Examines the plausibility of an asymetrical phonological inventory
where /b/ has no voiceless counterpart. Furthermore, argues that the loss of /p/ is not related to the lenition of voiceless stops as this phaenomenon postdates the Proto-Celtic period.