or intuitively felt, with reference to the formal “genius” of his own language; they cannot be carried over without loss or modification.  Croce[196] is therefore perfectly right in saying that a work of literary art can never be translated.  Nevertheless literature does get itself translated, sometimes with astonishing adequacy.  This brings up the question whether in the art of literature there are not intertwined two distinct kinds or levels of art—­a generalized, non-linguistic art, which can be transferred without loss into an alien linguistic medium, and a specifically linguistic art that is not transferable.[197] I believe the distinction is entirely valid, though we never get the two levels pure in practice.  Literature moves in language as a medium, but that medium comprises two layers, the latent content of language—­our intuitive record of experience—­and the particular conformation of a given language—­the specific how of our record of experience.  Literature that draws its sustenance mainly—­never entirely—­from the lower level, say a play of Shakespeare’s, is translatable without too great a loss of character.  If it moves in the upper rather than in the lower level—­a fair example is a lyric of Swinburne’s—­it is as good as untranslatable.  Both types of literary expression may be great or mediocre.

[Footnote 196:  See Benedetto Croce, “Aesthetic.”]

[Footnote 197:  The question of the transferability of art productions seems to me to be of genuine theoretic interest.  For all that we speak of the sacrosanct uniqueness of a given art work, we know very well, though we do not always admit it, that not all productions are equally intractable to transference.  A Chopin etude is inviolate; it moves altogether in the world of piano tone.  A Bach fugue is transferable into another set of musical timbres without serious loss of esthetic significance.  Chopin plays with the language of the piano as though no other language existed (the medium “disappears"); Bach speaks the language of the piano as a handy means of giving outward expression to a conception wrought in the generalized language of tone.]

There is really no mystery in the distinction.  It can be clarified a little by comparing literature with science.  A scientific truth is impersonal, in its essence it is untinctured by the particular linguistic medium in which it finds expression.  It can as readily deliver its message in Chinese[198] as in English.  Nevertheless it must have some expression, and that expression must needs be a linguistic one.  Indeed the apprehension of the scientific truth is itself a linguistic process, for thought is nothing but language denuded of its outward garb.  The proper medium of scientific expression is therefore a generalized language that may be defined as a symbolic algebra of which all known languages are translations.  One can adequately translate scientific literature because the original scientific expression is itself