Digital archaeology in the web of links: reconstructing a late Nineties web sphere

At the moment I have a chapter contribution to a book of essays working its way through the publication process. The abstract is below; I’d be very happy to share the paper privately, if people would care to contact me.

A shortened version of the paper is available as a video presentation.

Abstract
One unit of analysis within the archived Web is the ‘web sphere’, a body of material from different hosts that is related in some meaningful sense (following, broadly, the definition coined by Niels Brügger). This chapter outlines a method of reconstructing such a web sphere from the late 1990s, that of conservative British Christians as they interacted with each other and with others in the United States in relation to issues of morality, domestic and international politics, law and the prophetic interpretation of world events.

Using an iterative method of interrogation of the graph of links for the archived UK web, it shows the potential for the reconstruction of a web sphere from what is in effect an archive that has a finding aid, but one with only classmarks and without descriptions. It also demonstrates the kind of multi-source investigation necessary to uncover the archaeology of the early Web. Big data and small, printed sources, the traces of previous Web archiving efforts (even when unsuccessful), and echoes in the scholarly record itself: all these come into play.

I also propose a conceptual division of Brügger’s web spheres into two kinds, ‘hard’ and ‘soft’, as distinguished by the ease with which its boundaries can be identified, and the speed with which they change.

On digitisation and the visibility of historic journals

Here follows a tale of two journals, a cautionary tale of the degree to which the historical record is conditioned by the interaction of technology and the economics of publishing.

Firstly, the journal Theology, perhaps the leading general theology journal in the UK. It was founded in 1920, published by the Society for Promoting Christian Knowledge (SPCK), already a leading publisher of books with a particular focus on the Church of England. Although its tone and content changed over time, it has always tended to provide a forum for the publication of theological writing of a breadth of concern that would interest both professional theologians and clerical and lay readers. In it one finds work on the perennial themes of the discipline alongside writing that reflected on the issues of the day, as the Anglican Communion encountered radical theological change and the pressing practical issues raised by the ecumenical movement.

Secondly, the Church Quarterly Review. Though it began life in 1875 as a privately published journal for one party within the Church of England, the CQR became a more general journal, and it too was published by SPCK from 1920. It occupied a similar space to Theology, with substantial articles aimed at both professional theologians and the wider church, and on issues old and contemporary. From the Anglican scholar Eric Mascall (one of my particular preoccupations), the CQR carried articles on topics as varied as the Eucharist, the prospects of reunion with the Church of Scotland, and the impact of the Second Vatican Council, along with dozens of book reviews. (His work also appeared regularly in Theology). But where Theology survives to the present, the CQR does not. In 1968, the journal merged with the London Quarterly and Holborn Review (a Methodist title), but the resulting Church Quarterly ran only until 1971 and was not succeeded by another title.

Although Theology is published by SPCK, its online distribution was taken over in 2011 by SAGE Publications, and the entire back run has been digitised and made available via the SAGE site. As such, scholars may now access complete metadata and the full text of the journal back to its inception. By contrast, the CQR has no public online presence whatever. Unsurprisingly, a defunct journal held little attraction for potential buyers in the great consolidation of online journal publishing of the last twenty years. And, although several SAGE journals are included in JSTOR, Theology (and other SPCK titles) are not. As such, the CQR was not swept up in retrospective digitisation as other defunct titles from publishers involved in JSTOR have been. As it is, to read the CQR I must trouble the staff at my nearest university library to walk across to a store in a separate building and fetch the volumes for me.

There is, I think, an issue here that sits in the intersection of other questions of technology and practice which are better known. It is abundantly clear that current (or very recent) issues of journals that are available online have an advantage over those available only in print, and that the advantage is compounded when the journal is available Open Access. There is now also a great deal of stimulating reflection on the impact of digitised historic sources on historical practice. Within that, it has been observed that the digitisation of newspapers such as The Times earlier than other, equally prominent national newspapers risked skewing readers’ attention towards one source at the expense of another. Despite scholars’ best intentions – of leaving no stone unturned to get to the truth, no matter how heavy the stone – it is at least plausible that more easily accessible sources will be privileged. And the cases of Theology and the CQR suggest that the same might be true in certain fields of modern intellectual history, as the back issues of some current journals are digitised as a byproduct of current needs and others are not. That process of digitisation has tended to favour journals that survive over those that do not, and (in the case of JSTOR), defunct titles seem to stand a better chance if they were absorbed by one that survives.

