I wrote what there is of this essay in the mid-1980's and then put it aside to deal with more pressing projects and deadlines. For years after I looked for the time to to return to it and finish it, but without success. Now I put the finished parts on the web (December 2001) because I finally know that I won't return to it, at least not for many more years.

The Problem of Beginning
Peter Suber, Philosophy Department, Earlham College

"We're here together, so begin!"
Goethe, Faust, I.ii.269

Lovejoy's Lament

In the preface to The Great Chain of Being, Arthur Lovejoy remarked that the idea of which his book was a history was for many centuries one of the dominant ideas in western philosophy. He lamented that even hearsay knowledge of it had died away so completely that he had to explain to his readers just what the idea was.

The problem of beginning has suffered a similar fate. For many centuries it was a pervasive and foundational problem addressed in some degree by virtually every systematic philosopher. Failure to solve it satisfactorily was a ground of dismissal. Solving it, however, invariably required a few conspicuous odd kinks in the architectonic of the system, or a few conspicuous weak admissions, which affected the substance and plausibility of the entire doctrine. When we read a system in ignorance of what this problem was and what it required, we may seize on these kinks or admissions as defects simpliciter and forget that the obvious alternatives do not solve the same problem at all.

Unlike Lovejoy, I am not writing a history. I want to give a schematic picture of how the problem has been approached. Eventually, I hope to write an historical treatment of these approaches that will make better sense of the approaches themselves and show their degree of integration with the positions for the sake of which the problem was solved.

Unlike the great chain of being, the problem of beginning did not have a name. It was, in effect, a logical field force in which rigorous and, particularly, systematic philosophy, was conceived and written. To name it and to show the many traditional approaches to it, and to recognize the costs of solving it, should reawaken interest —and sympathy— for the thinkers who took it seriously. I hope it also reawakens some serious interest in the problem itself.

The Problem

Either one proves one's beginning or one does not. If one does not, then there is no reason for us to accept it. But if one does, then two problems arise. First, it seems that one has begun before one has begun, since the premises of the proof of the beginning must be logically prior to the conclusion. Second, one has a new beginning (the premises) which will either be proved or go without proof. Either one continues backwards in an infinite regress, or one stops with a self-justifying premise, or one stops with an unproved beginning. If one stops with an unproved beginning, then there is no reason for us to accept it. If one stops with a self-justifying premise, then apparently one commits a fallacious petitio principii. If one regresses infinitely, then we never hear the final reason for us to accept it.

In a nutshell, the problem is the dilemma of circular reasoning versus infinite regress. In some skeptical literature it is expanded to a trilemma, adding the possibility of a merely stipulated beginning, or an ipse dixit, which is equivalent to starting without proof. Because all are unattractive, and together, it seems, exhaustive, the philosopher who wants to establish her position seems torn on the horns of this trilemma. (For more on how the Greek skeptics used this trilemma and related arguments, see my essay, Classical Skepticism.)

See Sextus Empiricus, Outlines of Pyrrhonism, I.122-23, I.170-172, I.114-117, II.7-9, II.18-20, II.85, II.92-93, III.35- 36, III.53, III.67-69, III.241-242. Hegel, Lesser Logic, Sec. 1. Hans Albert, p. 18.

Some Distinctions

When we speak of beginnings, we may refer to the first members of many different series. We may refer to the logical prius, the first principle or premise from which all derivative conclusions follow by deduction or some other method of inference. We may refer to the temporal prius, the first moment of some process from which all subsequent moments follow by temporal progression. We may refer to the causal prius, the first cause from which all subsequent (hence, contingent or conditioned) events follow by causal necessity. We may refer to the first step in a voyage of discovery, the first glimpse of the eventual position from which the rest follows by a process of learning. Or we may refer to the first assertion in the written account of the position and its justification, the chosen introduction in the realm of language where —all lament— not everything can be said at once.

Hence, a beginning may be the beginning of at least the following orders or sequences: of proof (or justification), time, causation, discovery, and exposition. When I need adjectives I will refer to these as the orders of logical, temporal, causal, discovert, and expositional priority.

(On the word "discovert". The OED lists no adjectival form for "discovery", which is a sad hole in the fabric of English. "Discoveral" sounds barbaric to my ear. "Heuristic" comes from a different root and has acquired different connotations. I use "discovert" on the analogy of "cover" and "covert". As with "cover/covert", the accent must shift to the ultima in the adjective.)

These distinctions will help us to state both the problem of beginning and the approaches to its solution with greater precision. We are concerned primarily with the logical beginning. It will only rarely coincide with the causal, temporal, or discovert beginning, although it is often the expositional beginning. The problem arises from the question: how does one justify one's logical beginning? If one tells the story of one's path of discovery, one may have warmed the audience enough that it does not demand a proof of the logical beginning; but this will not solve the problem, of course. If one leaves the logical beginning unproved, then one's grand argument hangs on an unproved premise. But if one proves it, then one has shifted the actual beginning of the logical series one step back and replicated the problem there. The proof of the (original) beginning will have its own premises or logical beginning which will either be proved or go without proof.

With these distinctions in mind, we are saved from attractive but equivocal answers to the question: What is the (logical) beginning of so-and-so's philosophy? We cannot say: so-and-so's recollection of David Hume, awakening her from her dogmatic slumber; her first philosophy course in college; or her first suspicion that such-and-such might be the case. For at best these are causal, temporal, or discovert beginnings. We cannot answer: so-and-so's parents (but cf. Descartes, Meditation III, at p. 105); the big bang; or the first edition of her book. For at best these are causal or temporal beginnings. Nor can we say: the first word, sentence, paragraph, chapter, or philosophical assertion in so-and-so's book, since at best these are expositional beginnings.

When Kant said that the freedom was the ratio essendi of the moral law, while the moral law was the ratio cognoscendi of freedom (Kant, Critique of Practical Reason, Beck trans., p. 4.n.), he was saying that freedom is prior in being but posterior in discovery to the consciousness of the moral law. When he denies that synthetic a priori judgments are innate ideas, he is saying that they are logically but not temporally prior to experience. We must become accustomed to making distinctions of this kind when examining solutions to the problem of beginning. The distinctions are familiar to philosophers who acknowledge (in Kierkegaard's words) that we live forwards but think backwards. If time and causation unfold in one direction, we as inquirers swim upstream seeking their source or antecedents.

Because we are interested in the logical beginning, we are more concerned with grounds or foundations than origins. The problem, then, might well called the problem of foundations rather than the problem of beginning. I'd rather avoid calling it the problem of foundations, however, since that would imply that non-foundationalist epistemologies avoid the problem. They don't. In fact, they typically don't want to. They encounter the problem and have characteristic responses to it, just like foundationalist epistemologies.

