βιβλιαγράφος, βιβλιογράφος, βιβλογράφος
(Phryn. Ecl. 59, Phryn. PS 52.13, Poll. 7.211, Philemo [Laur.] 356, Philemo [Vindob.] 393.28, Orus fr. Α 20)

D. General commentary

The compound βιβλιαγράφος/βιβλιογράφος/βιβλογράφος specifically designates the scribe of a text – that is, the individual responsible for the mechanical process of copying a text from a pre-existing exemplar – as a figure distinct from the authors of the texts being copied. For a general discussion of the meaning of the term and its relationship to other compounds within the same lexical field related to books and writing, see Lobeck (1820, 655–6); Haeberlin (1890, 289–90); Mayser, Gramm. 1,1, 71; Haines-Eitzen (2000, 23–6); Caroli (2012); Blumell (2014, 33); Orth (2014, 69); Radici (2018); Lewis (2020, 162); Olson (2021, 23); Fiorentini (2022, 177–9).

In Ecl. 59 (A.1), Phrynichus compares the ancient and rare form βιβλιαγράφος, attested only in early classical Attic literature (C.1), with the tetrasyllabic form with o, βιβλογράφος, arguing that the former reflects how people in Attica pronounced the word for ‘scribe/copyist’. This same form appears to be prescribed by Phrynichus in his Praeparatio sophistica (A.2). Around the same period, the pentasyllabic form with α is also mentioned by Pollux (A.3), who cites C.1 as a source and contrasts it with another pentasyllabic variant, βιβλιογράφος, for which we have an isolated attestation before the 2nd century CE (C.2, quoted by A.3). From that period onwards, as shown below, literary usage consistently favours this pentasyllabic form with o (see C.3, C.4, C.5, C.6, C.7, C.8), which is also endorsed in lexicographical sources (Philemon, A.4 and A.5). The scantiness of the evidence makes it impossible to ascertain whether both pentasyllabic forms coexisted in the classical period, or whether the one with ο – either tetrasyllabic (βιβλογράφος) or pentasyllabic (βιβλιογράφος) – represents a later development from the form with α. Another dimension to the issue is that the tetrasyllabic variant βιβλογράφος is presented as the preferred form by some Atticist lexica (see Orus, A.6; the Vatican lexicon edited by Reitzenstein 1892–1893, B.2). Beyond the evident divergence between Atticist positions and literary usage – which in itself would not necessarily pose a problem – what proves more difficult to explain is the inconsistency among the various lexica regarding the form they prescribe as correct. While Pollux (A.3) merely notes the existence of the two pentasyllabic variants in α and o, attested respectively in Cratinus (C.1) and Antiphanes (C.2), Phrynichus (A.1, A.2) clearly prescribes βιβλιαγράφος and rejects the tetrasyllabic βιβλογράφος. On the contrary, both Philemon (A.4, A.5) and Orus (A.6) prescribe βιβλιαγράφος, but they differ in their preference for what they consider the reputable form (the former recommends the pentasyllabic one in ο, the latter the tetrasyllabic variant).

After briefly outlining a possible morphological explanation for the form with α, this entry presents an overview of the debate surrounding the variants of the compound, following a diachronic progression, and subsequently provides a detailed analysis of the literary attestations, also arranged in diachronic order.

The only extant literary attestation of βιβλιαγράφος is found in the lost comedy Chirons by the fifth-century BCE Athenian playwright Cratinus (C.1), as cited by Pollux in a digression on the semantic field of books and librarian activities (Poll. 7.211, A.3). Modern scholars generally explain the vocalism α by interpreting the first part of the compound as a neuter plural. In this regard, Haeberlin (1890, 290) and Thackeray (1909, 77) note the analogy with the similar compound βιβλιαφόρος ‘letter-carrier’, attested across a wide chronological span in both literary and papyrological/epigraphic sources (cf. e.g. Plb. 4.22.2; D.S. 2.26.8, 11.21.4, 11.28.5, 11.45.2, 11.54.3). Interestingly, βιβλιαφόρος is used also by Pseudo-Zonaras (451.6–7: γραμματοφόρος ὁ βιβλιαφόρος καὶ οὐχὶ βιβλιοφόρος, ‘You should say βιβλιαφόρος, and not βιβλιοφόρος, as a variant for γραμματοφόρος’), and, in discussing this compound, the Byzantine lexicographer appears to align with Phrynichus’ prescription of the vocalism α (Α.1), although he agrees with Orus (A.6) regarding the spelling βιβλογράφος.

