Introduction

Borne in Antiquity, the editorial mode of the anthology has come to represent what Milad Doueihi says is, “the ideal form and format of the digital civilization” (Doueihi 2011, 105). The canonical model of this “literary genre” (if any) is the Crown of Meleager, a collection of epigrams composed in the first century BC that represents the primary source of what is now called the Greek Anthology. Almost all collections of Greek epigrams have been preserved through a particular manuscript: the Palatinus 23 codex, kept at Heidelberg (Aubreton 1968).

However, the form “anthology” is by nature difficult to grasp: is it really a work? In the case of the palatine manuscript that brings together written epigrams from over sixteen centuries of literary production and by more than one hundred authors, this question is all the more relevant. If we cannot consider the anthology as a unified work, what is its nature? How then do we report it in an edition?

To answer these questions, we have initiated the project of a “collaborative edition” of the Anthology that is not only compliant with the standards of digital scholarly publishing but can also benefit from the collective intelligence (Lévy 1994) deployed on the web. Since the palatine manuscript is characterized by the linking of heterogeneous texts, how are we to give an accurate account of the anthological imaginary that cannot be limited by reproducing a classic critical and genetic edition attached to the establishment of one truth of the text? What is required rather is an emphasis on the innumerable connections, which exist between the anthological texts and between these texts and external cultural objects.

In this article, we will present the theoretical postulates and the editorial foundations we have adopted, as well as the methodological and technical processes put in place to meet the particular challenges posed by our objectives.

This work is part of the tradition of Digital Humanities, and in particular of the approach of Digital Classics, particularly related to critical editions of ancient texts that promote, through the use of digital tools, an appropriation of texts by contemporary readers (Solomon 1993), (Bodard and Mahony 2016).

Greek Anthology: a problematic definition

One of the first problems a publisher encounters when he or she examines the textual material of the Greek Anthology is how to define it.

Indeed, what is the Greek Anthology? The question may seem trivial, but it is very difficult to give a precise answer. We could say, as a starting definition, that it is a collection of epigrams composed in the tenth century. In fact, the manuscript on which the editions of this text are based is the Codex Palatinus 23 (“Cod. Pal. Graec. 23 Anthologia Palatina (Konstantinopel, 2. Hälfte 9. Jh. Und 1. Hälfte 10. Jh.)” n.d.), a manuscript dating from around 940 AD, kept in the Palatine Library of Heidelberg¹ discovered in 1606 by Claude Saumaise. This collection is based on a previous anthology, composed about half a century ago by Constantin Cephalas. According to our present knowledge (Aubreton 1968, 43), most of the epigrams found in the Palatinus codex 23 derive from the work of Cephalas. Therefore, the Greek Anthology would fundamentally be the Palatine Anthology.

Consulting the most current editions (Paton, Waltz) only helps us understand the complexity of this question because the fifteen books of the Anthology that correspond to the division of the codex Palatinus, contains works by the philologists of the Loeb edition produced in 1916 by Paton and the edition of Belles Lettres de Waltz in 1927 - add volume XVI under the title of Appendix Planudea. Maxime Planude had composed his own anthology in 1301 (Aubreton 1968) and it includes some of the epigrams of the Codex Palatinus to which are added other epigrams not present in this first manuscript.

The Greek Anthology would therefore be this object reconstructed a posteriori by the modern editors who, in doing so, attempt to give a somewhat exhaustive account of the Greek epigrammatic. But we must go back a little further in our chronology to the work of Cephalas whose work is based on earlier sources, the most notable of which are the Crown of Meleager, the Crown of Philip of Thessalonica and the Cycle of Agathias of Myrina, respectively from the 1st century BC and of the 1st century AD and the 6th century AD (K. Gutzwiller 1997). Let’s start with the oldest of these works from the Hellenistic period Meleager of Gadara who composed the first collection of epigrams together in a single volume (Waltz, 1928, p. In accordance with the taste of the time, Meleager says, “to weave” with these texts “a wreath of flowers” (literal translation of the Greek “anthologia”). In the proem of his crown - which we find at the beginning of the book IV of the codex Palatinus 23 - Meleager lists these poets by putting them each in parallel with a flower:

