The Latin language is uncharacteristically rich when it comes to describing witches.Footnote 1 A witch may be called a cantatrix or praecantrix, a sacerdos or vates. She may be docta, divina, saga, and maga, a venefica, malefica, lamia, lupula, strix, or striga. She may be simply quaedam anus.Footnote 2 The available terms are copious and diverse, and the presence of such an abundant differential vocabulary might suggest (incorrectly, I shall argue) that Latin made clear linguistic distinctions between various witch types. It would seem a reasonable expectation that praecantrices, a word evocative of those who sing of events before they happen (prae + cantare), would be concerned with divinatory practices,Footnote 3 while veneficae, given the term's close relationship to the word for poison (venenum), would deal in potions or philtres, leaving the lamiae (a Latinization of the Greek demon Lamia) or striges (personifications of the rapacious screech owl) to function as quasi-demonic bogeys posing threats to the lives of small children.Footnote 4 However, this expectation of semantic and morphological concordance remains unfulfilled following any concerted attempt to correlate a witch's title with her function.Footnote 5 Because of this disjuncture, this paper proposes to demonstrate not only the inaccuracy of the Latin vocabulary in articulating the functional differences between various witches, but also to assert the essential uniformity of witch characters in so far as each witch is, in essence, a blank canvas onto which a myriad of fears and anxieties may be mapped.
It is true that in many instances there exists a seeming concordance between the name attributed to a witch and the function she fulfils: Erictho is dubbed a Thessala vates as she reanimates a corpse to deliver a prophecy for Sextus Pompey (Luc. 6.651); Hypsipyle labels Medea a venefica for ensnaring Jason with special plants instead of relying on her beauty and good character (Ov. Her. 6.19, 93–4); and Horace, when advising against the composition of highly fictional literature, posits that removing a still-living boy from a lamia's stomach would be an example of excessive fictionalization (Hor. Ars 340). Each of the above names complements the witch's function: the vates is prophetic, the venefica plies potent herbs, and the lamia (a quasi-demonic figure)Footnote 6 has literally eaten a child. However, for each such occasion of apparent harmonic nomenclature there are multiple demonstrations of its discordance. In these instances, terms whose denotations are prima facie quite specific are deployed in situations that belie their problematic nature.
In an oft-cited anecdote, Pliny mentions the farmer Cresimus, who was accused of enhancing his own harvest while depleting those of his neighbours through the use of veneficia (Nat. 18.41–3); this accusation makes little sense if veneficia are merely potions or poisons, since the use of such materials provides no logical explanation for an increase in Cresimus' crops concomitant with a decrease in those of his neighbours. Even though the term veneficia is intrinsically related to the use of poisons (venena) and, while mundane poisoners are repeatedly referred to as veneficae or veneficii,Footnote 7 this is a clear instance wherein the use of veneficia must signify an act quite separate from the simple manipulation of poisonous materials and therefore must be expanded to encompass an ill-defined range of subtle, potentially supernatural, aggressive behaviours.Footnote 8 By extension, witches labelled as veneficae cannot be classified neatly as potion specialists or poisoners, when it is uncertain precisely what actions are and are not inscribed within the limits of veneficium.
This is but one example of the problems inherent in the Latin witch vocabulary. The extent of its nebulous nature is best demonstrated through the presentation of two sets of data. The first is an analysis of the range of attributes described by a single specific term: saga. Which women are labelled as sagae and what exactly are their purported abilities? The second data set examines the range of terms used to describe two specific witches, Meroe and Canidia, for whom there are significant extant narratives. In the first case, it will become clear that a single term can possess multiple, often mutually exclusive associations and can be applied to women who share few relevant attributes. The latter examination will illustrate Latin's lack of a consistent, descriptive witch vocabulary, as well as revealing the tendency of Roman authors to describe witches through the application of multiple terms that operate with conflicting denotative specificities. Each of these analyses will suggest that Romans were not making active linguistic distinctions between different witch-figures, a fact which promotes the idea that the Roman witch was persistently conceived of as a malleable and polyvalent entity.
QUAE SAGA?
