Data and data categories in I.Sicily

General remarks

All the data presented in the main table of I.Sicily is drawn from a set of XML files, marked up according to the TEI-EpiDoc standard. The data has been assembled over an extended period of time (since the year 2000), and is drawn from existing publications and a programme of autopsy (or wholly from autopsy, in the case of unpublished material). The data was originally assembled in an Access table. In the process of conversion to TEI-EpiDoc, the data has been cleaned, but also enriched. The process of cleaning and enrichment is ongoing. We here seek to make explicit the origins and handling of the data up to this point, so that users have some understanding of the nature (and likely quality) of the data before using it for their own research.

columnpicker
fields in the column picker

It is possible to view all the main data fields in the main table on the I.Sicily landing page – by default some of these fields are hidden – use the ‘Column picker’ button to the upper right of the table to choose which are displayed.

filtermenu
using the filters

All fields in the main data table can be filtered and searched using the small menu available top right of each column heading in the table. All data currently displayed in the table, filtered or unfiltered, can be downloaded as a .csv file at any point by clicking on the link at the bottom left of the table. Further guidance is provided on the ‘how to use’ page.

A more extensive set of data, in a more user-friendly format, appears in the full record page for each inscription, e.g. http://sicily.classics.ox.ac.uk/inscription/ISic000612. Each individual record can be accessed by clicking on any cell in the row for that inscription in the main table, or by clicking the identifier in the pop-ups on the map. Users should note that even the web publication of each text will not present every piece of information held in the EpiDoc file. In order to xmlbuttonssee the underlying EpiDoc, it is necessary to view the source XML, which is accessible through the ‘xml view’ and ‘xml download’ buttons at the foot of each record (all the XML files can also be downloaded in a single .zip file from the link at the foot of the main page).

File status

All records are given a ‘status‘ (marked up with the <revisionDesc status=”…”> element in the TEI), which at this point indicates whether they are simply the unchecked product of the original conversion (‘unchecked’), a file which has been provisionally checked for conversion errors and other basic mistakes or omissions (‘draft’), or a file which has been subject to more comprehensive editing and is in some sense ready (‘edited’). Files marked as ‘unchecked’ should be handled with caution – the data itself is as reliable as in other files, but errors and omissions consequent upon the primary data conversion and lack of editing are more likely; in particular any epigraphic text is likely to be an automated import and incomplete at best. Files marked as ‘draft’ are likely to contain a limited amount of data, or empty data fields, since the inscription may not have been thoroughly researched, but the epigraphic text itself will, as a minimum be a properly encoded copy of a reliable modern edition. Files marked as ‘edited’ may still not be complete, but are in general those that have been subjected to autopsy and/or further study.

Identifiers

The I.Sicily number is a unique identifier, of the form ISic000298, which in turn forms the tail of the http URI that is assigned to the full record of each text (e.g. sicily.classics.ox.ac.uk/inscription/ISic000298). The number is marked up with <idno type=”filename”> in the TEI. This number does not change if the record is updated, and will persist as an identifier even when, e.g., two previously unconnected fragments are joined. The numbers are assigned sequentially to inscriptions as they are added to the dataset, and carry no other significance. Ordinarily, a number is assigned to an individual inscription. On rare occasions where two clearly independent texts are inscribed on the same object they will normally be assigned separate identifiers, but particularly with fragmentary texts complete consistency is impossible. All inscriptions in I.Sicily are also recorded in Trismegistos (www.trismegistos.org), which also provides a unique identifier (I.Sicily and Trismegistos collaborate, so new texts will eventually be assigned numbers in both systems). Decisions as to what constitutes a single document for these purposes may on occcasion differ between the two databases, but the records will be aligned (cf. http://www.trismegistos.org/about_identifiers.php).

Citing records from I.Sicily

A fuller discussion of how to cite records from I.Sicily is provided under ‘How to cite‘. Here we simply urge users to consider the distinction between

  1. an abstract identifier for a text, whether the abbreviated handy reference such as ISic000612, or the full formal URI such as http://sicily.classics.ox.ac.uk/inscription/ISic000612.
  2. A permanently referenceable, citable edition, retrievable at any point in the future and the work of one or more individuals (as in a paper edition). For this purpose, I.Sicily editions are archived at each revision at Zenodo, with a unique DOI (e.g. http://doi.org/10.5281/zenodo.4021517). The reference details are provided in the current HTML version of any I.Sicily file.

Digital identifiers

(marked up as <idno type=”…”> in the TEI.)