Μοῦσα φίλα, τίνι τάνδε φέρεις πάγκαρπον ἀοιδάν; / ἢ τίς ὁ καὶ τεύξας ὑμνοθετᾶν στέφανον; / Ἄνυσε μὲν Μελέαγρος, ἀριζάλῳ δὲ Διοκλεῖ / μναμόσυνον ταύταν ἐξεπόνησε χάριν, / πολλὰ μὲν ἐμπλέξας Ἀνύτης κρίνα, πολλὰ δὲ Μοιροῦς / λείρια, καὶ Σαπφοῦς βαιὰ μέν, ἀλλὰ ῥόδα²…

The material harvested by Meleager is composed of texts, the first of which can be traced back to poets from the 6th century BC (eg Sappho) or the Vth - notably Simonide, one of the first poets to consider the epigram as a literary genre - up to Meleager himself, who inserts into his work several texts from his hand (K. J. Gutzwiller 2007). Throughout the centuries, following Meleager, the compilers succeed one another, adding to the old anthological material new epigrams, which develop and increase it. Some texts are epigrams in the proper sense of the term: texts designed to be inscribed on objects, tombs or buildings (Pfeiffer 1968). But most are poems that appropriate the epigraphic format playing with its rules: they are not real inscriptions, but poems that mimic the style of the first inscriptions. Indeed, possibly as early as the 1st century BC, the epigram became a recognized literary form. The popularity of this form exploded in the Hellenistic period when great literary figures like Callimachus and Asclepiades became very productive epigrammatists (K. J. Gutzwiller 1998). We cannot understand the meaning of this text except by thinking of it as living and moving material, which has resonated in the literary imaginary of several different societies: from archaic Greek culture to Byzantine society. Over the centuries, epigrammatists have written their own texts in the tradition of their predecessors, constantly echoing them³ and this play of stylistic and thematic amalgam takes the form of a poetic joust under the impulse of Agathias, who invites several contemporary poets to collaborate in the creation of his Cycle (Cameron 1993).

In total, the Anthology brings together a set of texts from the 6th century BC and in the 10th century AD it brings together more than a hundred poets - including several anonymous poets - who have lived in the interval of these sixteen centuries.

Is the Anthology a real text or simply a collective imaginary? We will and must come back to this question a little later because the different editions of the manuscript have taken little account of this question, as can be seen from the reviews of the various editions of the Anthology.

Critical editions of the Anthology

Since the discovery of the manuscript in Heidelberg by Claude Saumaise in 1606, many critical editions of the Anthology have been published. Before then, only Planude’s Anthology existed.

Claude Saumaise, who was planning to produce a critical edition of the Palatine Anthology but who never completely achieved this, did produce a collation of the P23 manuscript and composed a set of critical notes (Waltz 1928).

These notes were used for the first time in order to establish the text of the Anthology at the end of the eighteenth century. Thus, between 1772 and 1776, Richard-François-Philippe Brunck produced the first complete edition of the Palatine Anthology in his Analecta veterum poetarum Graecorum published in three volumes. In this edition (Tyrtée and Théocrite 1772), the epigrams do not follow the order that was found in the manuscript, but are grouped by poets, they chronologically classified. Friedrich Jacobs took over Brunck’s work a few years later and arranged under the same plan the epigrams grouping them into 13 volumes, under the general title Anthologia Graeca sive, Poetarum graecorum lusus ex recensione Brunckii. This is his first critical edition of the Anthology, published in Leipzig between 1794 and 1814 (Jacobs 1794).

