What is new since [nineteenth-century debates about poetry and theology] are theories of language that, in various ways, bypass Coleridgean questions about the truth of the imagination by asserting instead the truth of language. Perhaps best known is Michael Polanyi's claim that languages have an innate bias towards the truth. Clearly he did not mean that it is impossible to tell lies - he never disputed that much communication is intended to manipulate facts, or even to promulgate untruths - but, Polanyi argues, in the long run lies are usually seen to be just that. This is not a bid for access to absolute truths: rather, it entails the claim that unlike Orwell's Newspeak, or various technical languages operating with precise definitions, the innate fuzziness of ordinary speech has a long-term self-correcting tendency to revert towards the truth. Even though it may be possible to fool most of the people for most of the time, truth, like cheerfulness, will keep breaking through. If emperors fail to wear clothes, sooner or later someone is going to notice.
Though many postmodernists echo Plato's Thrasymachus in claiming that there is no such thing as truth, most poets wilfully persist in the conviction that they are somehow in the truth-telling business. Even the most scurrilous and cynical among them have usually insisted that they are exposing truths about human corruption and frailty. Indeed, if the creators of "fictions," poetic or novelistic, really believed in a total separation of language and truth, they would soon be out of business...
But the idea that language has an inescapable bias towards truth does not, of course, necessarily offer any moral guarantees, still less theological ones. It was Derrida, not Polanyi, who described the idea of innate textual meaning as "theological" - and it was hardly a compliment.
-- Stephen Prickett, "Religion Will Keep Breaking Through," TLS, November 11, 2011