Revelation and Epistemology

George I. Mavrodes

I first wish to express my thanks to Professor Clark and to Professor Nash for their invitation to me to add these comments on Gordon H. Clark’s essay. They represent a development and expansion of comments that I made extemporaneously on the occasion of Dr. Clark’s delivery of an earlier version of these lectures at Wheaton College in November 1965. My comments contain a large element of criticism of Clark’s views and arguments. This should not, however, be allowed to obscure the fact that in this essay he penetrates, and carries us with him, into some of the most important territory that can be explored by a Christian philosopher. I have found the consideration of this essay one of the most stimulating excursions I have recently made into the borderland between philosophy and Christian theology, and I hope that the joint consideration of these arguments and my own will lead to further explorations, and to a better understanding of the role and content of Christian philosophy.

My essay will be divided into four sections. In the first two I will discuss directly Clark’s attempt at axiomatization. In the third section I discuss the relation of sense experience to Biblical revelation, and in the fourth the possibility of extra-Biblical revelation. I believe that these latter topics, while perhaps broader in scope, are closely related to the type of project with which Clark is dealing.

I. What Can Be Deduced From The Axiom?

In sections 2 and 3 of his essay Clark sets forth what I take to be his principal theses. I will restate them here as the following five propositions:

(A) There is an axiom, called “The Axiom of Revelation,” from which, by itself, there can be validly deduced an important set of theorems.
(B) These theorems belong to a number of fields, such as theology, history, logic, ethics, politics, etc., and they constitute the Christian doctrine on these subjects.
(C) This axiomatic set of propositions is available to human beings as certain knowledge.
(D) This set contains the only certain knowledge, as perhaps opposed to opinion, available to human beings. (This is to be contrasted with the failure of other methods of obtaining knowledge.)
(E) The Axiom of Revelation is “The Bible is the Word of God.”

After the destructive criticism of the first section the reader of Clark’s essay is naturally attracted by theses (A) - (E), the claim that there is an axiom which will do what secular philosophy, science, and. common-sense fail to do. This axiom will, by itself, validly and systematically entail an important set of true theorems, and will thus provide us with certain knowledge in a number of areas, most importantly, of course, in theology. In fact, the power attributed to this axiom is so striking and so extensive that it cannot fail either to arouse us to the highest pitch of philosophic interest or else to make us profoundly skeptical. If theses (A) - (E) are true, then a competent logician who is provided with the axiom (it can easily be written on the back of a calling card) has no need of any other source of knowledge. He needs no laboratory, no encyclopedia, and no experience. Even more striking in this context, he will have absolutely no need of the Bible! For he can deduce every single bit of knowledge that a man can possibly have from the axiom by itself. I think the history of philosophy will provide us with few comparable claims of power for a single, short proposition.

Whether in hope, then, or in skepticism one naturally asks immediately, “What is this powerful axiom, and what are some of the more interesting deductions?” It is significant, however, that the Axiom of Revelation (I will call it simply “The Axiom”), though named and discussed throughout Section 2, is not stated until the very end of that section. Only there, do we learn thesis (E). Furthermore, not a single example is given anywhere of the deduction of any theorem, either important or trivial, from this Axiom. This is in striking contrast with the usual development of other axiomatic systems. The reader of a textbook in Euclidian geometry, for example, may expect to find most of the axioms stated clearly in the first five pages, and the deduction of theorems begins immediately. He can thus immediately begin to form his own judgment as to whether the claims for validity and deductive fertility are justified. In a short discussion such as this, of course, we should not expect the deduction of a complete volume on theology. But theses (A) - (E) are the very heart of Prof. Clark’s position. It is therefore most striking and significant that they are not supported by even a single example of the many important deductions, which they claim are possible.

In the absence of such an example, the task of discussing these claims is difficult. Immediately after stating the Axiom, however, Prof. Clark provides what might possibly be a schema, or pattern, for making the required deductions from it. He expresses this possible schema in a single sentence: “From the one axiom it follows syllogistically that such and such sentence in Scripture is true because ‘it is the Word of God.” I want to examine the possibility that this schema supplies the deficiency which I noted above, supplies it by telling us how to go about making the required deductions from the Axiom.

The first difficulty that faces us is one of interpretation. Can the schema be interpreted in such a way that (a) it is valid, and (b) it provides a pattern for the deduction of theorems of the sort required by thesis (B)? Let us try.

At the first sight the schema might seem equivalent to the following argument:
P1 The Bible is the Word of God. (axiom)
C1 Therefore, everything in the Bible is true. (theorem)
To put this argument into regular syllogistic form, however, requires the addition to it of another premise.
P2 Everything God says is true.
And this suggests, of course, that the conclusion is not derived from the Axiom alone, as Clark claims.

I do not know what account Clark might give of P2 in this argument. I can think of only one plausible expedient that might save the claim that the Axiom is the only premise of this argument. This would be to say that P2 is some sort of necessary truth, perhaps part of the concept of “God,” and hence need not be counted as an additional premise. This view may have some difficulties, but perhaps they can be overcome. At any rate, we might accept this reply (or some similar one) provisionally, and thus provisionally accept the short version of the argument as valid. An alternative would be for Clark to accept P2 as an additional axiom, and then to claim that his desired system follows from these two axioms. I have no objection to this emendation, and I think that no part of my succeeding comments will be seriously affected if it is made. (In fact, throughout the remainder of my paper, I will assume that the Axiom can be supplemented by P2 whenever desired.) Does this argument, then, assumed to be valid, support theses (A) - (E)?

It should be clear that it does not. C1 is a generalization about Biblical statements. But thesis (B) requires the deduction (from the Axiom) of specific doctrines, e.g., that David was a king of Israel, and not of generalizations over doctrines. Put in other words, C1 implies statements of the form, “If S is a Biblical statement, then S is true.” But this is a hypothetical statement about doctrine S, whereas thesis (B) requires the deduction of categorical judgments about doctrines, judgments of the form, “S is true,” or, more simply, just “S”.

Therefore, if we interpret the schema as being equivalent to either the short or the long form of the preceding argument it fails to provide the desired result. It may be valid (in the long form it is certainly valid), but it does not provide a pattern for the sort of deductions required by thesis (B).

