1. E.g., Michaud, G., Plaute (Paris 1920), i.157f.Google Scholar: ‘Voilà un drame complet: graves événements et dangers tragiques, beau dévouement et noble fidélité, tristesse et joies paternelles, reconaissances émouvantes, rien n'y manque, pas même le traître, pas même la vertu récompensée et le vice puni’; Stace, C., ‘The Slaves of Plautus’, G&R 15 (1968), 67Google Scholar: ‘[Tyndarus] even risks his life for his master-a most refreshing change after the normal Plautine slave-and is quite unPlautine in his nobility’; Viljoen, G. Van N., ‘The Plot of the Captivi of Plautus’, AC 6 (1963), 46Google Scholar: ‘…the character-type of the crafty slave assisting his young master in his intriguing, undergoes a veritable apotheosis in the role of Tyndarus, the real hero of the piece, with his noble self-sacrifice for the sake of his master’; similarly, though without using ‘noble’, Grimal, Pierre, ‘Le modèle et la date des Captivi de Plaute’, in Bibauw, Jacqueline (éd.), Hommages à Marcel Bernard (Brussels 1969), 398Google Scholar: ‘Tyndare, l'esclave, se hausse, par sa vertu, jusqu' à la qualité de sage [stoïcien]. Il échappe, par là, aux valeurs vulgaires—celles qui forment le monde d'Hégion—et se meut dans un autre univers.’ It is perhaps invidious to single out these authors; nearly all scholars write to similar effect on this play. The foregoing quotations also illustrate an unconscious acceptance of the play's class outlook (examples of this too could easily be multiplied), but here, from the other end of the political spectrum, is Dunkin, Paul Shaner, ‘Post-Aristophanic Comedy: Studies in the Social Outlook of Middle and New Comedy at both Athens and Rome’, Illinois Studies in Language and Literature 31, nos. 3-4 (Urbana 1946), 79Google Scholar: ‘To only one of them [slaves] did he [Plautus] grant nobility, and Tyndarus is still the only inspiritingly good man in comedy’; and (80) ‘[Tyndarus'] position is the more pathetic by virtue of the very nobility which he exhibits in maintaining devotion to duty in relation to such a master.’ My point is that it is easy for us to be unwary readers of ancient texts as we would not be, perhaps, of modern literature because they appear so remote, and that we can fall involuntarily into adopting whatever class perspective they seem to have; for a blatant example, see Legrand, Philippe E., The New Greek Comedy (Westport CT 1970Google Scholar; reprint of the 1917 edition), 106-16, 455f., 508. Most people nowadays are not, probably, supporters of slavery, and yet they regularly cheer the slave Tyndarus' self-sacrifice for his master (Dunkin at least avoids that uncritical posture).

2. Agnew, M.E., ‘Lessing's Critical Opinion of the Captivi of Plautus’, Classical Weekly 39 (1945–1946), 66–70CrossRefGoogle Scholar.

3. Segal, Erich, Roman Laughter: The Comedy of Plautus² (New York and Oxford 1987), 191–214Google Scholar—an appendix added to the second edition. On the manipulation of conventional patterns, see also Harsh, Phillip Whaley, A Handbook of Classical Drama (Stanford 1944), 348Google Scholar (who also appredates the play as comedy, if a ‘quiet’ one: p.346), and Dumont, Jean Christian, ‘Guerre, paix et servitude dans les Captifs’, Latomus 33 (1974), 507f.Google Scholar Viljoen, on the other hand, sees the typical roles transformed in this play (n.1 above, 46), and according to Eleanor Winsor Leach's sensitive but pessimistic reading, the characters' efforts ‘to preserve the real self are thwarted as they regularly ‘slip into the postures of the stock figures of comedy’: ‘Ergasilus and the Ironies of the Captivi’, C&M 30 (1969), 274fGoogle Scholar.

