In the famous and widely cited opening of his Satires (henceforth, Sat.) 1.4, Horace states (1–5):

Surprisingly, scholars have almost universally failed to notice that line 4 poses a problem.Footnote ¹ Ancient and modern readers can in fact find in the plays of Eupolis, Cratinus and Aristophanes a number of cases in which a thief (fur) or an adulterer (moechus) is mentioned by name; but a murderer (sicarius) is a figure they would never encounter in Greek Old Comedy in the context of ὀνομαστὶ κωμῳδεῖν,Footnote ² which is undoubtedly what Horace is referring to by si quis erat dignus describi and multa cum libertate notabant.Footnote ³

There is a reason why nobody is called a sicarius in Greek Old Comedy, and why we cannot suppose that Horace might have found a reference in a play which he could have read but which is now lost. The ‘law on slander’ (δίκη κακηγορίας; cf. Lys. 10, Against Theomnestos), which existed in effect in Athens at the time of Eupolis, Cratinus and Aristophanes, forbade certain kinds of insults or allegations to be made in public.Footnote ⁴ Amongst these was the accusation to be ἀνδροφόνος, a ‘murderer’, i.e. sicarius. The modern word ‘slander’ may be misleading. The δίκη κακηγορίας was not concerned with the truth or falsity of the allegation but with the use of ἀπόρρητα in public Athenian contexts, and therefore in the staging of dramas too.Footnote ⁵ The term ἀπόρρητα, like the Latin equivalent nefanda, refers to ‘the quasi-magical or quasi-religious sense that these were words which it was somehow inappropriate to utter whether or not they were true’.Footnote ⁶ ‘Murderer’ was one of those words, and this is why we find no references to killers in Old Comedies, either as epithets or as accusations.Footnote ⁷

Still, Horace himself states that he was a reader of Greek Old Comedies. In Sat. 2.3.11–12 Eupolis is mentioned amongst the authors whose works are ‘squeezed’ in the trunk that the poet always takes with him when he moves from Rome to his villa in Sabina. He must also have been a careful reader of those texts, as one can easily infer from passages where he declares that an excellent knowledge of Greek models is absolutely fundamental for the production of good Latin poetry (cf. for example Ars P. 268–9: uos exemplaria Graeca | nocturna uersate manu, uersate diurna).Footnote ⁸ Can we, then, suppose that Horace was simply wrong in pointing to some references to murderers nominatim in the plays of Eupolis, Cratinus and Aristophanes? Can we really see an imprecision in what Horace is saying at the beginning of Sat. 1.4, in a passage which is of pivotal importance for the contents of Horace's statement about the relationship between Greek Old Comedy and Lucilius as the inventor of a genre, a relationship which is so relevant for Horace's own satiric poetry?Footnote ⁹

Kirk Freudenburg has interpreted the passage as a parody of the alleged censorial function of Old Comedy. Noting the anomaly described above, he has suggested that the absence of murderers from the plays of Eupolis, Cratinus and Aristophanes can be compared to the Old Comedy's treatment of the theme of adultery:

This view is, I believe, misleading. Whereas, in fact, the Old Comic poets never name murderers in their plays for the legal reasons mentioned above, they do bring adulterers onto the stage. Aristophanes refers to μοιχεία quite often in his plays. In Thesm. 466–96, for instance, Euripides’ relative describes an adultery in all its details, from how the woman conceived it to how the sexual intercourse took place, and concludes by accusing all living women of being unfaithful to their husbands (549–50).