As for βιβλιογράφος, the noun first occurs in Antiphanes (C.2), quoted by Pollux (A.3: see Sonnino 2014, 170–1, 191–2, on Pollux’s interest for Middle and New Comedy), and then reappears in the imperial period (see below), eventually establishing itself as the most widely attested form of the compound in literary usage. If βιβλιογράφος is accepted as a variant attested from the late classical period (see Olson 2021, 23), one may follow the theory proposed by Lobeck (1820, 655), suggesting that later authors, such as Antiphanes – writing a century after Cratinus and the poets of Old Attic Comedy – ‘revived the proper, original form’ (‘formam propriam et primigeniam’) βιβλιογράφος. Unfortunately, Pollux merely references Cratinus’ and Antiphanes’ comedies as attestations of the two forms, βιβλιαγράφος and βιβλιογράφος, without providing any additional contextual information (neither quoting the full passage nor providing a paraphrase). This omission also prevents any meaningful evaluation of whether the choice between these variants by the two comic authors might have been motivated by subtler semantic distinctions. For example, in light of the possible interpretation of βιβλια- as a neuter plural, it remains unclear whether the usage of the compound was determined by the enunciative context – perhaps referring either to the copying of a single specific text (hence βιβλιο-γράφος) or to the copyist’s profession in general, thereby implying an undetermined plurality of books (hence βιβλια-γράφος).

The interpretative complexities raised by Pollux’s passage (A.3) are further compounded by textual issues affecting the manuscript tradition. The three witnesses preserving this passage (cod. Par. gr. 2670 [A]Par. gr. 2670; cod. Par. gr. 2646 [F]Par. gr. 2646; cod. Salm. 40 [S]Salm. 40) attest to two radically different versions of the text: A reports βιβλιογράφον δὲ παρὰ Κρατίνῳ ἐν Χείρωσιν· βιβλιογράφος παρὰ Ἀντιφάνει ἐν Σαπφοῖ· παρὰ δὲ τῷ νεωτέρῳ Κρατίνῳ, ἀπεμπολημένῃ βιβλιοθήκη, while F and S have βιβλίον γράφων δὲ παρὰ Ἀριστοφάνει ἐν Σαπφοῖ. βιβλογράφων δὲ παρὰ Κρατίνῳ ἐν Ὑποβολιμαίῳ (Ἱπ- in S) βιβλιοθήκην. Beyond the various discrepancies (the replacement of ἀπεμπολημένη in A with the comedy title in FS, and the different inflection of βιβλιοθήκη[ν]) and the respective textual errors (the mistaken repetition βιβλιογράφον-βιβλιογράφος in A, which prompted Lobeck’s correction to βιβλιαγράφον (Lobeck 1820, 655–6) based on A.1; the incorrect forms βιβλίων γράφων and Ἀριστοφάνει in FS), the most significant divergence lies in the omission of the reference to Cratinus’ Chirons in FS and in the merging of the second variant of the compound discussed in A (modified to βιβλογράφος) into the quotation from Cratinus the Younger (Cratin.Iun. fr. 11: see Caroli 2011 and 2014, 152–6; Mastellari 2020, 131–3) regarding βιβλιοθήκη (βιβλογράφων [...] βιβλιοθήκην, FS). Given the relevance of the testimony about Cratinus the Elder, also confirmed by the citation of the title of a play he actually wrote (see Kaibel apud Kassel, Austin, PCG vol. 4, 257), the text offered by A appears overall the most reliable, requiring only the emendation of βιβλιογράφον to βιβλιαγράφον. See below, F.1, for further commentary on Cratin. fr. 267 (C.1) and the possible identification of the βιβλιαγράφος alluded to by the poet.

Due to the scarcity of evidence from the classical period, the relationship between the two variants of the compound in -ια- and -ιο-, whether in historical-linguistic or semantic terms, remains substantially unclear. What can be established with greater confidence is that the noun βιβλιογραφία and the verb βιβλιογραφέω, both exclusively formed with ο, appear at a later chronological stage than the occurrences of βιβλιαγράφος and βιβλιογράφος cited by Pollux (A.3). βιβλιογραφία is attested from the late classical/Hellenistic period onwards (Pers. Stoic. fr. 435; Dsc. 1.85; D.L. 7.36), while the verb βιβλιογραφέω emerges in the Byzantine period, already carrying the distinct meaning of ‘writing books’ rather than merely ‘copying’ them (see below, E.).

Whereas Pollux’s discussion appears to be purely descriptive, lacking any prescriptive stance, his contemporary Phrynichus – both in his Eclogue (A.1) and Praeparatio sophistica (A.2, where the mistaken reading βιβλογράφος was corrected by de Borries 1911, 52) – clearly prescribes the form βιβλιαγράφος, based on Attic usage. For a discussion of the difference in approach between Pollux and Phrynichus, and the alleged rivalry between the two, see entry Iulius Pollux, Ὀνομαστικόν (Onomasticon), with all the relevant bibliography. Phrynichus focuses specifically on the contrast between βιβλιαγράφος and the tetrasyllabic form βιβλογράφος, without explicitly referencing the corresponding pentasyllabic form with ο, which is instead the subject of Pollux’s descriptive digression. Moreover, Phrynichus is the only lexicographer to propose that βιβλιαγράφος should be preferred as the pure Attic form. Already in a period roughly contemporary with Phrynichus (circa 200 CE), a preference for the compound form with ο, and specifically for βιβλιoγράφος, can be found in the Atticist grammarian Philemon (Philemo (Laur.) 356 = A.4, followed by Thom.Mag. 53.9 = B.3; Philemo (Vindob.) 393.28 = A.5). Although Philemon’s lexicon generally relies on Phrynichus (who thus provides a terminus post quem for dating Philemon’s work), this is one of the few instances in which he appears to diverge from his source. For a recent overview of Philemon’s chronology and work, see entry Philemon, Περὶ Ἀττικῆς ἀντιλογίας τῆς ἐν ταῖς λέξεσιν. This position is also supported by Herodian (B.1), where βιβλιογράφος is listed as an example of paroxytone accentuation in compounds derived from γράφω.