4. Norwood's scathing invective (‘was such a gulf of ineptitude ever plumbed before or since?’: Plautus and Terence [New York 1932], 89Google Scholar) only represents the extreme. The same literalist tendency is displayed by more sympathetic critics as well. See in general Hough, John N., ‘The Structure of the Captivi’, AJP 63 (1942), 26–37Google Scholar, and Viljoen (n. 1 above). I agree with David Konstan that anomalies of this type are best considered ‘not as defects in plotting but as economies which leave the fundamental situation of the play prominently visible and do not encumber it with distracting ramifications’: Roman Comedy (Ithaca and London 1983), 58fGoogle ScholarPubMed.

5. Cf. Hough (n.4 above), 31f; Viljoen (n.1 above), 56-58; Leach (n.3 above), 292f.

6. Presumably there was a Greek prototype for this play, but this may also have been atypical, for the same reason.

7. Fraenkel, Eduard, Plautinisches im Plautus (Berlin 1922), 231–50Google Scholar. On slaves in Plautus see Stace (n. 1 above), and above all Spranger, Peter P., Historische Untersuchungen zu den Sklavenflguren des Plautus und Terenz² (Stuttgart 1984)Google Scholar.

8. Spranger (n.7 above), 47-51; Segal (n.3 above), 137-69; Parker, Holt, ‘Crucially Funny or Tranio on the Couch: The Servus Callidus and Jokes About Torture’, TAPA 119 (1989)Google Scholar. Tranio at the end of the Mostellaria is probably the most memorable example: quid grauaris? quasi non eras iam commeream aliam noxiam.libi utrumque, et hoc et Mud, poteris ulcisci probe (‘Why are you complaining? As though tomorrow I won't commit another offence. Then you can justly punish me for both, this one and that’, 1178f.). Quotations from Plautus are taken from Lindsay's Oxford Classical Text edition (Oxford 1910). Translations are my own.

9. Parker (n.8 above). The quotation is his summary of part of this theory (p.235).

10. Parker (n.8 above), 238-40.

11. On these points, see the introductions to Konstan's two books on comedy (n.4 above, 1532; Greek Comedy and Ideology [New York and Oxford 1995], 3–11)Google Scholar. This viewpoint is not incompatible with Segal's argument (n.3 above) about Saturnalian inversion as typical of Roman comedy (although as an account of the genre that notion may be incomplete [Konstan (n.4 above), 29-31]), because the ultimate effect of such inversions is to reinforce the restrictions of daily life.

12. Franko, George Fredric, ‘Fides, Aetolia, and Plautus' Captivi’, TAPA 125 (1995), 155–76, esp. 172–74Google Scholar. What I say here goes beyond his discussion but is, I think, implicit in it.

13. In Rome, of course, a manumitted slave became a citizen, and manumission thus functioned as one way to integrate outsiders into the community; but until manumission the slave was property and was not viewed as a potential citizen (that would have been fatal to the whole institution of slavery). On slaves as outsiders in at least one aspect, see Konstan (n.4 above), 62. Thébert, Yvon, ‘The Slave’, in Giardina, Andrea (ed.), The Romans (Chicago 1993), 139–41Google Scholar, sees the opposition between citizen and non-citizen as the fundamental structural principle of the Greek city-state, which gave way by the Roman imperial period to that between slave and free. It is not clear where he would place the early second century BCE in this process. In any case, this view, though no doubt correct as a rough generalisation, poses the change too absolutely. The Greeks did not simply lump slaves with, say, metics as aliens: they were well aware of the difference between slave and free. It is true that Aristotle, whom Thébert invokes, in the Politics seems to equate the natural slave with the barbarian, but in my view that was simply his way out of the insoluble difficulties inherent in the concept of the natural slave. It is true in addition that slaves could also be viewed, and often saw themselves, as part of the family, and that in Rome they were, like freeborn sons, completely under the power of the paterfamilias (see Thébert 150f, Konstan 62). That does not mean, however, that the difference between slave and free was ever forgotten, even in this intimate setting and whatever affective relations developed. Aristotle, whom Thébert cites on this point too (‘…it is significant that all that Aristotle has to say about the slave in book I of the Politics is constantly intermingled with parallel remarks about the child and the woman’, 140), says clearly that the master's power over slaves differs in kind from his power over wife and children (Pol. 1259a37-59b1). I therefore find Thébert's discussion of the Roman change from a ‘patriarchal’ to a more systematic form of slavery overdrawn, even though there was a significant shift in the practice of slavery and it is important for our purposes that it was occurring when the Captiui was first produced. I would view the assimilation of the slave to a family member, to the extent that it occurred, as a disguise that served both to justify and mitigate the true relations of slavery.

