Formal Epistemology and Cartesian Skepticism

In Defense of Belief in the Natural World

Tomoji Shogenji

• Publisher: Routledge
• Year: 2017
• ISBN:
  9781351336543
• Edition: 1
• Topic: Philosophy
• Subtopic: Philosophy History & Theory

This book develops new techniques in formal epistemology and applies them to the challenge of Cartesian skepticism. It introduces two formats of epistemic evaluation that should be of interest to epistemologists and philosophers of science: the dual-component format, which evaluates a statement on the basis of its safety and informativeness, and the relative-divergence format, which evaluates a probabilistic model on the basis of its complexity and goodness of fit with data. Tomoji Shogenji shows that the former lends support to Cartesian skepticism, but the latter allows us to defeat Cartesian skepticism. Along the way, Shogenji addresses a number of related issues in epistemology and philosophy of science, including epistemic circularity, epistemic closure, and inductive skepticism.

1 Skepticism and the Method of Meliorative Epistemology

1.1 Meliorative Epistemology

This book examines the epistemic status of our belief in the natural world. The aim is to evaluate our basic view of the natural world; for example, that familiar objects such as rocks, trees and chairs exist in the natural world where we also reside, and we interact with these objects.¹ I am going to examine whether our basic view of the natural world withstands the challenge of Cartesian skepticism. The point of dispute is not occasional errors, but the possibility of a wholesale misconception of reality. The Cartesian skeptic questions our basic view of the natural world by introducing radical alternative scenarios, e.g. that all objects of our sense perception are illusions created by a supernatural being, or that they are entities in the virtual world to which we do not belong. So, to put our basic view of the natural world in negative terms, the objects of our sense perception are not illusions created by a supernatural being, not entities in the virtual world to which we do not belong, etc. I call it “the natural world hypothesis”.

In its original challenge (Descartes 1984) the skeptic introduces the possibility of a deceiver with an extraordinary power—we are actually bodiless souls but are deceived by a powerful demon into thinking that we live in a natural world.² For the purpose of exposition, however, I will use a popular variant discussed widely in contemporary philosophy that we are actually brains in vats (BIVs) kept alive and wired to a supercomputer that is programed to supply us with the kind of sensory input brains in skulls would receive from the natural environment. The BIV version does not deny the existence of a natural world—there are brains, vats, a supercomputer, etc.— but it denies that familiar natural objects exist in places where our sense perception locates them relative to us. What we take to be natural objects belong to the virtual world to which we do not belong.

A remark is in order on the extent of skepticism addressed in this book, especially on its impact on the scientific understanding of the natural world. Though Cartesian skepticism questions our basic view of the natural world, it does not bring back the discredited hypotheses of the past, such as the phlogiston theory and the geocentric theory. The Cartesian skeptic does not question the substance of currently accepted scientific theories, but only their interpretation. For example, the heliocentric theory may not be the correct theory of the natural world, but the correct theory of the virtual world to which we do not belong.

As stated already, I take this challenge seriously, and consider it an open question whether there is epistemic justification for accepting the natural word hypothesis. It is sometimes suggested that the cost of giving up the natural world hypothesis is too high, so that it is sensible to reject some element of the skeptic’s argument even if the argument appears compelling. It is not clear, however, what the high cost amounts to. There is a sense in which whether we accept the natural world hypothesis or the BIV hypothesis makes no difference to us because the two hypotheses are empirically equivalent—we will enjoy (or suffer from) the same sensory experience if we are BIVs instead of living in a natural environment. What is the cost there? Some people may mention the loss of familiarity and intuitiveness. Since we commonly take the natural world hypothesis for granted, any view that is inconsistent with it, such as the BIV hypothesis, is counterintuitive. But if the cost is the loss of familiarity and intuitiveness, its relevance to epistemic evaluation is questionable. Why are familiarity and intuitiveness important in epistemic evaluation?

Over the course of history, numerous theories that used to be familiar and intuitive were replaced by unfamiliar and counterintuitive theories based on new evidence and new conceptual tools. It may still be counterintuitive to some people that the earth is orbiting around the sun instead of the sun orbiting the earth, but the educated public have come to terms with the heliocentric theory. It is possible in a similar way that we may eventually come to terms with a counterintuitive theory with regard to the natural world as a whole. Indeed, our stance on the natural world hypothesis has already changed due in part to Cartesian skepticism. People with no exposure to philosophical reflection may still believe it is absolutely certain that the natural world hypothesis is correct. There is, however, a strong sense among epistemologists that we cannot rule out the BIV hypothesis with absolute certainty. This does not immediately lead to a surrender to skepticism. Most epistemologists no longer share the Cartesian aspiration for absolute certainty. Instead, they accept fallibilism and are content with the probabilistic evaluation of the hypothesis.

