Kiss me, kiss me, kiss me: for the Latin lover

. . . and then that last labiolingual basium might be read as a suavium if whoever the embracer then was wrote with a tongue in his (or perhaps her) cheek as the case may have been then. . .

— Finnegans Wake, p. 122.


Problem and method

Okay, so this post is not exactly short: but it is, I think, plain and readable. Steve from Languagehat wrote to me the other day, asking if I had any insight into the oft-repeated factoid that the Romans had three different words for kissing. I replied that I had indeed heard of such a distinction—in fact I mentioned it here—but that I had little first-hand knowledge of its basis. Here's the general line, as peddled by old man Wikipedia:

The Romans distinguished three types of kiss: osculum, a friendship kiss on the cheek; basium, a kiss of affection on the lips; and suavium, a lovers' deep kiss.

Here's a similar account in Spanish:

Los romanos distinguían tres tipos: 'osculum' era el que se daba en la cara, entre amigos; 'basium', el que se daba en los labios, y 'suavium', el erótico, el que se daba durante una relación sexua.

And in Dutch, for good measure:

Drie soorten Kussen, zei ik. Volgens Rome. Osculum is de vriendschappelijke Kus, te geven op het gezicht. Dan is er basicum [sic], de Kus van genegenheid die op de lippen belandt. En in suavium vinden we de Kus die Geliefden elkaar geven.

Notice the classificatory confusion between morphological and semantic criteria: are the three kisses defined by the physical action, or by the social relationship between the participants? I began to wonder just how much basis in classical texts this distinction has, and to find out, I turned to that great online collection of classical texts, the Perseus Project. I simply did a search for 'kiss' among the English translations, and rooted out the Latin originals for comparison. Now, this is not a rigorous method, and I occasionally found Dryden switching an embrace for a kiss, or even inserting a kiss from nowhere—but I think the results offer a rough idea of the classical kiss.

Data

By far the most hits for 'kiss' came from. . . guess who? Nope, not Ovid, but Plautus. In his Pseudolus, the protagonist promises his interlocutor 'a charming damsel, who shall give you kiss upon kiss': savia super savia (savium being an older form of suavium). In Truculentus, the same word is used, when the courtesan Phronesium complains to her returning lover that he will not give her a savium. In Amphitryon, however, Alcmene tells her husband that only the day before, she manum prehendi et osculum tetuli tibi: 'took his hand and gave him a kiss.' (Actually, it wasn't him, it was Jupiter in disguise.)

Plautus is the oldest of all the sources used here: next comes Cicero, Catullus, Lucretius, and then Vergil, Horace, Ovid, then Lucan and Phaedrus, then Tacitus, then Suetonius, and finally Jerome. From Lucretius to Jerome we find a universal reliance on osculum. In Book IV of De Rerum Natura—possibly my favourite poem of all time—Lucretius says that those in thrall to Venus' power 'clash with kisses', osculaque adfigunt. Vergil uses the word oscula when Jove kisses Venus, and later when various matrons superstitiously kiss thresholds to avert a storm. Horace in the Odes: fertur pudicae coniugis osculum, 'His wife's pure kiss he waved aside'. Ovid's Heroides: a wife while buckling her husband's armour multa tamen capies oscula, 'will however snatch many kisses'. In the Metamorphoses, Narcissus laments that he cannot reach his reflection: quotiens. . . porreximus oscula, 'how often I attempt a kiss'. The same word is always used in the Ars Amatoria.

Next the historians. Lucan in the Pharsalia describes a dying son seeking a kiss from his father: petit oscula. Tacitus mentions that a friendly meeting between a Roman captain and a Parthian prince 'ended with a kiss', osculo finitum. According to Suetonius, Domitian was so haughty that when Caenis, his father's concubine, 'offered him a kiss' (osculum), he merely extended his hand. St. Jerome renders Judas' traitor kiss (philema in the Greek) as osculum.

So what about Cicero, Phaedrus and Catullus? Well, Catullus, the erotic poet, universally uses basium. For instance, these examples from the Carmina (please excuse the foppish translations offered on Perseus):

me sinat usque basiare,
usque ad milia basiem trecenta

If any suffer me sans stint to buss,
I'd kiss of kisses hundred thousands three

nunquam iam posthac basia subripiam

After this never again kiss will I venture to snatch.

Catullus' Renaissance imitator, Johannes Secundus, similarly uses basium, and I translated one of his poems of that name here. The satirist Phaedrus describes a flute-player who facetiously blows kisses to a crowd, iactat basia. Which leaves Cicero. Like his successors, Cicero often uses osculum, and osculari, the verb: but he concludes one letter to his best friend Atticus with the injunction 'As Attica [Atticus' daughter] is inclined to be merry—the best sign in children—give her a kiss for me'. The word is suavium.

Conclusions

The most notable fact here is the overwhelming dominance of osculum, used for every kind of kiss, both in prose and poetry: superstitious (Vergil), mock-friendly (Jerome), chastely tender (Lucan), romantic (Ovid), erotic (Lucretius), official (Tacitus). The other two words are sufficiently unusual as to serve as a literary 'signature' in the earlier poets: savium in Plautus, and basium in Catullus, both of whom use the words in a sexual or romantic context. (According to the Wheelock's Latin website, Catullus 'introduced [basium] into the language from the north'). The real spanner, as far as I see it, is Cicero's use of suavium: how can he possibly have been asking his best friend, only four years older than himself, to give his daughter a 'lover's kiss'?

The distinction between these words cannot be primarily one of meaning (whether social or anatomical), but must rather be one of register. Perhaps the core sense of osculum—the kiss between friends—was replaced by a more euphemistic meaning, covering all varieties, particularly in poetry. (Here, too, the constraints of meter have a rôle.) I might have thought that basium had an anacreontic or 'low poetic' flavour, but for Phaedrus, whose facetious basia are not at all poetic. Similarly with suavium: Cicero's use of the word seems to dash the notion that it was essentially erotic in connotation, at least by the 1st century BC. The tone of Cicero's letter is familiar, but not bawdy.

*

Further, I was curious to know what became of the words after Jerome, so I checked the monumental dictionary of post-classical Latin compiled by Du Cange in the 17th century, still one of our best resources and online here. Interestingly, Du Cange has no entry for basium (or bassium, which Steve suggested as an alternative); but he does list savium—the mediaevals evidently resorted to the earlier form—as an 'osculum uxoriosum', a familiar (but not sexual) kiss. Osculum is still the basic word, however, with seven columns of definition, and its core sense remains a 'signum caritatis ac mutuae benevolentiae', a sign of affection and mutual well-wishing. The examples given are predominantly non-sexual.

The most interesting thing of all, as Steve noticed, is that the word picked up by the Romance languages was basium: bise or bisou in French, beso in Spanish, bacio in Italian. French, with its tendency to acquire diminutives (soliculum > soleil, rotula > roule, auriculum > oreille), might have produced something like ôcle. But instead, the one word not found in mediaeval Latin (according to Du Cange) was sufficiently present in the spoken argot to predominate in the Romance languages. I suppose that the growing rift in the spoken and written languages of Carolingian Europe, as noted by the 813 Council of Tours, may have unconsciously finalised the distinction between osculum and savium (written Latin) and basium (spoken vernacular).

Additional note: see the comments to Steve's link here, wherein is pointed out that the distinction is formally made in the Vergil commentary of Servius (c. 400 AD). Such a fact does not, however, invalidate the results of empirical research—yet another divergence between what was said about words, and what was done with them.