This aim of this study is to present the suprising occurrences of the voicing of voiceless plosives by a Murcian speaker. A quantitive account is given of the frequency of consonants that remain plosives and those which appear as approximants. In 15.85 % of the cases an approximant was pronounced, and full voicing was present 74.39% of the time (58.54% in the cases of the plosives). In addition, 95.74 % of the occurrences of voicing were produced in a non-initial context and not following an «s» in the coda, since in these contexts a voiceless stop is usually produced. The study concludes with a perception test which highlights the confusion of informants between minimal pairs of voiceless-voiced sounds. Informants strongly favoured the voiced phoneme when the words appeared isolated whilst there was a noted lack of confusion when they appeared in phrases. It is suggested that since there is a neutralisation of the features of voicing, tension and duration, it is the context which is important for the final interpretation of the phrase. Keywords: voicing, approximants