The coherence theory of truth states that a proposition is true if it is one amongst a set of coherent beliefs. As a necessary condition which must be satisfied in order for a philosophical system to be sound, the test of coherence – or to be more precise, consistency – amongst the pertinent propositions functions very well. Moreover, when coupled with a first principle, one is even able to avoid Russell’s assertion that two equally coherent systems could hypothetically exist on the grounds that such is incompatible with one’s first principle and that “…all other theories of truth lead to skepticism.” Such a belief may not be falsifiable, but then again, the idea that a proposition must be falsifiable in order to be justifiably believed is itself unfalsifiable. Instead, what is important is that which constitutes the preconditions for knowledge.
When not synthesized with a variation of foundationalism, however, the coherence theory of truth is open to numerous criticisms. Firstly,“…for the coherentist…The system of justified beliefs is detached from the world beyond, in a sense that justification is an internal matter totally unaffected by what lies outside.” That is, it is not clear that belief in the proposition “a coherent system does not necessarily reflect reality” is incoherent with a set of otherwise coherent beliefs, in which case the practical value of the coherence theory of truth is questionable.
Another point of concern is whether or not the justificatory method of coherentism is actually foundationalistic. If the method by which one justifies his belief is showing it is “a belief which coheres with a set of otherwise coherent beliefs,” such a definition is a precondition for belief that a given belief is justified. It may be that this definition is self-attesting, but either way, such would not preclude it from being a first principle, for it would in either case not be accepted on the basis of any other belief. Regardless, given the fact there exists a predication of beliefs within the coherentist’s framework, his theory of knowledge will contain a [set of] presupposition[s] from which all other non-arbitrarily accepted propositions are purportedly derived, and these presuppositions are foundationalistic first principles.
The most devastating critique of unqualified coherentism is provided by Bertrand Russell: “…this definition of truth is that it assumes the meaning of 'coherence' known, whereas, in fact, 'coherence' presupposes the truth of the laws of logic...if the law of contradiction itself were subjected to the test of coherence, we should find that, if we chose to suppose it false, nothing will any longer be incoherent with anything else.” Russell alludes here to the principle of explosion, the concept that if any contradiction is true, trivialism follows. The law of contradiction, then, cannot be subject to the justificatory test of the coherence theory of truth without simultaneously affirming and denying the validity of the test, and so the law of contradiction is actually a precondition for it.
Less well-known in comparison to the other theories of truth is Tarski’s semantic conception of truth: “a true sentence is one which says that the state of affairs is so and so, and that the state of affairs indeed is so and so.” It appears to be a sort of correspondence theory of truth which, as is obvious from the context of Tarski’s definition, is meant to include empirical observations and observation statements.
The problem with this theory is that it does not adequately explain the means by which one can justify what is “the state of affairs.” If the semantic conception of truth is only sensible when unified with a particular philosophical system, it is suspect to that system’s attendant difficulties. Logical positivists, for example, used this to connect syntax with semantics or meta-language with object-language, but since the semantic conception of truth is neither tautologous nor empirically verifiable, it fails the verification criterion of meaningfulness.