A Metaphysical System in Pieces

2

A System in Pieces
J.N. Findlay Lecture
Metaphysical Society of America
Sunday March 20 2016
Larry Cahoone

[NOTE: numbers at left correspond to slides in accompanying Power Point.]

1. There is something unnerving about giving a paper to the MSA. The reason is not a lack of hospitality, to be sure. It is rather, there are neither many young people, nor are there many members of well-established mainstream schools, that is, the cliques that dominate our business. One is stuck giving a talk to a group of mature individuals each of whom has formed, on the basis of wide historical and contemporary reading, her or his own established views. At mainstream meetings, everybody has imbibed a perspective in graduate school and slavishly kept within its borders all his or her life; kept there predominantly by social and professional pressure. Such people are narrow; you can run rings around them, although of course they won't notice because they can't see the rings. The MSA is different. I can't pretend to teach anyone here anything. But perhaps if I display my own way of thinking, some bits or pieces of what I do might be useful to the work others are doing. That's the hope, anyway. So I will describe my approach to metaphysics. I will have to speak in brief slogans and talking points, but that is the price of trying to do a lot in little time.

2. I assume whatever else philosophy is, it is a form of inquiry, continuous with all other forms of inquiry, from physics to literary interpretation. So it is trying to say what is true though using reasons and evidence. But philosophy is the most general form of inquiry, asking the most general questions. Consequently for the philosopher there are lots of things to inquire into, and lots of ways to inquire. My overall intuition is that proper method in the most general inquiries in philosophy – like metaphysics -- is tightening lug nuts on a tire. You know how it works: you can't tighten one all the way, then move to the next. You tighten one a little, then move to the most nearly opposing nut and tighten it a little, then choose a most opposing nut, eventually revisiting each in the same pattern. I feel that way about both subject matters – metaphysics, social-political philosophy, ethics, physics, biology, philosophy of mind, philosophy of culture, etc. – and about methods or approaches to those subject matters, from history to logic to positivism to pragmatism to phenomenology. Not that mine is the method of the polymath; it is the method of the constitutionally undecided.

3. My approach in metaphysics is structured by four hypothetical anchor points: fallibilism, pluralism, realism, and localism. None of these provide a foundation, just guesses as to the best way to avoid traps when discussing whatever we might need to discuss.
Fallibilism I take to be the view that nothing we know is certain or complete. And yes, I am not sure fallibilism is true, but if it is wrong, and some little bits of knowledge are certain or complete, we have lost little. For we should assert what is highly reliable or likely, and reliability, the highest possible accomplishment from a fallibilist perspective, is practically not so different from certainty.
By pluralism I mean both that there are many things, of many kinds, and many different ways of experiencing, handling, and inquiring into them. "Thing" by the way, at this point, just means "item," anything discriminable, not entities or substances. We inquire into entities, individuals, ensembles, relations, properties, possibilities, structures, processes, values, etc.
Third is a stronger, less obvious gambit: I assume realism until someone can show me it is wrong. That is, I assume everything about me and my mind and experience are little parts of something bigger; I assume the same of you; and I assume that most of my and your experience and cognition and judgment about the world reveal, in some way and however selectively, real things. They never reveal everything, and they can be wrong, but we already knew that.
Last, localism means that our goal is to understand what is known most robustly. By "robust" I mean what is known by multiple means and multiple inquirers. We have more robust access to what is closer to us. Localism means that what is known robustly or locally is valid even without knowing the most fundamental or most encompassing context in which it obtains. This does not mean truth or knowledge is relative. We are really here attending the MSA: we have lots of robust fallible evidence that is true. But we might upon laborious inquiry discover new things about this local context, and what is outside it, which thrown new light on it. Someone might say this MSA meeting is in fact an artifact of a universal mind or an accidental collocation of atoms. Localism does not deny the possibility that one of those claims is right – it merely says, even if one of them is true, it is still the case we are really here at the MSA. For the local order or robust context is still valid; indeed, we can only judge the validity of the more fundamental or encompassing contexts by their relation to the more robust or more local.
With this comes a denial, or at least an abstemious limit on our diet. We should not expect to come to the popular ends of inquiry: the Whole of everything, simples that compose everything else, or fundamentals that ground everything else. I will presume that we cannot know either the most comprehensive or fundamental contexts of everything. This may seem too large a concession, especially this early. To claim to know the whole is to claim to know something about ALL, and to claim that nothing yet unknown to us will make a significant change to our most inclusive or comprehensive knowledge, and a fallibilist cannot say that. Nor can we believe we know simples. A simple would be something that would remain what it is even if everything else were different, and no fallibilist could claim to know there was such a thing. Nor do we lose a lot if we drop foundations and the whole; we still can claim fallible knowledge, and an unreachable horizon for endless inquiry.
Now, there is only one 20th century philosophy that follows the lead of these four anchor points: for the moment I will call it Objective Relativism. It was in some sense present in the work of early 20th century thinkers like Whitehead, Dewey, and Mead, but was first articulated clearly by Morris Cohen, discussed by Ernest Nagel and John Herman Randall, and only systematically formulated much later by Justus Buchler. Buchler's is expressly a distributive objective relativism, a language by which any thing we can discriminate may be described. (I will explain this more fully later). Objective relativism is not, in my view, an adequate metaphysics: it merely provides the least exclusive way of talking about whatever we need to talk about. Using it as a meta-language, I will then hypothesize that an emergent naturalism, a naturalism that includes the notion of emergence, is the best, most approximately true way to account for whatever has been described in that meta-language. Thus on the basis of a pluralistic and fallibilist approach, which does NOT privilege naturalism, I think naturalism is the best metaphysics. But it has to be the right kind of naturalism.

