Feyerabend on empiricism and sola scriptura

In his essay “Classical Empiricism,” available in Problems of Empiricism: Philosophical Papers, Volume 2, philosopher of science Paul Feyerabend compares the empiricism of the early moderns to the Protestant doctrine of sola scriptura.  He suggests that there are important parallels between them; in particular, he finds them both incoherent, and for the same reasons.  (No, Feyerabend is not doing Catholic apologetics.  He’s critiquing empiricism.)

To understand Feyerabend’s comparison, we need to be clear on what “empiricism” is.  (Here and when commenting on sola scriptura I’ll be going a bit beyond what Feyerabend himself says, since some of his remarks are sketchy and merely suggestive.)  In the generic sense, empiricism is of course the view that all knowledge derives from experience.  But there are different ways to interpret that thesis, and the empiricism of Aristotle and Aquinas is by no means the same as that of Locke and Hume.  For the Aristotelian, Feyerabend says, “experience [is] the sum total of what is observed under normal circumstances (bright daylight; senses in good order; undisturbed and alert observer) and what is then described in some ordinary idiom that is understood by all” (p. 35).  It also involves interpreting what is currently perceived in light of “tradition” or “preconceived opinion” (p. 37).  Hence ordinary, everyday statements like “The gunman was wearing a ski mask” or “This apple is stale” -- which presuppose that we already know, from past experience, what a gunman typically looks like, what stale apples taste like, etc. -- would for the Aristotelian provide examples of the sorts of things we know immediately via experience.

But they are decidedly not the sorts of thing empiricism as it developed from Locke to the logical positivists regards as immediately knowable via experience.  Developing as it did in the shadow of Cartesian skepticism, modern empiricism holds that since you might be dreaming or hallucinating the gunman or the apple, what is immediately knowable from experience must instead be something that would remain true even if you were dreaming or hallucinating.  A first suggestion might be that what you know is that “It seems to me that there is a gunman wearing a ski mask” or “It seems to me that I am eating a stale apple.”  But this will not do, because even these statements presuppose all sorts of things which might be doubted. 

For example, they presuppose memory of recent events in light of which what you are experiencing now is best described in terms of a gunman or an apple.  But maybe where you now think you see a gunman, you thought, a few moments ago, that you were looking at your friend playing the part of a gunman in a play you are watching, and you have now forgotten about this context (under the influence of a Cartesian demon, say).  Or maybe a moment ago it was a circus clown that you thought was standing where the gunman now seems to be, and you have forgotten about that context (because of the LSD that someone put in your drink and that has just kicked in).  So why say “It seems to me that there is a gunman wearing a ski mask,” as opposed to something like “It seems to me there is a person (who may be a gunman, or my friend playing the role of a gunman, or a clown who for some reason suddenly looks like a gunman) wearing a ski mask”?  Indeed, why speak in terms of a person, since maybe instead it was a shoe or a ham sandwich you thought you saw there a moment ago (and then suddenly forgot about it under the influence of LSD, or a Cartesian demon, or whatever)?

So, the modern empiricist analysis of experience proceeds by abstracting out more and more of what common sense and Aristotelian empiricism alike regard as “experience.”  On this view, it isn’t statements like “This apple is stale” that we know immediately from experience, but rather something like “There is currently a reddish patch in the center of my field of vision” or even “I am being appeared to redly,” or some other bizarre sort of proposition, that we know immediately.  And to describe what it is that we know from these basic propositions, we cannot use our ordinary concepts but need to develop a new technical vocabulary and talk of “sense data,” “protocol sentences,” and the like.  Everyday statements like “This apple is stale” have to be somehow derived from or reconstructed out of these purportedly more basic statements -- as do all the propositions of science and whatever else we can truly be said to know.

Notoriously, attempts to reconstruct everyday knowledge and scientific knowledge from such purportedly more basic statements all fail.  Not only could modern empiricists not derive everyday and scientific statements from the purportedly more basic ones, they couldn’t agree on what the basic ones were supposed to be.  For the Aristotelian -- and for other critics of modern empiricism like the later Wittgenstein -- this is exactly what we should expect, for the whole project is incoherent.  Statements like “There is currently a reddish patch in the center of my field of vision” are not more basic than statements like “This apple is stale,” but less basic.  The notion of a reddish patch in the center of one’s field of vision (to stick with that example) is parasitic on the notion of everyday experience of objects like apples, an abstraction from such ordinary experiences.  We talk of reddish patches and the like precisely to describe experiences that are abnormal, cases where the ordinary course of experience has in some way broken down.

