In Poem 84 Catullus has a go at a man named Arrius.

Arrius’ fault is to aspirate, add an initial aitch to, unaspirated Latin words, turning insidiae into hinsidiae and commoda into chommoda, the ch not as in church but (W. Sidney Allen’s example, Vox Latina p. 26) quite like the initial sound of cot. Catullus’ target may be Q. Arrius, a lower-class, self-made orator snootily dismissed by Cicero in his history of oratory at Brutus 242-3. There are other open questions about the poem (what the point of calling Arrius’ uncle “free” is, for example, and whether the joke in the last line is just that Arrius and his aitches are on their way home), but Catullus’ objection to Arrius’ hypercorrection no doubt carries an edge of disdain from the upper-class (and self-consciously sophisticated) poet toward a social inferior, someone socially as well as phonetically aspirational. Romans were terrible snobs, needless to say.

I’m interested in something more specific, though. Line 8, audibant eadem haec leniter et leuiter, describes the halcyon aural conditions that obtain when Arrius and his aitches have left the country; literally, “they (either everyone, or everyone’s ears) heard these same words smoothly and softly” in Arrius’ absence, leuiter hinting at the spiritus leuis or “soft breathing”, the symbol that indicates a lack of aspiration over an initial vowel in Greek (Quinn, Catullus, the poems, ad loc.). But what there also is in line 8 is an example of the meeting of two words in a poetic line generally known as elision, but more accurately as synaloepha (“melding”): the four syllables eadem haec become three, because -em and hae- coalesce into one.

Synaloepha is a common feature of Latin poetry, and it happens when a vowel or diphthong at the end of one word meets a vowel or diphthong at the start of the next. In this poem, for example, we see it in line 2, dicere, et (pronounced diceret), and 3, se esse (sesse). In most cases it was not strictly a matter of “elision”, a syllable being entirely lost, but some kind of combination of the two vowels or diphthongs into one–hence the preference for “synaloepha” as the term to describe it. Eadem haec, one word ending with -m, the next starting with h-, doesn’t at first sight look like it should be subject to the same process. But a vowel followed by m in Catullus’ Latin was pronounced as a nasalized long vowel, while the aspiration of h was so weakly realised, in elite Latin at least, that it was as if the word simply opened with the diphthong ae.

That said, how exactly the blending of eadem and haec would have sounded is hard to reconstruct: a nasalized contraction of long e and ae, maybe (W. Sidney Allen, Vox Latina p. 81), or (J. Soubiran, L’élision dans la poésie latine p. 132, supported by Quintilian, Inst. 9.4.40) something like e^wae, the -em expressed by “rounding the lips as if to end with a -w” (A. Gratwick, Plautus, Menaechmi p. 251).

If that seems awkward, there are indications in the poetic use of this kind of synaloepha of -m, as Soubiran remarks, that it could be heard as an unpleasant sound: the textbook example (for ancients and moderns) is Virgil’s description of Polyphemus at Aeneid 3.658, monstrum horrendum, informe, ingens (monstru^worrendu^winformingens), a splurge of sound that destroys the elegant fabric of the heroic hexameter and is clearly designed to convey the ugliness of the Cyclops. That doesn’t exhaust the expressive power of synaloepha of -m, though: a line in many ways similar to the Cyclops line, Aen. 9.170, describing Latinus’ palace, tectum augustum, ingens, centum sublime columnis (tectu^waugustu^wingens), seems to suggest a building soaring beyond one’s ability to discern its structure, dissolving the structure of the hexameter for different effect, or so I once proposed (Musa pedestris p. 331). More persuasive is E.J. Kenney’s remark, in a review of Soubiran (CR 17 (1967), 325-8), that synaloepha can be “used, especially by Virgil, to produce an almost unlimited range of effects.” Elsewhere Catullus himself, at 17.26, ferream ut soleam tenaci in voragine mula, “as a mule [leaves] her iron shoe in the clinging mire”, “smears together” two normally separate elements of a composite verse to evoke, in tenacin, a horseshoe left stuck in the mud.

Returning to Catullus 84, what cannot assert itself in this encounter between eadem and haec is any aspiration of the h. The synaloepha in line 8 is tackled in a very acute reading of this poem by E. Vandiver (“Sound patterns in Catullus 84”, The Classical Journal 85 (1990), 337-40). Her suggestion is that it would have been pronounced eadhaec (a true elision of -em, in effect), and thus might evoke the aspirated consonants like ch that Arrius was in the habit of inflecting on everyone. In fact, though, as I’ve explained, that kind of elision of vowel + m isn’t generally considered the most likely account. But even if it were, an h with sufficient force to persist in this way wouldn’t admit synaloepha (by compromising the preceding vowel sound) at all.

So my suspicion is that something like the opposite is true. My polymathic colleague Jonathan Katz points out to me that unless Catullus is making a point about haec, there’s really no need to include this word at all. What is his point in introducing an h only to suppress it, then? Surely, rather than echoing Arrius’ crimes against good Latin, the line that describes a life (temporarily) free of Arrius’ aspirations is serving up an aitch pronounced as it should be pronounced, i.e. not pronounced at all.

For as long as Arrius is away, in other words, even when there’s an h on the page, no one ‘as to ‘ear it.