Latin sēnī crīnēs and the Hair Style of Roman Brides

L. LaFollette & R. Wallace: S ënl CrInSs43 Latin sënl crines and the Hair Style of Roman Brides Laetitia LaFollette & Rex Wallace The reconstruction of the hair style of Roman brides is based on the interpretation ofFestus' phrase sënlcrines1 (454. 23L) and artistic representations of Vestal Virgins, sen! crines is generally understood to refer to a hair style in which the hair is divided into six tresses.2 According to this view sënl is an adjective derived from the cardinal number sex (sënl < *seks-noi).3 Recently Giannecchini,4 in a note appended to an article on the hair style and dress of Roman women by L.Sensi,5 has challenged this interpretation ofsen!crines. Giannecchini investigates 1 The relevant passage from Festus 454.23L with accompanying translation is given below: senis crinibus nubentes omantur, quod his omatus vetussimus fait, quidem quod eo Vestales virgules omentur, quarum castilatem viris suis sponeoeant* nuptae a ceteris. "Brides are adorned with the sen!crines because it is the most ancient style for them; this is also whythe Vestals are so adorned, since this guarantees the chastity ofbrides totheir husbands." 2 For the interpretation of sënl crines as a 'six-tressed' hair style see O. Brendel, "The Great Frieze in the Villa of the Mysteries," in The Visible Idea, translation by M. Brendel (Washington, D.C. 1980) 119 note 60; A. Emout & A. Meillet, Dictionnaire étymologique de la langue latine (Paris 1959) 151; M. Torelli, Lavinio e Roma (Rome 1984) 33-39; and LB. Warren, "Roman Costumes," ANRW 1.4 (1973) 596. 3 For the etymology oí sënl see Emout-Meillet, Dictionnaire étymologique, 151, 621; and A. Walde and J.B. Hofmann, Lateinisches etymologisches Wörterbuch (Heidelberg 1965) 529. For the non-distributive use ofthis form see Horace Satirae 1.10.59, Ovid Metamorphoses 12.429, etc., and the dictionary entry in the OxfordLatin Dictionary (Oxford 1968) 1734. 4 "sënlcrines," Annalidella Faculta di Lettere e Filosofía Perugia-Sezione Studi Classici N.S. 4 (1980-81)91-2. 5 "Omatus e status sociale délie donne romane," Annali della Faculta di Lettere e Filosofía Perugia-Sezione Studi Classici N.S. 4 (1980-81) 55-102. 44SyIIecta CIassica 4 (1993) two etymologies that provide, in his opinion, more appropriate meanings for sënl than that offered by the interpretation 'six'. The acceptance of either of these proposals necessitates a major rethinking of the nature of Roman bridal customs. This article therefore examines Giannecchmi's proposals in detail and ultimately finds both unacceptable, the first on linguistic grounds, the second on the basis of literary testimonia on bridal hairdos. I Giannecchini's etymologicalproposals Giannecchini proposes two etymologies for sënl that fit the semantic requirements ofFestus' passage and do not contravene the rules of Latin diachronic phonology. According to Giannecchini's first proposal, sënl is a verbal adjective in *-no- derived from the full-grade of a Proto-Indo-European (PIE) root "bind' *seh,- (cf. dignus 'worthy' < *dek-no- [PIE *dek- 'take, accept']; lignum 'gathered wood' < *leg-no- [PIE *leg- 'collect'].6 Following this etymology, the phrase sënl crines refers to a style in which the hair is 'bound'. Giannecchini's second etymology is one in which sënl derives from the PIE root *sek- 'cut'. Once again the formation is that of a verbal adjective, but in contrast to the proposal described above, the suffix here is *-sno-, an isofunctional variant of the *-no- suffix attested by formations such as Latin luna 'moon' < *louk-sneh,.7 According to this proposal, the phrase sënl crines refers to 'cut hair'. ? Evidence against sënlcrines as oound hair' Giannechini's proposed reconstruction *seh,-no- 'bound' is well-formed morphologically and is acceptable on phonological grounds because PIE *h, is lost in preconsonantal position in Latin with compensatory lengthening of the preceding vowel, cf. semen 'seed' < *seh,-mn (*seh,- 'sow'). Nevertheless, a flaw remains. Giannecchini uncritically accepts the reconstruction of this root as *seh,-. Comparative evidence shows that it mustbe otherwise. The root at issue here is attested in several branches of the Indo-European family, among them Baltic (Lithuanian siëti 'bind' inf, Lettish siet "bind'), IndoIranian (Avestan hayeiti 'binds' 3 sg près, Sanskrit s(i)yáti 'binds' 3 sg près, osât "bound' 3 sg aor) and Anatolian (Hittite iShiianzi "bind' 3 pi près, Cuneiform Luvian hiShiianti lrind' 3 pi près). The Sanskrit aorist asat points to a PIE root ending in a laryngeal. This is confirmed by the Hittite and Cuneiform Luvian forms, both of which show a present