¹ Leumann-Stolz, §246 : Das Verhältnis von -ērunt -ĕrunt und -ēre ist nicht aufgeklärt. But there is a generally accepted theory of the origin of the three endings which is questioned by the present writer (§47–§51).

² Cf. Goetz and Schoell (1901), preface to volume three : Manuscript A was reexamined, and the testimony of Loewe (used in the large edition) was found to be unreliable.

³ A. Luchs, Quaestiones Metricae; in Vol. I of W. Studemund’s Studien 1-75; Berlin, T873. Lindsay ELV 105, 270-2.

⁴ Cf. §40. b.

⁵ E. H. Sturtevant, Accent and Ictus in Plautus and Terence; in Classical Philology 14.234-44 (1919).

⁶ Leumann-Stolz §163 c (page 188), with references. Cf. also Sturtevant, op. cit. 238-9.

⁷ I. e., in anapaestic meter, or according to Luchs’s Law (§46).

⁸ J. W. Duff, A Literary History of Rome 214; New York, 1928.

⁹ Yet the possibility that Terence also represents a tendency to develop a somewhat artificial literary language, much less pronounced than the same tendency in tragedy and epic, cannot be disregarded. Whether the cause is attributable to speech practices in the language of conservative classes, or to a beginning of an artificial literary language, or to both, the result (the fact that the language of Terence is somewhat more conservative than that of Plautus) is the same so far as the present study is concerned.

¹⁰ For a summary of the evidence which suggests a prehistory of Latin metrics on Italian soil before the time of Livius Andronicus, with particular emphasis on the iambic senarius and the trochaic septenarius, cf. F. Altheim : Die neuesten Forschungen zur Vorgeschichte der römischen Metrik; in Glotta 19. 24-48 (1931).

¹ E. g., to avoid the division between two words, of two short syllables which stand for a long. I admit that complete elision of finals of emphatic monosyllables and of some diphthongs may be doubtful, but such problems are outside the scope of the present study.

² Fleckeisen, Philologus 2.75 (1847), emended the hrst half of this verse to: quom quae in me fecērunt bene, to avoid the form fecĕrunt, because he found no other trisyllabic -ĕrunt perfect in Plautus. I can see no reason why the number of syllables in the word should have any relation to the perfect ending used. Cf. emĕrunt 1981, and fecĕrunt 74.

³ Mss. consuerunt; cf. §36. a.

⁴ Mss. consuere; cf. §36. a.

⁵ Goetz and others adopt meminerint, the reading of B. The jussive subjunctive meminerint fits the context better than the indicative, but the indicative with sarcastic force is possible. Coincidence of word accent and verse ictus favors nieminē’runt over meminerint, since quadrisyllablc perfect forms beginning with three short syllables are usually accented on the first syllable (§41. a).

⁶ Two cretics and two iambs; or, two cretics followed by . Lindsay, according to ELV 293-5, would call this verse a mi uric form of the curtailed cretic tetrameter. The curtailed cretic tetrameter consists of two cretics followed by ; the miuric form has a ‘short for the first long syllable of the final foot . In a number of Lindsay’s examples, I cannot see anything except- combinations of cretic with iambic or trochaic feet : me ire minime Mo. 693, est ut abeam Mo. 706, should certainly be scanned ; sic sine, Simo Ps. 1301 (similar to the end of the verse which is under discussion here) should be scanned , or perhaps . Cf. §40. a.

⁷ Ital. (apud Lindsay) ubi ubi, which is adopted by Leo and by Goetz, requires the scansion fŭĕrunt. B has ubique, probably an error in copying caused by the preceding quiquomque, and not for ubiqu’, as Lindsay suggests. The only apparent reason for changing to ubi ubi is objection to long u in fūērunt, but fū- occurs elsewhere in Plautus ‘. fūit Cap. 555 , fūerim Mi. 1364, fūerint Po. no. There is no other possible instance of fŭĕrunt in Plautus (§39). I follow Lindsay in the reading adopted.

¹ Rev. Phil. 31.230-3.

² References to perfect forms are here given according to the order in which they appear in the List (§9–§11), though Havet refers to play and line.

³ Plautus occasionally permits substitution in bacchiac meters : Lindsay ELY 290.

⁴ A pure iamb in this position is not absolutely required in Plautus : Lindsay ELV 275-6. Havet regards egērunt 67 as corrupt.

⁵ I am not acquainted with this metrical principle, and Havet does not give any other example to support the rule.