This suggests, of course, that we should try to amend the interpretation. And anyway, there is a hint within the schema itself that it is not equivalent to the preceding argument. For, the schema contains the phrase, “such and such sentence in Scripture,” which presumably is to be instantiated by some particular sentence from the Bible. But the preceding argument contains no expression that has this function.

It is not difficult, however, to see what sort of thing is needed. We require a way of passing from generalizations over doctrines to specific doctrines. Let us therefore supplement the preceding argument with the following:
P3 Everything in the Bible is true. (C1)
CF2 Therefore, S is true.
where S is to be instantiated by particular sentences, e.g., “David was a King of Israel.” This second argument (more properly, an argument form) provides us with conclusions of the proper form to satisfy thesis (B); that is, it yields specific doctrines as conclusions. And this argument form can properly be attached to the preceding argument, since its only premise, P3, is identical with the preceding conclusion, C1. It may seem, therefore, that the combination of this argument form with the preceding argument provides the required interpretation of the schema.

Unfortunately, however, this latter argument form is clearly invalid as it stands. The variable “S” appears only in the conclusion and can therefore be instantiated in any way we please, without reference to the premise. But there are many such instantiations, e.g., “David was a prime minister of Babylon,” which will make the conclusion false even if the premise is true. What is needed for validity, of course, is for “S” to occur in the proper sort of premise as well as in the conclusion, in a premise which will restrict “S” in a suitable way. The following premise is an obvious candidate, and is perhaps suggested in the schema itself:
PF4 S is a statement of the Bible.
Adding PF4, or something similar, will certainly make this form valid. But, of course, PF4 is an additional premise, and this subverts the claim of thesis (A) that the theorems can be deduced from the Axiom alone. And in this case, I fail completely to think of any stratagem at all, which gives the slightest promise of validly deducing any statement of the form PF4 from the Axiom (nor does Clark provide any hint of such a method). So far as I can see the Axiom contains no information whatsoever as to the specific content of the Bible. But PF4 requires a (partial) specification of that content. If this is so, then nothing of the form of PF4 can be deduced from the Axiom.

Furthermore, we cannot simply suggest, as before, that PF4 might be accepted as a second (or third) axiom of the system. For there is no one premise, PF4. For every statement whose truth we wished to deduce in this manner, we would require a specific instantiation of PF4, which stated that that particular statement appeared in the Bible. Therefore, if we wished to deduce all of the information contained in the Bible, the list of supplementary axioms required would be as long as (perhaps longer than) the Bible itself. But, Clark specifically rejects the suggestion that this system requires us to take the whole Bible, or any other large list of propositions, as axioms. He claims that it operates upon a single axiom, which he states.

Frankly, then, I fail to think of any way in which Clark’s apparent schema can be interpreted which both preserves its logical validity and also provides the required sort of conclusion. Nor can I think of any other schema that will serve. In fact, with the exception of the Axiom itself, I cannot think of a single interesting and important proposition, theological or non-theological, true or false, which can be validly deduced from the Axiom. As far as I can see, Clark’s Axiom is a remarkably sterile proposition. (Is it not significant that Prof. Clark himself does not deduce a single theorem from it?) If anyone thinks otherwise, however, I would be glad to see a deduction of the doctrine of the Trinity, or of the Deity of Christ, or even that David was a king of Israel.

At this point, someone may suggest that Clark has simply overstated his case, but that it nevertheless incorporates an important insight into Christian epistemology that can be salvaged. Perhaps he goes too far in claiming that Christian doctrine can be derived from the Axiom alone, but nevertheless we might suppose that the Axiom is an important, perhaps even an indispensable, premise in the deduction of Christian conclusions. Since the Axiom by itself entails nothing of interest we may as well follow up this suggestion, and see if it entails interesting conclusions when it is used along with other premises. And, in fact, we shall find that this suggestion leads to a somewhat surprising result. Let us, therefore, amend Prof. Clark’s thesis to allow the use of an indefinitely large list of supplementary axioms in conjunction with his special Axiom. We can make this alteration by replacing thesis (A) with: (A*) There is an axiom, called “The Axiom of Revelation,” from which, in Conjunction with other premises, there can be deduced, an important set of theorems.

This amended proposal brings us to the threshold of a problem that I wish to discuss in the third section of my paper. Here, however, we may note that it has at least one clear advantage over the original, but that it also gives rise to a curious apparent paradox. First, the advantage. The amended proposal certainly does allow, as the original did not, the valid deduction of many propositions of Christian theology, history, etc. In the case of Clark’s example, “David was a king of Israel,” we can construct the following argument:
P1 The Bible is the Word of God. (Axiom 1)
P2 Everything God says is true. (Axiom 2)
P4 The Bible says that David was a king of Israel. (additional premise)
C2 Therefore, David was a king of Israel.
This argument is valid, makes use of the Axiom (plus P2), and has the required conclusion. This certainly seems to be an improvement.

The additional premise, however, introduces what will seem surprising, at least to many. For P4 is not, so far as I can recall, a proposition which is expressed in the Bible (though it is expressed in many other books), nor is it deducible from any propositions which are expressed in the Bible. And the occurrence of non-Biblical premises of this sort will be, I think, a general feature of valid arguments in which the Axiom figures.

The claim that P4 does not occur in the Bible may at first appear startling to persons acquainted with Old Testament history, but this is because, in certain contexts, P4 may be confused with a related proposition:
P5 David was a king of Israel.
Now, P5 (or some rough equivalent of it) does occur several times in the Bible, but we must not allow this to mislead us into thinking that P4 occurs there. For P4 and P5 are by no means the same proposition, and some of the ways in which they differ are crucial for just the sort of topic we are concerned with here. I will mention three of these differences.

First, either one of these propositions may be believed without believing the other one. A person may believe that David was a king of Israel (P5) without having any idea at all of what is in the Bible, or even while believing that the Bible is concerned entirely with the history of Bolivian tin mining. On the other hand, a person might easily believe P4, a statement about what the Bible says, even if he thinks that every proposition in the Bible is a lie. These two propositions, then, are distinct candidates for belief.

