"a union of two vowels pronounced in one syllable," late 15c., diptonge, from Late Latin diphthongus, from Greek diphthongos "having two sounds," from di- "double" (from PIE root *dwo- "two") + phthongos "sound, voice," which is related to phthengesthai "to utter a sound, sound, raise one's voice, call, talk," which Beekes reports as of "no certain etymology. None of the existing connections with semantically comparable words ... is phonetically convincing." Related: Diphthongal; diphthongization.
In uttering a proper diphthong both vowels are pronounced; the sound is not simple, but the two sounds are so blended as to be considered as forming one syllable, as in joy, noise, bound, out. An "improper" diphthong is not a diphthong at all, being merely a collocation of two or more vowels in the same syllable, of which only one is sounded, as ea in breach, eo in people, ai in rain, eau in beau. [Century Dictionary]