Nota Bene: This page has been generated from an external URL:
http://www.kcl.ac.uk/humanities/cch/wlm/essays/know/
Copyright belongs to the author(s) of the text.

We would know how we know what we know:

Responding to the computational transformation
of the humanities[1]

Dr. W. McCarty?
Centre for Computing in the Humanities
King's College London
willard.mccarty@kcl.ac.uk

 Es ist schon ein großer und nötiger Beweis der Klugheit oder Einsicht, zu wissen, was man vernünftigerweise fragen solle. Denn, wenn die Frage an sich ungereimt ist, und unnötige Antworten verlangt, so hat sie, außer der BeschÃ? ¤mung dessen, der sie aufwirft, bisweilen noch den Nachteil, den unbehutsamen AnhÃ? ¶rer derselben zu ungereimten Antworten zu verleiten, und den belachenswerten Anblick zu geben, daß einer (wie die Alten sagten) den Bock melkt, der andere ein Sieb unterhält.

Immanuel Kant, Kritik der reinen Vernunft I.2.iii

To know what questions may reasonably be asked is already a great and necessary proof of sagacity and insight. For if a question is absurd in itself and calls for an answer where none is required, it not only brings shame on the propounder of the question, but may betray an incautious listener into absurd answers, thus presenting, as the ancients said, the ludicrous spectacle of one man milking a he-goat and the other holding a sieve underneath.


Abstract. Computing the data of the humanities began just over 50 years ago. Although mechanically aided treatment of these data is considerably older, our understanding of how knowledge in the humanities is affected remains inchoate and rudimentary. Transformation of the epistemology and sociology of this knowledge is proceeding at such a rapid pace, however, that we must figure out quite soon how our institutions might best respond. In this paper, I draw upon my own teaching, service and research over the last decade to construct a model of computing in the humanities as an interdisciplinary field in its own right. I show how this field relates to the existing disciplines, propose a tentative research agenda for it and argue that constituting it as a separate academic unit offers us the best opportunity for understanding and responding to the transformation now in progress. In other words, we must take humanities computing seriously in the best way we know how.

  1. Introduction
  2. Lessons from personal experience
    1. Teaching

    2. Service
    3. Research
      1. From the field of application
      2. From humanities computing
    4. Publication

  3. Humanities computing and the disciplines
    1. Primary disciplines
    2. Secondary disciplines

  1. Research agenda
    1. Mechanical primitives
    2. Encoding
  2. Institutional challenges

    1. Institutional home
    2. Service and the interrelationship of departments
    3. System-wide effects
    4. Instruction

  1. Introduction

    2. "He who replies to words of Doubt", the English poet William Blake wrote, "Doth put the Light of Knowledge out" (Auguries of Innocence 95f). This is not to say that all answers should be evasive, rather that in the humanities the fundamental questions are those without final answers, and that success in our pursuit of them is measured by our ability to formulate better questions. Such is my aim here.

    I am called upon by the Max Planck Gesellschaft specifically to set forth a "vision of scholarship in the humanities which takes up the challenges of the new electronic media and fully exploits their potential… including changes in the disciplinary and administrative structures." [2]. A great honour thus to be asked by one of the world's great research bodies, but also a most difficult assignment, not only because I must not and will not put the light of knowledge out, but also because this transformation of knowledge has so recently begun and our knowledge about it is so inchoate. Evidence for it in the humanities, the Geisteswissenschaften, often goes unobserved because those engaged in the affected research are pursuing specific disciplinary questions and tend not to notice how the computer is altering their work cognitively and methodologically. When they do, their observations tend to be desultory and to be published in the journals and books of the fields within which the primary research questions were framed. Thus isolated, the scattered bits of evidence for the effects of computing tend to be inconclusive or even seem trivial in themselves. [3] Thus the curiously persistent though obviously absurd statement that the computer is "just a tool" – as if the means of knowing had nothing whatever to do with what is known. [4] From my perspective, as an observer and practitioner of humanities computing – that is, the application of the computer to the arts and letters – it is as if the bits and pieces of a potentially coherent activity were going on independently in numerous scattered mountain villages, each isolated from the other, with little communication among them. [5]