The tetrasyllabic form βιβλογράφος appears in literary sources much later than βιβλιαγράφος and βιβλιογράφος. Its presumed first attestations are in Libanius (Lib. Or. 1.43 and 62: see the discussion below), but, as demonstrated by the entry in Phrynichus’ Eclogue (A.1), it must already have been in common usage by the 2nd century CE. In contrast to the theory advocated by Phrynichus, the 5th-century CE grammarian Orus (A.6) supported βιβλογράφος against βιβλιαγράφος. This entry is preserved in the Byzantine lexicon of Pseudo-Zonaras, but is attributed to Orus by Alpers (1981, 160) based on its clear anti-Phrynichean stance (for a recent discussion of Orus’ distinctive ‘anomalist’ position within the Atticist debate, and his frequent disagreement with Phrynichus’ doctrines, see entry Orus, Ἀττικῶν λέξεων συναγωγή). While it is possible that Orus drew on earlier lexicographical/grammatical sources (perhaps dating back to the 2nd-century CE Atticist debate), the main purpose of his entry appears to be to amend Phrynichus’ prescription. This is probably because the form βιβλιαγράφος seemed archaic and obscure to the 5th-century CE grammarian. Supporting this view is an unidentified Atticist lexicon (B.2) preserved in a 15th-century manuscript (cod. Vat. gr. 12), which explicitly attributes the form βιβλογράφος to Attic usage and merely records the attestation of the other two pentasyllabic variants (Reitzenstein 1892–1893, 4, also quoted in Alpers 1981, 160 and Fiorentini 2022, 179).

The question of the historical-linguistic relationship between the two variants with α and ο remains open (see above), as does the question of the relationship between the tetrasyllabic form prescribed by Orus/Pseudo-Zonaras (A.6) and the pentasyllabic βιβλιογράφος. In an attempt to justify the compatibility of Pollux’s testimony with the two prescriptive stances provided by Phrynichus and Orus, Olson (2021, 23) discusses the possibility that Pollux himself wrote βιβλογράφος, and that the letter ι was added by a later scribe, based on the example of the adjacent βιβλιαγράφος. Alternatively, it is possible that Pollux wrote βιβλιογράφος with the letter ι, and that he was reading from a copy of Antiphanes’ text in which the original βιβλογράφος had already been corrupted by the (mistaken) addition of ι. In light of the available evidence, it is more prudent to adhere to the testimony of the manuscript tradition of Pollux, which records the form βιβλιογράφος in A and the erroneous reading βιβλίον γράφων (‘much too uninteresting a phrase to have been collected and commented on by Pollux’: Olson 2021, 22) in F and S, which in any case must have arisen from a misinterpretation of the form with ι (on the textual issues of A.3, see above). Unfortunately, the absence of the original metrical context of the quotation precludes a decisive evaluation of whether βιβλιογράφος or βιβλογράφος was the original form. Nonetheless, assuming an iambic context, both variants are metrically admissible (see Olson 2021, 21), which further complicates resolving this issue. For the hypothesis of a possible semantic disambiguation between the forms βιβλογράφος and βιβλιογράφος, of which Libanius may provide an example within the same work (Lib. Or. 1), see the discussion below.

To sum up, lexicographers from the imperial to the Byzantine period did not agree on the spelling of βιβλιαγράφος/βιβλιογράφος/βιβλογράφος. While Phrynichus (A.1 and A.2) prescribed the rare and ancient βιβλιαγράφος, Philemon (A.4 and A.5) permitted only βιβλιογράφος, whereas Orus (A.6) favoured the tetrasyllabic βιβλογράφος. Bearing these conflicting opinions in mind, it is now worth examining in detail how prose authors contemporary to (or not much later than) imperial Atticism such as Galen, Lucian, Libanius, and Eusebius (as well as their manuscript transmission and modern editors) dealt with these forms.

