Introduction:
Broadly speaking, an argument from underdetermination is one that attempts to show that available evidence is insufficient to determine which of several competing theories is true. That is, many different theories might be able to explain the same evidence, hence any move to choose between theories must be “underdetermined,” i.e., not determined by the evidence. Within the class of such arguments, there are many that go a step further. These will often purport to show that for any body of evidence, there will always be an infinite number of different explanations that are consistent with that evidence.1

Since the start of the 20th century, these arguments have increasingly played a pivotal role in philosophy, often leading to radical conclusions. Yet even those who have resisted such conclusions have often found these arguments difficult to refute. Such arguments were not unknown to pre-modern thinkers, and yet they were generally considered fairly easy to dispatch. Hence, we are left with a bit of a mystery. Why did arguments from underdetermination begin to seem insurmountable?

In this post, we will try to answer this question. As we shall see, it is not that the ability of underdetermination to produce support for radical skepticism was lost on pre-modern thinkers. Rather, it was that different starting assumptions tended to render such arguments fairly innocuous. To illustrate this point, we shall explore a typical Thomistic response to arguments of this form.

Underdetermination and Skepticism:

Arguments from underdetermination have played a defining role in contemporary thought. There is for example:

• David Hume’s argument against causal inferences and explanations, as well as his hugely influential “Problem of Induction;”

• Ludwig Wittgenstein’s rule-following argument, as well as Saul Kripke’s influential reformulation of it;

• W.V.O Quine’s argument for the inscrutability of reference;

• Quine’s holist arguments for the underdetermination of theories by evidence, as well similar arguments for forms of theoretical underdetermination made by J.S. Mill and expounded upon by Pierre Duhem;

• Thomas Kuhn’s arguments about underdetermination at the level of scientific paradigms;

• As well as many others, including Feyerabend’s “epistemological anarchism,” Goodman’s “new riddle of induction,” etc.

It would not be an understatement to say that these arguments have been, in many ways, the defining feature of the Anglo-American, empiricist-analytic philosophical paradigm. It is also hard to overstate the radical nature of some of the conclusions that have been drawn from these arguments. Bertrand Russell, for instance, suggested that if Hume’s “Problem of Induction” could not be resolved then “there is no intellectual difference between sanity and insanity.”2 Richard Rorty and others, drawing from Quine and Wittgenstein, argued that truth and knowledge themselves would have to be radically redefined. “Truth,” rather than “being the adequacy of the intellect to being”—the view held, in varying forms, for most of history—would instead have to be understood in terms of a new belief’s coherence with our preexisting beliefs, or as merely a compliment we pay to beliefs we find good to hold. Similarly, these arguments have been used to argue against even modest forms of scientific realism, resulting in a radical anti-realism, which claims that scientific truth is merely a social label bestowed upon well-entrenched practices.3

Obviously, not all thinkers within this tradition have agreed with such radical proposals, yet they have often found it hard to counter them. Were a typical Enlightenment or Victorian-era philosopher transported to our time, I do not think it is a stretch to say they would find these skeptical conclusions shocking, a radical reversal of the faith in reason and empirical methods that dominated in their epoch. So too, we can well imagine that pre-modern scholastic thinkers would be astounded by shifts such as the attempt to eliminate causation from science (the very thing that, for them, defined science), or the various moves to radically redefine truth.

On this last move, the redefinition of truth, it is worth noting just how radical the paths that have been charted in contemporary thought might seem. We could consider here how Quine's more radical successors took their epistemological holism to require a shift to a coherence-based, pragmatic model of truth, where truth is a function of how well a new belief fits in with our “web” of preexisting beliefs. Such changes, and there have been many suggested, are of course, rarely presented as rejections of truth, but instead as necessary redefinitions (this being true even of most proponents of deflationary theories of truth). Yet we could well imagine our time travelers objecting: “but this is just equivocation. Truth is thought’s grasp of being. These redefinitions amount to denying truth, and then recommending some other thing to fill the gap. The equivocation just softens the blow. On your view, we are no longer capable of scientia (knowledge), but merely a system of conceptual and linguistic manipulation.” That is, the charge would be that this is really just radical skepticism and epistemic nihilism masquerading as an a mere “adaptation.”

