Feyerabend's "Classical Empiricism" (1970) draws on a 17th century Jesuit argument against Protestant fundamentalism.1 The argument is very general, and applies to any simple foundationalist epistemology. Feyerabend uses it against Classical Empiricism -- roughly, the view that what is to be believed is exactly what experience establishes, and no more -- which he identifies as among other things Newton's "dogmatic ideology".
I will first examine the argument, then ask whether it is perhaps too powerful, a sceptical argument which, if cogent, undermines any pretense to rational belief. Feyerabend explains why the argument could have had little force in the targeted Protestant community. But that is significant only if we can think of its epistemic stance as rational. Feyerabend's irony leaves us with doubts, and yet, perhaps, also with some hope of a genuinely promising development.
The Jesuit characterizes his Protestant opponent as what might today be called a fundamentalist (I take no responsibility for this characterization): this position is characterized by the rule of faith: sola scriptura. This rule says that on religious matters, and indeed on any thing on which the Scriptures pronounce at all, Scripture is the one and only source of information. Believe Scripture and Scripture alone -- that is the rule. What if Scripture has something to say about a certain subject, but leaves some questions about it open? In that case the rule is presumably at least negative: to believe only, or at most, propositions in accord with Scripture.
In philosophical terms this is clearly a foundationalist epistemic position. The position identifies a basis or foundation for all rationally permissible opinion or belief and for all knowledge-claims in a certain domain. In epistemology, as in ethics, we should distinguish between positions and policies and ("meta-") views concerning them. "Foundationalism" sometimes denotes just the metaview that foundationalist positions (in the present sense) are the only viable epistemic positions.
The Jesuit argument has three parts. First, it is not self-evident what is and what is not (genuine) scripture. Which texts belong to the Canon? Of those texts that do, are any parts due to errors or additions made in transcription? According to the rule of faith, this question is to be answered in accordance with Scripture. In fact, Scripture appears to say a good deal about it, as when the text records that some speaker is a prophet or says of itself that it was recorded by an eyewitness. But we cannot draw on this without circulariy. As long as the identity of Scripture is in question, we have -- by the rule of faith -- no basis for determining what Scripture says.
Secondly, the meaning of (putative) Scripture is not everywhere apodictically clear, hence requires interpretation. St. Augustine interpreted Genesis allegorically; this may be seen as violating the rule of faith. But the mere fact that competing interpretations are offered creates a question to be answered. Again, the rule of faith appears to be applicable, since (putative) Scripture includes many directions to the reader, by itself classifying some parts as poetry or song, some as parables, some as history. But the very passages which on one reading are directions to read something as history, biology, or astronomy, are on another reading examples of familiar narrative devices common to many forms of fiction and dramatic literature. We can not draw on Scripture to settle its own interpretation.
Thirdly, in the attempt to settle whether some belief is in accordance with Scripture, we need to know how to draw out its implications. Pure logic will not get us anywhere. It is not a matter of logic, that all humans have hearts, kidneys, knees, and elbows. Yet Scripture need not say explicitly that Abraham did, for us to take it that he did. In assimilating ordinary descriptions we draw consequences via general background and default assumptions of our own. However, we cannot interpret the rule of faith as allowing us to add all our own opinions wherever, as far as pure logic is concerned, Scripture is silent. (This is especially evident in the case of miracle stories.) To draw on Scripture we must settle what it implies; but we cannot draw on Scripture to do so.
To summarize: we cannot apply the rule without identifying, interpreting, and extrapolating from Scripture. But each of these admits of alternatives. The rule would ask the impossible, namely that choice between these alternatives should be on the basis of Scripture itself. Thus the formulated epistemic position is untenable. Its problems are in fact logical problems, having very little to do with what Scripture is or is meant to be.
Feyerabend extrapolates the Jesuit argument very quickly to apply to classical empiricism in which -- he claims -- experience plays the role of Scripture. We may doubt the analogy, given that experience and texts are prima facie very different. But before we examine the asserted parallel, let us look at the Jesuit argument itself.
