Modern scholars generally agree that in Geoffrey Chaucer's sound system, [e:] and [ε] were both full phonemes, a situation paralleled with the back vowels [o:] and [ɔ:]. Unfortunately, Chaucer's verse fairly frequently rhymes words that should have /e:/ (according to diachronic criteria) with words in which one expects /ε:/. The prevailing explanation for their occurrence insists that Chaucer created the faulty rhymes neither through ignorance nor poetic license; instead, he liberally employed common variant pronunciations from fourteenth-century London English. This paper argues that the study of rhymes has been heavily determined by a belief that this literary artifice, studied in a context informed by the knowledge of historical phonology, may permit us to recover facts of Chaucer's pronunciation that otherwise would be irretrievable. The connections between this presupposition and Derrida's critique of the spoken versus the written are revealed, suggesting that rhymes possess their own unique written properties. In a three-stage analysis, scholarly attention is redirected toward rhymes as a graphological phenomenon.

In einer Reihe von Artikeln haben Joseph B. Voyles und zwei deutsche Linguisten unter dem Einfluß generativistischer Theorien von neuem kritische Erklärungen des ahd. i−Umlauts und der nhd. Verteilung der Umlautsvokale unternommen. Diese Art von linguistischer Historiographie entspricht m.E. nicht der realen Sprachgeschichte.

Den Reichtum der Sprache findet man nicht, wenn man fragt: was hat das Wort x früher bedeutet? oder: woher kommt das Wort x?, sondern wenn man fragt: was sagt man alles für… ? (1)