Second, these two propositions are plausibly supported by quite different sorts of reasons. If someone is now asked “How do you know?” after his assertion of P5, then one appropriate reply he might make is “The Bible says so,” or something of the sort. In other words, P4 can plausibly be given as (part of) the reason for believing P5. And one way of developing and formalizing this reason might be to construct the argument we have just been considering. On the other hand, P5 cannot plausibly be given as a reason for believing P4, nor can one plausibly give “The Bible says so” as a reason for believing P4. One of the most common, and indeed one of the best, answers to “How do you know P4?” is “I’ve read it myself.” This type of answer seems to have no obvious relevance to the corresponding question about P5. We might say that these propositions are “epistemically ordered.” As part of our reason for believing P5, we may give P4. As part of our reason for P4, we may refer to some personal experience. But it does not make sense to invert this order. This difference recurs, from another angle, in section III, below.

The third difference to which I wish to call attention is that between the ways in which P4 and P5 function in arguments involving the Axiom. If, in the argument we have been considering, we replace P4 by P5, we get a new argument, which, like the old, is perfectly valid. The new one, however, unlike the old, is circular, since the new premise is identical with the conclusion. Moreover, and more important for our present interest, the Axiom performs no function in the new argument. It occurs there only vacuously, and can be dropped without damage to the validity or soundness of that argument. For the conclusion, being identical with P5, follows directly from P5 without need of the Axiom or of any other premise whatever. P4 and P5, then, function quite differently in arguments involving the Axiom.

We have seen now that we can use the Axiom, along with an additional premise, to construct a valid argument for one of Clark’s desired conclusions. But the additional premise was not a Biblical statement. On the other hand, when we replaced that statement with its Biblical analogue the argument remained valid but the Axiom became totally superfluous. We can also see, I think, why this superfluity of the Axiom must be the almost invariable result of using only Biblical statements as the additional premises. The Axiom ascribes a certain property (that of being the Word of God) to the Bible or to propositions occurring in the Bible. Ordinarily, therefore, it will play an essential role in a deduction only if another premise of the argument assigns to a certain proposition the property of being a part of the Bible. Now, the Bible contains a great many propositions, but only a very few of these are statements to the effect that some other proposition is a part of the Bible. That is, almost all of the propositions of the Bible resemble P5, not P4.

Very few propositions of the Bible, then, are candidates for deductively fruitful combinations with the Axiom. In almost all cases, if a given conclusion can be validly deduced from Biblical statements plus the Axiom that same conclusion can be validly deduced from those Biblical statements without the Axiom. For unless one of the Biblical statements resembles P4 rather than P5, an unlikely occurrence, there will be no “middle term” to relate that statement to the Axiom.

In a few instances the Bible may contain a proposition resembling P4, a proposition that can be combined fruitfully with the Axiom. Even in these few cases, however, the Axiom is not really necessary. For, if the Bible contains something like P4, it must also contain the corresponding P5 (at least it must do so if all of its propositions are true). Thus, even if there are Biblical statements, which can be combined with the Axiom to yield a conclusion, there must also be other Biblical statements, which entail that same conclusion without the Axiom. In no case, then, is the Axiom necessary or useful in the deduction of conclusions from Biblical statements.

We get, then, a curious and surprising result. The Axiom is a claim about the Bible. Roughly, we might suppose that it is intended to attribute to the Bible a high degree of authority, reliability, truthfulness, or something of the sort. By itself, however, the Axiom entails nothing interesting or important, either true or false, in the areas of theology, history, ethics, etc. In combination with other premises it does become deductively fruitful, yielding interesting conclusions. It is useful in this way, however, only if we combine it with the premises that are not found in the Bible. If we were content to use only Biblical propositions as the remaining premises, then we would find the Axiom a perfectly useless addition to our arguments. There may be something very important for Christian epistemology in this apparent paradox. We may see another aspect of it in section III, below.

II. Can We Axiomatize Christian Doctrine?

In section I, I concluded that Clark has failed to suggest a way to axiomatize our theological knowledge (or any other knowledge), and that nothing of importance follows from his Axiom alone. Apart from its bearing on this particular essay, is this failure of any consequence? I think not.

Suppose that we assemble a motley list of propositions, A, B, C…N, as random as we can find. Can we now axiomatize this random list? Certainly. A method of doing it can easily be taught to a child or programmed for a computer. Simply form a list of axioms that repeats the original list of propositions, A, B, C…N. Since every proposition entails itself the original list will follow deductively from this set of axioms. Presto! Axiomatization has been accomplished. But perhaps we are not satisfied. Why not? Is it desired to form a system that has fewer axioms than theorems, indeed only a single axiom? It is easy. Simply construct a single proposition, which is the conjunction of the original set, A · B · C · · · N, and take this as the single axiom. It validly implies each of the originals.

Are more complicated deductions from axiom to theorems desired? They can be supplied. For example, form the propositions A ⊃ B, B ⊃ C, C ⊃ D, … , M ⊃ N, N ⊃ A. Conjoin all these (in any order) along with A v N, and take this conjunction as the axiom. It entails all of the original set.

Is still more complexity desired? It can, of course, be provided. But the method should now be clear. It consists merely of working backwards from the desired theorems and building in whatever logical features are desired along the way. Furthermore, the methods of axiomatization, which I have suggested, and any further complications which arc carefully constructed, have an additional interesting feature. If all of the original propositions are true, then all of the axioms are true also. In that case the deduction of theorems from axioms will be a deduction of truths from truths.

None of these methods of axiomatization, however, depend in any way upon any special features of the propositions involved. Consequently, they can be applied to any set whatever. I suggest, then, that Prof. Clark’s failure to axiomatize Christian doctrine need not trouble us. For there is no doubt that it can be axiomatized, and very easily. In fact, like any other collection of propositions, it can be axiomatized in indefinitely many ways, and we can multiply such systems for as long as we care to work. Furthermore, if Christian doctrine is true, as I believe it is, then there are indefinitely many sets of true axioms that entail it. So we need not fear. Clark’s goal, though not his method, seems assured.

Another way of putting the point is to say that, if we can eliminate thesis (E), and perhaps also the name of the axiom from thesis (A), then theses (A) – (D) are true. Indeed they are trivially true.

But we may now wonder whether the goal itself is of great importance. Clark appears to believe that axiomatization makes a set of propositions into a system and, indeed, into a system of a very important sort. We can see now, however, that if we choose to say that every axiomatized set is a system, then we must use an almost unbelievably weak sense of “system,” a sense so weak that no collection of propositions, no matter how random, can possibly fail to be a system. I cannot imagine why it should be thought important to emphasize that Christian doctrine constitutes a system in this hyper-attenuated sense. The collection consisting of the thirty-second indicative sentence from each book in the Library of Congress constitutes a system in exactly the same sense. Furthermore, the actual construction of the axioms can hardly be a matter of any great importance. It can easily be done by a child who is taught to perform a few operations by rote, with no knowledge either of logic or of the subject matter involved. It can even more easily be done by a machine, which knows nothing at all. Once we understand that it can be done mechanically for every collection of propositions whatsoever, why should we trouble ourselves to do it at all?