GalenGalen records 20 occurrences of βιβλιογράφος (two in Gal. In Hp. Epid. I 2.85 Wenkebach = 17a.196.5–7 Kühn), only one of which was cited above for brevity’s sake (C.3). In all of these Galenic attestations, which systematically appear in commentaries on Hippocratic texts, the term is used to refer to the copyists of the work attributed to Hippocrates. These passages are particularly valuable from a historical-literary and philological perspective, as they touch upon characteristic features of manuscript transmission and textual criticism more broadly. C.3 describes a mistaken alteration of a syntagm in a Hippocratic passage by the first copyist as a possible explanation for its (unjustified) omission by the editor Artemidorus Capito (see Raiola 2018, 59); in Gal. In Hipp. De vict. acut. comm. 4.115 Helmreich (= 15.911.8–10 Kühn), the term occurs in relation to the omission of a passage; in Gal. In Hipp. VI Epid. comm. 1 (Pr.) Wenkebach (= 17b.794.14–7 Kühn), the βιβλιογράφος is said to have made a palaeographical error (θύραι > οὐραί); in Gal. In Hipp. Aphor. comm. 4.52 (= 17b.732.5–7 Kühn), the βιβλιογράφος is accused of having replaced the positive form of an adjective with the comparative; in Gal. In Hipp. Aphor. comm. 4.79 (= 17b.775.13–6 Kühn), the scribe has omitted a textual portion. See López Férez (2013) and López Férez (2019) for a comprehensive commentary on Galenic occurrences of scribal activity, and, more generally, on Galen’s interest in textual criticism. See also Manetti, Roselli (1994); Roselli (2012), and Raiola (2018) for a discussion of Galen as a witness to the philological treatment of Hippocratic texts in the 2nd century CE. Galen’s occurrences are especially significant as they constitute the earliest attestations of βιβλιογράφος after the isolated instance in Antiphanes (C.2), thus bridging a chronological gap of several centuries.

Another attestation of the term βιβλιογράφος from the time of imperial Atticism can be found in Lucian of Samosata’s Against the Ignorant Book-Collector (Luc. Ind. 24, C.4). In this regard, Fiorentini (2022, 179), presumably following LSJ s.v. βιβλιογράφος, argues that Lucian appears to be the only post-classical author to follow Phrynichus’ norm and employ the form with α, thereby providing a unique attestation of this specific form. Fiorentini is correct in claiming that the deliberate use of this form, endorsed by the Atticist doctrine of Lucian’s contemporary Phrynichus, is a plausible hypothesis, especially given Lucian’s well-attested familiarity with Greek comedy, his consistent adherence to Atticist linguistic norms, and the density of allusions and quotations from the classical Athenian literary tradition in this specific work. Nevertheless, although LSJ s.v. βιβλιογράφος cites C.4 as evidence of the form with α (see also Dimitrakou 1949, 1405, for a reference to βιβλιαγράφος as a varia lectio for this passage), all the witnesses and editions of Lucian’s text consistently report βιβλιογράφους (cf. Harmon 1921, 204; MacLeod 1974, 132; Valdés 2004, 47; Hopkinson 2008, 32), aligning with the literary usage of the author’s contemporaries. Furthermore, in light of the close connection between this passage and one from Galen (Gal. Ind. 2.13 Polemis, Xenophontos), which discusses the editorial activities of the same βιβλιογράφοι mentioned by Lucian, Callinus and Atticus (the latter likely to be identified as Cicero’s friend Titus Pomponius Atticus), it is reasonable to posit continuity between the two authors in their use of βιβλιογράφος: see Dorandi (2014, 12) and Dorandi (2016) for this parallel. In any case, the strongest argument against the plausibility of βιβλιαγράφος in Lucian, who was a far more Atticising author than Galen, is its complete absence from the manuscript tradition. Even if one were to posit the existence of an unknown source, whether manuscript or printed, from which the LSJ editors derived the form with α, it remains plausible that this reflects a hyper-Atticising correction of the transmitted reading. On the figure of Atticus and his autograph books (Ἀττικιανά / Ἀττίκεια), see Dziatzko (1896), where Lucian’s passage is also cited (again with the spelling βιβλιογράφους); Irigoin (1993, 59–60); Dorandi (2014); and Dorandi (2016). See also Boudon-Millot (2008, 105 n. 241), on the broader meaning of βιβλιογράφοι in this context as ‘librarians/editors’.

After Galen and Lucian, the compound reappears in Libanius, where it occurs four times in the first Oration (the so-called Autobiography: Lib. Or. 1.43, 62, 113 [C.5 and F.2], 155) and three in the Epistles: once in Lib. Ep. 263, addressed to a friend in Heraclea Pontica (Crispinus), in response to his request for copies of Libanius’ works; twice in Lib. Ep. 629 (C.6), addressed to the governor of Euphratensis (and former pupil) Priscianus, on behalf of a friend of Libanius’ scribe Maeonius. Regarding the occurrences of the term in Or. 1, most modern editions (see Foerster 1903–1927 vol. 1,1; Norman 1965; Martin, Petit 1979; Norman 1992) print the tetrasyllabic form βιβλογράφος, in contrast to the practice of earlier editors (e.g. Reiske 1791). This graphic practice adopted by modern editors is also reflected in the list of concordances of the term in Fatouros, Krischer, Najock (1989). In two cases, i.e. Or. 1.43 and 62, the manuscript tradition unanimously offers the reading without ι, while in the other two occurrences of the term in the Autobiography, Or. 1.113 (C.5) and 155, the tetrasyllabic form appears as a marginal annotation/correction added by a second hand in one witness of the tradition, P (cod. Par. gr. 2998).

