Vowel quantity before dentals: on the interaction between morphology and phonology in English monosyllables

Nikolaus Ritt (University of Vienna)

During the Early Modern period, vowels in English monosyllabic words were sporadically shortened in words such as blood , book , bread , cloth , cook , death , done , grit , lead , look , shred , soot , took , wet etc.. As has long been observed (see, for example, Hackmann 1908, Luick, 1914/21, Brunner 1960, Dobson 1968 or Görlach 1978), such shortenings seem to have occurred rather frequently before dentals such as /d/ or /T/. In Ritt (1997) it is argued that dentals may have favoured such shortenings not so much by virtue of their phonological dentalness, but rather because many inflectional suffixes were realised as dentals. This made EmodE /CVVD i / sequences (in which D i stands for dentals that express inflectional suffix morphs) morphologically ambiguous, in that they could stand either for {CVV}+{D i } or for {CVVD i }. Following Dressler (e.g. 1985), this would have made them semiotically dispreferred, as they would have violated the preference for bi-unique signs. Obviously, the same would not have been true of /CVD i / sequences. They must have been unambiguously monomorphemic, because a lexical {CV} morpheme would have been phonotactically ill-formed already in Early Modern English. This circumstance would have made /CVD i / sequences semiotically preferable over /CVVD i / sequences as representations of mono-morphemic word-forms. Accordingly, Ritt (1997) hypothesises the that this preference may explain the relative frequency of EModE vowel shortenings before word final dentals in monosyllabic lexemes.

Fortunately, the hypothesis is falsifiable and entails predictions that can be tested. First, it implies that CVVD i lexemes should generally be significantly less frequent than CVD i lexemes in synchronic samples of the Modern English lexicon. This is because the semiotic preference which Ritt assumes ought to have its assumed impact not only in the specific EModE vowel shortenings he deals with, but, more generally, on all processes that affect the inventory, and/or the shapes of lexical monosyllables. Secondly, Ritt's hypothesis implies that CV(V)D i monosyllables should behave differently from other CV(V)D monosyllables, i.e. from monosyllabic items with final dentals that cannot represent inflectional suffixes.

The present paper intends to test these predictions against the evidence of Early Modern English corpora and PDE dictionaries, and to discuss what the results of this test imply for Ritt's (1997) hypothesis.

References

Brunner, Karl. 1960. Die englische Sprache . 2nd edition. Tübingen: Niemeyer.

Dobson, Eric J.. 1968. English pronunciation. 1500-1700 . vol. II. 2nd edition. Oxford: Clarendon.

Dressler, Wolfgang U.. 1985. Morphonology. The dynamics of derivation . Ann Arbor: Karoma.

Görlach, Manfred. 1978. Einführung ins Frühneuenglische . Heidelberg: Quelle und Meyer.

Hackmann, Gottfried. 1908. Kürzung langer Tonvokale vor einfachen auslautenden Konsonanten in einsilbigen Wörtern im Alt-, Mittel- und Neuenglischen . Halle/Saale: Niemeyer.

Luick, Karl. 1914/21. Historische Grammatik der englischen Sprache . Stuttgart: Tauchnitz.

Ritt, Nikolaus. 1997. Early Modern English vowel shortenings in monosyllables before dentals: a morphologically conditioned sound change?. In: Hickey, Raymond and Stanislaw Puppel (eds.). Language History and Linguistic Modelling. A Festschrift for Jacek Fisiak on his 60 th birthday . Berlin: Mouton