Identifying manuscripts in social media
Some days ago a number of articles and blogposts appeared in my twitter timeline criticizing “twitter streams that do nothing more than post ‘old’ pictures and little tidbits of captions for them”1 , e.g. https://twitter.com/medievalreacts
Sarah Werner (whose blogpost I highly recommend!) and others rightly criticized these accounts for using unattributed and unidentified historical pictures for their own commercial purpose, making it impossible to access the underlying historical context of those pictures.
Following the debate on twitter, however, I stumbled across a few tweets (quoted below) that prompted me to scrutinize my own social media practice as a historian, especially the limitations and potential of scientific work using social media, and twitter in particular.
An increasing number of scholars not only employs the microblogging network for channeling private messages, but also to provide glimpses into their professional work and expertise. I have characterized this as a revival of the so called context of discovery (Reichenberg)2 .
Following both the logic of social media and the nature of historical research on the Middle Ages, a good deal of this information shared on twitter consists of digitized images of medieval manuscripts, to which sometimes additional information is added.
There are four types of such tweets:
- Existing images are shared or indicated
- Existing images are scholarly commented on
- New material is published directly on twitter
- Colleagues react by commenting or enquiries
The last point is particularly interesting, because twitter sometimes enables ad hoc collaboration on a small scale that leads to additional scholarly value preserved in written tweets for a distant scientific community to use. It works like a chat between experts on a conference, but is accessible from all over the world and for a long period of time.
@EdvanderVlist Ed, I almost fell off my chair: this is such an unusual hand, with lots of regional influences. Shall have to go to The Hague
— Erik Kwakkel (@erik_kwakkel) April 22, 2013
@EdvanderVlist have a look at CMD It Rome I, pp 29-30 (Rome BNC Farf. 1). Same type of "airy" round hand with v similar features.
— Erik Kwakkel (@erik_kwakkel) April 22, 2013
@EdvanderVlist is this KB? Can I use it in a lecture on Aristotle mss I am giving in May?
— Erik Kwakkel (@erik_kwakkel) April 23, 2013
The key to a fertile use of twitter for scientific purpose however is the ability to identify both the cited manuscripts and the tweets referring to them. Without an individual reference to the cited manuscript, the shared information cannot be used beyond the immediate context or linked to other information. As scholars, we are of course well trained in giving accessible references, but certainly not in an environment that supports only 140 characters.
.@humeannature your query underscores current discussion on @MedievalReacts and pics without attribution. Source: http://t.co/iMQgmZV2pC
— Erik Kwakkel (@erik_kwakkel) April 11, 2015
@erik_kwakkel @HumeanNature & the limitations of 140 characters! But yes, also That Guy.
— Medieval History (@medievalhistory) April 11, 2015
Social media were not designed for an academic purpose in the first place. Twitter does not support footnotes or other scientific systems of references. It only provides 140 characters in the textual body of a tweet to give both the crucial information and a proper traceable signature.
This creates a number of problems:
- The limitation to 140 characters may not be sufficient to give full reference
- Different people may use different styles or abbreviations in quoting shelfmarks
- It is difficult to connect multiple information properly, particularly if a number of tweets are referring to each other
- Searching on twitter becomes very difficult and sometimes arbitrary
So, although it is certainly important to give a correct signature of the manuscript in question in a tweet, it is not the ideal way of making reliable references. Twitter, however, does provide a system of reference that can be very powerful if used properly: the hashtag.
Hashtags are normaly used to tag tweets with a certain buzzword specifying the content of a tweet. They are also used as a marker for tweets commenting on a certain event, for instance a lecture or linking additional information to it.
A first step to enhance the scientific usage of twitter for historians (here is a very good introcduction to twitter by Mareike König) would be to apply hashtags as a method of identifying manuscripts dealt with scientifically in their tweets. By doing so different information on a certain manuscript could be collected to a digital folder: Images, life tweets from autopsies, literature and papers, blogposts and comments, whatever scholars on twitter may find worthy of micropublishing it. Using hashtags consistently could improve the scientific value of information shared.
The nice thing about twitter is its ability to connect a crowd of people around a shared topic of interests or an event marked by such a hashtag in a very short period of time. Usually, this works pretty well, because people interested in an ongoing event, let’s say #IMC15 (Leeds International Medieval Congress), agree on a convenient hashtag that is used exclusively by all participants in order to prevent two separated clouds of information.
