What interests me most in conducting this argument is the difference that is constantly appearing between the poetic and the prosaic methods of thought. The prosaic method was invented by the Greeks of the Classical age as an insurance against the swamping of reason by mythographic fancy. It has now become the only legitimate means of transmitting useful knowledge.

And in England, as in most mercantile countries, the current popular view is that 'music' and old-fasioned diction are the only characteristics of poetry which distinguish it from prose: that every poem has, or should have, a precise single-strand prose equivalent. As a result, the poetic faculty is atrophied in every educated person who does not privately struggle to cultivate it ... from this inability to think poetically - to resolve speech into its original images and rhythms and re-combine these on several simultaneous levels of thought into a multiple sense- derives the failure to think clearly in prose.

In prose one thinks on only one level at a time, and no combination of words needs to contain more than a single sense; nevertheless the images resident in words must be securely related if the passage is to have any bite. This simple need is forgotten, what passes for simple prose nowadays is a mechanical stringing together of stereotyped word-groups, without regard for the images contained in them.

The mechanical style, which began in the countinghouse, has now infiltrated into the university, some of its most zombiesque instances occurring in the works of eminent scholars and divines.

Mythographic statements which are perfectly reasonable to the few poets who can still think and talk in poetic shorthand seem either nonsensical or childish to nearly all literary scholars. Such statements, I mean, as: 'Mercury invented the alphabet after watching the flight of cranes' ... The best that the scholars have yet done for the poems of Gwion is 'wild and sublime'; and they never question the assumption that he, his colleagues and his public were people of either stunted or undisciplined intelligence.

The joke is that the more prose-minded the scholar, the more capable he is supposed to be of interpreting ancient poetic meaning, and that no scholar dares to set himself up as an authority on more than one narrow subject for fear of incurring the dislike and suspicion of his colleagues.

To know only one thing well is to have a barbaric mind: civilization implies the graceful relation of all varieties of experience to a central human system of thought. The present age is peculiarly barbaric: introduce [two scholars from different disciplines] and the pair of them would have no single topic in common but the weather or the war (if there happened to be a war in progress, which is usual in this barbaric age).

But that so many scholars are barbarians does not much matter so long as a few of them are ready to help with their specialized knowledge the few independent thinkers, that is to say the poets, who try to keep civilization alive. The scholar is a quarryman, not a builder, and all that is requried of him is that he should quarry cleanly. He is the poet's insurance against factual error.

...Fact is not be gainsaid; one may put it in this way, that fact is a Tribune of the People with no legislative right, but only the right of veto. Fact is not truth, but a poet who wilfully defies fact cannot achieve truth.

Robert Graves, The White Goddess - pp. 236-8 - Vintage 1948