History of humanities computing

Context note: this is a self-contained part of the METH4DH course. You can read this to gain an understanding of the historic development of humanities computing. However, if you want to understand how it relates to modern digital humanities, you are suggested to start from the three approaches to methods for digital humanities.

History of humanities computing

The birth of humanities computing is usually traced to the work of the Jesuit priest Roberto Busa. In 1949, Busa pitched to IBM the novel idea of using general purpose computers to create an index of the words appearing in Thomas Aquinas's works (see e.g. this article for a detailed historical account of the project).

At the same time, while Busa did manage to significantly decrease the labour in crafting such an index, his work is also already part of long traditions. To pick two examples, already in 1890, the U.S. Census was recorded on punch cards and tabulated automatically, while already in 1887 T. C. Mendenhall suggested in the then seven year old journal Science that quantitative analysis of texts (in this case, word length) could be used to determine their authorship. Indeed, indices such as Busa's had also already been compiled, both manually as well as through partial automation.

Thus, the birth of humanities computing as a field can also be seen just arising organically out of already existing (quantitative) approaches to humanities questions, made possible, or at least vastly less labour-intensive by the development of general purpose computers.

In any case, by the mid 1960's, the field was already well defined, as testified by the establishment, in 1966, of the journal "Computers and the Humanities" (in 2005 renamed to "Language Resources and Evaluation"). In the first volume of the journal, in an announcement on the third ACLS Program for funding Computer Studies in the Humanities, one even finds the following statement:

The Council has not considered, even if funds continue to be available, when it will bring its current program to a close. One can explore for just so long. Then it will seem that nothing more or nothing of special interest is going to turn up, and that better uses for the always limited resources should be found.

The ACLS program for computer studies in the humanities: Notes on computers and the humanities, Computers and the Humanities, September 1966, Volume 1, Issue 1, page 9 (in context)

Particularly singled out as "already solved" are stylistic literary analyses and quantitative historical studies:

For instance, many excellent plans to study a writing style, or to compare two styles, or even to study "influences" propose a use of computers which is routine, and too routine for an award in a program aimed at feeling for what new uses of computers there might be. Even many of those historical studies which collect and order vast quantities of information, say, about the members of an assembly or a voting group are now so settled in their method that, again, only if a proposal is of special moment is there much chance that it will receive support.

— The ACLS program for computer studies in the humanities: Notes on computers and the humanities, Computers and the Humanities, September 1966, Volume 1, Issue 1, page 8 (in context)

Whatever happened to humanities computing?

At the same time, the texts in the first volume of the Computers and the Humanities journal reflect hopes on how computers could and should transform the humanities as a whole. For example, the same announcement states:

We don't know what shapes studies in the humanities will take in ten or twenty years. Even our ideas about kinds of studies are changing now. The overlap of fields and departments of study seems only a surface over deeper stirrings. .... We don't know, of course, what the further subtleties will be as formalities are probed still more, how computers will incorporate them, or how the thought and instrument will enter and affect the humanities. For a long while, perhaps, the traditional or institutional divisions in the humanities will stay steady, so that within the established departments, along with other studies, there will also be some formalistic ones — studies in literature, government or history, for example, done with the aid of computers. It might happen, though, that our divisions themselves will change, that we will have departments of formal or computer studies, and that in these new departments there will be studies of literary and historical and other materials.

— The ACLS program for computer studies in the humanities: Notes on computers and the humanities, Computers and the Humanities, September 1966, Volume 1, Issue 1, pages 1–2 (in context)

Now, over 50 years later, the situation and expectations are seemingly exactly the same with regard to digital humanities. So, what went wrong? Why didn't the promised revolution come? Why would this time be any different?

Assignment

Optional: form a group with people from a similar disciplinary background

Read the following historical texts:

  1. From the 1966 inaugural issue of "Computers and the Humanities", read the ACLS announcement in full.

  2. From the second 1966 issue of "Computers and the Humanities", quickly browse through this article on "Computer Programs Designed to Solve Humanistic Problems"

  3. From the third, January 1967 issue of "Computers and the Humanities", read this article on a new "Center for Computer-Oriented Research in the Humanities and Social Sciences" at the University of Pennsylvania

  4. Then, browse this article from 1983 reflecting on how quantitative history had developed up to that point.

  5. Finally, read this article from 2012.

After reading the given articles, do your own searches and dig up the history of quantitative or computational approaches in your own field, and any analyses on how they ended up relating to the core field. You may also want to consult this timeline on the development of computers in general for context.

