¹ Oberhelman, Steven M. and Hall, Ralph G., ‘A new statistical analysis of accentual prose rhythms in imperial Latin authors’, CP 79 (1984), 114ffGoogle Scholar., and ‘Meter in accentual clausulae of late imperial Latin prose’, CP 80 (1985), 214ffGoogle Scholar.

² Hall, and Oberhelman, , ‘Rhythmical clausulae in the Codex Theodosianum and the Leges novellae ad Theodosianum pertinentes’, CQ 35 (1985), 201ffCrossRefGoogle Scholar.; ‘Internal clausulae in late Latin prose as evidence for the displacement of metre by word-stress’, CQ 36 (1986), 508ffCrossRefGoogle Scholar.; ‘Rhythmical clausulae in the letters of Augustine as a reflection of rhetorical and cultural goals’, Augustiniana, forthcoming.

³ Oberhelman, , ‘Clausular rhythms and the authenticity of the Pseudo-Sallustiana’, Latomus 45 (1986), 383ffGoogle Scholar., and 'The provenance of the prose style of Ammianus Marcellinus. QUCC, forthcoming.

⁴ For a bibliography of the cursus see Oberhelman, and Hall, , CP 79 (1984), 115 n. 5Google Scholar.

⁵ For the cursus mixtus see Oberhelman, and Hall, , CP 80 (1985), 216 n. 11Google Scholar.

⁶ I use the term miscellanei to denote any accentual scheme other than the three standard forms and th etrispondaicus. These would include the medius (one unaccented syllable between two accents with two unaccented syllables after the last accent: repúgnant spíritus); dispondaicus (two word-accents with one unaccented syllable after each accent: actiónes sólvat); and the dispondeus dactylicus (three unaccented syllables between two word-accents and two unaccented syllables after the last accent: firme constitútere and necessitátibus subícere). These categories correspond to the Ciceronian metrical clausulae of trochee-cretic, ditrochee and paeon-cretic/tribrach.

⁷ See the table in my forthcoming paper ‘The cursus in imperial Latin prose: a reconsideration of methodology’ in CP 83 (1988)Google Scholar.

⁸ See the tables in Janson, Tore, Prose Rhythm in Medieval Latin from the 9th to the 13th Century (Stockholm, 1975)Google Scholar, and Lindholm, G., Studien zum mittellateinischen Prosarhythmus: Seine Entwicklung und sein Abklingen in der Briefliteratur Italiens (Stockholm, 1963)Google Scholar. Janson, 104f., summarizes his data by demonstrating the existence of two different traditions of the cursus in the early Middle Ages, each tradition corresponding to one of the two traditions I have demonstrated for late Latin: the one with only a few forms (mainly the standard forms); the other rich in variants and by-forms. This double tradition may have survived in some manner from antiquity throughout the Dark Ages until the twelfth century, when the more restricted system prevailed.

⁹ Cf. Wilkinson, L. P., Golden Latin Artistry (Cambridge, 1970), 158fGoogle Scholar.; see also Oberhelman, and Hall, , CP 80 (1985), 216 with note 12Google Scholar.

¹⁰ For these texts see Oberhelman, and Hall, , CP 79 (1984), 118ffGoogle Scholar. Cicero's orations, which were the third of our control texts, are omitted from consideration here for obvious reasons.

¹¹ To verify the data collected from the medieval writers I also sampled 600 trispondaicus clausulae from Descartes and Calvin for metrical patterns. The percentages were nearly identical in every category. While consistency of control texts would have been achieved by sampling Descartes and Calvin in all accentual typologies, it was important to investigate occurrences of metre in at least one situation where accentual rhythms were deliberately employed.

¹² By ‘standard’ metrical form in this paragraph I mean the four standard forms (creticspondee, ditrochee, dicretic and cretic-tribrach) and the forms under investigation here: paeon-cretic/tribrach, trochee-cretic and paeon-spondee.

¹³ Oberhelman, and Hall, , CP 79 (1984), 122ffGoogle Scholar., and 80 (1985), 222f.; Oberhelman, , CP 83 (1988)Google Scholar.

¹⁴ The percentages of the totals with subscripts ‘e’ and ‘st’ denote the frequency of standard forms, either with or without coincidence of accent and ictus respectively, in relation to each accentual typology. Thus the percentage of T _t for Minucius signifies that 17·5% of all the clausulae were trispondaicus; on the other hand, the percentage for T _te in the next column means that 57·3% of all the trispondaicus clausulae contained a paeon-spondee, while the percentage for T _tst, says that 100% of the trispondaicus contained one of the standard metrical forms.

¹⁵ In cases of large corpora, such as those of Lactantius, Jerome, Ambrose and Augustine, three or four works were randomly selected. Of the Panegyrici latini I arbitrarily chose II, III, X, XI and XII.

¹⁶ Wilkinson, , op. cit. 141Google Scholar.

¹⁷ Ibid. 157.

