Roussou ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 Article copyright held by the author(s) and made available under the Creative Commons Attribution License CC-BY https://creativecommons.org/licenses/by/4.0/ Λέξεις ῥωµαῖαι κατὰ στοιχεῖον in Barocci 50 and Burney 124: On Greek Lexicographical Sources and the Suda Stephanie Roussou HE SHORT TREATISE Λέξεις ῥωµαῖαι κατὰ στοιχεῖον (henceforth abbreviated Λέξεις) consists of a small number of entries in alphabetical order, each of which contains a lemma and its gloss. Most of the lemmata are Latin loanwords used in Greek, but the treatise also includes a couple of words which are not attested anywhere else. The Λέξεις is transmitted in two manuscripts: the tenth-century Barocci 50 (fol. 109v–110v),1 kept in the Bodleian library in Oxford, and a seven- teenth-century copy of Barocci 50, namely London Burney 124 (fol. 122r–v),2 kept in the British Library. Barocci 50 is a parchment codex written in minuscule script, originally consisting of 387 1 For the date see H. O. Coxe, Bodleian Library Quarto Catalogues I (Oxford 1969) 70–78. The MS. is available online at https://digital.bodleian.ox.ac.uk/ objects/901e221a-988f-4202-8c4d-941509e2c868/surfaces/9f01102a- 802d-46f0-8e0e-a3119c058902. A. Adler, Suidae lexicon I (Leipzig 1928) p. V, mentions in her preface that in editing the text she found useful the material in the Λέξεις. For a codicological analysis of the MS. see F. Ronconi, “La miscellanea che non divenne mai silloge: il caso del Bodl. Barocci 50,” in R. M. Piccione and M. Perkams (eds.), Selecta Colligere II (Alessandria 2005) 295– 353. For discussion on the origin of this manuscript see e.g. N. G. Wilson, “On the Transmission of the Greek Lexica,” GRBS 23 (1982) 369–375, with further bibliography. Adler lists the Λέξεις among the grammatical sources she employed in her edition (I p. XVIII). She refers to this treatise as Explica- tiones vocum Latinarum, quarum pars in codice Barocciano 50 exstat. In her critical apparatus she reports the relevant readings and variant readings of Barocci 50. 2 For the date see The British Library Summary Catalogue of Greek Manuscripts I (London 1999) 71–73. T 342 ΛΕΞΕΙΣ ΡΩΜΑΙΑΙ ΚΑΤΑ ΣΤΟΙΧΕΙΟΝ ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 folios of which eight are now missing.3 It contains a large miscel- lany of texts, including grammatical and poetic texts such as Theognostus’ Canones, Choeroboscus’ Περὶ ὀρθογραφίας, Museus’ Hero and Leander, and the Batrachomyomachia.4 Burney 124 is made of paper and consists of 136 folios.5 It contains extracts from mis- cellaneous texts taken from various manuscripts kept in the Bod- leian Library, and could be the hand of some Oxford scholar. Some of the entries in the Λέξεις appear also in other lexico- graphical sources: mostly in the Suda and in a couple of cases also in Hesychius, Photius, Pseudo-Zonaras, and the Συναγωγή λέξεων χρησίµων. More specifically, my discussion below includes fifteen entries from the Λέξεις, thirteen of which are also in the Suda, two in Hesychius, two in Photius, one in Pseudo-Zonaras, and two in the Συναγωγή λέξεων χρησίµων. Inevitably, out of all these other lexicographical sources, my focus will be on the Suda. Some of the entries transmitted in the Λέξεις appear in the very same way in the Suda, while others differ in their lemma or their gloss. The differences between the Λέξεις and the Suda, and the two entries which are solely attested in the Λέξεις, have either gone unnoticed or understudied. This article highlights and offers an appraisal of this underexamined material. I aim to provide a better understanding of certain words used at least in written Greek and to elucidate the transmission history of the texts examined and the sources from which they derive their material. 1. Errors suggesting that the corruption in Barocci 50 and/or the Suda must have occurred already in their source(s), before the lemmata were alphabetized in the Λέξεις and the Suda In this section I study three entries whose glosses are the same (or nearly the same) in the Suda and Barocci 50 but whose lem- 3 K. Alpers, Theognostos Περὶ ὀρθογραφίας: Überlieferung, Quellen und Text der Kanones 1–84 (Hamburg 1964) 4, corrects Coxe, Bodleian Library Quarto Cata- logues 70, who writes that seven folia are missing. 4 See Coxe, Bodleian Library Quarto Catalogues 70–78. 5 See The British Library Summary Catalogue of Greek manuscripts 71–73. STEPHANIE ROUSSOU 343 ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 mata seem problematic. The first is transmitted under the lemma ἀπουρία (Suda α 3624; Barocci 50 fol. 110v.2), which is most probably wrong; the second under δηναδίπτρα in the Suda (δ 479) and under δωνάτιβα in Barocci 50 (fol. 109v.28); the third under προνκοµισ(σ)άριος (Suda π 2520; Barocci 50 fol. 110r.23), which might be wrong. If we correct the lemma ἀπουρία to σπουρία, δηναδίπτρα to δωνάτιβα, and προνκοµισ(σ)άριος to κοµ- προµισσάριος (as suggested below), then the corrected lemmata will be out of alphabetical order. This suggests that these errors must have occurred at an early stage in the textual transmission, before they were alphabetized. The sources for our texts must have already contained the wrong lemmata. 1.1. The entry on ἀπουρία (σπουρία?) Suda α 3624 reads:6 Ἀπουρία: ἡ ἐξ ἀνίσων γάµων σπορά, ἀπὸ δούλης ἢ δούλου γεν- νωµένη. Ἀπουρία: the offspring of unequal marriage, begotten from a female slave or male slave.7 We find the same entry in Barocci 50 (fol. 110v.2), with a small number of scribal errors. The diplomatic transcription of this passage is: απουρία ἡ ἐξ ἀνίσων γάµων σπορά ἠ ἀπο δούλης ἠ ἀπο δούλου γεννώµενον (γενώµενον in the text, with a superscript ν for γεννώµενον) The scribal error ἠ after the word σπορά leaves unclear whether the intended reading was the disjunctive ἢ or the article ἡ. If ἡ, one should change the transmitted γεννώµενον to γεννωµένη: ἀπουρία: ἡ ἐξ ἀνίσων γάµων σπορά, ἡ ἀπὸ δούλης ἢ ἀπὸ δούλου γεννωµένη. ἀπουρία: the offspring of unequal marriage, the one begotten from a female slave or male slave. 