A system of referencing manuscripts (or any other kind of historical sources) in tweets, though, has to endure for a longer period of time. If signatures are chosen arbitrarily by an ad hoc agreement, this decision cannot be reconstructed by someone who was not part of the agreement. This holds true for the citation of manuscripts in particular, because there are several slightly different conventions for quoting them. 10 Years from now, scholars cannot be certain if scholars in 2015 referred to a manuscript on twitter as #BSBMünchenClm250, #BSBClm250, #MünchenBSBClm250 or simply #Clm250 (or more likely all of them simultaneously in different languages, thus creating different clouds of isolated information).
Even if it may contradict the intuitive logic of social media, creating identifiers that follow certain standardized and reproducible rules becomes an imperative as we share more and more scientific information out of a semiofficial context of discovery – semiofficial because it has not been published “properly” in an article. Only by standardizing references we can make microinformation last beyond the social now, and prevent it from drifting through a digital space lacking resonance.
One possible way would be to use a system of place holders that are deposited somewhere, such as used in e-musicology, e.g. the cantusdatabase.
However, such a method would require an institutional setting, which is not how social media (should) work. To me, it makes more sense to reach an agreement on certain rules how to mark tweets referring to a manuscript, citing manuscripts as social scientific crowd, just as we do when creating an exclusive hashtag for a conference.
To start this discussion, I propose the following:
In general, tweets on manuscripts should always include the conventional shelfmark of the manuscript in question. The hashtag does not replace the shelfmark, but tags the tweet. The hashtag should be consist of the following parts based on the shelfmark:
- Two-letter country code ISO 3166-1 alpha-2: E.g. De for Germany.
- First two characters of the English name of the City, if the name consists of two words the first character of both words should be used: For instance Mu for Munich but NY for New York.
- Common Abbreviation of the library as used by the institution itself: E.g. BSB for Bayerische Staatsbibliothek; Bl for British Library; UB for Universitätsbibliothek.
- First three letters of the collection if there is one: For instance Har for Harley; Clm for Clm.
- Number of the manuscript
- Parts like cod., ms., hs. should be excluded, no dots should be used.
London British Library Harley 2799
München BSB Clm 250
Paris BNF Cod. lat. 4593
This system can provide both a short and individual reference to any manuscript mentioned in social media and it can link manuscripts to other information such as papers or literature.
This information can then be traced and found later without knowledge of an arbitrary ad hoc agreement of the crowd, but through applying the given rules of hashtag building.
People using social media, twitter in particular, usually like the absence of conventional constriction. They rightly advocate the benefits of open circulation of knowledge which in itself does not include the necessity of special knowledge to create hashtags. And this is a good thing, in general. However, if you share information you consider valuable from a scientific point of view, I think you should also secure scientific usage.
Of course, the method and rules suggestioned in this blogpost are far from perfect and open to debate. An important debate that does not only feature funny historical pictures and silly captions but the very condition of historical research in times of social media itself.
My favourite manuscript: Bl Harley 2799, part of huge and splendid Arnstein Bible http://t.co/Ts8lrwtLiG #GbLoBlHar2799 #manuscript #saec12
— Michael Schonhardt (@Quadrivium_Blog) April 17, 2015
Click here for searching other tweets dealing with the Arnstein Bible or simply click the hashtag used in the tweet above.
- Sarah Werner: it’s history, not a viral feed. Blogpost: http://sarahwerner.net/blog/2014/01/its-history-not-a-viral-feed/ 26.01.2015. [↩]
- Hans Reichenbach: Erfahrung und Prognose (Maria Reichenbach und Hermann Vetter (Hg.): Gesammelte Werke Bd. 4, Braunschweig 1983), S. 3f. [↩]
Thank you for this interesting post. I’d like to call your attention to RISM (where I work), an international cataloging project with a focus on musical sources, which has a long-established method of identifying institutions, made up of the country, city, and institution. GB-Ob is therefore Great Britain, Oxford, Bodleian Library. RISM’s sigla (library abbreviations) can be consulted at: http://www.rism.info/en/sigla.html. Using as an example something I recently saw on Twitter, GB-Ob Add. A44 would be #GBObAddA44 using RISM’s shorthand.