Moreover, Aristophanes’ plays warn the audience that adultery, by women or men alike, is the consequence of a departure from traditional morality. When pointing out to the Athenian audience what the advantages of having wings are, the chorus-leader in the Birds mentions the ease with which adulterers could be reached by their lovers (Av. 793–7)—a risqué as well as an immoral behaviour punishable by law.Footnote ¹¹

Beside such general forms of criticism of adultery as an immoral act, specific living figures were associated by Old Comedy playwrights with μοιχεία in the context of ὀνομαστὶ κωμῳδεῖν. In Lys. 725 Aristophanes mentions a man called Orsilochus, whom a woman in the play wants to visit, being desperate for sex while her husband is away. Orsilochus is explicitly called in the scholia a μοιχός.Footnote ¹² Similarly, the scholia to Eq. 608 and to Vesp. 43 inform us that Theorus, whom Sommerstein convincingly identifies as ‘the associate of Cleon who is frequently attacked in Ar[istophanes]’ other early plays’, was targeted by Old Comedy poets ὡς μοιχός.Footnote ¹³ In Av. 440, Aristophanes hints at the adulterous wife of a cutler, whom the Athenian audience is supposed to know for her infidelities. Euripides’ wife is likewise accused of adultery in Ran. 1048. Aristophanes was not alone in targeting Athenian adulterers nominatim. A scholium informs us that there was a μοιχός named Phormio whom Cratinus mentioned, and probably criticized, in his play Atalanta.Footnote ¹⁴ Athenaeus (369a) states that the comic playwright Telecleides accused Gnesippus of being ‘addicted to adultery’.Footnote ¹⁵ This very same Gnesippus, a poet whose literary genre is still debated,Footnote ¹⁶ is called by Eupolis the inventor of ‘the night-time songs for adulterers to call out their women’Footnote ¹⁷ in fr. 148 of the Helots, a passage which Prauscello has convincingly argued to be an ironic form of criticism against ‘the degenerate musical taste of the Athenian aristocratic hegemonic class’.Footnote ¹⁸

These examples should be enough to prove that Freudenburg's parallel between the (true) absence of murderers and the (alleged) absence of adulterers in Old Comedy is untenable. Consequently, his point on Horace's parody of Eupolis, Cratinus and Aristophanes seems less convincing, as the exclusion of murderers from the domain of their targets appears to be the only detail contradicting the statement in Sat. 1.4.1–5. Jennifer Ferriss-Hill has recently demonstrated that ‘[w]ith his claim that Lucilius “hangs entirely” from the writers of Old Comedy, Horace indicates how he and his predecessor adapted Old Comedy in forging their Roman Satire. A multilayered programatic statement, Satires 1.4.1–7 draws particular attention to the ways in which Lucilius aligns with Cratinus, leaving Aristophanes to be Horace's own preeminent model, and the incongruous trio of Old Comic censors to oversee his activities as scriba quaestorius.’Footnote ¹⁹

To this extent, there is no reason to maintain that Horace's representation of the Old Comic playwrights at the beginning of Sat. 1.4, a programmatic poem as well as the first in which Horace names the inventor of the genre and sketches its genealogy, is deliberately incorrect. On the contrary, in order to stress the lineage of Roman satire from Old Comedy, the poet needs to have a picture which his readers can understand and acknowledge as true.Footnote ²⁰ Thus, his explanation of the Old Comic poets’ role within society consists in a readaptation of that role to the Roman context of his readers: παρρησία becomes libertas, and the ὀνομαστὶ κωμῳδεῖν, which, according to the iambographic tradition (Old Comic playwrights, iambographers and moralizing Cynics), had the public function of ‘disinfecting’ Athens by pointing at the wrongdoers,Footnote ²¹ is rendered into the Roman censors’ duty of ‘branding’ (notare) the immoral behaviour of citizens nominatim.

Finally, it is worth noting that Horace's precision at the beginning of Sat. 1.4 also concerns his description of formal aspects of the comic-satiric tradition. As Ferriss-Hill rightly observes, Horace's assessment of Lucilius' metre at Sat. 1.4.7 (mutatis tantum pedibus numerisque) is technically accurate, because the latter poet wrote his earliest books in senarii and septenarii, the Roman equivalents of Greek iambic trimeters and trochaic tetrameters: ‘Lucilius could only have been thought of as reinventing Old Comedy as Roman Satire if he wrote in the same meters, which he did, at least initially.’Footnote ²²

If there are no sicarii in the plays of Eupolis, Cratinus and Aristophanes, if there was even a law which forbade them to brand anyone as a murderer, if the purpose of the paradosis of Sat. 1.4 is to sketch a trustworthy picture of the ancestors of Roman satire, how do we account for the word sicarius in Horace's list of the Old Comic poets’ targets?