In addition to the Trismegistos (TM) number already noted, we include the numbers for inscriptions recorded in the Epigraphic Database Roma (EDR), the PHI collection of Greek inscriptions, and the Epigraphic Database Clauss-Slaby (EDCS). EDR alignment is not total, but well progressed; likewise EDCS; PHI alignment is believed to be complete for all texts on stone (c.2000). Original alignment was undertaken manually, but automated alignment via the TM API is planned. Although some numbers are referenced for the Epigraphic Database Heidelberg (EDH) also, coverage of Sicily has been transferred to EDR (and EDH consequently only holds 193 records for Sicily).

It is possible to search and filter for Trismegistos, EDR, PHI and EDCS numbers on the main table.

Editions / bibliography

(marked up under <div type=”bibliography”> in the TEI.)

We have chosen to concentrate primarily on the inclusion of substantive discussion / editions of individual inscriptions, rather than citing every reference to an inscription as it occurs in e.g. historical discussion. Substantive discussions which do not constitute an edition will be referenced under a separate <listBibl> element, while bibliography that only pertains to the commentary will simply be included within the commentary. The primary bibliography (over 900 items) is maintained in a separate public Zotero bibliography (https://www.zotero.org/groups/isicily/items/). A copy of the Zotero library is cached locally within the website, in order to enable rapid searching and filtering within the web interface, on the separate bibliography page. All references to items from the bibliography are linked via the item’s Zotero key (which serves as a URI) to the original Zotero record, since export and use of the bibliographic items by users is easier and more powerful through the Zotero web application. Handling reference to editions in this way enables us to link inscriptions and bibliographic items directly, and so enables a powerful range of filtering and searching for inscriptions by bibliography. Selecting an individual item from the bibliography table (by clicking on the relevant row in the table) will produce a fuller record of the bibliographic item and a version of the main inscriptions table below listing the inscriptions recorded in that bibliographic item. In the main inscriptions table on the front page it is also possible to create ad hoc concordances between multiple bibliographic items (and to export these as .csv files).

We have classified some items in the bibliography as ‘corpora’ or ‘bulletins’ rather than simply as ‘publications’. Although from a pure data perspective this may not be entirely logical, we think that it is more intuitive to historians and epigraphers. A full list of the ‘corpora’ and ‘bulletins’, which includes key epigraphic gazetteers such as AE and SEG, as well as the main Sicilian catalogues, can be obtained either by clicking on the ‘filter by corpora’ box above the main table, or within the ‘column picker’ button. A list of the principal collections surveyed systematically is provided below. Note that it is possible to select multiple corpora in the ‘filter by corpora’ box  and then to apply filters to the new columns that are displayed.

corporadropdown
select one or more corpora for filtering

The ‘filter by publication’ box will instead provide an alphabetic dropdown of every publication in the main bibliography other than the ‘corpora’.

publicationdropdown

Users may find the search by individual publication easier to perform from the main ‘bibliography’ page instead, where the full bibliography is presented in a similar table format, and selecting any row will in turn generate a table of linked inscriptions.

I.Sicily has been developed beginning with the systematic incorporation of lapidary inscriptions that are from (or probably from) Sicily, as referenced in the following corpora:

  • CIL X.2.6976-7512
  • IG XIV.1-603
  • L’Année Epigraphique
  • Supplementum Epigraphicum Graecum
  • Agnello, S.L. 1953. Silloge di iscrizioni paleocristiane della Sicilia. Rome.
  • Manni Piraino, M.T. 1973. Iscrizioni greche lapidarie del Museo di Palermo. Sikelika 6. Palermo: S. F. Flaccovio.
  • Bivona, L. 1970. Iscrizioni latine lapidarie del Museo di Palermo. Sikelia 5. Palermo: Flaccovio.
  • Bivona, L. 1994. Iscrizioni latine lapidarie del museo civico di Termini Imerese. Vol. 9/8. Kokalos Supplementi / Sikelika serie storica. Palermo / Rome.
  • Korhonen, K. 2004. Le Iscrizioni Del Museo Civico Di Catania : Storia Delle Collezioni, Cultura Epigrafica, Edizione. Helsinki: Societas Scientiarum Fennica.
  • Brugnone, A. 1974. “Iscrizione greche del museo civico di Termini Imerese.” Kokalos 20: 218–64.
  • Bitto, I. 2001. Le Iscrizioni Greche E Latine Di Messina. Vol. 1. Pelorias 7. Messina: Dipartimento di Scienze dell’Antichità dell’Università degli Studi di Messina.
  • Amadasi Guzzo, M.G. 1967. Le iscrizioni fenicie e puniche delle colonie in Occidente. Rome: Istituto di studi del vicino Oriente.
  • Dubois, L. 1989. Inscriptions Grecques Dialectales de Sicile : Contribution à L’étude Du Vocabulaire Grec Colonial. CEFR 119. Rome: Ecole française de Rome.
  • Dubois, L. 2008. Inscriptions Grecques Dialectales de Sicile. Tome II. Geneva: Droz.
  • Bernabò Brea, L., M. Cavalier, and L. Campagna. 2003. Meligunìs Lipára. XII. Le Iscrizioni Lapidarie Greche E Latine Delle Isole Eolie. Palermo: Mario Grispo Editore.
  • Pugliese Carratelli, G. 1956. “Silloge dell’ epigrafi acrensi.” In Akrai, edited by L. Bernabó Brea, 151–81. Catania.

