1 The Linked Ancient Greek and Latin (LAGL) Project
Linked Ancient Greek and Latin (LAGL) is a project to extract linguistic data from ancient Greek and Latin sources and structure them as Linked Open Data (LOD) (Cayless, 2019; Middle, 2024). At the moment, the focus is on Ancient Greek and two kinds of data: (1) Named Entities (NEs) and (2) linguistic expressions containing bibliographic citations of authors and works.
Due to the limited availability of digital data for these two categories, the project is developing a workflow for the extraction, lemmatization, and token-level annotation of Named Entities. The workflow follows the classification schemes commonly employed in computational linguistics, such as PER and PERderiv for personal names and their derivatives (e.g., Mιλτιάδης ‘Miltiades’ and ’Eπικούρɛιος ‘Epicurean’), LOC and LOCderiv for place names and derivatives (e.g., ’Aλικαρνασσώς ‘Halicarnassus’ and Mυτιληναῖος ‘of Mytilene’), ORG and ORGderiv for organizations and related entities such as festivals and schools (e.g., Παναθήναια ‘Panathenaia’ and ’Aκαδήμɛια ‘Academy’), and OTH for other entities, including currencies, months, and work titles (e.g., ’Oκτόβριος ‘October’ and ’Iστορίαι ‘Histories’) (Jurafsky & Martin, 2025). These broad categories allow us to generate training data that can be used to train computational models for Named Entity Recognition (NER) and extract semi-automatically proper names from ancient sources (Berti, 2023; Berti, 2026).
In the LAGL project, Named Entities are used as linguistic anchors in the text to individuate bibliographic references, given that most of them are represented by proper names, as in the expression Λυκέας (PER) ὁ Nαυκρατίτης (LOCderiv) ἐν τρίτῃ Aἰγυπτιακῶν (LOCderiv/OTH) (Ath. Deipn. 13.10 = “Lyceas of Naucratis in the third book of the History of Egypt”). Forms like this are populating the LAGL Catalog of Authors and Works, which is based on the extraction of linguistic annotations of bibliographic citations found in ancient sources.
Data are stored in an SQL database and are currently accessible through the web resources of the LAGL Catalog. At the end of the project, data for Named Entities and annotations of bibliographic references will be released as open data in CSV, JSON, RDF, and TEI XML formats. Because part of the data-extraction workflow relies on the web-based annotation platform INCEpTION, the data will also be exported in the UIMA CAS JSON format (Klie et al., 2018; Eckart De Castilho et al., 2024).
According to the guidelines of the Journal of Open Humanities Data (JOHD) and the theme of the Special Collection in which this contribution appears (Wikidata Across the Humanities: Datasets, Methodologies, Reuse), this article is a discussion paper that presents the methods and challenges involved in integrating LAGL data with Wikidata statements concerning Classical literature. Its aim is not to describe the LAGL dataset itself, but rather to raise awareness among Classicists, Digital Humanists, and Computational Linguists about both the potential and the limitations of this form of integration.
1.1 The LAGL Catalog
The Catalog of the LAGL project collects and disambiguates linguistic data about authors and works with CTS URNs developed according to the CITE Architecture (Berti et al., 2016; Blackwell & Smith, 2019). The Catalog currently contains annotations of a total of 1,236 Greek and Latin authors extracted from the Deipnosophists of Athenaeus of Naucratis, the Lexicon of the Ten Orators of Valerius Harpocration, and the Byzantine lexicon of the Suda (Berti, 2023; Berti, 2024; Berti, 2025a; Berti, 2025b).
