late 13c., "sound made by the human mouth," from Old French voiz "voice, speech; word, saying, rumor, report" (Modern French voix), from Latin vocem (nominative vox) "voice, sound, utterance, cry, call, speech, sentence, language, word" (source also of Italian voce, Spanish voz), related to vocare "to call," from PIE root *wekw- "to speak."
Replaced Old English stefn "voice," from Proto-Germanic *stemno, from PIE *stomen- (see stoma). Meaning "ability in a singer" is first attested c. 1600. Meaning "expression of feeling, etc." (in reference to groups of people, etc., such as Voice of America) is recorded from late 14c. Meaning "invisible spirit or force that directs or suggests" (especially in the context of insanity, as in hear voices in (one's) head) is from 1911.