    The reports that do get through the mountain passes from village to village might be more credible and useful if they were not mixed up with inflated promotional claims. These have misdirected us, not so much by exaggerating technological progress as by confusing the scholarly value of computing with its engineering successes. As I will argue, it is rather the struggles and failures to which we must look for the source of vision. Although technology progresses by making new things work, we learn, we progress in the Geisteswissenschaften when they don't. The scholarly way ahead is a via negativa. Our greatest institutional challenge is somehow to accommodate creative failure. [6]

    Practical as well as philosophical steps need to be taken. For humanities computing these in logical order are the following:

    1. gathering evidence for the computational transformation of knowledge from the research where it has been and is being manifested;
    2. formulating the intellectual, scientific agenda of humanities computing from this evidence;
    3. making the case for its recognition by the academy on the basis of that agenda; and
    4. realising the result in specific institutional terms.

    It would be utterly impractical to follow this sequence, however, since we need the institutional resources to accomplish the first three tasks properly, just as we need the agenda in order to recognise the relevant evidence. With respect to the humanities, then, we must take shortcuts. I propose to present evidence, formulate an agenda and make a case for humanities computing based on my own experience as a scholar and on observations of what others are doing. At the end of my paper I will outline some of the challenges that I think we face in attempting properly to institutionalise the field.

  2. Lessons from personal experience

    1. Teaching

      I have taught humanities computing for the last 8 years to undergraduate and postgraduate students and to colleagues at Toronto (in Canada), Princeton University (in the U.S.) and now King's College London. [7] The participants' disciplinary diversity has taught me that the only possible academic subject is the methodologies we have in common. By abstracting these from specific, discipline-focused examples we discover the interdisciplinary common ground, even when (perhaps especially when) the discipline in question is not our own.

      Note that neither the technically-orientated "short courses" commonly offered by computing centres nor instruction in computer science focuses on the interaction of computing with the research materials, approaches and questions of the humanities. Neither is concerned directly with the methodological common ground of humanities computing and what we learn from the mechanical modelling of our cultural artefacts. More about this later.

    2. Service

      Much the same realisation of a common ground to humanities research comes by a different route through advising postgraduate students and colleagues across the disciplines. As an academic field, mine is unique for the significant role such collegial service plays in the activities of a department, extending from casual advising to full collaboration in research projects. Collegial service is based on equality of status proceeding from a recognition of applied computing as a subject within the Geisteswissenschaften, but also on the interdependence and cooperation of equal partners in a common research enterprise. [8] I will return to the idea of service later.

    3. Research

      My research for the last decade or more has been located in the intersection of classical philology and computer technology. My observations suggest that this research is in essential respects typical of what happens. [9]

      Specifically, my research has been directed at perhaps the most influential artefact to survive from Greco-Roman antiquity, the Metamorphoses of the Roman poet Ovid – a mythological compendium in 15 books, about 12,000 lines of Latin hexameter. Ovid's poem has not fared well in the hands of literary scholars during the last century: the putative coherence of its numerous, often bafflingly sequenced narratives has proven elusive to the literary-critical quest for unity, frustrating all attempts to gain consensus for how it works as a poem, how its parts fit together. The objective of my research has been to produce a scholarly tool that will give others a systematic, reliable means of working on this problem of coherence and numerous related issues.

      I chose to approach the problem by identifying names – not just proper names and the like but all devices of language by which persons are referenced: all nominal and adjectival words and phrases, all verbs, all pronouns, all personal attributes, to a total of about 60,000 appellatives, approximately 5 per line of poetry. I have identified names computationally by inserting metatextual tags manually into the text. These tags are then sorted, grouped and formatted to produce the reference book in print and electronic form.

      From a scholarly perspective the significant work is entirely in the tagging. To make a long, complex story very short, the defining ontological instability of the Metamorphoses, in which persons fade into and out of existence from first to last, makes tagging a fundamentally problematic act. The entities to be tagged are often ambiguous; tagging requires that all ambiguities be resolved. For a computer-generated reference work to be useful, this must be done in as close to a totally consistent way as possible. Hence a kind of grammar is required, i.e. a reasonable set of rules to impose on the unruly poetic phenomena.

      The paradigmatic question is, what happens when you do this? My experience with the Ovidian text suggests that there are two ways of characterising what happens when one takes – as computing requires – a formal, rule-based approach to any artefact in the humanities. One characterisation is defined by the discipline of application, in my case, classics; the other by the very perspective on research we are attempting to understand, namely humanities computing.