In the example provided in the previous section, Λυκέας ὁ Nαυκρατίτης (Lyceas of Naucratis) and Aἰγυπτιακῶν (History of Egypt) are extracted and identified with urn:cts:greekLit:tlg1469 and urn:cts:greekLit:tlg1469.ath001 respectively.1 The syntax of these CTS URNs contains information on the domain (ancient Greek literature) and a reference number as expressed in the Canon of the Thesaurus Linguae Graecae (TLG), where Lyceas is assigned the four-digit identification number tlg1469. In the CTS URN of Lyceas’ work, the string ath001 refers to the fact that Aἰγυπτιακῶν is the expression used by Athenaeus to cite the work of Lyceas that he was reading or that he found referenced in other sources. The three-digit identifier tlg002 of the TLG Canon is not used because it refers to the entire collection of Lyceas’ fragments in the edition of the Fragmenta Historicorum Graecorum (FHG) (= tlg1469.002).2
Latin authors are identified as belonging to the domain of Latin literature and receive identifiers with the prefix stoa or phi as in the Perseus Catalog and the Digital Latin Library (DLL)’s Catalog (e.g., Aquila Romanus: urn:cts:latinLit:stoa0030b; Gaius Suetonius Tranquillus: urn:cts:latinLit:phi1348).3 Authors who are not yet, or are no longer, accessible in the TLG Canon or the Perseus Catalog are assigned new numbers with the prefix lagl (e.g., Arignote: urn:cts:greekLit:lagl0309).4 A detailed discussion of the implementation of CTS URNs in the LAGL project is available on the website of the LAGL Catalog and in the relevant bibliography.5
1.2 The Wikidata LAGL Author ID Property
CTS URNs of LAGL authors are added to Wikidata under the Wikidata “LAGL Author ID” property P12869 using QuickStatements batches. Of the 1,236 authors currently recorded in the LAGL Catalog, only thirteen still lack a corresponding Wikidata item, due to uncertainties in their identification and limited biobibliographical evidence.6 A Wikidata LAGL Work ID property will be proposed in the future to add CTS URNs of works, to address format and subject type constraints that are different from those of authors. The goal of the LAGL Author ID property is to provide direct access to the linguistic annotations of the LAGL Catalog and to enable their integration into Wikidata once they have been completed and released in Linked Open Data formats, as explained in Section 1.
As part of the collaboration with Wikidata, the LAGL Catalog retrieves Wikidata QIDs of authors and works. The LAGL database stores Wikidata QIDs, making them available with their labels on the webpages of the Catalog, as it is visible in the case of Lyceas (Q11931866) and his History of Egypt (Q130367204).7 LAGL authors and works that are missing in Wikidata are added by creating new items. Experimentally and if available, Wikidata values concerning locations and genres of authors are displayed in the HTML pages of the LAGL Catalog by querying live the Wikidata properties for the “place of birth” (P19), the “place of death” (P20), and the “occupation” (P106). For example, at the time of writing this paper, Lyceas appears as born in “Naucratis” (Q137721) and active as “geographer” (Q901402) and “historian” (Q201788). Even if limited to a selection of a few properties, these experimental visualizations allow for assessing the current coverage of Wikidata statements about Classical authors and integrating them with values derived from annotated ancient sources.
The research currently conducted within the LAGL project demonstrates the potential of Wikidata properties for expressing structured statements about Classical authors and works, covering aspects such as literary genres, geographic distribution, chronology, and many others. These statements are contributing to the creation of a controlled vocabulary composed of property-value pairs capable of representing the richness and diversity of assertions concerning Greek and Latin authors as found in ancient sources, modern scholarship, and catalogs of national and international library systems. Given that this kind of research practice is gaining increasing attention but has yet to be fully integrated into the field of Classical Studies, the following section presents current LAGL experiments aimed at adding and visualizing Wikidata statements on the geography and literary genres of ancient Greek authors based on evidence from ancient sources.8
2 Wikidata Statements on Classical Authors
In this section, I examine two categories of assertions concerning Classical authors found in ancient sources: (1) those that describe geographical origin and movement, as reconstructed from parental background and biographical tradition, and (2) those that concern literary affiliation and intellectual pursuits. Ancient texts articulate these assertions in diverse ways, but their rich and complex language is often condensed in modern print editions, indices, and catalogs, where generic adjectives or other concise expressions, used mainly for disambiguation, tend to flatten more nuanced statements.9
In the digital environment, a clear example is provided by the “epithets” of the Canon of the TLG, which serve to disambiguate author names and to summarize the author’s place of floruit or principal place of residence (e.g., Syracusanus), as well as the kind of literary activity or genre in which the author engaged (e.g., historicus) (Pantelia, 2022, xxiii–xxix). In Wikidata, a similar loss is inevitable when such assertions are translated and converted into structured data. Even so, Wikidata makes it possible to link statements directly to their formulations in the original language of the context from which they derive, and to employ a richer, more flexible vocabulary. The main challenge lies in the fact that most Wikidata properties and values are designed for modern figures and concepts. This raises questions about how such models can be adapted for ancient authors, and how ambiguity, uncertainty, and conflicting evidence might be represented within a structured framework.10 The following subsections do not aim to resolve these challenges but to explore them through a series of experiments on representing Classical authors in Wikidata, using SPARQL queries and visualizations on maps and within database environments.
