¹ Hospitilla (Turnebus, Adversaria xii.1), Hypsithilla (Scaliger; diminutive of Hypsithea), Hispitilla (Vossius), Iphitilla (Guyet); Ipsitilla (Buecheler, Schwabe), Ipsicilla (Froehner, , Rh. Mus. 13 [1858], 148Google Scholar, Herz, ibid. 17 [1862], 325).

² A long time ago the present writer pointed out (Glotta 44 [1966/7], 174–6) that the technically most economical alternative to Ips-ic-illa would be I-psith-illa from Psithia ‘(kind of) Grape’ (LSJ ψθιος) with the connotations erotic and evocative of wine that go with names like Astaphium, Lesbia, Staphyla; for the Romans the major festivals of Venus were the Vinalia (April, August), and of course sine Cerere el Libero friget Venus (Ter. Eun. 732; cf. Otto, Sprichwörter der Römer s.n. Venus). In this case the diminutive would be formed in the same way as Septim-ille (from Septimius, 45.13), which presents no problem, but on this view we have to interpret the initial vowel as anaptyptic, a striking vulgarism only explicitly attested in much later Latin (ipsilion, ipsyllium = ψλλιον CGL iii.557.46, 558.64, 622.51, ipsatirus = ψαθυρϲ, Oribas. Syn. 4.1, Ypsichius = Psychius, CIL 13.3826 (a Christian epitaph). That is not a quite fatal objection, for ps-words are not common, and this kind of anaptyxis should in any case be distinguished from rather than compared with the well-known and increasingly systematic development in Imperial times of a supporting prothetic vowel in native words beginning sc- st- sp- in the spoken Latin from which the Romance languages descend (cf. Leumann, M., Lateinische Laut- und Formenlehre [Munich, 1977], p. 104)Google Scholar. For initial ps- in Greek names and loanwords may always have been presenting some Latin-speakers with a pronunciation-problem to which the anaptyxis would have been a ready solution (there is similarly late antique evidence for t- and pit- as Latin renditions of pt-, which, however, will again always have been a difficult combination for Latin speakers). If this analysis were right, then the name would ‘speak’ with linguistic irony: by her pronunciation, the speciously exotic Hortense de Paris betrays that she really comes from … well, some nearer and less romantic background. But this explanation, which I repeat here only to repudiate it, involves a linguistic assumption temerarious in the absence of better corroboration and I now agree with Morgan, M. Gwynn (Glotta 52 [1974], 233–6)Google Scholar, though not for the reasons he offers, that there is more outrageous point to the poem if the addressee is meant to be of superior rather than inferior social status, as indicated above and in what follows.

³ O is by and large a poorer representative of what Catullus actually wrote than X (the lost source of G and R), but is in some ways a more faithful representative of the common source V, for here and there it preserves the truth where the scribe of X has gone wrong, sometimes by accident, sometimes by over-ambitious ‘correction’ not attempted by the humbler scribe of O; cf. e.g. 11.5 arabasue O, arabasque X; 24.4 mi dededisses O, mi dedisses X; 26.1 uestra O, nostra X; 29.19 scit O, sit X; 34.21 sis … placet O, scis … placent X; 49.7 patronus O, patronum X; 63.27 notha O, nota X; 64.139 blanda O, nobis X, 179 ponti O, pontum X; 183 lentos O, uentos X; 219 cui O, quem X; 231 tum O, tu X; 353 messor O, cultor X; 67.72 ullo O, nullo X; 68.50 alii O, ali X; 79.4 notorum O, natorum X (this list is merely illustrative). In view of the easiness of the merely palaeographical confusion of lumen / limen it is impossible to say for certain whether liminis in X was just a slip or a ‘correction’.

⁴ Cf. V. Aen. 6.255 ecce … primi sub lumina (PR Servius, Claudius Don., Char. 307 B; limina FM, both readings Probi centonis 160) solis et onus sub pedibus mugire solum, cf. 8.69 aelherii … orientia solis lumina.

⁵ We are obviously not meant to bother about the practical details. Ground-floor windows in Greco-Roman town-houses were normally set much higher up and the street-aperture was typically much narrower than in modern houses, obviously for security (cf. Plin. N.H. 19.59); at Pompeii, they range from mere slits (2½″) to about 12″ in width and c. 30″ in height: at best, a tight squeeze for the sneak-thief! The inner aperture was of course much wider, to maximize illumination. The outer apertures of upper-storey windows could be appreciably wider, affording a good view of a procession (Ovid, Met. 14.752), a vantage point for an address to the crowd (Livy 1.41.4) or for soliciting (Martial 11.61.3), or affording an easy jump to suicide (Juv. 6.31). Upstairs and down, removable shutter-boards, single shutters hinged at the side or at the top, and double-hinged shutters are all attested by the literary and archaeological evidence, grills too, but in their absence, good stout bars were obviously essential for any but the narrowest downstairs windows. See Daremberg, C. and Saglio, E., Dictionnaire des antiquités grecques et romaines, t. 2 (1896), 1032ff., 1037–8Google Scholar (shutters), Mau, RE 6, 2180–5 fenestra, Herbig, , Rh. Mus. 44 (1929), 260ffGoogle Scholar., Cressedi, G., Enciclopedia dell' Arte antica classica ed orientate, ed. Bandinella, R. Bianchi, t. 3 (1960), 694ff.Google Scholar

⁶ The same conjecture was subsequently proposed, in several cases certainly independently, by Dousa, senior (Praecidanea pro Catullo, Antwerp, 1582)Google Scholar, Casaubon, , Gifanius, , Livineius, , Heinsius, , Bentley, , AND Mehler, E. (Mnemosyne 2 [1853], 180).Google Scholar

⁷ E.g. adiubant, Fronto p. 120 N, adiubauit, Itala cod. Ottob. gen. 49. 25, aiubante carm. epigr. 1807), and paronomasiae involving iubeo and iuuare become a mannerism in later Latin once intervocalic /w/ and /b/ had wholly coalesced as /β/, cf. TLL 7.583.76ff. For confusion of iubeto and iubebo cf. e.g. V. G. 3.329 (whichever there be right).

⁸ iubeo hoc / Mud, as well as both iube ueniam (cf. 3) and iubeo ut ueniat are constructions attested from Plautus right through Latinity beside iube me uenire, iubeo eum uenire; but iubeo ne ueniat (as opposed to ueto [ne] ueniat) first appears only in Tertullian, cf. TLL iubeo. That rules out the punctuation et si iusseris illud, adiubeto (or adiubebo) / ne quis … obseret tabellam, but readily allows iubebo illud, (scilicet), ne quis … obseret tabellam.

⁹ …paresque nobis sounds as though it should lead innocently to ‘…a pleasant little lunch à deux’ (e.g. ‘…priuatim leue prandium iocosque’ or the like); the comic surprise is not only in the substance but in the way it is expressed – a concentrated tricolon crescendo, words two and three of increasing length.

¹⁰ An equally massive three-word line, differently articulated. The close pairing -que … -que (itself mock-heroical) rules out the possibility of taking pallium as ‘blanket’: the point here is not that Cat. has just had breakfast in bed (Quinn ad loc.), but that he is dressed and ready to go – and in appropriately dandified garb. The last two words explicitly identify Don Juan as a greasy Graeculus, or a shamefully un-Roman Roman, cf. Cic. Ver. 5.31 cum iste cum pallio purpureo talarique tunica uersaretur in conuiuiis muliebribus, Suet. Tib. 13.1, etc. Dressing like that is a reproach in itself according to the conventional Roman grauitas at which this poem is so clearly cocking a cheeky snook.