What are book epigrams?

1. Orientation

Byzantine manuscripts have been one of the most important bridges through which the inheritance of Classical Antiquity has enriched our modern culture. Many of these medieval books however also contain poetical gems that have received far less attention. These poems are tightly connected to the contemporary Byzantine milieu in which the manuscripts originated. They are now known as "book epigrams", a textual genre that is as yet difficult to survey and consequently not fully explored.

The number of Byzantine manuscripts runs into the tens of thousands. One can roughly estimate that some ten per cent of them contain one or more book epigrams.

2. Definition(s)

The term itself is a recent coinage, but the genre has been known from the time Byzantine manuscripts were first investigated. Athanasios Kominis defined them as “epigrams in books and on books” (Kominis 1966:38), and this sums up well the most important characteristic of these texts. They are found in books, but at the same time they have these very books as their subject. In this respect, they can be compared to inscriptions on graves, religious images or buildings: they comment on the very object that displays them.

Marc Lauxtermann, who devoted a whole chapter of his book on Byzantine poetry to book epigrams, has defined them as “verses that are intimately related to the production of literary texts and manuscripts” (2003:197). In other words, they are the residues of the assiduous work of copying and transcribing; they are the texts in which the singularity of each manuscript becomes visible and in which the contemporary world of scribe, patron, and first readers comes to life.

Book epigrams can alternatively be defined as "metrical paratexts". Being true para-texts, they stand next to "main texts", helping the reader in his or her orientation as to what this main text means, and how the medium of the text (the book) came into being.

As a result, epigrams meant to be inscribed on other objects, cannot be considered as book epigrams proper. Nor can poems that stand on the same level as the main texts (for example, a poem on the Annunciation next to a hymn or oration on the Feast). Some poems can function as book epigrams in one manuscript, while just being part of an anthology in another. Our database attempts to be inclusive, and also contains paratextual epigrams that do not correspond to a strict definition of book epigrams (e.g. prayers).

3. Classification

The following classification can be made, based on the main actors that play a role in the communicative situation typical for book epigrams.


These epigrams concern the practices of writing and copying: the scribe announces that he has completed this book, in this or that monastery, at this or that date; he adds that he is glad that the work is done and attributes all glory to God, or thanks Him in a prayer. This category of poems is often the most formulaic. They are often called "poetic colophons", or "subscriptions".


These epigrams give information about the identity and motivations of the patron, who has "produced" this book, i.e. commissioned it and paid for it. In other words, these are the epigrams that inform us about the patronage of the object. As a rule, religious devotion, and / or the wish to repent for sins, are mentioned as the principal motives behind this patronage. These epigrams are perfectly comparable to similar inscriptions on objects or buildings. They can be formulaic, but often, they are new products specially conceived for the particular occasion. They are also called "dedicatory" epigrams.


Many book epigrams address the reader of the book. They prescribe certain reading habits and strategies or anticipate an expected response from the reader. Often, book epigrams emphasize the edifying qualities of the main text, recommending it for the spiritual well-being of the reader. Needless to say, in doing so, they offer us precious information about ideas and discourses on reading.


Many epigrams praise the author of the main text, and / or the merits of that text itself. The author is often an Evangelist or a Church Father, so that praise for the author often amounts to religious celebration. These epigrams are also called "laudatory" - although there are some that criticize (mostly pagan or heterodox) authors as well.


Many manuscripts include monostichs functioning as titles or as headings of table of contents, epigrams announcing the end of a work, short summaries, etc. They are extremely common, often formulaic, and in many cases, especially of single verses, scribes may not have been aware that they were actually copying verses.


Sometimes, book epigrams interact with miniatures in the manuscript, describing their features or explaining their meaning.

4. Authorship

For the most part, it is a futile undertaking to try to retrieve the names of the "authors" of book epigrams. The origin of many epigrams seems to lie in the anonymous "dark centuries". A scribe often borrows formulaic expressions from earlier models, sometimes only changing the name. The few book epigrams that are known as the work of a Byzantine poet, however, prove to be very interesting and tell us more about the circulation of "mainstream" poetry.

5. Further reading

  • F. Bernard and K. Demoen, “Book Epigrams”, in: A. Rhoby, N. Zagklas and W. Hörandner (eds.), Brill Companion to Byzantine Poetry (forthcoming)
  • F. Bernard, Writing and Reading Byzantine Secular Poetry, 1025-1081, Oxford 2014
  • D. Bianconi, “Et le livre s'est fait poésie”, in P. Odorico, M. Hinterberger and P. Agapitos (eds.), «Doux remède...» Poésie et poétique à Byzance, Paris 2008, pp. 15-35
  • K. Demoen, “La poésie de la συλλογή. Les paratextes métriques des manuscrits byzantins et le (vocabulaire du) recueil”, in C. Gastgeber, C. Messis, D. Muresan and F. Ronconi (eds.), Pour l'amour de Byzance: hommage à Paolo Odorico, Frankfurt am Main 2013, pp. 89-98
  • A. Kominis, Τὸ βυζαντινὸν ἱερὸν ἐπίγραμμα καὶ οἱ ἐπιγραμματοποιοί, Athens 1966
  • M. Lauxtermann, Byzantine Poetry from Pisides to Geometres. Texts and Contexts, Vienna 2003, pp. 197-212
  • P. Odorico, “Poésies à la marge. Réflexions personnelles?”, in F. Bernard and K. Demoen (eds.), Poetry and its Contexts in Eleventh-Century Byzantium, Farnham / Burlington 2012, pp. 207-224

Methodologies and practices

1. Scope


Our chronological scope falls between the date of the first manuscripts containing book epigrams and the end of the fifteenth century.


