Tag Archives: Data collection

2.5 Degrees from Ego

The correspondence network of Adriaan Reland (with data taken from Early Modern Letters Online) as a ego network taken to 2.5 degrees

Friday n° 37, June 28th, 2019

The last weeks have been packed with work, so that my last two blog posts had to be cancelled because I had to write chapters, presentations, papers, and other things (all about or issuing from the project, so that has all been working time) and was not able to communicate the state of work here for the time being. As promised, I am now returning to my schedule as the flood begins to sink and I’m no longer fearing to drown any moment, and will from now on again deliver my weekly research stats.

Hooray for data!

And to begin this week’s state of research, let me take the opportunity to advertise the ‘other thing’ I have been working at for the past weeks, which actually is a data set. Thanks to and in collaboration with the ERC Skillnet project I have been able to publish some of the letters I’ve been working as ‘The correspondence of Adriaan Reland’ on Early Modern Letters Online. 212 letters have either survived or can be inferred from those letters which I close-read with certainty, and, sadly, that’s all that is left. The metadata of these letters to and from Reland are now accessible via this collection, so I’d like to invite you all over to have a look at these. The added value of embedding them in a larger context as provided by EMLO of course lays in the possibility to explore the wider interconnections of this parcel of letters within the res publica litterarum of Reland’s time, so I thought, let’s give that a try for today.

A 2 degree Ego network

The 212 Reland letters have been sent to 36 correspondents, not so many in terms of scholarly correspondences, so it seemed a good idea to construct a second degree ego network from this in EMLO terms: I checked all of Reland’s direct correspondents via EMLO for their contacts among each other, to see how densely they were interconnected apart from their shared correspondence with Reland. So what you see in this visualization of the whole network is Adriaan Reland at the centre, highlighted in red, as befits a good ego network, and gathered around him his direct correspondents as green nodes, with the communication to Reland as green arrows also. The thickness of the edges depends on the amount of letters exchanged, whereas the size of the nodes is due to the total number of references to the scholar the node represents in terms of the whole network. Connections between Reland’s direct correspondents drawn from EMLO are visualized by black arrows, the thickness again keyed to the number of letters.

The missing 0.5 degrees…

But first of all there is a lot of blue and grey stuff in this diagram which I have not yet said anything about, and second I promised you a 2.5 degree network in the opening headline. So what about that?

Well, as you suspect already, both things are directly related. To start with the missing 0.5 degrees, this was due to the way in which I collected data on these letters. As my project is all about references to other scholars to track the circulation of information, I went through a part of these letters very closely, trying to identify each person and publication mentioned therein. Now EMLO does not support inclosing information on publications in their metadata on the letters, but they do support inclosing mentions of persons, so this is all in there. And that means these data were there to work with them in drawing up the ego network. So what I wanted to have a look at was if the people mentioned to Reland’s direct correspondents would line up with the rest of the network – that is, the contemporary people mentioned. That would give an indication as to whether this is something like an 18th century communication bubble and thus might provide a way to reconstruct missing epistolary evidence. All of these mentions I took to constitute a triad between a) the person who mentions a name, b) the person who is mentioned, and c) the person this name is mentioned to, and all of these relations are visualized in grey within this diagram, as they quite literally are somewhat shady in terms of ‘real’ network edges.

Click the image for a full view of the visualization graph.

What I then did was checking in EMLO if those people who were only mentioned in Reland’s direct correspondences had direct contacts with his correspondents or amongst each other. This would be something in between a real third-degree connection and a second-degree connection, so I decided to label it 2.5 degrees of separation. For those people and connections who had such connections, nodes and edges are coloured in blue to clearly distinguish them from the green nodes of the inner circle of direct correspondents and the green and black relations of these with Reland and amongst each other.

… and what they revealed

What’s really of interest now is of course: What did I gain from doing this? Is there anything new for my project in there? And yes, it was. There are 116 people mentioned in the letters between Reland and his correspondents, 97 out of which still lived in 1680, the year which I took as the point of demarcation between roughly ‘contemporary’ scholars and those whose correspondences I did not take to be relevant to Reland’s connectedness within the republic of letters in any way (this starting point is of course debatable, but one has to start somewhere). This testifies to Reland’s letters discussing more recent issues then debating past works and results; quite a large share of these 97 scholars mentioned whom I designated as ‘contemporary’ actually survived him, sometimes for several decades.

Of the 97 contemporaries mentioned, 26 had contacts either amongst each other – represented by blue arrows – or with members of the green inner circle of correspondents, represented by black arrows. Together with the interconnections of Reland’s direct correspondents between themselves, the second degree of the Ego network, the second-and-a-half-degree connections make clear that he actually was situated in an environment where most people directly or indirectly knew each other, and much of what he relates to in his letters has to be seen in this context. Although these people were spread out across the Netherlands, France, Britain, the Holy Roman Empire, Italy, and Denmark, it seems that they formed a remarkably tight-knit community. If this was a strength or a weakness of the network in question would have to be evaluated against contextual information which EMLO does not provide, but it gives hints where to look next.

Mind the gap!