Clark also appears to believe that it is very important to reason deductively, to use logic, etc., and to do these things correctly. Deriving theorems from axioms involves just those operations. Hence, presumably, it inherits their importance. But can this alleged importance survive examination?

Suppose that I believe some proposition B. However, instead of putting B forward on its own merits, whatever these may be, I deduce it from A with a flourish of logic. Suppose that my logic is impeccable, and even that I engage in a long discussion and vindication of the logical principles involved in the deduction. Will the fact that I deduce B by correct logic make B any better than it was prior to the deduction? Or will it serve, or even tend, by itself to justify my belief in B? Certainly not. Every proposition, false as well as true, is validly entailed by some other. Consequently, the construction or display of a valid entailment cannot serve, by itself, to distinguish true propositions from false ones. The deduction has no more implication for the truth of B than does asserting B in a Scotch brogue or writing it in green ink.

This brings us, however, to an objection and a further point. It is clear that many mathematicians, logicians, and philosophers have been very interested in axiomatic systems and have used great ingenuity in attempting to construct them. Why should this be so if they are as trivial as I suggest?

I think the answer is that while axiomatization without constraints is trivial, both in interest and importance, special sorts of axiomatization may be very interesting, very important (for certain purposes), and very difficult. In fact, I suspect that Clark himself is interested only in axiomatization of a special sort, but he does not make clear the constraints he may have in mind (perhaps only implicitly) nor the purpose they are intended to serve.

Traditionally, I think two related sorts of constraint upon the method of axiomatization have figured prominently. It was demanded that (a) the set of axioms should be in some sense simpler than the theorems to be derived from them, and (b) that the axioms should have more intuitive appeal, should be more “obviously” true or “self-evident” than the theorems. Both of these constraints are well exhibited by Euclid’s axiomatization of geometry. The axioms have at least the appearance of being few, simple, and intuitively obvious, while the theorems are certainly many, complex, and far from obvious.

The first sort of constraint is motivated partly by esthetic considerations, especially in mathematics where “elegance” is often explicitly used as an evaluation criterion. But it may also have a didactic motivation. A person who knows logic and who can be gotten to accept a small and simple set of powerful axioms can perhaps then easily be led to acknowledge a large and complex set of theorems without the introduction of further considerations. But as the complexity of the axiom set itself grows the comparative didactic advantage decreases. If the axiomatization is to be didactically useful, therefore, the axiom set must be kept simple.

The second sort of constraint is also motivated partly by didactic considerations. If one attempts to persuade an audience by deduction, it is desirable to begin with premises, which will be widely and quickly accepted. Hence, a person who wished to propagate a given body of doctrines might well seek for a set of propositions which entailed those doctrines and which also had a powerful and immediate intuitive appeal.

The second type of constraint, however, also has what may be called an epistemological motivation. Belief of every sort is subject to the erosion of ordinary doubt on the one hand, and, on the other, to that of the special skeptical arguments of the philosophers. In an effort to rescue the cognitive life, either in general or in some special field, the attempt was made to find beliefs that were invulnerable, or at least less vulnerable, to these erosions. The valid entailments of such propositions might also be thought to be immune to doubt in the same way. And so the axioniatization of an important set of doctrines in. terms of some “self-evident” truths could be considered an important epistemological achievement.

These constraints upon the permissible methods of axiomatization are governed by the goals, which were to be attained by means of the axiomatization. I do not intend to discuss these particular goals and constraints in detail here. They are, of course, not the only possible goals of axiomatization. (Clark himself refers to another possible goal, which I shall discuss shortly.) However, it may be interesting to discuss the extent to which Clark’s own attempt at axiomatization satisfies the constraints already mentioned.

Professor Clark briefly considers and rejects one proposed way of axiomatizing Christian doctrine, which clearly comes much closer to success than does his own attempt. The method he rejects would take all the propositions of the Bible as axioms. He apparently bases his rejection on the fact that this axiom is too complex, indeed not noticeably simpler than the theorems to be derived from it. This amounts, of course, to saying that this attempt does not satisfy constraint (a) above. His own attempt, using the single Axiom of Revelation, does seem to satisfy constraint (a). The Axiom appears to be simpler than many specific Christian doctrines, e.g., that of the Trinity, and especially simpler than the whole set of them.

The case is different, however, when we come to constraint (b). Is there any reason to suppose that the Axiom is more obviously true than many other important doctrines? I, at least, find it hard to think that the Axiom is nearly so “obvious,” or so likely to be accepted, as, e.g., that there is a God or that God created the world. I am sure that multitudes of people think it far more likely that David was a king of Israel than that the Axiom is true, and I would not be at all surprised to discover people who find it easier to believe in the Deity of Christ or His death for our salvation than to believe the Axiom, in the sense in. which I suppose that Clark intends it.

Furthermore, the Axiom has not, in fact, proved noticeably resistant to the erosions of either ordinary, or sophisticated, doubt. It is a common tactic of Christianity’s critics to direct some of their first and most effective attacks against the Axiom (or some very similar proposition). And anyone who talks with college students will find many of them saying that their very first doubts about Christianity arose over something like the Axiom. Therefore, even if the Axiom had succeeded logically in providing the desired entailments (as it did not) it appears that it would not have succeeded in. providing either the persuasive or epistemological power, which we might have expected from axiomatization.

I emphasize the word “might” in the preceding sentence because we might expect something quite different from axiomatization. Clark himself suggests something else. He says, “For some restricted purposes, particularly for elementary religious needs, a knowledge that something is so is sufficient. But a greater degree of faith, belief, knowledge, and understanding requires a grasp of why something is so. And since God’s revelation is rather ample, it would seem that he wishes us to have an ample understanding. Axiomatization, deduction, systematization are therefore to be considered desirable.” I think Clark is correct in claiming that knowledge of why a thing is so (“Why,” of course, has several senses) represents an increase of understanding. It is also possible that an axiomatization will provide that information. But if it is to do so, it must be constructed in a special and appropriate way. It does not appear, however, that Clark’s axiomatization is at all on the right track for this purpose.