2.1 Wikidata Location-related Properties for Classical Authors
Ancient sources preserve assertions about the geographic provenance of Classical authors which vary from the place of birth and death to the place of intellectual activity. These assertions may include ethnics (e.g., Διογένης Σινωπɛύς in Diogenes Laertius 6.20 and Suda δ 1143 for Diogenes of Sinope: Wikidata Q59180), prepositional phrases with an explicit reference to the relevant place (e.g., Λουκιανὸς ἐκ Σαμοσατῶν τῆς Συρίας in Suda λ 685 for Lucian the Martyr from Samosata in Syria: Wikidata Q510737), or also be part of extended passages (e.g., Ξɛνοκράτους μὲν ἐν ’Aκαδημɛίᾳ, ’Aριστοτέλους δ᾿ ἐν Xαλκίδι διατρίβοντος in Diogenes Laertius 10.1 for the activities of Xenocrates in the Academy and Aristotle in Chalcis: Wikidata Q214121 and Q868).
Wikidata offers many location-related properties for human beings. With the assistance of ChatGPT 5, I was able to individuate many of them with a focus on authors, scholars, and intellectuals.11 A selection of these properties is listed with their Wikidata descriptions in Table 1 together with other less common properties in Table 2. Wikidata also makes it possible to add statements for institution-related properties (with geographic relevance) that connect a person to institutions with defined locations, such as universities, employers, or academies. These can be linked through the institution’s property P131 or its geographic coordinates to derive spatial information.
Table 1
Wikidata location/place properties for authors, scholars, and intellectuals.
| PROPERTY ID | LABEL | DESCRIPTION (WIKIDATA) |
|---|---|---|
| P19 | place of birth | most specific known birth location of a person, animal or fictional character |
| P20 | place of death | most specific known (e.g. city instead of country, or hospital instead of city) death location of a person, animal or fictional character |
| P27 | country of citizenship | the object is a country that recognizes the subject as its citizen |
| P551 | residence | the place where the person is or has been, resident |
| P937 | work location | location where persons or organisations were actively participating in employment, business or other work |
| P263 | official residence | the residence at which heads of government and other senior figures officially reside |
| P276 | location | location of the object, structure or event |
| P131 | located in the administrative territorial entity | the item is located on the territory of the following administrative entity |
Table 2
Additional Wikidata location/place properties for authors, scholars, and intellectuals.
| PROPERTY ID | LABEL | DESCRIPTION (WIKIDATA) |
|---|---|---|
| P119 | place of burial | location of grave, resting place, place of ash-scattering, etc. (e.g., town/city or cemetery) for a person or animal |
| P66 | ancestral home | place of origin for ancestors of subject |
| P2842 | place of marriage | location where the marriage was celebrated |
| P2632 | place of detention | place where this person is or was detained |
| P2715 | elected in | election in which a person gained a position (as qualifier), or a position or legislative term is/was elected (as main value) |
Considering that this part of the LAGL project is still highly experimental and that there are location-related properties more suitable for modern and contemporary people, I have reduced the list to the following eight properties for extracting Wikidata statements about locations that primary and secondary sources connect to Classical authors: “place of birth” (P19), “place of death” (P20), “country of citizenship” (P27), “ancestral home” (P66), “place of burial” (P119), “residence” (P551), “work location” (P937), and “location of formation” (P740) that I added even if primarily intended for organizations and indirectly relevant in studies of authors and intellectuals.
In order to get an overview of the use of these properties and monitor changes in Wikidata, I created a SPARQL query to extract all Wikidata items with a LAGL Author ID (P12869) and at least one of the associated location properties that have geographic coordinates (P625). Locations are filtered according to the eight properties listed in the previous paragraph. The resulting Wikidata map is accessible in the main page of the LAGL Catalog under LAGL Authors Wikidata live map.12 When running, the query shows on OpenStreetMap the LAGL authors who are present in that moment in Wikidata with at least one of the eight location properties with coordinates for the places of birth, death and burial, the country of citizenship, the ancestral home, the residence, the location of formation, and the work location.13 Figure 1 shows a screenshot of the entire dynamic Wikidata map resulting from the SPARQL query. Figure 2 shows a detail of the map with markers of authors currently grouped in “Classical Athens” (Q844930) with a focus on the ancient Greek comedy writer “Hipparchus” (Q11925670), who is in the LAGL Catalog with the identifier urn:cts:greekLit:tlg0468.