In effect, the modern empiricist takes the most aberrant possible cases of “experience,” tries to find out what they have in common with all other cases, and makes of that lowest common denominator the baseline from which to reconstruct all experience.  It’s like a psychologist taking the thought processes of the most insane person he can find, teasing out whatever it is those thought processes might have in common with those of all other people, and then attempting to reconstruct a notion of “rationality” in terms of that.  The whole procedure is perverse, a matter of letting the tail -- indeed, a diseased, gangrenous tail -- wag the dog.  The correct procedure in the case of rationality is, of course, to start with paradigmatically rational thought processes and evaluate the various kinds of irrationality in light of those.  And the correct procedure where experience is concerned is to take the ordinary cases as paradigmatic and evaluate the aberrant cases in terms of those, rather than the other way around. 

Now, just as you are never going to derive everything that is constitutive of rationality merely from an analysis of the thought processes of which the most insane person is capable, neither are you ever going to derive everything that is constitutive of ordinary experience merely from the desiccated ingredients -- color patches in fields of vision, etc. -- to which the modern empiricist tends to confine himself.  There is simply far more to ordinary experience than that, and if you refuse to allow in anything but what can be constructed from the desiccated bits, you will inevitably undermine the very notion of empirical knowledge and end up in total skepticism (as Hume does).  And if you don’t end up in total skepticism, it is because you will surreptitiously be smuggling in elements to which you are not entitled given a modern empiricist conception of “experience.”

Thus, though hardly a philosophical traditionalist, Feyerabend judges that:

Aristotelian empiricism, as a matter of fact, is the only empiricism that is both clear -- one knows what kind of thing experience is supposed to be -- and rational -- one can give reasons why experience is stable and why it serves so well as a foundation of knowledge.

For example, one can say that experience is stable because human nature (under normal conditions) is stable. Even a slave perceives the world as his master does.  Or one can say that experience is trustworthy because normal man (man without instruments to becloud his senses and special doctrines to becloud his mind) and the universe are adapted to each other; they are in harmony.

This rational context which enables us to understand the Aristotelian doctrine and which also provides a starting point of discussion is eliminated by the ‘enlightenment’ of the sixteenth and seventeenth centuries.

…It is characteristic of this enlightenment that it constantly mentions new and undiluted foundations of knowledge… while at the same time making it impossible ever to identify these foundations and to build on them. (p. 35)

Feyerabend’s own main interest is not in what modern philosophers made of empiricism, though, but what early modern scientists like Newton made of it.  And he argues that, like any modern empiricism that does not dissolve into skepticism, the empiricist scruples of scientists like Newton were applied selectively and inconsistently. 

But what does this have to do with sola scriptura?  The idea is this.  Summarizing an early Jesuit critique of the Protestant doctrine, Feyerabend notes that (a) scripture alone can never tell you what counts as scripture, (b) scripture alone cannot tell you how to interpret scripture, and (c) scripture alone cannot give us a procedure for deriving consequences from scripture, applying it to new circumstances, and the like.  Let’s elaborate on each and note the parallels with modern empiricism.

First, there is no passage in any book regarded as scriptural that tells you: “Here is a list of the books which constitute scripture.”  And even if there were, how would we know that that passage is really part of scripture?  For the Catholic, the problem doesn’t arise, because scripture is not the only authoritative source of revealed theological knowledge in the first place.  It is rather part of a larger body of authoritative doctrine, which includes tradition and, ultimately, the decrees of an institutional, magisterial Church. 

This larger context -- tradition and Magisterium -- is analogous to the larger context within which both common sense and Aristotelianism understand “experience.”  Experience, for common sense and for the Aristotelian, includes not just sense data -- color patches, tactile impressions, etc. -- but also the rich conceptual content in terms of which we ordinarily describe experience, the immediate memories that provide context for present experience, and so forth.  Just as modern empiricism abstracts all this away and leaves us with desiccated sense contents as what is purportedly just “given,” so too does sola scriptura abstract away tradition and Magisterium and present (what it claims to be) scripture as if it were just given.  And just as the resulting experiential “given” is too thin to tell us anything -- including what counts as “given” -- so too is scripture divorced from its larger context unable to tell us even what counts as scripture.  The modern empiricist inevitably, and inconsistently, surreptitiously appeals to something beyond (what he claims to be) experience in order to tell us what counts as “experience.”  And the sola scriptura advocate inevitably, and inconsistently, surreptitiously appeals to something beyond scripture in order to tell us what scripture is.