Indeed, between 1813 and 1817, Jacobs realizes a second critical edition of this text, this time in three volumes (Jacobs 1813). The first two volumes include the epigrams of the Palatine Anthology, arranged in the order of the Palatine manuscript, to which are added the appendix of Planude’s epigrams, as well as 394 pieces from the ancient authors. In the third volume what has been arranged is the critical notes and indices. This second edition was based on the careful collation of the text, although it was not made on the palatine manuscript but on the copy of J. Spaletti (Waltz 1928, LXVIII)

Along with these philological works, critical studies of the Anthology and its manuscript tradition are multiplying. Johann-Friedrich Dübner integrates these research results in the edition he proposes in 1864-1877, continuing from Jacobs and renewing it. He thus edited the epigrams of the fifteen books of the Palatine Anthology, accompanied by a Latin translation, and followed by an Appendix Planudea and an appendix gathering the epigrams of the ancient authors, under the general title Epigrammatum Anthologia palatina cum Planudeis and appendix nova epigrammatum veterum ex libris and marmoribus (Dübner, Jacobs, and Cougny 1864).

It is precisely during the nineteenth century that philology is endowed with stable scientific methods. Twentieth-century philologists working on the Anthology claim to represent precisely those critical currents. Two major critical editions then emerged: that of WR Paton, published by the Loeb Classical Library in 1916-1918, and that of Pierre Waltz, published in Belles Lettres and continued, after the death of the philologist, by Jean Irigoin, Robert Aubreton and Felix Buffière. The whole of this edition (1928-1980) is composed of thirteen volumes.

In his edition, WR Paton relied on the manuscript of the Palatin Anthology as well as that of the Planude Anthology, showing the philological variants, observed in the other sources, where they differ from the established text (W. R. Paton 1916). It also includes the lemmata in the established text and accompanies its edition by an English translation of the epigrams. Pierre Waltz also wished to include, in his Greek Anthology, not only the epigrams of the Palatinus as well as those of the collection of Planude, but also the epigrams drawn from the papyrus and the metric inscriptions. It therefore takes into account the whole manuscript tradition of the Greek Anthology, but follows in priority the text of the manuscript Palatinus to establish the text of the epigrams contained in this collection. He indicates in the critical notes accompanying his text, the philological variants in an exhaustive manner (Waltz 1928, LXXXIII).

The UTET editions in Turin have recently published a last edition made up of the work of Fabrizio Conca, Nario Marzi and Giuseppe Zanetto, published in three volumes between 2005 and 2011 that propose an Italian translation of the text (Conca, Marzi, and Zanetto 2005).

These different critical editions all agree on the prevalence of the codex Palatinus 23, choosing it systematically as their main source. At the same time, they wish to bring a synthesis of the Greek epigrammatic work, including the annex pieces. Thus, philologists try to account for the character both of the abundant and scattered collections of epigrams and their editorial work to show the complexity induced by the anthological nature of these texts.

Strangely, none of the existing critical editions systematically establishes the text of the scholies⁴ and yet they are abundant in the Palatine manuscript. However, these scholies play a fundamental role, precisely from the point of view of the anthological form, insofar as they set up a particularly rich set of internal references. They relate epigrams that mention the same characters (for example, a girl to whom are dedicated two love poems) or connect the epigrams of the same author - for example, the scholia attached to the epigram 5.19 " τοῦ αὐτοῦ ἔρωτ (α) ἀλλόκοτον "which indicates that she is from the same author, but concerns another relationship. They also provide precise information on the composition and constitution of the anthological corpus, specifying the sources of the textual material. For example, the very first scholia of Book IV explain that the first poem of the book is the preface to the Crown of Meleager but it also gives us a number of biographical details. It is still the scholies who teach us that the source of the codex Palatinus 23 is the Anthology of Constantine Cephalas. Last but not least, scholies establish thematic links between different epigrams, thus drawing a series of topoi whose influence has proven to be vital in both artistic and literary history. We understand, therefore, how the scholies reveal the very essence of the anthological structure, suggesting to us that this text - more than a work - must be understood as a literary imaginary in its own right.