The list has been paralleled by scholars with a passage from the immediately preceding poem. Emily Gowers observes that ‘H[orace]’s characterization of Old Comedy's moral function recalls the embryonic legal system in his ‘history’ of the social contract at 3.105–106 (ponere leges, | ne quis fur esset neu latro neu quis adulter; cf. 3.103 notarent ~ 4.5 notabant).’Footnote ²³ Thus, one might be tempted to read sicarius in Sat. 1.4.4 as the equivalent of latro in 1.3.106, expanding the parallels fur–fur and adulter–moechus. However, a close analysis of the context of the passage from Sat. 1.3, as well as the results of a survey of the use of the term latro in Latin sources, exclude this possibility.

Thomas Grünewald's thoughtful study of a variety of ancient testimonies shows that ‘in Roman law … any person who involved himself in robbery could also be called a latro … though every latro was a robber, not every robber was a latro. In other words, the crime of robbery appears to have been a necessary, but not sufficient, criterion for being a latro. Latrones are emphasized as a special category of robbers through their use of weapons (uis armata), through their forming of bands (factiones, homines armati coactiue) and through their aim of plunder (spoliare) with malice aforethought (dolus malus).’Footnote ²⁴ This means that, in order to prove that sicarius in Sat. 1.4.4 is used as a synonym of latro, one has to demonstrate that the latro of Sat. 1.3 is a murderer, beyond being a robber. Gowers, discussing Sat. 1.3.103, rightly notices that ‘[t]he distinction between furtum (petty theft) and latrocinium (armed assault and burglary) foreshadows the distinction in the relative seriousness of crimes made at 116–117.’Footnote ²⁵ This latter passage, however, is totally concerned with stealing properties, not with murdering someone. Horace compares two kinds of thefts, a theft of cabbages from a garden and a theft of sacred objects from a temple. These are, respectively, the furta and the latrocinia of line 122; and they refer back to the fur and to the latro of line 106, just as lines 107–8 (nam fuit ante Helenam cunnus taeterrima belli | causa) refer to the adulter of line 106.

Nothing in these lines indicates that the parallel between the sicarius of Sat. 1.4 and the latro of the previous poem is sound.Footnote ²⁶ Finally, it is worth observing that references to murderers are not to be found elsewhere in Sat. 1.4 either. When Horace mentions the latrones Caelius and Birrius amongst the targets of the satirists Sulcius and Caprius in lines 65–70, the word has again the general meaning of wrongdoers.Footnote ²⁷

In addition to the unfitting inclusion of sicarii into the context and contents of the parabasis of Sat. 1.4, a close look at the text of the poem reveals that the word is problematic at the level of form as well. There are seventy-five occurrences of sicarius in the entire corpus of Latin literature. Of them, forty-seven are in Cicero's speeches, almost all in the legal orations.Footnote ²⁸ The rest of the occurrences are often to be found in legal contexts (for example, in the Digesta Iustiniani, in Apuleius’ Apologia and in references to the lex Cornelia de sicariis et ueneficiis occurring, for instance, in Sen. Apocol. 14.1). But the most striking result of this survey is another: if we exclude Hor. Sat. 1.4.4, the word sicarius is never used in Latin poetry. This is, I think, due to the fact that there is an aspect of technicality embedded in the word, the original meaning of which was ‘the man with the sica’, a particular kind of dagger.Footnote ²⁹ Such an aspect certainly did not prevent the meaning from evolving into ‘murderer’ in general (regardless of the weapon used), but made the word sicarius particularly appropriate to the semantic field of law.Footnote ³⁰