      1. The view from the field of application

        2. In the classicist's view, what happens in my project appears as a kind of translation, from the richly ambiguous language of poetry into the stark simplicity of unambiguous declarative statements (McCarty? 1994). At first glance, the result may seem impossibly poverty-stricken, worse than a waste of time. On more mature consideration, however, two sorts of benefits emerge. The first is pragmatic: if the maker does a careful, consistent job and the user keeps the crudeness of the translation in mind, the resulting tool can be useful, as indeed are the rules devised to construct it. The second, more interesting and consequential benefit is intellectual: what is lost in the translation reveals what is there in the original, which is to say that the discrepancy between what can be explicitly and consistently stated and what cannot, or not yet, leads to the fascinating question of how we know what we know. This is one manifestation of the via negativa I mentioned earlier: intellectual enlightenment coming out of a textual-engineering failure.

        More broadly, application of computing to my particular research question in classics results in a model of the poem I am studying: a manipulable device that is not true but which allows me to probe for an unreachable truth better than has been possible before.

      2. The view from humanities computing

        3. The second perspective on what happens in computer-aided research is determined by the device. This has already entered into my picture as the paradigmatic modelling that is in essence what the computer does for us. In the vague sense of having an idea about something and not taking the idea as the thing, modelling is of course not at all new to the humanities, but the use of an external, physical model is.

        Thus in contrast to the classicist, the computing humanist per se is concerned with the material culture of the scholarship and its effects, of which the classicist's loss is only one. He or she studies the characteristic properties of the source material as data, the computational techniques relevant to these data and how our view of the artefact is changed as a result. These properties are essentially portable and so potentially relevant to many other situations of research across the humanities. By observing what happens to each in its new contexts, the picture we get is continually refined. One might say that the research applications taking place within the various disciplines constitute the computing humanist's laboratory experiments.

        From the perspective of the material culture of humanities scholarship, one must question where the significant advances lie. More or less the same problems that I have described would arise were I using index cards, for example. Indeed, at root my problem is the very old one of categorisation. But without the need to be totally explicit, I would not need to operate on such a minute level of detail, and so would miss that detail. Indeed, I couldn't handle such detail otherwise. Without the means to assemble and so to see inconsistencies across a large number of cases, I would not have the discipline of consistency, its way of thinking, nor its result, a tentative grammar of ontological change for the poem. Without the ability iteratively to try out different schemes in a perfective attempt to reduce the variety to manageable proportions, I would not have a working model of Ovidian naming practice. Without the perfect recall of massive detail, thus forcibly systematised, I would not have the grounded perspective from which to take new sightings on Ovid's vexing poem.

        In short, the contribution of computational approach to humanities scholarship is its systematic formalisation of existing tendencies. [10] It is a new way of dealing with old problems, which up to now has been out of sight simply because we lacked the means of doing what we needed to do quickly and efficiently enough that we would in fact know do it.

    4. Publication

      5. The potential for fundamental change in the communication and publication of humanities research is, as we know, beginning to be realised. It is a vast and complex topic for which I do not have adequate time, but a few observations are so germaine to my purpose that I cannot let them go unmade.

      Serious engagement with the computer means internalising it so that it becomes, as I have implied, an agent of perception and instrument of thought, one of the Medien des Denkens. [11] This makes it a medium of communication in a more consequential sense than is usually realised.

      Scholarly editors, such as myself, have perhaps always felt that their practice was a kind of private spiritual exercise, only a small fraction of whose benefits could be passed on. Here the computer changes the picture radically. The electronic medium allows me as researcher, for example, to publish not only the finished work but also, because they are systematically formalised, all the materials that went into its production, including the marked-up text and the software used to generate the electronic book from it. In other words, I can publish my model of Ovidian naming, not simply one of its static states. This means that close editorial engagement, the creative psychomachia of the computational against the poetic, the core exercise of modelling, can be continued by the user. Solitary exercise thus becomes shared practice, hence scholarship more conversational, tentative, collaborative. Within the humanities the implications are profound and require our very close attention. (McCarty? 1997).

  3. Humanities computing and the disciplines

    4. My concern here, however, is the more specific question of how a perspective from the shared instrumentation of the humanities might point the way ahead to their methodological cross-fertilisation and so to new discoveries. I wish to root that perspective in concrete institutional terms.

