scientific truth is the quest for a single, formal model that accurately describes certain aspects of reality, whereas useful information doesn’t have to be modeled (via experience or intuition), and when it is (via logical reasoning) can include a variety of competing models. Most of our “common sense” knowledge consists of likelihoods of this sort rather than very specific models that attempt to make exacting predictions, often according to precise rules of cause and effect. Although we can’t prove that a scientific model is correct because our knowledge of the physical world is limited to sampling, all particles of a given type seem to behave identically, so we can effectively prove correctness in the physical sciences. The exact rules needed are still a bit too complex for us to nail down completely, but what we have devised so far works well enough that we can take it as true for all intents and purposes in the range of covered circumstances.

But how did science arrive at a consensus on general relativity and quantum mechanics as the theories to back? Their preeminence as the most correct theories available right now stems from science’s commitment to objectivity. Although our minds are inherently subjective, both by definition and simply because we must filter all our knowledge through our mental lens, science has given us a big appreciation of the value of objectivity. Science, and more importantly the technology it makes possible, has completely transformed our lives, giving us power over our circumstances that let us live as gods relative to our forebears. It owes this success to its ability to distinguish well-supported hypotheses from poorly-supported ones. This begins by instilling students with a belief in scientific truth, which is a bit ironic considering science is all theory and no truth. But the essentially perfect ability of the prevailing theories of physical science to predict the future seems about as reliable as mathematical truth, which is completely certain, so we take them as ever-so-slightly-qualified truths. Next, when practicing the scientific method, scientists know that they have to formulate any hypothesis in such a way that it is supported by all the relevant prevailing paradigms. If you want to flout a paradigm you had better be prepared to provide a better paradigm, which is a much higher hurdle to cross than just refining existing paradigms. It is for this reason that Thomas Kuhn felt that paradigm shift was the hardest but most critical step behind scientific revolutions. Well-formed and tested hypotheses are then scrutinized through peer review, which happens formally at publication time and also informally through the individual and collective opinion of the scientific community. While not perfect, these efforts to increase objectivity do usually result in the endorsement of scientific laws and principles that are more functional, i.e. carry greater predictive power by modeling more situations with greater accuracy than their alternatives. Our personal attempts to gather information in the world are nowhere near as exacting or detailed, though in practice they work pretty well considering that we deal with many situations of greater functional complexity than those addressed by science and we have to act with less study.

But the soft sciences are of more interest to us here. Do we even have prevailing paradigms, and if so, have they even been fleshed out enough for their boundaries to be well understood? Let’s look first at the sub-mental biological sciences, as anything about minds must build first on that. By now evolution is well-established as the shaper of all lifeforms, even if many details of its exact mechanisms are still uncertain. Most notably, the evidence and mathematical models support both punctuated equilibrium and gradualism as valid mechanisms, so we can’t say why evolution is sometimes very fast and other times very slow. But millions if not billions of observations now support evolution and no observations discredit it. Even so, the paradigms of biology don’t quite fully embrace evolution. The drawback of evolution is that it comes with too many caveats. Scientific models work best when they are simple and closed, making clear predictions in known circumstances. Evolution, on the other hand, is messy. Every individual is unique, every gene comes in many variants, and no gene can be said to have just one function that is crystal clear. It is just more practical to sort of put evolution on the back burner, as it were, and teach biology with simpler models, focusing on the predominant purposes of cells, tissues, organs, and organisms. This approach, while not as foolproof as physical science, is amazingly effective for many purposes and enhances the biology’s credibility. However, to some degree it hides biology’s dirty little secret, that it is a functional, aka soft, science, under the rug. But biologists need to be proud of this and need to shore up their metaphysical foundation by declaring that all function in the universe starts with and depends on biology, the first science of function.