⁶ Cf. §18. e, and §19. c.

⁷ For Haveťs remarks on the use of -ere in Terence, cf. § 30.

⁸ 332-41 (on Plautus and Terence).

⁹ 331, 334-5.

¹⁰ Inst. Or. 1.5.42-4. Quintilian himself opposes the theory. Havet did not, as Muller implies, suggest that -ēre was considered a dual at the time of Plautus; Havet believed that Nos. 111, 112, and 113, were non-Plautine, and suggested that -ērunt might have been changed to -ēre after the time of Plautus, by someone who believed that -ēre was a dual form. This is more probable than Muller’s theory, though I believe that the -ēre forms referred to are Plautine.

¹¹ 335-8.

¹² 340.

¹³ 333->4.

¹⁴ Plautinische Forschungen² 292-301; Berlin, 1912.

¹⁵ Leumann-Stolz, §156.

¹⁶ Leumann-Stolz, §155, §246. The loss of final -n, and of -m before consonants, is not a feature of the Latin of Plautus.

¹⁷ Lane-Morgan, §1575 : The third person (of the imperative), and the longer forms of the second person, are used particularly in laws, legal documents, and treaties, and also in impressive general rules or maxims.

¹⁸ Rev. Phil. 31.230.

¹⁹ The inference has been drawn, from Men. 407-12, that this comedy was staged before the death of Hiero, tyrant of Syracuse (270-215 B.C.). Cf. Brix-Niemeyer, 4th ed. of Menaechmi, pp. 5-6 (Leipzig, 1891); or H. N. Fowler, Introd. to edition of Menaechmi, pp. 23-4 (Boston, 1900). However, it is difficult to imply that old ablatives in -d were regularly used in normal colloquial Latin at the time when the play was written.

²⁰ CIL I² 581; but in agro tevrano, probably a titular addition, occurs at the end of the inscription.

²¹ Rev. Phil. 31.231 : Le miles raconte ses exploits sur un ton moins vulgaire, sans doute le ton de la tragédie (il serait même possible qu’il y eût ici parodie d’un vers tragique déterminé).

²² Muller, 335, associates with the use of fuēre the form deico (Po. 474) : forma per ei scripta perraro occurrit et senibus tantum sive personis amplis gravioribusque tribuitur. I consider the manuscript tradition on the spelling of the original eidiphthongs as entirely unreliable; at the time of Plautus, the original diphthongs were probably always written ei. Cf. A.R. Anderson, Trans, of the Amer. Phil. Assoc. 37.73-86 (1906), who says, page 85 : ‘ The ei-readings as a whole, far from going back to Plautus himself, do not go back even to the common archetype of our two families of Mss. ’

²³ But the disappearance of Carthaginian children may not have been as serious a matter to the Roman audience as it was to Hanno.

²⁴ Cf. Muller, 335, note 1 : Hoc quoque stili requisitioris est, qualis deo si ve daemone ‘ in ornatu prologi ’ dignus est?

²⁵ Lane-Morgan, §ı512, §1937, §52159.

²⁶ Cf. §20. b.

²⁷ Havet, Rev. Phil. 31.232.

²⁸ Cf. Mullers conclusion that -ērunt was well established by 200 B.C. (§14).

¹ Real or so-called (cf. §53). For want of any other term that would be clear, I shall regularly use the term ‘ contracted ’ of short forms such as those in -ārunt.

² Perfect stem, in this dissertation, regularly means the first person singular perfect indicative active minus the final -i : e. g., the perfect stem of dare is assumed to be ded-.

¹ The origin of -ēre cannot be explained from any Latin form and must therefore be considered to be primitive.

² Leumann-Stolz, §246; Sommer, 578.

³ F. Conrad : Vers-Ende und Sinnesabschnitt bei Plautus; in Glotta 15.28-44 (19²7)·

⁴ E. Vetter, in Glotta 14.26 note 2 (1925), seems to suggest that -ērunt and -ĕrunt should be explained as separate developments in Latin, because of Faliscan f[if]iqod CIE 8079 : ^‘ zeigt dasz das lat -ēre -ērunt, -ĕrunt, junge lateinische Sonderentwicklung ist, was ja wohl zu erwarten war ‘, Herbig, Indogermanische Forschungen 32.77 (1913), explains fifiqod as an aorist (as Leumann-Stolz, §246, takes the Oscan-Umbrian perfect ending -ens). Vetter thinks that -od is also quite distinct from -ens, and apparently takes fifiqod to be developed from an original perfect form.

