Date: 31 October 2001
Topic: Databases
Leader: Will Thomas
Business Items:
Andrea will repost schedule to email so that it's archived. Include URL
for hypermail archive.
John will post Powerpoint slide about his upcoming presentation in
Vancouver.
Seminar:
WT: Lots of work done in 80's and early 90's. Faded now. _History
Computing Review_, a journal once regularly released, is now only
occasionally published.
Is a relational database representative of history? Does it constrict the
researcher in his analysis and presentation of the historical source?
_Databases in Historical Research_ contains about 10 case studies of use
of databases in historical research.
Forcing history--which is messy, disorganized, etc--into ordered and
structured things such as databases is problematic.
Source-oriented data processing--doesn't force historian to decide at the
outset to choose her methodologies. That addresses the major concern of
historians, that the use of computers pre-defines the methods available to
the researcher.
--historical source should be entered as an ASCII transcription with as
little interference as possible; represent the source with as little added
as possible, many of these also include images; then layer on top of that
metadata, formatting, etc.
JM: You want to stay as close as possible to the original document as
possible.
DP: But what is as close as possible. Anything implies a structure. No
method requires choices at all stages of the game; there is no neutrality.
GR: What's important is that the researchers be clear about what the
structure is that their using. Self-reflection and knowledge of the
interpretations implicit in the process.
What's going on in these articles is that the historical researchers are
using statistics. Where the perceived problem lies is really the
statistics, not the databases, the querying options, the structures.
JU: In English Lit and Linguistics, quantitative methods are out of
fashion in the U.S. and in fashion in Europe. Is that the case in history
as well?
WT: Yes. In fashion in Britain and Canada. Out of fashion elsewhere.
JU: Why?
WT: Attitude of pervious quantitative scholars: we will bury you. Fogle
and Egerman (?) an example of the height of quantitative method.
Conclusion: slavery was profitable, capitalistic.
DS: Other historians reacted strongly to the "top-down" approach to
history. "People's histories" were gaining strength; narrative histories
of individual lives and conditions were seen as authoritative.
WT: And French and German philosophy. More like literary theory and less
like social sciences.
JU: Interesting that this split also exists in literature study.
JD: Split also exists in media studies.
SR: This is interesting, but there's a sheen of empiricism that's
disturbing.
JM: It's positivism.
DP: Too what extent is there a move to joining the two extremes--the
perceived objective and subjective approaches.
WT: Not really. Cultural history, which is not quantitative at all, is now
dominant. Environmental history, an emerging field, offers the
possibility for such a junction.
GR: What about GIS? Is this a place for union.
JU: Distinction between empirical and positivist. How do you place
yourself in on or the other camp?
SR: Quote that Will read. Issue of dates. Dealing with dates is a very
lossy process. And, in large part, it's a problem of chunking. But
aren't we doing this chunking all the time?
CR: But the table of values obscures the interpretations behind them.
DP: How can you make the vagueness that chunking requires still precise in
a way that allows the data to still be searched?
GR: What about gaming. Simulation? It seems like the ACH is interested
in play, in gaming, in simulation. Whereas ALLC is interested in making
literary theory a science.
JD: 60's and 70's, peak of popularity of AI. Minsky is misquoted as
claiming that 1972, Dreyfus publishes what computers can't do.
Hard-edged abstraction in late 60's; hard-edged graphics at same time.
Graphics get "modern" look, corporatization of images of these
technologies. Rand's graphics that make the IBM logo look completely
machine made. Machine-made things feel aesthetically pleasing and
stylish.
Computational capabilities at an industry saturation level that
computational methods are available to most. Penetration of research
institutions at that moment. 1960's.
JU: Each discipline, at this point, decides that we either need to pose
our problems in terms that can be calculated.
DP: Began in late 19th century. Machines available to aid computation.
SR: Levi-Strauss, at one point, says "too bad I don't have a PC." Derrida
attacked this.
JD: When did computational linguistics start to take form?
JU: Don't know for sure. 1975 - 80?
RD: Copenhagen, department in computational linguistics, translation is a
huge problem and expense. Very narrow conception of what we can do with
computers and computation. Thus it has lost its appeal, its interest.
Problem with "knoweldge representation." Current use of it is too
computational. But that's not what we're doing. Find another term.
JU: Disagree with Rune. We need to start with the constraints. We need
to try to fit these into the formal and then spend time talking about the
residue. We need constrained representation, not unconstrained
representation.
RD: Problem is with logic. And a very specific type of logic. But this
doesn't account for everything done with the computer.
SR: Central debate here: how do we philosophical theorize or define
quantitative methods? Can we use this in the class? How can we teach
this?
WT: You could use history _Time on the Cross_ vs. _Time on the Cross:
Reconsidered_.
_Without Consent or Contract_. Support of methods. 4 volumes of data.
Support original thesis.
GR: Let's try to locate the point of transition again. Let's not look for
a particular thinking. First thought it might be Godamer, but don't think
so now.
What are the turns we want to track? What are the words we'd want to look
for?
JU: quantitative
JD: computational, computer*
JM: antonym: critical
Project: send students into bibliographies. send them through the
literature, look at what words are being used in titles, subject headings,
etc. use this to track changes in methods.
DP: get LC subject headings. look at them historically. when are new
terms added, etc.
compare terms in titles and subject headings.
WM: scientific method vs. quantitative method. how are we using the
terms?
The debate about the differences btwn these is not only an issue in
humanities. It's a very important issue in sciences too. Many
physicists, biologists, etc. don't want to be called quantitative either.
computational is also different than quantitative.
coming up with average whippings per life. that's not computational,
that's quantitative. statistics is a key addition.
BR: It seems like we might be able to use GIS to teach students some of
these issues.
JU: what resolution of data do you need? this is very difficult to know,
and it affects the conclusions you're able to make later.
JU: what do I need from the data vs. what other uses might there be for
the data. how do I balance these two? you don't want to
WM: also issue of resolution in terms of what people can actually use or
process. what can users actually access? interpret?
RD: interface vs. computation. clash that is central to history of
computing. how users _interact_ with the machines vs. how the computer
uses data. augmentation vs. automation.
GR: interface goes back to fashion. these go through different styles.
designers creativity, motives aren't just efficiency, economics, etc.
"intellectual hemlines" rise and fall.
BR: where are the narratives?
JM: but these are all just forms of representation. narrative is one
form. tables is another. graphs are another.
JU: early writing was business and administrative. early computing was
business and administration. do disciplines that can be aligned with
business and/or administration--social sciences, political science,
etc--seem to embrace the computational technology more thoroughly and for
a longer time? is that what's ensured the long life of this technology in
these fields?
four histories: military, business, information/communication,
entertainment
This archive was generated by hypermail 2b30 : Thu Nov 01 2001 - 15:11:05 EST