To make this clear, let us consider again an argument, which results when we add to Clark’s Axiom enough additional premises to provide a valid deduction of one of the required theorems.
P1 The Bible is the word of God. (Axiom 1)
P2 Everything God says is true. (Axiom 2)
P4 The Bible says that David was a king of Israel. (Axiom 3)
C2 Therefore, David was a king of Israel.
This is valid, and it might convince us (if we accept all the premises) that David was a king of Israel. So far as I can see, however, this set of axioms (and, a fortiori, Clark’s Axiom by itself) contains no explanation whatsoever as to why David was a king. By going through this deduction, I learn nothing at all about the why of anything’s being so. I suppose this is because none of the premises make any reference to causes, purposes, or any other factors which are suitable for inclusion in the answer to a “why?” question.

Compare, however, the following axiomatization.
P6 God ordained that David should be a king of Israel. (Axiom 4)
P7 God brings to pass whatever He ordains. (Axiom 5)
C2 Therefore, David was a king of Israel.
This deduction is also valid, but, unlike the former, it does contain one sort of explanation as to why David became a king. If we accept this argument, we will be prepared to say, “David was a king because God ordained it and brought it to pass.” This axiom set, however, neither includes Clark’s Axiom, nor any axiom derivable from it. It is built along different lines, making no reference at all to the Bible. But, whatever may be its other merits or faults, at least it contains implicit references to causes or purposes, and thus contains one possible answer to the question “Why?”

It is possible, then, to construct such an “explanatory” axiomatization. (On the other hand, I think it also possible to provide explanation without axiomatization. But I will not argue this here.) However, if the attempt is to succeed, special constraints must be observed. At the very least, it appears that the axiom set must include references to causes, purposes, or other explanation-type concepts. And although explanation appears to be one of Clark’s goals for axiomatization his own attempt at it does not seem to satisfy this constraint.

Axiomatization, then, can have a point or purpose. Indeed, we can choose among a variety of purposes — esthetic, didactic, epistemological, explanatory, and perhaps others. This, however, suggests that before a serious attempt at axiomatization is undertaken, some preliminary work may be profitable.

(1) We should attempt to formulate as clearly as possible the goal (or goals) which we expect to achieve by means of the axiomatization.
(2) We should investigate whether perhaps some procedure other than axiomatization may not be more suitable for these goals. E.g., Clark mentions systematization as a goal. But “system” is not an end in. itself. There are many kinds of systems, and they are adapted to many purposes. The entries in a dictionary are ordered systematically. The system (by alphabetic arrangement) is essentially arbitrary, and has no relation to logical deduction. But, for the purposes of a dictionary, it is incomparably more valuable than a deductive ordering would be.
(3) We should formulate a set of constraints that will govern the method of axiomatization, and carefully investigate whether these constraints will, in fact, lead to the desired goals.
I suspect that if we make these preliminary investigations, we shall often decide not to continue with the axiomatization. But if we do go ahead, we shall have our goal clearly in mind, and we shall be supplied with a set of guidelines for our procedure.

III. Biblical Revelation and Sense Experience

So far I have not discussed Clark’s criticisms of alternative epistemological positions, and, in fact, I will have little to say about this part of his essay. One line of argument that he uses, however, leads us back to a topic upon which we touched in Part I of my paper. I want now to approach it again from this new angle.

I suspect that one of the important motivations for Prof. Clark’s project is that of discovering a source of knowledge, particularly for theological knowledge, which is more secure and reliable than is any ordinary source. It seems clear that the ways in which we ordinarily acquire the beliefs of everyday life are fallible. They sometimes lead us into error. Is there not a source, at least for the most important subjects, which is better than this? Many theologians have said that there is, and Clark belongs in this tradition.

I have so far expressed this tradition in very general terms. Some of these theologians, however, have been more specific and have in fact held two doctrines, which I believe cannot both be true. They have held, on the one hand, that sense experience is fallible and that any beliefs that rest upon it are tainted with uncertainty. The ground upon which they rest falls short of complete reliability. On the other hand they hold that the Bible is infallible and hence that it provides us with a type of knowledge, which rises above uncertainty and has a ground that is completely firm. Some comments of Clark’s make it appear that he also belongs in this tradition. At any rate, his position has such close affinities with it that I want to discuss it here.

Near the end of section 2, Clark replies to some conservative theologian who thinks that some “common-sense” ordinary knowledge is necessary even for the spiritual life. This theologian says “You cannot obey the seventh commandment if you don’t know who your wife is.” Clark, however, raises the question, “How do you know?” And it is clear that his reply is, “You don’t really know.” The general pattern of his argument, though not explicit, is also perhaps clear enough. He refers to a story about Louis XIV, who was allegedly surreptitiously replaced by his twin, and he reminds us of the arguments of the solipsists and subjective idealists. These he apparently takes to destroy any non-revelational base for this alleged knowledge. He thinks I cannot derive any knowledge about my marital status from the Axiom (he is surely right in this). Perhaps even more important, however, I cannot deduce such knowledge even if I took every proposition in the Bible as an axiom. So my belief about who is my wife has no firm basis whatever, and hence does not constitute knowledge. Clark does, however, allow that we may have opinions on such subjects, and that in some situations such opinions may serve us in place of knowledge.

Clark seems, on the other hand, to claim to know that David was a king of Israel. But how does he know it? He might reply that he deduces it from the Axiom. The deduction of this proposition from the Axiom, however, is certainly invalid. If Clark has no better reason than this for his belief about David, then his belief is in even worse condition than mine. I suspect, however, that he really does have a better reason than this. I think that under most conditions he, and many other theologians, would really make a quite different reply, one that claims that this knowledge is derived from the Bible itself, rather than from an Axiom about the Bible. For example, in his discussion of “Logic and Scripture” in section 2 Clark says, If the problem is to explain how we know in this age, one cannot use the personal experience of Moses. Today we have the Scripture. As the Westminster Confession says, “It pleased the Lord…to reveal himself…and afterwards…to commit the same wholly unto writing, which maketh the Holy Scripture to be most necessary, those former ways of God’s revealing his will unto his people being now ceased.” What God said to Moses is written in the Bible; the words are identical; the revelation is the same.