Second, even if what counts as scripture could be settled, there is still the question of how to interpret it.  Nor is it any good to claim that scripture itself interprets scripture.  If you say that scriptural passage A is to be interpreted in light of scriptural passage B, then how do you know you’ve gotten B itself right?  And why not say instead that B should be interpreted in light of A?  Inevitably you’re going to have to go beyond scripture in order to settle such questions.  Similarly, even if the modern empiricist can settle the question of which contents count as “experience” -- again, color patches, tactile impressions, or whatever -- there is still the question of what significance to attach to these contents.  Should we interpret them as properties of externally existing physical objects?  Should we interpret them instead in a phenomenalist way?  Is there some “natural” set of relations they bear to one another, or are all the ways we might relate them sheer constructs of the human mind?  However we answer such questions, we will be going beyond anything “experience” itself, as the modern empiricist construes it, could tell us.

Third, even if you can settle the questions of what counts as scripture and of what each scriptural passage means, scripture itself cannot tell you how to infer anything from scripture.  For example, when applying scriptural principles to scientific issues and practical problems, which background empirical, historical, and philosophical assumptions about the world should we employ?   In drawing inferences, should we use a traditional Aristotelian system of logic, or a modern Fregean one?  Which system of modal logic should we use?  What should we think about quantum logic, free logic and other such exotica?  Scripture itself obviously offers no answers to such questions.  Again, in drawing inferences from scripture we will be going beyond anything scripture itself says.  Similarly, “experience” as the modern empiricist construes it tells us nothing about how we are to infer anything from experience, so that in doing so we will thereby be going beyond experience.

Hence, just as Feyerabend thinks Aristotelian empiricism superior to the modern form, so too, on the question of how to understand scripture, he remarks: “We see how much more reasonable and human the Roman position has been” (p. 37).  But as I have said, he is not doing Catholic apologetics, but philosophy of science.  His point is that since sola scriptura is problematic, so is the classical empiricism in terms of which modern science was for so long interpreted.  Clearly, though, the sword cuts both ways.  If the parallels are as Feyerabend sees them, someone who already thinks sola scriptura problematic but is sympathetic to modern empiricism should re-think the latter.  (Cheekily, Feyerabend characterizes Baconian empiricism as “the second great fundamentalist doctrine of the seventeenth century” (p. 37))  But someone who is sympathetic to sola scriptura but already thinks that modern empiricism is problematic should re-think the former. 

Why, if these views are so clearly self-undermining, do their partisans not see this?  In answering this question, Feyerabend devotes much of his article to a discussion of the details of the history of the debate over Newton’s theory of color.  His aim is to provide an illustration of how the purported “success” of the empiricist interpretation of science -- which might seem to confirm that interpretation, despite its conceptual problems -- involves selective and inconsistent application of empiricist scruples, question-begging assumptions, ad hoc hypotheses, and so forth.  And once again he sees parallels with sola scriptura.  In both instances, Feyerabend thinks, partisans of the doctrines in question claim “success” by focusing their attention on cases they think confirm the “rule of faith” while dismissing problematic cases as relatively insignificant puzzles raised by heretics and other oddballs.  Though question-begging, this procedure seems reasonable to them because they are surrounded by a “community… which is already committed to a certain doctrine” (p. 38) and which thereby reinforces their perception that the doctrine is the one that is accepted by all reasonable people.  These communities inculcate a “party line” (p. 39) which determines how one perceives the weight of various objections, the significance of the relevant pieces of evidence, etc.  Hence the doctrines in question -- classical empiricism and sola scriptura -- “although logically vacuous, [are] by no means psychologically vacuous” (p. 38). 

(I’ve noted before -- for example, during the debate over Nagel’s Mind and Cosmos -- how contemporary appeals to the “success” of science as an argument for naturalism or scientism are similarly question-begging, but also have similarly powerful psychological support via the kind of groupthink Feyerabend is criticizing.)