There appear to be two possible responses. The first is to say that the Jesuit's three problems do not arise at all, and need no solution. The second is to say that he is right, but we have an additional source of information. Obviously he himself opts for the latter, with (Church) tradition as the needed arbiter.
How could it be that the three problems do not really arise? One could say: we do not need to try and identify Scripture because we already know what it is; we don't need to interpret it because its meaning is clear; and we have no difficulty deriving its consequences, which must be done by the same rules as for any other text. Feyerabend identifies this as the actual Protestant response, made possible because the rule sola scriptura is announced in a community in which everyone already uses the word "Scripture" to refer to the same text, and which already agrees in its reading of that text. But this means that there is in effect a second authorative source of information within that community, namely what it designates as "our" understanding -- i.e. its tradition. The existence or even possibility of a differently constituted community prevents that tradition from having the force of pure logic. Therefore the first response reduces to the second.
But does the Jesuit argument not work equally well against any and every putative foundation for knowledge? If so, it should apply to the view that correct opinion is opinion based solely on Scripture and Tradition, or on Scripture identified and interpreted according to one particular tradition. Surely the tradition cannot be used to answer questions like: what, precisely, is our tradition, what does it include and exclude, how is it to be understood? Or rather, it can so be used only with the same vicious circularity as was imputed above.
If that objection is cogent, then the Jesuit argument is more powerful than it looks, and destroys even what one might surmise to have been the Jesuit's own conclusion. But the argument still need not lead to scepticism. For its application is still solely to foundationalist positions. The proper conclusion is then: only non- or anti-foundationalist positions can survive this critique.
Now we see where Feyerabend is headed, or think we do. Some how the classical empiricist will be in the depicted Protestant's position with experience in the role of Scripture. The ensuing critique will be survivable only by a non-foundationalist position. But can any position still allow for rational mutual criticism if the basis for such criticism (Scripture, experience) is deprived of its epistemic primacy?
Classical empiricism points to a unique putative source to ground and test all judgment: 'experience is our sole source of information'. A naive appeal to experience assumes that there is never any question about what the deliverances of experience actually are, nor about their meaning or significance. It assumes furthermore that the implications -- namely, which theories are in accord with experience and which in conflict -- is evident and unequivocal. But these assumptions are not tenable, as becomes clear as soon as we turn from the philosophers' idealization to actual practice.
There are three main problems, ignored by such a naive point of view, not coincidentally parallel to the Jesuit's problems for the rule of sola scriptura. The first is the identification of what is experience in the requisite sense. In one sense, any report we can give of what we think has happened to us is a report on our experience. But that is useless without certain distinctions. Suppose I come in from the garden and report seeing a yellow flower. Perhaps what I actually saw was a candy wrapper, or perhaps I saw only grass but had a small hallucination or fainting spell. In all these cases my report was still presumably veridical in the minimal sense that indeed, it seemed to me that I saw a yellow flower. However, as basis for knowledge claims, as evidence, or as test for opinion, that is useless. For such a basis, we would need to identify those experiences which are veridical in a further sense, namely, indicative of what it was that was actually seen, touched, or heard.
Secondly, even after that identification is made or assumed, there remains a dubitable element of interpretation. Two people who have looked into a furnace will report on, respectively, oxidation and phlogiston escape. The terms learned at their mothers' knee are theory-laden, so the report is infested with theory. Thirdly, even if identification and interpretation are held fixed, an almost century long effort to codify evidential relations (so-called "confirmation theory") should have convinced us that "in accord with experience" is not a simple, uncritically usable, notion.
So far so good for Feyerabend. But the disanalogies are not to be ignored. The analogy looks fine if we simply use the noun "experience" -- a noun like "scripture", "source", "text". But this noun is a philosophical prop which we cannot afford to use naively. One major historical (empiricist?) confusion conflates experience in the sense of events which happen to us, and of which we are aware, with the judgments involved in this awareness. For example: I stepped on a garden hose, but I took it to be a snake -- I jumped and screamed, so everyone noticed my mistake and laughed at me. The event that happened to me, and of which I was certainly aware, differs from the content of my response: the judgment that I was stepping on a snake. The two are by no means the same, nor inseparable whether conceptually or really. This is clearest when the judgment is mistaken, but equally correct when it is true. This does not mean that one of these is the experience and the other something else; but we shall be confused if we either conflate the two or ignore one.
