Posting to Autocat
On 3/20/2016 6:48 PM, Tennant,Roy wrote:
On 3/20/16, 10:12 AM, “AUTOCAT on behalf of Daniel Stuhlman” wrote:
LCSH are Christian-centric. “Faith” without a modifier is consideredChristian. I know that is not fair, but the origin of theclassification and subject headings was WASPish. At the time of thecreation of the LC systems, no one paid attention to diversity. Alsoit was male oriented. That is why there are headings with “women”and not equivalent ones for “men.”
The status quo is completely unjustifiable. For a library topical controlled vocabulary that is widely used to have a point of view on matters like this is deeply disturbing. Why this has not been rectified by 2016 is incomprehensible.
While I sympathize with this, I wonder whether it really is a problem in practice. When records are considered in their entirety, items focused on “Christian faith” normally(should) have at least one extra subject that includes the term “Christianity” in some form, e.g.
“Faith seeking understanding : essays in memory of Paul Brand and Ralph D. Winter / edited by David Marshall.”
Brand, Paul W.
Winter, Ralph D.
Christianity and culture.
“100 days in the secret place / [introduced and compiled by] Gene Edwards.” with subjects:
Looking at the subject arrays for these records leaves no doubt that the items are about the Christian idea of faith.
So, I will admit that if someone looks only at the subject “Faith” there may be a problem, but humans do not do that. People look at records, and the moment they see the multiple subjects attached to an item, it becomes clear that a particular item is about Christianity. There are many such examples in our catalogs.
As a consequence, I agree that there is a theoretical problem with “linked data” where the single URI merges the concepts “Faith” with “Faith (Christianity)”. The machine sees only one subject without the others. But a real human being looks at an entire record for a real resource, and there is then no doubt that “Faith” is linked to “Christianity” in some way. To repeat, there may be a theoretical problem when the records are broken apart into separate bits of “data” (entification) but we shouldn’t forget that those bits will inevitably be recombined into something that genuinely describes the item.
There are a great many of these types of subjects that require more than a single LC subject to define a concept. I have mentioned before within the realm of library reference work, the reference interview. The two subjects used for the single idea of “the reference interview” are:
Reference services (Libraries)
Of course, I could come up with many, many, MANY more examples like this.
I can understand how the LCSH method of using multiple subject headings to describe a single concept can be difficult to fit into the concept of linked data, which is supposed to break a record apart into different bits of data to create a plethora of single URIs. But in the real world, we should include the fact that people look at entire records, plus the existence of a competent cataloger who will add subjects for Christianity (in some way), and there ceases to be a problem for a human being.
That said, I think it would be a good idea to split “Faith (Christianity)” from “Faith”, but that leads to the normal–and huge–practical problem of splitting any subject heading: what do you do with everything up to today? If we want “Faith” to be separate from “Faith (Christianity)” in our catalogs, we would have to redo the subject analysis of every record in the catalog with the subject “Faith” to determine if it refers to Christian faith or not. Catalogers do not have the resources to do that.