In spite of the reader's curiosity, the narrator will not break the secrecy of Isis’ mysteries. Precisely an act of curiositas had brought about Lucius’ transformation into an ass; at the end of the novel he demonstrates that he has learnt his lesson.Footnote ²

That something has dropped out, most probably between lingua and illae, is clear.Footnote ³ It has frequently been assumed that illae temerariae curiositatis, referring to aures, needs a corresponding expression of similar length and structure (deictic in nom. + adjective and noun in gen.), referring to lingua. This hypothesis is in line with Apuleius’ predilection for parallelism and stylistic richness. Van der Vliet, followed by Robertson, proposed incorporating ista impiae loquacitatis after lingua;Footnote ⁴ Nicolini, followed by Zimmerman and the recent Groningen commentary, suggested illicitae intemperantiae ista in the same position.Footnote ⁵ Corrections like these are elegant and perfectly plausible, but highly invasive.

I wonder whether the solution might not be much simpler. The risk is that, in a contrasting pair (dicerem ~ cognosceres), the narrator's tongue and the reader's ears may commit an ‘analogous crime’ (par noxa). The crucial point, therefore, does not seem to be so much the distinction between the two behaviours (emphasized by the supplements mentioned above) as their close collaboration, which results in the perpetration of the intellectual sin of curiositas by the reader. Along this line is S. Harrison's recent emendation sed parem noxam contraherent et aures et lingua, <ex> illa[e] temeraria[e] curiositate .Footnote ⁶ To my eyes, this is an excellent ‘diagnostic’ conjecture. It does not postulate a huge lacuna, thereby opening the way to supplements less speculative than the ones put forward so far. However, there is no need to suspect that so many terms were involved in the corruption.

I would suggest: sed parem noxam contraherent et aures et lingua, <anc>illae temerariae curiositatis.

The metaphorical use of ancilla and synonyms in reference to passions, urges or vices is well attested in Latin authors.Footnote ⁷ In Apuleius, cf. Met. 5.23.5 lucerna … amoris uile ministerium ; 4.22.7 (of Lucius as an ass) uentri tam profundo seruiens ; 7.27.3 uoracitati suae deseruit ; Apol. 12.2 animalium serua corpora (i.e. won over to lust); Met. 9.14.5 matutino mero et continuo stupro corpus manciparat . Psyche is often referred to as ancilla Veneris, which allegorically signifies the subjugation to love of the human soul (Met. 6.8.2, 6.8.6, 6.9.5, 6.10.2).

The image of the body or specific parts thereof as ‘attendants’ of the intellectual part is common all through the history of ancient culture.Footnote ⁸ One text in particular is worth singling out here, since it was famous and demonstrably known to Apuleius. When describing the function and reciprocal relations of the parts of the body in the Timaeus, Plato repeatedly has recourse to terms designating command and obedience: for instance, the head is the lord of the rest of the body that has been assigned to it as a servant (44d δεσποτοῦν … ὑπηρεσίαν); the chest, as seat of the courage, is meant to obey reason (70a τοῦ λόγου κατήκοον); through the blood the commands of the mind are conveyed to the limbs (70b παρακελεύσεων … ἐπήκοον … ἕποιτο … ἡγεμονεῖν); the genitals of males are disobedient and rule by themselves (91b ἀπειθές τε καὶ αὐτοκρατές). Most importantly for this paper, according to Plato the human tongue was expressly devised by the intermediary divinities so that the flow of speech might ‘serve’ the intelligence (75e γλώττῃ … ὑπηρετοῦν φρονήσει). Apuleius explicitly tells us that he knew and appreciated this section of the Timaeus (Apol. 49.1–2).