However, there is one group who might not be particularly surprised by this course of events: the original empiricists. After all, Sextus Empiricus, from whom the name “empiricist” is derived, was part of the explicitly skeptical Pyrrhonist school of ancient thought. A key idea within that school was that of equipollence. Equipollence exists when opposing arguments or theories appear to have equal strength or plausibility. In this way, it is quite similar to underdetermination. If we have attained equipollence, then we are not led to assert one conclusion over all others, and so we are not compelled towards any particular belief.

Rather than viewing equipollence as a sort of epistemic crisis (which was often how arguments from underdetermination were initially viewed in modern thought4) these skeptics strove to attain equipollence. The idea here is fairly straightforward. If we are not forced to any particular conclusions, then our passions will also not be raised by any particular conclusion. This allows us to achieve ataraxia, a state of serene calmness. By contrast, strong beliefs were thought to promote anxiety and distraction, the opposite of the state of dispassion sought by these thinkers. To be clear, the argument here is not that all arguments are equally strong, or that truth is, in principle, impossible to attain.5 It is rather that we can reason to the view that many arguments seem equally strong (which in turn does put the reliability of most truth claims in question).

It is useful to clarify here what made “empiricism” unique. It is not that it used sense observations to try to explain the world. If this alone made a philosophy “empirical,” then essentially all philosophy is empiricist. Nor was it a commitment to the proto-scientific methods of the ancient world. Aristotle, labeled as a dogmatist by the ancient empiricists, is often considered to be the father of many sciences. Of course, modern empiricists do sometimes claim Aristotle as one of their own on account of this fact. Yet, if empiricism is construed in these terms, then the overwhelming majority of all thinkers would be "empiricists." Scholasticism and its modern variants would be “empirical,” as would Hegelian philosophy, etc.

Rather, what made the ancient empiricists unique, aside from seeking equipollence, was their means for doing so: rejecting causal and metaphysical explanations. That is, they stuck to descriptions of phenomena and observed patterns. The early empiricists were physicians who employed a practice of recording case studies in order to keep track of patterns. Rather than posit causes that would explain why some cases improved and some did not, or why some treatment was efficacious, they simply guided their practices based on the consistency of the patterns they observed. This allowed for conclusions like: “treatment with this herb works consistently in these cases,” but not explanations for why this was so. Prima facie, this is actually somewhat at odds with the “scientific method” as popularly conceived, since one could hardly develop hypotheses to test without violating the prohibition on causal theses.6

Hence, it is perhaps unsurprising that a later philosophical tradition modeled on these thinkers would eventually find itself mired in (or perhaps, “set free by”) skepticism. Indeed, this connection was clear for some in the tradition, such as Hume, although it certainly became obscured as “empirical” was increasingly used as a synonym for “a scientifically-informed philosophy and world-view” (with “science” overwhelmingly being understood in realist terms in the 19th and early 20th centuries).7 For some, this skeptical shift, and a move towards anti-realism has been a positive outcome. Somewhat akin to the ancient empiricists, these thinkers think that the death of realism and certainty is an improvement, often because it is thought to “free us” by allowing us to think in new ways. Against this though, we might consider that, if realism is the case, our ignorance of reality would itself represent a limit on freedom.

At any rate, for most, the skepticism resulting from underdetermination has been seen as a serious threat and challenge. Yet responses have often involved “work-arounds” that might themselves been seen as skeptical. Here, it is worth employing Saul Kripke’s distinction between different responses to skeptical problems: straight solutions and skeptical solutions. The “straight solution” aims at showing that the skeptic is simply wrong, that their skepticism is unjustified. A “skeptical solution,” by contrast, allows that the skeptic has a point, but will maintain that “our ordinary practice or belief is justified because–contrary appearances notwithstanding—it need not require the justification the sceptic has shown to be untenable.”8 For example, Kripke’s own skeptical solution vis-à-vis linguistic meaning claims that language is not meaningful because of metaphysical facts, but only on account of shared behavioral regularities related to social norms and public agreement.9

Again, the difficulty here is that the “solutions” often seem quite skeptical, e.g., “words never refer to things,” “there is never a fact about exactly what rules we are following,” etc. Here, it is worth considering what it is one ought to do if one sees an argument with an absurd conclusion. The first things to do are to check that the argument is valid, and crucially, that the premises are true. I would argue that contemporary thought, particularly analytic thought, has far too often only done the first. Because it holds many empiricist presuppositions beyond repute (indeed, “dogmatically” might not be too strong a word) it has not generally questioned them.