One possible explanation for the instances in which the manuscript tradition consistently presents the form βιβλογράφος (Lib. Or. 1.43 and 62), and thus for the presence of both variants in the Autobiography, can be found in the subtle semanticSemantics distinction between βίβλος and βιβλίον, which may have been transferred to their respective compounds. According to this distinction, the former term more specifically refers to the copying of books intended for school use (Norman 1960, 124: ‘if any distinction is to be made, it is that βίβλος is almost always used to denote literary texts of common use in the schools or in the Christian and pagan religious teaching’): see Norman (1960, 124 n. 13), for some occurrences of βίβλος with this semantic nuance in Libanius. However, such a meaning does not apply to Or. 1.113, where the text being copied is specified (a panegyric for a public figure, the praetorian prefect Strategius), or to Or. 1.155, where explicit reference is made to the copying of Libanius’ own works and their intended wide circulation. In the case of Or. 1.113 (C.5), this remark would reinforce the probative value of using this passage as an exemplum in Thomas Magister’s normative statement (B.3). Even for Ep. 629 (C.6), the form with ι seems more appropriate, as it likely refers implicitly to copyists transcribing correspondence. In this case, it refers to the copyist Maeonius, mentioned by Libanius, who had personally prepared a neat copy of the letter (see below, F.3). Furthermore, the mention in Or. 1.43 of the past services performed by Libanius’ Cretan copyist during the author’s time in Athens (Κρητός τινος ἐπιεικοῦς πολλὰ τῇ δεξιᾷ καμόντος Ἀθήνησί τε καὶ πανταχοῦ) – specifically during Libanius’ upbringing in the city – supports the hypothesis of a semantic disambiguation through the use of the two variants. In this case, βιβλογράφος would carry an implicit reference to the scribe’s activity as a copyist of βίβλοι for Libanius’ own studies. This meaning could also be extended to Or. 1.62. Not only does this passage display clear thematic analogies with Or. 1.43 (accusations against Libanius and the possibility of torture being applied to one of his servants acting as a scribe/copyist), but its phrasing strongly suggests that the same person is referenced here: the bare noun βιβλογράφος is used, without any further identifying details, as in Or. 1.43 (βιβλογράφον καὶ αὐτὸς εἶλκε); above all, the text specifies that ‘he too’ (καὶ αὐτός, i.e. Libanius’ accuser) attempted to have the copyist arrested, clearly alluding to the episode already reported in Or. 1.43. See Norman (1992, 123), who translates ‘he too tried to have my copyist examined’ (i.e. the person mentioned earlier), and highlights the analogy with Or. 1.43 in the footnote to the passage.

Beyond this hypothesis of nuanced semantic disambiguation – whose clarity, of course, remains uncertain – if one prefers to regularise the textual occurrences as modern editors do, it seems more reasonable to adopt the form with ι. This choice is more consistent with its predominant attestations in earlier and contemporary literature (from Galen onwards; see also below on Eusebius, C.7), and with the very occurrence of the form with ι in the other three attestations in Libanius’ Epistles. Regarding the latter, in two cases (Ep. 263 and the second occurrence of the term in Ep. 629) the entire manuscript tradition consistently reports the form with ι. Moreover, in one case (the first occurrence in Ep. 629), two witnesses give the reading without ι, but in one of them a second hand adds the letter between λ and ο within the word itself. The presence of the immediately following βιβλιογράφων, preserved as such in all witnesses – with the sole exception of the erroneous reading βιβλία γράφων in cod. Mosq. gr. 459 (Mo); see Foerster (1903–1927 vol. 4, 577) – strongly suggests what the correct reading must also have been in the immediately preceding instance of the term in the same passage. Foerster himself, although he prints the form without ι in the Autobiography, nonetheless uses βιβλιογράφος for these two epistles (Foerster 1903–1927 vol. 4, ad loc.); the same choice is made by Norman (1960, 123 n. 4), who cites the passage from Ep. 629 with the regular pentasyllabic form. In his later editions of Or. 1 (1965 and 1992), however, as already noted, he standardises the occurrences of the noun to βιβλογράφος. Moreover, the addition made by a later hand in P at both Or. 1.113 and 1.155 can be interpreted as a varia lectio rather than a correction, particularly in light of the majority reading with ι. Furthermore, in the case of Or. 1.113, the very fact that this passage is cited by Thom.Mag. 53.9 (B.3) within the utterance of a lexicographical norm – albeit an argument a posteriori – provides significant support for the plausibility of this form as the correct one.