All I’ve written in this post so far has to be taken with a grain of salt, of course, because EMLO is – as great and wonderful a resource as it is – far from being complete (although I would think it is comprehensive). This means that most probably there are interconnections which I missed out on because they are not (yet) in there. And this becomes even more pronounced as I did not close-read all of Reland’s remaining letters but only about a third of them, so I have missed out on a lot of mentions which would have given me new leads to track in EMLO as well. From the point of a study of remembrance and forgetting such as mine this is of course something to emphasize: What is in there is what is still – at least in some way – in circulation, in this case triggered by current research into these phenomena; and the gaps are caused by processes of discarding and source loss, which are parts of structural forgetting. So by mapping out the gaps, I do get a better grip on what I am really at also. And it’s never bad to gain added benefits from something.

And: Mind the results nevertheless!

But the point was precisely this: To see what I could do with this resource even as it stands now and regardless of the patchiness of my own research so far. And that proved quite a success, because in all likelihood I would only have ended up with more connections in the end, not less, which strengthens my initial hypothesis that the 2.5 degrees are a useful way of coming to terms with gaps in the physical evidence. In Reland’s case, this worked out well and points to him as really being closely connected within a certain epistemic community (as I once termed this) of his day.

What now remains to be done is to see if this also works out for my other three protagonists, since this would provide a way to cope with the dearth of direct epistolary evidence. I’ll see to it that I can present some preliminary results from going in this direction during the coming weeks.

PS: Actually, it’s a funny coincidence that today is not only the 37th Friday of my project but also my 37th birthday. Makes it feel even better to be back on track again. And now I’m off for some cake. See you next Friday!

Questions Unsolved

Time tracking sample: Auction catalogue, entered 26/09/2018
Time tracking sample, 26 November 2018

Saturday, January 12th, 2019 (for Friday No. 15)

I am bit late this week with my state of research entry not because of one of the many good reasons to be brought forward at such an occasion – delays caused by family matters, urgent appointments, events which cannot be rescheduled, etc. – but because of something which happens to me only very rarely. I just did not know what to write. And to be honest, I am not entirely sure if I do now as I am writing this. The problem is that with the first three months of my project year now done, I should now know what to do and what to expect in the remaining nine months’ time, so that if any changes need to be made to the general design, I should recognize this now and implement them. But I am not so sure if that is really the case. So let’s have a quick overview over the project so far and then see what’s to be done.

The state of the project

At first view the project as such is looking quite healthy and running good. First of all I have collected a lot of data for my database (main categories, project start/now):

  • Persons: 530/1.511 (+981)
  • Letters: 720/737 (+17)
  • Publications: 230/1.215 (+985)
  • Institutions: 120/196 (+76)
  • Publishing houses: 224/644 (+420)

Which means that around 2500 entities have been added to the database, ranging from simple person entries only containing name, gender, date and place of birth and death, data source(s), external identifier(s) and confession (if available) to complex publications citing scores of other publications and people. While this may look impressive, the problem is that it is time-consuming, because I could not yet retrieve anything automatically. It all has to be entered manually. On average it now takes me around five minutes to identify a new entity and to add it to the database. Which means it took me around 210 hours of time only to enter my data (that’s five weeks of work). The work necessary to gain the data – finding possibly interesting archival collections, going to archives, reading through sources, taking notes; finding the necessary literature, getting that literature, reading through books and papers, taking notes – is not yet included in this figure. On a very rough estimate, it took me about as much time, perhaps a little less. So that’s another 200 hours, or another five weeks of work. The rest was spent thinking, writing, and talking it through with other people; and suddenly, three months are gone.  

As to the writing, always the other side of things, it does not look that bad, either. 100 pages are written, all chapters at least begun, so that around a third of the work is done there. The project has already spawned two chapters in edited volumes, one finished, the other still in the process of being written, and I am sure there are some other interesting shorter pieces hidden within the material.

So why am I complaining?

There is a downside to all of this: I am not nearly finished. As the timespan it takes me to identify a new entity and to add it to the database has been fairly stable throughout the last two months (I checked every now and then), I do not suppose I’ll ever become faster than these five minutes per entity. And as the project time is limited, this naturally limits the number of entities I will be able to still add to the database. There are a lot more things to do than just entering data, so that my calculations allow for about 3.000 to 4.000 entities to be added to what I have now; and that needs to be it. If the distribution of data over time would be as I had thought when starting the project – hyperbolically converging towards a relatively low level as approaching the present – this should suffice. But I am no longer sure if that really is the case, because most of my data still stems from the time between 1700 and 1750. The density of references diminishes but slower than I thought; and the epistemic communities within which such references were made turned out to be much more diverse than I thought, which means that the number of entities multiplies as the number of shared persons and publications between individual epistemic communities is lower than originally assumed. So apart from a few samples, from a database point of view I am still stuck in the 18th century but do have the 19th and 20th ahead of me still.

Those episodes from the 19th and 20th centuries I have digested so far I discovered without the help of the database, and for the most part they are not yet entered into it – although already covered in writing – because I wanted to keep at least data collection roughly homogenous to rather have good data for a lesser than unevenly dispersed data for a larger timespan.

What’s the plan?

So I fear that if I go ahead as planned, I just will not be able to finish in time. Which calls for a change of plan. But what kind of change?

There are some possible solutions, but none of those I have thought of so far are satisfactorily. I am not sure that I will have found a good one until next Friday, so it’s going be some anecdotal evidence from the sources again next week. And in time again this time. I hope.