View Message

I've provided for those apparent exceptions — key vowels in those words are frequently pronounced as the unvoiced ə̥ˈ (or similar) forcing the s (or c pronounced as s) to be unvoiced as well, and without a final voiced vowel the s may (or may not) also remain unvoiced as well. Besides "normal" implies there are exceptions such as those anyway, and dialects where s is unvoiced where in others it would naturally be voiced (as noted many Americans often pronounce s as unvoiced where it would naturally be voiced and is in other English dialects). Another factor may be whether the following consonant is voiced or unvoiced. This may help explain divergent pronunciations of Joseph — the ph is unvoiced, and the e may also be unvoiced, so medial s could be either voiced if the e is voiced, but not if the e is unvoiced (as the ph is unvoiced).