In addition to Libanius, Eusebius of Caesarea provides further literary evidence for the form βιβλιογράφος. In a passage from the Ecclesiastical History (6.23.2, C.7), Eusebius recounts an episode from the life of his distinguished fellow citizen Origen (mid-3rd century CE): more than seven copyists of servile status, along with young women trained in the art of calligraphy, were made available to Origen by his rich patron Ambrose, so that they might assist him in transcribing his commentaries on the Holy Scriptures. This anecdote, frequently cited in literary and exegetical testimonies from the Byzantine period (see below, E.), is particularly noteworthy as it offers – albeit in somewhat hyperbolic terms – a depiction of early Christian scribal practices. For a general discussion of this topic, see Haines-Eitzen (2000); for a commentary on Eusebius’ passage concerning this broader issue, see Blumell (2014, 37) and Schott (2019, 307 n. 93). More specifically, this passage highlights the clear distinction between βιβλιογράφος and another term from the same semantic field, ταχυγράφος (‘shorthand writer’), depicting two figures engaged in copying tasks in a logical sequence. The latter is the shorthand writer, who rapidly records dictated content in a rough form, while the former is the calligrapher, responsible for transcribing the rough material provided by shorthand writers into a refined, aesthetically polished format. The role of the βιβλιογράφοι, therefore – as also attested by Libanius (see below, F.2) – emerges as one of patient, meticulous labour, directed towards the creation of fully-fledged aesthetic artifacts.

In sum, an overview of the literary sources makes it clear that Phrynichus’ orthographic prescription was not the dominant position within ancient scholarship and literary usage, since the form with ο is much more frequently attested. Moreover, it is worth noting that Phrynichus explicitly questions only the form βιβλογράφος, in favour of the strictly Attic form βιβλιαγράφος, without clearly pronouncing on the legitimacy of the pentasyllabic form with o – which, as we know from Pollux’s citation of Antiphanes (A.3 = C.2), was indeed in use even in (late) classical Athens. Ultimately, it cannot be stated with certainty whether Phrynichus’ prescription was ever put into practice by literary authors, including those with pronounced Attic tendencies such as Lucian (who appears to conform to the more widely attested usage of βιβλιογράφος). Indeed, Phrynichus’ position emerges as merely one among several within the broader Atticist debate, as evidenced by Philemon (A.4, A.5) and Orus (A.6). Consequently, Atticist authors may well have followed one or another of these coexisting positions.

A general conclusion that can be drawn, given the insolubility of the more specific issues concerning the relationship between the three morphological variants of the compound (see above for possible semantic and historical-linguistic arguments), is that a debate on the legitimacy of these forms was already underway in the 2nd century CE. This debate had repercussions for literary/exegetical usage, as well as for lexicographical treatment in the following centuries up to the Byzantine period (Pseudo-Zonaras, A.6; Thomas Magister, B.3; the Vatican lexicon, B.2; John Tzetzes, B.4 and B.5). The antiquity of this lexical plurality – already evident in Athenian linguistic usage during the classical period (the vowel alternation α/ο between C.1 and C.2) – ultimately justifies the relevance of this compound as an object of discussion within the framework of Atticist lexicographical exegesis.

E. Byzantine and Modern Greek commentary

Byzantine Greek exhibits a coexistence of the two graphic forms of the compound with o, βιβλιογράφος and βιβλογράφος. In particular, the anecdote reported by Eusebius (C.7: see above, D.) concerning the transcription of Origen’s commentaries is repeatedly cited in Byzantine literary sources (e.g. Georgius Monachus Chronicon 454.23–455.3 de Boor = Constantinus VII Excerpta de virtutibus et vitiis 1.138.12–3 Büttner-Wobst–Roos, C.8; Zonar. Ep. Hist. 3.124) and in lexicographical works (C.7 is quoted verbatim in Su. ω 182, 618.6–10 Adler; C.8 is quoted in Su. ω 183, 621.7–9 Adler). Particularly noteworthy is C.8, which, compared to Eusebius’ account (C.7), places even greater emphasis on the nuanced distinction of roles and the sequence of actions among the various agents involved in textual transcription. This account thereby highlights the ultimate aesthetic, rather than merely practical, aims of the activity of the βιβλιογράφοι: first Origen dictated to shorthand writers (ὑπηγόρευε τοῖς ταχυγράφοις), and then the copyists produced an elegant transcription (καὶ οἱ βιβλιογράφοι […] ἔγραφον).

Moreover, during the Byzantine period, the term βιβλιογράφος came to mean ‘writer’ as well as ‘copyist’ of books, following the semantic shiftSemantic shift of the verb βιβλιογραφέω, which is only attested in this period (e.g. Eust. Opusc. 91.44, 180.74 Tafel; Eust. in Od. 1871.19).

