Philosophy, Film, Politics, Etc.

Sunday, July 11, 2010

Getting Past Gettier

Larry Nemirow once told me that I should try to publish my good ideas wherever I can, and not wait until somebody else gets credit for them first. This apparently happened to him several times when he was a grad student. But that was before the blog era. These days, it's a choice between waiting several months before getting a rejection letter (and, as a non-professional, I cannot expect anything other than rejection letters) or getting my ideas out in the open through my blog (or other online venues.) So that's what I'm gonna do now.

Below is a very short paper I wrote which was rejected by a reputable philosophy journal a couple months ago. (I'm not sure why, but it seems better not to say which journal.) The editor-in-chief gave me some instructive comments--journals don't always do that, and I'm truly grateful. I'd like to spend a month or so researching and rewriting the paper, but I don't know when I'll have time to do that. In the meantime, maybe some readers can give me a few suggestions or criticisms.

I'm posting the editor-in-chief's comment first, and then the paper:

Dear Dr. Streitfeld [I'm not a doctor, but it's nice of them to presume otherwise],

Thank you for submitting your manuscript "Getting past Gettier" to [journal]. While it contains some interesting observations, it seems to me that it does not contact an enormous amount of literature in linguistic semantics on very pertinent topics (for example, the analysis of definite descriptions with respect to de re and de dicto beliefs) which has been done since the references you cite. It therefore does not meet the standards for publication in [journal], but I would encourage you to contact some of that literature as you work further on this topic. Good luck in finding a suitable venue for your work.

I should also mention that I sent an earlier draft of this paper (which contained discussion of more recent versions of Gettier cases) to Stephen Hetherington, and he made a similar comment--specifically, that I should connect with Grice and Donnellan on definite descriptions. But I've looked into Grice and Donnellan, and it seems to me that any discussion of them would be too tangential. It would make the paper meatier, but wouldn't help me make my essential point. Though perhaps I've just missed something of relevance there.

Getting Past Gettier

It is commonplace to attribute beliefs with sentences of the form, “Jennifer believes that snow is white.” Philosophers are inclined to regard the subordinate clause as a sentence, which is often given a number:

  1. Snow is white.
Such sentences are said to express or represent the belief in question, and reference to sentences is not distinguished from reference to beliefs. The association between beliefs and sentences is ubiquitous and accounts for much confusion in philosophy. Unlike beliefs, sentences can neither be true nor false, but can be used to make true or false statements (Strawson, 1950). Confusion arises when philosophers regard sentences as either true or false whilst overlooking significant differences in their use.

I aim to show that just such confusion is responsible for the prevailing discourse concerning Gettier cases. It was not so long ago that Edmund Gettier made a stir by arguing that justified true belief is not sufficient for propositional knowledge, despite a long tradition of thinking otherwise (Gettier, 1963). It has since become commonplace for philosophers to discuss a variety of Gettier cases—purported examples of justified true beliefs which are not classifiable as propositional knowledge. It is commonly supposed that there is a problem with the traditional conceptualization of propositional knowledge, or that a better understanding of justification is in order. However, there is a stark lack of agreement over how to approach, let alone resolve, the dilemma. This is the Gettier problem.

I aim to dissolve the problem by arguing that the subjects in Gettier cases do not have justified true beliefs, but rather justified false beliefs. Confusion arises because the false beliefs are expressed with sentences which can also be used to express true beliefs—though not true beliefs had by the subjects in question. Only the philosophers discussing the cases have the true beliefs in question. By failing to observe the role of definite descriptions in these sentences, philosophers suppose that the sentences are true, because philosophers—having extra information—are naturally inclined to use the sentences to make true statements. By failing to distinguish between sentences and the beliefs they express, philosophers fail to distinguish between their own beliefs and the beliefs of the subjects in question. This accounts for the intuition to attribute justified true beliefs to the subjects in Gettier cases, and also explains the strong intuition to deny that the subjects have the relevant propositional knowledge—for it is obvious that the subjects in Gettier cases lack the information required to make the relevant true statements. I aim to demonstrate this through analysis of Gettier’s two original examples (Gettier, 1963). By accounting for philosophers' beliefs and intuitions, and revealing the source of confusion, I aim to preserve the conceptualization of propositional knowledge as justified true belief.