The diction and the syntax of Horace's Satires are affected by their generic status of sermo in verse, metrical prose, on which Horace remarks at 1.4.56–62. This leads him to revert to prosaic legalistic language in some passages of his Satires, such as in the formulae datis uadibus in Sat. 1.1.11 and licet antestari in Sat. 1.9.76. As it is well known, the sermo of satire consists in the use of common Latin idiom, and it is therefore opposed to ornate speech. This makes the language of the genre mimetic of the reality of everyday Rome; as Thomas puts it, ‘[s]ermo cottidianus … imports into its literary context the flavour of the street and is in the process transformed into a literary phenomenon.’Footnote ³¹ When assessing the use of prosaic diction in Horace, then, one should always consider in what kind of mimetic context it occurs, i.e. whose everyday language Horace is imitating.

The poet always makes the mimetic context explicit. The formula datis uadibus in Sat. 1.1.11, which imitates the language of a lawyer, is used by Horace precisely because he is bringing a lawyer onto the scene (iuris legumque peritus, 1.1.9). Likewise, the poet is asked licet antestari? in Sat. 1.9.76 because he is called by the bore's aduersarius as a witness to a legal procedure of arrest (rapit in ius, 1.9.77). These examples show that Horace uses prosaic, especially legalistic, diction in the Satires not as a simple cliché of the genre but as a form of imitation of the language of particular categories of people. The word sicarius in Sat. 1.4.4 is extraneous to the mimetic context in which it appears.

It has been noted that the opening of Sat. 1.4 echoes the parabaseis of Greek Old Comic plays; in fact, its sermo imitates the language of the Old Comic playwrights.Footnote ³² The poet's preference for the word moechus in Sat. 1.4 instead of the word adulter of Sat. 1.3 is consistent with the programmatic choice of a Graecism in a passage about a feature of Greek poetry which was typical of the democratic regime of Athens and which—Horace says—Lucilius imported into Rome from there.Footnote ³³ Unlike sicarius, the Graecism moechus is also perfectly suitable to the context, since it is alien to the language of prose but typical of the language of poetry in general and comedy in particular.Footnote ³⁴ Accordingly, the style of the passage is highly poetic. The sentence quod malus ac fur | quod moechus foret combines the anaphora of quod with an alliteration in parallelism of the sounds ‘m’ and ‘f’ (the latter echoed again at the end of the clause by famosus), all within two consecutive enjambements. Similarly, Fedeli points out that the combination of the construction dignus describi, also in alliteration, with the infinitive of a verb is common in the polished works of Catullus and the Augustan poets, probably as a form of Latin calque of the Greek ἄξιος + infinitive.Footnote ³⁵ In this picture, the profoundly Latin and profoundly prosaic word sicarius appears alien to the colour of the passage, and its use can hardly be justified by simply appealing to the nature of Horace's sermo.Footnote ³⁶

These formal observations are small but cumulative arguments which, when considered in the light of the major issue of the absence of murderers in Old Comedy, should—I believe—lead to our questioning the soundness of the reading aut sicarius in Hor. Sat. 1.4.4. My proposal for an emendation is the following:

According to my emendation, Horace's reference is to the Old Comic poets’ criticism of someone ‘for being a villain and a thief, for being an adulterer or an otherwise notorious man, despite being a more eminent figure’ than the Old Comic poets themselves (or, simply, ‘a rather eminent figure’).Footnote ³⁸

The reading etsi clarior, which solves all the problems concerning the form and content of the passage, fits Horace's description of the attitude of the Old Comic playwrights towards the frank criticism of individuals, which was then inherited by Lucilius (hinc omnis pendet) and ended up characterizing the genre of Roman satire. LaFleur rightly points out that, ‘as it was with the Old Comedy poets, with Lucilius, and the best of fathers, so with Horace, the aim of onomasti komodein, to furnish vivid examples of behaviour both to emulate and to avoid, is moral, constructive, prophylactic.’Footnote ³⁹ In order for public criticism to have this ‘therapeutic value for the state’,Footnote ⁴⁰ the targets must be eminent figures within society. Eupolis, Cratinus and Aristophanes were clearly aware of this;Footnote ⁴¹ and so was Lucilius, when he mocked or branded the members of the patrician families and ruling class of Rome.Footnote ⁴²