RISM library sigla are created for institutions that have music holdings, which admittedly isn’t every institution, but currently they number over 7,000 worldwide if I’m not mistaken. There are sigla for the institutions you listed in your examples, and so I offer you your hashtags using the RISM sigla:
London British Library Harley 2799: #GBLblHar2799
München BSB Clm 250: #DMbsClm250
Paris BNF Cod. lat. 4593: #FPnLat4593
The RISM sigla lose a bit of recognition when it comes to “common abbreviations” as you suggest, but they can all be looked up in the database, and they save a character or two.
Regarding the question of how to cite medieval manuscripts on social networking sites, I would like to refer your attention to the technique outlined on “Fokus Handschrift”, a new German-language Facebook page. This method, which could serve as a general standard in this area, helps eliminate the issue of source citation on such sites, though not on Twitter due to the 140 character limit imposed by the platform.
It seems to me brevity of hashtag will help with adoption of practice. Unique identifiers for institution and collection item have been used by natural sciences scholars. http://herpetologistsleague.org/dox/CollectAcronym-Sabaj10.pdf I’m wondering whether libraries already have a unique identifier that could allow you to eliminate the city initials.
You make an interesting point and you are right, that brevity would help. There is indeed a unique identifier for libraries (ISIL: International Standard Identifier for Libraries and Related Organisations). But there are two difficulties here. Firstly that identifier is not necessarily short – it has a maximal length of 16 characters and to them we’d still have to add the collection and definitely the shelfmark of the codex. The second difficulty is, that it’s difficult to remember or even to guess. While most medievalists already know the abbreviation for, let’s say, the French National Library – BnF for Bibliothèque nationale de France -, no one would really want to learn that the ISIL code for the BnF is FR-751131015 (or whatever number).
However – the pdf-list you linked to in your comment is quite helpful and, although there does not seem to be a consistent system behind those abbreviations there, a good reference point for a system that includes the libraries and collections relevant for medievalists and others working with manuscripts.
I find Michael’s idea of creating a big cloud of connected information instead of “different clouds of isolated information” intriguing. Certainly, twitter was not designed to cater for academic purposes in the first place. However, since more and more scholars are beginning to use it professionally, changes are in my opinion a natural step in the evolution of the medium. And frankly, I think that we should be helping to form it. Bend the web to your needs, you could say. I see Michael’s suggestion as an easy way of collaboration and co-working, as a means to carry academic matters again a wee bit further into the digital sphere, firmly adding a distinct “social” to both the big anonymous web that surrounds us and our often solitary research in eremitical spaces such as the manuscript reading room of a library. Since twitter is a continuous stream of information I quite like the idea to stick a stop sign in the air and delve into matters more deeply, to follow for a moment only one exclusive branch of information rather than having plenty of it streaming by. As Michael already pointed out, for conferences and workshops or any topic that might be of interest to a larger crowd the hashtag is already a well-tried means. The fact that a hashtag for manuscript identification needs a certain amount of “centralisation” or “regulation” (having a weakness for the anarchical and subversive of the medium, I heavily object to excessive regulation mania myself!) does imho not necessarily contradict the “democratic principles” of social media, because here we are, the manuscript lovers, to discuss it!
In the end, I go with Clemens: “So we should give it a try, use the pattern, and see how it develops”.
As user name @JBPiggin I tweet digitizations by the BAV in Rome on the day they appear. This is done purely out of enthusiasm: I have no official link to the Vatican. When I started up, I thought about the issues you raise, but had to come up with a more pragmatic solution.
From the start, I have always included the URL of the manuscript. The BAV has carefully scripted those URLs to be permalinks that are human readable and they will still be intelligible 1,000 years from now.
After a couple of days, I write a roundup on my blog, developing what is known about the manuscripts, and I tweet about this, with a URL back to my blog. The shelfmarks are in both the URLs and the blog posts to make this system backwards compatible.
But the whole point is that this needs to be a future-proof too.
The world is developing into an “internet of things” where every car, manuscript, elevator and pump will have its own URN. I think that we have to bear that future movement in mind, and not hang on to the shelfmark reference system at all costs. That’s why I so approve of the BAV policy and wish other libraries would follow it. (My nightmare case is the public library in Lyons, France, where I have yet to find any way to link directly to its digitized manuscripts, but there are plenty of other slackers out there.)
