It's all about the deep questions.
Epistemology Essay #1 A Conceptual Analysis of Knowledge
To be frank, I’m not a big fan of conceptual analysis. When we are participating in any type of conceptual analysis, what we are debating is how we use words by looking at how our intuitions square with definitions we decide to give our words. Debating about definitions seems almost like an oxymoron to me. The only point of language should be to facilitate discourse between two people. If a word is giving you a lot of trouble, then use a different word! I’m not saying conceptual analysis has no use however. If a word comes up in a field of study again and again (how knowledge does in epistemology), then it would be wise to clarify what one means by the word. This is the point of the present essay.
As we said before, the standard analysis is: S knows that p = (1) S believes that p (2) p is true, and (3) p is epistemically justified for S.
However, Gettier examples (see essay #0) give us cases where S justifiably believes the true proposition S, yet intuitively, S does not know p. The essay in the handbook goes through a lot of different revisions to what knowledge means (often ones that seem very contrived), but I will only present here 3 major contendors: the no false grounds approach, the defeasibility approach, and the causal approach.
No false grounds approach
S knows that p = (1) S believes that p (2) p is true (3) p is epistemically justified for S, and (4) S’s grounds for believing that p do not include any false propositions.
It gives the right results for various Gettier cases. Consider the example made in the introduction. Smith grounds for believing the proposition “John owns a Ford or Brown lives in Barcelona” include the false proposition “John owns a Ford”. Therefore, Smith does not know that John owns a Ford or Brown lives in Barcelona. It gives the right result.
First, it is too weak to rule out some Gettier cases. Suppose that Smith is justified in believing (j) Jones, who works in my office, has always given a Ford in the past, has just offered me a ride in a Ford, and says that he owns a Ford.
From (j), Smith deduces (k) There is someone, who works in my office, who has always driven a Ford in the past, has just offered me a ride in a Ford, and says that he owns a Ford.
On the basis of (k), Smith believes: (l) Someone in my office owns a Ford.
Now, suppose that Jones had only been pretending to own a Ford, but that someone else in Smith’s office, Brown, does own a Ford. Under the no false grounds definition Smith knows (l) since he believes it, it is true, he is justified in believing it, and the grounds by which he believes it (j) and (k) are true. Yet, Smith only got (l) right by luck, so intuitively, he does not know (l)
Second, it is too strong and rules out cases of knowledge that do seem like they should be knowledge. Suppose you justifiably believe something true on many grounds, of which one turns out to be false. The no false grounds approach says that that shouldn’t be in an instance of knowledge, when it really does seem to be.
A better revised definition would be: S knows that p = (1) S believes that p (2) p is true (3) p is epistemically justified for S, and (4) S’s grounds for believing that p do not justify any false propositions for S.
This would give the right result for the previous 2 examples. However, suppose Smith knows (m) Jones has always driven a Ford in the past, has just offered me a ride in a Ford, and says he owns a Ford.
Surely, he would know (k) There is someone in my office who has always driven a Ford in the past, has just offered me a ride in a Ford, and says he owns a Ford. The problem is (m) justifies the false proposition that Jones owns a Ford.
S knows that p = (1) S believes that p (2) p is true (3) p is epistemically justified for S, and (4) there is no true proposition, q, such that if S were justified in believing q, then S would not be justified in believing that p.
Works in all of the counter-examples we have considered thus far.
It is too strong. Consider the following example:
Stan’s wife notices that he is awfully quiet and seems trouble about something. She asks if he is upset or angry with Tim, a co-worker. Suppose that Stan introspectively is justified in believing (n) I am not angry at Tom. But suppose that earlier in the day unbeknownst to Stan, Tom had gotten into a nasty fight with Stan’s mother. Suppose that (o) Tom grossly insulted Stan’s mother, is true. According to the defeasibility approach Stan does not know (n) because there is a fact (o) that would make him angry at Tom. That’s the wrong result.
The causal approach
S knows that p = The fact that p is causally connected in an appropriate way with S’s believing that p.
My first qualm about this definition is the vagueness of the word “appropriate”. Given any counter example, one could say, “oh but that is (not) causally connected in the appropriate way.” Even so, this theory has attracted some prominent defenders. One further objection is our knowledge of universals, such as “All men are mortal.”
This is by no means a complete survey of the first essay. The first essay even analyzes the seemingly innocent requirements of truth and belief for knowledge! It also distinguishes between propositional and non-propositional knowledge. Furthermore, there are various strands of thoughts within each of the approaches I have summarized, and hopefully one will eventually turn out to be successful. Another interesting approach, made by Timothy Williamson, tells us that “knowledge” is conceptually basic and cannot be given a further analysis. Perhaps this is the right “analysis”.