Reliabilism: An Externalist Theory of Knowledge

Hello! Welcome to our lesson on reliabilism.

In our last session, we mapped out the great divide in contemporary epistemology between internalism and externalism. We saw that internalists locate justification within the mind's accessible states, while externalists argue that justification depends on factors outside the mind, primarily a connection to the truth. We ended by introducing reliabilism as the quintessential externalist theory, born from the need to forge a sturdy link between a belief being justified and its being true.

Today, we will delve into the details of this influential theory. Your learning outcome is to: Describe reliabilism as a prominent externalist theory of knowledge.

We will explore its origins as a response to simpler theories, articulate its core principles, and critically examine the significant challenges it faces. This will give you a robust understanding of one of the most important developments in epistemology after the Gettier problem.

1. The Road to Reliabilism: From Cause to Process

Reliabilism, championed by philosopher Alvin Goldman, didn't appear in a vacuum. It was an evolution of his earlier Causal Theory of Knowledge, which was a direct attempt to solve the Gettier problem. The causal theory proposed that for a belief to count as knowledge, the fact that makes the belief true must be part of the causal chain that produces the belief.

This seems to solve Gettier cases: in those scenarios, the justification is causally disconnected from what makes the belief true. However, Goldman himself discovered a powerful counterexample that showed the causal condition, while perhaps necessary, was not sufficient.

In the Fake Barn County case, Henry's true belief that he's seeing a barn is caused by a real barn. The causal chain is intact. Yet, our intuition is that he doesn't know it's a barn. Why? Because he could have easily been wrong. His method of belief-formation—glancing from the road—is unreliable in this specific, deceptive environment.

This insight is the key: the problem isn't the specific causal chain, but the reliability of the general process used to form the belief. This is the conceptual leap from the Causal Theory to Process Reliabilism.

2. The Core of Reliabilism

Reliabilism replaces the narrow focus on a specific causal chain with a broader evaluation of the belief-forming process.

The core idea is elegantly simple:

A belief is justified if and only if it is the product of a reliable cognitive belief-forming process.

A process is reliable if it tends to produce a high ratio of true beliefs to false ones. Examples of generally reliable processes include vision, hearing, memory, and valid deductive reasoning. Unreliable processes include wishful thinking, hasty generalization, and reading tea leaves.

Let's get a more formal definition. Given your background in mathematics, you'll appreciate that Goldman and other reliabilists aim for a reductive analysis with precise principles.

As the article details, reliabilism is:

• Externalist: The reliability of a process is an objective fact about its success rate in the world. The believer does not need to know, or be justified in believing, that their process is reliable. This is what distinguishes it from internalist theories.
• Historical: For beliefs based on inference, the justification of the conclusion depends on the justification of the premises. This traces back through a history of belief-formation, which must ultimately terminate in reliable processes that don't have other beliefs as inputs (like perception).

3. Challenges to Reliabilism

Like any major philosophical theory, reliabilism faces significant objections. A full description requires understanding its weak points.

Let's break down these challenges:

a) The New Evil Demon Problem

We discussed this in the last lesson as an argument for internalism. For reliabilism, it's a major objection. Your brain-in-a-vat twin forms beliefs using perception, but in that world, perception is completely unreliable. Reliabilism is forced to conclude that your twin's beliefs are all unjustified. This clashes with the strong intuition that the twin is being perfectly rational and epistemically responsible.

b) The Generality Problem

This is perhaps the most persistent internal problem for reliabilism. Any single belief-forming event (a process token) can be described as an instance of many different general process types.

• The process that formed Henry's belief could be "vision," "vision in broad daylight," "identifying medium-sized objects from a car," or "identifying barns in Fake Barn County."
• These different process types have different levels of reliability. "Vision" is very reliable; "identifying barns in Fake Barn County" is very unreliable.
• Which process type is the correct one to assess for justification? Without a principled way to answer this, reliabilism seems arbitrary.

c) The Lottery Problem

This problem is particularly interesting as it brings probabilistic reasoning to the forefront, a connection you might find compelling.

• Suppose you hold one ticket in a 1,000,000-ticket lottery. The belief "My ticket will lose" is formed through a process that is 99.9999% reliable.
• According to a simple version of reliabilism, this belief should be justified. You might even argue it constitutes knowledge if the belief is true.
• However, most people have the strong intuition that you don't know your ticket will lose until the results are announced. Merely having high probability is not enough for knowledge.

This problem forces reliabilists to refine what "reliability" means. Is it just a high truth-ratio? Or is it something more complex?

Comesaña's analysis shows that the simple idea of "tends to produce true beliefs" can be formalized in different ways, with different consequences for hard cases like the lottery. Defining reliability via conditional probability ($\text{Pr}(p|e)$), as Alston suggested, provides a more sophisticated tool than a simple frequency count. This move demonstrates the kind of formal rigor that analytic philosophers, especially those interested in the foundations of math and logic, bring to epistemological problems.

Test your understanding!

Imagine a scientist using a very sensitive instrument to detect a rare particle. The instrument has two modes:

1. Screening Mode: It is extremely reliable at telling you when a particle is not present (99.9% accuracy).
2. Confirmation Mode: It runs a slower, more complex analysis that is very reliable at confirming a particle is present (99% accuracy).

A scientist checks the screening mode and it reads "No Particle Detected." Is her belief "There is no particle" justified, according to process reliabilism? Now, suppose she switches to confirmation mode to double-check a positive reading. Is her belief "There is a particle" justified? How does this illustrate the Generality Problem?

Show answer

According to process reliabilism, both beliefs would likely be considered justified because they are produced by highly reliable processes. The process "forming a belief based on the screening mode's negative reading" is 99.9% reliable, and the process "forming a belief based on the confirmation mode's positive reading" is 99% reliable. Both exceed any reasonable threshold for justification.

This illustrates the Generality Problem because we had to choose the relevant process type. Is the process "using this instrument," or is it "using the screening mode" or "using the confirmation mode"? In this case, specifying the mode seems natural. But in many other cases, like Henry in Fake Barn County, the choice is not so clear. The justification of the belief depends entirely on which process description we select as the relevant one.

Conclusion

Today we've taken a deep dive into process reliabilism, the leading externalist theory of justification. It offers an intuitive, truth-connected account of what makes a belief justified, moving beyond the purely internal perspective of its rivals.

Key Takeaways:

• Reliabilism emerged from the failure of the Causal Theory of Knowledge to handle cases like Fake Barn County.
• Its core thesis is that a belief is justified if and only if it is produced by a reliable belief-forming process—one that has a high truth-conduciveness.
• As an externalist theory, the believer does not need to have access to the facts that make their belief justified (i.e., the reliability of the process).
• Reliabilism faces serious challenges, including the New Evil Demon Problem (an objection from internalists), the Generality Problem (an internal consistency issue), and the Lottery Problem (a puzzle about high probability).
• Advanced versions of reliabilism use tools from probability theory to offer more nuanced definitions of "reliability," demonstrating the formal and rigorous nature of contemporary analytic epistemology.

Preview of the Next Lesson:

We've seen how reliabilism attempts to build a theory of justification that is resistant to Gettier-style problems by requiring a reliable connection to the truth. In our next lesson, we will circle back to the Gettier problem itself to examine the crucial element at its heart: epistemic luck. We will analyze different types of luck and how they help diagnose exactly what goes wrong in Gettier cases, providing us with a more refined toolkit for evaluating any theory of knowledge.