If someone claims that X exists, where X could be, for instance, unicorns, that person is making an ontological statement, that is a statement about what does or does not exist. Ontology is a branch of the broader field of metaphysics, which is concerned with our understanding of the fundamental nature of reality. Part of that understanding, naturally enough, are statements about what is / is not part of that reality.
Now, if you say that unicorn exists, I am well within my rights to ask you: how do you know? In other words, I am asking you to provide your epistemic warrant for such an assertion. Epistemology, then, is the branch of philosophy that is concerned with the nature of knowledge.
Nice. Now, why on earth am I telling you this? Because people, including some professional philosophers, seem confused about the relationship between metaphysics and epistemology. And this confusion leads to all sorts of misunderstandings. Take, for instance, a recent discussion I've had over at Letter.wiki with my colleague Philip Goff, concerning the suddenly controversial topic of panpsychism.
Panpsychism comes in a variety of flavors, but one of its basic assertions is that consciousness is an elemental property of the universe. It isn't entirely clear what this means, but apparently it doesn't mean that electrons can think. Whatever your opinion about this debate -- and I urge you to read both my four letters to Philip and his four addressed to me -- you would probably agree that I was well within my right to ask him: how do you know that?
After a bit of back and forth, Philip conceded that there is no conceivable empirical evidence that would confirm or disconfirm the ontological notion that consciousness is elemental. I replied that if that were the case we had pretty much reached the end of the discussion, since one can make up an arbitrary number of ontological claims without the constraint of empirical evidence to back them up. Philip replied that I was being a "physicalist," and that my epistemological query assumed a particular metaphysical position. Since he rejects that position, we were at a standstill.
Let me restate the problem clearly: Philip argues that there is no epistemology without an underlying metaphysics. Conversely, I suggest that there is no metaphysics without an underlying epistemology, or -- at the very least -- that one's metaphysical claims should never be too far from one's epistemic warrants for those claims. Let's look at this a bit more closely.
The term "metaphysics" apparently comes from an anonymous first century editor of Aristotle's work, who collected certain entries under the heading "ta meta ta phusika," which literally means "after the Physics," with reference to another of Aristotle's works. Ever since, the major topics characterizing metaphysics have been the already mentioned ontology (what does or does not exist), the nature of identity and change, the nature of space and time, the nature of causation, and the difference between necessity and possibility.
Crucially for my argument, most metaphysicians themselves recognize that their work has to rely on some sort of epistemic foundation (and not the other way around, as Philip maintains). Such foundation lies in the deductive method of a priori reasoning, as distinct, say, from induction or from abduction (weird term, nothing to do with aliens and their anal probes, otherwise referred to as inference to the best explanation, or IBE).
In other words, metaphysicians ever since Thales of Miletus (the guy who suggested that the world is, ultimately, made of water) work like mathematicians and logicians: they begin with certain axioms, or assumptions, and logically deduce the consequences of such axioms. The major difference between metaphysicians and mathematicians or logicians is that the latter may do all their work with no reference whatsoever to the empirical world (though it turns out that both logic and mathematics are useful to scientists). Metaphysicians, by contrast, attempt to understand reality as it actually is, not the infinite ways in which it may be.
There are fundamentally two modern, highly contrasting, conceptions of metaphysics. The classical one, often referred to as "first philosophy," claims that metaphysics is in the business of understanding mind-independent objects and properties that are foundational to science and common sense. (Here is an example.) For instance, when scientists claim that electrons are objects characterized by the property of charge, metaphysicians inquire on the nature of the categories "object" and "property."
The second view, sometimes referred to as "scientific metaphysics" (example here), assumes instead that metaphysicians study human constructs: notions like "object," "property," etc. don't exist "out there," so to speak, but are useful inventions that human beings deploy in order to understand how the world works.
First philosophy type metaphysics is foundational to science; scientific metaphysics, by contrast, takes science as providing the raw material on which to work. Both of them, I maintain, assume a certain epistemology, because for both kinds of metaphysicians the question "how do you know that?" is perfectly intelligible.