We only include metrical paratexts. In cases of doubt (one-line epigrams can often be metrical by mere coincidence) we have chosen to be inclusive.

2. Occurrences and types

We have made a basic distinction between two kinds of textual material, corresponding to two different sets of text records in the database.

  1. ‘Occurrences’ reproduce the text of individual epigrams as faithfully as possible, i.e. as it is found in the manuscripts. Hence, the text is given with all the particular idiosyncrasies of the manuscript in terms of orthography, punctuation etc. Ideally, the text of the occurrences is based on consultation of (reproductions of) manuscripts. In many cases, however, we are still relying on manuscript catalogues as well as other related publications. These do not always apply rigorous standards and very often normalize the manuscript text.
  2. On the other hand, each of these ‘occurrences’ is linked to an overarching text, a ‘type’. These ‘types’ group one or more ‘occurrences’ that have an identical or similar text. This enables the user to browse between similar ‘occurrences’ and also makes it easier to obtain an overview of the text corpus. Ideally, type records offer a normalized, critical text, as well as interpretive data (translation, identification of the subject, comments, etc.).

3. Deficient texts

Many catalogues in fact do not reproduce the text of the epigram in the specific manuscript, but refer to another standard work that prints the epigram, but based on another manuscript. Since in these cases we are not sure about the exact textual appearance of the occurrence, one may find in the occurrence record a text box that only gives an incipit, or displays NA (not available). These blank occurrences are nevertheless linked to known types, so that the user can at least get an idea of the epigram.

When we do not have any clue about the text of the epigram (for example, just a mention ‘epigramma in S. Gregorium’) the abbreviation ‘NA’ appears instead of the incipit.

4. Own ‘editions’

The indication ‘text source’ indicates the publication from which we have adopted the text, exactly as it has been found. If the text source indicates ‘DBBE’, this means that our team had access to (reproductions of) manuscripts and has established the text. Transcriptions have been made as faithfully as possible, according to the standards of so-called diplomatic editions. We plan to add more texts based on our own consultation of further manuscripts.

5. Codicological descriptions

The user should not expect that our database includes all codicological information. Our focus is on the paratexts. The information on manuscripts that we offer now, is given in order to better understand the manuscript context of these paratexts. We do not identify scribal hands ourselves, neither do we date manuscripts ourselves, nor do we provide a detailed description of all the contents of the manuscript. The Pinakes database of the IRHT in Paris is the point of departure for this kind of detailed codicological information. As a member of the Diktyon network, we provide links to Pinakes in our manuscript records.

Functionalities of the database

Users can search or browse through either Manuscripts, Occurrences, Types or Bibliography tabs. These lead to search results, and from there to more detailed separate records, which form the backbone of our database.

1. Manuscripts

Manuscript identifications

(city, library, collection, shelfmark nr) basically follow the conventions applied by the Pinakes database, with some minor exceptions (e.g. DBBE: Athos, Vaticano vs Pinakes: Hagion Oros, Città del Vaticano).


Manuscripts are roughly divided into categories, based on the TLG’s generic epithets (with some modifications). These are meant to give a very general idea, and are based on sheer numbers of manuscripts; therefore, Gregory of Nazianzus is a separate author, and Menander is not. We refer to the Pinakes database for a more granular search of manuscripts according to the works they contain.


Exact year, or range, down to a quarter of a century.


Scribes’ names are given with references to M. Vogel-V. Gardthausen, Die griechischen Schreiber des Mittelalters und der Renaissance, (Leipzig, 1909, repr. Hildesheim 1966) (VGH) and E. Gamillscheg-D. Harlfinger et al., Repertorium der griechischen Kopisten (RGK). VGH reference is given as: ‘page number in VGH. Letter’. DBBE established the letter by giving each person a letter in alphabetical order on each page. e.g. 322.B denotes the person Μιχαὴλ ὁ Συναδηνός. RGK references contain the volume number, a ‘dot’ and its unique number, e.g. III.150. We decided to transliterate Greek proper names rather than Anglicising them (so Iohannes, and not John).


Variously indicated in sources as commissioner, dedicatee, sponsor, etc. Since it is very difficult to distinguish between these roles, we have ranged them all under ‘patron’: those who provided the funds and/or took the initiative to produce the manuscript.


Whenever available.