Figure 1
Wikidata live map of LAGL Authors.
Figure 2
Wikidata live map of LAGL Authors (“Classical Athens” Q844930).
Because the live Wikidata map locates LAGL authors but does not expose the underlying location properties — and a single, more complex Wikidata query would time out — I built a lightweight experimental Python workflow that reuses the query described above and outputs two self-contained HTML views: (1) a Leaflet map with one marker per place whose popups list LAGL authors and their properties at that place, and (2) a searchable database with columns for Wikidata Author and QID, LAGL CTS URN sourced from the Wikidata property P12869, Wikidata Property and Property ID, and Wikidata Place and Place QID. CTS URNs link to the LAGL Catalog, QIDs and Properties link to Wikidata, and the two views are cross-linked, so that it is possible to access the database from the map and vice versa. The database supports client-side search, sorting, and filtering by author, place, property, and CTS URN.
The resulting LAGL Authors map with Wikidata location properties and its database are accessible in the LAGL Catalog homepage and regularly updated to reflect changes in Wikidata.14 Of the 1,236 authors currently recorded in the LAGL Catalog, 772 are provided with at least one Wikidata location value including geographic coordinates, and are therefore visible in both the map and the database, corresponding to a total of 360 distinct places. Figures 3 and 4 show screenshots of the map for LAGL authors located in “Classical Athens” (Q844930) and “Halicarnassus” (Q5843680).
Figure 3
LAGL Authors geo-located in “Classical Athens” (Q844930).
Figure 4
LAGL Authors geo-located in “Halicarnassus” (Q5843680).
Authors can have multiple location statements, as they are stored in Wikidata from primary and secondary sources, and therefore appear in several places corresponding to different life events (e.g., birth, residence, work location). For example, the ancient Greek historian Herodotus (Q26825) is currently located in five places stored under various Wikidata properties: Halicarnassus (“place of birth” P19, “country of citizenship” P27, and “residence” P551), Samos (“residence” P551 and “work location” P937), Thurii (“residence” P551 and “place of death” P20), Pella (“place of death” P20) and Ancient Greece (“work location” P937). These Wikidata statements derive from secondary sources, like Wikipedia and external databases, and from primary sources such as the Byzantine lexicon Suda, whose entry on Herodotus (η 536) mentions four places where the historian was born (Halicarnassus), resettled and was intellectually active (Samos and Thurii), and died (Thurii, but Pella according to other unnamed sources).15
Many questions arise from this experimental map and the database, considering that they show a live view of data queried according to a selection of Wikidata location-related properties, whose values contain a mixture of elements coming from primary and secondary sources. This is certainly not the place to address and answer all these questions, but I will consider a few of them if we want to connect Wikidata with linguistic annotations of ancient sources.
The basic question is how we can convert the language of ancient sources into Wikidata statements expressed through property-value pairs and integrate them with assertions derived from secondary sources such as library catalogs and scholarly publications. The LAGL project is moving in this direction by semi-automatically extracting from ancient Greek sources linguistic forms that also refer to places associated with authors. Named Entities and CTS URNs disambiguate these forms within their context and connect them to the relevant authors, as in the case of the entry of the Suda on Herodotus (η 536). In that passage, the adjective ’Aλικαρνασɛύς is part of the onomastic of the author and annotated not only as a geographic named entity (LOCderiv), but also as urn:cts:greekLit:tlg0016, which uniquely identifies Herodotus of Halicarnassus in the LAGL Catalog. Other tokens in the Suda entry that refer to places associated with Herodotus’ life — for example Samos (Σάμῳ) and Thurii (Θούριον) — are not yet directly linked to Herodotus in the annotations, except insofar as they occur in the same context that contains bibliographic references to him.
