________________ reichen, wenn noch weitere individuelle Zuge der Systeme von Ve und vor allem Tr herausgearbeitet werden konnten. Insbesondere scheint die in Va und Tr vorliegende Gestalt der Lehre von der Vijnaptimatrata, wo ich bei anderer Gelegenheit zu zeigen hoffe, nicht nur bezuglich der Basis, auf der sie entwickelt wird, sondern auch in anderer Hinsicht von der Auffassung des klassischen Yogacara abzuweichen. Ich will daher dem endgultigen Ergebnis einer grundlichen Untersuchung aller unter dem Namen Vasubandhu uberlieferten Yogacarawerke hier nicht vorgreifen. Summary A closer examination of Vasubandhus Vinsatika shows that in this work the Vijnaptimatrata is not developed on the basis of the eightfold complex of mental series taught by the Yogacarins (Alayavijnana, Manas etc., all of which may simultaneously arise) but on the basis of the "one-layered" mental series of the Sautrantikas. This is clear from the following facts: 1. the container of the residues (vasana) of actions is not called "Alayavijnana", but is called merely "mental series" (vijnana-santanah); 2. the Sautrantika-term "(vijnana-) santanaparinama-(visesah)"; 3. the conception of sense-faculties espoused by Vimsatika; 4. the synonymity of "cittam", "manah", "vijnanam" and "vijnaptih"; 5. the distinction of sense-knowledge as "perception" (pratyaksam) from the succeeding Manovijnana as "vikalpah". - Such an idealism developed on the basis of the "one-layered" mental series of the Sautrantikas is to be met with also in Dignaga's Alambanapariksa and in Dharmakirti. - Vasubandhu's Trimsika, as against Vimsatika, advocates the eightfold complex of mental series taught by the Yogacarins; still unmistakable traces of the author's Sautrantika-past are noticeable in the work. - On account of these Sautrantika-elements Vimsatika and Trimsika may be attributed to the author of Abhidharmakosah. At the same time it appears that such elements are wanting in the commentaries on the works of Maitreyanatha and Asanga attributed to Vasubandhu, a fact which seems to support the "two-Vasubandhuhypothesis" put forward by Frauwallner. 136