We all know how justification can proceed once premises are given. The problem studied here is how we can justify our ultimate premises. If we dispense with a premise-conclusion model of justification, but think our position justified some other way, the problem is how to attain a kind of justification which will answer doubts even about the model of justification. Either way, the problem is how we can achieve final or complete justification. Hence it is about method, criteria, certainty, and system. It is about answering skepticism. (Hence also it has a prima facie application even to non-foundational epistemologies insofar as they want to be justified in their own ways.)

To take the problem seriously does not require us to assume that complete justification is attainable. Indeed, to deny that complete justification is attainable is one solution to the problem.

Solutions and Approaches

One may appreciate the pervasiveness and foundational character of the problem by recognizing it as one of the more piercing and enduring challenges of skepticism. The typology of the replies, sketched here, shows the need to go to ultimate or foundational philosophical strategies —to First Principles— in order to meet its terms or even to abjure them seriously.

Note that several philosophers appear in more than one section below if it seems that they used more than one approach to the problem of beginning. I will not try to give definitive textual interpretations here, but will classify philosophers wherever plausible readings permit. I am more concerned to give a good outline of types than exegetically accurate classifications, although I hope I have given both.

Moreover, I will not assess the adequacy of the solutions and approaches collected here. I will simply show how a solution meets the terms of the problem and what price is paid for that degree or kind of adequacy.

Here are the types to follow:

1.1  Justify the beginning
1.2  Begin with metaphilosophy
2     Leave the beginning unjustified
3     Begin with a truth which needs no proof
3.1  Begin with the self-evident
3.2  Begin with faith
4     Accept a circular beginning
4.1  Begin with a self-justifying principle
4.2  Justify the beginning by the middle or end
4.3  Identify the beginning with the end; prove the beginning by proving the end
5     Begin anywhere
5.1  Regress to the logical beginning from wherever one starts
5.2  Render coherent what you find
5.3  Reject ideal of certainty, insist on corrigibility
6.1  Begin with the presuppositionless
6.2  Begin with something negative
7     Begin with error, not truth
8     Begin with something non-cognitive, not knowledge
9     Begin with the idea of the first cause
10   Begin with realities, not the ideas of realities
11   Don't begin
12   Reject the problem

On first inspection, it is plain that there are two primordial approaches to take. One is to try to prove one's beginning and one is to admit that one's beginning is to go without proof. These are the first two solutions to be considered.

1.1.  Justify the beginning

There are two ways to justify one's beginning: one may justify it before or at the beginning, or one may do so after having begun. The first will be treated here; the second will be treated below (#4), on circular beginnings.

Surprisingly few philosophers may be found in this category. I will use Kant as my prime example.

Before writing the system of metaphysics that he envisioned, Kant (like Locke) had to examine the faculty of reason that would construct the system. The Critique of Pure Reason was the result. It is a "propaedeutic" to the system, not the system itself (B.xliii). There are important differences. The Critique, by and large, employs the critical method while "[i]n the execution of the plan prescribed by the critique, that is, in the future system of metaphysics, we have therefore to follow [the dogmatic method,] the strict method of the celebrated Wolff..." (B.xxvi). The Critique "is a treatise on the method, not a system of the science itself" (B.xxii). However, Kant said, "[s]uch a system of pure (speculative) reason I hope myself to produce under the title Metaphysics of Nature" (A.xxi). Not only the system of metaphysics, but also the "system of pure reason" and the "system of transcendental philosophy" (if indeed these three systems are different) are logically posterior to the Critique:

[W]e can regard a science of the mere examination of pure reason, of its sources and limits, as the propaedeutic to the system of pure reason. As such, it should be called a critique, not a doctrine, of pure reason....A system of [transcendental] concepts might be entitled transcendental philosophy. But that is still, at this stage, too large an undertaking. (B.25, emphasis is Kant's.)

In short, "[t]ranscendental philosophy is only the idea of a science, for which the critique of pure reason has to lay down the complete architectonic plan" (B.27).

Kant could not make a logical beginning in the system of philosophy he envisioned without first demonstrating that he had the right method, that the concepts he would use were sound, and that the plan of the system was within the reach of reason. Kant is responding to skeptical objections to the system before the system is written.

Kant was quite self-conscious in the first Critique in laying out the grounds for the acceptability of the results of the critical method. He was not as conscious of the question whether the Critique was itself known by the standards it laid down or whether its premises were inside or outside the system. If the Critique is an example of the kind of knowledge for which it is supposed to be the propaedeutic, then he may have collapsed levels inconsistently, in a way that is eventually circular. But if it is not, then it may not be knowledge at all, for the Critique spends little time identifying a type of assent which could capture the Critique itself. This uncomfortable dilemma is one way to state the price Kant pays for trying to justify his beginning.

There is another way to put it that does not depend as much on how the case turns out upon close reading. He has 'begun before he began' in the sense that he has retreated one or more logical steps prior to the beginning of the system to examine and justify his methods and criteria. Hegel's objection to this practice is famous:

Kant undertook to examine how far the forms of thought were capable of leading to the knowledge of truth. In particular he demanded a criticism of the faculty of cognition as preliminary to its exercise.... Unfortunately there soon creeps in the misconception of already knowing before you know, —the error of refusing to enter the water until you have learnt to swim. (Lesser Logic, Sec. 40.z, p. 84.)

Schopenhauer's complaint is equally vivid: Kant sang a song and danced to it afterwards. The remedy, in Hegel's words, is to recognize that "the examination of knowledge can only be carried out by an act of knowledge. To examine this so-called instrument is the same thing as to know it. But to seek to know before we know is as absurd as the wise resolution of Scholasticus, not to venture into the water until he had learned to swim" (Hegel, ibid., Sec. 10, p. 17.) To Hegel this will require an explicitly circular beginning, as we will see in #4 below.

There is a third price that Kant pays for his strategem of justifying his beginning at the beginning, or before: he opens the door to an infinite regress and creates incentives to pass through the door. If we ask why the Critique is justified, or what are the premises of its "apodeictic proofs", we will either be driven yet one further step back in the regress (and so on, indefinitely), or we will reach a self-justifying step, or we will come to the end of justification and admit an unproved premise. These are unattractive alternatives for most philosophers; I think it fair to say that they are especially unattractive to Kant.

It does not follow that he had an elegant way out. On the contrary, Kant's tendency seems to have been to cope with the problem of beginning by writing propaedeutics to the propaedeutics, retreating further from the beginning, as if for a running jump. In both the systems of morals and of natural science, Kant had critiques to examine the appropriate faculties, and foundations to the metaphysics of each, and for the system of metaphysics, prolegomena as well. From this angle Kant's intellectual career was like Tristram Shandy's: only toward the end of his life did he begin to finish the prefatory matter and come to the beginning.