⁵ 578.

⁶ Leumann-Stolz. §176.I.B.

⁷ Leumann-Stolz, §244.1.b; Sommer, 562-3. It might, however, be maintained that amā-runt is formed directly after the model of fecē-runt. Similarly, fecis-ti (s) may have had some influence on the formation of amās-ti(s), and perhaps even on that of dormīs-ti(s).

⁸ W. Corssen : Über Aussprache, Vokalismus, und Betonung der Lateinischen Sprache 1². 611-3; Leipzig, 1868.

⁹ There is no evidence in Latin that long vowels were shortened for this reason; cf. Leumann-Stolz, §31.

¹⁰ H. Osthoff : Zur Geschichte des Perfekts 213-4; Strassburg, 1884. 11. Leumann-Stolz, §27. b.

¹¹ Leumann-Stolz, §27. b

¹² W. M. Lindsay : The Latin Language 531-2; Oxford, 1894.

¹³ Leumann-Stolz, §246.

¹⁴ According to Leumann-Stolz, §128, the change to r was probably completed about the middle of the fourth century B. C. It need not be assumed that conservative expressions of ritualistic and formulaic character prevented the spread of -ĕrunt forms in colloquial Latin.

¹⁵ For the loss of -t and cf. Leumann-Stolz, §246.

¹⁶ This method of writing may have originated with the Etruscans. Cf. M. Hammarström, Beiträge zur Geschichte des Etruskischen, Lateinischen, und Griechischen Alphabets 15-20, 31-4; Helsingfors, 1920.

¹⁷ 7.14.15-15.7 K.

¹⁸ Cf. §35. There are too few examples tp support an assumption that Plautus avoided -āvē’runt, but it seems significant that a form in -āvē’runt never occurs.

¹⁹ Cf.·§6, and §23.

²⁰ I object to dormĭĕrunt for dormīrunt no, not primarily because of the ending -ĕrunt, but also because of the preceding -ĭ-; cf. §36. b.

²¹ Muller, 335-6 (with references). There may be one -ris form in Terence, Hecyra 317; this is discussed by Muller, 340. note 1.

²² I am not nclined to agree with Lindsay (ELY 13) that Plautus aimed at reproducing the tones of ‘everyday (educated) speech’. There seems to be no basis for assuming that Plautus was aware of any distinction such as that implied by the term ‘ educated Presumably the only notable distinctions in the speech at Rome in the time of Plautus were those attributable to dialectic influences, or to formal language of conservative or archaic character (as illustrated in the use of -ēre, §17- §21). A language may be quite regular, with few linguistic changes in progress at a given time, before there is any grammar or formal education in the language. The Sanskrit in the Rig-Veda, the Greek of Homer, and the English of Chaucer, are all essentially regular, but it is hardly conceivable that grammars or formal education in those languages ever existed before those works were written. I should therefore prefer to characterize the language of Plautus as ‘ normal ’ colloquia Latin.

¹ W. Meyer-Lübke : Grammatik der Romanischen Sprachen 2.299; Leipzig, 1894.

² 340-89·

³ ·H. Hagendahl : Die Perfektformen auf -ere und -erunt (ein Beitrag zur Technik der spätlateinischen Kunstprosa); Uppsala, 1923.

⁴ W. Baehrens, Philologische Wochenschrift 44.702-3 (1924).

⁵ Cf. E. Wölfflin, Archiv fūr Lateinische Lexikographie und Grammatik 22.166 (1902) : Die Endung -ēre ist nicht vulgär, sondern sie wurde von den Epikern metri causa begūnstigt, und floss durch Ennius den Historikern zu.

⁶ Except in the sections which deal with the usage in Phaedrus (§79-§82), -ēre will indicate both the elided and unelided forms of the ending throughout this chapter.

⁷ 354-61,363-8.

⁸ R. B. Steele : The Endings -ere and -erunt in Dactylic Hexameter; American Journal of Philology 32.328-32 (1911).

⁹ The -ere ending is frequent in a quotation given by Gellius 3.7.19; further details are given by Muller, 346-8.

¹⁰ Muller, 349-54.

¹¹ Muller, 377-89, uses the statistics of E. B. Lease : Livy’s Use of -arunt, -emut, and -ere; American Journal of Philology 24.408-22 (1903).