Given the precedence of the retreat to fallibilism, it is conceivable that the challenge of Cartesian skepticism pushes us farther away from the familiar and intuitive view to the reluctant concession that the natural world hypothesis is indefensible even probabilistically. As we will see in Chapter 5, a strong Bayesian case can be made for accepting Cartesian skepticism and retreating to the goal of epistemic evaluation up to empirical equivalence. In the investigation that follows, the import of philosophical skepticism is its possible role in the revision of our epistemic practice, including our basic view of reality and the prevailing format of epistemic evaluation. The pre-theoretical judgment to the contrary is not good reason to resist revision.

The basic question of methodology arises at this point: What is the basis of revising or reaffirming our epistemic practice if we cannot turn to our pre-theoretical judgments? My answer is meliorative epistemology, which is broadly a Cartesian approach to epistemology.³ Cartesian epistemology is not grounded in our pre-theoretical judgments, but is meant to guide them. As Descartes put it, “The aim of our studies should be to direct the mind with a view to forming true and sound judgments about whatever comes before it” (Descartes 1985a, p. 9). More generally, in meliorative epistemology epistemic evaluation is expected to guide our epistemic practice to better results, where the better results are understood alethically. When our cognitive state is qualitative—to accept a statement, reject it, or withhold judgement—the alethic goal is to increase true beliefs and avoid false beliefs. When our cognitive state is quantitative—to assign a probability to a statement—the alethic goal is to make the probability as close to the true value as possible.⁴ This is all done under the resource constraints, i.e. we pursue the alethic goals with the epistemic resources available to us, both empirical and conceptual. The format of epistemic evaluation should be determined so that it will best serve the alethic goals under the resource constraints.

The project of meliorative epistemology is similar to a project in engineering. Engineers seek the best design for achieving certain goals under the given constraints. Similarly, meliorative epistemologists seek the best format of epistemic evaluation to achieve the alethic goals under the constraints of epistemic resources available. We may call it conceptual engineering.⁵ As in any engineering project, the constraints can change over time. As new empirical evidence emerges or new conceptual tools are devised, the best format of epistemic evaluation may also change. The alethic goals are not set in stone either. We may aim at a more ambitious goal as the epistemic resources expand, but it is also possible that we need to give up the original goal in favor of a less ambitious goal if new evidence or new conceptual tools reveal that the original goal is unachievable. That is one of the open questions addressed in this book.

1.2 The Role of Intuition

I adhere to the policy of not relying on our pre-theoretical judgments in adjudicating competing theories of meliorative epistemology. Since the policy is not widely shared among contemporary epistemologists, I want to address likely objections to the policy in this section.⁶ Let us see first how an argument from pre-theoretical judgment typically works. In his celebrated article (Gettier 1963), Gettier challenged the traditional theory of knowledge, according to which S knows that p if and only if S has a justified true belief that p. Gettier used two imaginary cases (thought experiments) to undermine the theory. In one of them (the ten coins case) the description of the circumstance makes it clear that Smith—an applicant for a certain job—has a justified true belief that the man who will get the job has ten coins in his pocket, but the description also makes it clear that Smith does not know this proposition.⁷ The traditional theory of knowledge is thereby rejected. The key in this argument is the judgment that Smith does not know the proposition in question, and this judgment is not formed by conscious inference from the information given. The argument is supported by a pre-theoretical judgment, or “intuition” as it is called in philosophical discussions. The use of intuition in thought experiments is not confined to epistemology. The trolley case in ethics (Foot 1967) and the Gödel-Schmidt case in philosophy of language (Kripke 1980) are often mentioned in the discussion of intuition in philosophy.

It is usually conceded by its advocates that intuition as evidence is defeasible. It can be overridden by other intuitions, other types of evidence, or by some theoretical considerations. According to this view, pre-theoretical judgments are not non-negotiable data for the analysis. However, some people may think I have gone too far in rejecting any use of intuition in meliorative epistemology. Indeed it is doubtful that we can avoid the use of pre-theoretical judgments completely. How can we start any epistemological investigation, or any investigation for that matter, without relying on some pre-theoretical judgments? Even scientists in search of an explanation or engineers seeking the best design routinely rely on their intuitive judgments, and no logical inference from consciously identified information can replace them because intuitive judgments have different roots, most likely in the subpersonal network of associations.⁸ Of course, the kind of intuitive judgments that scientists and engineers make are informed by their rich past experience as researchers and are often abstract in nature, while intuitive judgments in typical thought experiments are non-technical and concrete. But they are still of the same kind in the sense that they are, presumably, both rooted in the subpersonal network of associations, and not derived by explicit inference from consciously identified information.