In his logical regressions he may have been in search of a presuppositionless beginning that would not require further regress, and a strong case could be made that he thought he found such a thing in criticism (see #6 below).

In addition, Kant probably thought he was innocent of all these charges. I quoted him at length on the distinction between the Critique and the future logically posterior system in order to show on what basis his critics could object to his method of beginning. But in fact, toward the end of his life, Kant made the following remarkable claim, inconsistent I believe with the textual evidence of the Critique:

I must remark here that the assumption that I have intended to publish only a propaedeutic to transcendental philosophy and not the actual system of this philosophy is incomprehensible to me. Such an intention could never have occurred to me, since I took the completeness of pure philosophy within the Critique of Pure Reason to be the best indication of the truth of my work. (Zweig, at p. 254.)


1.2.  Begin with metaphilosophy

An important variation on the method of justifying one's beginning is to start one's philosophy with metaphilosophy. The attraction is similar to what must have attracted Kant: by starting at a different level from the eventual system, one somehow avoids precipitating an infinite regress of justifications. This is much more a sense to be articulated than a self-evident principle. 'Meta' levels can regress infinitely, just as proofs can within any single level.

Fichte is one example of a philosopher who (on one reading) began with metaphilosophy. While he wants to say that the logical beginning of his system of idealism lies in an intellectual intuition of freedom or in an utterly free choice (discussed further below), his expositional beginning gives us a definition of philosophy that sets the terms for the logical beginning. After asking why some presentations are given to us accompanied by the feeling of necessity, he says that every philosophy must answer this question. (Fichte, p. 6.) He moves smoothly from here to the disjunction between realism and idealism as the only two possible systems, where his more interesting discussion of the problem of beginning occurs. It is fair, then, to see his metaphilosophical remarks as more than a mere expositional beginning, but as a logical beginning that shapes the question and answer of his position.


2.  Leave the beginning unjustified

Instead of justifying one's beginning, one may cheerfully admit that it has none. Usually one thereby signals, more or less explicitly, that one is happy to remain hypothetical or arbitrary. The price paid for this solution is obviously steep.

This is the 20th century attitude toward axiomatic systems in logic and mathematics. Formerly, axiomatic systems were not even attempted unless the axioms were thought self-evident, or at least true. But now we are happy to invent formal systems from arbitrary premises just to explore the logic of axiomatization, or to reduce a complex system to publicly scrutable primitives.

Logicians are also responding directly to the problem of beginning. No inquirers know better than logicians that justification comes to an end. Arguments need premises to certify conclusions. For purposes of logical analysis, the premises are taken as simply given, not as self-evident or as true. If one chooses to prove the premises, one is simply taking a new set of premises as given; and so on. Logicians accept that any argument must proceed from unproven premises, even if those premises could be proved in another argument. While even ancient logicians, such as the Stoics, recognized the need for indemonstrabilia to get off the ground, they tended to see their indemonstrabilia as self-evident; modern logicians are more humble. As a result, modern logicians rarely claim truth for their formal systems, just validity. This is another form of the price paid for this solution to the problem of beginning.

(The history of the transition —the fall— in logic and mathematics from truth to validity, or from the assertion of self-evident truth to contentment with the hypothetical is very well told in Morris Kline, Mathematics: The Loss of Certainty, Oxford University Press, 1980.)

A similar result occurs in ficitionalism. If a position does not claim to be true, then it need not prove its beginning to be true; thereby, perhaps, it evades the problem of beginning. Self-conscious fictionalists may not need to prove that their beginning has any particular virtue, but if they do, they need prove only utility, beauty, vitality, or some comparable virtue, not truth. Needless to say, fictionalism is also a price too high for most philosophers to pay willingly.

This is also the position of any 'existentialist' who begins with a sheer choice. The logical beginning is the act of will, not the 'reasons' that may or may not have guided it. This is one way to read Fichte. He argues that idealism and realism are each capable of explaining experience adequately, and are incapable of refuting one another (Fichte, pp. 77, 93). Hence, one faces a perfect balance and can only adopt one of the positions by choice. Because of the balance or skeptical isosthenia, the choice must be intellectually groundless (12-14, 35, 70, 77, 102). It is based only on the kind of person one is (16, cf. 75f) and interest (15, although this interest is rationally vindicated at 40). The fact that Fichte also argues that idealism is the superior choice (11-16, 18, 19, 27, 40, 58, 65, 71, 147, 197, 202) is just a symptom of his struggle with the problem of beginning, trying at once to prove his beginning and let it go without proof as an act of will.

Several prices are paid for this existential variation. One must admit that all reasons are limited and insufficient. The system is either not true but just chosen, or is true by pre- established harmony, or is true by arguments that appear inconsistent with the framework of the choice. It will probably be or appear to be hypothetical. It will probably be or appear to be non-cognitive. Further, one must give up trying to persuade the reader by argument. One must instead try to elicit the same idea in the reader even while admitting that it cannot be proved (11, 16, 33, 66, 70, 76, 78, 85, 91, 223, 250). If one has already forsworn ordinary argument for these reasons, and if one holds that one of the available options is actually preferable, then one may use ad hominem arguments and pity one's critics as immoral or immature (16, 24, 32, 43, 75, 79, 85, 251).

One way to read Kierkegaard and Sartre makes this kind of beginning seem necessary. If you say you begin for such-and-such reasons, they will ask why that is a good idea. Whatever criteria you give, they will ask why your criteria are good criteria. This will persist until it is clear that you have chosen your position to match your criteria, and that you have chosen your criteria to match some other desideratum, and that ultimately you have chosen the whole show or at least its logical beginning. On this account, one either admits the choice and takes full responsibility, or one denies one's freedom and lives in bad faith. If this works, then we might say that we all begin with choice in this sense, although not all of us know it or admit it, and hence not all choose this method as their solution to the problem of beginning. Of course, not all of us regard sheer choice as an adequate solution to the problem of beginning either.

One may also read the position of Johannes Climacus in the Philosophical Fragments in this way, but not for the reasons in the last paragraph. He seems to offer us a choice between the Socratic theory of recollection and its denial. He develops the denial into a form of Christianity but without explicitly arguing that it is for that reason to be preferred. (I do not read the Fragments in this way because I find in it implicit arguments for the denial of recollection.)

An important variation on the method of leaving the beginning unproved is the partial concession to skepticism. Many non-skeptics who are familiar with skeptical arguments happily admit that infallible, incorrigible, indefeasible, perfect certainty is probably attainable, or at least that they have stopped seeking it. When a dogmatist (non-skeptic) abandons the search for certainty and proof, she will also reject the demand for certainty and proof. Insofar as the problem of beginning makes these demands, this kind of reformed dogmatist will not feel threatened by it. Of course this does not solve the problem; it reaches an accomodation with the failure to solve it. So while other variations on this method cheerfully admit hypotheticality, or arbitrariness, or groundless choice, this variation cheerfully admits uncertainty. (See Sections 5.1, 5.2.)

