<x-flowed>
Humanist Discussion Group, Vol. 17, No. 549.
Centre for Computing in the Humanities, King's College London
www.kcl.ac.uk/humanities/cch/humanist/
www.princeton.edu/humanist/
Submit to: humanist@princeton.edu
Date: Wed, 21 Jan 2004 08:18:01 +0000
From: Manfred Thaller <manfred.thaller@uni-koeln.de>
Subject: humanities computer science
Dear Willard,
Well, yes.
I think I should start with a very small item of background information, as
the understanding of my reply to your statements on "humanities computer
science" may be enhanced by a bit of contextual information.
I myself have for some time been engaged in various rather technically
oriented developments in the field of computer supported projects in the
Humanities. Since a few years I hold a professorship with the totally
untranslatable designation of "Historisch-Kulturwissenschaftliche
Informationsverarbeitung", which I would much prefer to be called
"Geisteswissenschaftliche Fachinformatik" in the first place, which in turn
I render in English as "Humanities Computer
Science". Whether this is the only professorship of its type I do not know,
but whatsoever you call it, there are relatively few around which implement
a CS / Humanities study program going all the way from the [German
functional equivalent of a] BA right through to a PhD program. And we DO
assume that our students are at least as close to "true CS" as other
applied CS types of study; rather extensive programming projects being
actually the main reasons from dropping out from the curriculum, though, as
I should clarify, this is an ARTS FACULTY professorship.
While in earlier years a not infrequent participant in ALLC / ACH type
events - and at some stage president of the than in Europe highly visible
Association for History and Computing - I have in recent years insisted
quite forcefully, that the model which is implemented by the professorship
quoted above, and indeed the whole paradigm I try to follow, cannot be
meaningfully subsumed under the heading of "Humanities Computing", but
requires the definition of something provisionally called "Humanities
Computer Science".
For the benefit of my students I have invited Willard to a seminar last
summer, where we held a kind of medieval <i>disputatio</i> in front of the
students about whether or not such a discipline like "Humanities Computer
Science" actually exists, can exist, what it may mean etc.
I hope that does not read too vain, but somehow I thought the argument
which I am taking up with Willard's recent comments on " ... it would seem
to me that a term such as "humanities computer science" is rather
problematic ..." may be more transparent, if it is clear that it is part of
a discussion ongoing for some time, though not exactly publicly.
From background to substance:
I totally agree, that CS is an extremely broad field; I cannot resist the
temptation, however, to point out that, if that was sufficient to exclude
it from being a potential subject of a specialisation looking at it from a
specific knowledge domain, just about all fields of scholarship which claim
any significance for "the Humanities" (CHum comes to mind) would be
unapproachable as well, as "the Humanities" are themselves at least as
broad and heterogeneous as CS.
So, if we talk about the relationship between two very diverse areas, we
should first give the definitions each of us is using.
In my arguments - along with the German and much of the French and Italian
tradition in Europe - "the Humanities" include more or less all of what the
old "philosophical faculty" used to teach: literature, linguistics,
history, archaeology and all aspects of discussing the arts, just short of
the actual production or performance of art.
Similarly, when I talk about "Computer Science", I have to explain, that
this is a translation of the German term "Informatik". At first look that
term is just as shifting as the one which Willard describes (the homepage
of the German "Gesellschaft für Informatik" (= Association of CS) actually
refuses to give a precise definition of the term, pointing explicitly at
the constantly changing focus and increasing scope). But in actual usage it
is possibly slightly more focused, by implying a much clearer separation
from hardware (which is engineering) - and by its very name a closeness to
the handling of "information".
Indeed, most CS persons which I am aware of - and that also applies to the
larger world - would probably subscribe to one or the other variety of the
definition of CS as "the study of the representation of information on
computing machinery and the ways in which these representations can be
processed". Actually I do doubt that there exist very many curricula even
remotely related to CS, which do not include a quite extensive module on
"data structures and algorithms" right at the beginning of the course of
study. Different curricula may indeed lead from these "data structures and
algorithms" into totally different directions: You can mathematically prove
an algorithm as optimal, even if it is totally pointless from the point of
view of an applicability. You can study ways and means of embedding
knowledge about the properties of data structures into the architecture of
larger systems. You can study ways and means of building software tools
which make the workings of an algorithm opaque when concentrating a user
interface. But: Very few subfields of CS are not aware of these two concepts.
In any case - <i>petitio principium</i>: the CS I am talking about
<emph>is</emph> the "study of the representation of information on
computing machinery and the ways in which these representations can be
processed".
If we accept this as a non trivial subset of the possible meanings of CS, a
"Humanities Computer Science" defines itself as the study of those
structures needed for the representation of information and those methods
of processing them, which are (a) indigenous to the Humanities (in the
sense of "specific for their knowledge domain"), (b) and / or so far away
from the mainstream of CS, that they are not being studied there, (c) and /
or require a background in the Humanities to understand the specific
properties of such information.
Which, of course, to be meaningful, requires the Humanities to handle
information which fits the bill. Now, in my opinion such information
<emph>does</emph> exist - which in itself would need a much larger
treatment, but for brevity's sake let me just state that the following
properties of
information:
- vagueness,
- ambiguity,
- context sensitivity,
- incompleteness
and a few others are all incomparably more prevalent in the Humanities than
in the knowledge domains which CS has been applied to traditionally.