It may be thought that I am pragmatically incoherent because I must have relied on my intuition in some way in the process of constructing an argument against the use of intuition in meliorative epistemology. We may call this line of argument “the self-defeat argument”^.9 In response to the self-defeat argument, I want to call attention to the distinction between the context of discovery and the context of justification.¹⁰ I grant that we need intuition to conduct any investigation. Scientists in search of an explanation often rely on their intuition in generating new hypotheses or recognizing their potential shortcomings. Engineers seeking the best design also rely on their intuition to come up with new ideas or anticipate problems to overcome. Investigators in any field consult their intuition, among many other factors, to decide on the direction of their research. However, this is all in the context of discovery. Scientists and engineers do not (and should not) cite their intuition in the context of justification, e.g. when the scientists defend their theory in a scientific article or when the engineers explain the merit of their design to their client. It is inappropriate to say “I have the intuition that this is correct” in the context of justification.

The same is true in meliorative epistemology. In order to generate hypotheses about good epistemic practice or to detect their potential problems, meliorative epistemologists may rely on their intuition. They may also consult their intuition, among many other factors, to decide on the direction of their research. However, this is all in the context of discovery. It is as inappropriate for them to cite their intuition in the context of justification as it is for the scientists or the engineers to do so. This is also true of objections to some theory. The opponents of a theory can certainly rely on their intuition in the context of discovery, e.g. in their search for points against the proposed theory. However, it is inappropriate for them to cite their intuition as the ground of their objection.¹¹ The arguments in meliorative epistemology, both for and against a theory, must be based on the alethic goals of our epistemic practice and the constraints of our epistemic resources.

Some defenders of intuition may point out that we would have no good reason to use intuition even in the context of discovery unless it is genuine evidence, though perhaps weak and easily defeasible evidence. Why should we use it in generating hypotheses and deciding on the direction of research if intuition were not even weak and easily defeasible evidence?¹² Though scientists and engineers avoid citing intuition in their formal arguments and presentations, its weakness as evidence and easy defeasibility explain their reluctance, viz. formal arguments and presentations call for evidence that is strong and not easily defeasible. It does not mean that intuition has no evidential value at all. In philosophy, where strong evidence is hard to come by and we do not expect to remove all reasonable doubt, intuition is admissible even in the context of justification, it may be suggested.

I am willing to grant that intuition is not totally unreliable in the sense of being a random occurrence. However, it is impossible to measure the reliability of intuition in a meaningful way.¹³ The reliability of intuition depends on the amount and the quality of the relevant experience the particular individual has on the particular subject. We cannot extrapolate from the past successful or unsuccessful use of intuition by other people on different subjects. There may be some philosophers who think that their discipline is in such a desperate state that even evidence from a source to which no meaningful degree of reliability can be assigned is admissible in the context of justification. I disagree, but they are free to pursue their project as they deem fit provided they make it clear that their evidence comes from a shaky source, and their conclusion is speculative. In this book I will consult intuition only in the context of discovery, and discuss it only when I introduce an idea informally and explain the motivation. In the context of justification, I will disregard intuition whether it is for or against the proposed theory. I will often say that intuition plays no role in meliorative epistemology, ignoring its role in the context of discovery, because it is true of any intellectual project that intuition plays a role in the context of discovery.

1.3 Meliorative Epistemology vs. Epistemography

Intuition plays no role in the evaluation of a theory in this book, but this is not a polemical stance I take in the sense that I reject other approaches in epistemology. It may shed light on our epistemic practice to formulate general principles that govern our pre-theoretical judgments of the epistemic kind. If that is the project, which I call epistemography, then the formulated general principles should accommodate our pre-theoretical judgments. It is therefore perfectly fine for an epistemographer to cite our pre-theoretical judgments in adjudicating competing hypotheses. There may also be other valuable projects in epistemology such as deontological epistemology to determine our obligation as responsible epistemic subjects. I have no objection to any such projects, provided the advocates of each project are clear and upfront about the nature of their project—what they try to accomplish and what are the grounds for adjudicating competing hypotheses.

The worst approach is to mix together projects with different goals and grounds. For example, we should not adjust a format of epistemic evaluation in meliorative epistemology to accommodate our pre-theoretical judgments; we should not disregard pre-theoretical judgments in epistemography to make them consistent with some principles of epistemic evaluation. Some epistemologists defend adjustments of this kind in the name of reflective equilibrium, but it is unclear what project in epistemology is served by the method. When successfully executed, the method of reflective equilibrium may produce a coherent set of principles that accommodates most of our pre-theo...

Table of contents

• Cover
• Title
• Copyright
• Dedication
• Contents
• Acknowledgments
• Introduction
• 1 Skepticism and the Method of Meliorative Epistemology
• 2 Truth as Correspondence
• 3 The Myth of Epistemic Circularity
• 4 Dual Components of Epistemic Evaluation
• 5 A Bayesian Case for Skepticism
• 6 Divergence from the Truth
• 7 Cartesian Skepticism Defeated
• References
• Index