iv. ORTHOGRAPHY AND PHONOLOGY.--In the remarks upon Gower’s language which here follow there is no systematic completeness. Attention is called to such points as seem to be important or interesting, reference being made especially to the language of Chaucer, as dealt with in B. ten Brink’s _Chaucers Sprache und Verskunst_ (second edition, 1899). It is necessary perhaps to remark here upon a difference of procedure which distinguishes this investigation from those which have for their object the text of Chaucer or of other writers whose work is handed down to us in manuscripts which do not proceed from the author himself. In such cases we have first to ascertain what the author actually wrote, before we can draw any valid conclusions about the laws of his language. It may even be necessary to restrict the discussion to such forms as are authenticated by rhyme; but when we are compelled to do this, we must remember that we are accepting a rather dangerous limitation. The conclusions drawn from the rhyme-words of a Middle English author will probably not be precisely applicable to his language in general. The sphere of our investigations will be that in which the licentious and exceptional is most likely to be found. If he has any tendency to borrow from other dialects than his own or to use irregular forms, this tendency will be most seen in his rhymes, for it will probably be the exigencies of rhyme which suggest the variation. Chaucer repeatedly uses ‘here,’ in the sense of the modern ‘her,’ to rhyme with such words as ‘bere,’ ‘spere,’ but we should certainly not be justified in concluding that this and not ‘hire’ was the normal form of his language. Similarly in the case of Gower by examination of his rhymes alone we might be led to many very doubtful results. For example, we should gather that he almost always used the form _sinne_ rather than _senne_, _wile_ (verb) and not _wole_ or _wol_, _axe_ and not _aske_, _sek_ (adj.) and never _sik_, _hond_ and never _hand_, _couthe_ and never _coude_, _sente_ (pret.) rather than _sende_, the adverb ending _-ly_ in preference to _-liche_ or _-lich_. In these cases and in many others we might easily be misled, the forms of these words as used in rhyme being determined chiefly by the comparative frequency of the various rhyme-syllables. Most of the conclusions above mentioned, and others like them, have in fact been arrived at in a paper by K. Fahrenberg, published in the _Archiv für die neueren Sprachen_, vol. 89. The author of this paper, having only Pauli’s text before him, very properly confines himself to an examination of the rhymes, and within these limits most of his results are sound enough; but it would be very unsafe to treat them as generally applicable to the language of Gower. In our case it must be understood that the Fairfax manuscript is regarded (for reasons which will afterwards be stated) as a practically accurate reproduction of the author’s original