The Science of Function

I’ve made a case for functional existence being independent of physical existence and said that it came into existence when living things began managing information through DNA. But speaking more abstractly, what is the essential character of a functional entity? The essence of a thing is also called the thing itself or its noumenon. The noumenon of a physical thing is its form in spacetime. The noumenon of a functional thing is its function, which is what it can do. Functional things are independent of space or time, but information management systems make it possible for some physical things to perform functions. While we can talk about the noumena of physical or functional things, we can never really have any direct understanding of them because understanding is indirect. To define, characterize, or describe noumena we make observations of their phenomena. Physical phenomena can be observed using our senses, but can also be measured with much greater objectivity and accuracy using instruments. The descriptions we create of phenomena are themselves functional entities, where the function they serve is to help us understand things. We can describe functional entities, too, by observing their phenomena, and these descriptions are themselves functional entities as well and can help us understand what functional entities are. But what, then, is understanding? Understanding is the ability to make useful predictions about how something will behave. Such predictions require information, which is a record of observed patterns discovered from feedback that generalize repeating phenomena. The record of the pattern is not the pattern but is a simplifying abstraction that breaks the pattern down into categories or kinds. The information in a record is recorded physically but is not physical itself because of indirection: the connection between reference and referent is decoupled or abstracted when we allocate a specific kind to a general kind. So while both physical understandings and functional understandings are functional themselves, one can’t really understand what a functional entity will do using only a physical understanding of it; the physical part is disconnected from its function.

We are quite comfortable knowing that everything in our minds, including all our knowledge both scientific and otherwise, is entirely functional and not physical because we are designed to be comfortable with knowledge, even though we know we can never physically put a finger on it. This is a bit ironic because, at the same time, the dominant framework of science, materialism, says that only physical things exist, leaving no room for the very theory that says so. While I would say that functional things like the integers or scientific theories exist in a hypothetical sense whether anybody thinks them or not, for us to discuss them they do have to have a physical manifestation. It is very much a mistake to assume that this physical manifestation is the only sense in which they exist; it only facilitates our access to functional existence in a physical universe. For science to understand the mind, we have to acknowledge their non-physical existence and probe the intersection of physical and functional. Functional entities can’t be understood or predicted using physical laws; they can only be understood using information. Information about the physical world is always gathered inductively using feedback, but one can also create deductive information using strict rules of logic or cause and effect. While deduction only works in the formal systems that define it, the uniformity of nature lets us align deductive models to natural systems to achieve considerably better predictive power than induction alone can provide.

If we can’t observe functional entities using instruments, in what sense can we observe them? We can make observations about what they can do, which includes reflections about things they have done and considerations of things they might do. Reflections and considerations are not physical observations but functional ones; they create a model for understanding that ascribes explanations to phenomena. Thinking generally about the nature of functional entities, we can see that one essential characteristic of functional things is that they represent patterns using information, with which they can predict with better than even odds. For information to provide a statistical edge over noise, it must either have been confirmed by inductively testing or follow from deductive rules. While all physical knowledge is inductive, formal systems can be deductive and/or inductive. Function has one more essential characteristic: purpose. Functional things must have a telos, meaning a purpose or “final cause”. Aristotle rightly concluded that the purpose of an acorn is to grow into an oak tree. We now also know that the purpose of the oak tree is to create acorns, completing the cycle. Oak tree genes manage the information that propagates them forward in time by building trees to metabolize matter and energy and acorns to multiply. The trunk of the tree is tall and strong to provide a competitive edge over other plants. This and many other dependent functions are served by the oak genome. These purposes are intrinsic to the informational process that created the oak gene line. That information confers many functional features to oak trees that have general-purpose utility based on both the uniformity and diversity of environmental challenges oaks have faced in their history. To some extent, we can understand the purposes served by genes, but our understanding is itself a dependent function of the mind, which is another kind of information management system than that of the oak organism, and so can only approximately explain oaks, e.g. from perspectives most relevant to our interests. Still, we can safely say that oaks are functional and our understanding of them is functional, while rocks, streams and weather patterns are inanimate and do not manage information and are consequently not functional entities.

In the physical world, functions need an information management system to be executed. Biological organisms gather information in DNA to solve functionally similar problems in general-purpose ways. DNA is physical and the proteins it codes are physical, but their functions are not physical; functions are nonphysical objectives that organisms need to achieve to survive well. Functions are not specific actions but generalized capabilities that make a range of similar actions possible. One of the most common actions animals need is locomotion. Each organism has physical mechanisms that provide the kind of locomotion it needs, but the proteins and DNA have been selected based on their success across many situations. This iterative feedback, i.e. evolution, is entirely physical at any given moment, but over time develops mechanisms shaped by generalized outcomes (functions) and not specific sequences of events. Specific events provide feedback that assesses the value of functional features; in evolution, that assessment is survival — features that survive better are deemed more effective. This creates genetic purposes meaningful to survival but meaningless to atoms or other groupings of matter. Something nonphysical — function — comes to exist through a physical medium.

