Ⅰ．The challenge to the justification of knowledge

i．Agrippa's Five Modes

In the Outlines of Phyrronism , Empiricus put forward five modes leading to suspension of belief, and named them Agrippa's Five Modes. In Empiricus's view, the emergence of these five modes means that people are incapable of building knowledge that is not blamed by skepticism. They are disagreement, infinite regression, relativity, hypothesis, and reciprocity.

To simplify, Fogelin points out that “disagreement” and “relativity”mean that there may be many competing assertions about the natural world, and a justifiable epistemology system must be able to explain why some assertions are superior to others in knowledge. For example,different people and even different cultures may have different assertions and explanations about the same thing in the world. When people want to think about which opinions and explanations are knowledge, a reasonable epistemology should be able to explain why some assertions have epistemological priority or are closer to the concept of“truth” .

“Infinite regression”, “hypothesis” and “reciprocity” are used to illustrate that the reasoning process required by the former two is impossible. Infinite regression points out that people will rely on further proof when making proof. For example, the proposition “the sun rises in the east and sets in the west” itself requires people to believe that “the sun exists” . And “the sun exists” itself depends on the further proof process. Hypothesis shows that people's knowledge is not an objective truth, but only a kind of reasoning, so it doesn't have reality in the real sense. Reciprocity shows that a proof that tries to prove something or an event has its own problems. If we agree with Fogelin's classification of Agrippa's Five Modes, then a justifiable epistemology should be able to successfully answer the three modes of “hypothesis”, “infinite regression”and “reciprocity”, or at least explain to what extent these three modes are not applicable to the epistemology system.

Pyrrhonian Skepticism is different from Cartesian Skepticism.Pyrrho doesn't need people to take the most extreme reflection to question everything or whether the world is false in the most fundamental sense (such as the brain in a vat). On the contrary, Pyrrhonian skeptics can conduct their daily life well, and they are “moderate” skeptics.

ii．The Truth View and Challenge of Correspondence Theory of Truth

The truth view of correspondence theory of truth is a kind of truth theory which is very intuitive. Its basic proposition is that knowledge is the correspondence of facts. Correspondence theory of truth has many forms,such as Frege's referential theory, Russell's definite description, Tarski biconditional and Kripke's causal theory of reference. In order to simplify the discussion, I will discuss the correspondence theory of truth based on Tarski Convention next.

According to Tarski biconditional, the following statement can be made:

A proposition “A is F” is true if and only if A is F.

According to Douglas Edwards 'interpretation, we can understand Tarski Convention as: the object referred to by “A” belongs to the predicate“F” if and only if the object referred to by “F” to “A” is actually true. In everyday language, we can simply understand that the proposition “apple is red” is true and only if the predicate “red” is actually true for apple. In other words, apples are red, not green.

Note that the Tarski Convention itself can be interpreted in two ways. The first explanation is the correspondence theory of truth we just mentioned, which means that a proposition is true if and only if the corresponding facts described by the proposition exist in reality. Another explanation is as follows: “A is F” is true if and only if A is F, because the right half of the biconditional itself is a tautology of the left half, so it is strictly true.

Edward thinks that people believe in the correspondence theory of truth should accept the first explanation. Firstiy, the second explanation itself lacks the participation of the outside world. For the correspondence theorists, the most important thing is exactly what the facts of the external world itself look like. Beliefs that conform to the facts are true and may(“may” is because of the existence of Gettier's counterexample) become knowledge, while beliefs that do not conform to the facts are false.Secondly, Edward pointed out that the second explanation cannot give people knowledge, at least it is far weaker than the first explanation which includes the outside world. As Kant said, “Thoughts without intuitions are empty, intuitions without concepts are blind.” The second explanation just doesn't explain how people should acquire knowledge, and even arbitrarily implies that people can create countless meaningless true propositions. For example, “unicorns fly in schools” is true only if unicorns fly in schools.

As a phased summary, we can summarize the claims of Tarski's correspondence theory of truth as follows: a proposition “A is F” is true if and only if A is F, because the object referred to by “A” belongs to the predicate “F” if and only if the object referred to by “F” to “A” is actually true. If a proposition corresponds to reality, the truth value is satisfied (it's 1). If a proposition doesn't correspond to reality, the truth value is not satisfied (it's 0).

