Trismegistos People is a tool dealing with personal names of non-royal individuals living in Egypt in documentary texts between BC 800 and AD 800, including all languages and scripts written on any surface. Not included are pharaohs, emperors, and saints; people attested in texts outside Egypt (with the exception of some Prosopographia Ptolemaica entries) or people living outside Egypt (no consuls!); and names from non-documentary texts (again with the exception of some Prosopographia Ptolemaica entries).

Fig. 1: Schematic overview of TM People's tables

TM People consists of a complex set of prosopographical and onomastic tables. At the heart of the structure is the REF table, which lists attestations of people identified by personal names (currently 553,614 records). The PER table of individuals, which at present includes 379,581 person records, forms the prosopographical component. The onomastic structure consists of three tiers, dealing with names (NAM), name variants (NAMVAR), and declined name variants (NAMVARCASE) respectively. The Nam table currently has 39,847 names. Each of these standard names is connected to a set of variants, often in different languages / scripts, in the NamVar table (239,068 variants). For each of these variants, declined forms were created in the NamVarCase table. This last table is the largest, with 1,032,681 entries, and forms the link between REF and NAMVAR & NAM.

Each of these tables has its limitations, set out below in the following sections. REF does not yet cover all Trismegistos Texts and has not yet been checked for completeness or mistakes, while prosopographical identifications in PER are a never-ending and often difficult exercise (read more...); the names in NAM could and often should be grouped otherwise, while some adaptations to the Latin name system may still be needed (read more...); and the transliterated Egyptian variants in NAMVAR, as well as Greek accentuation (in all tables), are not always standardized (read more...).

We hope, however, that even in its current state the tool may prove useful enough to avert nemesis. Also, digital instruments such as TM People have the advantage that they can be updated and improved easily. We would therefore be very grateful if users not only show clemency, but also help us improve the quality: suggestions and mistakes can be reported by clicking on 'Report an error' in the header above.

Online databases tend not to be quoted, or only reluctantly. Often scholars will not document the use of digital tools and point to the (printed version of the) sources directly. Gradually, however, scholarship seems to enter a new phase where online edition is taking over the front position from paper copy. For this purpose, we have developed stable numeric identifiers for each entry in each of the TM People tables. For more information, please consult the 'How to cite' section below.

Trismegistos People now counts 553,614 attestations of people. Most of the Ptolemaic data was incorporated from the digital version of the Prosopographia Ptolemaica. This collection served as an important stepping stone for the 'Creating identities in Graeco_Roman Egypt' [CIGRE] project (2008-2012), during which technical advancements allowed us to collect data on a much larger scale, advancing into the Roman, late-Roman, and Byzantine periods. TM People is far from exhaustive, however, and the number of records will rise further, perhaps to 600,000 or more, since not all texts from Egypt in Trismegistos have currently been checked for personal names. The coverage varies per language /script, as does the method used.

Greek

Papyrological texts Greek papyri and ostraca have been 'parsed' semi-automatically using Named Entity Recognition [NER] on the basis of a Unicode XML version generously provided by the Duke Databank of Documentary Papyri ('DDbDP', now incorporated in Papyri.info) in October 2008. The precise procedure has been described in Journal of Juristic Papyrology 39 (2009), pp. 31-47. As a result of this cooperation we have been able to process a substantial portion of the roughly 45,000 texts in a few years with only few collaborators. The records are detailed (including information about the case etc.) but many have not yet been checked, and there are no doubt still many mistakes, some caused by human error, others by technical problems (e.g. wrong cases, places identified as people and vice-versa).

Epigraphic texts The names in Greek inscriptions have been entered manually, which is much more time-consuming. Also personal names have been added in the nominative and no information about the case attested in the text is currently included. Names in amphora stamps have been excluded, because they were stamped outside Egypt.