Mechanisms based on DNA and protein can solve a wide range of chemical and physical problems but are not so good at solving novel problems requiring tailored situations. Organisms consequently have an evolutionary incentive to gather real-time feedback to make on-the-fly decisions. Learning is the capacity to dynamically acquire, assess, and act on patterns rather than waiting generations hoping for that feedback to be incorporated into DNA. In any case, most dynamic situations are too intricate and uncommon for genetic evolution to be able to address. Learning gives an organism a way to develop a dynamic store of learned strategies during a single lifetime. DNA provides the same functions to all individuals in a species, while learning tunes functions to each individual s’s novel circumstances, making the species flexible across a much broader ecological niche. Probably all plants and animals have some capacity to learn (more on this later), but learning is principally managed through brains, with higher animals showing substantially more capacity to learn. While our knowledge of brains and nerves is still quite immature, we know quite a few things about them. Most notably, we know that the secret of nerves is in their interconnections and that the number and strength of these connections somehow encodes information. Nerves combine hardware and software into one — patterns are recorded using a network of nerves that reactivates to repeat the pattern. Our record of a pattern is not the same as the pattern (“the map is not the terrain”), but we can correlate or map it back to the phenomenon it represents. Although these correlations are always approximate in the sense that references list traits that might apply to many referents, we can include specificity among those traits to link a reference to a specific referent, the way a name or proper noun is specific. In this way, the mind catalogs specific cases as refinements of general cases. In other words, everything is ultimately general or relative and nothing is truly locked down.

Brains leverage the nervous and endocrine systems to create a dynamic information management environment. Minds are a separate information management system within brains whose specific purpose is to manage top-level decisions. The mind is a subprocess that manages a customized subset of brain information. Just as a software program like a browser uses distinct logic and data from the underlying operating system, so too is the mind’s logic and data distinct from the rest of the brain. Organisms, brains, and minds are all distinct kinds of functional entities that manage different kinds of information. Nearly all mobile animals have some need of a specialized top-level information management system to make them pursue different behaviors at different times. Andrew Barron and Colin Klein argue in “What insects can tell us about the origins of consciousness” that many lower animals with centralized brains, such as insects, have subjective experience:

In vertebrates the capacity for subjective experience is supported by integrated structures in the midbrain that create a neural simulation of the state of the mobile animal in space. This integrated and egocentric representation of the world from the animal’s perspective is sufficient for subjective experience. Structures in the insect brain perform analogous functions. Therefore, we argue the insect brain also supports a capacity for subjective experience.1

Their phrase, “neural simulation of the state of the mobile animal in space”, is the computational essence of conscious awareness. Many other functions contribute to our overall experience of consciousness, most of which are not available to insects, but an “integrated and egocentric representation of the world from the animal’s perspective” is the most fundamental attribute of consciousness. So it stands to reason that the basic neural structures supporting it should be found in all animal brains (specifically in the midbrain for vertebrates) and not in the cortex, which developed later. Animals brains first arose about 550 million years ago, and mammals and dinosaurs split off about 200 million years ago. These two lines evolved cortical areas which appear to support more advanced reasoning, but basic support for consciousness had a 350 million year head start.

Using our minds, we can create higher-order information management systems to address specific problems. Most notably, we built civilizations to manage large groups of people in a stable, organized way. Civilizations store information in institutional practices shared and carried out by their members. We have also designed many formal systems, for instance academic fields of study, that manage information by establishing premises and rules to manipulate them. And recently we have built machines (computers and their peripherals) that can simulate formal systems, which can give these functional entities a physical manifestation.

Simulation in the brain is the product of information processing, but more significantly than the simulation alone is the secondary process “watching” the simulation as subjective experience, i.e. consciously. Simulating mobile animals in space through a first-person “technicolor” experience is the approach evolution chose to control animals but has not yet been used to program robots because we don’t yet know how to design a program to do that. In principle, we could faithfully replicate our conscious experience in a robot given better sensing and programming technologies. Note that the other information management systems, i.e. organisms, brains, civilizations, and programs executing on computers, don’t use the simulation approach and consequently function without awareness or consciousness. The simulation approach requires two cooperation subprocesses, which I call the simulator and the simulatee. Both are processes in the brain, but the simulatee is special in that it is conscious, which is not actually magical or mystical as we like to think, but is simply how this subprocess works. The simulatee doesn’t just watch the simulation, it interacts with it. Because the simulation is fed live data from the outside, those interactions result in physical-world changes, which makes the simulatee an agent in that world and not just a brain in a bottle.

The job of the simulator is to twofold: first, to provide information that is as accurate and precise as is practicable so that the agent can trust it, and second, to inspire the simulatee to act. The first of these jobs is routine: to provide consistent and detailed information, the brain should develop specialized processing for each sense that always behaves the same way. The second of these is itself twofold: first, to provide preferences for desirable objectives, and next to help balance those preferences as effectively as possible. It may seem like the simulatee, being our conscious self, should be able to figure out what it prefers and make its own choices, but our conscious role is only supervisory. The inspiration is fed to us by the simulator as drives and emotions, which are specialized algorithms refined by millions of years of evolution to act in ways which have best promoted our own interests. We actually do supervise; the inclinations etched into our simulators are no guarantee of future success, so the simulatee has to pull the pieces together to arrive at appropriate reactions. But the two processes work closely together to streamline the process, perhaps most notably in the construction of belief. Belief is pervasive in all our thoughts — we believe that our senses and memory provide us with a seamless and complete picture of the world, even though our knowledge is really superficial and full of holes. But we, as simulatees, evaluate feedback to develop considered opinions which our simulator then renders back to us later as layers of certainty called belief, which we then use to make quick and effective decisions. We feel belief as trust, which differs in strength from hunches to facts, and the strength affects how likely we are to act on the feeling. Belief is an essential cognitive shortcut, not only to prevent analysis paralysis but to let us develop models of the world in the first place. We aren’t fooled; we know our impressions and beliefs are only tools to help map the simulation to external reality, but even so, our feelings are real to us because the simulator is external to us as well. We can gradually review and revise beliefs and so come to feel differently about things, but this feedback-based control of the simulator takes time.