Gettier’s first example is thus: Smith is applying for a job but learns from the president of the company that Jones will get it instead. Smith knows that Jones has ten coins in his pocket, and so forms the belief:

2. The man who gets the job has ten coins in his pocket.

It turns out that Smith gets the job, not Jones. It is also the case that Smith has ten coins in his pocket. Gettier claims that (2) is true and Smith is justified in believing it. Yet, there is a strong and widespread intuition to deny that Smith knows that (2) is true. We are thus tempted to conclude that justified true belief is not sufficient for propositional knowledge.

Sentence (2) can be used to express any number of beliefs. When philosophers say (2) is true, they are expressing a belief about Smith, knowing that Smith will get the job. Yet, when Smith believes (2), he is thinking about Jones. Smith’s belief is false, though it is expressed with the same sentence philosophers use to express their true belief.

Gettier might resort to a more abstract formulation, stipulating that Smith believes:

3. There is some man, X, such that X will get the job and X has ten coins in his pocket.

It is thus supposed that Smith’s belief is not about Jones or anybody else. However, Smith does not believe that there is some set of men such that one unspecified member of that set will get the job and has tens coins in his pocket. Smith has no justification for such a belief. In so far as (3) represents Smith’s beliefs, the “ X” refers to Jones, and only Jones. Smith is not ambiguous in his belief about who will get the job.

In Gettier’s second example, Smith and Jones are joined by Brown, who is in Barcelona. Smith has no reason to believe that Brown is in Barcelona. Yet, Smith has just cause to believe that Jones owns a Ford. He thus forms a variety of disjunctive beliefs, including:

4. Jones owns a Ford or Brown is in Barcelona.

Smith believes (4) because he believes that Jones owns a Ford. However, Smith is mistaken. Jones does not own a Ford. Yet, (4) is true and justified, it is said, because Brown is in Barcelona.

We must consider what it means to believe a disjunction. The belief represented by (4) may be restated as:

5. One of the following is true: Jones owns a Ford; Brown is in Barcelona.

If Smith believed that either Jones owned a Ford or Brown was in Barcelona, but was not sure which was true, the belief could be classified as disjunctive. However, as it stands, Smith is not ambiguous. For Smith, “one of the following” in (5) is a definite description and it refers to “Jones owns a Ford.” Thus, the belief is false.

We are perhaps naturally inclined to forget that a single sentence can represent different beliefs for different people, even when they are in almost identical circumstances. This explains the tendency to attribute justified true beliefs in Gettier cases, even when the beliefs in question are false. The above analysis also explains the strong intuition to deny propositional knowledge. We know that (2) is true, in so far as we know that we use (2) to express a true statement about Smith. Since Smith does not have our knowledge, Smith cannot know about this true statement. Therefore, we do not grant him the relevant propositional knowledge. Similarly, we say (4) is true because we know that Brown is in Barcelona. Since Smith lacks this knowledge, we do not attribute the relevant propositional knowledge to him. We only falter in supposing that (2) and (4) represent Smith’s true beliefs, when they only express our own judgments about the cases.

The above analysis accounts for the intuitions to attribute justified true beliefs as well as the intuitions to deny propositional knowledge in Gettier cases. At the same time, it preserves the conception of propositional knowledge as justified true belief. I submit these as grounds for rejecting the claim that there is a Gettier problem.


Gettier, Edmund (1963). Is Justified True Belief Knowledge? Analysis, 23, 121-123.

Strawson, P. F. (1950). On Referring. Mind, 59, (235):320-344.