Yesterday I noticed this comment in Danker’s Concise Lexicon in his entry for ζάω:
this form is an invention of grammarians; see ζῶ.
And that’s all you find in CL s.v. ζάω; the lexical entry is found under the spelling ζῶ.
(The same info may be implied in BDAG where ζάω is listed in square brackets, but it is not nearly as explicit.) With that clue, I’ve since discovered that Liddell & Scott also lists the word as ζῶ instead of ζάω. I’ve also canvassed Smyth’s discussions over several scattered sections. I’ve learned something about the morphology of this word that I probably should have known long ago.
The original formation was ζη-ιω (with consonantal iota) which contracted to ζῶ. Other than the lexical form, however, every other inflected form still shows the original eta in the stem. This word appears to act like an alpha contract verb (alpha lengthening to eta), but either the eta is retained from the original root, ζη-ι ̯ω (not a lengthened alpha), or this is a rare example of an eta contract verb which acts much like an alpha-contract except that the contraction of η + ε = η, and the contraction of η + ει = ῃ, thus: ζῶ, ζῇς, ζῇ, ζῶμεν, ζῆτε, ζῶσι(ν). (BDF§88 says the eta was retained due to the substantive τὸ ζῆν, which seems speculative to me.)
As a result it has long been traditional to treat the word as if it were spelled ζάω since that was thought to account for the eta (as a lengthened alpha). Most NT grammars, lexica, and tagged texts list the lexical form as ζάω despite the fact that this is a fictitious form. I have concluded that we ought to follow Danker’s example and list this word as ζῶ, not only in the lexicon, but also in first year grammars. (Danker’s CL is now by far the best lexicon to use for first year Greek. If you don’t use it, you should.) Unfortunately, since the word is so often listed incorrectly, students will also need to learn that ζῶ will usually be found under or listed as ζάω.
Carl Conrad tells me that this word appears in Homer as ζώω, but that this
was never a real form, rather it’s a case of what is called “diectasis” … : it is the artificial re-lengthening of a contraction into a double vowel for the sake of the meter (older grammarians used the Latin phrase
“metri gratia” — it’s sort of like someone exclaiming “boo-oo-oo-ti-ful!”
Thanks to Carl for helping me sort through this morphological puzzle yesterday.