Collegare risorse epigrafiche e fotografiche con HTML+RDFa

On the page for SCU 02386 one can learn of a record, by Maria Rosaria Stefanangeli, which gives a bit more information about the inventory number and we can see 4 different negatives, documenting clearly the text and the three visible sides of the statue base, which I reproduced here. Actually, if one still has in mind Julia Domna, named on the Fotothek's page, it is possible to modify the semantically meaningful URN and retrieve information about that as well. SCU 00049, which actually also provides a view of the statue base from the top. Foto Col 07642

The image of the inscription at the front is actually marked as "Malter" and may be the same as bhpd71012 in the Fotothek. Five negatives are named and for four of them a URN can be constructed.

• MC B/1909
• d/7290, Malter
• d/19905
• d/19906
• d/19907

The last two images associated to the statue base in the website of the Musei Capitolini show also that there are additional texts on the same statue base, on each side. Both are dated 1681 within the text itself (MDCLXXXI). The size of the photos does not allow to read much.

On the side to the left of the ancient text we read from d/19907

]NCISCUS FANUS COS
]AN SODERINUS COS
]BERTUS URSINUS COS
]CUS ANT DEGRASSIS PRI[?]
M DC LXXXI

On the side to the right of the ancient text we read from d/19906

FRANCISCUS FANUS[
FRAN SODERINUS [
RUBERTUS URSINU[
MARCUS ANT DEGRAS[
M DC LXXXI [

It seems plausible that these two inscriptions are copies of the same text, inscribed in the same circumstance and for the same purpose in one creative act, so without being one the copy of the other.

Claudio De Dominicis, MEMBRI DEL SENATO DELLA ROMA PONTIFICIA Senatori, Conservatori, Caporioni e loro Priori e Lista d’oro delle famiglie dirigenti (secc. X-XIX), Roma 2009 lists these names at page 56 and 119 of the PDF, which does not correspond to the pagination of the printed book, as notified in the PDF itself.

1681-1/7 - Francesco Fani della Regola, Co. Francesco [F. Antonio] Soderini di S. Eustachio,Roberto Orsini [Ursinus] di Pigna - (Cred. I, to. 35, c. 119v).

1681-1/7 - Marc’Antonio de Grassis [Grassi] de Monti priore, Co. Giovanni Francesco Vespignani di Trevi, Prospero della Molara di Colonna, Alessandro Bentivogli di Campo Marzo, Alessandro Fusti di Ponte, Giovanni Matteo Grifoni di Parione, Francesco Ferrini della Regola, Leonardo Severoli di S. Eustachio, Orazio Cardoni di Pigna, Pietro Zonca di Campitelli, Giulio Sinibaldi di S. Angelo, Giovanni Battista Vitale di Ripa, Nicolò Callimachi di Trastevere, Francesco Brandani di Borgo - (Cred. I, to. 35, c. 119v).

In this list, findable with a Google Search, Cred. I, to. 35, c. 119v stands for Credenzone I, tomo 35, Catena 119v, which is the collocation in the Archivio Capitolino.

COS stands for Conservatori. PRI for Priore dei Capo Rioni, often corresponding, like in this case, to the "caporione del Rione I" (Monti).

These four magistrates were top figures of the Rome administration system between 1223 and 1870. They always signed together all completed projects. Which activity was signed off on these monuments is however not known to me and I have to give up investigating this further. However, the PDF referenced above gives us some more important details, placing the dating of these inscriptions between July and October of 1681.

The following is a screenshot of the edition of these texts, kindly pointed out to me by Silvia Orlandi. Without this suggestion the present exercise would have had a very quick end.

Forcella is based on a quantity of manuscripts as the introduction details, which we are not going to investigate but probably include the ones referred to also from De Domenicis.

The Entry in the Musei Capitolini page, would have even more to say.

There are four links in the description, apparently to other entities in the Musei Capitolini, two of which land on parts of a "Cippo" of M. Valerius Messalla Corvinus: SCU 01018 , and SCU 01032 .

The first seems to be unrelated given the findspot.

The images of the latter instead are interesting, because they also carry inscriptions of the Magistratura, which seem to be even later than the ones on NCE 02386.

The PDF list takes us to the last quarter of 1698,

1698-1/10 - March. Giovanni Battista del Drago Biscia, Filippo Fonseca, Giuseppe Sorbolonghi [no Sorbelloni] - (Cred. I, to. 35, c. 194).

But this base cannot be the other statue base, or base of the column which held a bronze statue, as the Musei Capitolini Catalogue tell us, because it does not have the inscription. The actual base with inscription is there and can be found at SCU 01883 and the printed Catalogue claims it also carries on the sides inscriptions from 1681. But there are no photos of the sides.

We have found online until now three identifiable objects.

