Kramer

This week’s readings were interesting as they explored the implications of digital history on the way we view our profession.  Kramer suggests that perhaps “all” that historians do is add meta to metadata.  While at first I found this insulting, eventually I saw the value of this label.  History adds information to/about the primary source documents.  What we do is analysis and by placing the historian and the archivist together in this way the act of “doing” history becomes more clearly defined and accessible.

The historiography is where Kramer runs into trouble.  I do not believe that the entire historiography of a source can be placed into a primary source and be understood.  Perhaps arguments and counter arguments could be layered into the system; for instance, full access for people who would find that meta useful and common access for the average person who can not, without training ,understand the implications of what he/she is looking at.

Olsen and Argamon

Text mining is an interesting concept which seems filled with both problems and solutions.  When a historian works with primary sources they are limited in the number of texts they can use for any given project.  A computer is not human, therefore it can sort through many documents much more quickly which means faster work with a more complete body of sources.  This sounds like it would be extremely helpful in the humanities, but a computer is not human.

A computer can not understand the text it reads, it can not place value in circumstances and it can not feel the era it “reads”; a computer only recognizes symbols and their recurrence in a particular set of data. As Dr. Church said, it is difficult for the computer to mine symbols and metaphors in primary source documents.  This is a red flag on the use of text mining, because it is limited in scope;  the historian can view patterns through it, but computers have a limited capacity–perhaps this is comforting.

The article suggests that one goal of text mining is to make very large data sets “manageable and meaningful”.  In this pursuit, I believe that text mining provides the ability to organize and work with a large set of primary source documents, however, it does not provide meaning.  Meaning comes from human ability to grapple with data.  Meaning comes from the interpretation of the historian.  Text mining can lead the professional to explore different research questions and broaden his/her ability to work with sources, but can not be relied upon to give meaning to sources.