Alexander is said to have destroyed the books of the Zoroastrians, Hermippus of Alexandria is said to have read them.[38] But whether on the revival of the Persian religion and literature, that is to say 500 years after Alexander, the works of Zoroaster were collected and restored from extant MSS., or from oral tradition, must remain uncertain, and the disturbed state of the phonetic system would rather lead us to suppose a long-continued influence of oral tradition. What the Zend language might become, if entrusted to the guardianship of memory alone, unassisted by grammatical study and archæological research, may be seen at the present day, when some of the Parsis, who are unable either to read or write, still mutter hymns and prayers in their temples, which, though to them mere sound, disclose to the experienced ear of an European scholar the time-hallowed accents of Zarathustra's speech. [Footnote 37: Spiegel states the results of his last researches into the language of the different parts of the Avesta in the following words: 'We are now prepared to attempt an arrangement of the different portions of the Zend-Avesta in the order of their antiquity. First, we place the second part of the Ya_s_na, as separated in respect to the language of the Zend-Avesta, yet not composed by Zoroaster himself, since he is named in the third person; and indeed everything intimates that neither he nor his disciple Gushtasp was alive. The second place must unquestionably be assigned to the Vendidad. I do not believe that the book was originally composed as it now stands: it has suffered both earlier and later interpolations; still, its present form may be traced to a considerable antiquity. The antiquity of the work is proved by its contents, which distinctly show that the sacred literature was not yet completed. 'The case is different with the writings of the last period, among which I reckon the first part of the Ya_s_na, and the whole of the Yeshts. Among these a theological character is unmistakeable, the separate divinities having their attributes and titles dogmatically fixed. 'Altogether, it is interesting to trace the progress of religion in Parsi writings. It is a significant fact, that in the oldest, that is to say, the second part of the Ya_s_na, nothing is fixed in the doctrine regarding God. In the writings of the second period, that is in the Vendidad, we trace the advance to a theological, and, in its way, mild and scientific system. Out of this, in the last place, there springs the stern and intolerant religion of the Sassanian epoch.'--From the Rev. J. Murray Mitchell's Translation.] [Footnote 38: 'Lectures on the Science of Language,' First Series, p. 95.]