A BIG question for history (of economic thought)

I thought that answering Ben’s excellent and stimulating question (in the previous post just above this one) could justify the opening of a new post.

His question was, are there any new BIG questions for history? I’d say that there is at least one great challenge for history: the entry into “e-history”. By that, I mean the renewing of methods and research questions in history, implied by the recent digital data floods: internet, digital personal documents, online journals, transactional data (e.g., mobile communications, data on financial transfers, etc.).

Wait, is it really a BIG question, or simply a new set of tools?

Well, let’s consider what happened to sociologists, when the realization came that digital data could be of importance to their field. The following quote will make this post awfully long, but it gives such a dramatic view on the issue that I can’t resist:

Our sense of this impending crisis has crept up upon us as we have gone about our work in recent years. For Savage, an early sign was in 2004 when he attended the ESRC Research Methods festival. With colleagues Gindo Tampubolon and Alan Warde he was enrolled in a session designed to popularize social network methods. He talked about an ESRC-funded research project which mapped the personal connections and ties of members of three voluntary organizations using social network analysis. The project had proved time consuming and intensive.

A lot of time had been spent finding three organizations prepared to participate, a postal questionnaire had been sent to 320 members in total, with a very high response rate. Many members had been interviewed face-to-face to ask detailed questions about their social networks. Thirty life histories had been conducted. The resulting intensive study of the members’ social ties was amongst the most detailed ever carried out in the UK (see Ray et al., 2003; Warde et al., 2005). During the Festival Savage talked to other participants interested in social network methods. It turned out that one enthusiast was not an academic but worked in a research unit attached to a leading telecommunications company. When asked what data he used for his social network studies, he shyly replied that he had the entire records of every phone call made on his system over several years, amounting to several billion ties. This is data which dwarves anything that an academic social scientist could garner. Crucially, it was data that did not require a special effort to collect, but was the digital by-product of the routine operations of a large capitalist institution. It is also private data to which most academics have no access. To be sure, we can cavil about its limits. It does not tell us what the callers actually talked about. We can emphasize our superior reflexivity, theoretical sophistication, or critical edge. Fair enough – up to a point. Yet the danger is that this response involves taking refuge in the reassurance of our own internal world, our own assumed abilities to be more ‘sophisticated’, and thereby we chose to ignore the huge swathes of ‘social data’ that now proliferate.

Sociology is simply not the same when you rely on panels and surveys from a small sample of a population, or if you rely instead on massive and exhaustive records of the actual social transactions performed by all individuals of a population (this now classic article from 2007, from which the quote is taken, sounded the alarm first). Will digital data transform history as dramatically?

Now, transactional data might not be of prime importance for historians of economic thought. But other forms of digital data definitely are. Many living economists will donate their archives not in the form of 200+ linear feet of boxes, but simply in the form of a one kilo external hard drive containing a few terabytes of data.

The method that we know and practice – go slowly and methodically through each file of a box in the archives – might still work (simply demanding even more patience, I suppose), but the volume of data will invite to different forms of historical analysis as well.

The future look of Duke's Economists's Papers Project?

Indeed, these archives, we might expect, will make new kinds of material accessible: more complete series of drafts leading to the published version of a given work, with preserved annotations from all co-authors in the case they used the capabilities now afforded by most text editors. The underlying data in case of an empirical work, full correspondence with co-authors, detailed chronology / calendar of the work on a publication, all will become potentially available to the historian. But also drafts of grant applications, full bibliographies chronologically indexed by date of their entry in the database (now a common practice, which will be so useful for historians!), lecture materials, administrative records, … should become routinely available to historians, very soon.

The archivist receiving Paul Krugman's complete works, empirical data, personal files, and correspondence, one day.

Correspondence, in particular, will be much better documented. Whereas we had to find two economists close enough professionally, but far enough geographically, to expect a letter correspondence between the two, now the communication by email will surely make available a much richer record of exchanges, even between two colleagues working on the same floor.
Same for archives from universities and departments, institutes, governmental organizations. It will surely encourage the development of mid-range studies, at mid-way between the “big picture” history and the biographical approach.

BIG questions will (actually, are) now on the table:

– geography of science (through geolocalization data and content analysis)
=> can we observe a phenomena of Americanization of economics?
=> is it possible to map the coming of Keynesianism to the US?
=> can we represent the impact of the European refugees in American universities from the 1930s onwards?
=> can the “travel” of a concept, and its derivations, be tracked in publications?

– interdisciplinarity (through scientometrics and content analysis)
=> what was the dynamics of of interdisciplinary contacts between economics and the rest of social and natural sciences? At the individual level? At the organizational level?
=> was the interdisciplinarity character of a research center or a deparment a fact, and in which sense?