Some of them ARE studied by CS: Every time I listen to one of those
cultural critics who tell us, why the "digital principle" allegedly changes
our whole world, as computers understand only 'true' or 'false', I engage
in little day dreams of asking them to write ten thousand times "There is
such a field as fuzzy reasoning and I will not open my mouth again until I
have read at least an introductory text into the not quite so trivial parts
of CS". But none of that is even remotely close to what we would need to
process Humanities information in slightly more advanced ways.
Others are totally neglected: It is quite clear, that certain aspects of
"time" in cultural systems simply cannot be handled by our current ideas on
how to implement time. (Yes, some of these are purely notational - but
along a line of increasing complexity we land at charming concepts like
"cyclical time" where you simply cannot glue a paperback to the existing
data structures any more, but have to start from scratch.)
And in some cases it is painfully clear, that the separation of CS
knowledge and Humanities knowledge hurts. Just once in a while I am
dreaming what could have been done with the money Getty has burned for the
production of thesauruses over the last decade or so, if those programs
would not have been defined by art historians who believed what IT personel
told them, which did not understand the question in the first place.
So, Humanities Computer Science consists of the chasing of very remote
ideals, which have no relevance at all for the production of the results
which we need day by day? Well, my colleague in archaeology who finds it
quite attractive that we supply a steady stream of students who have taken
courses about abstract CS principles, have taken their archaology courses,
too, and are therefore quite useful as research assistants would not see it
<emph>quite</emph> that way. (And that together with the courses on C++
programming they are also trained in familiarising themselves quickly with
new types of application software does not hurt either.)
Willard has asked, whether a field of common interest between "CS" and "the
Humanities" could be defined. In an abstract way I think I have given some
examples of why I think the answer is clearly "yes". But I have given the
preceding excuse for not engaging in the totally futile, as a very obvious
corollary to Willard's statement could be added in response to the argument
above.
"Still, given that what Manfred says may be right", it would run, "still:
Why do we need somebody studying that set of problems within a Humanities
setting? Should we not simply wait until CS discovers that area and focuses
more closely upon it?"
The reason I raise this unasked question is, that my answers to it are
close to the heart of my own reasons, why I do not see myself as part of
the current field of "Humanities Computing" any more and try to champion
the notion of a "Humanities Computer Science".
The difference being is - leaving all anecdotal references aside - that in
Humanities Computing as it stands today there is a tacit and sometimes
vocal assumption, that there is a body of secure knowledge, produced by CS,
which just needs to be applied to the Humanities to lead to wonderful
results. This point of view turns Humanities scholars into consumers of
knowledge produced by others.
My understanding of CS as a field, which is currently amalgamating implicit
ways to handle information taken from various knowledge domains and
defining itself by the experience, precludes me from assuming that
"something is out there, which we just have to adapt". I do not know, what
these these new paradigms of handling information will ultimately lead to
and whether they will be called CS: But influencing these paradigms seems
to me to be of vital importance for all fields of study, if they want to
keep their status. Humanities who bring their abilities to handle specific
types of information into such an emerging field of study and contribute to
the emerging body of knowledge on how to handle that information, will in
my opinion be taken incomparably more
serious than Humanities which just consume what "CS" has produced
(shrink-wrapped by Microsoft afterwards).
You have to admit that it <emph>is</emph> a bit absurd: The Humanities have
over the centuries developed a pretty good knowledge on how to orient
themselves in bodies of contradictory, ambigous, suspicious and more
generally chaotic data. The WWW can be described as the transition from
clean and well defined sets of information towards a body of contradictory,
ambigous, suspicious and more generally chaotic data. But 99.99 % of all
the decisions
about the WWW are taken by people carefully trained to handle clear,
unambigous, reliable and more generally well ordered information. And the
current generation of Humanities researchers relies on those champions of a
clear, unambigous, reliable and more generally well ordered world to
rediscover much of what they themselves know since quite some time.
One of the most hilarious jokes in the development of information over the
last fifteen years needs to be retold. The notion of hypertext is one
deeply rooted in the Humanities. It was designed with the notion in mind,
that a text is a changing body of information and much of the impetus
behind the original thinking about hypertexts was inspired by the wish to
model a text in precisely
that way. To have all the versions in which it existed present and to have
the possibility to see how the argument developed miniscule modification
after miniscule modification.
Well, of course that would have been much to challenging actually to
implement by Humanities researchers ... so out of the notion of a hypertext
the current notion of the content of the WWW was derived; by a group of
people who were geniuses in programming but had all the naivety about texts
which is inherent in most hard science people.
The result? Many wise men (and women) complain that "information
technology" destroys the historicity of texts, keeping only the last
version and loosing all the former versions.
Is it not really amusing? The final product does exactly the opposite thing
of the original intellectual design - but, fortunately, nobody in the
Humanities needed to get his or her hands dirty. We just had to wait what a
science dominated IT did to our original concepts.
Which, in my opinion, is exactly what we get, when we rely on an anonymous
type of "CS" to produce solutions, which we than have "just to adapt to our
knowledge domain".
Or, to take up Willard's final quote:
"Let's build even more bridges to other disciplines and to industry;
let's amplify the special strengths of each."
A bridge needs two heads; on the various beaches of scientia many are
there who are ready to construct them. If we are not willing or able to
build our own bridgeheads, the bridges might never reach our shores.
Comments?
Yours,
Manfred
</x-flowed>
This archive was generated by hypermail 2b30 : Fri Mar 26 2004 - 11:19:38 EST