Towers and networks: on the differences between the sciences and the humanities

Having worked for several years in interdisciplinary environments, a thing often heard both from humanists and scholars in STEM disciplines is a kind of amiable incomprehension about the way the other works. It crops up in relation to publishing, in the so-called ‘reproducibility crisis’, and in many other contexts. Here, I propose a pair of metaphors which may go a little way, towards least, at bridging the gap.

One of the questions that was ever-present in the background of my recent little book about edited collections was this: why is it that the edited collection is so prevalent in the humanities and (to a lesser degree) in the social sciences, while it is very rare indeed in the hard sciences? I didn’t address the question directly, as I wanted to explore how the format worked in the places where it was to be found, rather than to account for its absence elsewhere. But it has come to mind once again, prompted by a 2014 post by Patrick Dunleavy of the LSE (which had eluded me until now) on the varying cultures of citation in STEM subjects, when compared the humanities and social sciences. (In short, the critique is that, in general, articles in the latter tend to cite many fewer publications by others; that this is a bad thing – even a failure of ethics – and should be changed.)

The two issues may seem unrelated, but I should like to suggest that the answer to both lies in understanding the structurally different way in which knowledge accumulates in the ‘hard’ sciences as opposed to the more humanistic disciplines. I suggest that the difference may usefully thought about by means of two metaphors for bodies of scholarly literature: the tower, and the network. The two metaphors doubtless oversimplify the matter, but contain nonetheless a real and durable distinction.

Consider for a moment a particular topic in medical research: the making of a vaccine for a virus. All the several teams currently racing to develop a vaccine for COVID-19 will – I strongly suspect – all be fully familiar with the same literature: that published on COVID-19 specifically since the crisis started; on vaccines already developed for other coronaviruses; on the development of vaccines in general. And although these literatures are doubtless very large, they may be visualised as a tower, in which each individual study forms a brick in the wall. The dependence of the scholar on each study is relatively direct – all the bricks are, as it were, load-bearing in the structure. The papers high up in the walls could not have been written without those below them.

Contrast this for a moment with an essay I have just this week completed on the conservative poet and critic C.H. Sisson, and specifically his understanding of the relationship of church and state in 1970s England. To be sure, there is a tower here, but only a small one – a group of half a dozen studies over the past forty years that have touched on the specific question of Sisson’s religious politics. My work builds on, refines, critiques and contradicts these as it proceeds.

But the tower metaphor only takes one so far, and so I propose a new, or at least additional, one, of the network. The essay is also a new node in a vast network of studies on the very broad fields of British religion, politics and the arts in the twentieth century. I’ve been reading in these fields for nearly twenty years, and am sure that my mind has been shaped in ways that I couldn’t possibly remember, by studies that I don’t even recall reading. Sisson was an example of many things: a certain kind of political conservative, a maker of trouble in his parish church, a resident of Kent and then Somerset – he has a local history – a fierce anti-Catholic, a working-class boy made good, a member of a profession (the civil service), a former soldier who had served in colonial India, a philosopher of religious language (of a sort), a poet with a certain set of ideas about how verse should work. Each of these aspects of his life and thought is touched upon in my account. Each of those aspects – English anticatholicism and the ecumenical movement, the local history of Kent, the Anglican parish system, the political history of English conservatism, the effects of the Empire on English thought, the evolution of religious language, the development of English poetry – has its own scholarly literature, some of them vast; much of it I have read, but I can hardly claim to know it all. All I can hope to do is to refer the reader to studies that most closely touch my work, or that exemplify current ways of understanding the period, and that provide ways into those literatures; to create new edges in the network that the reader can follow. To systematically cite them all is simply impractical, and arguably pointless. It is surely this structural difference between different kinds of knowledge that in part accounts for differing citation cultures. I have added a brick to a small tower, and a node to a vast network.

And the connection with the edited collection? Perhaps the reason that the edited collection is so rare in the hard sciences is that it is very difficult for it to contain bricks from the same tower. If bricks are to be laid on top of each other – to cite and develop each other – they can’t very easily be published simultaneously.

However, the edited collection format is ideal for the publication of a group of new nodes in a network, since they are not so dependent on each other. My study on Sisson is (I hope) to form part of a collection on his work, alongside essays on his verse, and on his translations of the Latin classics. I cite one of the other chapters as providing a complementary point of view; I could have written my own chapter without it, but they are now in relation with each other. With some modification, however, my essay might have equally well have found a home in a recent edited collection on the Church of England and politics, or in another on the developing relationship between the Church of England and the arts. The three volumes (two real, the last still just an idea in my mind) are each a snapshot of different sets of nodes in the greater network.

Like all metaphors, my tower and my network are a simplification; all fields will manifest elements of both patterns. And there are also structural and economic reasons why the literatures in different fields tend more toward either tower or network, which I don’t propose to go into here. But I would assert that literatures in the hard sciences in general are structured more like the tower, and those in the humanities as a network. If having such metaphors to hand does anything to lessen some of the mutual incomprehension between disciplines, then I shall have achieved my aim.

Just how important is the monograph for history ?

Back in January I posted about the visibility or otherwise of collections of edited essays, suggesting that whilst this form of publication may be as good as invisible in the sciences, we need to view the humanities (or at least history) differently.

In reply Dorothy Bishop (@deevybee) gave some intriguing numbers from an analysis of her citation patterns. In a recent review article in neuropsychology, she had cited 84 journal articles, 3 chapters, and one book. What’s more, she finds herself now very reluctant to cite anything that isn’t available online (which includes most books and edited collections.)

By way of a counter-example, I here extend the analysis of my own citation patterns that I began in that post. (For details of the data, read it here.) Here I’m interested in the balance between journal articles, books, and edited collections.

There has been much debate about finding a viable business model for Open Access monograph publishing in the humanities. Anecdotal evidence abounds, which chimes with my experience, that the ‘big book’ remains the Gold Standard for senior scholars; the once-in-a-decade intervention that changes the game. Lower down the food chain, the perception still rules that it is impossible for a young scholar to secure the vital first academic job without the book-of-the-thesis.

So much, so familiar. But there remains a real lack of data to back up this near-universal intuition, and to establish whether readers think as highly of the monograph as authors, publishers and research assessors think they do. And so, I looked at my own citation behaviour over the last few years, to answer the questions: how much do I cite monographs, as opposed to journal articles or papers in edited collections, and how old are those books when I cite them ?

The answer to the first of these even I found surprising. Here are the numbers:

Category Proportion of citations (%)
Books 59.4
Edited collections (and individual articles therein) 23.4
Journal articles 11.7
Theses 2.0
Other 3.5

I had expected the monograph to loom large, but not to the extent revealed in these figures. I was also very surprised that I cited more chapters in edited collections than articles in journals, which were an amazingly low proportion of the total.

It’s also interesting just how old some of these monographs I cite are. The mean age of a monograph (at the time I cited it) was 17.5 years, with a median of 14.5. Some, in specialist areas, are still current after 40 years. This would suggest that any embargo-based scheme of Green OA for monographs would have to include very long embargoes indeed to satisfy publishers, which suggests that Gold would seem the way to go.

If you’ve enjoyed this post, why not support the blog on Patreon?