The noun/adjective saga conveys at its most basic level a sense of knowledge. Cicero posits that one who is sagacious (sagire) is keenly aware (sentire acute); hence some old women are sage (sagae) because they know many things (multa scire).Footnote 9 That these are specifically old women seems important only in so far as their age has accorded them additional opportunities to obtain knowledge, and the term carries with it no pejorative connotations that Cicero mentions. In this sense, a woman who is saga need not possess any supernatural abilities, merely a sharpened awareness of her surroundings. For this reason, Cicero applies the root of this term equally to dogs (sagaces canes). Moreover, he makes a sharp distinction between this keen perceptiveness (sagire) and the ability to perceive the future (praesagire). As far as he is concerned, sagae cannot predict the future. One who perceives events not in the moment but in the moment beforehand (ante sagit quam oblata res est) is said to presage (praesagire), or to sense events before they occur (futura ante sentire). This does not apply to those who are simply sagae.
Conversely, Columella specifically groups sagae together with the divinatory haruspices and advises that both sets of practitioners be avoided lest their empty religious observation (vana superstitione) drive base individuals to needless expense and ultimate disgrace.Footnote 10 That sagae are here discussed in explicit combination with haruspices and are actively counselled against owing to their appeal to the uneducated (rudis animos) suggests that Columella's sagae are of an altogether different sort from those sage old women described by Cicero. If their association with the haruspices can be read as a reliable indication of their practice, these sagae appear to be a type of professional consultant, presumably available for a fee to offer predictions about the future. This interpretation would be commensurate with the description of a Syrian saga in the retinue of Gaius Marius; Frontinus explains that Marius kept this saga so that through her knowledge he could pretend that he himself knew the outcome of battles before they even occurred (Str. 1.11.12: dimicationum eventus praediscere simulabat). Although Cicero understood sagae to be older women possessed of knowledge separate from that held by those who could predict the future, Columella and Frontinus clearly considered sagae to be women who were explicitly skilled in the divinatory arts.
This difference regarding the powers of a saga might be attributed to the roughly one hundred and thirty years dividing Cicero's De divinatione from Frontinus' Strategemata; the intervening time could have allowed the term to evolve and expand from describing women who were merely wise to include those who were gifted with preternatural foresight. However, the evidence offered by a rough contemporary of Frontinus suggests otherwise. In Martial's Epigrammata a saga is mentioned on two separate occasions and in neither instance appears to possess any powers of divination: each saga is instead invoked as a purifier, specifically of dreams:
In so far as their function is concerned, these sagae are unrelated to those envisioned by either Cicero, or Frontinus and Columella; they possess no innate wisdom and they offer no predictions. They are drunken purifiers.Footnote 11 The second-century grammarian Pompeius Festus describes sagae similarly, although without mention of their predilection for wine, and includes them in a list of women adept at purification: ‘A priestess (sacerdos) experienced in rites of purification was called a purifier (piatrix), labelled by some as a shifter (simulatrix), by others a wise woman (saga) and others an expiator (expiatrix).’Footnote 12
These unique, concomitant perceptions of the saga indicate that the lack of any consistent definitions of this figure among Roman authors cannot be attributed solely to a temporal semantic shift. Martial knows sagae to be purifiers while his contemporaries understand them as diviners, whereas Cicero had previously drawn a clear distinction between diviners and sagae, yet made no mention of their work as purifiers. Moreover, Horace, as a contemporary of Cicero, would potentially have shared Cicero's understanding of sagae but instead cast them as either purifiers (Carm. 1.27.21–2: quae saga, quis te solvere Thessalis | magus venenis, quis poterit deus?) or scornful objects of empty and superstitious dread (Ep. 2.2.208–9: somnia, terrores magicos, miracula, sagas | nocturnos lemures, portentaque Thessala rides?)Footnote 13 Even the Roman elegists Tibullus, Propertius, and Ovid, all contemporaries of Horace, present yet another interpretation of the sagae. Within the elegiac corpus, these women are not diviners or purifiers per se but are regularly referred to as specialists of erotic magic. Tibullus twice mentions the powers of a certain saga, once when she provides a spell to deceive his lover's husband (1.2.43–66) and again when he curses the saga, this time for counselling the object of his affection to value wealth over love (1.5.47–60).Footnote 14 Propertius' Thessala saga was unable to rid him of his love for Cynthia (3.24.9–10), and Ovid considered that it might have been a saga at the root of his otherwise inexplicable impotence (Am. 3.7.27–36). Such erotic associations, while commonly attributed to sagae in Roman elegy, are simply not employed by Horace, Cicero, Martial, Columella, Festus, or Frontinus, whereas sagae with purificatory or divinatory skills are wholly unattested in the elegiac corpus.