6 Other than that οἷον appears before ἀπὸ δούλης, the entry is the same in Pseudo-Zonaras α 245.16–17. 7 All translations are mine except where otherwise indicated. 344 ΛΕΞΕΙΣ ΡΩΜΑΙΑΙ ΚΑΤΑ ΣΤΟΙΧΕΙΟΝ ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 However, it is hard to see how γεννωµένη could have given rise to γεννώµενον. On the other hand, if the disjunctive ἤ with a variant gloss was intended, one should add before ἀπὸ δούλης the article τό in agreement with γεννώµενον. This seems palaeo- graphically preferable and I therefore propose: ἀπουρία: ἡ ἐξ ἀνίσων γάµων σπορά, ἢ ⟨τὸ⟩ ἀπὸ δούλης ἢ ἀπὸ δούλου γεννώµενον. ἀπουρία: the offspring of unequal marriage or what is begotten from a female slave or male slave. ἀπουρία is attested only in the Suda (α 3624), in Pseudo- Zonaras (α 245.16-17), and in Barocci 50 and Burney 124. Whitehead in the Suda On Line notes: “Its etymology is left unelucidated, but the obvious link would seem to be with urine.”8 I argue instead that the intended lemma must have been σπουρία. Burney 124 includes ἀπουρία in the main text but pro- vides σπουρία in the margin. This is probably a conjecture be- cause it is preceded by the abbreviation ἰσ. for ἴσως (“perhaps”).9 In both Barocci 50 and Burney 124 the lemma ἀπουρία occurs where one would expect words beginning with σ- (in general these manuscripts follow an alphabetical order, at least for the first letter of the word). Thus, ἀπουρία either starts with the wrong letter (a scribal error) or has been wrongly alphabetized. Two reasons convince me that σπουρία in Burney 124 is the correct reading: (i) the word σπουρία fits the gloss attested in the Suda (and Pseudo-Zonaras) and in Barocci 50 and Burney 124, for σπούριος “bastard/false” is borrowed from the Latin spurius “son of an unknown father”;10 (ii) the corruption can be explained on palaeographical grounds with recourse to Latin, if one accepts that it occurred when the lemmata were still written in Latin. 8 http://www.cs.uky.edu/~raphael/sol/sol-entries/alpha/3624. 9 Burney 124 offers two other good readings for two lemmata reviewed below. I discuss the status of the good readings in the margin of this MS. in section 7 below. 10 See E. Dickey, Latin Loanwords in Ancient Greek: A Lexicon and Analysis (Cam- bridge, forthcoming) s.v. σπούριος. STEPHANIE ROUSSOU 345 ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 The change of s to a likely happened in the Latin script, e.g. in the Roman pen-written cursive of the second century A.D., due to similarity.11 Figure 1 shows first the letter a and then s in the Roman cursive of this period:12 Figure 1 Alternatively, it is possible that a scribe might have had in mind the word ποῦρος (or Latin purus) and might have misread ἀπουρία, with privative ἀ- affixed to Latin purus. The folk ety- mology would reflect the idea that “the offspring of unequal marriage or what is begotten from a female slave or male slave” was an “impurity” (ἀ + ποῦρ(ο)- + -ία “impurity” from purus “pure”). A TLG search of ποῦρος shows that it is not attested before the ninth century A.D., that it is infrequent, and that most of its occurrences come from the Basilica code of law. Thus, ποῦρος appears not to have been sufficiently established in Greek for the formation of ἀπουρία from ἀ + ποῦρος to be plausible. If I am right that ἀπουρία is a manuscript corruption of σπουρία, given the alphabetical order followed by the Suda we must conclude that it had already taken place in the Suda’s source before ἀπουρία was placed among lemmata beginning with ἀπουρ-. If the change occurred in the Latin script, the source for 11 See for example the table with the Latin cursive alphabet in E. M. Thompson, An Introduction to Greek and Latin Palaeography (Oxford 1912) 336. For a discussion of Roman cursive see H. B. van Hoesen, Roman Cursive Writing (Princeton 1915). The mixing of Greek and Latin letters has been studied by Gulielmo Cavallo, “La κοινή scrittoria greco-romana nella prassi document- tale di età bizantina,” JÖB 19 (1970) 1–31. See also P. Radiciotti, “Mano- scritti digrafici grecolatini e latinogreci nell’antichità,” Papyrologica Lupiensia 6 (1997) 107–146, esp. 111, and “Manoscritti digrafici grecolatini e latinogreci nella tarda antichità,” Papyrologica Lupiensia 7 (1998) 153–185. 12 Drawings after Thompson, An Introduction 336. 346 ΛΕΞΕΙΣ ΡΩΜΑΙΑΙ ΚΑΤΑ ΣΤΟΙΧΕΙΟΝ ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 this lemma was likely a Latin one; and if this script was the Roman cursive of the second century A.D., this would furnish a terminus ante quem for the source. The corrupt lemma was tran- scribed into Greek no earlier than the second century and even- tually found its way into the Suda and the Λέξεις. Although the seventeenth-century Burney 124 is the only wit- ness to σπουρία in the Λέξεις, the adjective σπούριος is attested in later Greek texts from the thirteenth century onwards.13 1.2. The entry on δηναδίπτρα/δωνάτιβα Suda δ 479 reads: Δηναδίπτρα: κατὰ Ῥωµαίους δωρεὰ βασιλικὴ στρατιώταις. Δηναδίπτρα: among the Romans, (it is) a royal gift to the soldiers. As Adler points out in her apparatus of sources, the Λέξεις have the lemma δωνάτιβα instead. More specifically, Barocci 50 (fol. 109v.28) reads: δωνάτιβα· δωρεὰ βασιλικὴ στρατιώταις. δωνάτιβα: a royal gift to the soldiers. The gloss δωρεὰ βασιλικὴ στρατιώταις seems appropriate for δωνάτιβον (or the plural δωνάτιβα; alternative forms are δωνάτιον or δωνατίουον), immediately borrowed from Latin donativum “money given to soldiers as a gratuity from the emperor.”14 δηναδίπτρα, attested only in the Suda and Pseudo-Zonaras’ lexicon,15 must be a scribal error for δωνάτιβα (although a palae- ographic explanation for the change of δωνάτιβα to δηναδίπτρα is 13 See E. Kriaras, Λεξικό της µεσαιωνικής ελληνικής δηµώδους γραµ- µατείας 1100–1669 XX (Thessaloniki 2016) s.v. σπούριος, with references to the texts where the word is attested. 