Work in Progress: Theory-Dependence and the Web of Belief
Friday 21 October 2022
I have been reading one of my notebooks from a dozen years ago, and I was interested to see that I made some comments about civilization at that time, when my conception of civilization was essentially inchoate, and, looking back from the present, I would not have thought I would have had anything worthwhile to say about civilization at that point in my life. For instance, near the beginning of the notebook I wrote this:
The idea of civilization unfolds as slowly as the concrete embodiment of the idea, and indeed each of these developments influences the other, so that the idea of civilization drives the practice and the practice drives the development of the idea.
And near the end of the same notebook, I wrote this:
There is a sense in which no civilization is based on an idea insofar as civilizations emerge organically from the way of life of a people, and the way of life of a people in turn emerges from a particular landscape with a particular climate — in short, the ecology of a particular biome. But implicit in a landscape, a people, a biome, an ecology, are those ideas eventually given voice by the people emergent from a region.
These are both sensible statements that I could endorse and defend today, even if I might formulate them a bit differently, and place them in a slightly different context. One of the ideas I was working on a dozen years ago was the idea that civilizations were based on an idea — I hadn’t yet adopted the terminology of central project, but that is more or less what I meant — and ideas could be incommensurable, therefore civilizations could be incommensurable. Without my knowing it at the time, I had restated a central thesis of Spengler about civilization.
The idea that civilizations could be toto coelo different from one another implies that there really couldn’t be a scientific study of civilization, and if we combine this idea with the idea that history has a distinctively ideographic method (distinct from the nomothetic method of the natural sciences), and that civilizations are intrinsically historical existents, then everything about a civilization defies the possibility of scientific schematism, and civilization could never be the basis of a body of scientific knowledge about civilization, except for the historical knowledge of civilizations, which is how civilizations have been studied and understood anyway.
I could have eventually taken this road, and used these arguments to defend the ideographic nature of civilizations and their histories, and used this as an explanation of why no science of history or of civilization has appeared. Indeed, insofar as no rigorous science of history or civilization has yet appeared, this argument might be held to be persuasive as it has the bulk of evidence on its side.
The idea of civilizations being based on radically different ideas is given expression in Ruth Benedict’s Patterns of Culture, where she quotes Spengler and emphasizes the contingency of the apparently trivial factors in human experience that get built up into grand cultural narratives with attendant rituals and ceremonies, and the trivialities of life built up into a culture by one population might be nearly totally neglected by some other population, which picks some equally trivial contingency of human experience as the basis for its culture.
From my present point of view, I would point out that all of this can be true, and yet all civilizations have institutional structures that can be identified independently of the cultures that supply all the colorful details of these institutions, and make them appear so different and so unique to the anthropologist. So while I was previously, without knowing it, reiterating Spengler’s account of “destiny ideas” that shape civilizations, or what Henri Frankfort calls the “form” of a civilization, and taking these things to define a civilization, I would now say that it is the institutional structure that defines what a civilization is — the genera of civilization — while the idea or central project is the differentia of each species of civilization. In this way I arrive at an Aristotelian definitional schema for civilization.
I frequently point out that there is no science of civilization, and no definition of civilization with any degree of consensus across disciplines. (There are definitions specific to particular sciences, but these sciences use the concept of civilization as a means to end of their discipline, and not as an end in itself.) Above I have suggested an argument by which this position might be defended, but what I came to realize in the past week, after both my reading of my notebook of a dozen years ago, and re-reading parts of Benedict and Frankfort, is that interesting, important, and relevant things can be said about something that we cannot yet define.
Once I made this idea explicit for myself, it seems blindingly obvious in retrospect, as I have often pointed out that biology has no consensus definition of life, psychology has no consensus definition of mind, and sociology has no consensus definition of society. All of these disciplines are almost entirely at sea when it comes to their foundational concepts, and yet there has been progress. One might quibble here, given the replication crisis, but certainly biology has made remarkable progress, even if progress in psychology and sociology has been slower and more ambiguous.