In Sat. 2.1.62–70, Horace himself refers to this attitude of the inventor of the genre in his reply to Trebatius’ warnings of the risks of criticizing powerful and important figures:

Fedeli interprets the passage as ironical, since Metellus and Lupus were personal and political enemies of Scipio and Laelius.Footnote ⁴³ This might be a sound interpretation; but Horace's main point is that Lucilius lampooned contemporary figures regardless of their social status, and even regardless of their relationship with him or his protector, as fragments from his Satires indicate. So, for instance, the poet mocks Scipio himself for his pronunciation of Latin words (frr. 963–4 M) and, quite embarrassingly, represents him in company with a cinaedus (frr. 1138–42 M, although the context is not clear). More ignominious attributes could also be used, such as operarius, sector and even fur (cf. Hor. Sat. 1.4.3) in frr. 422–4 M, all applied to Gaius ‘big-head’ Cassius, who was politically close to Scipio.Footnote ⁴⁴ If there is any, the irony of Horace's reply to Trebatius does not weaken or modify Horace's point: Lucilius, ‘entirely dependent’ on the spirit of Old Comedy (Sat. 1.4.6), is presented as a poet who has dared to attack the primores populi, as he had in fact programmatically declared to do in his Satires.Footnote ⁴⁵

Ancient scholars considered full democracy and freedom to be quintessential factors for the genre of Old Comedy.Footnote ⁴⁶ In the late-antique treatise On the Different Types of Comedy (Περὶ διαφορᾶς κωμῳδιῶν), based on Hellenistic sources, Platonius explicitly identifies the historical cause (τὰς αἰτίας, 1) of the transition from Old Comedy to Middle Comedy with the change of political regime in Athens.Footnote ⁴⁷ Since under the democracy of the seventh and sixth centuries b.c. there was freedom of speech for all (τῆς ἰσηγορίας, 6), ‘the writers of comedy were not afraid of lampooning the generals, the jurymen who judged wrongly, and some citizens who were greedy or lived in licentiousness’ (7–10); but then, when the oligarchy of the Thirty Tyrants was established, ‘fear fell upon the poets. In fact, it was no longer possible to lampoon anybody openly, because those who felt outraged by the poets arraigned them’ (19–21).

It is important to stress that Platonius is not presenting a theoretical picture of the link between comedy and society. He is outlining a history of the genre, pointing out that the change of political regime in Athens brought about substantial changes in technical aspects of the dramatic art. The comic poets’ abandon of choral songs, certain plots, the parabasis and masks portraying the citizens who were lampooned is what distinguishes Old Comedy from Middle Comedy and New Comedy (Diff. com. 23–75).

Horace sketches a similar history of Greek comedy in Ars P. 281–4:Footnote ⁴⁸

Although Horace is less explicit than Platonius on the role played by politics in the transformation of the genre, the issue is clearly at stake in the background. The poet points out that the limitation of the libertas of the Athenian comic poets did not occur as a spontaneous choice, but it was imposed by an official law (lege … lex, 283) which was introduced to prevent them from nocere (284). The object of the expression sublato iure nocendi is omitted, but it could be easily understood by any reader as a reference to the Athenian politicians who were harshly criticized by Eupolis, Cratinus and Aristophanes. In this sense, Horace's point is the same as Platonius’: ‘it was no longer possible to lampoon anybody openly, because those who felt outraged by the poets arraigned them’ (Diff. com. 21).