14 See Dickey, Latin Loanwords s.v. δωνατίουον (with further bibliography); Lex.Byz.Gr. s.v. δωνατίβαι. Cf. the note by Roth, Hutton, and Whitehead in the Suda On Line (https://www.cs.uky.edu/~raphael/sol/sol-entries/delta/ 479). 15 This was pointed out by Roth, Hutton, and Whitehead in the Suda On Line. A TLG search of δηναδίπτρα yielded only these two results. STEPHANIE ROUSSOU 347 ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 hardly obvious).16 In her edition of the Suda Adler seems aware of the reading δωνάτιβα and reports it in Barocci 50. But she prints δηναδίπτρα, which appears in alphabetical order among words with δην-. Thus, the corruption must have been already present in its source before the Suda alphabetized it. The gloss in the Λέξεις does not include the phrase κατὰ Ῥωµαίους: possibly a scribe or the author of the treatise considered it redundant in light of the ῥωµαῖαι in the title.17 The good reading δωνάτιβα transmitted in Barocci 50 deserves further thought. One may wonder if it is due to the scribe or to the source for this manuscript. If to the scribe, we would expect a manuscript largely free from such errors as a philologically skilled scribe could correct. But Barocci 50 contains a significant number of scribal errors, some of them straightforward; its scribe moreover has not been identified, nor do we know of any other manuscript by him. If we add to this that a correction of δηναδίπτρα to δωνάτιβα is hardly obvious, one may reasonably conclude that Barocci 50 got the good reading δωνάτιβα from its source and that it is not a scribal correction. This in turn implies that this source is different from, and predates, the one used by the Suda, which already contained the error. 1.3. The entry on προνκοµισ(σ)άριος Both the Suda (π 2520) and Barocci 50 (fol. 110r.23) include an entry on προνκοµισ(σ)άριος (with a single σ in the Suda and with two in Barocci 50). In the Suda the entry reads: προνκοµισάριος: ὁ ἐκ συναινέσεως δύο µερῶν µεταξὺ ληφθεὶς διαιτητής. προνκοµισάριος: the arbitrator taken in between [i.e. as a medi- ator to stand between them] by the consent of two parties. Barocci 50 (fol. 110r.23) reads: 16 δηναδίπτρα is not included as a lemma in Lex.Byz.Gr., nor in Dickey, Latin Loanwords. 17 See also section 4, on the absence of the phrase παρὰ Ῥωµαίοις in Barocci 50 or its source. 348 ΛΕΞΕΙΣ ΡΩΜΑΙΑΙ ΚΑΤΑ ΣΤΟΙΧΕΙΟΝ ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 προνκοµισσάριος·18 ὁ ἐκ συναινέσεως δύο µερῶν µεταξὺ ληφθεὶς διαιτητής. προνκοµισσάριος: the arbitrator taken in between [i.e. as a medi- ator to stand between them] by the consent of two parties. προνκοµισ(σ)άριος is otherwise attested only in the scholia to the Basilica, dated to the eleventh to twelfth centuries (specifically προκοµισσαρίου at 22.5.30.6). This form is thus not well attested. There are good reasons for thinking that the intended lemma was κοµπροµισσάριος (a rare word from Latin compromissarius).19 κοµπροµισσάριος is well attested and its meaning suits the gloss for προνκοµισ(σ)άριος in Suda π 2520 and Barocci 50. If so, προνκο- µισ(σ)άριος could be a scribal error: perhaps at some point -προµ- was omitted, then added above the line, and finally replaced before κοµ-, at which point the µ of προµ- was changed to the nasal ν before κο-. If this theory is right, that the Suda and Barocci 50 both present προνκοµισ(σ)άριος in alphabetical order implies that the corruption was already present in their sources. As regards the spelling of προνκοµισ(σ)άριος, leaving aside the mis- spelling πρωνκοµισσάριος in Barocci 50, it is worth noting that this same manuscript respects the two sigmas of the uncompounded κοµισσάριος.20 The Suda, on the other hand, has only one sigma. While one cannot preclude that the two sigmas are a scribal 18 I print προνκοµισσάριος instead of the transmitted πρωνκοµισσάριος, and συναινέσεως (a marginal reading of Burney 124 preceded by ἴσ[ως]) instead of the transmitted συνέσεως. For these good marginal readings of Burney 124 see section 7 below. 19 See Dickey, Latin Loanwords s.v. κοµπροµισ(σ)άριος, with further bibli- ography; cf. Lex.Byz.Gr. s.v. κοµπροµισσάριος. 20 The spelling with two sigmas must be correct, whether κοµισσάριος is a borrowing of an unattested comissarius (for which see H. Hofmann, Die lateinischen Wӧrter im Griechischen bis 600 n. Chr. [diss. Erlangen-Nürnberg 1989] 190), or it comes from the Latin verb committo (“entrust”), commissus (so White- head and Roth in the Suda On Line http://www.cs.uky.edu/~raphael/sol/sol- entries/pi/2520), or κοµισσάριος is only attested in the sixth century as a variant form for κοµπροµισσάριος in Just. Nov. 113.1.1 (for which see Dickey, Latin Loanwords s.v. κοµισσάριος). STEPHANIE ROUSSOU 349 ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 error, it is more likely that this spelling has been taken over from its source. Probably προνκοµισ(σ)άριος was understood as a borrowing of an unattested Latin procomissarius (by analogy with κοµισσάριος from an unattested comissarius). But in that case it is hard to see how the ν between προ and κοµισσάριος came about. 