Yet if an epistemology results in our having to affirm conclusions that seem prima facie absurd, and if further, it seems to lead towards radical skepticism and epistemological nihilism, or an ever branching fragmentation of disparate “skeptical solutions” and new “anti-realisms,” that might be a good indication that it is simply a bad epistemology. Indeed, an ability to at least secure our most bedrock beliefs might be considered a sort of minimal benchmark for a set of epistemic premises. Yet, due to the conflation of “empiricism” with “the scientific method,” as well as modern culture’s preference for iconoclasm, novelty, and “creativity,” the starting assumptions that lead to these conclusions are rarely questioned. With that in mind, let us turn to the realist responses that, in prior epochs, made these arguments seem relatively insubstantial.10


The Realist Response:

A Thomistic or Aristotelian response to arguments from underdetermination would deny the assumptions that make such arguments plausible in the first place. It is only because of the presupposition that no metaphysically substantial premise can enter into our epistemology that such arguments are able to gain momentum. This, paired with the demand that all relevant “evidence” must be “third-person” and “external,” arguably makes explaining knowledge (a first-person experience) impossible from the outset. Indeed, it should be unsurprising that such limitations result in theories of truth that claim that truth has no metaphysical import; such a conclusion has been simply assumed as a premise!

These different assumptions show up a key in difference in how sense knowledge is considered. For the empiricist, perception is merely raw, inchoate “sense data,” from which patterns can be derived. For Aquinas and Aristotle, there is a real form in things that is received by the senses, and then abstracted and known by the intellect. The latter process explains the phenomenon of our understanding anything at all. From the realist perspective, we start from a world populated by trees, rabbits, etc. When Quine concludes that we can never tell which word in a foreign language refers to rabbits, because there is always other stimuli that accompanies the presence of a rabbit that could be referred to instead, he is involved in a sort of question begging. He has assumed, from the outset, that there are no things with determinant natures for us to name. Yet if this is assumed, it can hardly be shocking that we are forced conclude that words cannot refer to things, for we have already decided that there are no things to refer to.11

Here, the empiricist might object: “that’s a fine thesis, but how can you prove that this ‘form’ or ‘actuality’ exists?” The key assumptions that underwrite these notions boil down to this:

1. Things do not happen “for no reason at all.” Things/events have causes. If something is contingent, if it has only potential existence, then some prior actuality must bring it into being. It will not simply snap into being of itself. Our experiences are contingent, thus they must be caused by something that is prior to them.

2. Being is intelligible, and to be is to be intelligible. Every being is something in particular. That is, it has a form, an actuality, that is determinant of what it is (as well as the potential to change, explained by matter). This actuality determines how a thing interacts with everything else, including our sense organs and our intellects. If this was not the case, interactions would be essentially uncaused, and then there would be no reason for them to be one way and not any other (i.e. random).

That is it. These are assumptions, but they do not seem to be particularly objectionable ones. Indeed, if they did not hold, if being were unintelligible and things did happen “for no reason at all,” we might suppose that philosophy and science are a lost cause. Of course, much disagreement here results from a misunderstanding of the nature of form, potentiality, etc. Sometimes, these are taken to be something like terms in a scientific theory, rather than a set of metaphysical principles. Yet a thing’s form is simply that which makes a thing anything at all. The basic idea is this. We perceive a world around us. It is not a world of indistinct sense data, but a world full of things, particularly living things that continually act to sustain themselves, to keep being what they are (i.e., to maintain their form). Such perceptions must come from somewhere, since they cannot occur “for no reason at all.” They have causes. Whatever acts on our senses, and is understood by our intellects, has a prior actuality that activates our potential to perceive and understand.12

This is not a naive realism. Perception can be thought of as “of” the interaction between the immediate environment and our sense organs. For instance, light waves carry the form of a tree to our eye.13 Perception is, in a sense, the experience of this interaction, but the interaction relates us directly to the objects known by the senses. Knowledge of trees, an understanding of what a tree is, comes from the presence of this form in our intellect after it has been abstracted from the senses.