Now, a critic might ask: Wouldn’t the Jesuit critique of sola scriptura apply to the Catholic position as well?  And wouldn’t Feyerabend’s proposed application of it to classical empiricism apply also to the Aristotelian conception of experience?  Indeed, wouldn’t this style of criticism undermine any proposed epistemological criteria, leading to a radical skepticism?  No, no, and no.  

Note first that there is no “sola” prefixed to the Catholic and Aristotelian positions, nor to many other possible epistemological positions.  Sola scriptura and early modern empiricism were both self-consciously revolutionary doctrines, intended decisively to rein in what their proponents thought to be epistemological excesses.  Hence they were formulated precisely so as to lay down an unambiguous line the crossing of which is strictly forbidden, thereby to take down in one fell swoop enormous bodies of doctrine (Catholic theology in the one case, Scholastic and rationalist metaphysics in the other).  They were, you might say, “weaponized” theses from the start.  That isn’t what is going on with positions like the Catholic one and the Aristotelian one.  To be sure, both clearly rule many things out, but neither was formulated with such polemical intent, and thus neither takes the form of a crisp and simple thesis that might lend itself to a charge of self-refutation -- of a weapon which might be wrestled from the wielder’s hand and immediately aimed back at him.  They aren’t trying to boil everything down to some tidy epistemological thesis which might be deployed as a cudgel against opponents, but rather trying precisely to capture the complexity of our epistemological situation, including the complexity inherent in appeals to revelation or experience.  Thus, if someone is going to accuse either position of somehow undermining itself, it will take considerable work to show exactly how it does so.

For another thing, there is a crucial feature of the sola scriptura and early modern empiricist positions that makes them open to the Jesuit/Feyerabend attack, but which the Catholic and Aristotelian positions lack -- namely, commitment to a “myth of the given,” as it has come to be called in discussions of empiricism.  In the case of early modern empiricism, the myth in question is the supposition that there is some basic level of sensory experiences whose significance is somehow built-in and graspable apart from any wider conceptual and epistemological context (as opposed to being intelligible only in light of a body of theory, or a tradition, or the practices of a linguistic community, or what have you).  Aristotelian epistemology not only does not commit itself to such a “given,” it denies that there is one.  In the case of sola scriptura, the myth is the supposition that there is a text whose exact contents and meaning are somehow evident from the text itself and thus knowable apart from any wider conceptual and epistemological context (as opposed to being intelligible only in light of a larger tradition of which the text is itself a part, or an authoritative interpreter, or what have you).  The Catholic position not only does not commit itself to such a scriptural “given,” it denies that there is one.

Now, the reason sola scriptura and early modern empiricism get themselves into trouble is that they purportedly limit themselves to the deliverances of a “given,” but where the existence of the purported “given” in question and the imperative to limit ourselves to it are not themselves knowable from the “given.”  This entails a kind of self-refutation to which doctrines that do not posit such a “given” in the first place are not subject.

Bas van Fraassen, commenting on Feyerabend in his article “Sola Experientia? Feyerabend’s Refutation of Classical Empiricism” (available in John Preston, Gonzalo Munévar, and David Lamb, eds., The Worst Enemy of Science? Essays in Memory of Paul Feyerabend), writes:

[T]he Jesuit argument does not lead to skepticism but only to a rejection of any position that posits a foundation representable as a text.  For we cannot draw on a text in any way without relying on something else, if only on our own language.  This is true equally whether we regard the text as being in our own language or as translated into our language.  But what we rely on is not itself representable as a text or body of information, so the same questions do not arise. (p. 33, emphasis added)

If either the Catholic position or the Aristotelian one “posit[ed] a foundation representable as a text,” then they would be open to the Jesuit/Feyerabend objection.  But that is precisely what they do not do.  The Aristotelian epistemological view does not conceive of “experience” in terms of a sensory “given.”  And the Catholic position does not merely posit a larger text or set of texts (one that would add the deuterocanonicals, statements found in the Church Fathers, decrees of various councils, etc.).  The trouble with texts is that you can never ask them what exactly they include, or what they mean, or how they are to be applied.  But you can ask such questions of an authoritative interpreter who stands outside the texts.  And such an interpreter -- in the form of an institutional Church -- is exactly what the Catholic position posits.

Anyway, I imagine Feyerabend might have sympathized with Ralph McInerny’s quip that “modern philosophy is the Reformation carried on by other means.”