AI, v. Haw., to eat; s. food, vegetable food, in distinction from ia, meat; ai-na, for ai-ana, eating, means of eating, fruits of the land; hence land, field, country. New Zeal., kai, to eat; kainga, food, meal, home, residence, country. Tong., kai, to eat. Sam., 'ai, to eat; ainga, family, kindred. Marqu., kaika, kainga, food, meal. Tagal., cain, to eat.

Zend., gaya, life; gaetha, the world; gava, land, country. Vedic, gaya, house, family (A. Pictet). Sanskr., ghâsa, food; ghas, devour. Greek, αία, γαια, γη, different forms occurring in Homer, land, country, cultivated land; γειος, indigenous; γειτων, a neighbour; ήια, provisions for a journey. Goth., gawi, gauja, country, region. Germ., gau. Lat., ganea, eating-house; ganeo, glutton. Lith., goyas; Ant. Slav. and Russ., gai, 'past-rage', nemus. Polish, gay, id.

Mr. A. Pictet, in his 'Les Origines Indo-Europ.', vol. II. p. 15, says that the Vedic and Zend gay 'n'ont surement aucun rapport avec le grec γαια.' This assertion evokes a doubt, inasmuch as, as late a in Homer's time, two other dialectical variations of this word existed in the Greek, viz. αία and δα or δη, in δη-μητηρ, contracted from some ancient form in δαια, as γη and γα, from γαια. As neither of these can be supposed to be derived from, or to be a phonetic corruption of, the other, it seems to me that they must have come down abreast from primeval times, thus indicating that the original root was differently pronounced by various sections of the still united Aryan stock; and I believe that this root, in its archaic forms, still survives in the Polynesian ai and kai, to eat.

The Sanskrit go, land, the earth, from which Benfey derives a hypthetical gavyd and a Greek γαfια - by elimination γαια - is probably itself a contraction from the Vedic and Zend gaya, as the Greek γη and γα, as the ancient Saxon gâ and gô, pagus, regio, and the ancient Slav. gai, nemus, are contractions from derivations of that ancient root still found in Polynesia. The above derivatives in sound and sense certainly refer themselves better to some ancient ai of kai, food, the fruits of the forest or the roots of the field, than to the Sanskrit go, bull, cow, cattle; for the Aryan family undoubtedly had one or more names for eating and for food before its various divisions applied themselves to the herding of cattle. The Sanskrit ghas, ghâsa, the Latin ganea, ganeo, point strongly to the underlying original sense of eating and food.

According to Professor A. H. Sayce, in 'Introduction to the Science of Language', vol ii. p. 19, it is probable that the Latin edere, to eat, is a compound word = e-dere, like ab-dere, con-dere, cre-dere, and others, thus leaving e as the root. How far that e may have been a dialectical variant or a phonetic decay of an older form more nearly allied to the Polynesian ai, kai, I leave to abler philologists to determine. (Fornander)