Goldman and His Critics

Brian P. McLaughlin, Hilary Kornblith

• Publisher: Wiley-Blackwell
• Year: 2016
• ISBN:
  9781118609170
• Edition: 1
• Topic: Philosophy
• Subtopic: Epistemology in Philosophy

Goldman and His Critics presents a series of original essays contributed by influential philosophers who critically examine Alvin Goldman's work, followed by Goldman's responses to each essay.

• Critiques Alvin Goldman's groundbreaking theories, writings, and ideas on a range of philosophical topics
• Features contributions from some of the most important and influential contemporary philosophers
• Covers Goldman's views on epistemology—both individual and social—in addition to cognitive science and metaphysics
• Pays special attention to Goldman's writings on philosophy of mind, including the evolution of his thoughts on Simulation-Theory (ST)

PART I
RELIABILISM, INTERNALISM, AND EXTERNALISM

1
INTERNALISM, RELIABILISM, AND DEONTOLOGY

MICHAEL WILLIAMS

1 Internalism and the Reliabilist Revolution

Since the 1960s, Anglophone epistemology has undergone a paradigm‐shift: “the Reliabilist Revolution.” The revolutionary‐in‐chief has been Alvin Goldman.

Reliabilism names a family of views. According to Goldman’s version, knowledge is true belief acquired and sustained by some reliable cognitive process or procedure: that is, a process or procedure that tends to produce true beliefs and to avoid producing false ones. Other versions avoid the reference to cognitive processes or procedures and equate knowledge with true belief that in some appropriate way counterfactually co‐varies with the facts: for example, had it not been the case that p, I would not have believed that p (“sensitivity”), or that in believing that p I could not easily have gone wrong (“safety”)ⁱ. I won’t be concerned with whether one approach is superior. Nor will I investigate the best way to formulate either.

Rather, my focus will be on a central element in all versions: the shift from an “internalist” to an “externalist” approach to understanding knowledge and justification. Goldman’s assessment of “internalism” is negative in the extreme. He argues not merely that his “externalist” understanding of knowledge is superior to any “internalist” rival but that internalism suffers from crippling defects: “fundamental problems that lie at the core of internalism.” Further, he claims to have challenged “the viability of [internalism’s] most prominent rationale” (Goldman 1999 [2002]: 3).ⁱⁱ He finds this putative rationale in what he calls the guidance‐deontological (“GD”) approach to justification, according to which justification depends on self‐consciously living up to one’s epistemic obligations. As Goldman notes, the GD approach has a long history. Arguably, it was the dominant approach prior to the Reliabilist Revolution: hardly surprising, given that in the pre‐revolutionary era virtually all epistemologists were internalists of one kind or another.ⁱⁱⁱ Indeed, there is a case for saying that the internalist/externalist contrast itself – or at least a clear awareness of its importance – is a product of the Revolution, though we would have to add that how the distinction is best understood is a matter of debate.

I agree that “internalism,” as Goldman understands it, should be repudiated. But as Goldman recognizes, the GD approach to understanding knowledge and justification does not, in and of itself, require us to adopt the kind of internalism he rejects. So I shall argue for two lemmas. The first is that Goldman has not adequately diagnosed the sources of the untenable internalism that is his principal target: additional commitments must be brought to light. The second is that dispensing with these commitments opens the way to an approach to knowledge and justification that is “internalist” by a standard that Goldman himself recognizes, yet free of the drawbacks he brings to our attention. My conclusion is that internalist justification needs refinement, not rejection. This means, in turn, that the GD conception of justification also survives.

2 Internalism’s burdens

According to Goldman, the distinction between an “externalist” and an “internalist” conception of justification turns on contrasting answers to the question of whether or not justification‐conferring properties must be accessible to the subject. He writes:

Externalism characteristically holds that beliefs acquire justificational status if they are produced by methods with certain “external” properties, properties that need not be known – and perhaps need not be knowable, or at any rate “directly” knowable – by the agent himself. Internalism takes issue with this claim about proper methods or pathways. It holds that all justification‐conferring properties (“justifiers,” for short) must be accessible to the agent. (PK: vii)

However, although we might reasonably take insistence on the “accessibility” of justifiers to be the essential feature of internalism, in doing so we would fall short of characterizing internalism as Goldman understands it. For Goldman, internalism is not just the view that justifiers be knowable. There is the further requirement that “only internal conditions qualify as legitimate determiners of justification,” so that “justification must be a purely internal affair” (PK: 3) “Internal” here means “internal to the subject”: justifiers, at least in the first instance, are mental states (or facts concerning such states). The “internalism” that Goldman repudiates is subjectivist internalism.