Of course, it may well be that the CQR is in fact a less significant journal for twentieth century religious history than is Theology. But historical matters become perceived as significant partly as a result of the attention they are paid. It is at least possible that the relative ease of access to Theology will in itself (over time) give it a significance greater than the CQR by a kind of default. If this pattern is repeated in other areas of twentieth century intellectual history, then it perhaps deserves more attention than it has received so far.

References

Josef L. Altholz, ‘The Church Quarterly Review, 1875-1900: a marked file and other sources’, Victorian Periodicals Review 17 (1.2), 1984, 52-7.

Adrian Bingham, ‘The digitization of newspaper archives: opportunities and challenges for historians’, Twentieth Century British History 21(2), 2010, 225-31.

Lara Putnam, ‘The trans-national and the text-searchable: digitised sources and the shadows they cast’, American Historical Review 121(2), 2016, 376-402.

SAGE publications, Press release from 2010 on the digitisation of Theology

Understanding national Web domains

Yesterday I was delighted to find in the mail my copy of an important new book of essays: The Historical Web and Digital Humanities: the case of national Web domains. It is published by Routledge and edited by Niels Brügger and Ditte Laursen.

I say it is important because it investigates for the first time a particular issue that is of immediate practical concerns for two quite distinct groups. The first – Web archivists in the world’s national libraries, and particularly those who work within a legal deposit framework – have sometimes to define and then certainly to work within a definition of the ‘national’ Web, and to understand how much of it they are able to archive. As the volume amply demonstrates, that task of definition is not straightforward, and has been dealt with in widely varying ways.

Outside the small but growing community of Web historians, there are many others (not least contemporary historians) who are not primarily interested in the Web itself, but in what a study of it can tell us about everything else. And the definition of the nation, of a shared but bounded space in which a political community speaks together, is the kind of question which has exercised historians of many periods and of other ‘new’ media. As I wrote in my own chapter:

The advent of the web presents historians with a new and somewhat perplexing question: where is it? What does it mean to think of the web in spatial and quasi-geographic terms? How may we write national histories of the web? Where did a particular website ‘live’? Of where was it a resident or citizen, so to speak?

The volume is important, too, because it explicitly tries to connect Web history with the larger field of digital humanities, where hitherto the two fields have been in only the loosest contact (rather to my surprise, I might add.) It is good to see the volume appear in Routledge’s series on Digital Research in the Arts and Humanities, which also carries work in more ‘traditional’ digital humanities areas.

Finally, the volume marks an important moment in the development of the discipline of Web history. Previous collections (in which my own work also appeared), all of them crucial in their way, have have been more specifically methodological in focus, and have been designed to make the case for the importance and the integrity of the discipline. Although each chapter made a contribution to its own particular field, those previous volumes did not contribute as a group to particular questions of history, or religious studies, or sociology. (See, for instance The Web as History (2017) or Web25 (2017), and the Sage Handbook of Web History (2018). This volume is the first for several years which speaks to a substantive issue of politics, history and sociology, as well as to archival science and the methodology of studying the archived Web.

The chapters fall into three sections: collecting and preserving national domains; methodological issues, and results and dissemination. I won’t try to summarise them here, save to say that each group of readers – archivists and scholars – should read each section, since their concerns overlap. As I’ve argued elsewhere, scholars need to understand more than they do about how archives come into existence, and (in this case) about the administrative histories of particular ccTLDs. Archivists will similarly gain a great deal from the discussions of method and dissemination in the second two parts, since those questions go to the heart of both archiving policy and the design of effective systems for discovery, playback and analysis of the archived Web.

Part One: Collecting and preserving a national Web domain
Kees Teszelszky on ‘reconstructing and saving the Dutch national web using historical methods’.
Sally Chambers, Peter Mechant, & Friedel Geeraert, on the PROMISE project in Belgium: ‘Towards a national web archive in a federated country’.
Ian Milligan and Tom Smyth on the Canadian .ca domain, and studying the web ‘in the shadow of Uncle Sam’.
Helen Hockx-Yu, Ditte Laursen, & Daniel Gomes on the curious case of the .eu domain.