stem lhi- with zero grade of the root SA-.8 Hittite h and 6 Giannecchini's reconstruction of the root as *sê-, following J. Pokorny, (Indogermanisches etymologisches Wörterbuch I [Bern-München 1959] 891-2), is out-of-date. In this paper *s8- is rewritten as *seh,- according to contemporary laryngeal theory. We note here that there is some disagreement among Indo-Europeanists about the number of laryngeal consonants in PIE and also about theirphonetic qualities. We accept and think it is fair to note that most Indo-Europeanists also accept the system ofthree laryngeals described by M. Mayrhofer in W. Cowgill and M. Mayrhofer, Indogermanische Grammatik, band I (Heidelberg 1986) 121-50. ' For brief discussion of Latin words formed by means of the *-sno- suffix see M. Leumann, Lateinische Laut- undFormenlehre (München 1977) 320. 8 For the formation of these verbs see H.C. Melchert, Studies in Hittite Historical Phonology (Göttingen 1984) 98-9. L. LaFoUette & R. Wallace: Sënl Crines45 Cuneiform Luvian A clearly show that the root-final laryngeal must be ?G? *h2.9 *h, can be ruled out immediately because this sound is lost in Hittite and Luvian in all contexts, with the exception perhaps of word-initial position before stops where the vowel a may be a reflex.10 ?G? *h, also appears to be lost in medial position in Hittite and Cuneiform Luvian, regardless of the surrounding segments.11 As a result, even though there are no examples of *h, after s (aside from the putative case of the *bind' root), it is unlikely that *h, is preserved after s in this position to the exclusion of all other medial environments. We can, then, with reasonable security also rule out *h, as the ?G? proto-segment for A in these Anatolian verb forms.12 The comparative evidence then guarantees that the *bind' root is to be reconstructed (in the full-grade) as *seh2-, not as *sehf or *sehj-.13 If the reconstruction of the 'bind' root is *seh2-, as the comparative evidence indicates it must be, then it is impossible from a phonological standpoint to derive Latin sënl from a formation *seh2-no-. In Latin the ?G? sequence *eh2C develops to âC, as is clearly proven by forms such as iänus '[god of] passageway(s)' < *yeh2- 'go' and mâtûrus 'seasonable' < *meh2- 'to be timely'.14 Still the reconstruction of the root 'bind' with *h2 does not automatically rule out a connection withsënl. Anotherpossibility exists.15 Since ?G? *h» does not alter the quality of a preceding long vowel, there is no phonological obstacle standing in the way of deriving sënl from the lengthened grade of the root *seh2-.16 A prehistoric form *sêh»-no- will in fact produce the attested Latin form sënl. However, a reconstruction with lengthened-grade of the root *sêlv must be ruled out of order, despite the fact that the phonology permits the derivation of this form. Verbal adjectives with the suffix *-no- are made to roots in the full-grade and the zero-grade but there is no comparative evidence, and certainly no evidence in Latin, to suggest that these formations were built on roots with lengthened-grade vocalism. Without morphological support, this proposal too is unacceptable. In sum, then, Giannecchini's first etymology must be rejected on linguistic grounds. The root *bind' cannot be reconstructed as *seh,-. Comparative evidence 9 The reflexes of *h, in Anatolian are discussed by N. Oettinger, Stammbildung des hethitischen Verbums (Nürnberg 1979) 546-50. 10See S. Kimball, "Initial *h,s- in Hittite," Studies in Memory ofWarren Cowgill (1929-1985) (New York/Berlin) 160-81. 1 1 The farts on *h, in Anatolian are discussed in S. Kimball, "*H3 in Anatolian," Festschift for Henry Hoenigswald (Tübingen 1987) 185-92; and H.C. Melchert, "Reflexes of *h3 in Anatolian," DU Sprache 33.1-2 (1987) 19-28. 12 We thank Craig Melchert (perlateras) fordiscussing this with us. 13For the reconstruction of this root as *seh,- compare Oettinger (above, note 9) 461, 466; J. Tischler, Hethitisches etymologisches Glossar (Innsbruck 1983) 384-88; and J. Puhvel, Hittite EtymologicalDictionary (New York 1984) 398-403. 14For discussion of the development of PIE laryngeals in Latin see M. Mayrhofer, "Die Vertretung der indogermanischen Laryngale im Lateinischen," KZ 100.1 (1987) 86-108. 15 We thank Brent Vine (p.c.) forbringing this possibility to our attention. 16See H. Eichner, "Die Etymologie von heth. mehur," MSS 31 (1973) 53-108; and Mayrhofer (above, note 6) 132-3. 