In 2019, Trismegistos and the Lexicon of Greek Personal Names [LGPN] at Oxford partnered up to start on a much-needed update of TM People. During the first phase, from September 2019 until August 2022, the team will focus on attestations in texts written in the Fayum and Lower Egypt. All existing REFs will be checked for mistakes and new readings, and new REFs will be created for new editions. For this latter purpose, a new round of NER was performed in the fall of 2019 on texts that had been added to Papyri.info between October 2008 and September 2016. In the fall of 2021, a second extraction will be performed from texts that were added after September 2016. The second phase of the collaboration, which is due to run from September 2022 until August 2025, will tackle the material from Upper Egypt.

Demotic

For Demotic papyrological and epigraphic material manual entry started at the end of the project 'Multilingualism and Multiculturalism in Graeco-Roman Egypt' ('MaMIGRE' from 2004-2008, see the general 'history of TM' page), and was continued in the course of CIGRE. This is the only set which has been partially checked. Much work remains to be done, however.

Coptic

Papyrological texts The Coptic papyrological texts have been added with the help of Alain Delattre, who worked on our project as a fellow for one year (March 2009 - March 2010). His Brussels Coptic Database included a field listing transliterated names for each Coptic text. We developed a way to integrate this information onomastically and to check how many times each name was attested in the text. It was not yet possible to include line numbers for the attestations: the first reference of each name is called occ1, the second occ2 etc. Since these are almost always short texts, we have assumed that identical names in the same text refer to identical people. The longer texts which also mention patronymics have not yet been added.

Epigraphic texts Coptic epigraphic texts have not yet been included.

Hieroglyphic and hieratic

Again all manual entries, partially done at the end of MaMIGRE and some during CIGRE.

Latin

Papyrological texts Latin papyrological material from Egypt – normally included in Duke – remains to be done: even the onomastic set used in the parsing procedure is now only in the early stages of development.

Epigraphic texts Latin inscriptions have been done manually, but since the onomastic set has not yet been developed, the entries may look peculiar.

Other languages and scripts

So far nothing has been added for any other language / script, although we have done some preliminary work for Arabic in the NAM table.

Prosopography

TM People is currently only partially a prosopography. For monolingual texts prosopographical identifications should in principle have been made within a single text, but even there we may have missed some, especially for very long texts. For multilingual texts very often prosopographical identifications still have to be added within a single text, even after a first check, because the name attestations have been added by different collaborators at different times.

This does not mean that there are currently no intertextual prosopographical identifications: since TM People builds on the Prosopographia Ptolemaica, most of the identifications made there have already been incorporated, and the Zenon archive has been done as well. All individuals with a double name, as well as their direct kin, have furthermore been processed during CIGRE. The Center for the Tebtunis Papyri (CTP) in Berkeley has started to build a prosopography for Tebtunis on the basis of Trismegistos People records, and we are currently also collaborating with the Aphrodito project (Paris), which is setting up an online guide to the Aphrodito papyri, including a prosopography. During the past years, much work has also been done by Willy Clarysse and Yanne Broux on specific officials (e.g. the praefecti Aegypti, στραγηγοί, βασιλικοὶ γραμματεῖς, various procurators, ...), and individuals bearing specific names (e.g. Sokrates, names of Hellenistic queens, ...). Protagonists of ancient archives are dealt with by Katelijn Vandorpe and her TM Archives collaborators.

We will continue to carry out as many prosopographical identifications as possible. But also we are eager to cooperate with people who aim to develop prosopographies for certain areas or periods, as we currently do with Berkeley and the Aphrodito project, as well as many individual researchers who send us prosopographical suggestions (for which many, many thanks!). Please contact us if you are interested in collaborating.