In addition to the lexicographical prescription in Pseudo-Zonaras (based on Orus, A.6), it is worth noting a particularly remarkable instance of the form βιβλογράφος in literary/exegetical practice. This occurs in the writings of John Tzetzes, who frequently uses the tetrasyllabic form in several annotations to ancient texts, which are composed either in prose or, more frequently, in verse. In particular, one of the main categories of the several iambic scholia that Tzetzes wrote in the margin of Thucydides’ text in E (cod. Pal. gr. 252, 9th c. CE) concerns his invective against the βιβλογράφος of the manuscript and his archaic script, which caused Tzetzes considerable difficulty in reading it. Specifically, βιβλογράφος appears in six scholia as the target of Tzetzes’ invective: in addition to B.4, see also schol. (Tz.) Thuc. 3.89.2; 4.1.1; 5.14.1; 5.15.1; 5.16.2. On Tzetzes’ verse scholia to Thucydides, see Scott (1981, 60–74); Baldwin (1982, 313–16); Maltese (1995, 370–1); Luzzatto (1999, especially 21–42 on Tzetzes’ comments against the copyist of E); Reinsch (2006, 757–8); Mészáros (2013); Kaldellis (2015, 65–79); Pontani (2020, 458–9); and Bértola (2022).

A similar polemic against the mistakes made by βιβλογράφοι appears elsewhere in Tzetzes’ exegetical work. In this context, it is worth mentioning a passage from an extensive poem of no fewer than 117 iambic trimeters, which is embedded in the second edition of his commentary on Aristophanes’ Wealth (schol. (Tz.) Ar. Pl. 137 [P22]: see Massa Positano et al. 1960, 41–6; Luzzatto 1999, 43–4; Pontani 2020, 459; Roilos 2021, 261–2; Pizzone 2022, 104–7). In this passage, the author presents himself as heroically facing the ‘wave of barbarisms’ disseminated by ignorant copyists (B.5). This hyperbolic self-representation, portraying himself as a sort of new Odysseus – the seafaring hero who confronts many formidable perils and enemies – echoes, to some extent, the metaphorical association of the copyist’s obscurity with the monstrous Charybdis in schol. (Tz.) Thuc. 5.18.5 (B.4): see Luzzatto (1999, 37–9, 43–4). Additional examples include Tzetzes’ annotations to his own Histories: for marginal scholia in verse like those mentioned above, see e.g. Tz. Inscr.H. 19.2; schol. Tz. H. 13.620a.3. See also Benuzzi (2019) on Tzetzes’ tone in his commentaries on Aristophanes, particularly against the σχολιογράφοι (criticised no less harshly than the βιβλ[ι]ογράφοι), and Savio (2020) on Tzetzes’ invectives against both ancient and contemporary authors as a strategy of intellectual/literary self-promotion.

The use of the tetrasyllabic form βιβλογράφος in the quoted instances is easily justifiable on metricalMetre grounds: especially in the case of iambic trimeters, this variant can aptly be employed within this metrical pattern due to its potential iambic scansion. Moreover, the cultural and scholarly background of a scholar like Tzetzes must be considered: his consistent use of the form βιβλογράφος may well reflect linguistic and grammatical purism, in line with Orus’ doctrine, of which Pseudo-Zonaras – approximately contemporary with Tzetzes – provides a testimony with normative value (A.6). Nevertheless, given that Tzetzes represents an isolated Byzantine instance of the use of βιβλογράφος, and that Orus/Pseudo-Zonaras does not explicitly reject the pentasyllabic form βιβλιογράφος, the metrical motivation remains the most compelling explanation.

F. Commentary on individual texts and occurrences

(1)    Cratin. fr. 267 (C.1)