¹² Muller, 381-2.

¹³ Sommer, 58ο; Muller, 376-7.

¹⁴ Muller, 376-7.

¹⁵ Sommer, 580.

¹⁶ Orator 47.157 ; writing of verba plena et imminuta : Nec vero reprehenderim ‘scripsere alii rem’ [et] scripserunt esse verius sentio, sed consuetudini auribus indulgenti libenter obsequor. Cf. Brutus 19.76, about Ennius : ‘Scripsere,’ inquit, ‘alii rem vorsibus’ — et luculente quidem scripserunt, etiam si minus quam tu polite. Cicero himself uses scripserunt, immediately after quoting the form in -ēre from Ennius.

¹⁷ E. Löfstedt : Philologischer Kommentar zur Peregrinatio Aetheriae 36-9; Uppsala, 1911.

¹⁸ E. Wölfflin, Archiv fūr Lateinische Lexikographie und Grammatik 14.478 (1906). .

¹⁹ Op.cit. 39; W. Baehrens, Phil. Woch. 24.702-5, with special reference to Hagendahl (cf. note 3).

²⁰ R. Klussmann : Emeudationes Frornoniae 77; Berlin, 1874.

²¹ Löfstedt, op. cit. 36.

²² Op. cit. 705.

²³ Cf. note 16.

²⁴ Inst. Or. 1.5.42 : Quanquam fuerunt, qui nobis quoque adicerent dualem ‘ scripsere ‘ legere ’; quod evitandae asperitatis gratia mollitum est, ut apud veteres pro ‘ male mereris ‘male merere’.

²⁵ 368-74; the conclusion mentioned here appears on page 373.

²⁶ 354-6.

²⁷ 337, note 2·

²⁸ I have used the text of A. Brenot (Phèdre, Fables; in the Assn. Guillaume Budé series; Paris, 1924). Forms which appear in the so-called ‘Appendix’ are also included, since the use of the perfect endings in these Fables does not differ from the usage in the poems usually attributed to Phaedrus. With regard to -enini forms which appear before a final iambic word, and forms of verbs whose perfect stem consists of two short syllables, I have followed the procedure indicated in §3 and §41. a.

²⁹ 1.20.3, 1.24.8, 3. 3. 4, 3.7.4, 3 .10.35, 3 .10. 39 (apud Brenot; petiere P, rogarunt NV, petierunt L. Mueller), 3.17.2, 4.1.7, 4.6.4, 4.19.10, 4.19.33, 4.23.18, 4.26.32, 5. prol. 6, 5.1.16, App. 20.4, App. 28.2, App. 28.3.

³⁰ S35, §37, §41·

³¹ E. g., Italian amarono, French aimèrent, from Latin amā’runt.

³² Cf. §20, §23.

³³ Sacerdos, 6.451.4-6 K : Barbarismus est vitiosa dictio unius verbi, per correptionem, steterunt te correpta quae longa est. Servius in Donatum, 4.444.256 K, in a section entitled ^ť De Barbarismo ’ : ‘ steteruntque comae et vox faucibus haesit’ (Aen. 2.774) pro steterunt; media enim syllaba naturaliter produci debuit. The term ‘ barbarismus ‘ is defined as an error of pronunciation in prose. A deviation from the standard form (i. e., -ērunt) in verse is called a ‘ metaplasmus ’ : Sacerdos, 6.451.14-6 K : Haec vitia cum dicuntur, barbarismi sunt; …. et cum dicuntur a nobis, vitia sunt; cum a poetis, metaplasmi. Servius in Donatum, 4.444. 5-11 : Sed tunc barbarismus, cum in uno Romanae linguae sermone peccamus. Ceterum si peregrina sint verba …. tunc barbarolexis vocatur. Praeterea si in prosa oratione fiat, tunc barbarismus dicitur; si autem in poemate, metaplasmus vocatur ………. si soluta est oratio, barbarismum vocaremus. The example given by Servius is therefore a metaplasmus rather than a barbarismus.

³⁴ Cf. note 16.

³⁵ Cf. note 24.

³⁶ Inst. Or. 1.5.43, among other examples of -ere forms : ‘ dixere ’… . praeco pronuntiet.

³⁷ The following Italian forms illustrate this development : amarono (amārunt), dormirono (dormīrunt), furono (fuĕrunt), fécero (Jecĕrunt), dissero (dixĕrunť), ébbero (habuĕrunt.)