Coverage estimates

As described above, extracting the personal names from TM Texts is a work in progress, and because of the varying methods, we are at different stages for each language and even for each text. We distinguish the following phases:

• phase 0: not yet done, personal names still to be added
• phase 1: some personal names added, but possibly incomplete
• phase 2: all personal names supposedly added, to be checked
• phase 3: all personal names added, first check done
• phase 4: intertextual prosopographical identifications investigated

Below is a table with estimates of the progress made for various languages / scripts (up until July 2021):

	Phase 0	Phase 1	Phase 2	Phase 3
Greek papyrology	30%	1%	56%	13%
Demotic (Pap. & Epigr.)	43.3%	7.5%	28%	21.2%
Greek Epigraphy	33%	6%	60%	1%
Coptic Papyrology	64%	–	36%*	–
Latin Papyrology	81%	9%	9%	1%
Latin epigraphy	67%	6%	27%	–
Coptic epigraphy	92.9%	0.1%	7%	–
Hierog. & hierat. papyrology	88%	8.8%	3%	0.2%
Hierog. & hierat. epigraphy	57.6%	35.3%	7%	0.1%
Aramaic (pap. & epigr.)	100%	–	–	–
Arabic papyrology	100%	–	–	–

*TM Texts currently contains many unpublished Coptic ostraca from the British Museum and other collections. These have obviously not all been checked for names. When excluding these 'ined.', the amount of texts in phase 2 rises to 56%

Amongst the inaccuracies and mistakes to be corrected in phase 3 are incorrectly recognized cases or incorrect accentuation for Greek, or unstandardized transliterations for Egyptian. The addition and correction of column and line numbers is a desideratum for all references, while the one-to-many and many-to-one relations between documents in Trismegistos and Papyri.info have also caused names to be attributed to the wrong texts. Please also note that in processing the DDbDP material, for technical reasons we were unable to add indications of abbreviations ‘( )’ , damaged names ‘[ ]’ or other diacritic signs. We are in the process of adding these "raw" forms, but please double check with Papyri.info or the Packard Humanities Greek Inscription website whether the form is reconstructed or not.

One of the difficult problems concerning names is grouping variants of a name: when does a variant become a name of its own? In current onomastic systems, there is a tendency to a different spelling as a different name. People called 'Yanne' often do not feel their name is the same as 'Janne', and certainly not the same as 'Johanna', from which it is originally derived. Still these names are etymologically related or even identical. A 'separatist' approach would create 3 NAM entries for these three names, an 'integrist' approach would group these three together under a single NAM entry.

In our database, we have opted for the latter, grouping variants, mainly because of the problems posed by the separatist approach when dealing with different languages. Pushing the separatist principle to the extreme, Greek Ἀπολλώνιος would be a different name from its Demotic transliteration ȝpwlnys. Also, because of the existence of a separate table of name variants, users can still treat each variant as a separate name should they prefer to do so.

But even within each language and script we have permitted some variation within a single name. For Egyptian names (i.e. names with an Egyptian etymology), we have grouped together variants with or without genitive '-n' or feminine and plural endings '.t' and '.w', as well as variants with and without an article inside the name, e.g. Ḥr-Ỉkš and Ḥr-pȝ-Ỉkš. For Greek names spelling mistakes such as Πτολλεμαῖος (for Πτολεμαῖος) or dialectal variants such as Ἀπολλωνίδας (for Ἀπολλωνίδης) do not warrant the creation of a separate name.

Perhaps less obvious is the choice for separation in the following cases: for Egyptian the addition of the article at the beginning of a name, e.g. Pȝ-ḥtr against Ḥtr (despite the problems posed by hieroglyphic texts); the possessive pronoun versus the article, e.g. Pa-hb against Pȝ-hb (although we do not claim consistency here, especially not considering the problems created by the possessive pronoun, in particular for the feminine forms Ta- and Tȝy-); for Greek names a variation in ending is enough to consider it a separate name, e.g. Ἀχιλλεύς against Ἀχιλλᾶς. Hypocoristica are also considered separate names, even though sometimes full name and hypocoristic are used side by side for the same person in the same text or archive.