2. Entries that do not occur in other lexicographical sources Barocci 50 contains two entries which are not attested in other lexicographical sources: θονγραµµασίγγελον and λεγίτιµος. The former is not attested anywhere else and is problematic. Both probably go back to an earlier source, unknown to us. 2.1. The entry on θονγραµµασίγγελον θονγραµµασίγγελον, nowhere else attested apart from the two manuscripts which transmit the Λέξεις (Barocci 50 fol. 110v.4, Burney 124 fol. 122v.7), is the lemma of the following entry: Θονγραµµασίγγελον: τὸν δρόµον τῶν κτηνῶν τὸν δηµόσιον. Θονγραµµασίγγελον: the public pathway of herds. This dubious word might even be a scribal error. Taking the gloss as a guide, one wonders whether αµµασιγγ ελο could con- ceal animalium uia(m).21 The corruption could have occurred by misreading Roman pen-written cursive of the fourth and fifth centuries A.D. To be specific, animalium could have been misread as αµµασιγγ by confusing -nim- with -µµ- (the ligature of -ni- resembling the Greek -µ-, and the -m- a µ),22 l with a Greek lunate sigma, u with a Greek γ, and m also with γ; via(m) could 21 I am grateful to David Langslow for this suggestion, which he made on the basis of letter shapes in tenth-century manuscripts. The letter shapes in Thompson, An Introduction 337, show that similar confusion was already possible in the fourth and fifth centuries. In fact, the earlier date is more likely because, as I will show in section 3, further corruption probably happened in the Greek cursive script of the sixth and seventh centuries. At that time, therefore, the lemmata were no longer in the Latin script. 22 On the resemblance of -m- with Greek µ see Thompson, An Introduction 337 (for Roman cursive) and 193 (for Greek cursive), and van Hoesen, Roman Cursive Writing 235. 350 ΛΕΞΕΙΣ ΡΩΜΑΙΑΙ ΚΑΤΑ ΣΤΟΙΧΕΙΟΝ ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 have been misread as ελο by confusing u (v) with a Greek ε, i with a λ, and a with a Greek ο (the final m of via(m), perhaps indicated by a horizontal stroke above the preceding letter, may have been finally omitted). Figure 2 shows in ink first animalium uia(m) in Roman cursive of the fourth or fifth century and then αµµασιγγ ελο in the Greek cursive of a similar date:23 Figure 2 There is a striking similarity between the first letters of the transmitted lemma, θονγ, and the letters bovu in the Roman pen- written cursive of the fourth and fifth centuries: the Latin b resembles a θ, and -vu- the Greek -νγ-. Figure 3 first shows bovu, then θονγ in the respective cursive styles: Figure 3 Perhaps then the word bovu(m) (“of the cattle”?) might have been intended (if so, the final m of bovu(m) could have been indicated by a horizontal stroke above the preceding letter that was eventually left out). Although the transmitted gloss does not include a specific Greek term for bovu(m), given the reconstructed animalium and the word κτηνῶν, we can reasonably assume that “cattle” was mentioned in this context. Given the likelihood of multiple corruptions, uncovering the original reading is hard if not impossible. 23 Drawings after Thompson, An Introduction 337 and 193. STEPHANIE ROUSSOU 351 ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 2.2. The entry on λεγίτιµος A lemma in Barocci 50 (fol. 110r.14) absent in other lexico- graphical sources is λεγίτιµος, a direct loan from Latin legitimus:24 λεγίτιµος:25 ὁ νόµιµος κουράτωρ κατὰ26 νόµους διδόµενος. λεγίτιµος: the lawful curator established according to the laws. The adjective λεγίτιµος (“based on law/statutory”) derives from Latin legitimus “of the law.”27 In this entry, however, it is not an adjective but a noun glossed by κουράτωρ “curator,” itself borrowed from Latin curator.28 As a legal term curator referred specifically to the “guardian appointed to administer the property of minors (pupilli), women, and insane persons.”29 The adjective “based on law” ended up being used as a noun specifically denoting the person to whom the office of curator was assigned by law. In other words, λεγίτιµος stood for λεγίτιµος κουράτωρ (“lawful curator”). Nowhere else is λεγίτιµος glossed by κουράτωρ. In a small number of legal texts which include the Ba- silica, Michael Attaliates’ Πόνηµα νοµικὸν ἤτοι σύνοψις πραγµατική (25.10‒12), the Prochiron Auctum (36.15.1‒3), and Konstantinos Harmenopoulos’ Ἑξάβιβλος or Πρόχειρον νόµων (5.12.3.3‒5), λεγίτιµος pertains to the noun ἐπίτροπος (“guardian, protector”). κουράτωρ in our passage could be seen as a synonym for ἐπίτροπος. The passages from Michael Attaliates and the Prochiron Auctum contain the phrase ὁ παρὰ τοῦ νόµου διδόµενος (“the one given by law”), which is close to κατὰ νόµους διδόµενος (“estab- lished according to the laws”). In Konstantinos Harmenopoulos we find the similar παρὰ τοῦ νόµου δοθείς (“given by the law”), 24 See Dickey, Latin Loanwords s.v. λεγίτιµος, with further bibliography; cf. Lex.Byz.Gr. s.v. λεγίτιµος. 25 The diplomatic transcription of this entry is λεγήτιµος ὁ νόµιµος κουράτωρ κατα νόµους διδόµενος. For λεγήτιµος I print λεγίτιµος. 26 I have supplied the accent missing in the MS. 27 See Dickey, Latin Loanwords s.v. λεγίτιµος; cf. Lex.Byz.Gr. s.v. λεγίτιµος. 28 See Dickey, Latin Loanwords s.v. κουράτωρ, with further bibliography. 29 OLD s.v. curator 3. 352 ΛΕΞΕΙΣ ΡΩΜΑΙΑΙ ΚΑΤΑ ΣΤΟΙΧΕΙΟΝ ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 while in the Basilica (2.2.213.12‒13) we read παρ’ αὐτοῦ τοῦ νόµου καλούµενος καὶ προχειριζόµενος (“called and assigned by the law itself ”). Judging from the occurrence of κουράτωρ in our text rather than ἐπίτροπος, as in the above-mentioned texts, one may think that our treatise got its material for this entry from a different source. And yet the phraseological nearness of κατὰ νόµους διδόµενος to its corresponding parallels should not be discounted, even if διδόµενος and δοθείς are to be expected for “given/established” by, or according to, the law. 3. Entries with different glosses In this section I discuss four entries whose glosses in the Λέξεις are different from the ones in the Suda (and in Hesychius, Photius, and the Συναγωγή λέξεων χρησίµων where these too are available). Their lemmata are: ῥελατορίαι/ῥελατορία, κυαίστωρ/ κοιαίστωρ, κεντυρίων, κώδιξ/κώδικα. Their implication is that our treatise followed a different source. 