We know nothing of his philosophical opinions, except that he held the soul to be a harmony of the body (Cic. Tusc. Disp. 1.10, 18; Lact. Instit. 7.13, de Opif Dei, 100.16), a doctrine which had been already discussed by Plato (in the Phaedo) and combated by Aristotle. (De An. 1.4.)

It is only in his character as a musician that Aristoxenus appears to have deserved and acquired a reputation for real excellence; and no considerable remains of his works have come down to us except three books of ἁρμονικὰ στοιχεῖα, or rather, as their contents seem to shew, fragments of two or three separate musical treatises. (See Burney, Hist. of Music, vol. i. p. 442.) They contain less actual information on the theory of Greek music than the later treatises ascribed to Euclid, Aristeides Quintilianus, and others; but they are interesting from their antiquity, and valuable for their criticisms on the music of the times to which they belong. Aristoxenus, at least if we may trust his own account, was the first to attempt a complete and systematic exposition of the subject; and he aimed at introducing not only a more scientific knowledge, but also a more refined and intellectual taste than that which prevailed among his contemporaries, whom he accuses of cultivating only that kind of music which was capable of sweetness. (Aristox. p. 23, ed. Meibom.) He became the founder of a sect or school of musicians, called, after him, Aristoxeneans, who were opposed to the Pythagoreans on the question whether reason or sense should furnish the principles of musical science and the criterion of the truth of its propositions. Pythagoras had discovered the connexion between musical intervals and numerical ratios; and it had been found that the principal concords were defined by simple ratios which were either superparticular (of the form n+1/n) or multiple (of the form n/1). From this fact, he or his followers inferred, that no interval could be consonant which was defined by a ratio of a different kind; and hence they were obliged to maintain (contrary to the evidence of the senses), that such intervals as the octave and fourth (the eleventh), for example, were dissonant. Aristoxenus justly blamed them for their contempt of facts, but went into the opposite extreme of allowing too much authority to the decisions of the ear, though without denying the existence of a certain truth in the arithmetical theory (p. 33). He maintains, for instance, not only that every consonant interval added to the octave produces another consonance, which is true; but also that the fourth is equal to two tones and a half (p. 56), the falsity of which proposition is not directly apparent to the ear, but indirectly would become evident by means of the very experiment which he suggests for the confirmation of it. (See Porphyr. Comm. in Ptol. Harm. in Wallis, Op. vol. iii. p. 211, and Wallis's appendix, pp. 159, 169; Burney, vol. i. chap. v.; Theon Smyrn. p. 83, ed. Bulliald. and note, p. 202.)

The titles of a good many other works of Aristoxenus have been collected from various sources by Meursius and others. (See Fabric. Bibl. Graec. vol. ii. p. 257; Clinton, F. H. vol. ii. appendix, 100.12.) Among them are lives of Pythagoras, Archytas, Socrates, Plato, and other distinguished persons; and several treatises on subjects connected with music, including one Περὶ Τραγικῆς Ὀρχήσεως, and one Περὶ Αὐλῶν Τρήσεως.