This list is very far from being the complete list of incorporated items, but simply indicates the initial publications which were systematically and completely incorporated. Every item that appears in the separate Zotero bibliography is included because it is referenced at least once in the corpus for the publication of at least one inscription. All of this material is being continuously extended by the addition of other publications (e.g. individual journal articles) and this is an ongoing process (we welcome all notices of additional material for inclusion). The corpus also already includes and will in future aim to include otherwise unpublished material wherever possible.

Specific limitations to note on the coverage (November 2020): references to BE are currently not systematic or complete. Material from older corpora such as SGDI and CIG, and the antiquarian collections such as Torremuzza is included but not yet systematically so (although CIG references are present for all material in IG XIV). A systematic survey of the Notizie degli Scavi di Antichità is in progress, which will enable the first index of Sicilian inscriptions in the Notizie (much of the material from the NSA is already included and indexed, but not all). The series of volumes Iscrizioni greche archaiche by R. Arena are extensively included but not yet systematically or completely. The biggest ‘gap’ in the existing set of records is the limited inclusion of Christian material, in particular from the catacombs of Siracusa. We are collaborating with the I.C.I. Siracusa project of the late dott.ssa M. Sgarlata, with the assistance of dott.ssa I. Gradante, as well as the Museo archeologico regionale Paolo Orsi di Siracusa, in order to rectify this as soon as possible.


OTHER CATEGORIES OF DATA

Date information

Very few Sicilian inscriptions can be tightly dated, whether by archaeological evidence or through internal evidence in the text. Dating inscriptions palaeographically (i.e. by letter forms) is a very inexact science, especially in a region where any one location has produced a small number of texts over a long period of time, and when few of those texts can be dated by other criteria. We suggest a date for the majority of the inscriptions here, and our aim is to offer a date for every text. In the majority of cases, the dates included are taken from existing publications and discussions, but where autopsy has been undertaken we will update that assessment if necessary/possible. In a great many cases, however, a date range of one or more centuries is the best that can be suggested. Where a date is missing, this is likely to be either because none has so far been suggested in existing publications, and/or because autopsy has not yet been possible, and/or because the inscription simply cannot be dated with any useful precision.

Dates are recorded both with a free-text description and numerically. The numeric encoding is always expressed as a range between two dates (in the TEI using <origDate notBefore-custom=”0000″ notAfter-custom=”0000″>). Inscriptions can be filtered on the basis of the numerical coding. In the original conversion from the old Access dataset the chronological range was never narrower than a 50-year period, even when the inscription could formally be dated to a single year on the basis of internal content. As we revise the individual epidoc files, this encoding will be made more accurate, as appropriate.

Although not originally included, newly edited (or re-edited) files will normally incorporate an indication of the basis for the dating decision. This is encoded on the origDate element, using the attribute “evidence”, with values following those in the EpiDoc guidelines on dating criteria.

Geographical information

Geographical information is included in a number of places in the I.Sicily files, relating to three primary categories of information:

  • the current location of the inscription
  • the original findspot of the inscription
  • the original location of the inscription when erected/displayed

If the current location of the inscription is in a museum collection or equivalent, or a maintained archaeological site/park, then this information is generally recorded in the <msDesc> element of the TEI header. For the treatment of museum collections in I.Sicily, see the separate section below.

If the original findspot of the inscription is known, then this information is recorded under <provenance type=”found”>, within the <history> element of the TEI header (and going forward this information will be further encoded in line with the EpiDoc guidelines). This information is recorded as plain text and presented on the individual inscription page under ‘provenance’ within the ‘Location’ section. Wherever possible, precise geographical co-ordinates (in standard decimal degree format) will be included, within a  <geo> element; these are marked as approximate only when necessary. If specific information regarding the findspot is available and recorded in this element, then these co-ordinates are visualised on the ‘find spot’ map on the individual inscription page, using a pin-marker. If no precise information is available for provenance, then the assumed original location will be shown on the map (using a large transparent circular marker to imply lack of precision).

