Let us take an instance. If we open our best Greek dictionaries, we find that the Greek αὐγή, light, splendor, is compared with the German word for eye, _Auge_. No doubt every letter in the two words is the same, and the meaning of the Greek word could easily be supposed to have been specialized or localized in German. Sophocles (“Aj.” 70) speaks of ὀμμάτων αὐγαί, the lights of the eyes, and Euripides (“Andr.” 1180) uses αὐγαί by itself for eyes, like the Latin _lumina_. The verb αὐγαζω, too, is used in Greek in the sense of seeing or viewing. Why, then, it was asked, should αὐγή not be referred to the same source as the German _Auge_, and why should not both be traced back to the same root that yielded the Latin _oc-ulus_? As long as we trust to our ears, or to what is complacently called common sense, it would seem mere fastidiousness to reject so evident an etymology. But as soon as we know the real chemistry of vowels and consonants, we shrink instinctly from such combinations. If a German word has the same sound as a Greek word, the two words cannot be the same, unless we ignore that independent process of phonetic growth which made Greek Greek, and German German. Whenever we find in Greek a media, a _g_, we expect in Gothic the corresponding tenuis. Thus the root _gan_, which we have in Greek γιγνώσκω, is in Gothic _kann_. The Greek γόνυ, Lat. _genu_, is in Gothic _kniu_. If, therefore, αὐγή existed in Gothic it would be _auko_, and not _augo_. Secondly, the diphthong _au_ in _augo_ would be different from the Greek diphthong. Grimm supposed that the Gothic _augo_ came from the same etymon which yields the Latin _oc-ulus_, the Sanskrit +ak-sh-i+, eye, the Greek ὄσσε for ὄκι-ε, and likewise the Greek stem ὀπ in ὄπ-ωπ-α, ὄμμα, and ὀφ-θ-αλμός. It is true that the short radical vowel _a_ in Sanskrit, _o_ in Greek, _u_ in Latin, sinks down to _u_ in Gothic, and it is equally true, as Grimm has shown, that, according to a phonetic law peculiar to Gothic, _u_ before _h_ and _r_ is changed to _aú_. Grimm, therefore, takes the Gothic _aúgô_ for *_aúhô_, and this for *_uhô_, which, as he shows, would be a proper representative in Gothic of the Sanskrit +ak-an+, or +aksh-an+. But here Grimm seems wrong. If the _au_ of _augô_ were this peculiar Gothic _aú_, which represents an original short _a_, changed to _u_, and then raised to a diphthong by the insertion of a short _a_, then that diphthong would be restricted to Gothic; and the other Teutonic dialects would have their own representatives for an original short _a_. But in Anglo-Saxon we find _eáge_, in Old High German _augâ_, both pointing to a labial diphthong, _i.e._ to a radical _u_ raised to _au_.[14]