14. Freyburger, Gérard, ‘La morale et la fides chez l'esclave de la comédie’, REL 55 (1977), 113–27, esp. 122-25Google Scholar; Dumont (n.3 above), 512. For a different view see Konstan (n.4 above), 62. Franko is aware of the incongruity (n.12 above, 158), and explains Hegio's entering into an agreement with his slaves as due to his assurance that one of them enjoyed high status at home (he is no less now a slave, however), and observes that because of the deception he is gulled into dealing on a footing of equality with the actual slave Tyndarus. But even if this is so, it hardly explains the role played by the concept of fides within the structure of representation in the play as a whole. From that perspective, it is essential that, as the audience knows all along and as the characters discover at the end, Tyndarus is really free-born, a citizen of Aetolia, and Hegio's son. This fact limits the tendency to level social distinctions in the spectacle of a master and his slaves entering on a reciprocal agreement, which can be played out harmlessly.

15. Spranger (n.7 above), 23-26.

16. Spranger (n.7 above), 26-28, 29f. He does emphasise, however, the concern in the play to assert and protect the distinction between slave and free-an important point.

17. Konstan (n.4 above), 57-72, a chapter earlier published as ‘Plautus' Captivi and the Ideology of the Ancient City-State’, Ramus 5 (1976), 76–91CrossRefGoogle Scholar.

18. Dumont (n.3 above); see esp. 512-16.

19. Grimal (n.l above). Grimal's use of alleged Stoic elements in the play to date its Greek model and identify the latter's author seems quite a stretch to me.

20. See Garlan, Yvon, Slavery in Ancient Greece (Ithaca and London 1988), 126Google Scholar. The way such universalism can disguise the actual relation between slave and master can be seen in this comment on Philocrates and Tyndarus by Michaud (n.1 above), ii.39: ‘…en ce commun malheur il n'y a plus ni esclave ni maître, il n'a plus que des amis.’

21. For discussion of these attitudes, see Bradley, Keith, Masters and Slaves in the Roman Empire: A Study in Social Control (New York and Oxford 1987), 26-31 and 33–39Google Scholar respectively.

22. On the vicissitudes of this justification of slavery in Roman culture, see Thébert (n.13 above), 164-66.

23. This is an extremely common phenomenon; on slavery as ‘fictive kinship’, see Patterson, Orlando, Slavery and Social Death: A Comparative Study (Cambridge MA and London 1982), 62–65Google Scholar.

24. Patterson (n.23 above), 336f.

25. According to Plutarch (Cat. Mai. 20.3), Cato's wife nursed slave children along with her own son deliberately to ensure their good-will to him: . See also Spranger (n.7 above), 58f., and Parker (n.8 above), 242.

26. The problem can be seen as implied by Roman comedy's usual response to it, which consisted in never representing the slave as morally superior to the master, as Spranger notes (n.7 above, 113). As he explains, ‘die Unfreiheit erweist sich als ein Mangelzustand der Seele, der Sklavenstand bedingt den Sklavencharacter, und verhindert die volle Entwicklung einer ethischen Persönlichkeit. Nichts ist kennzeichnender für den Character des Bühnensklaven als seine Anpassungsfähigkeit, seine Prinzipienlosigkeit. Schon deshalb können seine Grundsätze nicht in sittlichen Prinzipien verwurzelt sein.' He regards Tyndarus in the Captiui as only an apparent exception. I would say that Tyndarus is the exception that proves the rule: he cannot really be a slave at all, or remain in that condition. In this respect as in so many others, the play gives the implicit ideological background for the representation of slaves in comedy generally, and it is the contours of that ideology that I want to explore here.

27. κτῆμά τι ἔμψνχον, Arist. Pol. 1253b32; instrumentum uocale, Varro RR 1.17.1.

