Edited by me from Romantic Religion: a study of Owen Barfield, CS Lewis, Charles Williams and JRR Tolkien. by RJ Reilly (1971) - republished version of 2006 from Lindisfarne Books, Great Barrington, MA, USA; pages 214-5. My editorial interjections are in [square brackets].
Barfield sees both myth and language itself as existing in the form of unconscious meaning before the existence of any individual thinker. Both myth and language point-back towards the pre-human time when all that existed was spirit, un-individuated meaning; the original phase of the cosmic evolution.
These myths (the Paradisal myths, for example) therefore suggest truth in a quite literal sense: they allude to the original 'way things were'.
[That is, myths allude to Original Participation.]
Both Lewis and Tolkien, by contrast, speak of the possibility of all myths being 'true' in some other existence than our own. Williams, too, feeling the call of myth, goes so far as to adapt the Arthurian Myth as a kind of objective correlative for his religious views.
But, quite plainly, Barfield has explained the origin and force of myth in a way that the others have not. They have used myth in various ways and with varying degrees of effectiveness, but they have not really said why. Or rather, Lewis, Tolkien and Williams have have used myth, or they have made-up new myths, as a means of avoiding conceptual argument, or as a means of speaking symbolically rather than rationally.
There is nothing wrong with what Lewis, Tolkien and Williams do in using myths, so far as it works. But to the extent that it can be reduced to a set of rational propositions, it must strike the reader as making myth into something closer to allegory than to true myth.
True myth - in Barfield's terms, and in reality - is nearly impenetrable; because there are no 'ideas' in myth for the reader to penetrate to.
For Barfield, myth is the closest thing in Man's mental life to pure pre-logical thought; meaning which the rational intellect has not yet ordered. Myth is more of an experience than a 'thought' at all.
Barfield argues that the function of the imagination in the future will be to discover 'clear and distinct ideas', but it may discover these in the forms of William Blakean 'beings' rather than as concepts - these beings being explicable in something analogous to the way that the beings of the old allegories like Bunyan's Pilgrim's Progress are explicable.
[Examples of such 'beings' would, presumably, include Steiner's demonic influences Lucifer and Ahriman - which Barfield also describes in detail in Unancestral Voice.]
NOTE: The above passages strike me as a highly insightful analysis of a topic of vital importance.
In particular, I am impressed by Reilly's point that neither Tolkien nor Lewis (in particular) really justify the importance and specific value of myth in their writings; but Barfield does - assuming that one can accept, at some level, Barfield's point that, in some literal sense, myths refer to the nature of our experience during a previous state of spiritual reality.
For Barfield (and Steiner), this previous state would be modern people's earlier incarnations at an earlier point in the history of earth; for Mormon Christians and some others, myth could refer to our pre-mortal, pre-incarnated life as spirits.