It might be added that this is very much what the skeptical propounders of the problem have in mind.


3.  Begin with a truth which needs no proof

A synthesis of 1 and 2 is to claim that one's beginning is true but in a way that needs no proof. This is to be distinguished from a beginning that proves itself (see #4.1)


3.1.  Begin with the self-evident

This looks like the ideal solution. The regress of justification comes to a halt, but without a circular or self- justifying prius. The logical beginning needs no proof, and yet is not hypothetical.

As noted above, this was the theory of beginning in logic and mathematics for centuries. For example, it was the universal view of Euclidean geometry until the 19th century. For many philosophers this is the only way to explain the propriety of the principle of non-contradiction and other First Principles of logic and mathematics.

One problem with this approach is the historical record. Self-evidence has been claimed for principles which inquirers have almost unanimously rejected only a few decades or centuries later. Self-evidence seems to be a byproduct of culturation or paradigm-influenced perception, not a theory-free anchor by which to judge theories. New discoveries have forced us to unlearn the self-evidence of the commensurability of all numbers, the motionlessness of the earth, the parallel postulate of Euclid, and most recently the naive definition of a set as any collection of any elements.

If that is the diachronic problem with self-evidence, the synchronic problem is the frequent failure of unanimity on what is self-evident, even in the same generation in the same culture. What does it mean to call a principle self-evident and face disagreement? By the nature of the case, there can be no argument in support of the claim of self-evidence. Either the principle loses its certainty or the dissenters are diagnosed as defective. In the first case, the proponents of the principle (by their own standards) have lost their license to begin with it. In the second case, the logical rudeness to dissenters, even if justified, has almost the same effect. If to solve the problem of beginning means to satisfy the doubts of inquirers, then any approach fails which merely explains doubts away without answering them, or which stigmatizes doubters without satisfying them. (For more on this, see my essay on Logical Rudeness.)

Fichte is in this position. As noted, on one reading he began with the self-evidence of freedom. Yet he knew that not all philosophers were aware of their freedom. His solution was to describe a crude dialectic of enlightenment or maturity that could explain this benighted condition without validating it. This is just one way in which Fichte anticipated Hegelian dialectic. But while Hegel is magisterial and somewhat clinical in describing the inadequate standpoints of immature spirit, Fichte cannot resist the temptation to belittle his critics. "A person indolent by nature or dulled and distorted by mental servitude, learned luxury, and vanity will never raise himself to the level of idealism" (Fichte, p. 16). "The ground of this inability of theirs does not reside in any special weakness of intellect, but rather in a weakness of their whole character" (74). "[T]his is an observation so striking, that anyone who fails to grasp it, and is not thereby uplifted into transcendental idealism, must unquestionably be suffering from mental blindness" (105).

Spinoza found himself with no better responses than Fichte, although the logic of his beginning was quite different. After a very self-conscious discussion of the problem of beginning and his solution, Spinoza said:

If there yet remains some sceptic, who doubts of our primary truth, and of all deductions we make, taking such truth as our standard, he must either be arguing in bad faith, or we must confess that there are men in complete mental blindness, either innate or due to misconceptions —that is, to some external influence. (Improvement of Human Understanding, p. 17.)

Once the strategy is pointed out, one recognizes many cases of it: Max Scheler on the self-evidence of his scale of values and the "value blindness" of critics; some (not all) Marxists on Marxism and the ideological distortions of critics.

In mathematics, as noted, the theory of self-evidence has largely been abandoned. It remains, however, for some entire schools of metamathematics, and for some branches of mathematics that have resisted successful axiomatization. The outstanding example of the latter is set theory. There is still no demonstrably consistent and complete axiomatization for set theory. Lacking that kind of foundation, many are tempted to ground it in the self-evidence of its primitive definitions and operations. This worked until Russell's set paradox showed that our intuitions are naive and can harbor contradiction.

It should perhaps be noted that much of the metamathematical enthusiasm for recursive function theory today lies in its apparent fulfillment of the old dream of self-evidence. Its simple primitives and simple building operations seem to guarantee that no contradiction will arise in the systems built from them. The enthusiasm is also partly due to the mathematical value of building on this foundation, for it seems (by Church's thesis) that from this starkly simple beginning all the computable functions can be built. Note, however, that for many mathematicians this is self-evidence without truth; it is syntactic, not semantic.

A variation on the theme of self-evidence is to begin with clear and distinct ideas. As I read Descartes, clarity and distinctness differ from self-evidence in several ways. Clarity and distinctness must often be acquired by hard study (Descartes, at pp. 25, 86, 90, 99, 123), are a matter of degree (109, 115, 128, 131, 137), may sometimes be known only by inference (62), cannot occur to animal senses (89), and may err (77, 92). Like Fichte, Descartes can be seen struggling with the problem of beginning by trying out several inconsistent approaches. On the one hand, he wants the tidiness and certainty of beginning with a clear and distinct idea (of himself), but on the other, he can be seen beginning before he begins, trying to justify this beginning by arguing that clarity and distinctness are reliable because God is not a deceiver.

(For Descartes' method of beginning with doubt, see #6 below.)


3.2.  Begin with faith

This is a significant variation on the theme of self-evidence. One who begins with faith may rebuff the demand for proof, and yet still claim that the beginning is true. By holding it through faith, rather than knowledge, one is not hypothetical, arbitrary, or fictionalist, and yet one avoids the dilemma that seems to attach to any attempt at justification.

This position is common to an entire sub-tradition arising in Augustine and passing through Anselm (and Descartes, I would argue), to Kierkegaard. The position asserts the necessity of faith to knowledge: one cannot know until one believes. In some cases the position asserts only discovert priority of faith to knowledge. In others faith is clearly made prior to knowledge in both the discovert and logical series: knowledge cannot be justified as knowledge without premises taken from faith.

The chief problems with it are the same synchronic and diachronic problems associated with the claim of self-evidence. Diachronically, faith has historically supported claims wildly at variance with one another. This is not an objection to its use for any given claim, of course, but the inquirers who hear the claim may legitimately wonder why it should suffice here and now when it does not elsewhere and hitherto. Synchronically, faith can be found supporting virtually every position in the intellectual spectrum. For those trying to judge adequacy or truth, the logical beginning in faith is not at all a virtue or a discriminating feature. Perhaps it would not be even if only one position claimed it.

These two problems can be combined in one: the adequacy of beginning with faith can comfort only the faithful. To inquirers without faith (or without that faith) it provides no solution to the problem at all. If the point of solving the problem of beginning is to answer the skeptic, this method fails. It tells the skeptic that others have solved the problem but not in a way that can be shared or certified by argument.