28. Arist. Pol. 1.1255a3-1255b 15, 1256b20-26. In my view, Aristotle in the Politics was codifying long-held attitudes to slavery and trying to cope with their inner contradictions. A striking, if bizarre, example of a dominant culture's anxiety to combine these versions of the slave, and of the lengths to which it might go, is given by Gould, Stephen Jay, The Mismeasure of Man (New York and London 1981), 71Google Scholar. He tells of S.A. Cartwright, a nineteenth-century physician in the American South, who ‘wondered why slaves often tried to flee, and identified the cause as a mental disease called drapetomania, or the insane desire to run away. ‘Like children, they are constrained by unalterable physiological laws, to love those in authority over them. Hence, from a law of his nature, the negro can no more help loving a kind master, than the child can help loving her that gives it suck.’ Notice how Cartwright manages to naturalise both the slave's inferiority (‘like children’) and obedience, and to join with the latter a propensity for running away, here seen as aberrant but a disease to which slaves are naturally prone. That this is allegedly pathological also depicts domination and obedience as the norm and naturally right—a view that flatters the master class's self-image.

29. Therefore, although I sympathise with the impulse behind his vigorous argument, I cannot agree with Dunkin that the Plautine slave's trickery (behaviour in which Tyndarus engages) ‘is the natural result of his position: a man driven to cunning by ill treatment and given abundant opportunity to complain of that treatment', and that in Plautus' plays ‘may be seen the instinctive reaction of a vigorous poor man to an oppressive capitalistic system' (n.1 above, 86, 104). The relation of these dramas, and of most literary texts, to contemporary ideologies is far more complex than that (and pre-modern societies were not capitalist).

30. This is clear from Hegio's later orders to lighten the chains and let the prisoners walk around: … is indito catenas singulariaslistas, maiores, quibu’ sunt iuncti, demito./sinito ambulare, si foris, si intus uolent (‘Put those individual chains on them, and take off the heavier ones with which they are joined. Let them walk around, outside, if they want, or indoors’, 112-14).

31. Harsh (n.3 above), 348. Obviously it was necessary, as he says, to identify for the audience beyond all possibility of confusion two characters who had already exchanged identity before the play started. But this need does not preclude the spectacle's thematic significance, with which I am concerned here (it has not been much appreciated).

32. For a succinct summary, with bibliography, see Parker (n.8 above), 237. In fact, the Romans had been winning large numbers of slaves in war for several centuries. I simply emphasise the audience's recent experience.

33. This distinction is made by Volkmann, Hans, Die Massenversklavungen der Einwohner eroberter Städte in der hellenistisch-römischen Zeit (Wiesbaden 1961), 14Google Scholar.

34. Lindsay, W.M., The Captivi of Plautus (London 1900), 117Google Scholar. The compound astant (‘they are Standing nearby’) might present a slight problem, since it emphasises proximity rather than the mere fact of standing, which is what is needed.

35. Moore, Timothy J., ‘Seats and Social Status in the Plautine Theater’, CJ 90 (1994–1995), 118Google Scholar.

36. See Moore's discussion of these lines (n.35 above, 118f.). He does not quite make the point I wish to.

37. In this case, there remains what might seem to us an awkwardness with hic, ‘here’, in line 1. In fact, Plautus elsewhere shifts between demonstratives referring to the same person or place. For examples, see the appendix in Elmer, Herbert C., T. Macci Plauti Captivi (Boston and Chicago 1900), 159Google Scholar. They include Capt. 110-12 (where Lindsay silently changes the manuscripts' his to is).

38. Leo, understanding the lines in this way, refers to his apparatus at Aul. 207, where he gives a number of examples (Plauti Comoediae [Berlin 1958, reprint of the 1895 edition], 181, 100)Google Scholar. The transposition of lines 2-3 proposed by Fay, Edwin W., ‘Plautus, Captivi 1-3’, CR 12 (1898), 352–54Google Scholar, with light emendation of line 2, would yield the same result, and sharpen the joke by making the solemn assertion that the speaker will tell the truth precede the deflating punch-line.