The first challenge to external correspondence theory of truth comes from Frege himself, the inventor of correspondence theory of truth. Frege is keenly aware that correspondence must be the correspondence of a certain property. For example, for all true propositions, we can say that they all have a common attribute of “being true”; For all false propositions,they similarly have a common attribute of“not being true”. The truth of a proposition means that its proposition itself correspond to the truth condition of“being true”.

Here comes the question. Please look at the following three propositions and assume that they are all true. In order to approach the daily expression, the definition of Tarski is expanded here:

(P1) Lu Xun is the author of Call to Arms .

(P2) Yes, this beer is bitter.

(P3) Hosico is a cat.

It is known that these three propositions are all true propositions,so they all meet the truth condition of“being true”, and therefore refer to the attribute of“being true”. This implies that P1 to P3 have the same reference. Furthermore, the referential correspondence theorists face a criticism in the metaphysical sense: all true propositions have the same reference. Frege named it “the great reference.” This conclusion is obviously unacceptable.

Furthermore, Fogelin pointed out that the referential correspondence theory not only has fatal problems in metaphysics, but also it is too vague in epistemology, and there is a question-begging argument. This vagueness just stems from the Cartesian binary opposition between the external world and the internal world. The classic question arises: How can we get an “objective” understanding of the outside world beyond ourselves?Chisholm tried to answer this question by inventing a complex high-order cognitive system, but he finally got into the question-begging dilemma, and built his theory on the assumption that “we should believe that beliefs are true”.

Fogelin concluded that the attempt of correspondence theory was unsuccessful, because it not only had defects in metaphysics, but also fell into the question-begging argument in epistemology.

iii．Coherence Theory of Truth and Censure

Coherence theory of truth is a view of truth different from correspondence theory, which claims that another belief is the only basis to justify a belief. There are two kinds of coherence theory of truth: internal coherentism and Davidsonian coherenism. I will discuss the latter.

Davidsonian coherenism is very complicated, and he himself has not given a complete definition directly, so it is necessary to understand what the traditional coherence theory of truth is first. According to Laurence BonJour, we can express it as:

(1) A belief system is coherent if and only if it is logically coherent.

(2) The consistency of belief system is directly proportional to the degree of consistency of its probability.

(3) The existence of the reasoning connection between its constituent beliefs increases the difference of systematic beliefs, and increases proportionally with the number and intensity of such connections.

(4) The coherence of a belief system is weakened, so that it is divided into belief subsystems, which are relatively conflicted and disconnected from each other through reasoning.

(5) The consistency of belief system decreases proportionally with unexplained anomalies of the trusted content in the system.

According to this view, a belief system that is logically coherent and supported by sub-beliefs is closer to truth. For example, if there are two conflicting beliefs in a person's belief system, then it can be considered that this person's belief system needs to be revised. This basic idea was directly used by Davidson as a basis. However, different from the traditional coherence theory, Davidson uses a Tarski biconditional, namely:

A proposition “A is F” is true if and only if A is F.

Different from Tarski's basic position of correspondence theory,Davidson reinterpreted the conditions for the T condition to come true. In Davidsonian coherenism, “A is F” with quotation marks is the expression of a speaker in a given language. The second part of “A is F”is the interpreter's translation of “A is F” in a metalanguage (meaning of expression). If the speaker and interpreter successfully establish the T condition, then the expression is true. In Davidson's view, meaning is given by an objective truth condition. The truth condition is tested by coherence.Therefore, meaning is explained by coherence. In a word, justification is explained by coherence, followed by truth. In other words, the Cartesian binary opposition between the subject and the object should be surpassed,because meaning and truth condition are bound to each other in the criterion of coherence. Furthermore, Davidsonian coherenism allows the existence of a few false propositions, which will be constantly corrected in the dialog.On this basis, Fogelin reconstructed Davidson as follows:

(1) If a belief (true or false) is a member of a coherent set of beliefs,then it is a reasonable belief.