Functional entities can be said to divide into two categories, those that are fully autonomous, which include organisms, brains, minds, civilizations, and computers, and those that are the dependent on autonomous entities. The dependent functions of organisms include purposes of everything from proteins to cell organelles to organs. For brains, it includes a variety of unconscious bodily functions such as heart rate, digestion, and rate of respiration, as well as the many unconscious support functions needed to support the mind, such as 3-D vision processing and a knack for language. For minds, it includes feelings, subconcepts, and concepts. For civilizations, it includes the functions served by laws, institutions, and conventions. And for formal systems, it includes whatever functions the system defines, which covers both its formal implications and the less formal ways the system can be applied to appropriate situations. Any discussion we have about any of these dependent functions will itself be conducted using ideas, which are composed of feelings, subconcepts, and concepts, which situates these discussions entirely within our minds. Although function is not itself physical, functions can manifest physically by exploiting the uniformity of nature to gather information and feed that information back to achieve functional goals in an otherwise indifferent universe. This natural and perhaps inevitable consequence of the laws of nature means that both physical and functional existence are natural, and not just the more visible and measurable physical existence.

To describe how functional things work I will use concepts, which are functional things that can characterize other functional things. Ideas also include feelings and subconcepts, but these are intrinsic and not explanatory. Feelings leverage innate information management mechanisms to create conscious states that make us preferentially disposed to act in ways that have been helpful over evolutionary time. Feelings include sensations like pain, drives like hunger and thirst, and emotions like fear and happiness. Subconcepts comprise our whole experience categorized into generalized impressions. These impressions let us know whether things we encounter are likely to be helpful or hurtful, a good fit or a bad fit, etc., but they don’t explain why. We need feelings and subconcepts to understand things and keep our thoughts directed usefully, but for explanations backed by a chain of reasoning, we need concepts. While feelings and subconcepts inherently depend on each person’s unique psychological makeup and experience, concepts can provide a more objective understanding we can all agree on.

Let’s take a look at some concepts to get a better idea of how they work. The concept APPLE (capitalization means we are referring to apple as a concept) refers to the generic idea of an apple and not to any specific apple. It is not about the word “apple” or metaphors of apples, but to things we would classify as an apple based on being sufficiently apple-like. We each arrive at our understanding of what APPLE means from our experience with apples. Even though we each have distinct apple experiences, our concept of what APPLE means is functionally the same for most purposes. How can this be? APPLE is a generalization based on a variety of objects we encounter which we learn are called apples. Apples are the fruit of the apple tree, Malus pumila, and are typically red, yellow or green, about the size of a fist, and are often eaten in a succession of bites, except for the core, or sliced up and baked into apple pies. Although each of us has an entirely unique set of interactions with apples, our functional understanding, namely that they are white-fleshed fruits eaten in certain ways, is the same from the most general perspective. Some of us may think of them as sweet and others as sour or tart, but the functional interactions commonly associated with apples are about the same. Some of us may know that apples are defined to be the fruit of a single species of tree, and some may not, but concepts are generalizations, so they still work, albeit in fewer situations, even if only partially understood. The person who thinks that pears and/or apple pears are also apples has a mistaken APPLE concept relative to the more broadly accepted standard, but their more general concept is still valid for their purposes. One can endlessly debate the exact standard for any concept, but exactness is immaterial in most cases because only certain general features are usually relevant to the functions under consideration. Because any given word associates with a given concept in a given context, and the concepts associated are functionally equivalent for most intents and purposes, communication is possible. Many concepts have corresponding words, but many more do not. Many words have multiple definitions, each of which can fairly be called a distinct concept. It certainly helps when discussing a concept if an agreed word or phrase refers to it so as to avoid confusion or repetition. A fake apple made of wax is not an APPLE as it would not meet the dictionary definition, but it is common practice to refer to a look-alike of an object using the name of the object, given that it is understood it is an imitation. Borrowing a word in this way often leads to the creation of additional definitions for the word. It is ok for a word to have many definitions provided the context reveals the definition of interest.

HUNT is a concept for a functional entity, as hunting is an activity undertaken to achieve a purpose. There can be no physical manifestation of hunt, as the people engaged in a hunt cannot be said to be a hunt, because words defined in terms of their purpose are functional and not physical. Functional concepts include all mental states and purpose-based pursuits, like HAPPINESS, JUSTICE, and SUBTRACTION. Any extent to which physical manifestations facilitate their existence fail to capture the essence of their existence. Emotional responses are instinctive. Our capacity to use subconcepts and concepts is also instinctive, but our skill at using them and their breadth is a function of experience, which is learned and influenced by decisions we make. The arts and sciences are the product of focused efforts to develop ideas.

