A. INTRODUCTION

We have outgrown the tendency of which Cornford complained in 1907, to write the history of Greek philosophy ‘as if Thales had suddenly dropped from the sky, and as he bumped the earth ejaculated: “Everything must be made of water!”’ It was a sign of the changed outlook that, in preparing the fifth edition of Diels's Fragments of the Presocratic Philosophers in 1934, Walther Kranz put into effect a suggestion made by Diels himself in his preface to the fourth edition, namely to place at the beginning a chapter of extracts from early cosmological, astronomical and gnomic writings, which in the previous editions had been relegated to an appendix. This raises the difficult question whether the present work should follow the same plan. A strong reason against doing so is the endlessly disputed authenticity and date of the records of this ‘ pre-philosophicaT tradition, which are for the most part preserved only as quotations in writers of a much later period. We may be sure that Hesiod's Theogony (the only extant work of its type) preceded the Milesian philosophers, but when we come to the fragments of Orphic cosmogony, or of the Theogony of Epimenides, it is difficult to be sure whether they may be reckoned as an influence on the Milesians or, on the contrary, as owing something to the Milesians themselves. Thus Kern saw in the fragments of Epimenides the impact of Anaximenes, and Rohde maintained of the ‘Rhapsodic Theogony’ attributed to Orpheus that ‘in the very few passages in which a real coincidence exists between the Rhapsodies and Pherekydes, Herakleitos, Parmenides or Empedokles, the poet of the Rhapsodies is the borrower not the creditor’. Recent opinion is on the whole inclined to assign the main outlines of the world-view expressed in Orphic theogonical and cosmogonical fragments to the sixth century b.c., but doubts remain. All things considered, it seems best to proceed at once to an examination of the remains of those who are usually, and not without reason, called the first of the philosophers. The necessary reference to their actual or possible predecessors can be made where this examination demands it.