The anthological imaginary

Our main thesis is that the Anthology should not be considered as a “work” but as a dynamic and open collective imaginary. This hypothesis, at the base of our editorial work, draws on four considerations:

1. Historically (and genetically), the anthology is first and foremost the result of a massive and transecular contribution
2. In terms of content, it is characterized by a fundamental heterogeneity (in formal and thematic terms, in particular)
3. It is this heterogeneity, which is counterbalanced by a dialogical dimension that draws reading paths together at the structural level.
4. On the anthological plane itself, its vocation is to refuse any closure, presupposing its own enrichment

By examining the history of the constitution of Anthology, we could see that this text appears much less as a work (or a unit), which would have been transmitted through time, than as the product of a stratification of texts. More than a hundred poets during more than sixteen centuries (Cameron 1993) have contributed to create it. In fact, in terms of content, the Anthology is particularly heterogeneous, testifying to the personal preferences and subjectivity of each contributor. Alongside the few “real” inscriptions (to take up here the origin of the term “epigram”) appearing in the Anthology, there are innumerable epigrams created from scratch to integrate into the Anthology, which is therefore transformed in literary play. On the formal plane, the elegiac couplet (privileged form of the epigram) coexists with other metrical forms (iambic trimmers, dactyl hexameters, etc.). The range of themes is spread from erotic epigrams to Christian epigrams, testifying to societal tastes and issues that follow one another.

The texts manage to answer each other over time due to the fact that the dialogue that is woven between them counterbalances their heterogeneity structurally (K. J. Gutzwiller 1998). Thanks to the scholies in particular, we observe the highlighting of topoi that not only cross the manuscript, but have also had a strong resonance in literature, from medieval troubadour songs to our contemporary songs and in Hollywood movies. Consider, for example, the topos of the carp diem, very present in Hellenistic epigrams like those of Asclepiades (see AP 5.85), later taken up in the Odes of Horace, and whose traces are also found in the poetry of Ronsard (Sonnets for Hélène) and Laurent de Médicis (“The Triumph of Bacchus and Ariadne”), or even in the songs of Brassens (“Saturn”). Thus, the anthological nature in essence refuses any closure and on the contrary, encourages a continual enrichment of the text and its references. Even today, the Anthology remains an open text that refers, internally, to other pieces present in the manuscript but also echoes external texts potentially related to the material of the Codex Palatinus.

The material of the Anthology still inhabits our literary topoi far beyond the explicit references made to its texts. To publish then means to try and account for this thousand-year-old circulation, the diversity of the material involved as well as the richness and heterogeneity of the references between the texts.

But how do you account for this richness and complexity? The traditional methodologies and experiences of scholarly publishing seem not to be appropriate for these objectives. Indeed, while it is quite possible to contextualize the text of the Palatine manuscript and to make a critical edition of it, it is very difficult to give an account of the overall cultural value of this material and in particular of the impact it has had on the collective imaginary over the centuries. In other words, we could of course make a critical edition of the Codex Palatinus, but it is impossible to do that of the Greek Anthology because this work is overflowing with many different transmission contexts, making it impossible to limit.

Editorialization: a conceptual solution to editing the Anthology

How does one edit the anthological imaginary? In order to answer this question, it seemed only appropriate to refer to the theory of editorialization. This theory proved to be useful in understanding how the dynamics of production and the circulation of contents, in a digital environment, help us interpret and propose an edition of an object, such as the Anthology. Editorialization then became a founding principle of our editorial approach.

The term “editorialisation” (translated as editorialization) for nearly a decade now, has come to be used more frequently by the French-speaking scientific community and refers to the production and circulation of content in digital environments(Vitali-Rosati 2018). However, it is sometimes difficult to grasp the exact meaning that researchers attribute to this concept. To give a precise definition, we can say that:

Editorialization is the set of dynamics that produce and structure digital space. These dynamics can be understood as the interactions of individual and collective actions within a particular digital environment.

In other words, editorialization is an open and collective process, which produces contents and visions of the world that organize our relationship to reality.