With this distinction in mind, how do the Jesuit arguments transpose? I have no problem of identification for the events that happen to me -- at least, not in the sense that I must find some criterion to isolate the events that happen to me among all the events that happen. On the other hand, I have no problem of interpretation for the judgment I make in response. For any such spontaneous judgment, if explicit, is made in my own language. But this judgment does identify -- in the sense of classify -- the event in question. It also involves an element of interpretation, because it is couched in my own language which even I myself, on reflection, recognize to be heavily laden with old beliefs and theories. Worse: the text comes shrouded in uncertainty, ambiguity, and inconsistency.
So what does the rule of sola experientia prescribe? What I have experienced, in the sense of what has really happened to me, is the touchstone for all theory. But theory is, in me, confronted only with the text of my spontaneous judgments, that is, my immediate judgmental response to what befalls me. This text must be divided into dreaming and waking; and it must also be subjected to a critique in which I isolate at least a first layer of interpretation in the text. What shall I take as guide for these tasks?
In practice I will certainly trust and rely on my prior opinion and theoretical commitments. I will police my own data, and not accept any immediate, spontaneous, unreflective responses as ultimate authority. This is the analogue of relying on tradition. But doing so, I can no longer pretend to empiricist foundationalism: sola experientia was not a rule I could follow strictly at all. Foundationalists will now tell you that at this point I will inevitably slip into a debilitating scepticism or relativism.
The reason that the 'tradition' response looks cogent at first sight may indeed be because it can be taken as an attempt at a non-foundationalist epistemology. The idea is this: the rule of faith itself is a (prescriptive) statement, and has no meaning at all if regarded simply as a bit of syntax. Like any statement, it is meaningful only if heard as a statement in our language, with both sense and reference, therefore, fixed for us to some reasonable (or sufficient) extent. This is equally true of the questions that can be raised about it. We cannot think of our understanding or interpretation as a further text belonging to a language with less semantic structure than our own, which we could then believe or disbelieve, doubt or dispute.
In this way the objection is countered: the Jesuit argument applies to any foundation which can be represented as a text in a language in which the sense and reference of that text are not already fully determined. Relying on tradition does not mean believing in an additional text which informs us of the meaning of all other texts, but consists rather in the indispensable reliance on our understanding of our own language. That reliance is a precondition not only of the meaningfulness of the rule of faith, but equally of any discussion thereof. Of course, our tradition, in this sense, has no power of legislation against other identifications and interpretations, offered from outside -- except for us, whose tradition it is and who persevere in our adherence to that tradition!
In some respects this response is clearly rather feeble. As a Protestant reply to the Jesuit, it would seemingly give up on the use of Scripture as an incontrovertible arbiter between them. Maintaining the rule of faith as understood would admittedly amount precisely to self-assertion, a matter of will, namely determined adherence to one's own tradition. As Catholic alternative to Protestantism it would amount, after explicit addition of the texts codifying their tradition, to an exactly similar stance -- namely, determined maintenance of their own form of understanding of Scripture.
There is a possibility for both to make their differences explicit, and attempt to engage in a dialogue "on common basis", that is, with those differences suspended. How optimistic can we be about such a dialogue? The usual dilemmas of (even moderate) relativism appear: how to respect the coexistence and right to life of alternative beliefs and attitudes with out giving up one's own?
Yet it is clear that we have not necessarily arrived at a debilitating or self-destructive form of relativism or scepticism here. Rather, from both an intellectual and (in a broad sense) political point of view, this is our actual situation. The philosophical problem is not to find the sort of foundation that would rescue us from this aspect of the human condition. It is rather to account clearly for how we can live and function epistemically perfectly well (as we sometimes do!) under these conditions.
To summarize then: the Jesuit argument does not lead to scepticism but only to a rejection of any position which 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 represent able as a text or body of information, so the same questions do not arise. On the other hand, it clearly admits of alternatives, and so this way out does not allow for Scripture -- or any other source -- the sort of role which foundationalists wanted a foundation to play.