The ongoing process of converting these annotations into structured statements through Wikidata location properties remains experimental. It involves not only the inclusion in Wikidata of inflected and lemmatized forms in ancient Greek, but also the selection of the most fitting properties. In the context of the entry of the Suda on Herodotus, we have enough information to align the annotations with specific Wikidata location properties. However, in many other cases, references in ancient sources are generic, and the mapping to Wikidata properties and items can be somewhat arbitrary, depending on the annotator’s interpretation. The ethnics attached to author names in other entries of the Suda exemplify this ambiguity, since it is frequently uncertain whether they indicate the place of birth, residence, or intellectual activity; in such instances, the use of the broader Wikidata property P276 (“location”) would likely be preferable.16
Finally, we must also take into account the toponyms themselves and their corresponding geographic coordinates. The first issue concerns the identification and disambiguation of places. For instance, ethnics such as ’Aλɛξανδρɛύς and Λαοδικɛύς must, whenever possible, be disambiguated, as they may refer to homonymous places: “Ancient Alexandria” in Egypt (Q1244206) or “Alexandria Troas” in Turkey (Q1393407), and “Laodicea” in Syria (Q11931162) or “Laodicea on the Lycus” in Asia Minor (Q849709).
A related issue involves the use of ancient versus modern place identifiers for Classical authors in the values of Wikidata properties. One example is the inconsistent use of values referring to “Athens” (Q1524) or “Classical Athens” (Q844930); to “Roma” (Q18287233) or “Ancient Rome” (Q1747689); and to “Samos,” whether as the ancient Greek city-state (Q13580795), the Greek town on the Samos island (Q1018150), or the island itself in the Aegean Sea (Q156882). Such inconsistencies become evident in map visualizations, where authors appear scattered across multiple coordinates within the same geographical region. The authors situated around Athens in the LAGL Authors Map with Wikidata location properties provide a clear example of this phenomenon.17
2.2 Wikidata Work-related Properties for Classical Authors
Ancient sources preserve information on literary genres and intellectual activities of Classical authors. The expressions containing this information vary considerably, and their structure and richness also depend on the genre and the chronology of the texts that have transmitted them to us. For example, scholia, lexica, and commentaries contain more genre-related expressions than other sources.18 For the purposes of this paper, I can briefly mention adjectives of genre (e.g., κωμικώς ‘comic poet’, ῥήτωρ ‘orator’, or γραμματικώς ‘scholar’), forms referring to teacher-pupil relationships (e.g., μαθητής and διδάσκαλος followed by the name of the pupil or the teacher in the genitive), and expressions referring to school belonging (e.g., Kαλλιμάχɛιος ‘Callimachean’ or oἱ πɛρὶ ’Aριστοτέλην ‘those around Aristotle’).19
As in the previous cases, the question is how we can export genre-related assertions from ancient sources into the structured knowledge base of Wikidata, selecting suitable properties, and combining them with assertions derived from secondary sources. With the assistance of ChatGPT 5, I individuated Wikidata work-related properties with a focus on authors, scholars, and intellectuals; a selection of them is listed in Table 3 with their Wikidata description.20
Table 3
Wikidata work-related properties.
| PROPERTY ID | LABEL | DESCRIPTION (WIKIDATA) |
|---|---|---|
| P106 | occupation | occupation of a person |
| P101 | field of work | specialization of a person or organization |
| P39 | position held | subject currently or formerly holds the object position or public office |
| P1416 | affiliation | organization that a person or organization is affiliated with (not necessarily member of or employed by) |
| P463 | member of | organization, club or musical group to which the subject belongs |
| P135 | movement | literary, artistic, scientific or philosophical movement or scene associated with this person or work |
| P69 | educated at | educational institution attended by subject |
| P742 | pseudonym | alias used by someone |
| P136 | genre | creative work’s genre or an artist’s field of work (P101) |
| P1066 | student of | person who has taught this person |
| P802 | student | notable student(s) of the subject individual |
| P50 | author | main creator(s) of a written work |
| P800 | notable work | notable scientific, artistic or literary work, or other work of significance among subject’s works |
In order to experiment the use of these properties with the LAGL Catalog, I selected four of them to visualize statements about genres and literary activities of Classical authors that are currently stored in Wikidata: “occupation” (P106), “field of work” (P101), “movement” (P135), and “genre” (P136).21 A SPARQL query extracts all Wikidata items with a LAGL Author ID and at least one of these four properties, and the same query is used in an experimental Python workflow to output a self-contained HTML view for searching and filtering LAGL authors by name, CTS URN, Wikidata QID and Label, and Wikidata work-related property (“occupation” P106, “field of work” P101, “movement” P135, “genre” P136). This resource is accessible in the home page of the LAGL Catalog under LAGL Authors Wikidata work-related properties and is regularly updated to reflect changes in Wikidata.22 Of the 1,236 authors currently recorded in the LAGL Catalog, 1,067 are provided with at least one Wikidata work-related value and are therefore represented in the database.
