In a recent article in LANGUAGE,1 Hans Kurath criticized some remarks of mine in the same journal.2 I have profited from his criticisms, some of them quite justified; a personal or bad-tempered reply would be inexcusable. On the other hand, the issues which Kurath has raised are basic; his arguments are not always either clear or conclusive; and if they go unquestioned, they may be accepted, not because they are sound but because they are his. Since inquiry among my linguistic friends suggests that some explanations might be of general interest, I wish to ask a number of questions. 1. Kurath proposes3 that 'if two phonemic solutions appear to be equally probable or equally doubtful, the solution that facilitates the tie-up between the synchronic and the diachronic treatment of the language should be adopted.' In the light of this principle, his historical statements are at least indirectly relevant to a discussion of Modern English, and certainly they are of great interest in themselves. Some of them, both in his critique of 'the binary interpretation of English vowels' and in his discussion of the Middle English fricatives,4 are puzzling. 1.1. According to Kurath,5 'clOOO1 Old English had contrastive long consonant phonemes only in the sequence /Tccv/, i.e. between a stressed or half-stressed short vowel and a following unstressed vowel, as in settan, hycgan, spinnan, etc. and in bitter, appel.' But 'this statement refers only to the occurrence of /vcc-v/, in which both consonants belong to the same morpheme. Complex forms such as the preterits raed-de 'advised', met-te 'met', do not have long consonants, but sequences of identical short consonants; i.e. such forms must be taken as /vvccv/.'6 If these things were so, what was the nature of /-/? Can anything be said about the PHONETIC difference between a long consonant and a sequence of identical short consonants? Did the tt of sette, preterit of settan, represent a long consonant or a sequence? If the tt of settan and the tt of sette represented different phonemic entities, is any comment needed on the identical spellings? 1.2. In Old English, Kurath finds three sequences 'in which a single consonant, long or short, is flanked by vowels': /vccv/, /vcv/, and /WVcv/.7 This situation was altered, and the three sequences reduced to two, when in some dialects of Middle English 'the sequence /vca/ changed to /vvca/ in words of two syllables, through the lengthening of ALL short stressed vowels in this position'.8 Now,