Yet sagae as practitioners of erotic magic do occur in Apuleius' second-century novel, the Metamorphoses, although with considerable additional powers. The witch Meroe, described by the traveller Aristomenes as wise and godlike (saga et divina), is known for her ability to compel men the world over to fall passionately in love with her (1.8), and another unnamed saga is sought out to reconcile a wife to her husband (9.29). The similarities that these women bear to their elegiac counterparts end there. The anonymous saga, after failing to assuage the husband, summons the ghost of a violently slaughtered woman to kill the man instead (9.30–1) and Meroe's other supernatural skills include transforming her enemies into animals, preventing a woman from giving birth, locking an entire city indoors, and transporting a man's home to a barren mountain peak (1.9–10). Other sagae appear in Apuleius' Metamorphoses as nameless, menacing shape-shifters (versipelles) – vile women who gnaw off the faces of corpses and attempt all manner of evil to satisfy their lust (2.21–30).
Although both Apuleius and the Roman elegists present sagae in specifically erotic contexts, it is clear that the sagae of the Metamorphoses are operating on a different plane from those of Roman elegy. The elegiac sagae are professional practitioners of erotic magic, consulted and paid by both men and women either to help or to hinder amorous relationships, whereas Apuleius' sagae are not exclusively magical professionals available for hire. His sagae possess skills far beyond simple amatory magic, and their behaviour – including corpse defilement, mordant facial disfiguration, and the commission of dispassionate murder – borders on the monstrous.Footnote 15 None of these sagae, neither Apuleius' nor those of Roman elegy, can be reconciled with those conceptualized as mere diviners or purifiers.
Even the attempts of the third-century scholiast Porphyrion to define the term saga ultimately only serve to further confound its nature. He glosses Horace's purificatory saga of Carmen 1.27:
sagae sunt praecantatrices, quae vel arcessere carminibus mala hominibus possunt vel expellere. (In Hor. Carm. 1.27.21)
Sagae are praecantatrices who, with their songs, can either provoke or prevent evils for men.
Porphyrion's definition is problematic for a number of reasons, not the least of which being that praecantatrices is a hapax legomenon. Still, it is a likely variant of praecantrices, which is found in both Plautus and Varro. In Plautus, a praecantrix is named first in a list of female divinatory professionals (praecantrices, coniectrices, hariolae, haruspicae) and can thus be presumed to fulfil that same role in this instance (Mil. 693). Varro, however, perceives them more as healers since, in a passage on child-rearing, he relates that many women prefer to consult praecantrices rather than doctors.Footnote 16 Thus Porphyrion's gloss generalizes sagae as diviners (or perhaps he means purifiers?) who can bring either injury or aid. While sagae have been variously described as monstrously lecherous women, professional practitioners of erotic magic, purifiers, diviners, or intelligent women who are specifically distinct from diviners, Porphyrion has here removed virtually every division and condensed a multitude of mutually exclusive interpretations into one nebulous definition. The polyvalence of Porphyrion's gloss is a prime example of a term that possesses almost as many definitions as it does attestations.
As a definition, the term saga is all but useless. It has at least five specific denotations (wise woman, diviner, purifier, figure of superstitious dread, erotic magic professional), of which many are simultaneously activated and mutually exclusive. The subset of all Roman witches described as sagae is so collectively disparate that only the broadest of generalizations can be made, namely that their power (whatever that may be) derives from some sort of ambiguous knowledge. Were certain outliers overlooked, it could be concluded that many sagae use spells (carmina, cantus, etc.) in the application of their art and they tend to be able to effect reversals in the natural order (stopping rivers, drawing down the moon, raising the dead, and so forth). Nevertheless, these observations do not account for all Roman sagae, nor are these observations different from those that can be made regarding characters described as magae or veneficae. In short, the Latin vocabulary cannot provide a reliably accurate method for categorizing or conceptualizing the Roman witch.