• SCU 02386, the statue base on which we can see the first text (later we will see that this can be identified as EDR093684), and Forcella 178 on the two sides, with the statue of Giulia Domna on top, for which we have photos of all sides, except the one facing the wall from the Musei Capitolini.
• SCU 01032, the back half of the statue base with text (later we will see that this can be identified as EDR103369), and Forcella 198 on the sides, for which we have photos of all sides from the Musei Capitolini.
• SCU 01883, the front half of the statue base with text (later we will see that this can be identified as EDR103369), and Forcella 198 on the sides, for which we only have a photo of the front side in EDR but none of the sides and nothing by the Musei Capitolini.

If this consideration was right, because we have dated the texts on the sides from the lists of the Conversatori, we could say that the dating provided in the Museum catalogue for this text does not correspond, which would be however of minor gain. Also, it remains a mystery why the image of the text is related to the atrio of the musei, but we do not have photos of those, while there is documentation of what seems to be the back of that.

Being in Rome, it did not cost much to check, and here we have the missing images to confirm that the text edited by Forcella is this and that the supposition that S 1032 is the back half of this stone may be correct.

I have tried to match the images with the commonest tools, and the result is ugly but provides the argument with a depiction of the complete texts.

There is no database of inscriptions from Rome of this period as far as I know, otherwise I could have contributed there if they did not know yet.

With this two images, also for the two versions of Forcella n. 178 we have almost all the text.

We also find the lists referenced by the XVII century inscriptions, because the Fasti Consulares are there in the Museum.

So, going to the Museum was very useful. All pieces of the bases could be found. Of course without knowing from the web-based searches what to look for, those would have remained supports for nice statues and nothing else.

Of course the quality of my pictures is turistic and it does not even stand close to any sort of documentation photography.

Now, we can go back to the record we started from, because there is another piece of information from the Fotothek's object "Statuensockel mit Inschrift des Caius Cestius", which can be explored, that is, the CIL reference.

We could get a similar effect to that of the manuscript overview, with Google maps.

The web is a sea with many islands and you can jump, by design, from one island to another, seemingly entirely unconnected, as in no other sea. Even better if the page is indexed by search engines and a search with parts of the information available leads directly to the page we do not yet know about. Scholarly resources still do far too little of the far too simple search engine optimization practice to allow for this to happen.

So let me jump to another resource, with a completely different perspective on the same objects of research.

I happen to be quite sure that this inscription must be part of the Epigraphic Database Roma (EDR). Because this is a different field of research, Ancient History and namely Ancient Epigraphy, it is of no surprise that the Fotothek does not point to this resource and viceversa.

The bibliography is here vaster and we find links to two other web resources:

• EDH https://edh.ub.uni-heidelberg.de/edh/inschrift/HD031696
• The Ohio State University Center for Epigraphical and Palaeographical Studies https://kb.osu.edu/handle/1811/100272

We will explore those resources, as we have done above exploring links from the Fotothek, but now let us finish looking at the digital editions of the texts of the inscriptions in EDR.

The most important bit of information we find is

Textus gemellus invenitur ad EDR103369

and, in facts, there it is. This, as the sister record, carry a inventory number which can be compared to that of the Fotothek and the Musei Capitolini. NCE 2386 is associated with EDR093684 only. EDR103369 is associated to NCE 2385, of which we have already said, which makes the information nice and tidy.

EDR093684 (and not EDR103369) provides a link to https://kb.osu.edu/handle/1811/100272. This entity, decorated with a stable identifier (http://hdl.handle.net/1811/100272), publishes 5 different photos without any other metadata about them, as depictions of an epigraphic entry. We get to know that the photos are part of the Arthur E. Gordon and Joyce S. Gordon Collection

Unfortunately, neither the metadata or the images or the bibliography from this resource seem to increase our knowledge of the inscriptions.

The same inscription is classified differently, but aligned to the same EAGLE vocabulary for type of inscription, in the aggregated record for HD031696, as we learn from the Europeana Aggregator.

• Epitaph (EAGLE)

We should have started from here perhaps. Europeana aggregates data from the Fotothek at the Bibliotheca Hertziana as well as data from EDR and EDH. We would have found all in one place, already aligned to one model. So I tried.

While Pharos Vocabularies actually link up the resources they align, the EAGLE Vocabularies leave this to their users. Let us see how much connection we can get and especially if we do learn something new from these connections, except for opening up to more examples of entities tagged in the same way, which is of course already useful for some purposes.

While the entities for Photographer, Current Location and Institution are a very useful local navigation hint, the other terms, which are present in other resources offer ways to connect and align.