The original location of the inscription is included for almost every text, but in a great many cases the information is only a generic attribution to a particular site or city, since more precise information is not available. Where an inscription is confidently known to be still in situ, then this information will be the same as that under <provenance type=”found”>. Original location information is recorded in the TEI under <origPlace>. Wherever possible, two sets of information are then included, the <placeName type=”ancient”> and <placeName type=”modern”>. The former, when it is known, is then supplemented by the addition of a reference to the Pleiades URI for that location and the ancient name in plain text. A Geonames URI is provided on the modern place name, provided there is an equivalent modern placename. The geographical coordinates for a representative point at the approximate centre of that location are additionally included to facilitate mapping.
For example:

<origPlace>
<placeName type=”ancient” ref=”http://pleiades.stoa.org/places/462213“>Gaulus
</placeName>
<placeName type=”modern” ref=”http://www.geonames.org/2563049/gozo.html“>Gozo
</placeName>
<geo>36.04528, 14.24167</geo>
</origPlace>

In any case where the original provenance is known to a greater degree of precision, that information is added under <provenance> as described above. It should be noted that any more specific information of this sort is currently not included in the primary table for filtering and searching, which is limited to searching by the main original settlement to which the text is attributed. Findspot data is only presented fully on the individual inscription page. The map on the main inscription search page displays inscriptions according to this <origPlace> data, i.e. according to the settlement where they are presumed originally to have been displayed.

Museums (work of M. Metcalfe)

One of the main aims of I.Sicily is to provide accurate information on the current location of inscriptions. Since the majority of texts are held in museum collections, this entails detailed documentation of the museums, inventory records and display information. Some information on the museums and sites of Sicily is available, for example via the Regione Siciliana website, which maintains a list of museums, galleries and sites. However, this list only relates to those under the authority of the Regione, and to our knowledge no full, reasonably up-to-date listing is available. We have therefore decided to build our own database of museum collections in Sicily (the primary inspiration and all of the data collection and entry was undertaken by Dr Michael Metcalfe). Doing so enables us to include all known collections of relevance (we have attempted to include all museums with an archaeological collection of any description), and in turn to maintain citable URIs for the museums. The database is presented through a web-interface at http://sicily.classics.ox.ac.uk/museums. This can be searched and filtered in the same way as the main inscriptions page, and the table downloaded as a CSV file. Clicking on an individual record will open the dedicated page for that museum (e.g. http://sicily.classics.ox.ac.uk/museum/1), where basic information about that museum is presented, together with a table (in the same format as the main inscriptions table) detailing the inscriptions known to be held in that museum. We are in the process of extending this database to include maintained archaeological sites on the island also, and the museums and sites table can be toggled to show either sites or museums, or both (separate records are included for antiquaria on a site and for the site itself where such double entities exist, since inscriptions may be held in the antiquarium or still be in situ on the site). The vast majority of the museums listed in the database have been visited and checked (by Metcalfe). A separate field, populated manually, provides information on inscriptions known to be on public display at the time of the most recent visit. The primary piece of information that we aim to include is the inventory number for each inscription.

This information is marked up in the <msDesc> section of the TEI header, as per the following example for Palermo archaeological museum:

<msIdentifier>
<country>Italy</country>
<region>Sicily</region>
<settlement>Palermo</settlement>
<repository role=”museum” ref=”http://sicily.classics.ox.ac.uk/isicily/museums/064“>Museo Archeologico Regionale Antonino Salinas</repository>
<idno type=”inventory”>3508</idno>
</msIdentifier>

Deprecated or alternative inventory numbers will also be recorded under the <altIdentifier> element, and any further specific information such as the exact location within the museum (e.g. the particular store-room or gallery) is marked up in the <additional><adminInfo> element of the <msDesc> section, as free text.

This structure with individual museum pages enables us to develop an online catalogue for any individual museum’s epigraphic collection, and we are already collaborating with a number of museums to develop fuller records of their holdings.

Materials, inscription and object types

All three of these elements have been marked up in the TEI with the appropriate EAGLE vocabulary URIs, in order to impose consistency and to enable future linking of data. Where such URIs have been added, the data in the web page for the inscription has been rendered as a hyperlink to the existing EAGLE pages, which generally provide multilingual translations of the equivalent terms.