    In his book Image and Logic: A material culture of microphysics, Peter Galison conceptualises the interrelations among subfields in that discipline by focusing on their shared instrumentation (Galison 1997). He adopts the anthropological—linguistic metaphor of a "trading zone" between radically divergent cultures, where people lacking a common speech cobble together a temporary, semantically restricted language, or "pidgin", to negotiate the exchange of objects. These objects, he points out, tend to have very different meanings and uses in the two cultures. From the wholistic view of each culture, translation of the object into the other denudes it of its local, original meaning, but from the trader's view there is no loss, only change in how the shared object is put to use. The trader's concern is perforce with communication of a common object. [12]

    Galison's notion of a trading zone, where "knowledge instruments" are exchanged, supplies what I need to move from anecdotal evidence toward a cogent model for the assimilation of computing into the humanities. If, that is, we think of the emergent professional role as analogous to the ancient Phoenician merchant's, trading methods objectified in software, then we have the beginnings of a job-description. I would like now to fill in the details, specifically by asking how humanities computing might relate to the fields with which it has dealings. I mention the Phoenicians because they invented the alphabet out of the exigencies of commerce – which I think suggests just about the right mixture of pragmatic and intellectual activity.

    In my intellectual trader's view, then, humanities computing relates in two ways to the disciplines on which it impinges. The first or primary relation is to those fields whose ways and methods constitute part of the basic equipment whatever the discipline of application. The secondary relation is to all the disciplines by the type of data they ordinarily handle, as I will explain.

    1. Primary disciplines

      2. In my sense, then, the primary disciplines are history, philosophy and sociology, as follows.

      1. History

        2. History, specifically of technology, is central to the agenda because we need to understand how technologies interact with the societies in which they emerge, and so by inference what kind of a thing the computer is and what we are attempting to do with it. Locating it in the history of automata, for example, brings a wealth of insight to bear on fundamental questions, including for example the centrality of computing to the human imagination, beginning with Hephaistos' robotic tripods in the Iliad of Homer (18.376-7). As a communications technology the computer is comparable to the telephone, whose history brilliantly illustrates the struggle, still very much unresolved for the computer, to understand a new invention and the balance between contingent and intrinsic factors. In brief the historical context draws us well beyond Turing's machine to the factual basis for a deep cultural understanding of the device.

      2. Philosophy

        3. In the broad sense, philosophical questions naturally arise out of a machine that mediates knowledge and whose modelling of cognition reflects back on the question of how we know what we know. Philosophical training would seem a sine qua non because of its disciplined and systematic focus on logic and critical thinking skills, as well as a concern (as old as Plato, if not Parmenides) with how to interpret diverse representations of knowledge, including what philosophers and literary critics jointly refer to as hermeneutics. [13] Under the rubric of philosophy I would also include the practical skills of rhetoric, argumentation and deployment of evidence, which we need to rethink in the context of information technology, especially because of the volumes of raw information newly available in many fields. We need to teach our students and colleagues, indeed to understand ourselves, how to think with and argue from evidence that as it grows in volume challenges our ability to have any single view of it. [14] In some fields the mutability of data is a feature, not a temporary problem; in many the pragmatics of making information available precludes traditional scholarly care. How do we navigate in these waters?

        Philosophy is also basic to humanities computing because, as Professor Donald Gillies argues, it is historically fundamental to computing. In a recent paper (Gillies and Zheng [forthcoming]), he shows that the philosophy of mathematics in its foundational period, from Frege's Begriffschrift in 1879 to the developments from Hilbert's programme into the late 1930s, formed the basis for computer science. Especially in the the schools of logicism and formalism the failed attempts to deliver a secure basis for mathematics found their true application in making a science of computing. One might say that as in other areas of application the escape of mathematical reality from the tight constraints of formalisms, as shown by Gödel, gave to computing its raison d'être in the via negativa of enlightening failure.

      3. Sociology

        4. Sociology [15] is by definition not one of the humanities. Nevertheless the sociology of knowledge is a fundamental problem in the application of the computer to any field. In the humanities this problem resolves into at least two genuine research questions. The first is how use of the computer, by altering basic mechanisms of research, publication and instruction, exerts system-wide pressure for change within our institutions, as I suggested earlier. Those of us intimately involved both with the humanities and computing are needed to help answer this question and so to work out the consequences for the kinds of work our colleagues do. The second question arises chiefly in the context of software development, i.e. the construction of actual models for research. The logical place to begin is with what people actually do and how they interact. Few of us, however, are methodologically self-aware enough to be reliable informants; much of what we in fact do is habitual, learned by apprenticeship and practice rather than by explicit instruction. To triangulate on research behaviour is not easy, and it makes little sense for us to reinvent effective ways of doing this when proven sociological methods are available.

      These primary disciplines – history, philosophy, sociology – I would suggest as contributors toward a postgraduate programme in humanities computing. I will return to the question of training at the end of my paper.

    2. Secondary disciplines

      3. The secondary disciplines, as I noted, are defined with respect to computing by data-type (rougly: text, image, sound). Thus, for example, linguistics, philology, literature, history, philosophy and theology cluster by words; archaeology and several kinds of history by images; music, musicology, aspects of linguistics and areas in cultural history by sounds. The research areas centred on verbal data in turn subdivide methodologically according to whether they take running text or discontinuous chunks as the norm, and so tend to use concordancing and other text-analytic or database approaches, respectively. Within each cluster, methodological exchange is relatively easy, though the mental barriers may be formidable. In some cases, demonstrated by the recent incursion of imaging techniques into literary and textual scholarship, the availability of new techniques may show the relevance of formerly ignored data.

      What of computer science and the humanities? As far as I know, this question was first raised publically in the English speaking world two years ago, at a meeting convened by the National Science Council of the U.S. The roundtable discussion, subsequently published by the American Council of Learned Societies (ACLS 1997), made clear that the interests of the two broad areas of research were coming into alignment but that so far a "common language" of negotiation was wanting. In other words, we do not yet have anything like a satisfactory answer. We understand the process, however: as in a Galisonian trading zone, its job, though hardly begun, is clearly to import products from computer science or commercial intermediaries into the humanities, and to formulate basic problems revealed in humanities field-work for computer scientists to tackle. If I heard the computer scientists at that meeting correctly, such problems are most welcome.

  4. A research agenda for humanities computing

    5. The remaining question I have set for myself here is the research agenda of humanities computing. As David Hilbert noted in his famous address, Mathematische Probleme (Hilbert 1900), "As long as a branch of science offers an abundance of problems, so long is it alive; a lack of problems foreshadows extinction or the cessation of independent development.". Problems in mathematics can be solved in a sense neither possible nor desirable for comparably important problems in the humanities, but Hilbert's point holds: we also need an identifiable research agenda of problems that, as Hilbert further observed, must be at the right level of difficulty, each challenging enough to spark genuine involvement yet not so inaccessible, "lest it mock our efforts".

    Humanities computing is, however, but a new sprout in a very old garden. Our research agenda is in the early stages of formation, with a mere 50 years behind us, and much of that very sparsely populated. Work has been done across the disciplines for the last 50 years, but as I said at the beginning, seeing how the computer changes the terms in which we think and resolving what belongs to humanities computing proper is still new. I propose, then, to identify two very broad and complementary areas out of which an extensive research agenda is to be mined. My intention is to provoke discussion among computing humanists across the disciplines, so that they may ask, each of his or her own speciality, what current problems are most productively susceptible to computing. Knowing these will help us to extract the details of our agenda. [16]

    As you will see, these two areas may be considered aspects of a single defining problem, viewed either as process or as product.

    1. Mechanical primitives

      2. If we begin with a conception of research as what scholars do, then the fundamental problem on the research agenda for humanities computing is to define what I call the mechanical primitives of research, i.e. a basic set of combinable algorithmic processes that represent common steps in the manipulation of research data.

      The pragmatic goal of the attempt to define primitives is to create a kind of programming language on a sufficiently high level that the individual scholar can model his or her research problem without recourse to technical help. Such a language would allow the working scholar to become his or her own tool-maker, and thus bring into play the majority of humanists, who must now struggle with software most of which has not been written for the purpose in mind, and almost all of which is too difficult to adapt as research proceeds. It would allow scholars to a far greater degree than now possible to engage in heuristic play, and so for modelling to be fully realised as a central activity of their work. It would allow for the development and communication of methodological ideas, and so foster progressive change on many levels.

