That Greek θ does not legitimately represent a Sanskrit, Latin, Slavonic, and Celtic _d_ is a fact that ought never to have been overlooked by comparative philologists, and nothing could be more useful than the strong protest entered by Windischmann, Schleicher, Curtius, and others, against the favorite identification of Sk. +deva+, _deus_, and θεός. Considering it as one of the first duties, in all etymological researches, that we should pay implicit obedience to phonetic laws, I have never, so far as I remember, quoted θεός as identical with _deus_, together with the other derivatives of the root +div+, such as +Dyaus+, Ζεύς, _Jupiter_, +deva+, Lith. _deva-s_, Irish _día_. But with all due respect for phonetic laws, I have never in my own heart doubted that θεός belonged to the same cluster of words which the early Aryans employed to express the brightness of the sky and of the day, and which helped them to utter their first conception of a god of the bright sky (+Dyaus+), of bright beings in heaven, as opposed to the powers of night and darkness and winter (+deva+), and, lastly, of deity in the abstract.[7] I have never become an atheist; and though I did not undervalue the powerful arguments advanced against the identity of _deus_ and θεός, I thought that other arguments also possessed their value, and could not be ignored with impunity. If, with our eyes shut, we submit to the dictates of phonetic laws, we are forced to believe that while the Greeks shared with the Hindus, the Italians, and Germans the name for the bright god of the sky Zeus, +Dyaus+, _Jovis_, _Zio_, and while they again shared with them such derivatives as δῖος, heavenly, Sk. +divyas+, they threw away the intermediate old Aryan word for god, +deva+, _deus_, and formed a new one from a different root, but agreeing with the word which they had rejected in all letters but one. I suppose that even the strongest supporters of the atheistic theory would have accepted δεός, if it existed in Greek, as a correlative of +deva+ and _deus_; and I ask, would it not be an almost incredible coincidence, if the Greeks, after giving up the common Aryan word, which would have been δοιϝός or δειϝός or δεϝός, had coined a new word for god from a different root, yet coming so near to δεϝός as θεϝός? These internal difficulties seem to me nearly as great as the external: at all events it would not be right to attempt to extenuate either.