A denial of the transmission of form would essentially be a denial of the idea that any prior actuality in the things we perceive and known makes it to our minds. Yet if this were the case, it hardly seems that perception could be of anything. If perception is not caused by the prior actuality of things, it would seemingly be uncaused. More importantly, it could not involve the “appearance of” anything; rather any such appearances would simply be free standing appearances that are “of” nothing in particular (in which case, the appearance/reality distinction seems to collapse, and it turns out that phenomena just are reality).

However, it is clear how confusion can arise here. We do experience error. We can, for instance, mistake a fake apple for a real one. However, crucially, this is because the fake apple has the form (at least partially) of a real apple. Likewise, even if we were to experience seeing an apple due to some sort of electromagnetic stimulation of our visual cortex, we would still experience the sight of an apple because something carrying the form of an apple (or something very much like it) is interacting with us. The point here is that perception emerges because we possess a determinant actuality (a nature) that interacts with other determinant natures. No perception occurs in a vacuum. Placing a healthy, experiencing human body in an absolute vacuum would result in death; consciousness requires a quite narrow range of environments. It is the entire system (object, environment, and person) that is required for sensation. Error can occur within our thoughts, but it will always be an error with, in principle, distinct causal origins.

The second key realist assertion is that the intellect is able to abstract this form. This “abstraction” is not merely a sort of pattern recognition. In abstraction, the form of what is known becomes present in the intellect; the intellect is in a sense identical with the thing known. This explains the act of understanding. This does not imply that we come to know everything about the actuality of the form. Indeed, we will never know everything about any sort of thing. As Aquinas’ famously put it: “all the efforts of the human intellect cannot exhaust the essence of a single fly.” Nonetheless we know what a fly is. We understand it. It is this phenomenological experience of understanding that is the key datum which epistemology is supposed to explain. Empiricism, by rendering this experience off-limits, is essentially sawing off the branch on which any epistemology must hang.

Form and its transmission explain two crucial things. First, it explains why we perceive anything at all, why we experience one thing and not another, and how we come to possess understanding. Second, it explains why we experience things that are “outside” of ourselves. Much more can be said here. In depth accounts have been given of how these principles might work in terms of the mechanics of perception. However, this basic understanding is all that we need to critically weaken many famous arguments from underdetermination. It is not the case that anything at all (or nothing at all) can be responsible for our understanding. Understanding must flow from causes, from a prior actuality (form). If the cause of my understanding of apples is not apples, it must still be something the contains the form of an apple. Likewise, an understanding of number (dimensive quantity)—of magnitude and multitude—presupposes some understanding of a measure, which must be defined in terms of distinguishable qualites (i.e., virtual quantity). To know “three ducks” or “half a duck” requires the measure “duck,” and so it is for all quantities, although we can abstract quantity from substantial form.14 Our understanding of number then, must come from a prior actuality that is abstracted from the senses.

To be sure, we might always be mistaken about the things which we only partially understand, or how and why different things interact. Nonetheless, there must be actualities that are responsible for our understanding and we seem to have absolutely no good reason to think these actualities are anything other than trees, ants, flowers, stars, etc. (i.e., all the things known by the senses and understood by the intellect). When we learn something new about these entities, e.g., when combustion theory replaces phlogiston theory as an explanation of fire, we are still involved in understanding the same entities and natures. What is important to keep in focus here is that theories, models, etc. (as well as ideas and words) are how we know, not primarily what we know. What we know, the form, remains the same.