At the time of writing this paper, only the LAGL authors Aristophanes (the Athenian comic playwright) and Plato (the philosopher) have complete entries across all recorded properties in the dataset. Aristophanes (urn:cts:greekLit:tlg0019) has the following values for each property: “occupation” P106 = comedy writer, playwright, poet, writer; “field of work” P101 = drama, literature; “movement” P135 = Old Comedy; “genre” P136 = comedy. Plato (urn:cts:greekLit:tlg0059) has the following values for each property: “occupation” P106 = epigrammatist, philosopher, philosopher of language, writer; “field of work” P101 = ancient philosophy, classical antiquity, education, philosophy; “movement” P135 = Platonism; “genre” P136 = Platonic dialogue.
The database and the examples of Aristophanes and Plato raise several questions that cannot be addressed in this paper but nonetheless reveal the potential of Wikidata’s structured language for representing statements derived from ancient and modern sources. The issues are similar to those discussed for locations in the previous section, although, in the context of linguistic annotations, many work- or genre-related statements are not covered by Named Entities and could be perhaps added through co-reference resolution to identify those expressions that are related to contextual real-world entities (for example, Bημάρχιος … σοϕιστής ‘Bemarchius … the sophist’ in Suda β 259). Moreover, values currently available in work-related Wikidata properties do not always reflect the semantic range of ancient Greek terms in their context.23
3 Results and Discussion
The experiments discussed in this paper demonstrate the potential of Wikidata as a structured environment for representing information on Classical authors derived from ancient sources. By linking the LAGL Catalog of linguistic annotations with Wikidata through the LAGL Author ID Property (P12869), it becomes possible to visualize and query ancient sources according to two main dimensions: (1) their geographic associations and (2) their literary or intellectual activities. This integration represents a significant step toward connecting critical editions and philological annotation workflows with large-scale, collaboratively curated knowledge graphs.
Of the 1,236 authors currently recorded in the LAGL Catalog, 772 have at least one Wikidata location value with geographic coordinates and are therefore visible on the LAGL Authors Map with Wikidata location properties for a total of 360 distinct places, identified through a selection of Wikidata location-related properties. Although the present version of the map does not yet incorporate all the statements from the complete set of annotations of the LAGL Catalog, it provides an overview of how Wikidata is currently representing Classical authors and how these values can be refined and expanded through the integration of linguistic evidence from ancient sources.
Integrating linguistic annotations into structured data environments generates numbers of conceptual and methodological challenges. The disambiguation of ethnics, homonymous toponyms, and ambiguous references to places requires close philological assessment; translating such expressions into Wikidata statements often involves interpretive decisions; inconsistencies between ancient and modern place identifiers further underscore the need for alignment with domain-specific gazetteers such as Pleiades to ensure historical accuracy and spatial coherence.
In the domain of literary and intellectual classification, 1,067 authors in the LAGL Catalog are currently associated with at least one Wikidata work-related property. The examples of Aristophanes and Plato demonstrate how structured vocabularies can begin to represent relationships among occupation (P106), field of work (P101), movement (P135), and genre (P136) applied to ancient authors. However, many ancient Greek terms lack precise semantic equivalents in the Wikidata ontology. This highlights the need for domain-aware extensions that reflect the cultural specificity and historical nuance of ancient terminology, while remaining interoperable with modern linked data frameworks.
By fostering two-way interoperability between philological annotation environments and Linked Open Data infrastructures, the LAGL project aims to transform linguistic evidence from ancient sources into a reusable, semantically rich, and methodologically transparent layer of digital scholarship. The highly experimental approach demonstrates how semantic modelling and data interlinking enable new forms of analysis beyond traditional publications, supporting interoperability, long-term reuse, and future integration with AI-driven research in Classical Philology and Graeco-Roman Studies.