In this sense, the theory of editorialization can help us understand the essence of the Greek Anthology. Indeed, this textual material is an open and collective process, which has no strong unity, and yet has a structuring effect on our social imaginary. The set-up of editorialization mechanisms found in the Codex Palatinus is one possible way to underline the anthological culture (Crane, Seales, and Terras 2009). The following three aspects of the concept of editorialization have proven to be the most relevant in fulfilling our goal:

1. Editorialization is not a closed process but an open and dynamic one, where a content that circulates in digital environments can always be used, modified and reused, to fulfill another function. This is exactly what characterizes the process of constitution in the Anthology, as we have analyzed it.

2. Editorialization is a collective process, as is the Anthology, and because neither has a single author, what results is textual material made up of a multiplicity of voices and hands.

3. Editorialization is a process that cannot be controlled because it frees itself from its own instigators and goes beyond the intentions that led to its creation in the first place.

It is on these theoretical foundations that we have built the collaborative publishing project of the Anthology. We have specifically designed a digital environment in which scholarly publishing is conceived as a form of editorialization: an open process, based on the interaction between different teams of researchers and other contributors (a contributive model) and which invites each participant to appropriate the textual material and modulate it or use it according to his or her own needs and wishes - far beyond the scientific requirements of a scholarly edition.

Methodology and technical aspects

From a methodological and technical point of view, we decided to set up a digital infrastructure that best reflects the theoretical hypotheses previously described. The desire to leave the canonical paradigm of scholarly publishing led us to make three strategic choices:

1. Create an open database queryable via api
2. Allow a free structure of data, to avoid imposing a predetermined epistemological approach
3. Animate communities to initiate the collaborative process

The choice of an open database, searchable by an API, plays with the very nature of the anthological model, which ultimately relies on a principle of aggregation of fragments selected by an author or a compiler. In the same way, we chose to structure data in JSON rather than XML / epidoc. This decision may seem to break the rules of the digital critical edition, but according to us, it stimulates the appropriation and reconfiguration of the contents, in harmony with the spirit of the anthology. The flat structure of the JSON, unlike the tree structure of the XML / epidoc, allows the contributor/editor to reorganize data according to different paradigms of interpretation.

Finally, in order to stimulate exchanges and interactivity between the different contributors - in the best tradition of the dialogue that has been woven, through the centuries, between the various poets of the Anthology - we have created a digital editorial environment based on a contributive logic, in order to establish a real community of contributors. In particular, we have launched a vast educational project, in partnership with secondary schools where Ancient Greek is heavily taught. Supervised by their teachers, several groups of pupils translate the text and publish it online. This project has many educational advantages: - language learning - text analysis - learning digital publishing tools.

The work produced by a novice community then facilitates the task of the scientific community by supplying usable contents for research. In addition, this choice is based on a strong epistemological positioning, since it erases the boundaries between the work of the scientific community and the contribution of novices. This part of the project is totally in resonance with the theoretical and methodological positioning recommended in the Digital Classics(Melissa Terras 2010), (Christopher Blackwell, and Thomas Martin 2009). As we said above, the very idea of “truth” of the text - that is usually based on a critical or genealogical approach - is not, in the case of our editorial project, an objective to be achieved. On the contrary, we seek to bring out the pluralities of perception of the textual material - because it is this plurality, at the origin of the collective imagination woven around the Palatine Anthology, which constitutes, according to us, the essence of this text.

Thanks to its multiple tables, our database should allow us to pursue and complete this philosophy specific to the anthological genre, highlighting not only the intertextual connections, but also the networks of references built through the centuries, between the anthological texts and any other cultural or artistic object (Coffee et al. 2012). On our contributive platform, we will find different translations (a literal translation that is faithful to the word-by-word, a literary translation allowing itself to be more poetic). The highlighting of translations, present in several languages in our database (mainly English, French and Italian for the moment), allows us to emphasize the importance we attach to forms of appropriation of textual material. What constitutes the characteristic of the Anthology is also its capacity for temporal and spatial circulation: this circulation, guaranteed in the Hellenistic period by the fact that the Greek, as Koiné dialektos, was a universal language, must be ensured today by multilingualism that allows a comparable degree of universality.