– production of knowledge (through many methods, including network analysis and web mining)
=> what are the invisible colleges in the history of economics? Offline, and online?
=> trace back the precise genesis of a land-marking paper in a given field
=> trace back a very rich biographical portrait of a given scientist

One might object that these questions are not new. True, but a new question is most often a classic one, asked in a new way. Look at the abstract of the communication by Esther Duflo, cited by Ben in his post: her big question for development economics is all about suggesting different empirical methods for classic questions. Same for history of economics and the new methods from e-science, I’d say.
I do not pretend that the questions and methodological innovations brought by e-science should replace our existing practices. Neither do I claim that they will develop easily. They have begun to raise serious issues for librarians and curators, be it the delineation between private and professional documents contained in the digital files donated by a scholar, or the question of the preservation of digital files coded in file formats no longer used after a few years or decades. Even the “simple” question of the communication of these documents to the visiting or online historian appears to be a difficult one in practice. But they will have an impact – each of us can realize that – users as we are of emails, text editors, blogs, etc.
So what? If we agree on the above, then one important consequence emerges, I’d argue. PhD students, and all of us, ideally, in history (of economic thought, and else) would benefit from a training in the methods to address this new empirical material. Yes, it means learning a bit of programming and computing science. A New Year’s resolution for 2011?


10 thoughts on “A BIG question for history (of economic thought)

  1. Your analysis largely rely on your assumption that “Many living economists will donate their archives….in the form of a one kilo external hard drive containing a few terabytes of data.”
    And sitting in the Duke archive reading room a few weeks ago, browsing the well ordered files of some great economists presumably put together and assembled by their administrative assistant (who sometimes played an explicit and active role, and in the case of Solow and Samuelson) I was pondering this very assumption.

    They will donate… ? What exactly? Your analysis requires as a preliminary a bit of sociology of e-behavior. Great economists are not unmistakably great at first, even not to themselves, and when they become so, do they use administrative assistants to classify their huge correspondence and research material now that these are electronic files?

    If they sort it out alone on a daily basis, do they keep everything? Of course we will certainly get wonderful “grant application” folders, but what about correspondence including some economic controversies, casual account of economic events, or administrative/academic/ educational deliberations and debates, recommendations, etc? Will the electronic material retained by economists be the same as the paper material 50 years ago? It may be simpler to create new folders to classify all this, but it’s also much simpler to just “delete” if the material does not seem important to the subject.

  2. Oh yes.

    Whereas paper and pencil documents could be stored in garages and attics, dusting and being forgotten until being ultimately donated (or binned!), now the delete button make it much easier to selectively retain and forget, or “preemptively manage” the archives by their authors.

    So yes, a “preliminary bit of sociology of e-behavior” is requisite. As I see it, not just the “e-behavior” of economists should be examined and not taken for granted, but every aspects of this new regime of data.

    There are thousands of questions to be examined, that’s so exciting! One of them, related to your comment: if the cost of collecting and keeping a scientist’ archives drops dramatically, thanks to savings of space in particular, does it mean that we can now start collecting much more archives – and not just the “rich and famous” of the profession?

    In my fantasies, a center like Duke or else would set up an agreement with Harvard (or else), inviting each of Harvard’s faculties to donate a copy of their professional archives when their position terminates (with a buffer period for access by researchers, and all the other necessary conditions). There are so many possibilities!

  3. I think there is a lot of exciting things to come out of this change, and I really like the idea of partnerships between a library and a university faculty to exchange digital copies of their archives when tenure finishes. That could be a really exciting long term project.

    But also, there is an astounding amount of data out there today, which there has always been, but now it is all a lot easier – cheaper – to connect, and as a result a lot of it is already connected, allowing new and exciting research to happen. (I am thinking here of the telecoms researcher Clement mentioned. Twenty years ago, British Telecom or any national carrier had all that data, but they would have never connected it – now we can, and do as a matter of course) Of course, with this ‘data deluge’ I’d expect us to see a lot of spurious empirical results – like the length of skirts vis-a-vis the stock market – but one has to take the good with the bad, or at least the good with the funny.

