摘要: In the literature on sound change, much has been made of neogrammarian doctrine that changes operate without exceptions. Without some such hypothesis any description would be a long list unsystematic correspondences, with no assurance same under comparable conditions not change into variety different sounds, governing principle whatsoever. This point view, which Hockett termed 'regularity hypothesis', richly rewarding in historical research; as celebrated part heritage modern linguistics, needs elaboration here. When irregularities appear to leak through net postulated phonetic laws, there should an explanation for them. words Karl Verner, 'There must rule irregularity; problem is find it.' searching explain certain residues Grimm's Law, Verner found it necessary go beyond segmental environment accentual systems IndoEuropean and Germanic, though condition he discovered still phonetic. recent years, impressive body evidence accumulated are diachronic rules depend altogether nonphonetic, e.g. factors morphological syntactic.' That this so surprising: phonology, frequently leave counterparts form synchronic rules, numerous instances where obviously dependent 'grammatical prerequisites'. Furthermore, stock morphemes language often partitioned several layers according non-phonetic criteria, these exhibit phonological behavior. To large extent, partitions correlate sources various layers, Romance versus native English, or Chinese Japanese.