a word which they understood in the sense in which it was meant to be understood.¹ Again, it is unlikely that Muhammad turned to Ethiopic for a word to denote "angel". It is more likely that the word mal'ak, with its characteristically Ethiopic form of plural mala'ika, had already crept in from Abyssinia and was known to the people of Mecca. Grimme has cited a number of words in the Qur'an which he regards as of South Arabian origin. With the exception of rahman, which for a time Muhammad used as a name for God, and which I would rather regard as of Aramaic origin, none of these are of much religious importance. The investigation of the language of the South Arabian inscriptions is not as yet far enough advanced for us to found much upon it. But these words which I have mentioned as derived from Aramaic and Ethiopic—the list is by no means complete—show whence the language of religion came. The probability is that Muhammad found this language largely ready to his hand.

With the language must have come also some knowledge of the ideas. What was meant by a prophet, a holy book, revelation, prayer, and praise, cannot have been entirely unknown to the Arabs. A more thorough search of the old

poetry would probably disclose even to critical eyes, a wider range of religious ideas referred to as already known if not assimilated. Professor Horovitz, in a paper which I have not been able to procure, has, I understand, shown that nearly all the details of Muhammad's description of Paradise are to be found referred to by older and contemporary poets. If he has been able to prove that, it will add force to the contention that there was an atmosphere of Jewish and Christian ideas pervading Arabia at that time.

If from the ancient poetry of Arabia we cannot draw any conclusive evidence that these ideas had made any deep impression upon the Arab spirit, we can at least draw from it the negative conclusion that the old pagan religion had lost its hold. References to it are just as rare as to Christianity. Names of heathen deities occur in oaths and similar phrases. But even more commonly the name Allah is used. Making all allowance for the possibility that, as the poetry was not collected till the second century of Islam, the name Allah may sometimes have been substituted for that of some other deity, it is certain that Allah is frequently used as the name of God in pre-Islamic poetry. Wellhausen has remarked that, within his own territory, the king is sufficiently designated by his title, his proper name being unnecessary ; and that the custom of thus referring to the tribal deity, combined with the wandering habits of the tribes, may have opened the way for the general idea of God, which is what Allah, if it be Arabic, denotes, and thus for the conception of one