¹⁸ See the table in Oberhelman, and Hall, , CP 80 (1985), 223Google Scholar.

¹⁹ In my third CP paper (83 [1988])Google Scholar I have shown with certainty that all these works are accentual.

²⁰ Here I use ‘standard’ in the same sense as in n. 12 above.

²¹ Oberhelman, and Hall, , CP 80 (1985), 224Google Scholar, and CQ 36 (1986), 519ffGoogle Scholar.

²² Oberhelman, and Hall, , CP 79 (1984), 130Google Scholar, and CP 80 (1985), 226fGoogle Scholar.; and Hall, and Oberhelman, , CQ 36 (1986), 525fCrossRefGoogle Scholar.

²³ See Hall, and Oberhelman, , CQ 35 (1985), 201 n. 4CrossRefGoogle Scholar.

²⁴ Janson, , op. cit. 104Google Scholar.

²⁵ See the tables in Oberhelman, and Hall, , CP 80 (1985)Google Scholar, and Hall, and Oberhelman, , CQ 36 (1986)CrossRefGoogle Scholar.

²⁶ This is elaborated at great length in the Augustine paper cited above in note 2.

²⁷ This is the important conclusion of my paper in CP 83 (1988)Google Scholar.

²⁸ It now becomes even mor e clear that simple statistical counts and single hypothesis-testing will not provide reliable answers to complex questions concerning prose rhythms. AH typologies, accentual and metrical, must be tested individually, and not lumped into one statistical category in an ‘either-or’ test. Thus the methodologies of the first two CP papers, while significant for their efforts in placing the study of late Latin prose rhythms in a new scientific light, are too weak in their ability to gauge sensitive differences in prose style to be of wide application.

²⁹ Thecursus system of Ammianus is not included in this survey; as I have demonstrated in my forthcoming paper in QUCC, Ammianus borrowed a Greek practice.

³⁰ This generalization is based of course only on what works I have studied thus far. I have also omitted from consideration here certain texts that I have previously labelled ‘cursus’: spurious works of uncertain date (parts of the Apuleian corpus) and the grammarians Donatus and Servius. Michael Reeve in private correspondence has shown to my satisfaction that the grammarians proper pose such special problems as to make statistics derived from their work negligible.

³¹ See the table in Oberhelman, and Hall, , CP 80 (1985)Google Scholar. The table with discussions in text in Oberhelman, , CP 83 (1988)Google Scholar, is important, as some decisions about rhythm in the 1985 table are revised.

³² I owe much of this paragraph to Robert Kaster. My deepest thanks to Professor Kaster for his continued personal and professional help in clarifying my thoughts on ancient prose rhythm; his acute suggestions and criticisms have always proved invaluable and profitable.

³³ Hall and I made much of the ‘mass audience’ and ‘African untrained laity’ in the first two CP papers and our studies in CQ for 1985 and 1986.

³⁴ Cicero, , Or. 213–14Google Scholar; cf. 195.

³⁵ See Hörandner, W., Der Prosarhythmus in der rhetorischen Literatur der Byzantiner (Wien, 1981)Google Scholar. There is no tradition of the cursus mixtus in Greek.

³⁶ Dewing, H., “The origin of accentual prose rhythm in Greek”, AJP 31 (1910), 323fGoogle Scholar.

³⁷ If this scenario is correct, irony abounds. The Greek cursus, first evident in Themistius and Himerius, seems to have been taken from the Latin cursus mixtus. But if the late fourth-century Christian writers borrowed the concept of accentual-only clausulae from the Greek patristic authors, then the Greek cursus was the inspiration for an alternative to the cursus mixtus. Thus, we have this cycle of influence over the course of about six hundred years: Greek metrical system → Latin metrical system → Latin cursus mixtus → Greek cursus → Latin cursus.

³⁸ Comm. in Is. 6.1; Ep. 1.1 and 53.8.

³⁹ Cf. De trinitate 15.27.48 and Retract. 2.32.2.

⁴⁰ See 1.6.24; 5.24.84; 5.24.92; 6.1.1. For the cursus of this work see my 1988 CP paper. Augustine's captivation with Ambrose's sermons in Milan is commonplace, as is his pleasure at Ambrose's exegeses.

⁴¹ It is in the De doct. christ. that Augustine advises Christian professors to imitate the style of Paul and other scripture or of the best writers like Ambrose.

⁴² Aili, H., The Prose Rhythm of Sallust and Livy (Stockholm, 1979)Google Scholar.

⁴³ See Hall, and Oberhelman, , CQ 36 (1986)CrossRefGoogle Scholar, table with extensive discussion passim.

⁴⁴ Stephens, L., ‘Some remarks on prose rhythm’, Helios 15.1 (1988)Google Scholar, forthcoming.

⁴⁵ Sheerin, D., ‘Medieval Latin: In media Latinitate’, Helios 14.2 (1987)Google Scholar, forthcoming.

⁴⁶ This paper has benefited greatly from the comments of the anonymous reader of CQ.