Scientific theories are also functional entities and don’t reveal the actual nature or truth about physical entities, but only functional aspects of what we could expect from interactions with them. Boyle’s Law says that the volume and pressure of a gas at constant temperature are inversely proportional. But volume and pressure are macroscopic “properties” of the “gas”, which itself is a macroscopic conception of how molecules of the same substance behave when their kinetic energy is too large for intermolecular (chiefly electrostatic) forces to collapse them into a liquid or solid. Volume and pressure are useful collective generalizations but are not fundamental properties of spacetime. Under our current best understanding of physics, the Standard Model of particle physics, fundamental properties can be attributed to subatomic particles, of which there are a few dozen varieties, and scientific “laws” regarding groupings of such particles, from atoms to molecules to solid/liquid/gas formations, are somewhat approximate in their accuracy and precision for any number of reasons. Still, we can measure volume and pressure almost perfectly most of the time, and in these cases the law applies almost perfectly as well, so we can feel pretty comfortable trusting that it always works when applied within appropriate parameters. Despite our inability to prove it, we can label it “true” as settled science and use it without fear. We have built many machines that work perfectly all the time provided they are kept within operating parameters. When they do fail, we can, in principle, always trace the failure to falling outside those parameters or a failure to set good parameters rather than to a failure of scientific laws, although in practice even settled scientific laws may sometimes need refinement.

Given these caveats about how science can only characterize subjects and can’t reveal their true nature, let’s take a look at which sciences study functional subjects. Viewed most abstractly, science divides into two branches, the formal and experimental sciences. Formal science is entirely theoretical but provides mathematical and logical tools for the experimental sciences, which study the physical world using a combination of hypotheses and testing. Testing gathers evidence from observations and correlates it with hypotheses to support or refute them. Superficially, the formal sciences are all creativity while the experimental sciences are all discovery, but in practice most formal sciences need to provide some real-world value, and most experimental sciences require creative hypotheses, which are themselves wholly formal. Experimental science further divides into fundamental physics, which studies irreducible fields and/or particles, and the special sciences (all other natural and social sciences), which conceive aggregate properties of spacetime which are presumed by materialists to be reducible in principle to fundamental physics. Experimental science is studied using the scientific method, which is a loop in which one proposes a hypothesis, then tests it, and then refines and tests it again ad infinitum.

Alternately, we can separate the sciences based on whether they study form or function. Physical forms are more objectively observable, given that we can use instruments, but as noted above we can claim to observe functional entities through reflections and considerations of their functional claims. For example, the formal sciences establish functional entities through premises and rules from which one draws implications (which are functional). The formal sciences include logic, mathematics, statistics, theoretical computer science, information theory, game theory, systems theory, decision theory, and theoretical linguistics. They are named after formal systems, in which “form” means well-defined or having a well-specified nature, scope or meaning. However, that definition of form actually means function, because definition, specification, and meaning are functional. I will restrict my use of “form” to physical substance: a form is a noumenon that can only be known through observed phenomena, though that knowledge is itself functional. Functional noumena always have a purpose which can be known independent of phenomena, and purposes have no physical substance. This is another way of seeing that form and function are distinct classes of entities. We can study formalized functions either deductively or inductively (as I mentioned above). Deductively, we just apply the rules we provide to the premises and draw out the logical implications. Inductively, we can use simulations to test more complex implications than those that follow directly from the rules. Some formal sciences (e.g. weather modeling) are arguably more experimental than formal because they depend so much on inductive simulations. We often think of the formal sciences as pure and not bound to applications, but this is an illusion; formal sciences are functional and function requires applicability. It is true that an infinite range of formal sciences are possible, and any arbitrary set of rules may be chosen for any of them, so if we had no criteria for studying one over another then their scope would resemble an infinite set of monkeys trying to produce the works of Shakespeare. But that doesn’t happen because formal scientists are very concerned about the potential for applicability, even if they are unconcerned about specific applications. Formal science pursues pure, generalized, logical systems with provable implications because they have a high chance of being useful. In Mathematics Form and Function, Saunders MacLane proposed six possible foundations for mathematics: Logicism, Set Theory, Platonism, Formalism, Intuitionism, and Empiricism. But these foundations are all wet — all the formal sciences really aim to do is maximize functionality. Whether they align with logic, sets, ideals, forms, hunches, evidence, or any arbitrary rules is not the point; applicability is the point. Math, and all the formal sciences, are not random flights of fancy; they have the unspoken goal of supporting applied science.

While the experimental sciences, being physical, life, social, and applied science, are concerned with physical form, they are more focused on function. The physical sciences study form alone, specifically physical forms in our universe bound by the theory of everything, which is not yet a concrete theory but the hoped-for unification of general relativity and quantum field theory. However, formal models constitute the theoretical framework of the physical sciences, so its explanations are functional constructs. The life sciences principally study function, specifically functions that result from the theory of evolution. Of course, lifeforms have physical bodies as well, which must also be studied, but nearly all understanding of life derives from function and not form, so looking at the form is basically only helpful in furthering the understanding of function. The distinctly identifiable functions that living things perform gradually evolve through complex, interwoven, and layered feedback that gradually causes physical mechanisms that further functional goals to “emerge”. Functional existence leverages physical existence, and depends on it to continue to exist physically, but is not the same as physical existence. The social sciences also focus on function, specifically functions produced by the mind, despite the lack of a theory that explains the mind itself. Nearly all the understandings of the social sciences were built by observing patterns in human behavior. Behavior is a complex consequence of how minds function. Finally, applied science is principally concerned with functions that help us live better, but uses both form and function to develop technologies to do so. All the experimental sciences depend heavily on the formal sciences for mathematical and logical rigor. To summarize, understanding physical forms is central to physical science but is only incidental to the other experimental sciences and is irrelevant to the formal sciences. Understanding function, however, is central to the formal, life, social, and applied sciences, but it also underlies the theories of the physical sciences. After all, the reason we try to understand physical forms is to gain functional mastery over them (via laws of physics).