Description of the editorial infrastructure

Our infrastructure can be described according to three aspects:

1. The relational database;
2. The API;
3. The different displays made from the API.

1.The database is built from the notion of entity - in a repository aligned to the Perseus URIs - which constitutes an ideal textual unit. Each entity corresponds to an epigram, understood as an anthological fragment and not as a textual manifestation attached to a specific manuscript or edition. This entity then takes shape in various textual manifestations. In our database, we can therefore associate each entity with several pieces of information and, first and foremost, with text versions of the epigram. Our epistemological model led us to consider on the same level the different versions of the Greek text and the set of possible translations. All these versions can be aligned with each other, thanks to a software editor, which allows structuring the correspondences between the Greek texts and the different translations. Each entity can be attributed to one or several authors, and can be tagged by keywords, which include information on the themes of epigrams, on literary genres, on the characters mentioned, or on the reading path established by users. Each entity can also be linked to one or more scholies whose texts are written in Greek or in a translated version. Entities can also be associated with iconographic material or other texts that echo with the corresponding epigram. This possibility allows us to underline what we have called the “weak links” between the text and artistic works, in order to enrich the anthological material. If we go back to the example mentioned above, we can insert, in the database itself, the links between the epigram of Asclepiad - which evokes the topos of Carpe Diem - and the famous Ode of Horace or the poem of Laurent de Médicis. Finally, each entity can be aligned with the facsimile of the corresponding manuscript.

2.The database is exposed in JSON through an API, which makes it possible to carry out queries to display the data, as well as to enrich the database - since the API is open for reading and writing.

3.The API obviously allows several forms of appropriation of textual material. As an example, in the case of our project, we propose a “scientific display,” which makes it possible to visualize all the information available on the base, a public display, which offers a navigation by thematic reading paths (these paths being thought precisely to narrativize a group of epigrams), and finally, a tweet bot which post epigrams on the social network, associating them with the corresponding image of the manuscript.

Conclusion and expected benefits

Our project is currently in its second year of development. We were able to appreciate the enthusiasm of the various communities involved in the editorial activity of this project, making this material resonate beyond the academic sphere. The potential for appropriation of the Greek Anthology by contemporary readers is therefore very strong and proves the vitality of this material in contemporary culture. Given its current development, we can identify four expected benefits of the project:

1.Although our project is primarily an experiment in Digital Classics and our publishing model is a prototype, we are convinced of its potential as a reusable model for other projects because it renews the philological approach to embrace the anthological project.

2.By allowing the users of the API a link to the Greek epigrams and any literary or artistic objects that resonate with them, our project leads contemporary readers to appropriate these ancient texts. This not only refreshes their reception of the Greek Anthology but their reception is updated, illustrating the way in which literary topoi cross centuries.

3.By renewing the classic philological approach, we have shifted the boundaries between scholarly practices and those of the novice, both in terms of production and dissemination of editorial content. The boundaries between scholar and novice now must undergo reconsideration.

4.It goes without saying that our project not only contributes, but also promotes the advancement of knowledge, therefore serving the interest of the scientific community. The novice’s appropriation and playful use that our approach allows does not dismiss how our database can also be structured in scientific ways. In particular, our project contributes to the improvement of the Perseus database (our partner) which will collect in xml / epidoc format all the texts, translations and alignments produced, as well as other relevant information, thanks to queries on the API (at the moment, Perseus does not have, for instance, French, English and Italian translations of the epigrams, nor the Greek text of the scholies) (Crane 1992). In addition, the project contributes to the production of semantic data thanks to the alignment of our data with linked data repositories (Matteo Romanello 2009). The API queries Wikidata on the fly to retrieve existing data (for example, information about an author or a city). All the data collected on the platform will then be re-injected on Wikidata via the API.

In accordance with the principles of the Digital Classics and Digital Humanities (Schreibman, Siemens, and Unsworth 2016), we are feeding encyclopedic projects by taking an active part in the pooling of knowledge.