[…] Do the phonetic conditions of a sound law – whatever their form might be – remain constant over a period of time? I would like to answer that here and now with a concrete example. Modern French 'e' […] in 'chef', 'fève', 'pré',' tel', 'mer', 'nez', 'ème', 'lène' = 'caput', 'faba', 'prato', 'tale', 'mare', 'naso', 'amat', 'lana', corresponds to Gallo-Vulgar Latin 'ā' […] The following consonant appears to be a matter of indifference. This was not so in Old French. This fact is reflected in present-day orthography: 'chef', etc., but 'áime', 'láine'. If now 'ā' before 'm' and 'n' has become 'e' passing through the stage 'ai', could this not likewise have happened before the other consonants? And if 'chaif', 'faive',' tail', 'mair' were originally pronounced, it is conceivable that in a more recent period 'chaif', 'faive',' tel', 'mer' occurred. Thus different stages of conditions were present with respect to the monophthongization of the 'ai' that developed from 'a'. (p. 55)
- Schuchardt (1972)