THE WITCHES MEROE AND CANIDIA
The following explorations of Apuleius' Meroe and Horace's Canidia will demonstrate not only a muddled confusion of generic terms but also the stark contrast between these characters and other witches with overlapping descriptive vocabulary. In this section, the capability of witches to effect inversions of natural order will not be addressed; literary descriptions of such abilities occur frequently and are ascribed to such a wide range of characters that they would appear to have little purpose other than to act as a generic indication of supernatural power – Daniel Ogden goes so far as to dub these ubiquitous descriptions ‘thumbnail sketches’.Footnote 17 These sketches are rarely informative of the capacity in which a given character serves and the litany of natural inversions effected is not merely limited to witches: Ovid employs a similar list to describe the potency of his own carmina (Am. 2.1.19–28) and such powers are also (although to a lesser extent) attributed to the poet Orpheus (Ap. Rhod. 1.23–31; Ov. Met. 10.86–105). Thus, although this descriptive language is widespread, its very prevalence negates its utility in understanding the nuances of the witch characters to whom it is applied.
Apuleius' Meroe
Meroe is a stereotypically lustful innkeeper in Apuleius' Metamorphoses (1.6–19), whose primary narrative function is the seduction of a traveller by the name of Socrates. Their relationship is described by Socrates in terms of chronic illness – ‘from our first moments together I caught a sickening, long-lasting disease’ (annosam ac pestilentem condicionem contraho, 1.7) – and it is clear to Socrates' friend Aristomenes that Meroe is keeping him as her lover against his will, while his wife and family at home mourn him as dead. Meroe's other powers are equally significant and include the ability to transform her enemies (former lovers among them) into animals, lock entire cities inside their homes, and transport a single house to the peak of a distant mountain. Her behaviour is monstrous: upon discovering Socrates' plans to escape, she breaks into his room, slits his throat with a sword, drains his blood, removes his heart, then stops the wound with a sponge before urinating upon his cowering friend and vanishing into the night. Marvellously, her magic is powerful enough to restore the butchered Socrates to the image of health, but dissipates the instant he drinks water, leaving him a ghastly and bloodless corpse.
The behaviours and abilities here attributed to Meroe are significantly different from those found among the majority of other witches in Latin literature:Footnote 18 the commission of murder for its own sake is rare and the use of a sword for such an act is elsewhere unattested;Footnote 19 the power to transform other humans into animals is reserved only for the gods;Footnote 20 her ability to transport individual homes over long distances and to lock the population of an entire city indoors are wholly unique to Meroe; and the debilitating effects of her sexual relationship with Socrates, in conjunction with her deliberate exsanguination of him are, so far as I am aware, otherwise undocumented in the Latin literary corpus.Footnote 21
Despite her distinctive character, Apuleius does not describe Meroe through the use of any unique vocabulary; the words he does employ are too vague to offer any but the most banal of interpretations. As an old woman (anus) Meroe is in the company of a host of different characters, from a practitioner of erotic magic whom Ovid suggests his readers avoid (Rem. 254) to the Cumaean Sibyl (Fast. 4.15) and even the venerable Penelope (Prop. 2.9.8), and while many witches labelled as anus are noted for their physically repulsive aspects,Footnote 22 Socrates evidently finds MeroeFootnote 23 still somewhat attractive (admodum scitulam).Footnote 24 She has godlike abilities (divini potens) and she makes use of graveyard spells (devotiones sepulcrales) as well as the power of supernatural forces (numinis ministerium), but these are no more than generic descriptors for magical practices. Although she is labelled as a saga, the word's wide range of meaning and Meroe's impressive battery of diverse skills creates significant difficulties in determining the precise significance of the term in this context.Footnote 25
The most intriguing term that is applied to Meroe, lamia, ultimately provides little useful comparanda (1.17: at ille, odore alioquin spurcissimi humoris percussus quo me lamiae illae infecerant, vehementer aspernatur); it is clear from passing references in Lucilius and Horace that the term refers to a demonic female figure characterized by her hunger (specifically for human flesh)Footnote 26 but no parallels exist wherein another witch is likewise described as a lamia. For all Meroe's uniqueness as a Roman witch, the Latin vocabulary is bereft of terms with which to detail her nature specifically. Lamia comes the nearest, although Apuleius' reapplication of the word to Psyche's mundane, albeit vindictive, sisters (5.11: si posthac pessimae illae lamiae noxiis animis armatae venerint) indicates that the term may simply be one of general opprobrium that is without deeper insight into Meroe's character.Footnote 27