46Syllecta Classica 4 (1993) points to a root *seh2- and it is phonologically impossible to derive the root element in Latin sënl from such a pre-form. sënlcrines then cannot refer to hound hair'. IH Evidence against sënlcrines as 'cut hair' Unlike Giannecchini's first proposal, derivation of sënl from the root *sek- 'cut' cannot be rejected on linguistic grounds.17 The proposed pre-form *sek-snoi yields Latin sënl by regular sound changes (cf. lüna above for the phonology of the cluster *-ksn-). Accepting Giannecchini's second hypothesis, sënl crines as 'cut hair", has radical implications for our understanding of Roman bridal practices. As Sensi's study makes clear, in Roman times a young girl's long hair was her ornätus, part of her beauty; yet he accepts the association of the sënlcrines hair style with the ritual cutting off of the bride's hair, a deliberate, if temporary mutilation intended to instill a sense ofshame in the bride-to-be.18 This interpretation deserves close scrutiny, since such a punitive act seems hardly in keeping with the tone of other Roman bridal customs, such as the weaving of the bridal tunic by the bride, which showed off her skill at what would become one of her major domestic responsibilities, or with the fertility and constancy symbolized by the color and associations of her wedding veil, theflammeum.19 The supporting evidence for the interpretation of sënl as 'cut' comes from the Greek practice on the one hand and on the other from the Vestals.20 Let us examine each in turn. The analogy with Greek practice is misleading, since the best attested examples involved only the cutting and dedication of a lock of hair of the bride-to17Despite the fact that this etymology is possible, there are considerations that make us uneasy about its plausibility. For example, we wonder about the relationship between the formation ofsent and signum 'mark', a nominal form which is connected etymologically by some (and cited approvingly by Giannecchini) with the root *sek- 'cut'. UsSnT is to be derived from *sek-sno-, and *sek- is also accepted as the etymological source for signum (< *sek-no-), then we must ask for some discussion ofjust how two semantically isofunctional suffixes came to be added to the same root, a phenomenon which is not otherwise attested for these suffixes in Latin. We realize, of course, that it is possible to avoid this problem altogether by deriving signum from the root *sek"- 'follow' (e.g. Walde-Hofmann, Wörterbuch 535 and Ernout-MeiUet, Dictionnaire étymologique 625, and noted in a footnote by Giannecchini). Further, the existence of (dialectal) Latin asignae 'flesh cut into parts' (CGL ? 24,6) is perhaps potentially damaging to the view that sënl is to be derived from a pre-form *sek-sno-. Although the source of the prefix a- is not clear (though possibly the same as that found in anhelo pant' with loss of the nasal in the environment ns, cf. côsul < *konsul), the morphology of the rest of the word is straightforward, namely a full-grade root *sek- followed by a suffix -no-. If this is a legitimate Latin word and not a borrowing from Oscan-Umbrian (cf. Marrucinian asignas, see E. Vetter, Handbuch der italischen Dialekte [Heidelberg 1953] note 218), then it suggests that the *-no- adjective made from the root *sek- in Latin was not built with an extended *-sno- suffix. 18 Sensi (above, note 5) 55-6 (importance of woman's head and hair), 68-9 (sënlcrines and ritual haircut). 19Weaving ofthe bridal tunic: Festus 364.21L, cf. Pliny Naturalis Historia 8.124;flammeum of bride associated v/ühflaminica Dialis and her fidelity to one man: Festus 82.6L and 79.23L For her costume as a sign of the newly empowered status of the bride, see L. LaFollette, "The Costume of the Roman Bride," Roman Costume as a Historical Source, eds. L. Bonfante and J. Sebesta (Madison, forthcoming). 20These are the two sources adduced by Sensi (above, note 5), the only authority to our knowledge to argue for such an interpretation. L. LaFoUette & R. Wallace: Sënl Crines47 be.21 Nor was this done only for the bride. The practice is attested for the groom and also for the deceased at funerals. The custom of such dedications to virgin deities, male or female, emphasized the loss of virginity in marriage, and the changing of status, and thus was not designed to instill shame in the bride. Even at Sparta, where the young girl's hair was indeed shorn, the purpose, like that of other Spartan customs such as her exercise and training, was to root out effeminacy and to toughen the young woman like her male counterpart.22 No such attitudes existed at Rome. The evidence for the short hair style of the Vestals is much stronger, however, and clearly important since Festus states that the sënl crines was an attribute of Vestals as well as brides. The Vestal's hair was definitely cut upon the occasion ofher induction into the College, and Pliny reports