Part Two: Methodological challenges
Jane Winters on the many archives of the UK web space.
Anat Ben-David on Palestine, Kosovo and the quest of national self-determination on the fringe of the Web.
My own chapter on Northern Ireland and the limitations of the ccTLD as proxy for the nation.
Niels Brügger, Ditte Laursen, & Janne Nielsen on establishing a corpus of the Danish web.

Part Three: Results and dissemination
Valérie Schafer explores the French web of the 1990s.
Rebecca Kahn on locating a national museum online (the British Museum).
Niels Brügger proposes a way towards the creation of a national web trend index.

Where is the national Web, exactly? A case study

[A summary of my chapter in The Historical Web and Digital Humanities. The case of national web domains, edited by Niels Brügger and Ditte Laursen.
The full title is ‘Understanding the limitations of the ccTLD as a proxy for the national web: lessons from cross-border religion in the Northern Irish web sphere’
See also the full text (PDF)]

The writing of modern history has often depended on a stable idea of the state; on the idea that persons have some form of citizenship, a legal identification with a political unit. Even if they may hold more than one, each citizenship may stand on its own without legal ambiguity. Another fundamental assumption is that geographical space (at least on land) can usually be clearly divided into units under unified and monopolistic systems of law and government. To elaborate an insight of Max Weber, in order for a state successfully to enforce a monopoly on the use of violence, it must first know where its boundaries are.

Scholars have also been interested in the interactions between states and their peoples across borders, but still (by and large) supposing a fixity in those states at any one point in time. Studies of migration presuppose a point of origin and a point of arrival. Printed publications may circulate freely, but their publication is still governed by a national legal framework; something similar may be said of television and other broadcast media.

The advent of the web presents historians with a new and somewhat perplexing question: where is it? What does it mean to think of the web in spatial and quasi-geographic terms? How may we write national histories of the web? Where did a particular website ‘live’? Of where was it a resident or citizen, so to speak?

In most cases, the task of defining a national web domain has begun with one or more country code top-level domains (ccTLD) even if it has not ended with them. Here I examine the nature of the .uk ccTLD as a proxy for the UK web by means of a case study of the web estate of the Christian churches in Northern Ireland.

The society of Northern Ireland is marked by an interlinking of religious and national identity, which may be unique in Europe if not in the world. The chapter uses publicly available data, and including that provided by the British Library, to reconstruct the link relationships between churches in Northern Ireland, examining the regional, national, and cross-border relationships that they imply.

Due to its very particular religious and political history, Northern Irish society has been characterised by an exceptional sensitivity to symbols, to history, and to place. How far has that sensitivity to space and symbol been transferred online? Amongst the churches, Catholic and Protestant, in a province where the symbols of national identity have such prominence, does the location of a website within or outside the .uk domain carry any symbolic weight? Might those churches most associated with unionism be more likely to register in the UK ccTLD than Roman Catholic churches?

Based on the patterns of domain registration for the churches of Northern Ireland in 2015 and 2016, it would seem that Roman Catholic congregations were likely to register domains outside the UK, a finding broadly in line with the initial hypothesis. However, the converse – in relation to the Protestant churches – is not borne out; no particular prioritisation of registration within the UK ccTLD is evident in the data. Both conclusions point to important areas of future research on the nature of national webs, and the limitations of the ccTLD as a proxy for them. If organisations that might be expected to want their web estate to reside within a particular national domain do not in fact register their domains there, it suggests that the ‘gravitational pull’ of the ccTLD is weaker than might be supposed.

The second half of the chapter takes the case of one of the Protestant denominations in Ireland in order to investigate the mapping (or lack of it) between the nation and the ccTLD. It recreates the networks of links between individual Baptist churches on both sides of the border, and asks: are these link networks influenced by the fact of the ccTLD, or are there more geographic and cultural factors in play that determine their shape? It is based on an analysis of the .uk link graph for the period 1996-2010.