2. Occurrences


The text of the occurrences ideally represents the text of the epigram exactly as it is found in the manuscript (see also methodologies and practices). Transcriptions have been made as faithfully as possible, according to the standards of ‘diplomatic’ editions. The following editorial conventions have been used:

  • (…) indicates that either the text is not readable or our source (e.g. the manuscript catalogue) does not provide the whole text.
  • Abbreviations of nomina sacra are resolved and the missing text is placed within parentheses, e.g. θ(εο)ῦ.
  • The distinction between majuscule and minuscule letters is normalised.
  • The punctuation found in the manuscript is reproduced.
  • Misspelled words and wrong accents or breathings are kept.


If present in the manuscript, the title (sometimes called ‘lemma’) of the epigram is given.

Text source

Acknowledges the source on which our database relied in order to give the text. In cases when more than one source is available, preference is given to the most reliable publication. If the text source is ‘DBBE’, this means that we have established or checked the text ourselves.

Text status

Indicates the certainty with which we can establish the text. If we know the text in its entirety the text is marked as ‘completely known’. For incomplete occurrences (epigrams of which the text is not given completely by catalogues, but only by means of incipit or as a reference to other publications) ‘text status’ is ‘text partially unknown’. In cases where we know that there is an epigram, but we have no access to it at all, text is marked as ‘completely unknown’.

Patron and/or scribe

See above, in manuscripts. Mostly (but not in all cases!) patron and/or scribe of the epigram will be the same as those of the manuscript.


Based on the most recent or authoritative publication. The date of the epigram can of course be different from the date of the manuscript.


Leads to the relevant type. See methodologies.


Our labels refer to the actor/agent that occupies the central role in the communicative situation established in the epigram. Most labels can be related to the genres of book epigrams given by M. Lauxtermann, Byzantine Poetry from Pisides to Geometres, vol. 1 (Vienna 2003), pp. 197-198.

  • scribe: colophons. This category includes also prayers.
  • author: laudatory.
  • patron: dedicatory.
  • reader: gives advice to the reader or emphasizes the value of the book.
  • text: this category mainly comprises book epigrams that structure or summarize the ‘main’ texts in the manuscript, e.g. metrical titles and hypotheseis.
  • image: epigrams that explicitly refer to images (miniatures) in the manuscript.

For the time being each epigram is linked to one genre. In the forseeable future it will be possible for an epigram to be linked to multiple genres.


The main subject of the poem: often a saint, divine person, or the author of the book. Provided to enable thematic search queries.

Place in manuscript

Either a precise folio number, or else a more general indication.

Paleographical information

Indications on script, color of ink, other paleographical features.

Contextual information

In the literal sense of ‘con-text’: surrounding texts, when relevant.

Other comments

Other information that may be relevant.


The user will find here either credits to an image containing the book epigram in a (print) publication, or a web link to an image in an online repository. Our team made use of many more images, which unfortunately we cannot publish online due to copyright restrictions.


(other than text source): ‘Primary’ refers to those publications containing (part of) the text under discussion, and ‘Secondary’ is used for all relevant publications commenting on the occurrence, but not containing Greek text.

ID number:

A unique number for each occurrence. Please use this number in your publications to cite DBBE records of occurrences.

3. Types


We ask users to be cautious when using the texts of the types. In the future, these are expected to evolve into reliable editions based on careful collation of the available sources. In the bèta release of our database, this is not yet the case.


Whenever available, a translation into a modern language is included.


More refined than ‘subject’, rather referring to some recurring motifs. Meant to enable specific thematic queries. Also this feature is provisional, and is meant to be expanded in the future.


This field often includes Vassis’ Latin description of the poems. See I. Vassis, Initia Carminum Byzantinorum (Vassis 2005) and id. Initia Carminum Byzantinorum Supplementum I (Vassis 2011).

List of related occurrences

This enables the user to browse through all occurrences linked to a certain type.

List of related types

This still needs to be completed. Ideally, this includes epigrams to which the present epigram is related, because it combines several existing epigrams, includes other epigrams inside its text, etc.


Types also have unique ID numbers but for reference purposes. Since most of our types are as yet by no means critically edited texts, it is not advisable to cite them at this stage.

4. Bibliography

The user can browse through the bibliography in alphabetical order. By clicking on a single item, one gets the relevant information concerning the bibliographical reference (author, title, kind of publication) and the occurrences and the types which are connected to each publication.

In the comments and in the ‘text source’ and ‘bibliography’ fields of the records, we use a short bibliographical reference, which normally contains name of the author(s) and the year of publication. Full reference is given by clicking on these keys.

How to cite

Each type and each occurrence are assigned with a unique ID. Its form for the types is: plus a 4 digit ID number. So the full link is: The form for the occurrences is: plus a 5 digit ID number. So the full link is:

You can find the ID number in the address bar. The ID numbers are also mentioned on the occurrences and types pages.

Since at this stage the types are provisional, you are encouraged to cite only occurrences from DBBE and not types. The full bibliographic description of the DBBE as a whole is Database of Byzantine Book Epigrams (2015) accessible at; the standard abbreviation is DBBE.
Occurrences should be cited with the date on which they were consulted, and you can use the permalink provided for each entry: so DBBE (consulted 1.xi.2016), <>