As is often the case in Pollux’s Onomasticon, the quotation from Cratinus is completely lacking in context. Therefore, it is impossible to offer a definitive interpretation of the identity of the βιβλιαγράφος mentioned by the poet. See also above, D., for a discussion of the textual issues posed by the manuscript transmission of Pollux’s passage (A.3). Meineke (1839b, 159) suggested that the βιβλιαγράφος could be the architectus iudicialium libellorum Pandeletus, based on a scholium to Aristophanes’ Clouds (schol. Ar. Nu. 924h = Cratin. fr. 260) which comments on the phrase Πανδελετείους γνώμας (‘the maxims of Pandeletus’), indicating that Cratinus referred to this figure in his Chirons (roughly datable between 443 and 431 BCE). This Pandeletus appears to have become a notable figure in Periclean Athens, sufficiently well-known for Aristophanes to make an intelligible reference to him even years after the performance of Cratinus’ comedy (the Clouds was first performed in 425 BCE). See Sommerstein (1990, 206) on the fame of Pandeletus in 430s Athens; Traill (2005, 46) for a discussion of the name Pandeletus and his uniqueness within the Athenian prosopographical landscape; Lambert (2010)    on the hypothesis that Pandeletus may have been a fictional character of comedy. Caroli (2012, 105) provides a more detailed explanation of this identification hypothesis, concluding that Pandeletus might indeed be a βιβλιαγράφος, given his role as proposer of challenges to recent ψηφίσματα (see also Buis 2019, 85). On the figure of Pandeletus as a writer of decrees, see also schol. Ar. Nu. 924f.α (γράφων ψηφίσματα). Alternatively, Caroli prudently hypothesises that the term could refer to the person who transcribed the Χείρωνος ὑποθῆκαι of Cratin. fr. 253 (see Caroli 2012, 99), which some scholars have identified as a reference to a pseudo-Hesiodic poem known in antiquity under that title: see Bergk (1838, 225); Meineke (1839b, 154), after more cautious remarks in Meineke (1839a, 75); Bothe (1855, 48); Kock, CAF vol. 1, 84; Luppe (1963, 214–5); and Kassel, Austin, PCG vol. 4, 257. Fiorentini (2022, 125–6) takes a more cautious approach, arguing that although Cratinus and his contemporaries were familiar with Hesiodic poetry (e.g. Cratin. fr. 349, quoting Hes. Op. 299–300; see Olson, Seaberg 2017, 138–9), and although other authors of the time likewise demonstrate knowledge of this tradition (see Dunbar 1995, 403–4, on Ar. Av. 609; Bagordo 2013, 117–20, on Telecleides’ Hesiods), it cannot be assumed that this particular expression alludes to the pseudo-Hesiodic poem, since there is no certainty that verses circulating under that title were already known at that time. The earliest attestation of the title in the form Χείρωνος ὑποθῆκαι is provided by Aristophanes of Byzantium (Ar.Byz. fr. 407), who attributed the work to Hesiod, assuming that Quint. Inst. 1.1.15, quoting Aristophanes himself, faithfully follows the latter’s usage of the term ὑποθῆκαι. Even more tentatively, can may argue, following West (1978, 23), that the passage alludes to Hesiod’s Titanomachy (Hes. fr. 6). For a comprehensive discussion of comic sources on the book trade in classical Athens, alongside Caroli (2012)    on C.1, see Caroli (2011)    and (2014, 152–6, esp. 154–6) on Cratin.Jun. fr. 11 = Poll. 7.211 (A.3); Mastellari (2020, 131–3), also on Cratin.Jun. fr. 11; and Maggio (2023, 212–13).

(2)    Lib. Or. 1.113 (C.5)

In this passage, explicitly cited by Thom.Mag. 53.9 (B.3) as an example of the occurrence of βιβλιογράφος, reference is made to the employment of ten copyists by the praetorian prefect Strategius for the transcription of a panegyric composed by Libanius in his honour. The aim was to ensure the dissemination of the work throughout the principal cities of the eastern Empire, and consequently its widespread distribution from these urban centres (see briefly Wiemer 1995, 160). This passage highlights the prestige that copyists’ work could confer upon a text whose broad transmission was desired, particularly when it had political-institutional significance (in this case, a panegyric for a praetorian prefect), by ensuring its high aesthetic quality and legibility. For a commentary on this passage in relation to the extent of the book trade and copying activities in 4th-century CE Antioch – on which Libanius’ statements throughout his oeuvre provide a contradictory overall picture – see Norman (1960, 122): ‘[…] it is regarded as both honorific and extravagant that the prefect, Strategius, should put ten copyists on the printing of a panegyric for its distribution […] especially in view of the delays in copying which private individuals had to endure at this time before texts of any new oration could be sent elsewhere’.

(3)    Lib. Ep. 629 (C.6)

In this letter addressed to Priscianus, a former student of Libanius and governor of Euphratensis at the time of its delivery (datable to the summer or autumn of 361 CE: see Bradbury 2004, 163), the rhetor intercedes with the governor on behalf of a certain Maeonius, one of the scribes in Libanius’ service, who, according to the author, was also known to the addressee. The author’s explicitly stated concern for the private needs of Maeonius and his copyists (ὅτι μοι βιβλιογράφων ἀμελεῖν οὐκ ἔνι, καὶ τοῦτο οἶσθα) confirms the importance Libanius attached to such figures. In the letter, a clear distinction is drawn between two roles within the semantic field of book-related labour: the copyist (βιβλιογράφος) and the letter-carrier (βιβλιοφόρος). It is also highly likely that Maeonius himself penned the letter, particularly given its destination to an institutional figure like the governor Priscianus. Moreover, Maeonius himself may have recommended the letter-carrier to Libanius. On the employment of βιβλιογράφοι in the copying of letters and official correspondence, see Blumell (2014) . This letter not only provides insight into the personal relationship between the author and the work of his copyists, but also more broadly illuminates the role of such figures in late antique Antioch. The servile status of most copyists in Libanius’ service is implied in Or. 1.43 and 62, where he refers to the potential use of torture on one of his scribes (here βιβλογράφος: see above, D.) as a means of obtaining confessions to support (false) accusations brought against him on two separate occasions.