Let us put the issue of Foundationalism behind us once and for all. We can't be foundationalists, least of all in epistemology. We have to accept that, like Neurath's mariner at sea, we are historically situated, relying on our pre-understanding, our own language, and our prior opinion as they are now and go on from there. Rationality will consist not in having a specially good starting point, but in how well we criticize, amend, and update our given condition.
Fine, but now the Protestant and the empiricist still come along with their rules of faith. Could we see some meaning in them that makes them relevant in our post-foundationalist condition?
Feyerabend correctly points out that the sola scriptura rule plays two very effective roles in the community when it holds sway. First of all, because there is a tradition which identifies and interprets Scripture, the rule tells everyone who lives in that tradition not to depart from it. So it reinforces and maintains orthodoxy. He correctly likens it to Newton's famous Fourth Rule of Reasoning in Philosophy -- let me state it here in today's idiom:
Is this a rationally compelling rule? Surely not. It is a counsel of epistemic conservativism, for which we may see some good practical reason. But prudence comes in degrees, we all have our personal risk quotient, and it is hard to see how logic could bend us one way or an other.
But besides this obvious first role the rule has a second equally effective role in the critique of accepted opinion. Should the need to revise our understanding arrive, the means are at hand. For suppose we want to change our belief from A to B. There is exactly one acceptable form of argument. We admit first of all that A was indeed solidly based on Scripture (or on experimental results, as the case may be). But then we point out interpretative elements in A, extrapolation or generalization not logically implied by that source. Suspending that part of A we are left with a weaker proposition A*, compatible with B. Now we must, as third step, point to part of Scripture (or experimental results) which together with A* provide support for B.
This is clearly a format that relies heavily on a shared pre-understanding, while making it possible to effect an agreed change in that understanding. Note the illusion traded on, namely that the correct opinion is uniquely determined by the source. Given this illusion we can point to the interpretative elements in A as a mistake, an 'un-Scriptural' or 'unscientific' addition. This is how popular science and school texts now treat Newton: he did not arrive at Einstein's relativity because, enslaved by old ideas and seeing his own results with myopic eyes, he extrapolated his facts in a biased way. Einstein pointed out how Newton had gone beyond the deliverance of experience, removed Newton's metaphysical additions, and thus made way for the right theory, truly true to experience.
The form of argument I described was for a way of 'talking oneself into' a new view at odds with the old. Disturbingly we see here (and Feyerabend trades on this) the possibility of justifying any such change in view in this way in retrospect. Earlier experience cannot logically imply what later experience will be like. Therefore if some opinion based on earlier experience does not fit the new, it must have gone logically beyond what was given -- of course!! By its failure it stands convicted of overinterpretation, for experience (or Scripture!) itself could not have been the source of error! The pattern Feyerabend elicits here is familiar and convincing. But that the argument form has this use for retrospective rationalization does not mean that it does not also have actual use as a format in which the community can co-operatively talk itself into a new view.
So in its second role, the rule prescribes the acceptable form of critique. Can that role be justified? Not a priori. It expresses an attitude, approach, or stance taken by this particular community. In addition, it too admits of degrees as to how much it takes for a critique to become sufficiently weighty to carry the day.
In its two main roles, the rule of sola scriptura or sola experientia is apparently at war with itself. In one role it maintains orthodoxy and forbids heeding the alternative interpretations ingenious minds can concoct. But in its other role, it devalues any aspect of orthodoxy that can be identified as interpretative -- and thus licenses a procedure which can successively peel off layer after layer. In the hands of logicians run wild, it would destroy the shared understanding altogether. But in a stable intellectual community, logicians are not allowed to run wild.
We are here face to face with an aspect of the social dimension of reason that philosophers are not happy to grant significance. The communal opinion remains stable because it is only philosophers who take ideas to their logical extreme, and happily no one listens to philosophers. In the hands of "reasonable" people, this dual-role rule is a great boon. It will maintain the status quo as long as there are no serious anomalies or new situations which throw the tradition into crisis. When such crises do happen, that same rule (in its other role) shows the way to reasoned and proportionate change, providing a rational form for consensual revision. The philosopher, at least if of rationalist or analytic bent, will be taken aback to see that reason, within a community, becomes a matter of politics, broadly speaking, and not of logic. Reasonable is not an epistemic but a bourgeois virtue.