Now, this suggests to me that Clark’s real reply to our question, “How do you know that David was a king of Israel?” will come down to, “The Bible says so” Very good. But not yet good enough. And it is important that we understand why it is not yet good enough. The mere fact that the Bible says something is not, by itself, sufficient to account for Clark’s knowing that thing. If it were, then it would be certain that Clark knows everything that is said in the Bible, which seems unlikely. In fact, if it were sufficient then it seems (since there is no reason to suppose Clark unique in this respect) that- everyone in the world would know everything that is in the Bible. The Bible’s saying it would make everyone know it. But there are many things said in the Bible that I do not know. Perhaps I never read them, or perhaps I have forgotten them, etc. This suggests, of course, that if “The Bible says so” is to function as a reason for a given person then it must be combined with something else. It must at least be combined with that person’s belief, and his reason for believing, that the Bible does say so.

Now, with regard to David’s kingship, Clark no doubt does believe that the Bible says so. Let us therefore repeat our question, “How do you know?” with respect to this second belief. As I pointed out in section I, the question now requires a different sort of answer. Conceivably he might reply that teachers of theology told him that the Bible says so, in which case we could shift our attention to them. But there seems to be no reason to doubt that Clark could reply more directly, “I have read it myself.” Of course he has, and so have I.

But perhaps this last sentence is too hasty. A philosopher who can envision the possibility that a French king and, indeed, his own wife, might be surreptitiously replaced by imposters surely will not boggle at the possibility that one book might be substituted for another. Perhaps Clark’s Bible has been replaced with a cunningly disguised substitute that contains a few crucial false statements. Perhaps the real Bible does not say that David was a king after all, and Clark has been misled. In fact, perhaps he has been misled by this false Bible on more important matters. (Is this a place where mere opinion could serve him as well as knowledge?)

Of course, he may appeal to his pastor for the confirmation or correction of his views. But, alas, pastors may be replaced as easily as kings or wives. Perhaps the grave and reverend gentleman in the pulpit is not the pastor but an imposter. And anyway, the substitution of the Bible may have taken place a long time ago, centuries or millennia in the past. Perhaps the pastor is deceived too, along with Augustine, Aquinas, Luther, and Calvin. Any alleged Biblical knowledge that Clark may claim, then, regarding the fortunes of David must fail for just the same reason, which he raises against knowledge regarding his wife; i.e., both of them fail because of the possibility of substitution.

On the other hand, suppose that Clark’s Bible is genuine. Can he be sure that he has read it correctly? Probably all of us know how easy it is to overlook the word “not” in a sentence, to our subsequent discomfiture. Perhaps Clark’s Bible says that David was not a king. Of course, he can read it again. Will that give him knowledge? Could he identify his wife with certainty by looking at her again? Apparently not. It is possible to make the same sensory mistake twice. Mere repetition of sense experience cannot confer epistemological certainty. So Clark’s apparent Biblical knowledge must also fail because of the fallibility of the sense. At best, mere opinion remains.

And, of course, as Clark says, there remain the prior difficulties of solipsism and subjective idealism. They apply to books as well as to anything else.

At least one point should now be clear enough. Whatever general difficulties or weaknesses infect beliefs derived from sense experience must also equally infect beliefs derived from the Bible. For sense experience is required for the derivation of such beliefs. Therefore, if Clark is correct, in thinking that he cannot get any knowledge from sense perception, then he cannot get any knowledge from the Bible either. Since we have already seen that he cannot get any from the Axiom, it is not immediately obvious where else he might look. It would be unfortunate, however, to stop here, passing this off as a curious inconsistency generated by Clark’s extreme skepticism regarding the senses. It has a consequence of importance for every attempt at Christian epistemology. The fact is that every consistent epistemology which assigns a role to the Bible in the acquisition of theological knowledge must assign a role of equal scope, and in precisely the same area, to sense perception. For whenever the Bible forms a link in an epistemological chain, then sensory contact with the Bible must form the very next link.

This point is perhaps worth emphasizing. The Bible is not a magic talisman that makes men “wise unto salvation” merely by containing true propositions and standing upon a shelf. It has epistemological results when men read it, study it, hear it preached and expounded. Missionaries understand this well enough. They do not merely carry the Bible in their pockets. They translate it, print it, sell it, give it away, and preach it. In short, they bring the people into contact with the Bible and its message, and they make this contact through sight and hearing, through sensation. A Christian epistemologist (at least one who includes the Bible in his system) has by no means finished before he has incorporated sense experience into his system. And he cannot relegate such experience to an appendix, as the mode whereby we obtain “worldly” knowledge (or opinion). To whatever extent God has chosen to make His revelation available in the Bible, to that same extent He has chosen to funnel exactly that same revelation through the senses. Sensation must therefore be just as central to the acquisition of theological knowledge as is the Bible itself.

It is therefore of no use for the philosopher to tell us that the senses are weak, variable, fallible, etc., unless he also has the courage of his convictions and allows that the knowledge he derives from the Bible is infected with this same weakness and fallibility. For the Bible is not an epistemological competitor of, or substitute for, sense experience. On the contrary, its epistemological function is dependent upon sense experience. If God has chosen to reveal his truth in the Bible, then he has chosen to reveal it through the senses, and the Christian philosopher must make his peace with them. If they are fallible, then they are fallible, and it will be the task of a true epistemology to show how the fallible senses fulfill their role in conveying to us the sure knowledge of God. If to do this is to be empiricist, then Christian philosophy must be empiricist.

IV. Revelation and The Westminster Principle

In the preceding section I discussed the theory that the Bible provides a ground of knowledge that is superior to all other humanly available grounds. In this section I want to discuss a related theory, one that claims that all the knowledge necessary to Christian faith and life can be derived from the Bible. Clark may have held that theory at one time (or he may indeed still hold it), and the theory, which he expresses in his essay here, may well be a psychological descendant of that theory. At any rate, it is so similar to Clark’s view in some ways, and it has so deeply influenced most of us who are evangelical Protestants, that I want to consider it briefly here.

The theory I have in mind is stated in the Westminster Confession of Faith with such admirable precision and clarity that I propose to call it “The Westminster Principle.” In that confession we read: “The whole counsel of God concerning all things necessary for his own glory, man’s salvation, faith, and life, is either expressly set down in Scripture, or by good and necessary consequence may be deduced from Scripture; unto which nothing at any time is to be added whether by new revelations of the spirit or traditions of men.”