Horace's Canidia
The witch Canidia is a character who frequently recurs throughout both Horace's Epodes and Satires. While Horace's portrayal of her shifts significantly between poems,Footnote 28 her core identifying characteristic – that of being an aged witch – remains largely constant throughout her most extensive appearances,Footnote 29 even though in all her guises she is never as powerful as Lucan's super-witch Erictho, nor is she a commonplace professional of erotic magic. At her most powerful she is the leader of a small collection of witches gathered together to starve a Roman child to death in order to create a love potion from his desiccated viscera (Epod. 5), and at her meekest she can be frightened away from her magical conduct by a flatulent Priapic garden statue (Sat. 1.8). Horace constructs her alternately as an object of derision or a cause for fear, yet at times simply invokes her in passing as a generic bringer of mild misfortune (Epod. 3; Sat. 2.1, 2.8). However, although Canidia is addressed by name in six poems, Horace rarely reuses any of her descriptive terminology and never explicitly labels her with a noun or adjective that would positively designate her as a witch. Conversely, two scholiasts commenting on Horace's work provide ample witch terminology for Canidia, but the meaning of those diverse terms quickly clouds as the scholiasts apply them with seeming interchangeability.
Just as with Meroe, anus is the only term that is applied to Canidia on more than one occasion (Epod. 5.98, 17.47). As a label, anus is ill-informative in its generic nature and conveys little about Canidia that is not covered by other descriptions of her. The unnamed victim of Epode 5 names the assembled witches ‘hideous old women’ (obscenas anus, 98). It is a parting insult that adds little colour to the women whom he had previously characterized as bestial (uti … | belua, 9–10), savage (truces, 4), animalistic (ut marinus … | echinus aut currens aper, 27–8), and impious (impias, 84). In Epode 17 the narrator professes somewhat cryptically that Canidia is not ‘an old woman skilled at scattering ashes’ (prudens anus | novendialis dissipare pulveres, 47–8). It is unclear whether this denial is of her age (has she outlived so many people as to become adept at performing funeral rites?) or of her unsavoury character (does she frequently produce corpses that must be disposed of?), but in the light of everything else that had been said of her in Epode 5 and Satire 1.8, this description of Canidia as an anus does not enhance our understanding of her either as a character or as a witch.Footnote 30
In addition to being called an anus, Canidia is compared to (but never explicitly labelled as) a venefica and to the practice of veneficium, but this terminology, much like that of the sagae, is imprecise and its implications uncertain. In Epode 5, she laments that her dread potions (dira | venena, 61–2) have failed to win over the object of her affections and assumes that the man has been freed from her influence by a more knowledgeable venefica's spell (veneficae | scientioris carmine, 47–8). In this context, the word venefica might serve as a term of derision whereby Canidia insults another witch whom she sees as being in a separate, lower class of magical practitioners.Footnote 31 On the other hand, it could be a professional term describing a rival witch with similar abilities to Canidia, or it could even refer to a woman whom Canidia perceives as belonging to a more elevated and powerful class than she. If either the first or third readings are favoured, then Canidia is not a venefica and the only thing brought to light is a venefica's nondescript ability to extricate someone from the influence of poisons. If the second reading is favoured, Canidia and her magical opponent are titled veneficae, but the implications of such nomenclature remain unclear.
The reference to veneficium at Epod. 17.56–9 is no more revealing. There, Canidia asks the anonymous narrator:
The terminology again raises more questions than it answers. The pontifex venefici may have been invoked because such a figure would ostensibly have been immune to Canidia's vengeance. It may also have been the reverse case, wherein the pontifex venefici was a character of considerably lower ability than Canidia, making the addressee's vain assumption that such meagre powers would suffice for protection quite laughable in Canidia's eyes. In either case, the power and scope of veneficium, as well as Canidia's relationship to the practice, remain inexplicit.