Information on the type of stone used for inscriptions is generally very poorly recorded in existing publications. We claim no particular expertise at this point in petrology (although it is a future ambition to conduct a more systematic and scientific analysis of the stones used for Sicilian inscriptions), and so in general types of stone are no more specific than e.g. ‘limestone’; and on where the type of stone has not been identified the generic ‘stone’ or ‘local stone’ will be used instead, usually following existing publication in default of autopsy. For example:

<material n=”volcanic” ref=”http://www.eagle-network.eu/voc/material/lod/28.html“>volcanic (pietra lavica locale)
</material>

Categories of inscription and object have been included wherever possible (on the basis either of existing publication or autopsy). Where classification is either ambiguous or uncertain, then either a second alternative classification will  be included and/or a cert=”low” attribute added to the element. For example:

<objectType n=”stele” ref=”http://www.eagle-network.eu/voc/objtyp/lod/250“>stele</objectType>

and

<textClass>
<keywords scheme=”http://www.eagle-network.eu/voc/typeins.html“>
<term n=”funerary” ref=”http://www.eagle-network.eu/voc/typeins/lod/92.html“>funerary</term>
</keywords>
</textClass>

Information on the text form/layout

Information is included in the EpiDoc on the layout of the inscription (a plain text description within <layoutDesc>), on the form and height of the letters and the interlinear space (under <handNote>, but only where the information is available), and, following the categories of the EAGLE vocabulary, on the execution technique of the inscription (within <layout><rs type=”execution” ref=”http://www.eagle-network.eu/voc/writing/lod/1“>Engraved</rs></layout>). Where autopsy has been undertaken this information is likely to be fairly complete; in default of autopsy, the information is limited to that available in existing publications, but this may not always have been transcribed in the initial round of work.

Texts

Although this may seem perverse to some, and we acknowledge that it will be frustrating to many at this stage, the inclusion of fully edited texts was not a priority in the initial work on I.Sicily. In part this is due to the fact that texts were already available for many of the inscriptions in the other online databases (particularly EDCS, PHI, and EDR), where including a text has been the priority. However, this has also been because waiting to achieve this would have placed an unacceptable and unhelpful delay on publishing the existing available material. The first ambition of I.Sicily has been to be comprehensive in identifying and locating the inscriptions of Sicily, and to include as much information about the inscriptions and their material support as possible. Careful editing of the existing published texts, wherever possible supplemented or replaced by autopsy constitutes a second stage in the work flow, and we have decided to put the records online before this work has been completed, in the belief that the corpus is already useful in its current form, and to encourage further contributions. A further hindrance to quick inclusion lies in the fact that the texts currently available online in other resources are, in the majority of cases, not yet available in the EpiDoc format (although we acknowledge and thank the EAGLE and EDR projects, and particularly Pietro Liuzzo, for making the EAGLE generated EpiDoc conversions of the EDR texts available to us).

As a first step in the inclusion of texts, unedited drafts of conversions of published texts were uploaded for approximately 2000 inscriptions. These were however mostly the result of automated conversion routines, intended to transform texts from standard rich-text format with Leiden annotation into generic EpiDoc texts. Consequently, until these texts have been checked and edited, most of them are likely to display incorrectly and show errors. We ask for patience while this process is underway, and welcome offers of help! (We are using this situation as a training opportunity, and multiple students have already contributed to this editorial work.)

Once a substantial proportion of the texts have been incorporated and edited to a basic standard, it is our ambition to start exploring the possibilities of more detailed TEI mark-up, within the framework of the Crossreads project.

Once edited, texts will be supported by a formal apparatus (using the appropriate TEI division), which will provide information on specific problems of readings and alternative restorations in other editions.

Translations

In parallel with the inclusion of texts, it is our ambition to include a translation for every text in I.Sicily (except where the text is too fragmentary to permit a meaningful translation). Published translations of Sicilian inscriptions are few. We aim to include translations in both English and Italian, but welcome offers to add translations in other languages also. At this stage we do not propose to align the translations, but this remains a possibility for the future, as part of the aim of improving the depth of mark-up on the original texts also.

Commentary

A full commentary will be added gradually to individual texts, and will be used as an opportunity either to explain a text or to offer a more in-depth analysis, with further bibliography and referencing.

Images

Wherever possible, one or more images of the inscription are included, and we encourage those with access to texts to send us an image where one is currently not included – but please only do so if you have the appropriate permission. We aim to include images of the support in the round also, and not only the inscription text itself, although the text is inevitably our priority. We are in dialogue with many of the major museums in Sicily for permission to include images of materials in their collections, and the responses have been extremely positive and supportive so far. All images are made available purely for the purposes of study, and if you require an image for publication, please contact the relevant museum directly.

Images are presented using the IIP image server, which uses the tiled TIF format. They are viewed in the browser using the Open Sea Dragon javascript viewer. This enables the easy presentation of high resolution image files over the internet. A JPEG version of the image is also made available, for the purposes of printing and saving of individual records.

%d bloggers like this: