i did a lttle on the subject as part of my degree, & the origin of 'language' as a form of communication isnt clear cut. however, it is thought that amongst primates, especially well documented in baboons, that the cultural development of verbal communication owed much to the habit of mutual grooming. 'lip-smacking' and other exaggerated facial expressions accompanied with grunts etc form the basis of different phonemes. the coordination required to hunt prey in hominids -eventually humans -greatly accelerated the evolution of language, & the ability to represent abstract comcepts verbally.
as for words themselves; the shape of the mouth when the word is spoken is often analagous to the object or concept the word describes. an example of this is the word 'large', which requires you to open your mouth wide for an extended period -literally making your mouth large, and in contrast the word 'tiny' or 'little' sees the mouth remain relatively small with lips together for a short time. this is true across language barriers, 'grand' & 'petit' in french for example. the physical act of saying many basic words can usually be linked in this way, many swear words are short hard consonant words -e.g. sh*t & f*ck, implying an unpleasant or violent movement, just as words such as 'beautiful' and 'nice' etc. are almost pleasant to say and are usually spoken softly. newer, more developed words may not neccessarily follow this trend. over time, thousands of years, groups of humans came to accept standard words as representing concepts, greatly helped by the innateness of linking facial expression especially in the mouth, as i have dedscribed, with objects and concepts.
& technically words are generated from air expelled by the lungs which is manipulated by the vocal chords, tongue & lips.