Synchronic rules of word-final devoicing are unnatural to the extent that consonants are devoiced not only before pause (or a voiceless consonant), but also in the intervocalic position, preceding a vowel-initial word, a context that should favor voicing instead. A common assumption is that devoicing starts diachronically in the prepausal context, where it is phonetically motivated, and from there it spreads to the phrase-medial context by analogy. In order to test this hypothesis, I examine the optional devoicing of final /-d/ in the Spanish of northwestern Spain (Old Castile), using a corpus of conversational speech. In this Spanish variety, /-d/ variably devoices to a voiceless fricative or deletes. The results show that voiceless fricatives are found in all phrasal contexts, but with significantly higher frequency before pause than before a vowel, lending support to the hypothesis of extension of the devoicing from the former context to the latter. The devoicing of /-d/ is neutralizing. Voiceless fricative realizations of /-d/ do not differ from those of phonemic /-θ/ either in amount of voicing or in duration. The two phonemes, however, are not systematically neutralized word finally, since, as mentioned, another common option for /-d/ is its deletion (other realizations, including stops and voiced approximant are uncommon in our corpus). Phonemic /- θ/, on the other hand, is not subject to deletion.Whereas the deletion of /-d/ has some lexical exceptions, its devoicing does not. Among the majority of /d/-final words, for which deletion is possible, the relative frequency with which they undergo deletion vs. devoicing appears to vary substantially depending on the lexical item. That is, both position in phrase and lexical identity probabilistically determine the realization of /-d/.