3.1. The entry on ῥελατορίαι/ῥελατορία In Barocci 50 (fol. 110r.27) we find the lemma δαιλατογίαι, a scribal error. With the help of the transmitted gloss and the marginal note ἴσ[ως] ῥαίλατ- in Burney 124, I have restored ῥελατορίαι (“receipts brought back after delivery”), a rare word derived from Latin relatoria (“receipt”) or relator (“one who registers”) + -ία:30 30 See Dickey, Latin Loanwords s.v. ῥελατωρία, ῥελατορία, with further bibli- ography; cf. Lex.Byz.Gr. s.v. ῥελατορία. P. Chantraine, Dictionnaire étymologique de la langue grecque (Paris 1999) s.v. ῥελατωρία, derives the word from the Latin relator “one who registers”; R. Beekes, Etymological Dictionary of Greek (Leiden 2010) s.v. ῥελατωρία, also writes that the word is borrowed from relator “registrator.” The confusion between the ε of ῥελατορίαι and the αι of the transmitted ῥαίλατ- is easily explained, since from the Roman period on- wards both ε and αι were pronounced [e]. See e.g. F. T. Gignac, A Grammar of the Greek Papyri of the Roman and Byzantine Periods I (Milan 1976) 191–193, and G. Horrocks, Greek: A History of the Language and its Speakers (Chichester 2014) 144–147 and 154. For the notion that ε and αι had already fallen together in pronunciation by the mid-second century B.C. see Horrocks 167. STEPHANIE ROUSSOU 353 ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 ῥελατορίαι·31 τῆς ἀποκαταστάσεως τοῦ δηµοσίου φόρου αἱ ἀποδείξεις. ῥελατορίαι: the proofs of the restoration of public payment. The corruption of the opening r to d (i.e. from relatoriae to delatoriae) likely happened in the Latin script, e.g. in the Roman cursive of the first century A.D. Figure 4 shows first d and then r in Roman pen-written cursive of this period:32 Figure 4 The corruption of ρ to γ (i.e. from δαιλατορίαι33 to δαιλατογίαι) likely happened in the Greek script, e.g. in Greek cursive of the sixth and seventh centuries. Figure 5 shows first γ, then ρ in Greek pen-written cursive of this period:34 Figure 5 Two pieces of evidence support the conjecture ῥαίλατ- in the margin of Burney 124: (i) the transmitted gloss, which suits the meaning of ῥελατορία/relatoria (“receipt”); and (ii) the palaeo- graphical elucidation of the scribal error in Barocci 50 and sub- sequently in Burney 124. 31 I print ῥελατορίαι, an emendation based on the reading ῥαίλατ- (Burney 124 in margine). See further section 7 below. 32 See e.g. the Latin cursive alphabet in Thompson, An Introduction 336; drawings after Thompson. 33 It is more economical to include the corruption of ε into αι already at this stage, although δελατορίαι may have preceded δαιλατορίαι. 34 See e.g. the Greek cursive alphabet in Thompson, An Introduction 194; drawings after Thompson. 354 ΛΕΞΕΙΣ ΡΩΜΑΙΑΙ ΚΑΤΑ ΣΤΟΙΧΕΙΟΝ ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 The lemma ῥελατορία is attested in Photius (ρ 80), the Συναγωγὴ λέξεων χρησίµων (versio antiqua, ρ 24), and the Suda (ρ 96), but the entry, differently glossed, must go back to another source: ῥελατορία· ἀναφορά. ῥελατορία: carrying back. 3.2. The entry on κυαίστωρ/κοιαίστωρ Both the Suda and the Λέξεις include an entry on κυαίστωρ, a direct loan from Latin quaestor,35 but their glosses are different. Barocci 50 (fol. 110r.12) glosses κυαίστωρ with ταµίας alone: κυαίστωρ· ταµίας. κυαίστωρ: holding the office of quaestor. The Suda (κ 2533) in turn has the alternative spelling κοιαίστωρ and includes a more extensive gloss without the term ταµίας:36 Κοιαίστωρ. τὴν τοῦ κοιαίστωρος διέπων ἀρχήν, ἣν οἶµαι ἀπὸ τοῦ ἀναζητεῖν ὧδε λελέχθαι παρὰ Ῥωµαίοις. Κοιαίστωρ. Managing the office of quaestor, which I think derives from the fact that this was the word for “investigate” among the Romans. Despite the brevity of the gloss ταµίας in our treatise, one can safely conclude that its source is other than the Suda’s. The Suda’s source is the sixth-century historian Menander Protector, De legationibus Romanorum ad gentes 11.1–4: ὅτι δὴ στέλλεται κατὰ τὴν Περσῶν χώραν πρεσβευτὴς Τραϊανὸς ἐν τοῖς βασιλείοις πατράσι τελῶν καὶ τὴν λεγοµένην τοῦ κοιαίστορος διέπων ἀρχήν, ἣν οἶµαι ἐκ τοῦ ἀναζητεῖν ὧδε 35 See Dickey, Latin Loanwords s.v. κουαίστωρ, with further bibliography; cf. Lex.Byz.Gr. s.v. κυαίστωρ. 36 That the word is transmitted with the spellings κυ- and κοι- is not sur- prising, for both υ and οι were pronounced [y] (this sound has not yet lost its lip rounding to become [i]). For the distinction between [y] and [i] until almost the beginning of the late medieval Greek period see D. Holton, G. Horrocks, M. Janssen, T. Lendari, I. Manolessou, N. Toufexis, The Cambridge Grammar of Medieval and Early Modern Greek I (Cambridge 2019) 11. STEPHANIE ROUSSOU 355 ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 λελέχθαι Ῥωµαίοις. that Trajan is sent as an ambassador to the country of the Persians, being a member of the noble patricians and managing the so-called office of quaestor, which I think derives from the fact that this was the word for “investigate” in Rome. As Roth and Whitehead point out in the Suda On Line,37 for ἀπὸ τοῦ ἀναζητεῖν to make sense one needs to convert the Greek ἀναζητέω into the Latin quaero. 3.3. The entry on κεντυρίων The Λέξεις include a lemma κεντυρίων directly borrowed from Latin centurio,38 also attested in Hesychius, Photius, the Συναγωγὴ λέξεων χρησίµων, and the Suda. The gloss in Barocci 50 differs slightly from the one in the other lexicographical sources. In the Suda (κ 1345) (and in Hesychius κ 2235, Photius κ 572, and the Συναγωγὴ λέξεων χρησίµων, versio antiqua κ 274) the lemma reads: κεντυρίων· ἑκατόνταρχος. centurion: commander of a hundred (soldiers). In Barocci 50 (fol. 110r.9) the corresponding lemma is: κεντυρίων· ρ̅ στρατιωτῶν ἄρχων.39 centurion: commander of a hundred soldiers. Instead of the single ἑκατόνταρχος Barocci 50 transmits a periph- rasis. This gloss, not attested elsewhere, most probably comes from a different, unknown source. 