Experimental science has had its greatest success in the physical sciences because (a) nature is very uniform, (b) we can measure it with impartial instruments of great accuracy and precision, and, most importantly, (c) relatively simple theories can predict what will happen. That nature is uniform and measurable is, thankfully, a given, but the last point is a subtle consequence of conceptual thinking. I mentioned above that all concepts are generalizations. A generalization is an identification of patterns shared by two or more things. Categorizing them as the same kind of thing lets us make predictions based on our knowledge of that kind. But there is a hidden challenge: every pattern has many features, and the number of combinations of features grows exponentially, resulting in an almost infinite number of kinds. How do we avoid a runaway proliferation of categories? The answer is Occam’s razor, which is both a deeply seated innate skill and a rule of thumb we can apply consciously. Occam’s razor says that “when different models of varying complexity can account for the data equally well, the simplest one should be selected”2. Simplicity trumps complexity for two reasons. First, the universe seems to follow a small set of fixed laws. While we can never know what they are with complete certainty, we can approximate them most effectively by trying to build as small a consistent set of laws as we can. And second, as the complexity of a pattern increases, it becomes less general and so works in fewer situations. Complexity is great for specific situations but bad for general situations. So again, to build a small but consistent set of laws, we should simplify wherever possible.

Consider an example. Theories to explain the motions of celestial bodies across the sky have been refined numerous times with the help of Occam’s razor. Ancient astronomers attached moving celestial bodies to geocentric (Earth at the center) celestial spheres. Ptolemy then explained the apparent retrograde motion of Mars as seen from Earth by saying a point on Mars’ sphere revolves around the Earth, and Mars revolves around that point along a smaller circle called an epicycle. (Note that additional epicycles were used to make small corrections for what were later learned to be elliptical orbits.) Copernicus realized that by using heliocentric (Sun at the center) celestial spheres, the epicycles used by Mars and the other planets to explain their retrograde motion would be unnecessary because this illusion is just a side effect of looking from one planet to another in a heliocentric model. Fewer epicycles were simpler to calculate and also made the model simpler to understand. Kepler’s discovery that orbits were ellipses replaced the celestial spheres with celestial ellipsoids and eliminated all the epicycles, again simplifying calculations and comprehension. Newton eliminated the need for celestial ellipsoids by introducing the universal law of gravitation, which both reduced all matter interactions to one equation but also explained gravity on Earth. However, Newton’s gravity depended on action at a distance, which seemed supernatural to Newton, who felt that interactions should require direct contact. The influence of gravity creates an exception to the rule that objects in motion move in a straight line. Einstein’s theory of general relativity eliminated that exception, allowing planets and all things to travel in a straight line once again, but now through curved space. And we still have a ways to go to reach a deeper and unified understanding spacetime, gravity and matter, which I will revisit later. But the important point is that our laws of gravity and physical science are now good enough to make nearly perfect predictions about many of the physical phenomena we have identified.

But life is a far less tractable subject. Billions of years of adaptations have piled on complexities orders of magnitude harder to decipher than those of nonliving physical systems. That complexity is driven by feedback to provide general-purpose functionality instead of the specific causes and effects studied in physics and chemistry. Functionality can’t be studied directly with instruments, but requires analysis — information can only be understood by doing more processing on the information. Consequently, theories about life can never achieve the same level of formality and closure enjoyed by the physical and formal sciences. But we do have some objective sources of information about function in living things. Chiefly, we know life evolved, and we know a number of the mechanisms that made that possible. More significantly than the mechanisms, we know that it was driven by the value of function to survival. Function was selected for, and the mechanisms that made it possible were only along for the ride. It was not the genes that were selfish, but the functions of the genes. In other words, their physical form mutated along the way to preserve an unending chain of function, making their functional existence paramount and their physical existence incidental. Their functions are informational constructs whose true depth hides in the full history that led up to each gene surviving to the present day, and can’t be fully grasped just by discovering, say, the apparent primary role of the protein the gene encodes. The whole context of its function across a wide variety of circumstances contributes to why the gene is exactly the way it is. But one can still try to guess the functions and run tests to verify them, so we can still adapt the scientific method to minimize subjectivity and maximize reliability and broad applicability. I said above that the two basic information management systems these sciences study are life (overall) and the mind. In the following quick review I will further subdivide the mind into instincts, subconcepts, and concepts:

Genes. The bodies of living organisms are their physical manifestations, but the bodies exist to fulfill evolutionary functions, and in particular the function of each gene. Genes either encode proteins or regulate when genes turn on and off. Proteins engage in pretty specific chemical reactions which usually reveal at least the primary purpose of the gene. We can conclude that their chemistry ties pretty closely to their function. This doesn’t resolve finer details of their function because it is impossible to predict all the side effects the production of a chemical might have. By studying variants of the gene and situations where the gene is not expressed normally, under normal or stressed situations, we can come to appreciate its net value to the organism better. But we have found in many cases that chemical knowledge translates pretty well to functional knowledge in the case of genes. A common-sense mantra of science says that “form determines function,” which means that one can guess at the function by considering the form. In Darwinian evolution, the form (variation) precedes the function (as determined by selection), but the net result is that the function causes the form to be selected and then to persist, so function determines form. While it is not fashionable (yet) for life scientists to declare that function is their primary object of study because physical monism is still the dominant paradigm and physical evidence is still seen to trump functional evidence, they still must think about function first and physical mechanisms second. Having demonstrable physical mechanisms behind the functions makes it easier for the life sciences to claim objectivity. In this way, they can still primarily study function and teach biological functions even though function has no metaphysical basis in the physicalist philosophy of science. It has widely been fashionable over the past century to wear the absence of philosophical support as a badge of honor because science gets things done while philosophy seems to spin its wheels beaming increasingly irrelevant, but that willful ignorance is wearing thin because it is starting to impede further scientific progress. Philosophy does not stand apart from science but is implicit in and essential to all rational thought. It is true that philosophy is a massive accumulation of ideas, many of which are now outmoded. This makes it harder for newer philosophies to push through, but we desperately need a firmer foundation for science. I’ll discuss that further in the next chapter.