      The basic idea is of course not new. [17] It has, however, been pursued neither systematically nor from a broadly interdisciplinary perspective, with predictably disappointing results. Valuable clues are to be found in existing software, in publications detailing individual research methods (computational or otherwise) and in what we know or can discover sociologically about actual practice. By the nature of humanities research this set of primitives is unlikely ever to be complete, but closure is not required. In practical terms what we need is a sufficiently useful number that others will be discovered in practice by their lack, and so our methodological self-awareness will grow under many hands, with a healthy diversity.

    2. Encoding

      3. If, however, we conceive of scholarship primarily as product rather than process, then the fundamental problem of humanities computing is encoding, i.e. rendering phenomena computable by addition of metadata that unambiguously state what is.

      With encoding we are in much better shape than with the mechanical primitives. This is because in recent years considerable progress has been made in specifications for textual encoding sufficient to meet the demands of humanities research. The result is the Guidelines of the Text Encoding Initiative, an international collaboration involving many scholars and their research projects (Burnard & Sperberg-McQueen 1994). The TEI has perhaps made the single most important contribution to the field in the last decade by providing a unified framework within which the attention of numerous active researchers could be focused on the central problem of how to represent complex textual data in computable form. The problem of encoding is fundamental and complementary to the problem of mechanical primitives in three ways: encoding takes over where analysis by algorithm fails; it allows us to model an interpretative process; and in the rigorously systematic approach a good encoding requires, it identifies phenomena that more sophisticated algorithms may be able to handle.

      Encoding thus proceeds from the intersection of human and artificial intelligence within the context of a humanities research problem. The locus is, of course, in the mind of the encoding scholar, and the operation is by nature an intellectual, interpretative act. Encoding illuminates and on occasion discovers research problems in the fields of application, as I illustrated earlier with my own project. A scholarly encoding is also, however, apt to challenge severely the metalanguage itself, and so raise new research questions in the metatextual representation of verbal phenomena. Solutions to these, when they are forthcoming, open up new possibilities in the fields of application, and so goes the progressive spiral.

  5. Institutional challenges

    6. What, then, are the challenges that applied computing poses to our administrative and disciplinary structures? In conclusion allow me to look beyond my own immediate concerns here to four institutional problems that face us.

    I deliberately avoid details of implementation. Although roughly the same job needs to be done everywhere, institutional structures are notoriously local and idiosyncratic. So I dare not presume to say how the implications of applied computing are to be worked out in each circumstance, especially in a country whose academic culture is foreign to me. The broad outlines are clear, however, and so is the urgent need for administrative as well as technological and scholarly imagination. This point is absolutely crucial: institutional structures so influence the minds of those who work within them that without administrative imagination, and the concomitant political will, scholarly and technological imagination may be thwarted, blasted or the bearers migrate elsewhere. It's not a question of resources as such: without the proper structures, crucial lines of thought and kinds of work may be impossible to think or do.

    The challenges to this administrative imagination are as follows:

    1. Institutional home

      2. Define an institutional home that fosters the pursuit of humanities computing as a coherent interdisciplinary activity. Computer science, as I noted, has the wrong focus. The argument for embedding it separately within each of the existing disciplines overlooks the uneven distribution this implies (strong departments can afford it, others cannot) and the severely debilitating fragmentation of a unified activity many of whose benefits accrue from its interdisciplinary pursuit. We are left, then, with the task of creating a new unit and new appointments, with all the problems those imply.

      In countries where tenure separates academic from service functions, establishing this home has proven very difficult. Opposition frequently takes form in the question, "is humanities computing a discipline?", as if those fields we are currently untroubled to call disciplines could pass a rigorous, principled test. The historical evidence suggests otherwise, indeed that many of them are pragmatic arrangements. Far more productive and less wasteful of human and material talent, it seems to me, is to ask, "is humanities computing a coherent intellectual activity?" – of which, I have attempted to demonstrate, can be no serious doubt.

    2. Service and the interrelationship of departments

      3. Rethink and reimplement the notion of service so that what scholars in applied computing do for their colleagues can be properly recognised and rewarded as an academic activity. [18] A collegial computing service, which I described earlier, challenges a commonplace status quo in the humanities, in which departments are isolationist and computing intellectually trivialised. Furthermore, such service is one aspect of the broader trend toward the collaboration that computing appears to urge on us in the humanities. The changes and consequent problems are not few or trivial.

