Published online by Cambridge University Press: 06 February 2009
Icelandic has a number of ‘surface minimal pairs’ between voiced and voiceless nasals. Facts and arguments are presented that speak for the phonemic status of these sounds, in contrast to the widespread view that voiceless nasals in Icelandic are derived. It is shown, among else, that stops after voiced and voiceless nasals are phonetically identical. Due to this fact it is not possible on synchronic grounds to ascribe the distinction between voiced and voiceless nasals to a distinction among the following stops, unless phonological abstractness is used as an analytical tool. The rarely mentioned case of single word-initial voiced versus voiceless nasals (where no subsequent stop occurs) would require an even greater commitment to phonological abstractness if the derived status of voiceless nasals were maintained. Implications for linguistic phonetics and the IPA are discussed.