(2) In addition, if this belief is true (if knowledge is proved to be a true belief), then this belief is regarded as knowledge.

The second premise will be challenged, because Davidson's “reality”here seems to presuppose the external world challenges faced by the correspondence theorists. Skeptics may ask: Is it possible for a person to form a coherent belief system about real world mistakes?

Here is Davidson's response: What is needed to answer skeptics is to prove that a person with (more or less) coherent beliefs has reason to think that his beliefs are basically not wrong. More importantly, in order to avoid the challenge of the outside world, this process of justification does not resort to experience. Davidson pointed out:

Third person argument: “A correct understanding of a person's language, beliefs, desires, intentions and other propositional attitudes will lead to a conclusion that most of a person's beliefs must be true.”

Davidson believes that this third-person view is a necessary prerequisite for the possibility of everyday language. Only by assuming that most of the propositions in a person's belief system are true can the interpreter's understanding of that person's belief system be possible. This premise is called the principle of charity: one should believe that most of one's beliefs must be true. If most of the beliefs of a speaker's belief system are true, then people can respond well to the challenge of skeptics.

Of course, many people think that the principle of charity looks more like Davidson's own wishful thinking, a belief, than a rigorous conclusion based on rational argument. However, the principle of charity itself is not an arbitrary belief: there are at least three reasons why people should accept it.

Firstly, only if most beliefs in a speaker's belief system are true,can inter-translation be possible. In Davidson's theory, the translation of T convention is accomplished through radical translation. Quine has an example of aborigines and a linguist: a linguist comes to a desert island and sees a group of aborigines hunting rabbits. Every time they find a rabbit, they shout out “Gavagai!” Through numerous observations of these indigenous “behaviors”, the linguist has come to such a conclusion that “Gavagai” in the indigenous language not only represents “rabbit”in English, but also represents the process of rabbit hunting. How does the linguist come to the conclusion? You know, the linguist doesn't have any knowledge about indigenous language befone, but the linguist and the aborigines have two things in common: first, they share a world. Second, the linguist can observe the relationship between the indigenous behavior and language expression. It is through the observation of behavior that the linguist can gradually try to understand the indigenous language and even their belief system. For example, when the aborigines find rabbits countless times and cry “Gavagai”, the linguist has the reason to think that the word represents rabbits. To test this conjecture, the linguist can further find a rabbit and say “Gavagai” to the indigenous people. If the aborigines express a positive attitude, then the linguist will establish an understanding of the indigenous language and even its proposition system.

However, this process must depend on the premise of the principle of charity. If the linguist assumes from the beginning that no matter what they say, the indigenous people will deceive him, then it is impossible to establish an understanding of the indigenous language. Or, if there is a fundamental difference between the belief system of the linguist and that of the aborigines, then the linguist can't establish such understanding,because no matter how he interprets the aborigines, he can't form a positive propositional attitude. Therefore, the principle of charity is presupposed by the intertranslatability of the languages between different races, cultures,and people.

Secondly, the intertranslatability is related to the outside world through the causal relationship between behavior and expression, so it is objective. Therefore, we should believe the truth theory established by the principle of charity. Let's go back to the example of the indigenous language. The linguist builds understanding of the word “Gavagai” through the consequences of indigenous behaviors. Therefore, the process of establishing T convention is causal and normative. Davidson believes that there is no reason to presuppose an external world. The significance of the world to us can only be grasped through the inseparable causal relationship between this kind of behavior and language.

Thirdly, Bayes Theorem provides some empirical evidence to support the rationality of the principle of charity. The probability that people can predict some events relatively accurately through Bayesian probability theory reflects the fact that most of our beliefs are true. If one believes that there is complete chaos in the system, that the solar system revolves around a mysterious cat, that every time the wind blows, apples will fly back to the trees, and so on, then the probability prediction based on these beliefs will also be chaotic.

The above views were criticized by Fogelin. He accused Davidson of falling into the question-begging argument and having no way to face the challenge of skepticism. I will reconstruct Fogelin's criticism in the third section, and point out that these criticisms are all based on the misreading of Davidson.