To go back to the opening lines of Sat. 1.4 as they have been preserved, it should now be evident that what is missing from Horace's picture of the role of ὀνομαστὶ κωμῳδεῖν in the plays of Eupolis, Cratinus and Aristophanes, as well as in the satires of their Roman heir Lucilius, is the core element of that criticism, namely the fact that its socio-political function could operate only if the targets were eminent figures. Instead of a reference to this aspect of the comic-satiric genre, of which Horace demonstrates awareness elsewhere in his works (Sat. 2.1.62–70 about Lucilius; Ars P. 281–4 about Old Comedy), we find the odd and unjustified inclusion of murder into the list of the accusations which were moved by the Old Comic poets, an accusation which is in fact foreign to the comic-satiric criticism of Greek and Roman antiquity. My emendation is in line with the convincing interpretation of Sat. 1.4 given by Emily Gowers, who reads the poem, and especially its opening lines, ‘as political in the widest sense … because it considers the crucial and timely question of the place of the satirist in relation to newly regulated society’.Footnote ⁴⁹

To conclude, some brief remarks about when and how the text of Sat. 1.4.4 may have been corrupted. The fact that all the manuscripts but one agree on the reading aut sicarius should not lead us to exclude a priori the possibility of scribal error. It is likely that the editions which scribes could access in the ninth century, when the oldest manuscripts of Horace we possess can be dated to, already contained contaminations and interpolations, which probably occurred between the time of the editio princeps by Probus and the Carolingian age, if they were not already present in that very same editio princeps.Footnote ⁵⁰ Therefore, although the text of Horace has generally not given reason to scholars to emend it, the possibility that the whole manuscript tradition includes mistakes does exist.Footnote ⁵¹

If we assume that the mistake occurred in an era prior to the era of the archetype(s) of the manuscript tradition, we can hypothesize what might have generated or contributed to the corruption of the original reading etsi clarior into aut sicarius:Footnote ⁵²

• • Instead of aut sicarius, an eleventh-century manuscript, Monacensis Lat. 14685 (E), has the lectio at si carus, which cannot be accepted for metrical reasons but which suggests that the sequence of letters tsic could have been fragmented in the wrong way in the manuscript tradition: (au)t sic(arius) instead of (e)tsi c(larior).

• • The transformation of etsi into aut si might have occurred at two different stages. Initially, the original sequence foretetsi might have become foretsi owing to haplography. A later scribe might have then restored the missing syllable et before si, but might have inserted different vowels than the original (foretaut instead of foretet), possibly attracted by the second aut of the sequence (since he was certainly aware that aut is often used in Latin in correlation: cf., for example, Hor. Carm. 1.7.1–2, laudabunt alii claram Rhodon aut Mytilenen | aut Epheson bimarisue Corinthi). Alternatively, it is possible to interpret the reading at si of Monacensis Lat. 14685 (E) as the result of an earlier confusion between a and e, which is common in some minuscule scripts. A second scribe might then have transformed at into aut owing to the attraction of the second aut of the sentence.

• • To account for the transformation of (etsi)clari(or) into (aut si)cari(us), two hypotheses can be advanced. The first and easiest is that the l of clarior disappeared in a pre-Carolingian edition owing to an erroneous omission by a scribe. The second is more complex. Given that c and l are written in a very similar way in some minuscule handwritings, an early-Carolingian scribe might have written the sequence of letters siclar in a way which a later scribe might have copied as siccar. The latter scribe, or even a later one, might have then omitted a c, thinking of dittography or realizing that the word did not make any sense when written with cc.

• • Finally, an ancient or early-Carolingian scribe reading sic(l/c)arior might have changed the –ior ending into –ius as a form of follow-up of the abundant masculine nominatives in –us in the sentence (dignus, malus, moechus, famosus), and/or because the word sic(l/c)arior does not exist.

Of course, as always in textual criticism, this is all speculative. But, as I have tried to show, the reading aut sicarius in Hor. Sat. 1.4.4 is problematic, both at the level of form and at the level of content. Whether my emendation etsi clarior convinces or not, there is clearly an issue in the text which needs to be taken into account and addressed.