Goldman is surely right that subjectivist internalism has enjoyed a long run in epistemology; and he argues convincingly that its run deserves to come to an end. He advances three major conclusions:

1. Internalism leads to skepticism. In its strong form, internalism restricts legitimate justifiers so severely that beliefs that intuitively count as justified turn out to be unjustified. Attempts to liberalize internalism do not repair the damage.
2. Internalism incorporates an untenable methodology of epistemology. Internalism “standardly incorporates the doctrine that epistemology is a purely a priori or armchair enterprise rather than one that needs help from empirical science.” However, the hallmark of epistemic justification is truth‐conduciveness; and the truth‐conduciveness of the procedure of guiding one’s beliefs by internal states is an empirical matter. (We have learned that some of our natural ways of thinking are not always truth‐reliable.)
3. Internalism lacks a cogent rationale. The main (only?) rationale for internalism derives from the GD conception of justification. But in itself, the GD conception does not support internalism. Goldman calls this the “core dilemma” for the “Three Step Argument” for internalism (PK: 17–18).

There is something to each of these points, though not as much as Goldman claims.

On the first point, I agree that many traditional forms of internalism lead to skepticism. So I can deal briefly with Goldman’s arguments to that effect.

Internalism, as Goldman understands it, confines justifiers to facts concerning mental states. “Strong Internalism” (SI) is even more demanding, restricting justifiers to facts about occurrent, conscious mental states. Thus:

(SI) Only facts concerning what conscious states an agent is in at time t are justifiers of the agent’s beliefs at t. (PK: 8)

Strong Internalism is no straw man. As Goldman says, Chisholm – surely one of the most influential American epistemologists of the last century – says many things that commit him to such a view.^iv

Strong Internalism falls to the problem of stored beliefs. Most of one’s beliefs are stored in memory; and even when a belief is consciously entertained, it does not come accompanied by all beliefs relevant to its status as a justifier for other beliefs. Given Strong Internalism, then, almost none of our beliefs is justified. Internalists might reply that justification for a belief requires only a disposition to generate conscious evidential states as justifiers for consciously entertained beliefs. However, it is doubtful whether a Strong Internalist can accept this suggestion. The fact of possessing a disposition is not the sort of thing that can be known at an instant by introspection. Allowing dispositions to count as justifiers threatens to open the door to externalism.

Suppose that we liberalize internalism by expanding the class of justifiers to incorporate stored mental states: this Weak Internalism falls to the problem of forgotten evidence. We do not normally suppose that for a belief to remain justified I must be able to recall all the evidence or reasons that justified it at the time of acquisition, and imposing such a requirement would have extensive, skeptical implications. As Goldman says, a belief can be justified (and in the absence of countervailing considerations remain justified) by virtue of being acquired in an epistemically proper way. But “past acquisition is irrelevant by the lights of internalism,” strong or weak, since all past events are “external” to an agent’s current mental states and capacities (PK: 10).

Internalism has been defended in both foundationalist and coherentist versions. However, both versions make appeal to logical or probabilistic relations, either to connect basic with non‐basic beliefs or to contribute to the coherence of one’s overall belief system. Since such relations do not concern either occurrent or stored mental states, it is unclear how internalists are entitled to count them as justifiers.

A further liberalization seems called for: we must credit an agent with some computational capacities that allow him to determine whether a targeted proposition stands in appropriate logical or probabilistic relations to other relevant belief‐states. But now we face the problem of the doxastic decision interval. Suppose that one’s belief‐system contains 138 atomic propositions: just checking for consistency, using the truth table method and employing an ideal computer working at top speed, would take 20 billion years, which makes nonsense of the idea of being justified at a time. Furthermore, algorithms like the truth‐table method are not in the conscious repertoire of most human beings: does this mean that most people have no justified beliefs? No doubt we have sub‐personal computational skills and procedures that are reliable enough in ordinary circumstances. But the possession of such skills is precisely the kind of external fact that internalists exclude from justificational relevance. Not surprisingly: what skills we possess, and how reliable they are, are manifestly empirical matters.

Finally, there is the question of whether the class of justifiers should include epistemic principles. Goldman argues that it should: “epistemic principles are among the items that determine whether a belief is justified, which is just how ‘justifiers’ was defined” (PK: 16). But internalists have a special reason for so treating them. The rationale for internalism rests on the GD conception of justification, according to which justification depends on fulfilling one’s epistemic obligations; and an agent’s knowledge of her internal states “will not instruct her about her epistemic duties and entitlements unless she also knows true epistemic principles” (PK: 16). Unfortunately for internalists, it is doubtful whether ordinary agents...

Table of contents

• COVER
• TITLE PAGE
• TABLE OF CONTENTS
• LIST OF CONTRIBUTORS
• FOREWORD
• PREFACE
• PART I: RELIABILISM, INTERNALISM, AND EXTERNALISM
• PART II: EPISTEMOLOGICAL TROUBLE‐SHOOTING AND SOCIAL EPISTEMOLOGY
• PART III: COGNITIVE SCIENCE, PHILOSOPHY OF MIND, AND METAPHYSICS
• APPENDIX: ALVIN GOLDMAN’S PUBLICATIONS
• INDEX
• END USER LICENSE AGREEMENT