30 But it would be plainly to cheat the original if we were guided only by the contemporary taste for the colloquial. Some compromise is necessary. And even there, one runs the risk of making parody. When Wallace Stevens had recourse to traditional high style, it was often for the sake of kidding it, or at least putting a left-handed irony into his tone. My own compromise was to try to keep the diction modern where I could, particularly as to its simplicity and concreteness, but to make locutions fairly complicated; for, in fact, high style is less a matter of diction or figure than of syntax, at least in English. The results are uneven, since this method was more applicable to some verses than others. The Vedas, I think, come off pretty well because of their unequivocal remoteness, their powerful forward thrust, and their austere surface. Clearly, in American, they are not "modern," but at their best, they are alive in the context of the modern. Some of the verse from the mdhakavyas is least successful because it depends so heavily on a tightly-woven, deeply-napped texture of diction and figure. Such verse seems to invite a sort of Keatsian decor, mistakenly because that suggests a personal intensity not in the original. The question of the personal constitutes another problem. While modern Hindi poetry is modern poetry and written by a deliberately idiosyncratic person, traditional Indian poetry was deliberately moved toward a general style. Brilliance within that style or extraordinary management of that style is one thing — and to be praised; deviation from it to express the private feelings of the poet is another, to be dismissed as sub-poetry, if, indeed, anyone ever tried it. In short, the translator has to resist a very contemporary impulse to "make it new" while trying somehow to do just that, because for us originality means an individual poetic voice heard for the first time ever, rather than a closer approximation to some ideal general style. For the traditional Indian poet, originality would have meant, as Eiroy Bundy has said in another connection, finding oneTs origins, in this instance, finding the best models of the tradition^ After working up the flat text, I turned the results back to my collaborators for close criticism. That meant usually pointing out where I had misread an implication, put weight on the wrong place, over or undersaid, or simply failed to get the tone. I never thought consciously about msa, but assumed that my texts and my collaborators would guide me to the proper mood. I cannot imagine that Sanskrit poets consciously worried about rasa every minute either. Absorption in the act of composition does^t leave the serious poet much time for critical theoryo He could count on his tradition to take care of that. It was a most dependable collaborator. I found prosody another problem. The Vedas were easiest; listening to them chanted, one knows their music cannot get over into our language. But they have a kind of loosened iambic feel when spoken and are, therefore, closer to us metrically than the often heavily trochaic pattern of much modern Hindi verse (someone is going to say here: "Good grief, man, Indian verses don1t have stress.T" I think they do; and in any case, I hear stress in them because my ear is American). The loosened iambic also seemed to apply to the mahakavyas and to verses from plays, but these