Now, as much as I hold biology near and dear to my heart, to the point where my first ambition was to be a geneticist, my interest here is with the second science of function, the mind. From there I will eventually move into the tertiary sciences of function, the social sciences, which are secondary ramifications of minds. (Note that I haven’t forgotten formal science (including math and computer science), which I call the zeroth science of function as it needs no empirical evidence. But our access to it is only through our minds, so it is derivative from my perspective). The question is whether we have a prevailing paradigm of the mind. Interestingly, it was the absence of paradigms in the social sciences that led Kuhn to recognize the existence of them in the natural sciences and then write his seminal book. And to this day, the social sciences are not bound by any paradigms that could act to prevent the “controversies over fundamentals” Kuhn noticed1. Each subfield of each social science is instead built on a school of thought which has been developed taking certain assumptions for granted that support its theories. So long as their theories are not overtly incompatible with science at large, their foundation is taken to be adequate. Despite the absence of a firm foundation, social sciences can still claim to be scientific if they can reliably predict outcomes with better success than chance alone. Many theories can hit this bar given even a very vague or biased perspective because high-level patterns appear in all systems despite ignorance of their deeper structure. The social sciences detect patterns without having to explain all the forces that cause them. But although the social sciences are practiced without paradigms today, I believe we can gradually change that by developing cognitive science.

This brings me back to the question, “Do we have prevailing paradigm of the mind?” We have nothing as concrete or broadly accepted as is found in the natural sciences, but there is much, taken primarily from our knowledge of biology, that is nearly settled. So then, whether it is a paradigm or not, there are at least some strongly prevailing views, which I will now present without further ado. The mind is a process of the brain that proceeds neurochemically, collecting information using senses and producing actions that control our bodies. We have a single train of thought or stream of consciousness running internally that sometimes appears to us subjectively like a movie running at live speed. We can rather easily swap between trains of thought we have set aside, giving us some facility for multitasking. Without conscious effort, memories and hunches come to us from stimulation of our senses or relevance to our thoughts. We have preferences and emotions that motivate us, i.e. influence our decisions. We know quite a bit about the specific range of our senses and physical capabilities. Humans also have some highly-developed mental skills, most notably language, which we know are largely genetic. These skills also include thinking, imagination, creativity, recognition, appreciation, sense of agency, theory of mind (the ability to see agency in others), and more. We acknowledge these skills and what we can do with them, but we can’t say how we came by them or how they work. We find ourselves in the curious circumstance of being able to use our minds but not knowing how they work. It sounds like a problem warranting immediate attention, and yet we have gotten by just fine without knowing. But there is undoubtedly value in knowing. As technology expands the scope of our physical activities, a better understanding of how and why we act are probably our best tools for preserving ourselves.

While the above views are widely accepted by scientists based on a huge amount of evidence across many fields, they lack the kind of firm foundation that one would expect of a paradigm. My sole ambition here is to elevate these prevailing views to the status of an actual paradigm for thinking about the mind. Nearly everything posited above about the mind came from introspection. We struggle to develop more objective explanations for consciousness, memory, thinking, language, etc., but without our subjective awareness of these things and our having observed them in others, we would have no idea that they even exist. So if it just comes down to hearsay, how can we speak scientifically about their existence? Is there any way to see the man behind the mirror? I am saying that there is and that it starts with our ability to see non-physical things, namely functions, for what they are. By addressing function as the elephant in the room we can start to develop scientific conversations about mental states, whose physical basis still eludes. Once again, keep in mind that the physical basis of thought will not prove sufficiently illuminating once we have figured it out because the physical can’t explain the functional. Function leverages but is not caused by its underlying physical mechanism; it is caused by the reinforcement of negative and positive feedback to produce a functional outcome. So the study of function in natural systems has to be based on just what kinds of functions feedback will create.