4. Now let us start over. We have an indefinitely large number of different subject matters, but we must conceive them as falling under some small number of kinds, because we can't think about many things at once. The same goes for a finite but multiple set of perspectives we can take on those subject matters – recognizing that perspectives are themselves rooted in features of subject matters to which they grant priority. Our goal is to understand each subject matter and how it is related to the others. I think of a map of what is as a finite set of collections, which are then connected by relations or lines, forming a kind of net. Some are strongly related, some weakly, some not at all; we cannot assume everything is related to everything without stretching the word "relation" to something negligibly significant. And we assume that there is far more that we don't know about; so the net or map represents or captures a limited depth, a small percentage of what there is to know. What we don't know could be very different from what we know, but we have no practical choice but guess our sample is representative of what is unknown. So imagine the map is a surface of a three dimensional volume. The subject matters on the map are the lugs of the tire. Each is associated with set of methods of investigation and a history and current state of inquiry into it, all of which ought to be taken into account. That is an impossible task for any individual. But we have no choice but to do the best we can.

5. There are doubtless an indefinitely large number of different perspectives in which we can view our subject matters. One set of historical categories or orders of description can be mentioned, not because it is adequate or exhaustive, but to ensure we do not limit ourselves unduly. First, the notion of the linear development of events in its relations to society, nature and gods was one of the early products of human literacy. That is history, or historiography. Things can be regarded as the outcome of, or occupying a place in, a history. Second, what is probably the product of the Axial Age, is logic, by which I mean, broadly, examining the concepts of things and their necessary relations. Much of the history of metaphysics seeks to find the most fundamental or encompassing concepts and determine their essential natures and relations. Third, natural science, which arose in the West in the 17th century, is experimental and naturalistic; it seeks reals acting according to rules which cause publicly available experience. We can find in more recent Western philosophy at least three more perspectives: personal experience, from early empiricism to phenomenology; social agency as the context for meaning, which appeared under many guises until being called "common sense realism" and later pragmatism; and lastly, regarding objects in their role as cultural signs, common to hermeneutics, semiotics, and post-structuralism. That is to say, at the very least, our inquiry can approach the items we discriminate, or some subset of them, in terms of: historical development, the relations among their concepts, their causal physical interactions, as events of a person's experience, in terms of the action of social agents, or in their semiotic role among cultural signs. Our inquiry can be guided by one of these, or any combination of them. We can't consider any of them first or last in an absolute sense, but we can on the basis of some one or some combination argue for the priority of any one of them. As with the subject matters they are connected to, we must use the lug-tightening method, meaning, start anywhere and describe everything from one perspective, then do it again from the point of view of an opposing perspective, etc. Try to come to some organizing, fallible conclusion, recognizing that there is no way to summarize or conclude without prioritizing one of the perspectives.

6. So what we eventually want is something like this: a way of drawing connections among subject matters, employing various and multiple alternate methods, knowing that whatever we thereby reliable comprehend is still only a small slice of what there is to know.

7. Now we start all over again, gaining more detail about the meta-language for connecting and discussing all of this. It could be called "distributive objective relativism," which is a mouthful, or "ordinal pluralism" in Buchler's terms. This regards any item whatsoever as Morris Cohen's put it, a "complex in relations" and applies that analysis distributively to each thing rather than collectively to things as a whole. It may be true that no non-trivial judgment or claim can be made that is not from some perspective; that there is no "view from nowhere." But a distributive approach is a view from anywhere, which is a different thing. So we regard as a "complex in relations" any X, any of X's properties or parts, the process that led to or maintains X, X's environment, etc. We reject any possible end to the analysis, either a Whole of all complexes, or any simple, or something that is the basis or cause of all complexes. Items are only related to some other items, not to all, but we assume no item can be unrelated to some other items.
I might point out, despite what Buchler says, this mode of analysis is not naturalist in any significant sense. It is just pluralist. Objective relativism lets us describe, or talk about, anything whatsoever (except the whole, foundations, or simples). It is the least restrictive ontological language we can get. But for that very reason it tells us little: it underdetermines our world. It would equally describe and hold for a realm of ideas or universals, a domain where pure spirits communicate telepathically, a planet on the back of a turtle, a universe of atomic billiard balls, or a quark plasma. Our world does not seem to be like any of those.