While this method seems to avoid the demand for proof without Fichte's rudeness, replacing it with a frank and humble confession that no proof is available, it may in the last analysis be ruder still. The unfaithful must be exlained. Either they are defective (sinful), which is Fichte's position, or they have not been blessed with a gift, in which case they are unable to help themselves. One can only reply to their questions with sorrow or pity. (For more, see my essay on Fichte's Ad Hominem Arguments.


The price paid for the first solution, proving the beginning, is apparently a logical circle or an infinite regress. Hence the next two solutions to be considered will be frank espousals of these options, with serious attempts to make them unobjectionable.

4.  Accept a circular beginning

There are several important variations here, of which only three will be treated in detail. One may regress in the logical series until one comes upon a premise that is self-justifying (4.1). One may start initially without proof but then prove the logical beginning with the logically posterior system (4.2). Or one may deny the linearity of this model of derivation, and arrange for the logical beginning to be the end also (4.3).


4.1.  Begin with a self-justifying principle

Gaston Isaye, a Belgian Jesuit and professor of philosophy at the Facultes Universitires de Namur, has worked particularly on the problem of grounding philosophical systems with self-justifying premises. His method is called retortion. The name comes from the retort the philosopher can make to the skeptic who would question the first principle. For example, Isaye's own inspiration for this method is Aristotle. If Aristotle asserts the principle of non-contradiction and a critic denies it, Aristotle can retort that the critic presupposes the principle in the denial since the critic wants the denial to be true to the exclusion of its opposite (Aristotle, Metaphysics 1005b-1006a, 1007b, 1008). If the critic will merely question or doubt, rather than deny, then other retortable principles could be used, such as "Error is possible". Doubt of it seems to presuppose that error is possible. Retortion is possible with any principle that is presupposed by its own denial. Many of these are of use as first principles. (Isaye's position and explorations of retortable principles is well summarized in Martin X. Moleski.)

Descartes' cogito can be restated as an argument by retortion, although many argue that he did not intend it to be an argument at all. At times Descartes seems to retort to the demon in perfect Isayean style. Descartes in effect offers the assertion "I exist" to the demon, who tries to deceive him on this question. Descartes' retort is that he must exist to be deceived. (See Meditation II, p. 82.) To Descartes, the proposition of his own existence is presupposed even by his doubt or denial of it.

In any case the cogito is a self-justifying principle. If it is immune to doubt because even doubt confirms it, then its establishment is circular in a way that Descartes would insist was benign. [cite D's letter to Sorbonne]

Fichte —again— can probably be found in this category. One way to state his basis for choosing idealism when the choice was supposed to be balanced is this: when picking a first principle, pick the one that explains one's freedom to pick a first principle. [Cites from Fichte essay]


4.2.  Justify the beginning by the middle or end

By the fruits of the beginning shall we know it. Fichte says that this is the method that must be used by the realism he rejects. Its first principle is the existence of a Kantian thing in itself. Fichte believes there is no direct proof for the thing in itself, which might be Kant's position as well. Hence, Fichte writes, the thing in itself "cannot be looked upon as anything other than a pure invention, which expects its conversion into reality only from the success of the system" (Fichte, p. 11).

A similar method could be used for any first principle not subject to direct confirmation, such as the reality of god, freedom, immortality, or astral projection. Atoms were not within the arena of direct empirical testing when Dalton posited them in the first modern atomic theory; he believed in them only from "the success of the system" or their success in explaining the integral ratios of chemical combination.

This may look like a simple begged question: because the logical consequence is used as evidence for the logical beginning, the conclusion (here the beginning) presupposes itself. But if it is a vicious circle, it is not a simple one. One may even defend this kind of circularity by seeing it as mutual reinforcement by the parts of a coherent story. There is nothing apparenty vicious in the fact that witness A supports witness B and vice versa: on the contrary, if they are telling the truth about a non-contradictory reality, we expect their testimonies to support one another. To prevent viciousness we need this innocent mutuality, not one-sided dependency that is cunningly closed in a loop. (For more, see my essay, Question-Begging Under a Non-Foundational Model of Argument.)

Bertrand Russell seems to approve this non-vicious kind of circularity in An Essay on the Foundation of Geometry (p. 58):

Knowledge once existent can be analysed, but knowledge which should have to win every inch of the way against a critical scepticism, could never begin, and could never attain that circular condition in which alone it can stand.


4.3.  Identify the beginning with the end; prove the beginning by proving the end

This is one of several ways to read Hegel —a philosopher obsessed with the problem of beginning. In the Phenomenology he says that the True "is the process of its own becoming, the circle that presupposes its end as its goal, having its end also as its beginning; and only by being worked out to its end, is it actual" (10). The movement of his dialectic "is the circle that returns into itself, the circle that presupposes its beginning and reaches it only at the end" (488). When in the last chapter he proves —in the dialectical sense of "proof"— that spirit knows itself immediately, he says that spirit is thus "the certainty of immediacy, or sense-consciousness —the beginning from which we started" (491).

In the Lesser Logic Hegel put it slightly differently (pp. 27-28; cf. 379):

The very point of view, which originally is taken on its own evidence only, must in the course of the [philosophical] science be converted to a result —the ultimate result in which philosophy returns into itself and reaches the point with which it began. In this manner philosophy exhibits the appearance of a circle which closes with itself, and has no beginning in the same way as the other sciences have.

Because the end of the dialectical progression is the same as the beginning, but more explicit or unfolded, the proof of the end is equivalent to a proof of the beginning. The elegance of this solution is enhanced when we notice that the method of proof of the end is just the unfolding or explication itself. As he puts in in the Lesser Logic, "[t]he absolute idea may in this respect be compared to the old man who utters the same creed as the child, but for whom it is pregnant with the significance of a lifetime" (375). Or as T.S. Eliot put it in "The Four Quartets": "And the end of our exploring / Will be to arrive where we started / And know the place for the first time."

The unfolding that takes us from the beginning to the end has a triadic structure: not thesis-antithesis-synthesis, but thesis, negation, negation of negation. The third is thus a return to the first, but at a higher level or more explicit and determinate form. Hence each triad is also a circle of departure and return. "The whole of philosophy in this way resembles a circle of circles" (Lesser Logic, p. 24; cf. 25, 28, 29, 229, 266f, 272, 273, 310, 323).

In the Greater Logic Hegel includes a chapter entitled "With What Must the Science Begin?" (67-78) in which he covers these and some other points (see #7 below). It is significant that this chapter is not part of the preface or introduction, but is part of Book One itself. Self-conscious that he is talking about beginning after beginning, and perhaps remembering how he criticized Kant for talking about beginning before beginning, he says at the end of chapter: his insight about how to begin "requires no preparation or further introduction; and, indeed, these preliminary, external reflections about it were not so much intended to lead up to it as rather to eliminate all preliminaries" (78).