The Westminster Principle differs from Clark’s view, as expressed in his essay here, in at least three ways. First, Clark professes to make deductions from a single Axiom; the Westminster Principle calls for deductions to be made from the whole set of propositions expressed in Scripture. Second, Clark claims to derive all knowledge in this way; the Westminster Principle appears to limit its claims to religious knowledge.

Third, although the equivalent of the Westminster Principle could be stated in terms of an “axiomatization” of Christian doctrine this terminology is not usually used, nor is the principle generally recommended as providing for”systematization.”17 Instead, this theory is usually put forward as providing a very high principle of religious authority, recognizing nothing but God’s Word as authoritative, and as providing for a fixed and purely “objective” standard against which all religious proposals may be judged. These, indeed, are the claims in which I am interested.

Nevertheless, the principle shares certain prominent features with Clark’s procedure. (1) Both of them profess to deduce an important body of knowledge, knowledge at least sufficient for all the needs of Christian faith and life, from a fixed set of propositions; (2) Both of them deny absolutely the possibility of achieving this knowledge, or any additional knowledge in the same area, by means other than this deduction; (3) Both of them appear intended to connect this knowledge with the concept of “revelation.” I believe that Clark’s own attempt does not succeed. But what shall we say of the procedure authorized by the Westminster Principle?

I have no doubt that a very important set of propositions are either “expressly set down in Scripture” or can be deduced from those. Nevertheless, we must not assume, without investigation, that the Bible provides a deduction for every doctrine, which we may have heard from our childhood. As in the case of the Axiom, we must investigate to see what deductions can actually be made. For the Westminster Principle is really quite stringent, and if theologians attempted to follow it some of their books would no doubt be shorter than they are. Indeed, if the authors of the Westminster Confession had followed this principle, that confession would have been substantially shorter than it is. I believe that, in fact, it is not possible to deduce all that is necessary for Christian faith and life in the only way that this Principle allows. Consequently I believe that there must be a source of theological knowledge other than that which the Principle recognizes. While several examples could be given of theological knowledge for which this must be true, I will here discuss only a single important case, one that is prominently illustrated in the Westminster Confession itself.

In that confession, we read “Under the name of Holy Scripture, or the word of God written, are now contained all the books of the Old and New Testament, which are these:” At this point there is a list, by name, of the 66 books of the Bible. And then we read, “All which are given by inspiration of God, to be the rule of faith and life.”

Now, the Westminster Confession frequently appends to its statements a list of references that are supposed to provide the Scriptural justification for these statements. The references cited here, however, are somewhat surprising. They are Luke 16:29, 31, Ephesians 2:20, Revelation 22:18, 19, and 2 Timothy 3:16. Putting the most charitable construction upon them, we might suppose that we have here references to “all scripture,” Moses, prophets, apostles, and “this book” (i.e., Revelation) as being parts of the Bible. There is not a word in these passages to specify what particular books belong to “all Scripture,” nor to specify any particular book as being “apostolic.” Yet we are presumably to believe that the Westminster divines found it possible to deduce the complete canonical list of 66 books from these six verses “by good and necessary consequence!” I do not believe it.

Perhaps they used other passages? Perhaps, though in view of the importance of the matter, it is surprising that they did not cite them. And who will cite them for us? Who will give us a set of Biblical statements, rather than statements from the Fathers, which entail “Mark is an apostolic book”? Modern “introductions,” discussing the canonicity of Mark, cite Papias, Justin Martyr, Irenaeus, Tatian, Clement of Alexandria, Tertullian, Origen, and Eusebius in support of its Marcan authorship and Petrine source. But they cite no Biblical statements to that effect. Will it be said that the question of authorship is not crucial? Very well. I do not contend for it. Who will show us a set of Biblical statements that entail “Mark is part of the written Word of God?” These statements, too, have apparently escaped discovery up to now. And so have the Biblical statements from which we could deduce the canonicity of Matthew, Hebrews, James, 2 John, etc., etc. I suspect that there is a very good reason why these verses are neither cited by the Westminster divines, nor by theologians who accept the Westminster Principle, nor by Biblical scholars who discuss canonicity. It is simply that no such verses are to be found anywhere in the Scriptures. The plain fact is that the list of canonical books is nowhere “expressly set down in Scripture,” nor does the Bible contain anywhere any set of statements from which that list can be deduced “by good and necessary consequence.

I want to make two points clear at this juncture. First, I am not saying that the Westminster list of Biblical books is incorrect or unjustified. I am saying that the list of books simply does not conform to the Westminster Principle, and therefore cannot consistently be accepted along with that Principle. But we may, of course, decide to accept the list and to reject the Principle.

Second, I am not arguing that the failure occurs because it would be circular to deduce the canonicity or authority of books from statements occurring in those books. I think, indeed, that there is a sense in which such deductions would be circular and epistemologically useless. But I am not arguing here along those lines. Rather I am arguing that, even if this objection were waived and totally circular arguments are accepted, the required deductions still cannot validly be made.

Now, I frankly have much more confidence in the Westminster list of canonical books than I have in the Westminster Principle. Since I cannot consistently have both, I am inclined to reject the principle and let it go at that. But perhaps we can go further. If one accepts the Westminster Principle, then it appears to be very important to find out just what books are in the Bible, to identify it accurately. After all, we are told that it “contains the whole counsel of God concerning all things necessary” to salvation, etc. Nothing, absolutely nothing, is to be added to it at any time, not even by the Holy Spirit himself! This conviction is reinforced when we read even earlier in the confession that “the holy scripture is most necessary, those former ways of God’s revealing his will unto his people being now ceased.” If Holy Scripture is most necessary, then it would seem that getting it right would be very important. And, indeed, it seems plausible to suppose that if a theologian who accepted the Westminster Principle should alter his opinion on the Canon, coming to believe, say, that James, John, Romans, and Isaiah were not canonical, while Tobit and II Maccabees were, then his faith and practice might also be somewhat altered.