Pharos Sockel is aligned to Getty AAT 300264092, an Object Facet but has no declared direct correspondence. The AAT value for Sockel provided by the MIDAS vocabulary is 300080499 as we have seen and the EAGLE Vocabulary for Basis aligns to a DAI Vocabulary term, which cannot be retrieved. But we know that EDR and Europeana as well as the Fotothek and Pharos refer to the same conceptual thing. We know we are speaking of the same thing, and the SKOS version of the MIDAS vocabulary adds the missing alignment to GND.

Skulptur is not used in the Epigraphic databases but both Pharos and the MIDAS Vocabulary refer to the GND entity. For Pharos however this is identical to AAT 300047090, while MIDAS claims is only related.

Marmor is aligned in EDR with a concept from an unreachable resource of the DAI, https://archwort.dainst.org/thesaurus/de/vocab/?tema=2383. In Pharos it is considered identical to AAT 300011443. and so it is by the MIDAS Vocabulary. We can safely say the same for the EDR and EDH.

The discordant typology of inscription is relevant only in the epigraphic database.

Unfortunately, if we were hoping to see the ends join up via the vocabularies, we have not found what we were looking for, or at least not quite.

Vocabularies offer however many further ways to navigate related information, whose relevance is limited for the purpose of this paper.

I did not enter the business of addressing all the bibliography provided in the related entities explored until now in the Fotothek and in Epigraphic Databases, in relation to the case study in object.

I did not do the research part dealing with the scientific publications, which would have most likely made obvious most of the bits of information found out above, because the aim of the experiment was to test the effectiveness of the online resources and their informativeness, coherence, connection and usefulness to facilitate the progress of human knowledge.

We could encode all bibliography and match entries by shared bibliography, or keywords associated to the titles cited of course, not to mention the people authoring and contributing to those resources.

In most cases it is also the case that in the explored resources there is no link to follow for any of the bibliographic entries, even where there could be one, at least to a catalogue entry, if not a permanent link to a openly available publication (or any of the many configurations in between these two extremes).

Bibliography, the queen of all types of scientific linkings, remains thus not actionable in the online resources explored. I have listed above the two references of the Fotothek's Object page, here is a further bibliography list from the other resources I have found.

• CIL 06, 01375, cfr. pp. 3141, 3805, 4688-4689. Linked and discussed above, with additional notes from Géza Alföldy at EDH
• ILS 0917a Inscriptiones Latinae Selectae is available at the Internet Archive, which allows also to link the exact opening https://archive.org/details/inscriptioneslat01dessuoft/page/202/mode/2up
• A.E. Gordon and J.S. Gordon. Album of Dated Latin Inscriptions (Berkeley and Los Angeles 1958-1965) no.16-17
• E. Nash. Pictorial Dictionary of Ancient Rome (New York and Washington 1968) vol. II, 321-323, fig. 1089
• I. Calabi Limentani. Epigrafia Latina. 4th ed. (Bologna 1991) 207-209, no. 32bis
• Boll. Arch., 13-15, 1992, pp. 8-10 (R.T. Ridley)
• Walser, Römische Inschrift-Kunst (2. Aufl.: 1993), S. 50f. can be searched in Google Books.
• AE 1994, 0103, available on JSTOR for those who have access, refers to the above literature.
• SupplIt Imagines - Roma 01, 0155
• F. Feraudi-Gruénais, Ubi diutius nobis habitandum est. Die Innendekoration der kaiserzeitlichen Gräber Roms (Wiesbaden 2001) 137, Anm. 832.
• F. Feraudi Gruénais, Inschriften und Selbstdartellung in stadtrömischen Grabbauten (Libitina, 2), Roma 2003, p. 110 nr. 157
• A. Kolb-J. Fugmann, Tod in Rom. Grabinschriften als Spiegel roemischen Lebens, Mainz am Rhein 2008, pp. 57-58
• Musei Capitolini. Le sculture del Palazzo Nuovo, 1, Milano 2010, p. 223, nr. 32
• Musei Capitolini. Le sculture del Palazzo Nuovo, 1, Milano 2010, p. 220, nr. 25

The only observation I can make towards the aims of this experiment is that similarly to vocabularies the potential augments with the age and relevance of the cited resource. CIL has the most connections to it and remains the primary source of information.

This comes also as no surprise, no one can expect an "exhaustive" bibliography for its own research questions on a helpfully available online record for a document of interest, it is simply not reasonable.

With RDF one can say anything, it is an extremely generic format. Here I have only identified Digital Objects and their relations to Research Organizations using schema.org vocabulary (the model used for Search Engine Optimization by Google). I have tried to be systematic about the relations among the entities and I have provided relations to the main available vocabularies.

