The epithet "four-eyed" may possibly contain a tentative coagulation of the two dogs in one. The capacity of the two dogs to see both by day (the sun) and by night (the moon) may have given the myth a slight start into the direction of the two-headed Greek Cerberus. But there is the alternate possibility that four-eyed is but a figure of speech for "sharp-sighted," especially as I have shown elsewhere that the parallel expression "to run with four feet" is a Vedic figure of speech for "swift of foot."[17] Certainly the god Agni, "Fire," is once in the _Rig-Veda_ (i. 31. 13) called "four-eyed," which can only mean "sharp-sighted."
The two dogs of Yama derive their proper names from their color epithets. The passages above make it clear that Çy[=a]ma (rarely Çy[=a]va), "the black," is the moon dog, and that Çabala, "the spotted, or brindled," is the sun dog. In one early passage (_Rig-Veda_, x. 14. 10) both dogs are named in the dual as Çabal[=a]u. But for a certain Vedic usage one might think that "the two spotted ones" was their earliest designation. The usage referred to is the eliptic dual: a close or natural pair, each member of which suggests the other, may be expressed through the dual of one of them, as when either _m[=a]tar[=a]u_ or _pitar[=a]u_, literally, "the two mothers," and "the two fathers," each mean "the two parents."[18] From this we may conclude that Çabal[=a]u means really Çabala and Çy[=a]ma, and not the two Çabalas, that is, "the two spotted ones."
More than a hundred years ago the Anglo-Indian Wilford, in the _Asiatick Researches_, iii., page 409, wrote: "Yama, the regent of hell, has two dogs, according to the Pur[=a]nas; one of them named Cerbura, or varied; the other Syama, or black." He then compares Cerbura with Kerberos, of course. The form Cerbura he obtained from his consulting Pandit, who explained the name Çabala by the Sanskrit word _karbura_ "variegated," a regular gloss of the Hindu scholiasts.
About fifty years later a number of distinguished scholars of the past generation, Max Müller, Albrecht Weber, and Theodor Benfey, compared the word Çabala with Greek [Greek: Kerberos] (rarely [Greek: Kerbelos]), but, since then, this identification has been assailed in numerous quarters with some degree of heat, because it suffers from a slight phonetic difficulty. One need but remember the swift changes which the name of Apollo passes through in the mouths of the Greeks--[Greek: Apollôn], [Greek: Apellôn], [Greek: Appellôn], [Greek: Apeilôn], [Greek: Aploun][19]--to realize that it is useless to demand strict phonetic conservation of mythic proper names. The nominative Çabalas, translated sound for sound into Greek, yields [Greek: Keberos], [Greek: Kebelos]; _vice versa_, [Greek: Kerberos?] translated sound for sound into Vedic Sanskrit yields Çalbalas, or perhaps, dialectically, Çabbalas. It is a sober view that considers it rather surprising that the two languages have not manipulated their respective versions of the word so as to increase still further the phonetic distance between them. Certainly the burden is now to prove that the identification is to be rejected, and, I think, that the soundest linguistic science will refuse ultimately to consider the phonetic discrepancy between the two words as a matter of serious import.