(On Hegel's method of beginning see Kierkegaard, Concluding Unscientific Postscript, at pp. 16-17, 101-07.)


5.  Begin anywhere

Many positions permit themselves to begin anywhere. Two variations will be considered here: to search for the logical beginning from wherever one finds oneself, and to do one's best without a certified first principle.

5.1.  Begin anywhere; regress to the logical beginning

Socrates began by talking to cobblers and sought first principles in the ensuing dialogues. He did not remain silent until he had found his logical beginning. This attractive path may not be a solution to the problem of beginning, however. It is not a way to justify the logical beginning or the system built on it; it seems to be a way to avoid the necessity of doing so at first, hence merely postponing the problem until a discovery is made.

However, if one admits that the discovery of the logical beginning does not come first in the discovert, temporal, or expositional series, then one may begin these latter series wherever one finds oneself in media res. Once begun on these series one may practice philosophy by looking for the logical beginning rather than from it. One does not know at this point whether one will regress infinitely or not. One may hope not. At least one is committed to regressing indefinitely until a satisfactory logical beginning is found.

Not many philosophers admit they are on this path, for by the time they communicate with us they have usually found something which satisfies them. But stray sentences in many works suggest this position. For example, this from Marx and Engels:

The removal of these difficulties is governed by premises...which only the study of the actual life-process and the activity of the individuals in each epoch will make evident. (German Ideology, p. 43)

That is, the premises of one kind of inquiry are discovered on the way, not beforehand.

Spinoza was on this path to the extent that he held we could begin with any true idea (Improvement of Human Understanding, p. 17.) This, however, was probably his discovert beginning. When he discusses the logical beginning, Spinoza is consistent in holding that the beginning must be one particular true idea, that of god (see #9).

Pascal urges us to recognize the necessity of this method when he says "You are already embarked" (Pensèes, No. 233, p. 66). "The last thing one settles in writing a book is what one should put in first" (No. 19, p. 7).

If philosophy is the 'search for presuppositions', then it seems to fall in this category. The problem of beginning will tighten its grip only when presuppositions are found that content the inquirer enough to build on them, or when building starts even without conscious reflection on presuppositions.


5.2.  Begin Anywhere: Render coherent what you find

For Nicholas Rescher, philosophy begins with pre-philosophical beliefs. These may originate in custom, tradition, authority, 'common sense', or science (Rescher, p. 18). These beliefs may be false or incomplete, but that is all we have to start with. Moreover, not all the attractive positions we inherit in this way are consistent with each other. We make the turn to philosophy when we notice this and try to render our many positions consistent (pp. 20f).

Rescher gives a most interesting account of the options open to philosophers in rendering their positions consistent. A point minor for him, but major for us, is that the result is merely the smelting and systematization of our inherited beliefs. These beliefs may still be false or incomplete, but now they are consistent with each other. This is attractive only for those who have already abandoned the correspondence model of truth for coherence. Adequacy lies not in representing reality, but in plausiblity and consistency. Rescher himself points out that there are many possible sets of mutually consistent beliefs. Each is a sort of island inconsistent with the others. While he uses this account to explain the inevitability of philosophical disagreement, we see in it a short way with the problem of beginning. Rescher's "orientational pluralism" is a confession that one's island of consistent beliefs is not better grounded than another. The demand for proof beyond coherence is senseless.

Even if Rescher may be right about this he faces the same problem as the fideist: he is not solving the problem of beginning so much as explaining why he is content not to solve it.

It is important, then, that on this method one does not really 'begin anywhere'. Rather, one begins wherever one finds oneself. The advantage of the method is that there is no problem in beginning, since everyone is somewhere when they decide to refine or educate their beliefs. The disadvantage is that one may have begun with inadequate beliefs; rendering a subset of them consistent by purging the rest may add the virtue of coherence but leaves unpurged any seed of doubt that may have infected the original beliefs.


5.3.  Begin Anywhere: Reject certainty, retain corrigibility

One would not start down an indefinite regress in search of a logical beginning unless one thought it might be attainable and would be worth attaining. Those who permit themselves to start anywhere and don't bother to regress in search of foundations will deny either that certified foundations are attainable or necessary —or both.

Critical realism in the philosophy of science is a good example of this kind of position. Certainty is either not attainable, or it is not attainable at first. One should start with plausibilities supported by experience, and revise one's position and its premises in light of new experience. One's hope lies not in a proof of the first premise, but in the corrigibility of the position.

Corrigibility is the essential alternative to certainty, but it is not sufficient. One must also have a method by which to interpret experience and revise one's position. And of course one's method and methodology must be corrigible.

The only obvious weakness of this position is that, like some fideist positions, it does not solve the problem of beginning but lives with it unsolved. It capitulates to the skeptic who finds no certainty. It lowers aim, and declares (or elects) contentment with generally plausible and reliable ideas, especially if they are subject to revision. But it cannot tax skeptics for stubbornness in refusing to lower their aim as well; for on the whole skeptics do not deny that many ideas are adequate to many human ends; they only fail to find certainty.


6.  Begin with the presuppositionless

This is the stated objective of many philosophers who face the problem of beginning explicitly.

The idea of presuppositionlessness, however, is either ambiguous or mystical. It could refer to an idea that has no logical antecedents, that cannot be derived from any other ideas. Or it may refer to some other kind of primacy. That is the ambiguity. The mystery is that the first kind cannot exist and the alternatives are not easily conceived. Every idea (proposition) can be derived from other ideas (propositions). For example, call the idea x, and plug it into this derivation:

If Socrates is mortal, then x.
All humans are mortal.
Socrates is human.
Therefore, x.

The triviality of this example is just the point. Premises can be found that suffice for any given rules of inference to derive any conclusion. If a presuppositionless idea is one that does not need logical antecedents, then we are back to self-evidence or self-justification.

In any case note that a presuppositionless logical beginning must be knowably presuppositionless to answer the doubts of inquirers and in that sense to solve the problem of beginning. But we must also know this presuppositionlessness without any paradoxical pre-beginning argumentation. Again, we seem to be thrown back to self-evidence or self-justification.


6.2.  Begin with something negative

This is the most significant variation on presuppositionlessness. If one's logical beginning is not an assertion, then the demand for proof is inapplicable. There are several ways to avoid assertion in the first step (see #8 and #10 below). This one is to start destructively, not constructively. One hopes, of course, that a dialectic will save the system and turn destruction into construction at some point without the help of dishonest manipulation.