After going through your sources, reflect on the general questions above about the development or non-development of humanities computing, 1) in general and 2) from the perspective of your own field.

To report on your work write a couple of paragraphs of text on what you've discovered, also citing your sources. Post the result on the #dhhistory channel.

My own take is that on the one hand, there really is nothing new under the sun, apart from continuous incremental improvement. All through history, methodologists have developed better analytical, mathematical, statistical and computational methods for exposing and validating phenomena in data, while others have sought to apply them to whatever material and questions they have had. Which is all as it should be.

Also similarly then as now, these practitioners were not really the ones painting juxtapositions, revolutions and the doom and gloom of traditional scholarship (for historical instances of references to these fears, see e.g. these highlights). Instead, right from the start, while they indeed were excited about the prospects offered by computing, they were also mindful of its limitations:

To imagine future uses for computers in the humanities, we have to remind ourselves of the notions that are central to computer operation, and to think how they might be interpreted even more imaginatively than they have been up till now. One of these notions is the notion of the countable item. A countable item can have a place in an array, be isolated, be combined or removed from combination, and compared with other items. The great advances which have so far been made with computers have been in those fields where we find countable items or have ready substitutes for them. The real or seeming extraneousness of computer studies for the humanities is owed to the fact that, in the humanities, what are most important are, if items at all, items that we can't count, or can count only most artificially. We know, for example, how little definite we mean in saying that we have two or three ideas, that there are four themes in a play, or that there were this or that number of historical events. Our "counting" is not the counting of items that were somehow there separate, waiting to be pointed out; it is a "counting" in which judgments themselves mark out what come to be the items that we count. Apart from the judgments, there are no separate items. Therefore, no technique of counting such items so as to yield, for the first time, a judgment or a summary is possible at all. But, granting that this sort of limitation is inescapable, computers could, it seems, still come to have a more vital use in the humanities than we have seen so far. The first, though the hardest, step in having this come about is simply to find or make countable surrogates for non-countable items.

— The ACLS program for computer studies in the humanities: Notes on computers and the humanities, Computers and the Humanities, September 1966, Volume 1, Issue 1, page 4 (in context)

At the same time, for reasons I cannot really explain, this basic interdisciplinary interaction did often end up being cut off in one way or another. Regarding linguistics, computational linguistics branched off from humanities computing to strike out on its own, later further devolving on the one hand into language technology which developed methods seemingly without notion to their use, and on the other hand into corpus linguistics, which stuck to established tools, measures and processes without much ambition to develop them. In social sciences similarly, quantitative statistical explorations were confined into certain subfields, among them economics and population statistics.

Further muddying the waters, with the coming of personal computers in the 1980's and the Internet in the 1990's, what was meant by humanities computing and later digital humanities became muddled, cutting off our view to this incremental progress of interdisciplinary interaction. All of a sudden, computers were no longer all about computation, but instead aids in all kinds of work, be it using word processors to edit texts, using email to communicate or publishing digital critical editions on the web.

When does then work fall under the heading of humanities computing, and when is the use of a computer incidental? Are you a digital humanist if you use Excel, Email or Word to do your work?

For the contents of this course, the answer will be that whenever you are utilising a computerised process to yield new understanding on questions of import for the humanities, you are engaging in the digital humanities this course is meant to support. Thus, from the above, at least Excel starts counting immediately after you move from merely entering values in a spreadsheet to summing them across dimensions or projecting them as graphs to better understand them. On the other hand, publishing digital critical editions on the web does not count. But the act of using a computer program to more easily compare versions in the building of such a critical edition does - as does utilising a well designed interface to a multilayered critical edition to yield new insight on the textual tradition. As a final example coming back to the start, Busa's original Index Thomisticus in itself also doesn't count, but all research done using it as a tool does.

Note also that this delineation rules out all the advances and development in computer science that were required in manifesting the Index Thomisticus, or any new computational tool. These are certainly part of humanities computing more broadly (and exactly my personal interest), but just not a part of this course, which is aimed at students coming in from a humanities background.