From practice, I have to say Richard Fitch’s point resonates with me too.
To be honest, it’s hard to say anything in 140 characters about a non-illuminated 12th century pontifical that makes people go, “Wow, I must look.” It’s even harder to say it in only 100 characters or less, after you lose space to the URL link, the image link and the hashtag #digitised.
Would I enjoy the challenge of adding #VaRoBAVArch.Cap.S.Pietro.F.11.pt.A to the post and saying something intelligent in 80 characters? No.
Thank you very much for your comment!
You raised an important point by mentioning URN. Stable links are the condition for scholarly work using the internet in general, so I totally agree with you there.
However, I had a look at some of the tweets you mentioned, f.i. this one: https://twitter.com/JBPiggin/status/583347492932206592
It refers to the manuscript BAV Vat. Lat. 39 as indicated in the Url of the manuscript (http://digi.vatlib.it/view/MSS_Vat.lat.39/0094). Now the problem is: twitter cannot find your tweet, because it doesn’t search hyperlinks. If you search the manuscript on twitter, you will only find two tweets by @marenostrum2 (https://twitter.com/search?f=realtime&q=vat.lat.39&src=typd).
So, here you can see the issue in a nutshell: We have different tweets on the same manuscript, but they are not linked together and yours can’t be found at all (which is very unfortunate, because you link the full digitized manuscript). By using some hashtag (and I am aware of the imperfection of my suggestion: #VaVcBavLat39), your tweets would be connected. For the sake of argument, let us assume I am going to write a blogpost on that manuscript in a few months (or years). By using the same hashtag, I can tweet a link to that blogpost and connect it to your tweets. Or I can ask “twitter” a question about it (let’s say an issue in reading some words) and the answer would be linked as well.
As you said: “But the whole point is that this needs to be a future-proof too.” As much as I can understand your reservation towards my suggestion, I think the way you (and I myself, to be honest) tweet your information is not a very future proved way. Which is very unfortunate, because you obviously devote a lot of time and energy into your tweets and they clearly are of importance to others.
However, I agree that the reference you added in the end is not very appealing. I would abbreviate it as follows: #VaVcBavArcF11 which leaves you 129 characters to spend. Still not very comfortable, I admit, so I would welcome any better suggestions.
Well, IF a manuscript is online, ideally hosted by the library or archive that owns it and ensures its digital preservation (the German term “Langzeitarchivierung” suits better, though), a link or, better, a persistent identifier (like you already suggested) would be helpful indeed. But, without actually knowing any numbers, I’d guess, that not even one third of all manuscripts is digitised. It’s often a self shot image from a manuscript or only the mention of a manuscript, that is tweeted.
In that case it could be considered possible to include at least a link to the online catalogue of libraries. But again, in quite many cases there is simply no catalogue entry for a manuscript, because although there might exist a manuscript description made by a medievalist, it need’s a librarian to catalogue it the way the respectivly binding rules (in Germany RAK/WB, soon RDA) for catalogisation dictate it – and, unfortunately, both activities are not always matched with each other.
In short: if a manuscript is not entirely digitised by a library/archive, that ensures its digital preservation, and if it does not not have an adequate catalogue entry either, it’s hard to include an appropriate link.
And as long as there is no norm data for manuscripts (I totally support that demand Clemens Radl made earlier!), an abbreviation system seems to me to be the best way to consistently reference medieval manuscripts on twitter.
Last remark: A blog would, of course, be a good place to extend tweets. And likewise, Twitter is quite often used only to link (to) more extensive blogposts. But Twitter is also a distinct way to communicate, even in academia. Sometimes there is no time or no need or no desire to leave this communication channel and to relocate the conversation to a blog. If you want to use Twitter in that manner, again a system of how to reference manuscripts like the system proposed by Michael can be a useful way to tell about what manuscript a tweet is about – especially when those references are “easy to create and to understand without having to consult third party databases”, like Clemens Radl said in his comment.
One of the main criticism’s of the whole medievalreacts affair on Twitter was the lack of attribution for images that would allow non-academic interested readers to be able to follow up the image, see it in context and thus kindle an interest in marginalia, manuscripts and history in general that perhaps may not have existed before seeing the respective Tweet. Is proposing a codified hashtag for referencing manuscripts etc not adding a layer of obscurity for these casual viewers rather than clarification?