The Latin name system with its praenomen, nomen gentilicium and cognomen or cognomina presented a supplementary methodological complication for the semi-automated processing of material from Duke (currently Papyri.info). We have developed a customized solution for this, which treats the praenomen and nomen gentilicium as ‘pseudo names’ or ‘name prefixes’, unless nothing else follows: Marcus Aurelius Sarapammon is thus listed among those having the name Sarapammon.

Similar problems arise with Coptic Apa and Abba, for which the 'pseudo' field is also used. For Arabic we have identified the problem with polyonymy, but since it currently falls outside our scope we have not yet taken action to deal with it.

Ghostbuster, Willy Clarysse's database of so-called "ghostnames", is currently still a separate feature in TM. It will be integrated into TM People in the future. For now, you can still consult it here.

Despite the efforts to remedy it, there are unfortunately still several systems to transliterate Egyptian. Even for Demotic, although there is agreement on the historic transliteration system, there is still variation in minor points, which is confusing for outsiders (e.g. d or t, j or y or i, ...). For hieroglyphs and hieratic often different standards are in use. Since most of our material is Demotic, however, we have chosen to take that as a standard and adapt hieroglyphic and hieratic transliterations accordingly.

In principle, we have transliterated group writings 'historically', even when the phonetic realizations no longer distinguished between voiced and unvoiced. We add a full stop '.' to separate gender and number endings from the root, and an equals sign '=' for suffix pronouns. The lexical components of a name are separated by a hyphen '-'. Names of deities and places are capitalized, as are the personal names themselves. We thus write:

• Pȝ-šr-n-nȝ-ḫty.w
• Pȝ-dỉ-Ḫnsw-pȝ-ỉ.ỉr-sḫy
• Tȝ-šr.t-n-ʿšȝ-ỉḫy
• Dnỉ.t-n-Ḥr
• Ḏd-Ḫnsw-ỉw=f-ʿnḫ

We have tried to be consistent, but since this is work under construction and new names are being entered continuously (often by students, as they are transliterated in the edition!), there is no doubt room for improvement. Please note that we currently do not attempt to distinguish variants with genitive '-n' or feminine ending '.t' from those without.

In Greek texts accents are almost never written before the 2nd century AD, and even after that only rarely (in some literary texts). Modern editors just add them in their editions. For Greek words, there are fixed rules which are well-known, although not every Greek papyrologist applies them flawlessly. For Egyptian words, and names in particular, matters are far less clear. Willy Clarysse argues for accentuating according to the Greek declensions, and this is also accepted for Latin names by Johannes Kramer. In this view, no accents should be put on non-declined names (e.g. the month names). Most papyrological editions continue, however, to use the traditional accentuation based not only on Preisigke’s Namenbuch, but also on our Latin pronounciation, e.g. -οσῖρις and Ἀμᾶσις instead of -όσιρις and Ἄμασις, as Clarysse would have it.

Because of the uncertainty we have generally omitted Greek accents from names of Egyptian origin. This rule may not always have been followed, and sporadically Greek accents may still appear on names, especially those that have been processed automatically from Duke through Named Entity Recognition (see above). In no case should the variants attested in TM People be seen as models for those wanting to accentuate ‘à la Clarysse’.

There is another complication with accents. The rules for their placement are complicated in that they take vowel length into account, and accents thus shift according to case endings. When building up our table of declined forms of name variants (NAMVARCASE), we started from an attested form, then created the nominative and the other cases. For this we never took the accents into account. As a result, many declined forms may not have the accent in the right place. The automated recognition process, however, made abstraction of the accents and kept the accentuation attested in the edition. The accentuation displayed in the REF records (attestations of names / individuals) is therefore that of the (online) edition. In the NAMVARCASE-table, on the other hand, some very strange and incorrect forms may appear. We hope to correct this in the future.

For a detailed description of how the onomastic (Nam) and co-occurence (Per) networks displayed on the respective detail pages work, please visit the TM People networks page.

For a video tutorial made during the TM People workshop held on September 2 2021, please visit the TM Academy page.