Notes
[1] The annotations with metadata are visible at https://catalog.lagl.org/index.php?what=urn:cts:greekLit:tlg1469 (last access December 2, 2025).
[2] On TLG numbers of authors and works, see Pantelia, 2022, xxiii–xxv and xxviii–xxix.
[3] Prefixes currently used in the LAGL Catalog are lagl, phi, stoa, tlg, fhg. New prefixes are added if available in external reference resources like the Perseus Catalog and the DLL Catalog: see Babeu, 2019 and Huskey, 2019.
[4] The reference resource for TLG identifiers is Pantelia, 2022, who collects more identifiers than the online version.
[5] Berti, 2024; Berti, 2025b.
[6] These include lagl0355 (Apollonius), lagl0375 (Athenodorus), lagl0412 (Myris), lagl0413 (Nicoteles), lagl0425 (Rhodon), lagl0447 (Battos), lagl0453 (Posochares), lagl0454 (Botrys), lagl0455 (Philaenis), lagl0460 (Domnus Judaeus), lagl0465 (Gregorius), lagl0467 (Caesarius), and tlg1139 (Anonymi Historici).
[7] See https://catalog.lagl.org/index.php?what=urn:cts:greekLit:tlg1469 (last access December 2, 2025).
[8] Even if not specifically on Classical Studies, see Farda-Sarbas & Müller-Birn, 2019; Bargioni et al., 2021; Zhao, 2023; Lubin & Fischer, 2024; Pellizzari di San Girolamo, 2024.
[9] Reference resources, among many others, for a discussion on this kind of language are Pfeiffer, 1968 and Montanari, 2020.
[10] Cf. Di Pasquale, 2024.
[11] The ChatGPT 5 prompt is available here (last access December 2, 2025). This conversation was helpful in providing an overview of the variety of location-related properties on Wikidata. However, the results had to be proofread, as ChatGPT generated a few mistakes. For example, P21 is the property for “sex or gender”, not “place of burial”, which is P119. This was incorrectly listed as “place of internment”.
[13] An example of the use of “ancestral home” (P66) is “Thrace” (Q41741) for Dionysius Thrax (Q344124) based on Suda, δ 1172. An example of “place of burial” (P119) is “Orchomenus” (Q13574822) for Hesiod (Q44233) based on Pausanias (9.38.4).
[14] Last access December 2, 2025. Currently, the update is carried out manually whenever new authors are added to the LAGL Catalog, or whenever I add or edit locations associated with authors in Wikidata.
[15] The text of Suda η 536 with a translation is available in the corresponding entry of the Suda On Line project.
[17] In another contribution I will discuss the extraction and visualization of places through coordinates of domain specific gazetteers like Pleiades, whose identifiers are progressively stored in Wikidata under P1584. See Schmidt, 2022 and Elliott et al., 2023.
[18] This is one of the reasons why the LAGL project began by extracting data from Athenaeus of Naucratis, Harpocration, and the Suda: see Berti, 2021; Berti, 2024; Berti, 2025a; Berti, 2025b.
[19] Cf. Pfeiffer, 1968 and Montanari, 2020. On the ambiguity of the preposition πɛρί with accusative, see Savio, 2019.
[20] The ChatGPT 5 prompt is available here (last access December 2, 2025). Also in this case results had to be proofread to find mistakes, as in the case of P8259, which is not the property “student of”. See note 11.
[21] “Student” (P802) and “student of” (P1066) are interesting properties that could be added, given that in ancient sources we have clear references to these relationships, as for example in Suda, α 97 and 571.
[22] Last access December 2, 2025. Currently, the update is carried out manually whenever new authors are added to the LAGL Catalog, or whenever I add or edit genre-related assertions associated with authors in Wikidata.
[23] Cf. Di Pasquale, 2024.
Acknowledgements
I acknowledge the contributions of colleagues and students who provided valuable observations and insights on Wikidata during courses and workshops. Special thanks are due to Camillo Carlo Pellizzari di San Girolamo for his assistance in proposing the Wikidata LAGL Author ID property and for his contributions to related discussions within the Wikidata community.
Competing Interests
The author has no competing interests to declare.
Author Contributions
Conceptualization, Data curation, Formal analysis, Funding acquisition, Methodology, Software, Visualization, Writing – original draft.