The two classical cases of this type are Cartesian doubt and Kantian criticism.

[More later]


7.  Begin with error, not truth

The problem of beginning only haunts positions that want to begin with truth; only those positions must decide between proving the beginning or letting it go without proof. Hegel neatly sidesteps this threat by consciously beginning with untruth.

In the Phenomenology Hegel starts with the most naive or inadequate and untruthful standpoint of spirit, sense-certainty. The logical structure of the book is then to show how spirit educates itself and steadily improves the adequacy of its standpoints until it attains absolute knowing. Because the logical structure is genetic or developmental in this way, he must start with the point furthest removed from the point of highest adequacy, just as humans develop from maximal immaturity to maturity. (See Phenomenology, p. 50.)

In his two Logics, he begins with the least adequate, most abstract notions of being, then shows how they give rise to more and more adequate notions until the absolute idea is attained. So while the Phenomenology begins with sense-certainty that knows only that objects are, or their being, the Logics begin with sheer being, being without predicates or properties, being so abstracted from determinate beings that it is equivalent to nothing.

In the Lesser Logic he is explicit about the need to begin with error.

Why then, it may be asked, begin with the false and not at once with the true? To which we answer that truth, to deserve the name, must authenticate its own truth... (Lesser Logic, Sec. 83.z, p. 155)

[W]here knowledge by thought is our aim, we cannot begin with the truth, because the truth, when it forms the beginning, must rest on mere assertion. The truth when it is thought must as such verify itself to thought. (ibid., Sec. 159, p. 285; cf. pp. 352, 358)

In both the Phenomenology and the Logics Hegel is careful to make this logical beginning coincide with the expositional beginning, if we don't count prefaces and introductions. The narratives of each of his dialectical journeys begin expositionally as well as logically with the most erroneous state of consciousness and develop into the most adequate. Scholars disagree on whether Hegel also intended these beginnings to coincide with the causal beginning, that is, whether he held that the world was created by dialectical development out of the least real. What is less controversial is that Hegel intended the logical and expositional beginnings to coincide with the temporal beginning at least of the history of explicit thought in philosophy. The Phenomenology is much like a universal history of philosophy that is clearly intended to be mapped against the actual unfolding of figures and books. The case is similar with the Logics:

It is sufficient to mention here, that logic begins where the proper history of philosophy begins. Philosophy began in the Eleatic school, especially with Parmenides. Parmenides, who conceives the absolute as Being, says that 'Being alone is and Nothing is not.' Such was the true starting point of philosophy... (ibid., Sec. 86.z, p. 160)

Fichte is here too. He describes a quasi-dialectic by which the unenlightened dogmatist can be diagnosed as immature and the idealist as mature. He does not describe the transition, but he does recognize that "it is the business of philosophy to demonstrate and explain genetically how the self comes to think of itself in this fashion" (Fichte, p. 66).

One problem with this method of beginning is that there are several kinds or at least several degrees of untruth for Hegel. If we may begin with any old untruth, much as Socrates began with the ignorant, then no special argument is needed. But Hegel wants to begin with the maximal untruth. This, however, seems to require some argument. Why is his candidate really the least adequate standpoint? If Hegel argues, then he begins before he begins, and if he does not, then he has left skeptical questions unanswered.

It is worth a second to sketch Hegel's likely answer to this objection. He takes pains in the Introduction to the Phenomenology to show that philosophical science cannot pause at the beginning to answer skeptical objections. If it does, then it will not begin and the skeptic will 'win' even if beginning could have produced true and demonstrable science to the discomfiture of all skepticism (Phenomenology, pp. 47, 48, 51- 52). The validity of the beginning is proved in the end, by the return to it in the absolutely adequate end and by the absolute adequacy of that result.

Apart from Fichte and Hegel, this category should include any position that "kicks down the ladder" after climbing it. The 'ladder' climbed by writer and reader must not only be inadequate, but incommensurable with the adequate result; otherwise climbing it would suffice and kicking it down would be unnecessary. (Hegel by contrast insists that the ladder of inadequate positions climbed by spirit in its self-education must be preserved in the result.) Wittgenstein in the Tractatus is not the only "kick down the ladder" position, although his is the most famous. Without much distortion one could find logically similar positions in Sextus Empiricus, Kierkegaard, and Emerson.


8.  Begin with something non-cognitive, not knowledge

In #3.2 we looked at beginning in faith. If one begins without knowledge one may avoid the problem of beginning. If one finds another non-cognitive point of departure, one may perhaps avoid the problems of beginning with faith.

Duns Scotus began with love. Love prepares the way for faith, which in turn prepares the way for knowledge. [More later]

If this works, then starting with other 'passions' may also work. For example, here we should include Plato's claim that philosophy begins with wonder. This is usually read as a claim about the discovert or temporal beginning, but by analogy with Duns Scotus, one may convert it to a logical beginning to address the problem of beginning.


9.  Begin with the idea of the first cause

This is Spinoza's way. The logical beginning is not the causal beginning but the idea of the causal beginning. The idea of the first cause is the idea of god. Hence we must begin with the idea of god (Improvement of Human Understanding, pp. 14, 15- 16, 18.) "[A]s soon as it is possible and rational, we should inquire whether there be any being (and if so, what being), that is the cause of all things, so that its essence, represented in thought, may be the cause of all our ideas" (ibid., p. 36). "The nature of God...should be reflected on first, inasmuch as it is prior both in the order of knowledge and the order of nature..." (Ethics, p. 90).

Behind this is Spinoza's belief that the order and connection of ideas is the same as the order and connection of causes (Improvement of Human Understanding, 15-16). As Hegel will say later, Spinoza is saying that the order by which the system grows from its logical beginning is the same as the order by which the world grows from its first cause. This follows from and reinforces the propriety of beginning with the idea of the first cause. Hence, this method of beginning places as much stress on the order after beginning as on the beginning itself.


10.  Begin with realities, not ideas of realities

Only ideas (principles, thoughts, theories, theses) need proof or raise suspicions when asserted without proof. Hence one can evade the problem of beginning by beginning with the ideata of ideas rather than with ideas themselves.

Not many philosophers mean this literally when they speak this way. When some speak of realities they may mean merely that one should begin with certifiable truths, or concrete truths, or specific truths, or empirical truths (all still ideas).

Feuerbach speaks this way, and is not so clear that we can say that he meant what he said.

The new and only positive philosophy is the negation of all scholastic philosophy...the negation of philosophy as an abstract, particular, i.e., scholastic, quality; it has no shibboleth, no particular language, no particular name, no particular principle; it is the thinking man himself...

("Preliminary Theses", p. 169).

All speculation concerning right, will, freedom, and personality without regard to man; i.e., outside of or even above man, is speculation without unity, necessity, substance, ground, and reality. (ibid., p. 172; cf. 160, 164, 170).