Instincts. Instincts are behaviors that respond to triggering stimuli in a way fixed by genes rather than experience, i.e. they are innate behaviors. An instinct strongly nudges the mind to react in a given way rather than being a hardwired reflex. We can distinguish instinctive from learned behaviors by seeing if the behavior happens without experience. Beavers that have never seen a dam can still build one, so we know it is instinctive. Humans who have never heard a language can quickly create one, so we know that language acquisition is instinctive. We don’t yet know which genes create instincts, and many instincts depend on many genes, but we know instincts are genetic because they are not learned. Once we have demonstrated that a behavior is instinctive, we could, in principle, discover the underlying genetic mechanisms that support it. In practice, this is somewhat beyond our current capabilities since we have found genes for few if any instinctive behaviors, but it still helps in our quest for objectivity to know that a behavior is instinctive because it means we objectively know it functions independent of learning.

Subconcepts. Before we get to subconcepts, let’s take a look at percepts. Percepts are the sensory feelings that flow into our minds continuously from our senses. The five classic senses are sight, hearing, taste, smell, and touch. Sight combines senses for color, brightness, and depth to create percepts about objects and movement. Smell combines over 1000 independent smell senses. Taste is based on five underlying taste senses (sweet, sour, salty, bitter, and umami). Hearing combines senses for pitch, volume, and other dimensions. And touch combines senses for pressure, temperature, and pain. Other somatosenses include balance, vibration sense, proprioception (limb awareness), hunger, erogenous sensation, and chemoreception (e.g. salt, carbon dioxide or oxygen levels in blood). Awareness and attention themselves have a feeling of time and space. We feel all of these things without reflection; they are immediate and hard-wired subconscious mechanisms that bring external information into conscious awareness.

We know that we can consciously reflect on our perceptions to find patterns, and when we find a pattern we will remember it and keep an eye out for it in the future. What is less clear to us because we have no conscious awareness of it is the degree to which our subconscious mind can find patterns and leverage them. However, we can use process of elimination to spot cognitive tasks we know we must be performing for which we can’t take conscious credit and attribute them to the subconscious. What I want to focus on here is the kind of stored experience the subconscious uses to do that, which can be divided into two categories, procedural and informational. Procedural memory, called muscle memory when used to help us move, focuses on helping us repeat useful behaviors. We consciously preapprove procedural memory to start or continue so long as situations warrant, so can appear to be driven entirely subconsciously, but conscious approval is required. Informational memory, which I call subconcepts, help us understand the world. Consciously, we just look at something and recognize it, but we know from an informational standpoint that it must require millions or billions of comparisons to work. We can conclude that recognition is a massively parallel subconscious process that accesses this large store of information. Recognition is so continuous and pervasive that it gives us an ongoing feeling of familiarity. Anything unfamiliar will stand out to our attention process, which is itself a subconscious process designed to make us focus conscious attention on anything novel or unexpected, i.e. not recognized by our subconceptual database. While procedural memory can be overridden, recognition can’t be suppressed; once we see (recognize) something, we can’t unsee it. We usually associate recognition with visual objects, but we can recognize many aspects of situations. If the recognition is faint, we get a sense of déjà vu, but usually we get comfortable familiarity. Recognition in this broader sense is more generally called intuition and includes everything we know without consciously reasoning about it. Intuition roughly divides into the confident knowledge we call common sense and the more speculative knowledge we call hunches. Let me go over concepts before considering objectivity again.

Concepts. As I have noted before, the hallmark function of conceptual thinking is problem-solving, which is much more powerful than recognition and intuition because logical reasoning chains causes and effects together, making multi-step procedures possible. Recognition and intuition can essentially only go one step deep, albeit very powerfully. Though it may sometimes seem to us multi-step solutions appear to us by intuition, I think it is more likely that we subconceptually store many things about steps and multi-step solutions we have used before, and intuition recognizes links we couldn’t see from conscious effort. With logical reasoning, we build simplified models of the world in which concepts form the subjects, predicates, and premises that we link together logically to reckon conclusions. Since formal models can achieve perfect objectivity because they define what is true, logical reasoning seems at least superficially to be the pinnacle of objectivity. This breaks down a bit because models must be applied to the real world, and they don’t always model it perfectly, and in fact can never do so perfectly. While a model itself can be spelled out in great detail and agreed to by any number of people quite objectively, how we apply it is always somewhat more circumstantial and a function of more intuitive, recognition-based matching, for which it is hard to be completely objective. Also, we rarely do spell models out in so much detail that the logic is perfect, so many of the steps in our thinking fall far short of being mathematical proofs and may, in fact, depend on unknown subjective processes. But the quest for objectivity should look first to the clear, widely accepted models, especially scientific models, and then to more subjective models and mechanisms which can vary from person to person.