39. For the slave's complete dependence on the master, to the point of matching the latter's moods, see Amph. 959-61: atque ita seruom par uidetur frugi sese instituere:/proinde eri ut sint, ipse item sit; uoltum e uoltu compare:/tristis sit, si eri sint tristes; hilarus sit, si gaudeant (‘That's how a good slave ought to behave. He should be just as the masters are; his face should take its expression from theirs. He should be sad if the masters are sad, merry if they are joyful'). A slave, being an instrument, is not supposed to have feelings of his or her own, or any subjectivity.

40. [Erg.] nunc hic [Hegio] occepit quaestum hunc fili gratia/inhonestum et maxume alienum ingenio suo (‘now for his son's sake [Hegio] has begun this trade, dishonorable and completely foreign to his character’).

41. ‘Implied audience’ is used here after the model of Wolfgang Iser's ‘implied reader’ (The Act of Reading: A Theory of Aesthetic Response [Baltimore and London 1978], 34–38)Google Scholar.

42. Lowe, J.C.B., ‘Prisoners, Guards, and Chains in Plautus, Captivi’, AJP 112 (1991), 29–44Google Scholar. Although explaining anomalies in the text by less than smooth adaptation of Greek models is a topos of Plautine criticism and sometimes taken to implausible lengths, Lowe's hypothesis seems to me quite credible.

43. See the similar description of Ergasilus' role by Konstan (n.4 above), 67 n.17. For fuller discussion and a different perspective on it, see Leach (n.3 above).

44. Cf. Men. 457-59, spoken by Peniculus returning in disgust from the forum where he was detained by an assembly: adfatim est hominum in dies qui singulas escas edint,/quibu’ negoti nihil est, qui essum neque uocantur neque uocan:/eos oportet contioni dare operant atque comitieis (There are enough men who eat just one meal a day, who have no business, who neither are invited nor invite others to dinner. Those are the ones who ought to get involved in meetings and assemblies’). Compare the complaints of the first Menaechmus, also returning after a hard day at the forum (Men. 571-601).

45. This for liber captiuos, rather than ‘a free man who has been taken captive’, fits the general context of the play, which is concerned less with the practice of taking prisoners in war than with the servitude that results. See Lindsay (n.34 above), 152, whom I follow here but who gives no reason for his preference.

46. One way of making a slave content and therefore relatively trustworthy is to give him some authority. There is, of course, an implicit irony in the Lorarius' position (not, however, marked by the text). A slave himself, he is assigned to guard other slaves and punish any attempt at escape.

47. So the text for line 199 printed by Lindsay, with Nettleship's emendation. The manuscripts read eamque et erili imperio, and it is also conceivable that the Lorarius would say that slaves' accommodation to servitude makes the master's authority an easier matter for him as well as for them.

48. So in this play Tyndarus was raised as virtually Philocrates' brother, and Philocrates-as-Tyndarus tells Hegio that his servitude in Elis was not all that bad: ‘for me it was as if I were the son of the family’ (nec mihi secus erat quam si essem familiaris filius, 273). This provokes the real Tyndarus’ admiring aside, ‘how he has wittily adapted his speech to servitude!’ (ut facete orationem ad seruitutem contulit, 276). Of course, it is part of the complexity of the drama's playing with social roles through the exchange of identities that Philocrates is actually stating the truth: he really was ‘the son of the family’. This fact sets limits on the levelling effect of the situation of captivity in the audience's minds and is an example of the devices for ‘containment’ through which the play avoids genuine questioning of the social order.

49. For parallels, see Lindsay (n.34 above), 228, ad 444. For discussion, see Segal (n.3 above), 111-16, who emphasises the Saturnalian character of this inversion.

50. On role-playing as an issue in this scene, see Leach (n.3 above), 276f. She is concerned with questions of essential identity; I emphasise instead social identity, or class.

51. See, for example, Franko (n.l2 above), 159, who correctly describes the ambiguity of the word in this passage but ascribes it to the prisoners' manipulation of a single prescriptive notion, whereas I would say that there is a contradiction within the notion itself which is identical to that between the two models of slavery.