Marx and Engels are much clearer and more explicit in The German Ideology:

The premises from which we begin are not arbitrary ones, not dogmas, but real premises from which abstraction can only be made in the imagination. They are the real individuals, their activity and the material conditions of their life, both those which they find already existing and those produces by their activity. These premises can thus be verified in a purely empirical way. (German Ideology, pp. 36-37)

In direct contrast to German philosophy [idealism] which descends from heaven to earth, here it is a matter of ascending from earth to heaven. That is to say, not of setting out from what men say, imagine, conceive, nor from men as narrated, thought of, imagined, or conceived, in order to arrive at men in the flesh; but setting out from real, active men, and on the basis of their real life-process.... (ibid., p. 42; cf. 43, 46)

One problem with this method of beginning, if we take it literally, is how to use the ideatum as a logical beginning? How can a reality, as opposed to an idea of reality, ground a theory? Not only does this look impossible (a category mistake), but hasty and question-begging as well. How do we know what is real without some inquiry first? The only answer I can see for Marx and Engels solves the category-mistake problem but not the question-begging problem. They are not using a reality to found a mere theory; a new vision of philosophy is emerging here. Philosophy is not just the thought of the world, but action within the world.


11.  Don't begin

This position is the default of not solving the problem. Hence it is not a solution, even if it is good philosophy. It is one way to read the position of the Greek Pyrrhonian skeptics.

In one sense, anyone who does not build a system with a logical beginning has avoided beginning in a saving way. But perhaps it is fair to chase these people down and look for an unstated logical beginning beneath their unsystematic beliefs. If the problem of beginning reaches out for this kind of non-philosophical believer (as well as for the unsystematic philosopher), then the only genuine way to avoid beginning in the saving way is to avoid belief, assertion, or commitment. This is exactly the position of the Pyrrhonian skeptics.

It is not equivalent to the position of Jains who starve to death silently after enlightenment. When Pyrrhonians talk to dogmatic (non-skeptical) philosophers in their pursuit of truth they are speaking ex concessis, with the premises and standards of their opponents. If skeptics are allowed to speak in this and other ways (which I have argued elsewhere), then they may begin expositional, discovert, and of course temporal and causal series, just not the logical series.


12.  Reject the problem

There are many ways to deny the applicability of the problem. Here are a few, in outline only. [More later.]

One difficulty with all these attempts to evade the problem is that either they will just say (or assume without saying) that the problem is inapplicable to a certain kind of philosophy, or they will argue it. If they just say (or assume) it, then they give us no reason to accept the conclusion. If they argue it, then they appeal to premises which are either proved or left unproved.

And we begin again.


Works Cited

Albert, Hans. Treatise on Critical Reason, trans. Mary Varney Rorty. Princeton University Press, 1985.

Aristotle. Metaphysics, trans. Apostle. Indiana University Press, 1966.

Descartes, Rene. Philosophical Essays, trans. Laurence Lafleur. Bobbs-Merrill, Library of Liberal Arts, 1964.

Feuerbach, Ludwig, "On [J.F. Reiff's] 'The Beginning of Philosophy,'" pp. 135-44; "Preliminary Theses on the Reform of Philosophy," pp. 152-173; in The Fiery Brook: Selected Writings of Ludwig Feuerbach, ed. and trans. Zawar Hanfi. Anchor Books, 1972.

Fichte, J.G. The Science of Knowledge (Wissenschaftslehre), trans. Heath and Lachs. Appleton, Century, Crofts, 1970.

Hegel, G.W.F. Hegel's Science of Logic, trans. A.V. Miller. George Allen and Unwin, 1969. Cited in the text as the Greater Logic.

Hegel, G.W.F. The Logic of Hegel, trans. William Wallace. Oxford University Press, 1892. Cited in the text as the Lesser Logic.

Hegel, G.W.F. The Phenomenology of Spirit, trans. A.V. Miller. Oxford University Press, 1977.

Kant, Immanuel. Critique of Practical Reason, trans. Lewis White Beck. Bobbs-Merrill, 1956.

Kant, Immanuel. Critique of Pure Reason, trans. Norman Kemp Smith. Macmillan, second impression, 1933.

Kant, Immanuel, "Open Letter on Fichte's Wissenschaftslehre," August 7, 1799. Public Declarations No. 6, Prussian Academy edition of Kant's Werke, XII.370-71.

Kierkegaard, Sren. Concluding Unscientific Postscript, trans. Swenson and Lowrie. Princeton University Press, 1968.

Kierkegaard, Sren. Philosophical Fragments, trans. Hong and Hong. Princeton University Press, 1985.

Kline, Morris. Mathematics: The Loss of Certainty. Oxford University Press, 1980.

Marx, Karl, and Frederick Engels. The German Ideology, trans. anon. Unabridged ed., Progress Publishers, 1976.

Moleski, Martin X., S.J. "Retortion: the Method and Metaphysics of Gaston Isaye," International Philosophical Quarterly, XVII, 1 (March 1977) 59-83. Includes a bibliograpy of Isaye's writings.

Pascal, Blaise. Pensèes, trans. W.F. Trotter. E.P. Dutton, 1958.

Rescher, Nicholas. The Strife of Systems: An Essay on the Grounds and Implications of Philosophical Diversity. University of Pittsburgh Press, 1985.

Russell, Bertrand. An Essay on the Foundation of Geometry. Dover Publications, 1956.

Sextus Empiricus. Outlines of Pyrrhonism, trans. R.G. Bury. Harvard University Press, Loeb Classical Library, 1933.

Spinoza, Baruch. On the Improvement of Human Understanding and The Ethics, trans. R.H.M. Elwes. Dover Publications, 1955.

Suber, Peter, "A Case Study in Ad Hominem Arguments: Fichte's Science of Knowledge," Philosophy and Rhetoric, 23, 1 (1990) 12-42.

Suber, Peter, "Classical Skepticism"

Suber, Peter, "Logical Rudeness," in S.J. Bartlett and P. Suber (eds.), Self-Reference: Reflections on Reflexivity, Martinus Nijhoff, 1987, pp. 41-67.

Suber, Peter, "Question-Begging Under a Non-Foundational Model of Argument," Argumentation, 8 (1994) 241-50.

Wittgenstein, Ludwig. Tractatus Logico-Philosophicus, trans. Pears and McGuinness. Routledge, 1961

. Zweig, Arnulf (ed. and trans.). Kant: Philosophical Correspondence, 1759-99. University of Chicago Press, 1967.

[Blue Ribbon] Peter Suber, Department of Philosophy, Earlham College, Richmond, Indiana, 47374, U.S.A.
peters@earlham.edu. Copyright © 1986-2001, Peter Suber.