Our direct knowledge of the management of information in the mind, via instinctive, subconceptual, and conceptual thinking, comes entirely from the use of our own minds and is consequently entirely subjective. How then can we develop objectivity about it? First, consider that all knowledge outside the formal sciences is subjective and that objectivity is just a tool to make some of it more reliable and broadly applicable. Evidence from instruments can be very precise and has been very helpful in the physical sciences, but the study of function doesn’t necessarily need that kind of precision. Working with the information sources available, we should still be able to develop reliable and broadly applicable theories. Next, we can still make hypotheses and test them even if it is harder to isolate the specific mental phenomena we want to test. But to get started I think we can find many things about mental function we could agree on without having to conduct tests at all. Finally, we should consider the ways the social sciences try to achieve objectivity considering their sources are purely subjective.

So yes, our thoughts are superficially subjective. While instruments can’t measure thoughts (although MRI and other brain imaging tools have done wonders to localize where different kinds of thoughts in the brain occur), most of the objective power of science comes from the use of the formal sciences, mainly math and logic, which can be applied to the mind as well. This is especially appropriate for understanding the mind, because the mind is what we use to create formal models. Introspective descriptions of our thoughts and feelings may be entirely subjective, but models supported by the available scientific information that are consistent with our subjective thoughts are objective. Theories that start on objective ground, even about subjective things, remain objective until proven otherwise. Anyway, our minds are designed to pull objectivity out of subjectivity: we all discern external objects the same way and believe it is generally possible to distinguish truth from falsity. Language itself depends on developing a shared understanding of concepts, which are abstracted objects. We mostly see the world as a place of facts, not opinions.

Our only evidence of our minds beyond our personal experience of them is our behavior. However, if we only had behavior to go on, we would be very hard-pressed to guess anything about the mechanisms of our minds. In fact, without our own first-hand experience of consciousness, we would have no reason to suspect that minds even existed. We would just see robots moving about getting things done, not unlike ants. To the extent ants can be said to have minds at all, which is pretty debatable, they are certainly not remotely as functionally complex as ours. Arguably the robots would claim they had minds, but, like us, they would still need to prove they existed and were not just charming affectations built into their programming but incidental to their operation. We can’t argue that minds are fundamentally necessary, either, since it is clearly possible to design a mindless brain to perform any given task we wish to give it using brute-force machine learning approaches that consume vast quantities of data and experience. While it is tempting to suppose that such a zombie-like robot would not be as adaptable to new circumstances as us, it is at least theoretically possible to program it to have a range of adaptability sufficient to handle any situation humans have yet faced. While such robot humans may not need art or entertainment (unless these turn out to play an important role in developing general-purpose adaptability), they would procreate and advance civilization as well or better than we would. The problem with these zombie scenarios is that while they are theoretically possible, they are not the route life took, which is probably because it is not as effective a way to design a system to do what we do. While I can’t prove this, it seems reasonable to assume that the path life took, which was to unleash considerable general-purpose intelligence in humans and attach emotions and drives to make sure we use it to enhance competitiveness and procreation, takes fewer genes and is therefore more direct than developing more instinctive responses to a wide range of situations. In other words, it was feasible for beavers to develop instincts for building dams, but it was not feasible for humans to develop instincts for building houses, cars, etc. And it is my contention that to the extent animals do employ general-purpose capacities (aka intelligence) to solve problems, they experience consciousness to do it. Consciousness exists to support general-purpose decision making as opposed to instinctive behavior. All earthly animals with centralized brains have some features of consciousness, which I propose happens because some degree of general-purpose thinking is so useful and minds are the way it works. So the real question is what capabilities of consciousness have made it so successful.

The umbrella answer is that consciousness is function made animate through agency. We know that the brain’s role is to control the body, but control is a functional construct because it has the goal of attaining purposes. This functional perspective of the brain is called the mind. The strategy earthly minds use to coordinate their control is agency, which means that the mind interprets its inputs and outputs as the feelings and actions of an agent. This concept of an actor or agent is entirely functional and has no physical meaning. This view intrinsically feels like it is first-person because the inputs contain self-information about the body and not-self-information about the outside world that combine into a running story of events that seem to center around the agent. The agent consequently adopts the philosophical stance that, although it may technically be an abstraction, it exists as an acting entity in the world. Enabling a portion of the brain, namely the conscious mind, to adopt this perspective brings survival benefits proven useful by evolution. I’ll start to tackle what these benefits of the conscious mind are in a few chapters. For now, we need only know that top-level decisions are not made by a mindless prioritization algorithm as one might expect but by a very mindful algorithm called consciousness. This subprocess “believes” that it is an autonomous agent in the world that “feels” inputs and uses the way they feel to select appropriate outputs. That we observe others acting purposefully may not prove that they experience agency, because some purposeful behavior is instinctive and does not depend on agency. But it is pretty safe to say that humans and other higher animals do experience consciousness as agency. Can we objectively describe what subjective experience feels like? It is possible, provided one keeps in mind that an explanation can only characterize the phenomenon and can’t capture its noumenal feel (i.e. the map is not the terrain). Put another way, we can describe it but a description can only reveal functional aspects or elicit related feelings and won’t actually let you feel the experience being described. So I can explain why redness exists and what role it serves, and I can draw analogies to other feelings, but I can’t help you feel red through descriptions. But I will dissect conscious experience into pieces in a few chapters, which will help us understand why it feels the way it does. The special qualities of the way our experiences feel are not imaginary, in the sense that they are created for our conscious minds by subconscious processes, but they are imaginary in the sense that they only exist as information in the brain. The result is that things “feel like their function”, i.e. what they make possible. Red and yellow evoke a warm feel while green and blue evoke a cool feel. That function is baked into how we perceive them. My stance, called functionalism in philosophy, holds that redness is not arbitrary but results from the function of some red things. Because all our sensory experiences have functional roles, our subconscious minds create appropriately customized feelings for them that are passed to our conscious minds. Many of these functional distinctions are learned, “acquired tastes” which we come to appreciate, but most are innate, the product of millions of years of evolutionary pressures mapping function to feeling.