Beside more known conceptual reference models like CIDOC-CRM and widely used ontologies like BIBO for the bibliographic information, named vocabularies and sources, I have used properties and classes from the ontology of La Syntax du Codex, which I developed some years ago and tested in another research environment. This is a structural guide to the description of manuscripts. Its applicability to any written artefact has been clear since its publication, and here I have applied it in a very simple way to the two statue bases and their inscriptions.

Another less known model, the so-called epigraphic ontology, for the description of inscription is the result of a large collaboration, which I helped write up some years ago, and is tailored to the need of inscriptions. This helps especially in the distinction of units related to text and support, with a core descriptive distinction based on the recognition of the binary nature of an inscribed object as text on a support and as support with a text. So, I have used that as well.

Additionally, I have added connections among the described resources, and alignment of the vocabulary.

A so-called Spaghetti Graph with all nodes in this presentation is not very useful, or readable. It may convey a sense of unity and coherence, but does not tell us more than what the data has already said, especially in this case, where the triples reproduce the content of the simple flow of the text.

Network Analysis on such a dataset would also be of little relevance due to the network variety in qualification of the relations and classification of nodes, as well as its limited size.

Let me present instead two graphs, one with only the relations which can be extracted from the data available online, and one with only the triples which are added here, excluding also all those triples which are functional to the presentation (title, author) and not related to its contents. I have accomplished this, by adding a custom data attribute to the triples, @data-type

• equal to akg for what has been added here,
• equal to source for what was available elsewhere and is only replicated here, and
• equal to functional for all what could not fit the previous two categories.

I could not think of another way to do this, given I have used the same vocabularies as my sources.

The source data of the following representations is hence the very same text you are seeing, and a javascript library (vis.js) is used to parse it and represent it as a graph. Color represent different types of entities.

In the previous graphs we can observe something which we knew already: the content of a contribution is coherent (hopefully), and the resources out there are only some times kept together by sporadic links.

If we also clean up relations which are not directly relevant semantically to a researcher, e.g. URLs, we have very little, but a quite clear picture of what resources we actually have.

What we take home from this reduction process is that all what really matters are qualified links between existing entities with their information baggage. Explicit and direct ones.

For example, the single link added by virtue of sending an email was more important than any long process to integrate resources in platforms and aggregators.

We can see that Epigraphic databases are more and better connected among them, yet there is only that one link joining the lonely Fotothek to them. And another lonely link to join the Fotothek to the resources of the Capitolini, making of the Fotothek, unexpectedly, the actual bridge among all these resources, hence enforcing the user-centred experience. If someone else than me had done the same experiment, the result would have been different.

This paper lists more than 40 available images online related to one another in several ways and adds to those other 30, while it imports even more from other resources but the documentation of the object in question remains perceptibly and factually incomplete.

I have assigned relations of the type "depicts" (crm:P62_depicts) to the sources, assuming their implicitness.

We know we have double images of the most ancient texts and partial images of the texts on the sides. We do not get to know from the online resource even which is which or what text is on what stone and what their relationship is. We need an argument to make sense of that information and many more statements, which I have tried to add and document here.

But of course this example alone is not a paradigm of the general situation, which is definitely much more complex as the items become more complex, multilayered and interconnected.

The actually important information, the one adding up knowledge of the objects, discovered by following up links provided and initial information found online, came up primarily in my opinion from:

• Visit to the Museum (Musei Capitolini)
• Hidden Web resources (Musei Capitolini)
• Personal conversation (see Aknowledgements)
• Paragraphic information (notes in the Fotothek)
• Hand-written notes on paper (Fotothek)
• Connection of resources in physical collections (Fotothek)

I have run another small declarative experiment on the data I have encoded here to see if we can evaluate which kind of information has more to say.

I have tried, always with vis.js javascript, to split up the data in this paper in networks by vocabulary of the statements.

It seems plausible to say that, smaller, more focused models have a better chance at consistency, readability and at the end of the day results. From the same data one could show all depicted sides of the object at the different identifiable stages of their transformation.

Links to non-replicating resources, or to replicating resources which provide added navigation and content value, are very important for findability and for a richer network. Fortunately they have to be curated, hand-picked, selected and encoded manually to guarantee quality information, as well as any other piece of information. One needs to know where and which to pick and use.

We need links, to join up in the wider worldwide network. And not only internal links to make better sense of our own contents, but especially links to other resources, reaching out and letting the users reach out to other relevant resources. And yes, a generic link to a specific resource is more immediately useful than an alignment to a third one. Especially when crossing disciplines, the paths of a mapping can be longer, but indeed more fruitful than a generical and mechanical alignment, which can only be made relevant buy this network of connections.

A curated link, a wilfully structured semantic statement or group of statements, individually curated, is much more valuable than any number of machine-generated links and models.

The missing link could be a network of curated links which only question driven scientific research can produce. These need, on one side, ways to facilitate this linking, and on the other side reliable curated linkable resources.