δρώδης, βορβορώδης (Plato), ἐργώδης (Plato). This group, however, is too small to be of any significance for our study.
We may note in passing that adjectives in -ώδης are unusually frequent in Thucydides[1] and Theophrastus.[2]
In order to form a proper estimate of an author's vocabulary we should examine not only the individual words he uses, but also the kinds of words. In comparing Menander's diction with the Koine we should study his use of certain types of words known to have been favorites in Hellenistic Greek. Such a test involves a statistical comparison on the one hand with certain writers of classical Attic, and on the other with typical writers of the Koine. For the former I have selected Aristophanes, Thucydides, Plato and Demosthenes; for the latter, Polybius and Plutarch.
Since a careful estimate of the amount of the extant fragments of Menander gives a total of c. 3393 full lines, or c. 113 pages of 30 lines each, an equal amount has been taken from each author with whom a comparison is made.[3] The selections of passages in the test authors were made almost at random, care being taken, however, to avoid any works that might furnish misleading results: e.g., the comedies of Aristophanes aimed at the sophists were not taken, because of the great frequency of -ικός words in those plays. A page of 30 Teubner lines was taken as standard. It has not been thought worth while to attempt to equalize the quantitative discrepancy between prose and verse.
- ↑ Wolcott, New Words in Thucydides. Trans. Am. Phil. Assn. 29 (1898) p. 154.
- ↑ Hindenlang, Sprachliche Untersuchungen zu Theophrasts botanischen Schriften, diss. Strassburg, 1910 (Diss. phil. Argent. sel. XIV 2) p. 188.
- ↑ At first, the average number of examples of each formation was made the basis of comparison, by obtaining the total number of such words used by each author and dividing by the number of hundreds of pages. This method of testing has been used frequently, and is perfectly accurate in many kinds of investigation: e. g., when all the examples of any phenomenon are taken into account. But after a good deal of deliberation it was decided that in this kind of work such a system is incorrect, inasmuch as an increase in the bulk of an author does not bring with it a corresponding increase of different words of a given type. Thus, the more prolific a writer was, the smaller his proportion per 100 pages of different words of a certain formation is likely to be.