3.4. The entry on κώδιξ/κώδικα Hesychius and the Suda share a lemma κώδιξ (accented thus with an acute, not a circumflex), a direct loan from the Latin codex/caudex,40 whereas in Barocci 50 we find the lemma κόδικα 37 https://www.cs.uky.edu/~raphael/sol/sol-entries/kappa/2533. 38 See Dickey, Latin Loanwords s.v. κεντ(ο)υρίων, with further bibliography. 39 I print ἄρχων instead of the transmitted ἄρχω. The diplomatic tran- scription of this entry is: κεντυρίων. ρ̅ στρατιωτῶν ἄρχω. 40 See Dickey, Latin Loanwords s.v. κῶδιξ, with further bibliography; cf. Lex.Byz.Gr. s.v. κῶδιξ. 356 ΛΕΞΕΙΣ ΡΩΜΑΙΑΙ ΚΑΤΑ ΣΤΟΙΧΕΙΟΝ ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 (sic).41 Hesychius κ 4780 reads: Κώδιξ· βιβλίον νόµιµον. Codex: a law book. The Suda κ 2215 includes the lemma κώδιξ and its genitive with- out a gloss: Κώδιξ, κώδικος The corresponding entry in Barocci 50 (fol. 110r.12) reads: κώδικα·42 βίβλος περιέχουσα νόµους. codex (is) a book containing laws. The lemma κόδικα requires a brief discussion. Leaving aside the orthographical error of omicron for omega, that the gloss begins with the nominative βίβλος raises the possibility that κόδικα is intended as nominative rather than accusative singular. Barocci 50 once again provides a periphrasis, περιέχουσα νόµους as op- posed to the single νόµιµον in Hesychius. This implies different sources for the material. 4. Absence of the phrase παρὰ Ῥωµαίοις in Barocci 50 or its source Some entries in Barocci 50 and the Suda share the gloss except that the former does not contain the phrase παρὰ Ῥωµαίοις (“among the Romans”). Probably a scribe or the author of the treatise left it out judging it redundant in light of the ῥωµαῖαι in the title of our treatise. The entries on φελλαγωγία, ἀντιµίσσιον, ληγατάρις, and πριµοπειλάριος furnish instances. 4.1. The entry on φελλαγωγία The Suda φ 187 contains the following entry on φελλαγωγία: Φελλαγωγία: πανήγυρις ἐπιτελουµένη παρὰ Ῥωµαίοις, ἐν ᾗ διέ- βαλλον ἑαυτούς. 41 The confusion between ω and ο is common throughout the Roman and Byzantine periods given the loss of vowel length. See e.g. Gignac, Grammar of the Greek Papyri I 275–277. For the notion that ω and ο had already fallen together in pronunciation by the mid-second century B.C. see Horrocks, Greek: A History of the Language and its Speakers 118, 167. 42 I print κώδικα instead of the transmitted reading κόδικα. STEPHANIE ROUSSOU 357 ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 Φελλαγωγία: a festival celebrated among the Romans, in which they used to slander themselves. In Barocci 50 (fol. 110v.9) the corresponding lemma reads: φελλαγωγία· πανήγυρις ἐπιτελουµένη εἰς Συρίαν ⟨ἐν⟩43 ᾗ δια- βάλλουσιν ἑαυτούς. φελλαγωγία: a festival celebrated in Syria, in which they slander themselves. To my knowledge, φελλαγωγία is attested only in the Suda, Barocci 50, and Burney 124. If this is the intended term, what might be the first member of the compound? A scholion to Aristophanes Nub. 71 in the 1498 Aldine edition (Ἀριστοφάνους κωµῳδίαι ἐννέα µετὰ σχολίων πολλῶν παλαιῶν καὶ ὠφελίµων = Aristophanis comoe- diae novem cum commentariis antiquiis admodum utilibus), probably by Marcus Musurus, observes that the Athenians have a festival called Φέλλος and that there is a place in Attica called Φελλεύς: ἐκ τοῦ Φελλέως: τόπος οὕτω καλούµενος ἐν Ἀττικῇ· οὐ µὴν ἀλλὰ καὶ περὶ τὸν Διόνυσόν ἐστί τις ἑορτὴ τοῖς Ἀθηναίοις Φέλλος καλουµένη, ἥτις τῶν Διονυσίων ἄρχεται. Λουκιανὸς δὲ ἐν τῷ Περὶ τῆς συρίης θεοῦ “φαλλούς,” φησίν, “ Ἕλληνες τῷ Διονύσῳ ἐγείρουσιν· ἐπὶ τῶν καὶ τοιόνδε τι φέρουσιν, ἄνδρας µικροὺς ἐκ ξύλου πεποιηµένους µεγάλα αἰδοῖα ἔχοντας. καλέεται δὲ τάδε νευρόσπαστα.” καὶ παρακατιὼν δέ “ἐν τοῖς προπυλαίοις τοῦ συρίης θεοῦ ναοῦ φαλλοὶ ἑστᾶσιν, οὓς Διόνυσος ἐστήσατο.” ἐκ τοῦ Φελλέως: a place so called in Attica; but the Athenians also have a festival connected with Dionysus called Φέλλος, which begins the Dionysia. Lucian in his On the Syrian Goddess says “the Greeks erect phalli for Dionysus; on which they carry something such as this, small men made of wood and having large private parts. And these are called puppets” and going further down “in the entrance to the temple of the Syrian goddess stand the phalli which Dionysus set up.” Whitehead points out in the Suda On Line44 that φελλαγωγία could 43 I supply the missing ἐν. The diplomatic transcription of this entry is: Φελλαγωγία. πανήγυρις ἐπιτελουµένη εἰς συρίαν ἧ διαβάλλουσιν ἑαυτούς. 44 https://www.cs.uky.edu/~raphael/sol/sol-entries/phi/187. 358 ΛΕΞΕΙΣ ΡΩΜΑΙΑΙ ΚΑΤΑ ΣΤΟΙΧΕΙΟΝ ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 be a mistake for φαλλαγωγία (“carrying of the phallus”) or φαλλαγώγια (phallic festival). The mention of φαλλός by Ari- stophanes (Ach. 243, 260) and Lucian (Syr.D. 16) supports this proposal. Whatever the original reading, whether φελλαγωγία, φαλλαγωγία, or φαλλαγώγια, none of these appear to derive from Latin. The absence of παρὰ Ῥωµαίοις from Barocci 50 is probably motivated by εἰς Συρίαν and not by any sense of redundancy. 