52. There, however, line 411 has a different kind of ambiguity, created by an impersonal verb: fecisti ut redire liceat ad parentes denuo (‘you contrived that it be permitted to return to [my/your] parents’, by revealing my birth). Hegio is supposed to think that the speaker refers to himself as beneficiary; the audience knows that he refers to the actual Philocrates, who has acted ultimately in his own interest. The ambiguity points up the question, raised by the whole context, of whether Philocrates will also remain true to Tyndarus.

53. And so Hegio feels entitled to command Philocrates-as-Tyndarus, as the latter's new master, to obey his old master and arrange for the exchange of prisoners (362f.).

54 Konstan (n.4 above, 67) points out that Hegio here assumes the stock comic role of the misanthrope who withdraws from humanity. The play, I think, translates this stereotype into its own terms: Hegio, who has embraced the ‘benevolent’ view of slavery and got burned, reverts to the ‘suspicious’ view.

55. Konstan (n.4 above, 59) suggests that Tyndarus ‘must choose between two obligations, loyalty to Philocrates and obedience to Hegio, and he is bound to violate his responsibilities to the one in acknowledging the claims of the other.’ It seems to me, however, that Tyndarus is never presented with a choice, if that implies that there is ever any doubt about what he will do, in his own mind or the audience's. Konstan himself later (62) says that Tyndarus escapes the dilemma by denying it, but that almost amounts to saying that it never existed, as a moral problem—for him, at least, though the audience might see it that way. Rather than an ‘ethical dilemma’, however, I prefer to speak of contradictory notions of slavery, which the play tries to accommodate.

56. On the effects of this scene, see Harsh (n.3 above), 349, Leach (n.3 above), 278f., and, in more detail, Franko (n.12 above), 159-65.

57. Konstan (n.4 above), 63f.

58. See Slater, Niall W., Plautus in Performance: The Theatre of the Mind (Princeton 1985)Google Scholar. Although the Captiui is not among the plays discussed at length there, I owe a general debt to that book for awareness of metatheatre and other aspects of Plautus' dramaturgy.

59. One result is that when Tyndarus is sent to the quarries his suffering seems quite genuine, for neither he nor we can be sure that Philocrates will return and gain his release.

60. This ‘dramatic irony’ is, of course, evident elsewhere in the play as well (see Harsh [n.3 above], 349, who refers specifically to line 310, and Viljoen [n.l above], 51), but it seems particularly important as one of the layers of communication in this scene.

61. This is especially true if the lacuna in line 420 is to be filled with erus nunc, so that the full picture of reciprocity is expressed. See Lindsay's apparatus to his OCT edition on this line. The supplement he conjectures in his separate edition of the play, <seruus>, would have the same effect, though without the elegance of chiasmus stressing the apparent symmetry in the relationship.

62. See Segal (n.3 above), 201f., and Franko (n.12 above), 165, neither of whom really focuses on the class issue here.

63. On this significance of the gesture here and the suggestion that it alludes to the manumission of Tyndarus (hence a further dimension of meaning), see Freyburger (n.l4 above), 125f.

64. The scene in which Aristophontes unmasks Tyndarus brings out a further aspect of slavery: the distinction, which he still maintains in captivity, between one born a slave and one reduced to slavery as the result of war. See lines 574-77 and 591, and compare Epid. 106-08 and Amph. 176-79. This distinction simply transposes the difference between the two models of slavery by juxtaposing the causes to which they attribute slavery, nature and vicissitude. Aristophontes, by the way, is a freeborn man who evidently remains in captivity after the play has sorted out the ‘proper’ position for the other characters. He thus appears to represent something of a loose end, although the audience would most probably forget all about him after the scene in which he figures.

65. Cf. Viljoen (n.l above), 56: ‘One could almost say that Tyndarus’ nobility of conduct demanded some explanation, and in itself already suggested that it belonged to an ingenuus.’ Why ‘almost’?

