Another first principle is – That the natural faculties, by which we distinguish truth from error, are not fallacious. If any man should demand a proof of this, it is impossible to satisfy him. For, suppose it should be mathematically demonstrated, this would signify nothing in this case; because, to judge of a demonstration, a man must trust his faculties, and take for granted the very thing in question.

– Thomas Reid, An Essay on the Intellectual Powers of Man VI.V.

The Ohio State University

2000

Dissertation Committee:

Approved by

Professor George Pappas, Adviser

Professor Marshall Swain _____________________________

Adviser

Professor George Schumm Department of Philosophy

ABSTRACT

Epistemic circularity is, roughly stated, a property of an argument such that its conclusion must be true if one may be said to have a justified belief in its premises. An example is an argument for the general reliability of sense-perception that makes use of sensory beliefs among its premises; as William Alston has pointed out, epistemic circularity poses a difficult problem for defending the reliability of sense-perception. It is also a key element in for a related (and broader) meta-epistemological problem, dubbed here The Problem of Meta-Justification . First we pose a question: how can we ultimately justify our standards of justification? The difficulty can be neatly stated in the form of a Meta-Regress Argument similar to the classic regress argument for foundationalism. The options offered by the Meta-Regress Argument are: self-support meta-foundationalism, meta-coherentism, meta-regressism, strict particularism, strict methodism, and meta-skepticism.

One might attempt to defuse the threat of epistemic circularity by attempting to show it to be "virtuous," rather than vicious. But no one has adequately argued that epistemic circularity is indeed virtuous, and several arguments can be deployed showing it to be vicious. Meta-coherentists, drawing on insights related to the Method of Reflective Equilibrium, might try to find ways to mitigate the viciousness; but their attempts fail. Varieties of particularism and methodism, two positions on the Problem of the Criterion, might also be offered as a way to escape epistemic circularity; but these views too fall prey to serious objections.

The results of Chapters 1-3 of this dissertation, sketched above, appear to support meta-skepticism. It is possible, however, that there are some beliefs that are epistemically rational but nonjustified (i.e., neither justified nor unjustified). Such beliefs can support justification standards without themselves being justified. In this way, meta-skepticism can be avoided. This solution to the Problem of Meta-Justification is developed in Chapter 4 in a way that owes a heavy debt to the epistemology of the great Scots philosopher, Thomas Reid.

Dedicated to my mother

ACKNOWLEDGMENTS

I am grateful to my adviser, Professor George Pappas, for his help, criticism, and patience. I would also like to thank the other members of my dissertation committee, George Schumm and Marshall Swain. All three committee members made comments that led to substantial improvements. A few, but far from all, of these comments are noted in the text and footnotes. George Schumm also made numerous proofreading corrections, for which I am grateful.

I would also like to express my gratitude to some other philosophy department professors and to some graduate students (including Robert Kraut, Joe Salerno, and Ty Lightner), as well as members of an internet discussion group, the Association for Systematic Philosophy.

Hearty thanks to Jimmy Wales for allowing me time off work to finish (and for giving me an extra incentive to finish), and thanks to Jake Rogers for help proofreading a penultimate draft.

Thanks to Carolyn Cutler for encouraging me to teach fiddle for a living for a while; following that good advice gave me the time and money to finish.

Finally, I wish to thank my mother, Lana Sanger, and some other people (especially Margarita Zagirova), for encouraging me to finish.

VITA

July 16, 1968 Born Bellevue, Washington

1991 B.A. Philosophy, Reed College

1995 M.A. Philosophy, The Ohio State University

1992-6, 1997-8 Graduate Teaching Associate,

The Ohio State University

FIELDS OF STUDY

Major Field: Philosophy

I've agreed to include this dissertation as part of Enlightenment at the request of Carolyn Ray. I never intended to publish it in any form.

I did not write this as a work of Objectivist philosophy, and I do not regard myself as an Objectivist. But I'm sure that I share many views in common with Objectivists, and I believe Objectivists might find quite a bit to agree with in this dissertation.

As the text itself should make tolerably clear, the specific details of the solution to the Meta-Justification Problem that I propose in Chapter 4 are tentative at best, and probably some fairly dramatic revision thereof is closer to the truth. I feel much more confident of the conclusions of Chapters 1-3.

I hope that anyone proposing to write a reply to this will do me the favor of actually trying to understand what I've written. Very many philosophical disputes consist primarily of different generally well-meaning parties failing to understand each others' meanings. I do not wish to engage in any such disputes, however.