Whatever Canidia's association with the words veneficae and veneficia, it was strong enough for the scholiast Porphryion to refer to her repeatedly as a venefica (In Hor. Epod. 3.8, 17.2; Sat. 1.8.1). Yet the connotations of this term remain solidly unclear. Porphyrion uses it elsewhere of Medea (In Hor. Carm. 2.13.3) in a roughly analogous sense (both are women who seek to influence men sexually through supernatural means), but the term is frequently used to describe women who exercise significantly different powers from either Canidia or Medea: venefica is used by Lucan of the necromantic Erictho (6.581),Footnote 32 by Ovid of an unnamed bringer of impotence likewise labelled a saga (Am. 3.3.79),Footnote 33 and by Curio of a woman whom he blamed for an embarrassing lapse of memory while in court (Cic. Brut. 217). More confusingly, Pliny specifically draws a distinction between the magical arts (magicas artes) and those employed by veneficae/venefici (veneficas artes), stating that he considers the magical arts loathsome, false, and empty, yet finds in the arts of the veneficae some actual efficacy (HN 30.18).
As if the definition of venefica were not sufficiently ambiguous, the scholia from Porphyrion and Pseudo-Acron employ several other terms in relation to Canidia that they treat as operating on a semantic par with venifica. Porphyrion, who is more consistent in his witch terminology than Pseudo-Acron, clearly prefers the venefica label but does once refer to Canidia as a malefica.Footnote 34 (Malefica is no less troublesome a term than venefica – the objects of its application are similarly broad and the word carries with it no absolute definitions.Footnote 35) Meanwhile, Pseudo-Acron not only describes Canidia as both venefica and malefica but also refers to her as maga and saga.Footnote 36 None of these words is used with any more specificity than the next; each of these four terms can be used with equal validity to describe Canidia, or potentially any other Roman witch. Although these scholiasts are writing several centuries after Horace, the generic and non-specific nature of the Latin witch vocabulary that they demonstrate is applicable throughout.
CONCLUSION
The vast collection of witches attested in Latin literature is a highly diverse group. Figures such as Circe, Medea, Deianira, the women of Virgil's eighth Eclogue, Canidia, Dipsas, Acanthis, Meroe, Pamphile, Petronius' strigae, Martina, and Erictho share only the broadest common denominators – that of being a female in possession of vaguely defined supernatural abilities – and beyond that the list is highly diverse: divinities appear alongside mortals, professional and lifelong practitioners with desperate amateurs, youthful maidens with elderly hags, loving wives with grave-robbers, and fictional characters with historical figures; appeals to their unity on the grounds of geography, morality, and intentionality are likewise defeated. Conversely, given the polyvalence of the Latin witch vocabulary, we must acknowledge that on some level there existed an absence of a Roman conceptual differentiation between one witch and another, regardless of their potential disparity.
However, the lexical ambiguity used of these witch characters fits well with the presence of their many observable differences if we understand these figures to be operating on the same functional level as demons – that is to say, as generic figures with supernatural powers that vary widely based on context. Sarah Iles Johnston describes demons as
[the] clay with which people mold images of their fears and anxieties; in order to express the fears and anxieties of the moment effectively, that clay must remain malleable. It is not until those who stand outside of a community begin to make lists of demons (i.e., demonologies) for their own purposes that any real consistency of traits and imagery is obtained, and it is an artificial consistency, born of a scholar's desire to organize, a magician's desire to control, or a missionary's desire to devalue and eventually overcome.Footnote 37
It is in this capacity that the Roman witch is operating – as a character onto which authors can project the ‘fears and anxieties of the moment’; it is only we scholars who have created a sense of ‘artificial consistency’ that purportedly encompasses the witches' varied incarnations when the only significant consistency uniting these characters is their tendency to behave so inconsistently.Footnote 38
In the light of this, I suggest that, rather than relying upon the demonstrably inaccurate Latin witch vocabulary as a means by which to ascertain the nuances of a particular character, we instead look past the terminology and assess witches primarily on the basis of the specific attributes that they manifest in a work. It is of the utmost importance to recall that any traits that these witch characters possess are the direct result of a conscious decision made by their author and, as such, they need not be seen to reflect some adherence to a ‘real’ external source.Footnote 39 Therefore, by analysing precisely how an author has chosen to portray a certain witch in a certain situation (rather than the terminology deployed to describe her), we may expect to gain a deeper understanding of that particular character in that particular context. Not only will this approach yield more accurate analyses of the roles and functions of specific witch characters, but the interpretations resulting from these analyses will also provide more nuanced readings of the texts in which these witches appear.
