[In the following essay, Abbott explains how Petronius expresses both the individuality and culture level of his characters through their vocabulary, colloquialisms, pronunciation, word-formation, and inflectional forms.]

The character and culture of a man are revealed by his dress, his conduct, his attitude toward the world, by the subjects in which he shows an interest, and by his manner of speech, and upon the use which writers of fiction have made of these indications of character depends the clearness with which we conceive the essential qualities of the people whom they depict. Among the Latin writers no one has equaled Petronius, it seems to me, in the portrayal of character, and the purpose of this paper is to call attention to one method which he has used with great success in attaining his end. His book, especially the main extant episode, while not lacking in external incidents of lively interest, is essentially a character study. In holding the attention of the reader he relies less, for instance, than the other great ancient novelist, Apuleius, upon description. When Petronius gives us an account of the house or the dress of Trimalchio, the appearance of Habinnas, and the jewels of Fortunata, it is incidental to his main purpose. In Apuleius descriptions are freely given for their own sake, as the writer naively remarks on one or two occasions. The subjects about which people in the Goden Ass and in Latin comedy talk, and the attitude which they take toward the world in their remarks, throw some light upon their character. This is true, for instance, of the miser and Thelyphron in Apuleius and of Pseudolus in Plautus, but the situation usually determines the subject of conversation and gives it its color. The tricky slave, for example, is bent upon thwarting the procurer, and his whole mind is centred upon this subject. But in Plautus, in Terence, and in Apuleius there is perhaps nothing quite comparable with the unrestrained flow of conversation at Trimalchio's table, where each speaker, with the cockles of his heart warmed by the hundred-year-old Opimian, talks about the things which make up the real interests of life for him and frankly states his optimistic or pessimistic view of gods and men. The gloomy philosophy of Seleucus, the rough-and-ready standards of Phileros, the querulous senility of Ganymedes, or the prosy optimism of Echion comes out with marvelous clearness in the choice which each makes of a subject and in the sentiments which he expresses about it. But it is in the careful distinction which he draws between the speech of the cultivated and that of the illiterate that the preeminence of Petronius as an artist comes out most clearly. The people who figure in the Satirae fall into two well-marked classes. In the first group are men like the poet Eumolpus, the rhetorician Agamemnon, the anti-hero Encolpius, and his attendant Giton. They all live by their wits and are unscrupulous and vicious, but they are all men of some education and taste, who are fairly versed in the practices of good society. Eumolpus was of course a professional man of letters; Encolpius is characterized as a scholasticus (10), and remarks to Ascyltus et tu litter as scis et ego (10), so that the writer represents them as men whose training would enable them to speak good Latin. Trimalchio and his freedman friends make up the other group. They are quite illiterate, as every one knows, and their conversation smacks of the junk shop and the stonecutter's yard. It is not easy in a brief space to bring out in its full significance the differences which Petronius makes between the language and style of the one set of men and those of the other, but a comparison of passages of similar length in which men of the two respective groups are talking will illustrate the point with sufficient clearness. I have chosen for examination the conversation running from p. 27, 1. 37, to p. 31, 1. 12, in Biicheler's fourth edition, and for comparison with it the two conversations which extend from p. 68,1. 9, to p. 71,1. 12, and from p. 73, 1. 16, to p. 74, 1. 16. The freedmen Dama, Seleucus, Phileros, Ganymedes, and Echion are talking in the first passage; Eumolpus, Encolpius, and Giton are the speakers in the other sections.

From these chosen pages I shall set down with very slight comment the words and phrases which belong to the people's speech and are at variance with formal usage. My sole purpose is to show one point in the technique of Petronius, to bring out the way in which he makes men of different degrees of culture reveal themselves in their language.

Taking up the passage from p. 27 to p. 31, and discussing, first, vocabulary, pronunciation, word-formation, and inflectional forms, we find that it contains the following colloquial elements:

1. Of words belonging exclusively to the plebeian vocabulary there occur bucca (44), [The references are to sections.] burdubasta (45), caldicerebrius (45), cicaro (46), filix (45), linguosus (43), miscix (45), merus (45) and pullarius (43) with transferred meaning, staminatus (41), sestertiarius (45), tertiarius (45), and frunisci (44).
2. baliscus (42), balneus (41), caelus (45), fatus (42), and vinus (41) appear as masculine nouns; schema and stigma as feminines of the 1. declension, and nervia (45), and librum (46) as neuters. The occurrence of these nouns in -us and the assignment of Greek nouns in -a to the first declension, in the speech of the illiterate in Petronius, furnish one of the earliest indications we have of the elimination of the neuter gender in vulgar Latin (cf. Appel De genere neutro intereunte in lingua latina; W. Meyer Die Schicksale des lat. Neutrums im romanischen; Suchier "Der Untergang der geschlechtlosen Substantivform," Archivf lat. Lex. III, p. 161). The analogical nominatives librum (cf. cultrum for culter) and nervia (pl.) are like many similar vulgar forms against which the author of the Appendix Probi warns his readers.
3. The syncopated colloquial forms bublum (44), caldicerebrius (45), calfacio (41), and cardeles (46) need no comment. In the same field of colloquial pronunciation are oricularios (43), percolopabant (44), plodo (45), and plovebat (44). The forms oricularios and plodo attest a well-known vulgar pronunciation of au. plovebat, which from Festus (330. 29 Th. de Ponor) we infer is an archaic form, is one of the interesting cases in which popular Latin has retained a form which has dropped out of the literary speech. As for percolopabant, one may well ask if Petronius is not nodding here. We should not be at all surprised at hearing the unaspirated pronunciation from an illiterate Italian in Rome or northern Italy, but in Cumae, in the mouth of the Greek Ganymedes, like tisicus in 64, it seems to be a slip on the part of the author.
4. The favorite popular endings -arius, -atus (Stolz Hist. Gram., pp. 424 f.), which are used eleven times in the Cena alone, -ax, -osus, -im, and the diminutive ending, are well represented by pullarius (43), sestertiarius (45 bis), tertiarius (45), staminatus (41), abstinax (42), salax (43), linguosus (43), urceatim (44), bellus (42), servulus (46), and amasiunculus (45).
5. arguto (46) and verso (41) perhaps show the colloquial fondness for iteratives, and arguto and delector (45), with an active meaning, illustrate the freedom with which deponent verbs admit an active form or vice versa in colloquial speech. vinciturum (45) is perhaps a clever plebeian attempt, like similar cases in Plautus, to distinguish the corresponding tense forms of vinco and vivo, while diibus (44) finds parallels in the plebeian inscriptions.
6. There are several interesting deviations from formal usage in the matter of syntax, and they are all indications of the breaking-down of the synthetic system which had reached such a high point of development in formal Latin. Of the prepositions those which take both the accusative and the ablative are most liable to confusion, and fui in funus (42) does not surprise us, since it is a short-hand phrase for the full expression "I went to, and was present at the funeral." The same careless brevity is responsible for the use of for as in the sentence nunc populus est domi leones, foras vulpes (44). unus de nobis (44), while not very unusual, is a precursor of the analytical form of the partitive expression. In twenty-five or thirty other cases in the Satirae the partitive idea is expressed by de or ex with the ablative. prae literas (46) illustrates a more general confusion of the accusative and ablative than fui in funus does (cf. Suchier Archiv III, p. 165). Analogy explains sufficiently tepersuadeam (46), meos fruniscar (44) and quodfrunitus est (43). The popular mind grasps the broad truth, with slight regard for the subtle limitations which the grammarian puts on it in practice, that the accusative is the case of the direct object, and this general principle it is likely to extend to any transitive verb in the language, and in this category fall persuadere and frunisci, although perhaps in the case of frunisci one may say that of the two constructions allowable in early times after frui and its derivatives popular speech retained the accusative, because this construction conformed to the general case usage after transitive verbs. In a similar way aediles male eveniat (44) is probably to be explained, although Biucheler regards the use of the accusative here as a Grecism. male evenire conveys an idea sufficiently transitive to justify to the popular mind the case commonly used after transitive verbs. belle erit (46) illustrates a colloquial usage common enough from the time of Plautus on. The indicative in the indirect question, nemo curat, quid annona mordet (44), and quod with the indicative for the accusative and infinitive in subolfacio quod … daturus est (45), and dixi quod mustella comedit (46) are early instances of a deviation, common enough later, from formal modal usage.

   From this brief examination of the favorite forms, words, and syntactical usages of Dama and his friends we come now to a discussion of their style:

7. They are rather fond of long words like frunisci (43) and argutare (46).
8. They are very free in the use of such epithets as aediles trium cauniarum (44), sestertiarius homo (45 bis), burdubasta (45), loripes (45), discordia non homo (43), stips (43), terrae filius (43), servi oricularii (43), pullarius (43), and linguosus (43). Of a complimentary character are cicaro meus (46), fortunae filius (43), omnis minervae homo (43), and piper, non homo (44).
9. Their favorite oath is mehercules (43 bis, 44, 45) which Cicero condemns (Or. 157) as an undesirable form.
10. modo, modo me appellavit (42), and the double negative in neminem nihil boni facere oportet (42), as 1 have tried to show in the University of Chicago Studies in Classical Philology III, pp. 72 f., are both instances of the colloquial use of duplication for emphasis. Perhaps olim oliorum (43) belongs to the same category.
11. A far more striking colloquial characteristic of the passage which we are considering than any of those thus far mentioned is the extraordinarily liberal use made of popular and proverbial expressions like amicus amico (44), udi mures (44), pro luto esse (44), micare in tenebris (44), urceatim plovebat (44) serva me, servabo te (44), animam ebulliii (42), and habet unde (45). In these four pages there are no less than seventy-five such phrases.
12. Equally noteworthy is the use by the speakers of asyndeton and of short, co-ordinate sentences and paratactical expressions. Some of these characteristics may be illustrated by a few lines from the remarks of Phileros (43): plane Fortunae filius, in manu illius plumbum aureum fiebat. facile est autem, ubi omnia quadrata currunt. et quot putas illum annos secum tulisse? septuaginta et supra. sed corneolus fuit, aetatem bene ferebat, niger tamquam corvus. The frequent employment of nam and of et merely as narrative particles to introduce a new statement, of which one case occurs in this passage (et quot putas, etc.), is characteristic of all the speakers of this group.

At pp. 68-71 and 73-74 it will be remembered that Eumolpus, Encolpius, and Giton are talking together. Using the rubrics adopted above, we notice the following colloquialisms: (a) excanduit (100) in a figurative sense; (d) tremebundus (100) with a favorite colloquial ending; (e) the colloquial phrases quid ergo (102) and quis nobiscum Hannibal navigat (101), although these expressions would not surprise one in formal Latin, and (i) the expletive per fidem (100). No other deviations from formal usage seem to occur anywhere in the four pages. The contrast, therefore, between the Latin which Dama and his friends use in conversation and that which Petronius puts into the mouths of his more cultivated characters is very striking.

To make sure that this difference in language is characteristic of the whole Satirae, let us take another passage, this time of about two pages, in the Buicheler edition, including 26-30. The speaker is the narrator Encolpius. Of colloquialisms we find: (a) unus (26), perhaps used as the indefinite article; (c) the form hoc (26) for huc; (e) the favorite colloquial periphrasis with coepi in errare coepimus (27), interrogare coepi (29), rogare coepit (30), and perhaps conaremur intrare (30) should be mentioned in this connection, and finally (h) deliciae (28) may be noted, although the word is not uncommon in formal Latin. The few words quoted from Trimalchio in 26, and the brief remarks of slaves in 26, 30 and 31 have, of course, been left out of consideration.

If Petronius has not exaggerated the peculiarities of his freedmen, there is no piece of Latin literature which shows in so interesting a fashion the difference between the sermo urbanus and the sermo plebeius, and, what is more to our purpose, no writer has so clearly indicated the standing or the culture of his different characters in their speech as Petronius has done.

He seems, too, to give individuality to his characters by showing their fondness for certain words or phrases. Friedlander, in his edition of the Satirae, p. 218, has already noted that ad summam, recte, and curabo with the subjunctive are favorites with the freedman Hermeros. In our first passage Seleucus, within a half page, uses quid si non twice, and both the cases of duplication noted above occur in his remarks. Seven of the eleven plebeian words are used by the rag-dealer Echion, while two of the four instances of mehercules are found in what Phileros says. Of the other two speakers Dama says too little to reveal his peculiarities, and perhaps the most marked quality of Ganymedes is his staccato style.

Finally it seems possible to detect certain differences between the styles of Trimalchio drunk and Trimalchio sober. Plautus has brought out some of the comic aspects of drunkenness in his plays. The sentimental, helpless attitude of Callidamates in the Mostellaria, and his thick-tongued utterance furnish one type of the drunken man; the sternly moral tone and the fluent discourse of Stasimus in the Trinummus illustrate the effect of stimulants upon a different temperament, but no Latin writer has made so interesting and accurate a psychological study of the effects of intoxication as Petronius has in the case of Trimalchio. Under the mellowing influence of the wine the host forgets his dignity as land-holder and sevir Augustalis. With the gradual change in his conduct and manner, which Petronius skilfully depicts, we are not concerned here. But his language, as well as his manner, undergoes a subtle change and loses the few suggestions of the polite world which it shows in the early part of the dinner. The admission of the cook and of the other slaves to share in the festivities (76) is typical of the frank relapse of Trimalchio's demeanor and speech to those of his former servile condition in life, and the gradual change in them is very cleverly indicated by Petronius; but it is a change whose subtlety can be appreciated only from one's own reading and not through an analysis made by another.