When authority is wrong

“J Sakai” is a name describing at least two people. “J Sakai” with the “J”, so far as I know, standing for nothing in particular (likely a pseudonym) is the name of a revolutionary leftist theorist.

“Sakai, J”, apparently short for “Jun’ichi'” is also the name of a (or at least one!) physicist.

They seem to all share the same LC authority record, that is titles authored by both have been assigned to the same authority record.



I am as reasonably confident as one can get that the author of Settlers: The Mythology of the White Proletariat (which I recommend, by the way) is not the author of Phase conjugate optics (which I probably couldn’t understand enough of to have an opinion on).  If the Identities page is to believed, under “Associated Subjects”, we have an author who’s publications touch upon “Anti-imperialist movements” and “Working class whites”, as well as “Columns, concrete — testing” and “Bridges — foundations and piers — mathematical models.”  Such a renaissance man may exist, but being somewhat familiar with the political Sakai’s work, I’m pretty sure it’s not he.

[I also have to wonder if it’s but one or in fact two Japanese physicists responsible for works on structural engineering theory; as well as on optics and solar flares. But I’m not familiar with those authors/works, and at least that’s all physics. And then we’ve got some books on geology and fossils too; a third scientist?]

But at some point some cataloger (or machine process?) appears to have seen the authority record for Physicist J Sakai, and decided it suited for revolutionary communist theorist J Sakai.

So I’ve noticed this, let’s say I’m a naive non-librarian user, or heck, let’s not and just say I’m me. What would I do about this?  The naive user might think since they’ve discovered it on worldcat.org, they should report it to OCLC.  But what’s OCLC going to do about it?  We know that OCLC is really only displaying the aggregated collective recorded data of an international community of catalogers.  So I can… what, find a cataloger who has NACO authority to change these records (and create a new authority record for the communist Sakai), AND has the time/interest to do so, just out of general community spirit, even though it doesn’t really matter for their institution which is paying them and expects them to catalog X records per hour that d matter to them?

Kind of makes one reconsider again the idea of ‘wiki-like’ editing of our collective cataloging corpus. Would we be worse or better off if an interested user could, upon noting this issue in Worldcat Identities, immediatly and easily,  right on that web page make a correction? Even make a correction resulting in a new authority record for the ‘missing’ Sakai?  Even if the interested person isn’t NACO-certified, or isn’t even a cataloger, or isn’t even a librarian?   As opposed to requiring not only a librarian, but a select elite of certified librarians, to open up a completely alternate interface, find the individual records in question, and submit changes?

The idea that limiting editing access to an esoteric elite results in perfect information is simply a myth.

The physicist Sakai, Jun’ichi 1936-, if he happened accross the Identities page, would probably be none too pleased to find the cover images presented as ‘representative works’ of his ouvre being politically charged works by a completely different guy, Confronting Fascism: Discussion Documents for a Militant Movement, and Looking at the White Working Class Historically (actually by David Gilbert, but with a foreward by the leftist Sakai).


12 thoughts on “When authority is wrong”

  1. While I take your point about the wikification of cataloguing data I would like to point out that both my current and former places of work have had several patrons report cataloguing errors to us which we subsequently fixed. In a number of those cases, the errors were authority related.

    I’ve never worked in a NACO or SACO institution but I’ve always known what the nearest such institution was so that we could ask for assistance. When the authority errors were reported to us, they were reported to the nearest NACO or SACO institution and promptly fixed. In one case, the patron was able to help untangle a very large mess around some Chinese authorities.

    So, errors do get fixed now. We just have to provide a way for people to report them.

  2. Good on you and your library, and the NACO catalogers who assisted you. That’s encouraging to hear, thanks.

    I’m not sure that most libraries would be similarly responsive to something ‘out of their control’, making sure the changes end up in OCLC/LC, not just in the local catalog (if even that). I’m not sure most libraries have the staff to afford to do that, under the present regime where a significant amount of time (potentially from several different institutions) is required to respond to such reports.

    Anyone want to give me the name of a friendly NACO/SACO cataloger who would be interested in hearing about such errors when I personally happen to run into them?

  3. Jonathan – OCLC/RLG programs is looking at the wiki-fied ability to fix exactly this type of problem See info on the project on Networked Names Identity Hub http://www.oclc.org/programs/ourwork/renovating/leveragevocab/idresource.htm

    They have completed writing use-case scenarios for such a hub, but I haven’t heard yet on progress towards a prototype. Karen Smith-Yoshimura could tell you more, however. The advisory committee’s final report is available

  4. Jonathan, I will ask my current and former NACO resource libraries if they would be willing to receive error reports from you or, failing that, if they know of one near you that would be willing.

    I’ve always had good experiences with the area NACO libraries but perhaps they take pity on me as a poor public library. ;)

  5. I’m a naive non-librarian patron , but I don’t have any doubt that Jonathan is correct that, “limiting editing access to an esoteric elite results in perfect information is simply a myth”. I’d love to see hard data on the data quality of any catalog, particularly a public library in a medium-sized town, but my impression is that the union catalog that the local public library belongs to isn’t very good.