8. Now begins the naturalist hypothesis. I will shamelessly say that we have lots of robust evidence that modern natural science must be true in some sense; we can't think it is completely, utterly wrong. But the same must be said about our personal experience, the social world of human agency, and features of reality prominent for other perspectives. So one significant question is, can a naturalism that takes modern natural science seriously both be based in objective relativism AND be compatible with essential knowledge that comes from those other perspectives? The search for such a naturalism is not, of course, new, but I think starting with objective relativism gives my approach a leg up. My claim will be that an emergent naturalist metaphysics can overcome the problems of other naturalisms and be best, most complete, account of whatever can be discriminated and described using the meta-language of objective relativism.
We can start, like Aristotle, with entities. Given our objective relativism, entities have no ontological privilege over process, relations, universals, etc. But there are entities (we fallibly know, because we discriminate them), and it is easiest to start with them.
We can regard each entity as a system. There are kinds of systems, and systems have properties, or traits. Some systems have a high degree of coherence, making them individuals, like most organisms, planets, and atoms. Individuality, like entity-hood in general, is a matter of degree and is maintained only over some set of conditions. Some systems have minimal structure, making them ensembles, like clouds or ecosystems. Some do not have true components at all, like fields.
Not everything is a system, of course. Hydrogen is not a system, but a kind of system. Triangularity, velocity, liquidity, desire, and happiness are not systems but properties of systems. So, somewhat like Aristotle, we could say we have systems, properties of systems, and kinds of systems. But here we might add something provided by our objective relativism: things, relations or structures, and events or processes must be equally real and fundamental. So should say that a system is equally and co-primordially an entity composed of other entities called components, a set of processes that maintain it, and a structure or organization of components. Neither entity nor process nor form is prior in general. In the kinds of systems to which we have the most robust access, just as we do not find systems without structures or not undergoing processes, we also do not find processes without something undergoing the process, or structures without something structured. The causal and explanatory role of each can be more or less important depending on the particular case.

9. Emergence
So, natural systems, whether individuals or ensembles, typically have components, a structure, and undergo a process – which keeps the components in the structure – in interaction with comparable systems, and are themselves components of more encompassing systems. In each case the system could be more or less structured, and its behavior more or less sensitive to some range of environmental conditions. Some system properties may be adequately explained by component properties and their interaction rules. That is reductive explanation. But some properties can only be explained by the system's interaction with comparable systems; I call that systemic explanation, explanation at the level of the system. (To explain the dent in my fender by the impact of the other car is not reductive, it is systemic.) Other properties can be explained only by the system's role in an encompassing system; that is functional explanation. A single system may exhibit all three types of properties. Whenever a system property cannot be reductively explained, that is emergence.
Here are 18 donuts. Together they are a system which has the property of triangularity. The same 18 donuts could be arranged to make a rectangle. In each case the system has a property none of the components has: triangularity or rectangularity. The reductionist wants to explain the triangle, or rectangle, by the properties of the individual donuts, which would mean specifying the spatial location of each donut, e.g. one has the property of being one inch from the left-hand margin of the slide and five inches from the bottom, the next one to the right has a position slightly to the right, etc. (In effect, that means downloading relations onto the individual parts.) That would indeed determinately construct the triangle or rectangle. But what explains the donuts having those spatial properties? Neither the nature of donuts nor their interaction rules. The organization of the whole is why each donut is where it is; and in this case, I am the cause of that organization. The triangularity cannot be explained without reference to items other than the properties of each component. When the properties of components are causally explained by the system or its role in a larger system – one might say, when the contribution each component makes to the system is itself causally affected by the system or its environment – that is called downward causation.
I should say that the discussion of emergence and downward causation in philosophy has been largely mucked up by the obsessive concern with one particular issue: the emergence of mind, or more precisely the human mind, and mental causation. The human mind presents special problems. If we could suspend our narcissism for a bit and put this issue aside, it would be obvious how extremely implausible a consistent reductionism would be, on purely natural scientific grounds. Does anybody think the properties of the quarks and leptons that constitute a live beaver – including their momentum and spacetime location – could be explained without reference to any of the concepts of biology? Ethology, ecology, zoology, and a host of other ologies, along with a particular historical, functional principle, called natural selection, will all be required to explain the very existence of that collection of quarks and leptons we call the beaver. Is anybody waiting for quantum field theory to explain natural selection?
The point is that reduction and emergence are not in conflict. Each is a matter of degree. Reduction is crucial; components matter. There are mechanisms in the world. Some system properties are almost always explained by component properties. My mass is nothing but the mass of my components, at every level – that is, the sum of my subatomic particles, my chemical substances, my organ systems and fluids, etc. But my life is not, since my sub-cellular (e.g. molecular) components are not alive. As the somewhat gruesome joke goes: place a live mouse in a blender and turn it on. The biophysicist is interested in the system properties that are conserved in that transformation. Life will not be one of them. But neither do we know any living organisms that are not also chemical, material systems.