To give an example, as we survey an ordinary scene in front of us, we are calm and nothing stands out to our attention, even though we can distinguish any number of discrete objects in the scene. But if anything in that scene becomes bright, or flashing, or red, or fast-moving, or loud, etc., our pulse will quicken and our attention will immediately be drawn to it. Those stimuli have the function of warning; they are different from each other, but any of them can trigger the warning reaction and so in many ways feel the same to us. Red and yellow stand out more in any context than other colors because in our ancestral environment objects of these colors were more likely to warrant attention than green, blue, brown or gray objects. This doesn’t mean color alone alarms us, but it is a factor, and importantly, it affects how these colors feel to us. Blues and greens are calming, while reds, oranges, and yellows are a bit unnerving. It is not unpleasant; it is just part of the quality about them that we feel. If we could devise a set of glasses that could invert greens to reds and blues to yellows3, I believe that after a week or two we would come to invert them back, feeling red things like leaves as if they were green and green things like blood as if they were red. While this experiment has not yet been done, this result has been found with similar experiments that flip the image to the brain horizontally or vertically. I am not saying flipped or inverted qualia revert to being indistinguishable from before. No, superficially the quality remains inverted — people who view inverted scenes know they are inverted but can interact with them on that basis as if it were perfectly normal, and it doesn’t seem strange to them. Similarly, people would know that red things like leaves are being perceived as red, but they would trigger calmness and other feelings the same way that green things used to feel. Much of the feeling we get from colors and other qualia isn’t about their superficial distinctness but about our beliefs about their function. Our memory of how colors used to make us feel would matter more to us than the way they appeared now, and we would remap our feelings about the colors back to what they were before. The reason I believe the brain could do this color feeling inversion is that it does this sort of thing all the time; changes in lighting can make the same color appear quite different, yet our feeling about it remains the same. The brain is constantly trying to interpret inputs into functional buckets, correcting for variations in the signal. Ultimately, redness, brightness, loudness, etc., are about how information is hooked up in our minds, not about what is happening outside them, and the way it is hooked up is all about what how that information can help us, i.e. what its function is. In other words, the brain is functional and not literal, so it has many mechanisms to look past literal changes to the underlying function.

The perception that physics and chemistry are fundamentally more objective, provable, and definitive has led to them being called hard science, while social science, which is seen as more subjective and less provable and definitive is called soft science. Biology has both hard and soft aspects. The distinction really derives from our intuition that hard science is a fixed or closed system while soft science is not. A closed system can be modeled as perfectly as you like with a logical model that explains all its fixed components. A variable or open system includes feedback loops which continually impact and adjust the design and capabilities of the system itself. When an open system is implemented using a closed system, as the mind uses the body, there will be an underlying fixed physical explanation for what is happening particle-wise at any given instant, but the physical explanation will reveal nothing about the functional capacities of the system. Physically, information doesn’t exist; fluctuating signals traveling on wires or nerves exist, but divorced from any concept or purpose they have no relevance to anything. It only acquires relevance when an information management system gathers and uses information about something else, summarized and analyzed at practical levels of detail. Cells manage inherited information through genes, which summarize metabolic information, mostly about proteins, in a practical way. Minds manage information summarized from sensory inputs using both inherited (natural) and learned (artificial) mechanisms. Because cells and organs have a very fixed structure and behavior for any given species at any given point in its evolutionary history, the study of these structures from a physical standpoint can often be done with clean, logical models that explain all the fixed components. These models can often be experimentally verified to a high degree of confidence. Although we know such models of biological systems are inherently less fixed than those of nonbiological systems, they are quite comparable for most intents and purposes. They do posit a function for each kind of tissue, which is necessarily a subjective or soft determination, but the primary purpose of most tissues seems very clear, so while some appreciation for multifunctional tissues is lost in this kind of summation, it still works pretty well. But this approach mostly breaks down when studying the brain because its functionality is so highly integrated across many levels. We have identified primary functions for many parts of the brain, but we also have to accept that almost every function of the brain includes substantial integration across many areas. Not only do different brain areas and functions work together to achieve overall function, but they also incorporate feedback across multiple timelines. Instinct gathers feedback over millennia, long-term memory gathers it over a lifetime, and short-term memory gathers it for the scope of a problem at hand. And then there is the matter of the processing, or thinking, that we do with the collected information. While this bears considerably more discussion, for now it is sufficient to say thinking is quite open-ended and impossible to predict. So the distinction between the hard and soft sciences, or more accurately between the physical experimental sciences and the functional experimental sciences, is quite significant. However, it is misleading to characterize it as hard vs. soft; the distinction is really between fixed, closed systems and variable, open systems.

  1. What insects can tell us about the origins of consciousness, Andrew B. Barron and Colin Klein, PNAS, 2016
  2. Structure learning and the Occam’s razor principle: a new view of human function acquisition, 2014 Sep 30
  3. Alex Byrne, Inverted Qualia, Stanford Encyclopedia of Philosophy, 2015

Leave a Reply