I conclude that although less than half of the Baptist web in Northern Ireland is registered in the UK ccTLD, the links between churches show in fact a very tight geographic concentration on the domains of churches in the eastern counties of Antrim, Down and (to a lesser extent) Armagh. Detailed local studies are needed to establish why this might be the case, although some lines of enquiry might be advanced. Is this a representation of a wider divide between rural and urban churches, or a reflection of the greater resources or perceived influence of churches in certain areas, particularly Belfast? Or is the prominence of certain individual churches merely the product of their particular local circumstances and understanding of their role? For whatever reason, the link graph shows little sign of sentiment regarding the common identity of all the Baptist churches in Northern Ireland.

These churches are linked together in a single organisation, the Association of Baptist Churches in Ireland: what evidence is there of link networks in the archived web that might reflect a sense of an all-Ireland identity? Approximately a quarter of Irish Baptist congregations are located in the Republic. What of the links from churches in the north to those in the south? The link graph connects only four Northern Irish congregations to twelve in the Republic, a very small proportion. Little all-Irish sentiment is to be detected in the northern Irish Baptist web.

Why might that be? Is the weakness of link connections between north and south characteristic of all churches in Northern Ireland, or only the Protestant churches, or is it unique to the Baptists? Is the network particularly weak in the Baptist case because of the relative weakness of its national organisational structure? These questions could in part be answered by the application of the approach used here to the web estate of the other churches.

More generally, a history of the web is required that also asks what it is that causes the human actors in control of websites to link to others. A substantial project of oral history interviews and fine-grained examination of individual websites is needed to understand the communicative strategies organisations adopt and their evolution over time. That said, I show what may be observed at a distance with a new kind of data. Macro-level analysis of the web such as this offers an additional tool for historians and other scholars to deploy alongside their existing methods.

The chapter has also pointed out a particular challenge that historians and analysts of national webs face. In the Baptist case, a network of links that is very tightly geographically concentrated is at the same time spread across four different TLDs. Studies of particular web spheres such as this are so far very few. However, if the kind of pattern I have outlined is at all typical of other web spheres, it suggests that for web archivists and scholars alike the ccTLD is a weak proxy indeed for the national web.

In addition, it brings into sharp relief one of the structural disadvantages of the division of world web archiving activities into national programmes. Though many web archives collect national material beyond their ccTLD, no organisation has any statutory responsibility to archive the non-geographic domains such as .com and .org as a whole. Unless and until it becomes possible to access web archives on a transnational basis, scholars will continue to work with fragmentary and non-commensurable data from several archives to reconstruct the national web.

Existing Web archives: an orientation

Web archives are fast becoming the fundamental source with which the history of the Web is written. Scholars coming to them for the first time are in need of some orientation, however, since those archives are brought into being by many different organisations for varying purposes and by different means. Their scope and structure also vary widely, as do the means of first locating and then using them.

My chapter in the new Sage Handbook of Web History aims to provide just that orientation.

It begins with a brief historical sketch of the development of Web archiving over the last 20 years, which I discussed at greater length here. It then moves on to outline the different means by which these archives are created, and what implications those differences have for how they must be interpreted. It outlines the varied kinds of collections in existence, and the different questions of method that this variety raises for scholars. Finally, it details the means by which scholars may first locate archived Web content, and (once located) how it may be used.

Along the way, it raises several points of necessary critical engagement for Web historians regarding the archived Web as a new class of primary source. Some of these issues have their analogues in print, manuscript or other sources; a scholar needs to understand who produced an object, whether it be a book, a manuscript, a painting or a PDF. But some of the issues presented here are peculiar to the archived Web, and must be thought through afresh.

The technologies that are used to create archived Web resources fundamentally shape those resources, and so understanding those technologies is a prerequisite to understanding the archive. Crucial also is an understanding of how the archive is structured: along national lines, by the institution or sector that created the content, by format or by a more general subject.

Finally, users must also understand something of the means by which they discover, search within, view and analyse archived objects, since those means are both relatively new and in a state of flux and development. That thinking will be greatly enabled by close collaboration between scholars and archivists: a partnership of mutual benefit which shows welcome early signs of growth.att

[See also, in the same volume, ‘Religion in Web history‘, my essaying of an agenda for the religious history of the Web.]