Ontology concerns the way things are in actuality. Science, by contrast, attempts to establish principles that lead to predictions which are consistent with how the universe behaves; building models of reality which have predictive power. There is an important distinction here. Ontology concerns a description of the totality of existence, of which a subset is the universe. We can conceive of other universes which do not in any way interact with our own or have any implications on our own in terms of logical necessity. They are conceivable, although fundamentally unknowable.
I cannot conceive of an experiment that proves the definitive existence of a universe, let alone a logically consistent universe which is amenable to understanding. Nevertheless, assuming there exists something external to oneself (and also that one’s self exists), we have methods of building models of reality. These models of reality, where they move beyond merely empirical descriptions, posit a specific view about the way of things. In our daily lives, we generate a world view, a model of the way things are. A scientific world view provides a subset of a coherent view; it is part of a world view that is consistent with the data that have been gathered, and has methodologies to prevent self-deception.
In what follows, I propose an a la carte view of ontology, where a specific ontological view is chosen within a specific scientific framework, provided it is logically consistent to do so.
Limitations on ontology
Ontology, in its broadest form, is the attempt to uncover the singular true model of reality that captures an understanding of the fundamental nature of reality. It seems then, that it must be the case that a true understanding of the ontology of reality is fundamentally unrealisable for three reasons.
Firstly, Science provides a specific subset of knowledge. A world view which makes claims in the domain of science, but which is itself not viewed as scientific, can be refuted on scientific grounds. That is, science creates coherent descriptions of reality, makes predictions which fit the data we have thus far and has some predictive utility. In the worst cases, limitations are acknowledged when these are known and the theory fits some empirical window.
For example, we can take astrology, many of whose proponents claim is not a science whilst others claim it is. Assuming we are charitable, we will grant, temporarily, that it is logically consistent and that the astrologers all talk with one principal voice when making predictions. Each specific astrologer can sometimes be convinced to make specific predictions (as in the famous Carlson test), which can then be compared with the empirical results, and thus support, or refute, the astrological hypothesis under consideration. Of course, someone might note that auxiliary hypotheses could be invoked etc to save the “theory”, but in practice this doesn’t happen with most pseudoscience. The proponents merely reject the existence of the results, make an ad hoc rationalisation to explain away (by post-hoc explanations) a specific result that allows their cognitive dissonance to dissipate, and continue on exactly as before. They ignore falsification rather than adapting their hypotheses, or as in the case of Gauquelin’s Mars is good for sports claims, they cherry pick the piece that agrees with themselves, whilst ignoring the part that completely refutes their proposal; a classic demonstration of confirmation bias. Kuhn pointed out that astrological endeavours merely involve explaining away events and failure, but never actually solve puzzles; there has never been any advancement, historically, on problems to astrology which are raised by empirical reality. Astrologers make predictions which are inconsistent with each other and at odds no greater than chance. Since astrology lacks a plausible mechanism it also provides no interesting insights whatever. One can only conclude that there is no utility, in explanatory terms, in astrological belief.
Secondly, coming back from the segway, it is possible for different assumptions about ontology to have the same implications, the various interpretations of Quantum Mechanics being the obvious example. For any specific view about reality, it is possible for another view to have exactly the same predictive power in terms of observation. For another example, conceive of some view about the nature of the universe, then add the additional premise that there exists another separate universe which does not interact in any way with this. Now, of course, there is no scientific utility to this since it does not increase the predictive power of the proposal nor make any difference to anything whatever, but it is conceivably the case. Since it is epistemically unwarranted to either dismiss or accept the existence of the conceivable other universes, one should augment this view of reality to highlight this uncertainty; thus realising no complete knowledge is possible about things.
Thirdly, while there is a fact of the matter with regards to science, there is no possible method of having full epistemic certainty of whether we actually have grasped the solution. This is because the only reliable method of learning about the nature of the universe is through experiment, which provides specific data about reality,
A la Carte Ontology
I think there exists an external universe, which I am in, and I’d wager you do as well, since you are necessarily interacting with it and assume it’s existence if you are reading this. There is no logical proof of it’s existence through reason alone, nor that what happened before is an indicator of what will happen in the future. That is, by interacting with me, you are conceding that, whatever your view of reality, acting as though you can communicate with me is a more likely view of reality than one in which I do not exist. If you thought it was more likely that you were the only thing in existence, say, there would be no reason to eat food, or talk to others, as well as no normative distinctions in your actions since it’s mere inventions of your mind. You necessarily assume it or else you don’t live for long; you don’t stick your hands in the fire, because you expect to get burnt.
As mentioned earlier, a model can not definitively establish whether it’s ontological assumptions are necessarily valid and unique (in that there are no rival assumptions which lead to the same conclusions). This is not a problem for science though. The conceptually simplest model should be used for a discussion. For example, when discussing standard atomic physics, where some interaction takes place and some observable is measured, a Copenhagen-like interpretation is perfectly serviceable as a framework for discussing what is happening in an experiment, for someone discussing classically sized objects, it makes perfect sense to talk in a Newtonian framework, while for a galactic astrophysicist, it makes sense to think in the conceptual framework that comes with general relativity. There are 3 examples out of many, but they also highlight something important. While currently there is an incompatibility between quantum mechanics and general relativity, even if that were resolved, and there was a grand unified theory, it may still be useful to keep exactly and unmodified the same ontological pictures as before in the specific domains of study, just as we keep a classical picture in our heads in the physics of the everyday.
The scientist should choose a metaphysical framework which has utility in being the conceptually easiest metaphysical view he can think of in which to explain the theory. The theory doesn’t change, rather the way we tell the story changes; the science is the bones of the explanation while we use the metaphysics to flesh out things. The language may then, not correspond to some fundamental nature of reality in physics, but nor does any language in any specific scientific discipline and they appear to get along just fine. There is no concept of biology at some fundamental constituent level in the physical world, rather biology is what occurs through a physical medium but which can not purely be described in terms of that physical medium (i.e I think the strong reductionist thesis is fundamentally impossible) But it is certainly meaningful to label things as people, and chairs and tables even if they don’t have an existence beyond being an aggregate of particles; the concepts have scientific utility.
So we have again chosen an ontological view, whilst ignoring the actual nature of reality and no matter that a deterministic view of, say, choices made, is seemingly at odds with a possible fundamental indeterminism in quantum mechanics. Neither can say something fundamentally true about ontology, except to say what is probably not true.
Of course, a minimal ontological view itself is likely to cause unanswered issues in itself, and in this way, may itself prove an interesting venue to fuel research to see where that analogy becomes a disanalogy.
1. Incidentally this causes confusion for one particularly badly thought out demarcation of science which demarcates non-science from pseudoscience on the basis of whether proponents claim they are doing science.
2. As an aside this also refutes the intelligibility argument of Pope Benedict XVI, which is premised on the claim that we can have knowledge of the fundamental nature of reality.
edit: Corrections by Anto and Cormac.