that the shorn locks were hung on a special tree.23 Whether or not the hair of the priestess remained short during her tenure has been debated, since many representations of Vestals do not allow careful examination of the hair style beneath their elaborate headdress of ceremonial veil and turban-like fillet. However, a recent re-examination of the portraits and several reliefs has demonstrated that the Vestals did indeed wear their hair short.24 Yet this does not prove a similar ritual haircut for the bride. The short hair style of the Vestals lends even greater support to Jordan's generally accepted suggestion that the priestess' fillet served as a cloth substitute for the sënlcrines of the bride.25 The need for such a substitute was obviously due to the short hair of the Vestal, which Jordan himself failed to recognize. Thus Festus' association of the sënl crines the bridal hair style, with the Vestals is still valid, with the minor modification that the Vestals imitated the bridal hairdo with a cloth fillet rather than their own hair. There is in fact yet another even more compelling reason to reject the notion of a bridal haircut, namely the literary evidence on the hair style-of the bride. Plautus (Most. 226) expliciüy refers to the bride putting up her hain The thing to do is to humor him, and fix up your hair for the wedding.' In the Miles, Unes 790-93, Plautus describes the hairdo at greater length: 'You are to take her home to your house at once, then bring her back here, all got up like a married lady, the usual headdress, hair done high in ribbons, and she's to pretend she's your wife.' There is no mention here of the ritual haircut argued by Sensi. The passages refer not to cut hair, but to (long) hair piled up on top of the head. These literary sources suggest that the interpretation ofsënl as 'cut' should be rejected. 21Troezen: Euripides Hippolytus 1424-26; Delos: Herodotus 4.33-35. For discussion, see J. Redfield, "Notes on the Greek Wedding," Arethusa 15 (1982) 190-91. 22Plutarch Lycurgus 15. 23Pliny Naturalis Historia 16.235, cf. Festus 50.12L. 24See LaFollette (above, note 19). The crucial evidence is the representation of the Vestal on frieze B ofthe Cancellería reliefs, F. Magi, / rilieviflavi del Palazzo della Cancellería (Rome 1945) 28 (figure 31) and 92 (plate 19 with discussion). 25H. Jordan, Der Tempelder Vesta unddasHaus der Vestalinnen (Berlin 1886) 47-8, accepted by, e.g., H. Jucker, "Bildnis einer Vestalin," RhM 68 (1961) 96; and S. Walker, "A Portrait Head of a Life-sized Statue of a Vestal," British Museum Occasional Paper 22, Dept of Greek and Roman Antiquities, New Acquisitions 1976-79 (British Museum, London 1981) 17. 48Syllecta Classica 4 (1993) Instead, the literary testimonia support the numerical interpretation of sënl, rejected by Sensi and Giannecchini. Particularly relevant is a previously unnoticed passage in Martial (12.32.2-4): I saw them Vacerra, I saw those bags they would not hold as payment for two years' worth of rent: your red-headed wife, the one with the seven-tressed hair (septem crûtes ), walked off with them. The expression septem crines has been taken to mean bald by some commentators, but this does not convince, since the wife is also described as a red-head.26 It seems far more plausible to suggest that the phrase septem crines, one of the two instances inLatin literature where the word crines is accompanied by anumerical adjective, is a play on the sen!crines mentioned by Festus. Thus Martial would be sketching the portrait of an unfaithful spouse by means of uncanonical bridal attributes.27 The Martial passage thus clearly indicates that by this period sënlwas understood numerically. IV Conclusion Our examination has shown that the two interpretations for sënl proposed by Giannecchini must be rejected, that as 'bound' on linguistic grounds, that as 'cut' on the basis ofother literary testimoniafor the bridal hairdo. Although our conclusion is hardly radical, since we return to the interpretation of sënl as 'six', we have adduced a previously unnoticed passage in Martial to support this meaning, and have also clarified the meaning of Festus' phrase with which we began. Even more importantly, we have demonstrated that the notion of a ritual shearing of the hair of the Roman bride for her wedding has no basis. Department of Art History Department of Classics University of Massachusetts 26M.V. Martialis, Epigrammata ad codices parisinos accurata recensita, vol. 3 (Paris 1825) 30 note 4: "septem crines: id est uxor tua est calva." 27Compare Juvenalis Satirae 6.224-26, where the poet puns on the worn-out bridal veil (flammeum) ofthe unfaithful wife. ...