66. As it happens, the one example of boni as ‘men of substance and social standing’ that the Oxford Latin Dictionary gives from early Latin occurs in this play, Capt. 583-which, in its context (note the contrast with eges, ‘you are needy’, at 581) seems to show conclusively that the word could be a class term this early. See also Plaut. Cure. 475, Mil. 1288, Poen. 39, Trin. 299; Ter. Phorm. 115; Cato Orat. fr. 58.6. In pointing out the class overtones of Capt. 1034,1 mean to underscore the contrast between my reading of the play and moralising interpretations, which it has often been thought to support.

67. Cf. Leach (n.3 above), 293: ‘His sudden appearance…tends to emphasise his function as a thematic speaker.’ I would say instead that it makes his ideological significance stand out.

68. A sign of the play's polarised representation of the ‘good’ and the ‘bad’ slaves, who are basic to the opposed ideological models of slavery, and hence of its concern with slavery itself, is that Stalagmus is the only slave ‘in all comedy…who is a real criminal’ (Legrand, n.l above, 110).

69. In the text as we have it, this is Tyndarus' own suggestion, and some writers have been troubled that he should be so vengeful. In the context of the relations of slavery, however, it makes perfect sense, and what to modern liberal sensibilities seems deplorable might have seemed unremarkable to an ancient audience living those relations. The line occurs, however, in a passage (1016-23) suspected of being a post-Plautine addition to replace lines that now precede and follow it. One manuscript in fact omits the passage. See the apparatuses of Leo (n.38 above, 220) and Lindsay (OCT). But because someone clearly thought it appropriate that Tyndarus should take satisfaction in Stalagmus' punishment, the line may still be said to reflect ancient attitudes to slavery.

70. Ketterer, Robert C., ‘Stage Properties in Plautine Comedy II’, Semiotica 59 (1986), 111–18CrossRefGoogle Scholar, remarks briefly and in a general way on this significance of fetters, but most of his discussion concerns the way they function as pointers to the identities, qualities, and moods of the various characters.

71. On both the manicae and these bonds, see Allen, Fredric D., ‘On “Os Columnatum” (Plaut. M.G. 211) and Ancient Instruments of Confinement’, HSCP 7 (1898), 41Google Scholar.

72. See Allen (n.71 above), 48f.

73. Allen (n.71 above), 44f., 57. That the shaft extended down the slave's front is indicated by Ergasilus' crude sexual joke about the fettered Stalagmus: boiam terit:/liberorum quaerendorum caussa ei, credo, uxor datast (‘he's rubbing the boia; I think he's been given a wife for begetting children’, 888f.).

74. For the very tenuous relation between slaves in Roman comedy and their historical counterparts, see Stace (n.1 above), 72-75, and Spranger (n.7 above), 90-113, 116-18.

75. For cross-cultural patterns in slavery, both its practice and attitudes arising from it, see Patterson (n.23 above). In this case, one could of course claim that the similarity between these texts is due to their use of a common traditional narrative pattern, perhaps with roots in folk-tale. But that simply pushes the question one step back. Traditional story patterns themselves are ideological productions, connected to the forms of the societies in which they arise and then are adapted. So the issue remains a particular way of representing slavery and the needs to which it responded.

76. On the date of the play, see Wellesley, Keith, ‘The Production Date of Plautus' Captivi’, AJP 76 (1955), 298–305Google Scholar, who feels able to specify not only the year but the month (September, 189 BCE); Grimal (n.l above), 413 (190 BCE); Dumont (n.3 above), 516-21 (c. 198 BCE); Franko (n.12 above), 169f. (late 190's BCE). Arguments for the date are mostly quite tenuous. The best evidence is the reference to the Boii in line 888, which seems to indicate a date late in this decade.

77. The facts surrounding the influx of slaves and their revolts are well summarised by Parker (n.8 above), 236-38, with bibliography. On the social and economic consequences of the increased exploitation of slavery, particularly the shift to slave labour as the foundation of agriculture, see Alföldi, Geza, The Social History of Rome (Totowa NJ 1984), 42f., 56–59Google Scholar, and Brunt, P.A., Social Conflicts in the Roman Republic (London 1971), 17–19Google Scholar. Cf. Thgbert (n.13 above), 146f., and Dumont (n.3 above), 520-21. War, of course, was not the only source of slaves, but it was a major one especially during this period.