Minds are specifically designed (taking this feedback cycle as the implicit designer) to control bodies, but not so much at the cellular level as at the aggregate level, which in higher animals has come to mean devising and prioritizing a single stream of coordinated actions. In principle, we can reverse engineer how minds work using evolutionary psychology. As it is currently practiced, however, evolutionary psychology has drawn a conservative boundary that includes just instinctive traits. Recall that instinct covers all information-based behavior except experience or reasoning. Our capacities to learn from experience and to reason also evolved and are instinctive, but they are general-purpose mechanisms that base behavior on stored information and not just on current sensory inputs. It is much harder to predict what a general-purpose mechanism might do, so it is a lot harder to study scientifically. Science traditionally derives its value from helping us predict, so this is not even an area that appears to be of value. Contrast the value of understanding learning and reasoning with the value of helping us learn or reason better, which supports the education industry. We spend a lot on education, so we also invest quite a bit in educational theory and techniques.

By establishing a new all-encompassing paradigm for science based on form and function dualism instead of having form-based paradigms for the physical sciences and no paradigms for the rest, I think we can develop a new-found appreciation for the study of function. We already recognize the value of studying function abstractly through the formal sciences. Mathematics and computer science yield algorithms that give us much greater control over the world than we had before. We can’t always predict what general purpose algorithms will do, but we don’t need to if we understand their function and know that they will relentlessly act to accomplish their function. They become predictable on a functional basis rather than in terms of their detailed actions, which are just the means to an end. We also recognize the value of functional descriptions of the universe, considering this has been the triumph of physical science. The physical world is not functional, but how we group objects and their behavior using cause and effect is. Our minds, as functional entities, can then leverage scientific laws for functional ends. And we recognize the value of functional descriptions of life, seeing genes as bearers of all the traits that make life work. Finally, we recognize the value of function in society and founded the social sciences to tease out patterns in the ways we have applied ourselves. Function is all about capacity, and our greatest and least understood capacity is our general-purpose intelligence. We have to get our heads out of the sand and start identifying the functional building blocks of the mind if we want to have control of our own future.

It’s not that we haven’t been making steps in the right direction, but more course correction is needed. As recently as the 1930’s, psychologists hoped that behaviorism might explain all2. But by the 1950’s general-purpose thinking could no longer be ignored and the cognitive revolution arose to address higher mental function head on. By the 1970’s cognitive science had become a formal discipline and today most universities have departments for it. Unlike other disciplines where one or at most a couple of paradigms back up all research, cognitive science has avoided firm paradigms and has instead embraced a cross-disciplinary approach where paradigms from different fields support different aspects. While it is wiser to admit the limits of knowledge and work with what you know, it isn’t helping to establish a consensus paradigm. Practicing scientists can’t work outside accepted paradigms, so they end up working on details and hoping the big picture just emerges at some point. Cognition is recognized as a process; the mind as an entity remains ineffable. But if we just say the mind is a functional entity we can start to study functionality as a kind of existence instead of looking only at physicality and processes. Furthermore, all the functional constructs of the mind, from instincts to subconcepts to concepts to mental models are functional entities as well, whose physical implementations are secondary and not primary causes. The failure to address function as the underlying entity to be studied has created confusion and splintered study into dozens of subdisciplines. For example, many of these fall under the umbrella of postcognitivism, which holds that “algorithmic” cognitive approaches are short-sighted and one must consider a variety of more embedded mechanisms. That’s great, but I would argue that cognition as a process was never the point; function is the point of the mind, whether it is achieved through genetic, instinctive, subconceptual or conceptual information management. Furthermore, until function is acknowledged as one of the two fundamental categories of existence, science will proceed on the tacit assumption that eliminative materialism will eventually reduce the social sciences to physical explanations. That isn’t possible, and even leaving it on the table as a possibility creates a blind spot that hampers scientific progress.

  1. Thomas Kuhn, The Structure of Scientific Revolutions, The University of Chicago Press, 1962, p X of preface
  2. In the 1930s, B. F. Skinner suggested radical behaviorism to explain the mind entirely in terms of observed behavior.

Leave a Reply