Larry Sanger, August 27th, 2000

INTRODUCTION

Ever eager to enforce a requirement that all philosophical claims be justified, philosophers are apt to run into a difficult predicament: that of requiring argument for the reliability of the most fundamental sources of our knowledge. Why this is a difficult predicament can be explained in short order, with a sort of example that will be familiar to most philosophers.

Suppose John is defending the possibility of perceptual knowledge against Dave's skeptical attacks. Dave brings up round towers appearing square at a distance, vivid dreams, evil deceiving demons, and brains in vats. The embattled John continually falls back on one sort of assertion: if it seems to him he perceives something, and the conditions are as good as he could ask for, then he is justified in believing what he perceives, despite strange doubts to the contrary. Dave finally decides to call John's bluff: "You keep asserting that your best perceptual appearance claims are veridical; all right, now give me an argument for your assertion. I don't have to keep bringing up different grounds for doubt; all I want you to do is to justify your assertion."

John agrees (no doubt due to his lack of experience with the likes of Dave). "Let's take the sense of sight as an example. Well, just yesterday I was walking home and I thought I saw my good friend Mary on a doorstep nearby. As I got closer I confirmed that it was Mary. And, once, I was wandering around in the biology department and saw something greenish through a glass doorway that looked like a frog. I went into the room and both smelled and heard, and got a good look at that frog. And of course these are only a few examples. The point is that over and over, it appears to us that we see, or hear, or smell something, and then later we come to a position where we can confirm that what we thought we perceived was indeed the case. This is how anyone knows that his best perceptual appearance claims are veridical. None of your skeptical doubts constitutes any good reason to doubt this."

Dave, having received an expected answer, springs his trap: "But look here. You used your senses to confirm what your senses told you; in order to confirm each perceptual appearance claim, you made one or more further perceptual claims. So you are begging the question. If you are permitted simply to assert that you'confirmed' (by your senses) that it was Mary in the doorway, then you are taking for granted, indeed, that your best perceptual appearance claims are veridical. But that's precisely the claim I asked you to support."

John's mode of arguing involves what has, within the last twenty years, come to be called epistemic circularity . On one rough formulation, an argument is epistemically circular when having a justified belief in the premises of the argument requires already having a justified belief in the conclusion. Thus the argument is not circular in the ordinary sense. But an appropriate response to an epistemically circular argument is always, "What is the point of offering this argument? If you think it is successful, you didn't have to present it at all."

This dissertation is a study of epistemic circularity, especially as it applies to arguments for fundamental standards of justified belief. The focus is on epistemic circularity per se, an abstract or reoccurring phenomenon that problems such as Cartesian skepticism and the problem of other minds have in common. In this regard my project might be compared to a study of regress arguments: such arguments can be found in epistemology, logic, philosophy of religion, philosophy of science, and no doubt other areas, but surely the fact that the argument type is widespread is no reason to think that its study need be unmanageably lengthy. Moreover, if the argument type is widespread, precisely that is an excellent reason to focus on it and for a time put aside the particular details of its particular varieties.

The discussion has been organized in terms of various competing solutions to a single basic problem, the Problem of Meta-Justification (PMJ ): very roughly, how are standards of justification themselves ultimately justified? As I will explain in Chapter 1 and elaborate in Chapters 2 and 3, answers can be organized, surprisingly neatly, as responses to a Meta-Regress Argument . Along the way various theories and methods merit discussion, including (among others) inductive justifications of induction, reflective equilibrium and related coherence methods of theory confirmation, particularism and methodism, and skepticism.

My own solution, developed after laying out criticisms of the competition, is distinctly Reidian: there are fundamental cognitive processes involving sense-perception, memory, reasoning, etc., that, we have no choice but to assume, give us mostly true beliefs. The practice of forming the most obviously true beliefs based on these most fundamental cognitive processes is known as (in an old sense of the term) reason. The core intuition behind my solution is this: it is silly to ask that we use reason to support the claims of reason's reliability, or to think that anything important is established thereby. As Reid wisely suggested (see the dissertation's frontispiece), we can take reason's reliability for granted, and there is nothing whatever wrong with such a move.

Issues surrounding the adjudication of these theories and concerning the PMJ have dominated the work of influential philosophers in this century at least as much as in prior centuries. But the PMJ has been, perhaps, of particular importance in recent years. W ith renewed interest in the epistemology of Thomas Reid, and attempts to get clear on the nature and merits of the Method of Reflective Equilibrium, not to mention the important work of William Alston on the topic of epistemic circularity itself, there is a need to formulate the PMJ clearly and discuss the constraints that may be placed on its solution. I hope this dissertation makes a positive contribution in that regard.