There is disturbing insinuation of double-think, bad faith, and self-deception within Feyerabend's praise for this Janus-faced rule of faith. The mariners-at-sea image depicts our real situation, but can it sustain itself only through the illusion of foundations? Is authenticity not possible? Are we dealing here merely with the last defence of a waning orthodoxy? Or can we reframe this practice so that it becomes at once a reasonable reliance on what we trust ourselves to have gained so far and also allowing of a sufficiently detached critical stance toward our past?
Feyerabend discusses this here only briefly, cryptically, ironically. If the rule of faith is actually empty but available for polemical use, we gain a terrible new freedom. Social factors, party lines, will drive debate when rationally compelling grounds are missing. But:
This is still Feyerabend at least half way into ironic mode, but there is a serious conviction at the core of his irony. In his own work, going forward from this critique of bygone epistemologies, we see him struggling with this task. The task not yet understood in the seventeenth century is as yet unfinished in the twentieth. Feyerabend put his shoulder to the wheel with the rest of us, even if he liked to chastise us scathingly from time to time for our laggard ways.
Feyerabend's argument does refute the position he characterized as classical empiricism. Whether or not that is an actual historical position, the argument is important because it affects a large range of conceivable positions. As we have seen, there are ways in which its targets might escape its force or amend themselves so as to become tenable. Since Feyerabend often enough called himself an empiricist, he might have been in sympathy with this attempt to see what empiricism could be come, in view of his demonstration of what it could not (tenably) be.
I am willing to follow Feyerabend some ways in this view of our epistemic life. In the normal course of things we merely update our opinion logically (non-ampliatively) in response to experience. A crucial role in this process of spinning out the implications of our prior opinion is that of 'expert sources'. These are the explicit occupants of the role of 'tradition', but they are expert for us in virtue of our so regarding them. (This attitude toward science, tradition, newspapers, statistics gathered for certain reference classes, etc. is also a type of epistemic attitude, along with the 'lower level' ones of simply factual opinions and beliefs.) But we enjoy a certain freedom which allows us to take up, at any mo ment, a more detached critical attitude toward our epistemic life so far -- and, given a certain accumulation of disappointments, to proclaim it a failure in certain respects, ready for pruning, slashing, or burning. There is however no rule or recipe, and there can be no such rule, to dictate how we must do this. We can demarcate certain forms of prudence, and insist on coherence, but in the end we face, in such 'revolutionary' moments, radical and unmediated responsibility for our own epistemic future. Feyerabend has pointed us here to notions of rules 'governing' (in some weaker sense than univocal prescriptive dictates) the negotiations and practices with which -- perhaps? -- we steer ourselves through such critical episodes.
Feyerabend, P. K. (1970), "Classical empiricism", in R. E. Butts and J. W. Davis, The Methodological Heritage of Newton. Oxford: Blackwell, 150-166. Reprinted in P. K. Feyerabend (1981), Problems of Empiricism: Philosophical Papers, vol. 2. Cambridge: Cambridge University Press, 34-51.
Popkin, R. H. (1979), The History of Scepticism from Erasmus to Spinoza. Berkeley: University of California Press.
0 Send reprint requests to the author, Department of Philosophy, Princeton University, Princeton N.J. 08544 USA. E-mail fraassen@princeton.edu. The author wishes to thank Elisabeth Lloyd, Philip Kitcher, and Alvin Plantinga for helpful discussions.
1 Feyerabend (1970). Feyerabend draws on the account of Popkin (1979, 70-82). I shall discuss Feyerabend's construal of the Jesuit's (François Veron's) argument, with no presumption that it is historically accurate, nor even that it matches Popkin's understanding of the same material. My interest is in how Feyerabend's argument bears on the possibility of empiricist positions in epistemology.
URL http://www.uwichill.edu.bb/bnccde/_e&ae/feyerab.html
HTML last revised October 16th, 2000.