Not only do biologists disagree on the definition of life, they don’t even agree on a definition of the central concept of species. I have read Marc Ereshefsky’s The Poverty of the Linnaean Hierachy with its critique of a univocal species concept, and I find this persuasive, but I have also read Ernst Mayr’s response to criticisms of his definition of species (in What Evolution Is, which I discussed in newsletter 140 and 141, inter alia), which critiques he seems to regard as trivial, and I find this persuasive also. I would have to spend a lot more time on philosophy of biology before I could assess these conflicting claims on a level at which I could understanding the various theoretical motives that underlie the differing positions on the definition of species.
The above being said, that even a science as advanced as biology cannot define life or species, and yet it seems to made epistemic progress, I think that there is a different kind of scientific progress that happens when fundamental ideas can be given a reasonably unambiguous definition. Probably the history of clearly defined sciences is more linear, while the histories of less clearly defined sciences is closer to exemplifying Kuhnian paradigm shifts and revolutionary change, and, when there is a revolutionary change in science, much more is cast aside with less well-defined sciences, whereas the more clearly defined sciences retain more of their concepts, which might be said to be symmetrical in regard to theory change.
A couple mornings ago I woke up thinking about related matters of revolutionary science, but I failed to take any notes upon getting out of bed, and it soon left my mind. But now, writing this, some of it is coming back to me. Before going to sleep I had been reading Peter Achinstein’s Concepts of Science: A Philosophical Analysis, which I only recently acquired. It is a satisfyingly rigorous text. I had been reading the section of theory-dependence and I realized that there seems to be a relationship between theory dependence and something like what Quine called the web of belief.
Quine argued that some concepts are nearer the center of our web of belief, and we should hesitate to revise them unless absolutely necessary, because if you make changes in the center of the web of beliefs, this affects everything else in the web. One ought to first try revising beliefs closer to the outer edge of the web of belief, only moving inward closer to the center when absolutely necessary.
One also could express this idea hierarchically, rather than in terms of a web, but asserting that some claims lie at the foundations of knowledge while other claims are higher up the chain that leads down to foundational ideas, so that we ought to first try to revise later and somewhat superficial ideas before we start to mess with the foundational ideas.
However we choose to express this, we can see that either the center of the web of belief, or the foundations of our epistemic hierarchy, are ideas that are widely used in many different theories, so while they are theory dependent in a sense, since they are shared by many theories if any one theory has to be abandoned, we still retain these central or foundational ideas, even if they have a slightly different cast in different theories. Ideas like temperature and mass may be given highly specialized definitions within different theories, but we retain the ideas of temperature and mass across many different theories and these ideas possess a kind of conceptual symmetry so that they are, at least to a degree, invariant in respect to theories.
What I am suggesting here is that we could map a web of theory-dependence, or a hierarchy of theory-independence, so it isn’t a simple matter of being theory-dependent or not, but a matter of degree of theory-dependence. Ideas that are near the center of the web of theories, or which lie at the epistemic foundation of all theories, are very little theory-dependent, whereas if you move outward from the center of the web of theories, or upward away from the foundations of our knowledge, you come to progressively more theory-dependent concepts.
The concept of life is an inexact but foundational (or central) concept of biology, and is found alike in Aristotelian biology and in contemporary astrobiology. It is at the foundation of knowledge, or the center of the web of belief, but it remains opaque despite being foundational or central. The concept of temperature is now understood as the excitation of atomic structures, and the clarity with which it is understood affects all the structures that are built upon it, which extend from its central position in the web of theories. Both concepts are central or foundational, but one is clearly defined and the other is not.
However, as we can see, not being clearly defined does not exclude an idea from a foundational or central role in theory, though it almost certainly bears upon the use to which that concept is put. Both the clearly defined and the inexact central concept can be equally theory-independent (relatively speaking), though I suspect that the theories built on inexact foundational concepts diverge more rapidly and more radically than those built on clearly defined concepts.
The concept of civilization is, like life, a central or foundational concept, though unlike life it is a concept for a discipline that does not yet exist. Biology can make progress despite controversies over the definition of life and of species. I would even say that one of the forces pushing biology toward conceptual clarification, however slowly, are the debates over the species question. Thus the study of civilization could make progress despite disagreement over its definition, and the conceptual clarification of a science of civilization could follow from controversies within the discipline. My reflections on civilization from a dozen years ago might yet have some value, even though I had no guiding idea at the time for the study of civilization. And yet, we have a science of biology and no science of civilization. Why is that?