In the section of his De dogmate Platonis Footnote ⁹ devoted to the human body (1.13–18) Apuleius himself repeatedly employs master/servant terminology: the soul is the master of the body (esse animam corporis dominam , 13); the less noble parts of the body are ‘ruled’ by the better ones and merely provide a nutritional ‘service’ (inferiora reguntur optimatium praestantia et ipsa ministerium suggerunt uictuale, 15); everything below the neck ‘serves’ the head (cetera enim membra ancillari et subseruire capiti, 13; pedes ceteraque umerorum tenus capiti oboediunt , 15); the various organs of the head (no doubt including ears and tongue) are ‘instruments’, the function of which is to ‘assist’ the mind ( machinamenta … in conspectu rationis, ut intellegendi ac persentiscendi ueritas adiuuetur , 13); one page later, as in our passage, ears and tongue are mentioned in sequence with special emphasis put on the service they provide to the intellect (auditu … per quem prudentiam sapientiamque condisceret … lingua … promptuarium potius rectae rationis, 14). For the notion of parts of the body as ‘servants’ in Apuleius, cf. also De mundo 27 manus ad Footnote ¹⁰ ministerium praesto erunt.

Two parallels for my proposal come from the Apologia. At 3.7 Apuleius says that certain ranting speakers are used to hiring out (the service of) their venomous tongue to the acrimony of their clients (solent linguae suae uirus alieno dolori locare ). The concept is similar to that presupposed by our passage: in both cases the tongue is depicted as being wilfully subordinated (ancilla ~ locare) by its owner (Lucius ~ speakers) to a negative urge (curiositas ~ dolor) of someone else (reader ~ clients).

Even more compelling is a later passage in which the tongue is ironically presented as an ‘attendant’ serving up lies and bitterness: Apol. 8.3 lingua, mendaciorum et amaritudinum praeministra . This passage is particularly significant owing to its structural correspondence to ours, with the synonyms ancilla ~ praeministra both being nouns in apposition immediately following the nominative lingua and both governing genitives that specify the type of vice ~ unpleasant things.

Also interesting for their structure are De dog. Plat. 9 detur enim uenia nouitati uerborum, rerum obscuritatibus seruienti #Footnote ¹¹ and De mundo 36 omnium animalium agrestium et cicurum, pinnatarum et pedestrium et aquatilium natura gignitur, nutritur, absumitur, parens caelestibus institutis #: as in the Met. passage, a larger sentence is closed by a clearly marked-off three-word explanatory addition (here a participle, in the Met. a substantive in apposition) designating ‘service’ or ‘obedience’.

For appositions consisting of a noun governing a genitive at the very end of a sentence (the apposition being attached to the subject) in Apuleius, cf. Met. 2.15.6 mensula … et calices … et lagoena …, prorsus gladiatoriae Veneris antecenia #; 5.22.7 iacebat arcus et pharetra et sagittae, magni dei propitia tela #; 7.10.1 sic ille latronum fisci aduocatus nostram causam pertulerat, uirginis et asini sospitator egregius #; Apol. 77.6 matri suae, praeter ingenium placidissimum immobili, iram quoque sua expostulatione accessisse, non mediocre pertinaciae alumentum #; Flor. 22.2 quin eo tempestiue Crates interueniret, litium omnium et iurgiorum inter propinquos disceptator atque arbiter #. A further element of resemblance between these cases and ours is that the nouns in apposition refer to analogously ‘ancillary’ functions (prelude, instrumentality, support, assistance). At Met. 6.9.2 ubi sunt, inquit, Sollicitudo atque Tristities, ancillae meae? #, we find the same sequence as in our passage: two feminine nominatives connected through ‘and’ (cf. aures et lingua) > appositive ancillae > genitive ~ possessive. Cf. also (in a different case) the final apposition of minister at Met. 6.31.4 … de isto asino semper pigro quidem, sed manducone summo, nunc etiam mendaci fictae debilitatis et uirginalis fugae sequestro ministroque #.

The loss of anc- can be explained in at least two different, yet not necessarily incompatible, ways: owing to the semantic autonomy of the segment -illae Footnote ¹² and owing to haplography with respect to the preceding syllable. In the latter case, whether the corruption of lingua into linguae took place before or after anc- was lost would be immaterial, since an haplographic dropping of anc- might be facilitated both by the contiguity in scriptio continua of the same letter (linguaãcillae) and by the paleographic similarity, in many minuscules, of ae and ac (linguaeãcillae).