10. Emergence implies that natural systems are hierarchically organized. Systems with irreducible causal properties studied by one science are asymmetrically dependent on, or composed by, systems with different properties studied by a different science. The science that studies the littlest components of things, and the largest background context in which things arise, is called physics. These smallest and largest systems are relatively simple, far simpler than the things based on them. The science of matter is not physics, it is chemistry; physics studies a lot of things that are not material and have no mass, like spacetime and photons. Rocks, which have to be studied with chemistry, are far more complex than and subject to process that do not apply to their microphysical components, just as cells have to be more complex than their chemical parts.
The different kinds of systems in the universe fall into orders of different scales that exhibit different kinds of processes, forces, and energy levels. A pond can freeze but its atoms cannot; freezing is a process that applies to the relations among molecules. The relative independence of entities and processes at different scales is a crucial feature of the middling scales of nature, that is, the objects of geology, biology, and psychology. These are much more complex than their physical and chemical components. And they have to maintain their existence throughout great variation of those lower level components. What is called multiple realizability requires that some wholes maintain themselves despite chaos or noise at lower levels, so that the particular set of components or conditions on components are irrelevant to the stability of the system. The pond remains the pond despite molecular chaos, as I remain me despite vast turnover among my cells and my microbial biome. Sometimes God, or the devil, is in the details, but sometimes the details are noise. It must be so. As the physicist Richard Feynman put it…

11. We are only now ready for the center of gravity of the theory: I think we can include all discriminable complexes in a naturalism that accepts emergence and regards nature as comprised by at least the following five orders: physical, material, biological, mental, cultural. This is one set of robustly distinguishable orders, related in multiple ways, but in particular by emergent dependence. I do not have confidence that these are the proper or only five; we could certainly divide nature much more finely. But I think we need at least five.
The physical is, roughly, the objects of basic physics, meaning microphysics, the theory of spacetime, and thermodynamics. The material is the types of atomic matter and the systems composed of them, the most impressive cases of which are star systems. The biological is cellular and multicellular life. Animals with sufficient neural complexity have minds, hence feelings, perceptions, and experiences. Humans alone make things that mean, that is, are cultural.
This hypothesis of a small set of hierarchically arranged orders is particularly robust because it happens to match the current scientific view of the temporal evolution of our universe. There was nothing we commonly call matter when our universe began approximately 14 billion years ago; that is, no atoms. They, and hence everything we call material, arose only much later,, most impressively, star systems. Billions of years afterward at least one rocky planet became the stage for life, which over four billion years evolved millions of different and increasingly complex forms, eventually mind. Half a billion years later there evolved at least one hominin species with culture. Even if we find life and culturally intelligent life elsewhere in the universe, it is difficult to imagine a very different sequence.

12. So one way of depicting the five orders is something like this. The orders of nature exhibit both a structural and historical hierarchy. The narrower depends on the historically earlier and structurally more basic. Of course, as we saw, there is downward causation. Living cells produce the macromolecules that constitute them. Nevertheless, the molecular components of those macromolecules can exist without life, but not vice versa, and they came first.

13. Here is the more detailed picture, but less colorful. Notice: a) the physical expansion of the universe goes off to the side; b) the line about the physical is dotted; I am not claiming there is nothing independent of the physical universe; c) time and complexity are inside, but I couldn't draw them inside. d) Also note that this account requires accounts of each of the five orders, their properties, and how they interact. The validity of my overall account depends on the validity of all those subordinate accounts. Any could be wrong. We can talk about any of them.

14. So that is the System in Pieces. It is not based on a foundation, on simples, on the priority of one kind of thing or one non-distributive concept, or a claim to know the whole. It could fit inside a more comprehensive vision, if that were justified. Each of its modular orders is open to revision. The account is built to accept additions and remodeling. But I think is a start.