4.2. The entry on ἀντιµίσ(σ)ιον Both the Λέξεις and the Suda (α 2685) contain an entry on ἀντιµίσ(σ)ιον. The gloss in these sources is unlike the one in Lex.Byz.Gr. s.v. ἀντιµίνσιον, “portable altar, altar cloth” (see also the Suda On Line,45 where Whitehead and Roth note that the altar cloth contained relics of saints and substituted for a portable altar). The entry in Barocci 50 (fol. 109v.20) reads: ἀντιµίσσιον· τράπεζα πρὸ τοῦ δικαστοῦ κειµένη.46 ἀντιµίσσιον: a table placed before the judge. Suda α 2685 glosses the lemma ἀντιµίσιον thus: Ἀντιµίσιον: παρὰ Ῥωµαίοις τράπεζα πρὸ τοῦ δικαστηρίου κει- µένη. Ἀντιµίσιον: among Romans (it is) the table placed before the court of law. Barocci 50 differs from the Suda in that the phrase παρὰ Ῥωµαίοις is missing and it says πρὸ τοῦ δικαστοῦ instead of πρὸ τοῦ δικα- στηρίου. A table lying “before a court of justice” makes no sense if the court is indoors or if the table lies outside. Probably the author of our treatise or a scribe found the gloss incompre- hensible and replaced δικαστηρίου with δικαστοῦ. As the lectio difficilior πρὸ τοῦ δικαστηρίου is the more likely original reading. Its sense must be “in front of the people of the court,” rather than in front of the court itself. This lemma is not attested in Hesychius, Photius, the Συναγωγὴ 45 https://www.cs.uky.edu/~raphael/sol/sol-entries/alpha/2685. 46 The diplomatic transcription of this entry is: ἀντιµίσσιον· τράπεζα προ του δικαστοῦ κειµενη. STEPHANIE ROUSSOU 359 ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 λέξεων χρησίµων, or the Etymologicum Genuinum (the Suda’s main sources).47 We are probably dealing here with an unknown, possibly lost source. δικαστοῦ in turn is also best attributed not to the scribe of Barocci 50 but to an unidentified source.48 4.3. The entry on ληγατάριος/ληγατάρις Suda λ 403 includes the following on ληγατάριος (borrowed from Latin legatarius and sometimes spelled λεγατάριος):49 Ληγατάριος· εἶδος ἄρχοντος παρὰ Ῥωµαίοις. Ληγατάριος: a kind of official among the Romans. Barocci 50 (fol. 110r.15) ends the lemma in -άρις50 and glosses it without the phrase παρὰ Ῥωµαίοις:51 Ληγατάρις· εἶδος ἄρχοντος. Ληγατάρις: a kind of official. While the exchange of -ιος and -ις is common and could have been effected by the scribe, different sources cannot be pre- cluded. 4.4. The entry on πριµοπιλάριος/πριµοπειλάριος Suda π 2288 says the following on πριµοπιλάριος (an alternative form of πριµιπιλάριος and a direct loan from Latin primipilaris/ primipilarius/primopilaris):52 πριµοπιλάριος· ὁ τὸ µεῖζον ἀξίωµα τῶν στρατιωτῶν ἔχων παρὰ Ῥωµαίοις. πριµοπιλάριος: among the Romans the one who has the higher rank among the soldiers. 47 For the sources for the Suda see e.g. Adler, Suidae lexicon I v–vi. 48 On the scribe of Barocci 50 see section 2 above. 49 See Dickey, Latin Loanwords s.v. ληγατάριος, with further bibliography; cf. Lex.Byz.Gr. s.v. λεγατάριος. 50 For the pattern of alternation seen in ληγατάριος ⁓ ληγατάρις see Gignac, Grammar of the Greek Papyri II 25–29. 51 Already pointed out by Adler. 52 See Dickey, Latin Loanwords s.v. πριµιπιλάριος, with further bibliography, and Lex.Byz.Gr. s.v. πριµιπιλάριος. 360 ΛΕΞΕΙΣ ΡΩΜΑΙΑΙ ΚΑΤΑ ΣΤΟΙΧΕΙΟΝ ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 Once again Barocci 50 (fol. 110r.21) omits παρὰ Ῥωµαίοις and spells the lemma differently:53 πριµοπειλάριος· ὁ τὸ54 µεῖζον ἀξίωµα τῶν στρατιωτῶν ἔχων. πριµοπειλάριος: the one who has the higher rank among the soldiers. Both spellings are equally acceptable.55 5. Same gloss but different lemma We now consider entries in the Suda and the Λέξεις that transmit the same gloss under different lemmata. Suda π 802 reads: πατρωικὸν δίκαιον· τὸ ἐκ προστάτου δίκαιον. πατρωικὸν δίκαιον: the right of a patron. Adler notes the omission of δίκαιον in the corresponding lemma of the Λέξεις (Barocci 50 fol. 110r.19): πατρωνικόν·56 τὸ ἐκ προστάτου δίκαιον. πατρωνικόν (is) the right of a patron. Here πατρωνικόν (a derivative of patronus via πάτρων + -ικος) is used as a noun, while the Suda uses the adjective πατρωικόν with δίκαιον. πατρωικόν is nowhere else attested.57 Roth and Whitehead suggest that this entry came from a commentary on Aristotle’s Nicomachean Ethics (1134b8–9), where we find: τὸ δὲ δεσποτικὸν δίκαιον καὶ τὸ πατρικὸν οὐ ταὐτὸν τούτοις ἀλλ’ ὅµοιον. 53 Adler already notes the omission. 54 I have supplied the missing accent. 55 Iotacism makes the exchange of ει and ι common during the Roman and Byzantine periods: Gignac, Grammar of the Greek Papyri I 189–191, and Horrocks, Greek: A History of the Language 167–168. 56 I have accented the word as πατρωνικόν instead of πατρώνικον as trans- mitted. On its meaning see Dickey, Latin Loanwords s.v. πατρωνικός, with further bibliography, and Lex.Byz.Gr. s.v. πατρωνικόν. 57 Already noted by Roth and Whitehead, in the Suda On Line (https://www.cs.uky.edu/~raphael/sol/sol-entries/pi/802). A TLG search confirmed this. STEPHANIE ROUSSOU 361 ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 Justice between master and slave and between father and child is not the same as absolute and political justice, but only analogous to them.58 I do not believe that the wording sufficiently supports depen- dence on Aristotle, since neither the Suda nor Barocci 50 mentions δεσποτικὸν δίκαιον or includes the word πατρικόν. The Lex.Byz.Gr. includes an entry πατρωνικόν in the sense of “jus patronatus.” πατρωνικός (“of or for a patron”) is attested in documents of the second and third centuries A.D. (SB 12533.6, P.Oxy. IX 1205.6, PSI IX 1040.17)59 and in sixth-century literary sources like Justinian’s Novels. Thus, the lemma in Barocci 50 was in use eight centuries before the date of the manuscript. On the other hand, the ‘ghost form’ πατρωικόν in the Suda could be due to the conflation of πατρωνικόν and πατρικόν or to a scribal omission of -ν-. 6. Different spelling of the same lemma Suda σ 966 includes an entry on σπόρτουλλα, the plural of σπόρτουλον (usually spelled with one lambda60 and a direct loan from the Latin word sportula): Σπόρτουλλα· δῶρα ἐπὶ πάντων διδόµενα παρρησίᾳ. οἶδας δέ, ὅτι καὶ σπόρτουλλα ἡµῖν χρεωστεῖς· ἀλλ’ ὅµως τῶν πτωχῶν ἕνεκεν εἰς τοῦτό σοι διαλυόµεθα. Σπόρτουλλα: gifts given lavishly on all occasions. And you know that you also owe us gifts; nevertheless, on account of the poor we absolve you in this regard . The diplomatic transcription of Barocci 50 fol. 110v.6 on σπόρ- τουλα reads as follows: σπόρτουλα· δῶρα ἐπι πάν διδόµενα παρρησία. σπόρτουλα: gifts given lavishly on the whole. 