    3. System-wide effects of computing

      4. Work out the system-wide effects of computing on our institutions and the consequences of these for the academy as a whole and for specific disciplines. Although I mentioned briefly the implications of conversational, collaborative work in the humanities, I have avoided saying much about this topic here because it requires sustained attention from someone who understands the particular academic culture in detail (see, however, McCarty? 1997). I note it in passing – with a strong recommendation that you find such a person – because these effects follow from the shift in perspective that, I have argued, computing the humanities represents.

    4. Instruction

      5. Design academic courses and degree programmes in applied computing. Particular attention needs to be paid to training the next generation of computing humanists (at the M.A. or Ph.D. level); we urgently need to guarantee that our now fragile strength is not lost. To design such courses and programmes means understanding the nature and scope of the field, and so carries on from such arguments as I have conducted here. Institutionally it also means, however, figuring out how to gain the active support and understanding of affected departments, both those that want their students to have academic training in the technological aspects and those whose contributions to the programmes are implied by the interdisciplinary scope of the field. For humanities computing itself, I have identified history, philosophy and sociology as primary, but others are possible. For some fields, such as cyberculture and postmodern studies, association with humanities computing offers the technological grounding and actual contact with the revolutionary intellectual effects these study.

    Again, the key to our way forward institutionally is the combination of an educated imagination and concomitant political will to foster what is already happening and shape it to our ends. None of us is alone in this – such decisions as you are about to make are being weighed and made around the world. The hundred flowers are starting to bloom. Ongoing international discussion, as at ACH/ALLC, the international humanities computing conference [], and through electronic seminars such as Humanist [], help in all the obvious ways and must not be ignored.

    At root, however, what all this turns on is scholarship, and what that turns on for humanities computing is the questioning spirit that sees a model fail and asks why. When I was a small boy, Padre Roberto Busa, the Jesuit scholar who began humanities computing about 50 years ago, started asking such questions as I am raising here (see Busa 1980). The major difference between then and now is perhaps that many more of us today are asking them, and we have much better machines and considerably greater institutional resources with which to pursue them. Let us get to work.

Notes

  1. The essay is for Harold Short, who has made so many things possible, and Jed Buchwald, who told me I had to write this long before I knew how. The first part of my title is indebted to the title of a paper by R.W. Hamming, "We Would Know What They Thought When They Did It", in Metropolis 1980, cited by Mahoney 1999.

  2. Letter from Prof. Dr. Jürgen Renn, Max Planck Institut für Wissenschaftsgeschichte (Berlin), 12 March 1999..

  3. In N America the view of academic computing from within the disciplines and at particular institutions has taken the form of statements in recognition of computing-related work toward promotion and tenure, for example the "Guidelines for Evaluating Computer-Related Work in the Modern Languages", by the Modern Language Association of America []; the "Statement of Computer Support" by the American Philological Association []; the "Report On Academic Computing Recognition", by the Faculty of Humanities at the University of Victoria (Canada) []. (The "MLA Statement on Computer Support" typifies the problem in N America; it quotes Brian Hawkins: "a new kind of support person [is needed] in most of our computing organizations", one of those "who understand both the technology and the methodologies and disciplinary content of a given faculty member". The commonplace distinction between "support person" and "faculty member" is seriously limiting; see note 6, below.) Far more promising work is being done by organisations that stand outside the individual disciplines, in particular the American Council of Learned Societies [], especially through its publication series; the National Initiative for a Networked Cultural Heritage (NINCH) [].

  4. By the phrase "just a tool" I refer to the reductionist's dismissive view that the intention or purpose of the user is unaffected by using the tool. As Michael Sperberg-McQueen has pointed out to me (private e-mail, 27/5/99), a more serious objection to the kind of scholarly attention I am advocating can be made by observing that the microscope, telescope, Bunsen burner and Florence flask are also tools that have effects on what we know and how we know it, but a serious argument that we should have academic fields dedicated to them would seem rather difficult to make. Certainly, a department of the Florence flask would be a Monty Pythonesque creation, but many serious departments are concerned with the material culture of science. The intellectual polymorphism of the computer means that it is many tools, i.e. a modelling device, and so offers a sufficient basis for a very rich material culture for research in the humanities -- hence the need for humanities computing.