The term "epistemology," was introduced in modern philosophical parlance by James Frederick Ferrier in his Institutes of Metaphysics, published in 1854. There he writes: "This section of the science is appropriately termed the Epistemology—the doctrine or theory of knowing, just as ontology is the science of being. ... It answers the general question, 'What is knowing and the known?'—or more shortly, 'What is knowledge?'"
Notice the contrast with ontology, and the fact that epistemology is a translation of the German wissenschaftslehre, which literally means theory of science, further confirming my take that it is epistemology that is foundational to any discipline seeking to make knowledge claims, including of course both science and metaphysics (and mathematics, and logic), however we may conceive the latter.
Going back to my discussion with Philip Goff, then, it is an effective but misguided rhetorical move to claim that I cannot ask for evidence for panpsychism on the grounds that in even making such a request I am assuming a particular metaphysical view, therefore begging the question. To drive the point home, consider two sharply diverging metaphysical frameworks and how they relate to evidence: physicalism and idealism.
(Note that it is not at all clear what metaphysical framework is represented by panpsychism, as different authors have wildly divergent opinions: Galen Strawson says that panpsychism is a form of physicalism; David Chalmers describes it as an alternative to both materialism and dualism; and Philip Goff maintains that panpsychism is an alternative to both physicalism and substance dualism. Could such confusion among proponents be a reflection that the concept itself is hopelessly confused? I wonder.)
Physicalism is the metaphysical thesis that everything in the world is either physical or "supervenes" on the physical (meaning that any change in a non physical thing, like a concept, or a mathematical construct, is the result of an underlying physical change, for instance a human brain). It is a form of ontological monism, claiming that there is only one kind of stuff at the bottom of everything. That stuff is assumed to be whatever the best current guess of fundamental physicists is (particles, strings, fields or what not). Physicalism is, for all effective purposes, the same thing as materialism, in a philosophical sense or, better, it's a scientifically-guided extension of materialism, because physics has showed us that there is more than matter (e.g., energy) in the cosmos.
Idealism, by contrast, is the metaphysical thesis that everything in the world is the result of thought, though idealists differ sharply on whose thought, exactly, we are talking about. It could be human thought (subjective idealism) or the thought of God or similar transcendent entity (objective idealism).
Now, neither physicalism nor idealism can strictly speaking be subjected to empirical confirmation, because both frameworks are completely compatible with the facts of the world as we see them. This is why the famous "refutation" of idealism by Samuel Johnson is cute, but makes no sense. In discussing idealism with one of its major 18th century proponents, George Berkeley, an exasperated Johnson kicked a stone and grandly uttered the words "I refute you thus!," meaning that the stone had to be physical, and not a product of thought, or he couldn't kick it.
But idealists maintain that the world would appear to be as it does regardless of whether it exists in the mind of God or not. Causes would still precede effects, and kicks would still move stones -- if the stones are not too large. The problem is that Berkeley's victory was a pyrrhic one, because if there is never a way to empirically distinguish among different metaphysical frameworks then the entire project of first philosophy becomes utterly irrelevant, a parlor game that nowadays requires a PhD in philosophy. and, I would say, a project that has very successfully delegated to science.
That's a major reason I sympathize with scientific metaphysics, which is in the business of making sense -- in ways that human beings can understand -- of what we know and how we think about the world, taking both science and common sense as its inputs. Individual sciences, such as physics, biology, psychology and so forth, can't do that, because they are by their nature focused on isolated portions of the overall puzzle.
So panpsychism construed as a general metaphysical framework in alternative to physicalism, idealism, and so forth, is at the very least irrelevant. Worse, it's chief specific -- ontological -- claim, that consciousness is an elemental property of matter, is either incoherent (we don't know what it means) or empirically untestable, thereby lacking epistemic warrant.
Contrast this with standard scientific claims, which come with epistemic warrants attached: when a physicist says that electrons have charge and you ask them how they know, they can provide you with a working definition of the property of charge, as well as with replicable measurements of that property, which anyone with the necessary knowledge and instruments can verify. That is neither incoherent nor useless. I rest my case, without having to kick any stones.