He says the book "claims to present, not merely a theory of poetic diction, but a theory of poetry: and not merely a theory of poetry, but a theory of knowledge. It is as such, I see, that it must be judged. Apparently, the author was determined that the title should at least be unassuming" (14).
The current generation of people, Barfield notes (23 years after the first ed.), is not quite so interested in the "various theories of language, life and literature" that Barfield was concerned to address (rationalizing his own observations in terms of these) when the book was written. This is because it now regards them as "irrelevant to its own more maturely skeptical philosophy". It is confidently asserted of Kant, Berkeley, and Locke, for example, that they were not so much wrong, but "asked the wrong questions" (14, 15). He says that if he were doing the book today, he would criticize Hume and his more recent disciples in the Appendices instead, "not the less so because [now] the fashionable method is to analyse language itself - which is the heart of my matter" (15).
He notes that any readers coming "fresh to the subject" may at this point want to start reading at Chapter One, coming back to the Preface later (you have been forewarned!) It becomes clear why he says this: he gets into very philosophical territory. He discusses and dismisses the ideas Dr. I. A. Richards and the Logical Positivists, starting with this:
"Now in a footnote to page 113 in this book it is pointed out that 'logical judgments, by their nature, can only render more explicit some one part of a truth already implicit in their terms'. And in another, to page 131, that the logician is continually seeking to reduce the meaning of his terms, and that 'he could only evolve a language whose propositions would really obey the laws of thought by eliminating meaning altogether'" (16).
He goes on to say that "I do not think it too sweeping to say that the doctrines of linguistic analysis, or as it has sometimes been called, Logical Positivism, are no more than an extensive gloss on this principle". The Logical Positivists basically said that "all propositions except those from which some observation-statement can be deduce are, it is averred, meaningless, either as misuse of language, or as tautologies" (16, 17), (me: therefore eliminating the claimed importance of religious language, for example, which they believed was completely unverifiable)
Barfield then naturally moves on to talking about Locke and Hume, who attributed great importance to observation, or sense-impression, and dismissed any notions of "innate ideas" or innate thoughts. For Hume, he notes, "man, as knower, is above all a passive recipient of impressions [or ideas, which for Hume are the perceptions]. Such is also the assumption on which the edifice of physical science is erected" (17).
My summary: The whole idea of science has (so far at least, when he is writing this) been based on the idea that we can, through the power of our sense impressions and the careful use of descriptive language, "pin things down" (I say: to capture and use the truth). In order to do this, they often have to reduce the meaning of the terms they use. Furthermore, Logical Positivists insist that this is the only kind of real knowledge there is.