Whilst I wholeheatedly agree that some system of referencing that is univerally understoond and allow searching of Twitter postings to link conversation and subjects together would be ideal I just don’t see it being easily achievable with the 140 character limit; to my mind the “accepted” notion of including the @ handle of the image holder and the ms ref are the best of an acceptably bad lot. If space in the initial tweet doesn’t allow then a second appending the reference is the next best option, or possibly with the new way (as of mid April 2015) of quoting tweets where the original tweet is appended by twitter with no impact to the 140 character limit quoting the initial tweet and including the reference then would keep all the information together so that subsequent retweets were less likely to separate image and reference…not ideal I know, but a reasonable solution to that problem.
Clearly, as we have discussed on Twitter (where I post as Tudorcook) it would be possible to have a site listing the abbreviations used within any proposed hashtag, but whilst a good solution for those looking at this from an academic perspective, one that fails from the point of view regarding wider non-academic connections with the subject. Suggestions of including the codified hashtag as well as the full reference would most likely fall foul of the 140 character limitation as well as adding redundancy to the system. Yes it would increase searchability but so does a full reference.
I hasten to add that I am very much in favour of this idea, I am merely reticent with regards to how a wider public would interact with the proposed hashtag; though perhaps this is something that is in all likelihood a concern over nothing. From my experience the wider public is interested in the references not because they themselves wish to follow up or take the idea further but rather they are concerned to know that “we the experts” are doing due diligence to the subject and can show that the history which they trust us to look after and investigate is actually in safe, knowledgeable hands.
Thank you very much for your comment. As I said earlier on twitter, I think we generally agree. I might have chosen a slightly misleading introduction, though, so I try to clarify my approach now.
My main issue is not the attribution of pictures on twitter, though I used it as some kind of a peg here. My focus lies on the traceability of information that might have scholarly value.
I always liked twitter because of its ability to break through the ivory tower and address a wider audience and I do not think it would be desirable to restrain this communications by any rules.
However, people use twitter more and more for scholarly communication as well. I remember a couple of tweets by Lisa Fagin Davis and Eric Kwakkel, where they identified a number of fragments (actually, I can’t find them now, which kind of proves my point). Such tweets clearly do have scholarly value, and they are of special interest to other scholars. The more we share such information, we should think about preserving them.
I like to think about twitter as some kind of a huge social archive. Of course, in order to search this archive (now and in a few years) one can use the implemented twitter search (which is not very sophisticated). However, that does by no means exhaust the potential of twitter: to access a number of linked information quickly.
The problem with just giving a full (or more likely abbreviated) reference is, that it can be found, but not connected to other information. It is like tweeting about the superbowl but not using a hashtag. It works, but it is unlikely your tweet will become part of the digital discourse.
Maybe we can agree to this: Tweets on manuscripts don’t have to use hashtags like I proposed by all means. Most of the time, it is perfectly fine if they just delight people and make them aware of how interesting medieval culture is. Int his case, they just should provide proper attribution. However, if the content of a tweet is likely to be crucial to other scholars, using some kind of a hashtag based identifier seems to be a good way to preserve it for further use.
I think you are correct that we do indeed generally agree, the need for some overarching referencing system for the way that academia and associated users have come to use Twitter would be extremely useful.
It’s a shame that the tagging of images can’t be used to add a hashtag for cataloguing purposes as this would be an ideal solution; adding a searchable tag that doesn’t take up valuable tweet real estate.
I really like this approach: It is simple and effective, even if it does not provide unique identifiers in each and every case: #FrVeBm35 could be (Verdun|Vesoul|Vendome), Bibl. Municipale, Ms. 35. So in the long run the scientific community will need universally unique identifiers for manuscripts just like VIAF/GND/LOC numbers for persons and places. But how many of us would want to include something like #VIAF66806872 in every tweet about Augustine? And who would be willing to look up people in VIAF just to tweet about them?
The interesting thing about the recommendation given here is that the “reference” is easy to create and to understand without having to consult third party databases, just like most historians would know how to resolve “MGH Conc. 3 p. 324”. So we should give it a try, use the pattern, and see how it develops.