    It’s not just distinguishing between physicists and revolutionaries. I checked “King, Stephen, 1947-” and found that five of his books had wandered off in search of “King, Stephen”, who according to LC is at least one other person. It’s really a mark of fame and success to have duplicate name records. “Scalzi, John 1969”, who is inexplicably missing his hyphen, has only recently made the grade, with three books under “Scalzi, John”. And how many Drupal books have the right LC subject heading — “Drupal (Computer File)”?

    It would be great to have a big red button on every page so that I could report errors, but what then? I agree that the possibilities are either adding a note to the catalog, or reporting the error to one of the catalogers to fix in his / her copious spare time, but I don’t see these possibilities as distinct.

    To allow patrons to add notes to the catalog, I don’t think a simple wiki is the right model. The missing ingredient is some sort of reputation system, like the slashdot comment system. That would allow other users to choose whether to see patron corrections at all, and if so to rate them up or down, with the result that patrons would develop more or less credibility within the system.

    And a reputation system also works well with a system that reports errors to catalogers who would make “official” changes. Giving patrons (or for that matter other librarians) a karma score allows corrections to be submitted to the catalogers in order of karma.

    As for WorldCat getting a pass because it’s nothing more than the sum of its parts, well, no. Once OCLC decided to field WorldCat — a patron interface — it got responsibility for the quality of the data it displays. If it’s a patron interface, I want to see a big red button, whether it’s the local catalog, WorldCat or LC.

    The other issue that WorldCat HAS to solve is records from the local union catalog that don’t make it into WorldCat, so that WorldCat reports false negatives. That REALLY ticks me off.

    I can see that the sort of top-down work being done by (e.g.) RLG to develop global authority records is necessary and important, but as a patron at the bottom, I’d like to see some focus down here, too. It feels like it will be a long time before top-down efforts make their way all the way to the bottom.

  6. The problem seems to be in the way WorldCat Identities lumps things together. The authority record for Sakai, Junichi, 1936- says nothing about the political writer J. Sakai. The bib records for the political writer use the heading Sakai, J., as he or she appears on the books. However, that one has no authority record because his works probably haven’t been cataloged by a NACO library. Some records for patents also use the name J. Sakai for an inventor. There is another authority record for just Sakai, Junichi, for a scientist at Berkeley. I wouldn’t blame authority records or catalogers for the mess. It’s pretty much inevitable when there are hundreds of millions of records and untold people with the same or similar names.

  7. Interesting John, thanks. My fault for assuming these shared the same authority record. Apparently they don’t, but Identities groups them anyway. So it is Identities that should be notified, although I don’t know if they could do anything about it, because presumably the grouping is algorithmic, and no algorithm will be perfect — identity disambiguation is a hard thing.

    [Also interesting that there appear to be not two, but three or even possibly four J. Sakai’s grouped on that Identities page!]

    But, this brings up another question. Okay, we can’t ‘blame catalogers’ for collocating works that belong to different authors. But the fact remains that our collective cataloging corpus did not in fact have authorities at all for these different Sakai’s, and/or did not always assign bibs to an authority that did exist. Of course, the actual body of catalogers is only so many people with so much time, they can only do so much.

    But this means what we have to compare is not trained catalogers assigning authorities vs. “the crowd”, but rather the crowd vs _nothing_, or the crowd vs. algorithmic attempts like Worldcat Identities. Maybe trained catalogers will do better than the crowd, but which will do better at closing the gaps that trained catalogers miss? Seems to algorithmic attempts like Identities, which end up putting together possibly several scientists and a revolutionary in one identity, could use some help from the interested crowd.

  8. If you’re interested in telling people which person in a bib record should be associated with which heading in which authority record, PERSNAME-L is a forum established for this purpose. PERSNAME-L is a Listserv(R)-hosted electronic distribution list. To sign up, visit http://lists.ou.edu/cgi-bin/wa?A0=PERSNAME-L

  9. This message is found on the Library of Congress PSD page

    How may I contact LC with cataloging questions, comments, and queries?

    Please see the Library of Congress “Ask a Librarian” Web page at http://www.loc.gov/rr/askalib/. Complete the General Inquiry Form to send general questions and comments and the Error Report Form to report catalog or authority record errors. An alternative is to send an email message to the ABA Policy and Standards Division at policy@loc.gov.

  10. If I see an error (or something I think is an error) in worldcat.org… how do I know if LC is an appropriate agency to contact about it? I guess I could try to look up the same record in catalog.loc.gov, and see if it’s there, and has the same (perceived) error?

  11. I think checking the LC catalog and contacting them is the best approach. There is a link to an error report form at the bottom of each page in the LC online catalog.

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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 )

Google+ photo

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

Connecting to %s