58 Transl. H. Rackham, Aristotle Nicomachean Ethics (Cambridge [Mass.] 1968). 59 See Dickey, Latin Loanwords s.v. πατρωνικός/πατρονικός. 60 For further variants see Lex.Byz.Gr. s.v. σπόρτουλον and Dickey, Latin Loanwords s.v. σπόρτουλον. 362 ΛΕΞΕΙΣ ΡΩΜΑΙΑΙ ΚΑΤΑ ΣΤΟΙΧΕΙΟΝ ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 One might think that πάν here stands for πᾶν, but the parallel passage from the Suda suggests that πάντων was intended and that the end -των was simply omitted in error. I therefore propose: σπόρτουλα· δῶρα ἐπὶ πάντων διδόµενα παρρησίᾳ. σπόρτουλα: gifts given lavishly on all occasions. Other than this error, the identical wording of the first sentence points to a common source. The second sentence in the Suda comes from John Chrysostom’s Letter 217 to Valentinus (PG 52 731.23), except that σπόρτουλα is spelled there with a single lambda and that ἐκείνων ἕνεκεν replaces τῶν πτωχῶν ἕνεκεν:61 Οἶδας γὰρ ὅτι καὶ σπόρτουλα ἡµῖν χρεωστεῖς· ἀλλ’ ὅµως καὶ ἐκείνων ἕνεκεν εἰς τοῦτό σοι διαλυόµεθα. For you know that you also owe us σπόρτουλα; for all that, even on account of those we absolve you in this regard. Given the spelling of sportula,62 one would expect the Suda to transmit the loanword with a single lambda, just as Barocci 50 does. Because this is not so, and because the Suda alone quotes Chrysostom, we must again assume different sources for the re- spective entries. 7. The good readings in Burney 124 In this study we have come across three good readings pre- served in the margin of Burney 124. These informed my emenda- tions of ἀπουρία to σπουρία in section 1.1 and of συνέσεως to συναινέσεως in section 1.3, and my restoration of ῥελατορίαι from the corrupt δαιλατογίαι in section 2.1. A brief discussion of the status of these readings is now in order. These marginal notes may be scribal conjectures or they may come from a source other than Barocci 50. To affirm with confidence that they are the scribe’s own conjectures one would need to establish his pos- sessing commensurate philological skills. Although we do not 61 Cf. A. Favuzzi, “False attribuzioni e nuovi riconoscimenti nella Suda,” Annali della Facoltà di Lettere e Filosofia, Università degli Studi di Bari “Aldo Moro” 51 (2008) 60. 62 See Dickey, Latin Loanwords s.v. σπόρτουλον. STEPHANIE ROUSSOU 363 ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 know his identity, to judge from the texts included he must have had philological interests. Among other works we find a copy of Choeroboscus’ and Theognostus’ works on orthography. That all three readings are preceded by the abbreviation ἰσ. (for ἴσως, “perhaps”) supports the view that they are conjectural. Had they come from another source, one would expect γράφεται instead. Even if these readings are not the scribe’s own, his knowledge of Greek led him to appreciate their quality and to include them in his manuscript. Despite its late seventeenth-century date, Burney 124 has turned out to be a valuable aid for the reconstruction of the text. 8. Some conclusions In section 1.1 I argued that ἀπουρία probably was a scribal error and that the original lemma was σπουρία. This proposal provides an important correction to LSJ and the DGE, both of which include ἀπουρία as a lemma. If I am right, ἀπουρία would be a ‘ghost word’, a scribal error in the manuscript tradition of the Suda, Pseudo-Zonaras, and the Λέξεις. In section 1.2 I argued that the good reading δωνάτιβα, trans- mitted in Barocci 50, was not a scribal emendation but came from a source that was different from, and earlier than, the one avail- able to the Suda. Thus, Barocci 50 emerges as the oldest witness to this lexicographical entry. This highlights the importance of the treatise Λέξεις ῥωµαῖαι κατὰ στοιχεῖον, which turns out to pre- serve authentic material from earlier lexicographical sources that would otherwise have been lost. In section 2.1 I sought to understand the word θονγραµ- µασίγγελον. Although obscurities remain, if bovu(m) animalium via(m) lies behind it, this would suggest an ultimate Latin source from the lemma. The evidence reviewed that points to a Latin source for the Λέξεις, at least for the lemmata transmitted in Latin script, provides valuable information for the layout and languages in which both our treatise and other lexicographical works with Latin loanwords may have circulated. If I am right that θονγραµµασίγγελον resulted from misreading Roman cursive 364 ΛΕΞΕΙΣ ΡΩΜΑΙΑΙ ΚΑΤΑ ΣΤΟΙΧΕΙΟΝ ————— Greek, Roman, and Byzantine Studies 62 (2022) 341–364 of the fourth and fifth centuries, we would know approximately when the lemmata of the Λέξεις (perhaps also of other treatises) were transcribed from Latin into Greek. If we take into account all the corruptions involving scripts, Latin or Greek, we find that misreadings of Roman cursive probably date to the first, second, and fourth or fifth centuries; while misreadings of Greek cursive probably date to the sixth or seventh century. While this may be accidental, when the potential misreading of bovu(m) animalium via(m) as θονγραµµα- σίγγελον is added to our consideration, the period from the fourth or fifth century until the sixth or seventh century emerges as a transitional period during which the lemmata of this treatise were gradually transcribed from Latin into Greek script.63 September, 2022 Aristotle University of Thessaloniki stephanie.roussou@gmail.com 63 I am grateful to Eleanor Dickey and Nigel Wilson for their feedback on an earlier version of this article; to Philomen Probert for expert advice on linguistic matters; to two anonymous referees for their useful comments; and to José M. González for generous help with form and content.