  4. Two sobering comments in the midst of such new year’s enthusiasm:
    – IT technology transforms the problem of history, but does not answer it no more than more traditional technique. It transform it because the problem is now how to process and make sense of so much information (from whatever sources, digital or not), whereas the traditional problem was more like where can I find useful data on such and such historical problem? There is an excitement of seeing loads and loads of data that were non-existent or buried coming in (and you can sense it in Clément’s post), but what you may do if you had nearly infinite resources and complete liberty to use them is not what you can do in a normal academic lifetime with a normal – meaning – research budget. As I see it (and your quote makes it clear), whether you have a lot of resources but you are very constrained to use them in a peculiar way because you are part of a money making business (or finance by it), or you have little resources but much more liberty to define your own research program as in the academia. I see no easy way to get around this dilemma.
    – My second point is that all these new data and old data in new clothes (like digitized texts) are driving you to certain questions and certain tools to answer them and again the quote is quite clear on this: to a large extent the rise of network analysis is a produce of IT technology (try to it with a pencil and a slide rule and you will see what I mean), just as cliometrics would not be thinkable without computers. Whether it is a good thing or not for history is not so clear: are we really sure that the number of links is always or even generally a good proxy for the importance of a relationship? I see my neighbors or department colleagues and interact with them quite often, I am not sure what that really means on a deeper level. But I can cite a couple of books or persons that had a deep impression on me, even if the interactions I had with them that can be measured are meager from a quantitative point of view.

  5. I see your point Loic.

    I tried in my post not to discuss the issue in an “either / or” fashion. The question is not, “history as we know it vs. history with computers only.” The two should be of course accommodated. Again, an either / or debate would lead us in sterile debates about “what is more fruitful, quantitative or qualitative analysis?” – this leads nowhere in my view.

    To take the example of the sociologists and their social network data, the availability of mobile communication networks with billions of ties, but without any info on the content of the communications, will not replace carefully drawn case studies of networks with a few dozens of nodes. Of course not.

    The point of my post is rather, when we look at our current practices, publications, conferences, and technologies, historians of ideas seem not to have given much consideration to digital empirical data. This will change, I believe.

    On money matters. Things are changing for the better, I would say. The EU funding programs, and also local national research agencies, are very much interested in large scale, international (and by consequence very expensive, in the order of millions of euros) projects for the building of digital historical databases. An example I recently heard of was a European project awarded to build a single framework to digitize and host demographic data from worldwide sources since their origins, properly annotated and fully searchable for researchers interested in economic history and else. Things are moving!

    1. I do not think that you see my point Clément. Let me try to be clearer.
      I am not making a case for qualitative analysis, but I am saying that those brand new and shiny tools are not neutral: they favors some historical questions over others. One example, why there is so much literature on monetary and financial stuff in cliometrics; simple answer: because record of prices and in particular financial asset prices are the easiest quantitative data available. Does it means that they are much more important than say the price of bread to understand what happened in the past, I do not think so? But at the same time, if I am a young cliometric turk, what I need is data, and fast…
      On money, things are not changing at all. EU funding wants you fund EU related project with EU related teams, if that is not narrowing the perspective… Moreover, they are prefering funding policy-related or at least present-day related issues; if you are working post-war economics that is good for you, if you are not it’s not… Besides my sentiment was that in today’s academic setting, historical issues tend to be side-lined in favor of more useful subjects.

    2. Thanks for the clarification! I think I get it better this time.

      I agree on the first point: a set of tools and techniques brings its new perspectives, but also its dead angles. But this buttresses my claim for plurality: it is by multiplying the methodologies that dead angles are minimized. So e-history should be developed, along with the preservation of existing methods.

      I disagree with some of the points you make on money, although I must admit that I surely don’t have a full view of the existing funding schemes, not or of the fine politics involved. But on the specific points you make:

      – funding of projects related to the EU: no, not necessarily. It is not a formal necessity, and in practice projects without a EU-geographical base or purpose are accepted. And there is always the possibility to craft a project with a European leg, the other leg being elsewhere.

      – EU teams: yes, certainly. You say it narrows the perspective… if you compare it to fully international teams, I guess. I rather have the impression that it expands the perspective by incentivizing researchers to reach out for collaborations outside of their national frontiers.

      – Finally on the historical period and the policy-related issues: yes, you’re certainly right! It has not much to do with e-history, though, and that’s more of the eternal problem of the underfunding of the humanities.

    1. Again, you misread me, you are talking about creating a dead angle, I am saying that it creates strong incentives to read history from a certain angle. Meaning that plurality is more difficult to attain. About your last sentence: there was a time when research in the social sciences was very time-consuming but relatively cheap; we are now in an age where funding (read money or “capital” if you want to portray me as a marxist) makes an important difference because the tools you mentioned (and the people you need to run them) are much more costly than pens and paper, because you have to present your research in conferences, seminars (and you have to fund these travels) if they are to make a significant impact, and this too is costly, etc. In a word, an age where money counts much more that it did. Finally, humanities were much better funded at the end of the 19th and up to WWII than it is now say the historians of sciences…

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s