The Perils of Human Interpretation


Steventheblog really hit on an interesting dilemma in his blog post this week when he writes:

“The assumption running through this chapter seems to be that that the metadata librarians write guidelines for other librarians to read and interpret. But this isn’t always the case, sometimes digital librarians have content submitters or creators submit metadata for digital collections.”

I think his observation speaks to one overarching problem related to the use of metadata.  Human interpretation.  Standards like Dublin Core are useful in so much as they are universally followed.  Sure they can be modified to suit you libraries needs, but what happens when you metadata is being interpreted and produced by a variety of people with vastly different results.  Steventheblog,  lsteckervetz and I completed a group project last Spring for LIS 551 and chose to look at the University of Wisconsin’s Digital Collections’  Historic Theater Posters collection

One impression we had was the various individuals responsible for inputting the DC metadata made interpretive decisions that affected the ability to search the database successfully.    For example, when completing the “publisher” field for one poster, one individual chose to use “Strobridge Lithography Company: Cincinnati, OH”  the next individual considering a poster created by the same publisher used “Strobridge Litho Co: Cincinnati”  So the result was that when searching for “Publishers” we would play with the searched terms and reach vastly different numbers of results.  Sometimes 30 or so, sometimes 2.  Before I belabor this point, I was left with an alarming feeling that even the slightest deviation from a standard guideline for how one submits metadata can in effect hide the item from discovery and compromise the effort to help locate items easily.  We concluded that some of the metadata was submitted by content submitters and other was added by volunteers.

Has anyone who has worked closely inputting metadata encountered this issue and felt uneasy that they might be creating a problem?


2 responses to “The Perils of Human Interpretation

  1. It looks like your example is a typical Information Retrieval issue. It seems some library scientists are interested in applying machine learning theory to improve IR and other DL issues.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s