  5. The most recent efforts have been the bibliographies by Lancashire 1991 and by Adamo 1994; see also McCarty? & Hockey 1996. An ongoing effort and more of a synthesis is required, though the problem is a very difficult one. Not only are relevant publications scattered across the disciplinary journals and useful material buried within publications on other subjects, but the barrier of language inevitably results in further isolation. Lack of recognition for the field is discouraging and results in a significant amount of that which does get published appearing as "grey literature", which is of course more difficult to find because it is less widely circulated through established channels. Furthermore, since the kind of work recognised as scholarship varies by discipline or even by school of thought within a single discipline, reaching consensus on the quality or scholarly relevance of work in humanities computing is not a simple matter.

  6. Unsworth 1997 stresses "the indispensability of mistakes, and the necessity of recognizing, documenting, and analyzing our failures", according to the scientific model of Karl Popper. He quotes from Conjectures and Refutations (1960; 1968), that "the deeper the original problem and the bolder its solution" the more it reveals new unsolved problems, and so returns us to the Blakean notion of intellectual doubt with which I began. My emphasis here is rather different from Unsworth's; I am concerned not so much with the nature of scientific statements in the progress of humanities computing research as with the negative illumination of what we know but cannot yet account for.

  7. At Toronto, in the now defunct Centre for Computing in the Humanities, Faculty of Arts and Sciences, under the aegis of the School of Graduate Studies, 1991-96; at Princeton, in the Summer Seminar of the Center for Electronic Texts in the Humanities (CETH), 1992-96; at King's College London, in the Centre for Computing in the Humanities, School of Humanities [].

  8. As Pelikan 1992: 62 argues, "even such a term as 'providers of support services' is becoming far too limited to describe both the skills and the knowledge required of those who hold such positions. Scholars and scientists in all fields have found that the older configurations of such services, according to which the principal investigator has the questions and the staff person provides the answers, are no longer valid, if they ever were; as both the technological expertise and the scholarly range necessary for research grow, it is also for the formulation and the refinement of the questions themselves that principal investigators have to turn to 'staff', whom it is increasingly necessary — not a matter of courtesy, much less as a matter of condescension, but a matter of justice and accuracy — to identify as colleagues in the research enterprise.

  9. For a brief description of my research toward An Analytical Onomasticon to the Metamorphoses of Ovid, see McCarty? 1996. A provisional version of Onomasticon itself is available online [].

  10. For discussion of formalisation in the humanities with respect to computing, see Orlandi 1998. It would seem that philosophy is the natural discipline for a study of this formalisation per se, but the study of the effects and implications of formalisation in the humanities as a result of computerisation belongs to humanities computing.

  11. The title of a lecture series offered by the Max Planck Institut für Wissenschaftsgeschichte (Berlin) [].

  12. As Luigi M Bianchi (York, Canada) suggests in Humanist 13.31 [], "the question of language in... the exchange of 'knowledge instruments' in the 'trading zone'" is an important one. It should perhaps be added to the agenda for discussion of formal training in humanities computing. It would seem necessary to recognise, however, that requiring very many languages for active participation in the field would cut off all but a few, including those whose background, training and/or area of employment has not presented them with the opportunity and motivation for mastering them.

  13. For this observation I am indebted to Professor Charles Ess (Philosophy, Drury College, U.S.A.), and for his gift of lines from Homer on another occasion. See also Bynum and Moor 1998.

  14. J.C. Nyíri argues that philosophical investigations into the disunity of knowledge date at least from Nietzsche; see his excellent discussion of the philosophical background in Nyíri 1996.

  15. By "sociology" here I mean "The application of sociological concepts and analysis to the social context of other disciplines or fields; a particular sociological system" (OED s.v. 1.b); thus, as the quotations cited in the OED make clear, we can speak about the sociology of law, religion, knowledge, science, language, art, literature and so forth..

  16. See also Aarseth 1997; Orlandi 1998.

  17. The development of text-analytic tools from the early concordancers illustrates in a haphazard, desultory way the kind of methodological research I have in mind. The UNIX "toolbox", with its pipes and filters, is an early implementation along similar lines, and more recent scientific visualisation tools carry the notion into the world of GUI interfaces.

  18. Boyer 1990 discusses service as "scholarship of application", though he has in mind application outside the academy. His remarks are germaine, however: that when service is "tied directly to one's special field of knowledge... [it] is serious, demanding work, requiring the rigor--and the accountability--traditionally associated with research activities" (22). He notes also that "New intellectual understandings can arise out of the very act of application.... theory and practice vitally interact, and one renews the other" (23).

admin