It is, therefore, entirely in keeping with the requirements of the Westminster Principle that a list of Biblical books should be specified in the very first chapter of the Confession. If the Principle is correct, nothing can be epistemologically more basic, more fundamental, than that list. No theologian who accepts the principle has any right to go any further until he has definitively settled the question of the canon, the question of the reliability of the present text, and the question of the accuracy of the translation (if any) that he will use. These questions, however, are simply the serious and substantial analogues of our somewhat extreme speculation in section III that someone may have replaced Clark’s Bible. They are the technical and detailed way in which we ask whether we (or the Westminster divines) have the authentic, complete, and unadulterated Bible.

Now, the Westminster Principle recognizes only one standard for answering questions, which bear upon our faith; deduce the required proposition from the Scripture! The Confession provides an answer to our question on its own authority, but it does not provide the required deduction. The Bible just does not have the required statements in it. The Westminster Principle then makes the question of canonicity crucial, and simultaneously makes it unanswerable.

If a man wants to continue in the Christian life, then, he seems to have two alternatives. He can explicitly reject the Westminster Principle, and then go on (if he is philosophically inclined) to construct an epistemology upon some other lines. Or he can continue to accept that Principle while making surreptitious exceptions to it as they become necessary. Furthermore, some of these exceptions must be made, not in peripheral matters, but at the crucial points, as the foundations themselves are laid. No doubt many people live in this way (the Westminster Confession certainly appears to have been written in this way). But sometimes, even if only vaguely, the inconsistency must become apparent even to the unsophisticated believer. If he opts for the Principle at this point, he faces a dreadful conclusion. He must believe that, whatever may have been true of Isaiah or Paul, his own faith rests entirely upon a merely human judgment. He has read and studied a book. He has used it to guide his faith and life. But is that book which he has used truly the Word of God? He has the testimony of Eusebius, Tertullian, Augustine…right on up to the modem bookstore clerk and the modem scholar. But, since the required proposition is not in the Bible, the Westminster Principle requires him to believe that he has no word from God on this question. To assure him that he does have the Bible, he has nothing but the word of man, from beginning to end. At the cornerstone of the whole edifice of his faith, the Westminster Principle allows him to recognize nothing but the purely human.

The Westminster Principle, then, promised a very high view of religious authority. Its actual effect, however, when consistently followed, is to drive one to a very low view of that authority. It professed to give us complete objectivity and a fixed unyielding standard. It ends by driving us to accept, often at secondhand, purely human opinions in a difficult and recondite science as a preliminary to our having any standard at all. It seemed to exalt the Word of God. But in the end he who follows it consistently must believe that his faith is built upon nothing but the word of man. The Westminster Principle, in short, effectively removes the concept of revelation, as an actually operative, working concept, from Christian epistemology.

Now, it seems to me that I can no longer be satisfied with a principle, which has to be evaded in the first chapter. Nor can I any longer believe that God is silent when the foundation is laid, and begins to speak only when the roof is being erected on the edifice of faith. I suggest, therefore, that one of the most pressing problems now facing Christian philosophers and theologians is precisely the reversal of the effect of the Westminster Principle, and the effective re-introduction into Christian thought of the concept of revelation. I mean its re-introduction as a working concept, one with a present tense, one actually to be used in giving an account of how the knowledge of God is acquired today.

I suggest that we try applying it to the simple case of the believer, and ask if God has spoken to him, as well as to his fathers. If so, how? Let us apply it to the hard case of the questions about the canon and the text of the Bible. Does God speak to these questions? Where, and how? (Is it, perhaps, in the testimony of the church, or in the books of the scholars? Does God’s word occur there, as well as the word of man?) And what is the proposition that He reveals? We can learn by applying it to the humble case, the case of the missionary who says, “God called me to Peru.” He finds it natural to use the same form of words as Luke did (Acts 16:10). Did God call him to Peru? If so, then what God revealed to him cannot be deduced from everything in the Bible. But if he had been unfaithful to that word of God, it would have been a tragedy. It would have made all the difference to his life. What was the mode whereby God revealed that word to him?

I know there will be objections. I know that someone will fear that if we do this, we will lose infallibility, and some things will be mistakenly counted as revelation when they are not. And this is true. But we must not forget that the senses, too, are weak, variable, and prone to error. Their fallibility is a commonplace of philosophy, but it is known to every untutored peasant also. And yet every bit of knowledge that we glean from the Bible is filtered and funneled through those fallible senses. And after that it is exposed again to the fallibility of our own minds. The Westminster Principle has by no means prevented mistakes and disagreements over what was revealed. We will not lose this sort of infallibility, for we have never had it.

Some will ask, too, how the revelation of something, which is not in the Bible, is to be distinguished from demonic deception or from psychological quirk. A true, detailed, and informative answer to that question is very difficult. I do not know how to answer it. But a useful propaedeutic to an answer is to ask the questioner how Isaiah or Paul, when they received revelations which went beyond what was ‘already written, distinguished them from demonic deceptions and psychological quirks. How did Abraham know that it was God who called him to Mt. Moriah, rather than the devil or his own complexes? The answer is also very hard, and yet it seems that the thing itself must have been done. How was it done? Whoever gives a true and informative answer to that will contribute a great deal toward answering the question that will be asked of us.

Clark’s essay, then, has driven us far out into the problem of epistemology, into a consideration of how we come to know God. In some respects his essay is extreme and radical; that is part of its provocativeness. But basically, I think, it lies within a philosophical tradition, which has dominated evangelical thought for some time. His epistemology, as expressed here, is like that of the Westminster Principle in a fundamental way. We might call them “linear” epistemologies. If they were to be pictured, they would represent the process of knowledge, or of revelation, as a line or chain in which everything hangs from a single link. In Clark’s essay, the single link is the Axiom; in the Westminster Principle it is the Bible. I find myself dissatisfied with that tradition. Let me therefore conclude by suggesting a different picture. My picture is that of a net which is attached to God at many points, all around its circumference. In its center is the present man, who also touches it at many points. The net is formed of many strands, and of many kinds of strands; the Bible (no doubt one of the more important), history, tradition, science, scholarship, sense experience, mystical experience, logic and deduction, joy, pain, crisis, and perhaps many others. They cross and recross each other, knotted, twisted, braided, spreading apart, converging and diverging many times before they span the gap between God and man. And God’s contact with the man, His speech to the man, is communicated through the whole net, through the whole world and life that He has created.

It is, of course, only a picture. But perhaps for someone it will help to break the domination of the picture of the chain. That in its turn may lead us to a fuller understanding than we have yet achieved of the way in which God speaks, not only to our fathers, but also to us.