The phonologically voiceless fricatives of Old English (noch einmal)

Abstract The phonological status of Old English (OE) fricatives has been a vexed one, the general agreement being that the distribution of voiced ([v ð z ɣ]) and voiceless ([f θ s x]) fricatives was allophonic (Fulk 2001; Minkova 2011, 2014). We argue that OE was a fortis–lenis language specified for [spread] in terms of laryngeal realism, or ‘glottal width’ (GW) (Avery & Idsardi 2001). We discuss OE lenis and fortis stops, the structure of voiceless geminates ([ff], [ss], [tt], etc.) and voiceless geminate‑like structures ([sp], [st], [xt], etc.) and conclude that OE had phonologically marked fricatives for GW, found as the first member of phonetically voiceless (partial) geminates ([ff] /f GW f 0 /, [sp] /s GW b 0 /). Unmarked singleton fricatives, by contrast, were phonetically enhanced with GW in strong positions (foot‑initially in trochees and degenerate feet) and with ‘glottal tension’ (GT) in post‑tonic foot‑internal intersonorant position, which is less controversial. They were, however, unenhanced word‑finally and when couched between unstressed vowels, and thus phonetically variably voiced. We explore some of the consequences of entertaining such ideas.