Arguments to the effect that we can never know from our own experiences alone which sorts of arithmetic operations we are preforming rely on denying the interior act of understanding as a valid datum in epistemology. Yet, prima facie, epistemology, the study of knowledge, is precisely the field that is supposed to explain this phenomenon. That is, it is supposed to explain how we know what we know, and secondarily, the possibility and causes of error. Excluding all noetic experience makes the main goal impossible by default, leaving only the secondary goal, elucidating error. In such a context, skepticism is inevitable because the only valid topic is error. Of course, another irony here is that “external justifications” themselves always rely on experiences that are first-person, and which require an act of understanding to be intelligible. When the intellect is disrupted, as during a stroke, sense data is still received, and yet without the action of the understanding there is no knowledge (as described by victims, who can no longer recognize everyday objects or words).15

This helps explain a huge disconnect in modern epistemology. Through advances in quantitative methods, computing methods, and instrumentation, we can now make predictions far better than ever before. We are better able to collect data, and to use this data to forecast the future, or to predict the outcome of a particular action. We can explain prediction, but not understanding. Indeed, we seemed pressed ever closer towards either a sort of anti-realist pragmatism bordering on epistemic nihilism or an ontic structural realism that claims that the quantities of our predictive models just are all that exists. Yet hasn’t this outcome been inevitable? We have removed the act of understanding from the datum of epistemology. Next, we removed all the metaphysical infrastructure that explained how such an act was possible, on the grounds that it was extraneous to explaining prediction and pattern recognition. The result is that the underdetermination of sheer prediction becomes unanswerable, and skepticism reigns.*

On a closing note, I would just suggest the possibility that, on further investigation, give the strict epistemic presuppositions often in play, it could be revealed that many arguments for underdetermination are in fact themselves underdetermined. That is, their truth might be underdetermined by the evidence we can muster. In particular, this is likely to affect arguments from phenomenological underdetermination (the idea that false beliefs can "feel like" true ones).

---

1. This is akin to the idea that, for any finite set of points on a graph, there will always be an infinite number of functions that will pass through every point.

2. Bertrand Russell. A History of Western Philosophy. (1946) pg. 699. Russell, of course, also extended Hume’s thought on causation to propose that causation itself should be wholly eliminated from scientific discourse

3. E.g., the work of David Bloor and others working from the “sociology of scientific knowledge” paradigm.

4. Recall Russell’s statement about Hume’s conclusions collapsing the distinction between sanity and insanity.

5. The ancient Academic skeptics did utilize similar methods, mostly focused on phenomenological underdetermination (i.e., that false beliefs feel like true ones) to argue against our ability to grasp truth.

6. Thus, the emergence of a strong “anti-metaphysical” movement within early-20th century empiricism was a pivot back towards empiricism’s roots (sometimes a conscious one).

7. Indeed, this conflation of “empiricism” with “science” is probably the key reason that the epistemic presuppositions of empiricism go unchallenged regardless of the apparent absurdities they lead to.

8. Saul Kripke. Wittgenstein on Rules and Private Language. (1982) pg. 66

9.E.g., we do not “mean ‘tree’” by “tree” because the form of trees is present in our intellects and signified our utterance.

10. Another intriguing explanatory connection here is the historical linkage between empiricism and the now hegemonic political ideology of liberalism. Liberalism itself often relies explicitly on skepticism about human nature and the human good for its justification.

11. Of course, the intelligibility of Quine’s argument requires that we understand what a rabbit is.

12. See: St. Thomas Aquinas. Summa Theologiae I, Q.84-86, De Veritate Q.2, A.6 & Q.3, A.3 and Aristotle’s De Anima, Books II-III (as well as St. Thomas’ commentary).

13. This can be conceived of in terms of a triadic semiotic relationship. Nathan Lyon’s Signs in the Dust: A Theory of Natural Culture and Cultural Nature (2019) includes a detailed description of the mechanics here, relying on Aquinas’ notion of “intentions in the media.”

14. See: Aristotle. Metaphysics. (Book X, Ch. I)

15. Indeed, in some forms of phenomenology, a preference for immediate experience, and a bias against any sort of conceptual understanding seems to elevate the experience of the stroke victim or infant to a sort of ideal. The sage and scientist are said to be “lost in abstractions,” while the unformed or damaged mind becomes an idealized epistemic goal.