For Whom the Ringer Tolls

“I’m a millennial, so I don’t like labels.”
—Tomi Lahren
Back in the days when libraries were more known for their contemplative spaces than as venues which allow food, let alone feature coffee shops, the publication cycle of print journals followed by print abstracts made for some now curious information-seeking behavior, namely the browsing of a current periodicals section in the hopes of stumbling upon not-yet-indexed items of interest.
That situation is these days more or less reversed. We don’t have a current periodicals section anymore, due to the digital turn in publication formats, and thanks to e-mail alert services and the like, there are better ways to keep up to date on your research topics. Of course, as with any new technology, there are tradeoffs of benefits and hazards alike whenever there are improved means that can nonetheless go wrong.
I receive a steady stream of error reports from our discovery layer because it contains records for articles which are not yet loaded into our full-text databases. Other items are still not even published in the print version of the journals (“epub ahead of print”) and inexplicably therefore not included in our licensed online subscriptions. The range of records included in our discovery layer (e.g., book reviews, conference poster abstracts) also make for some similar end-user frustrations, especially when coupled with our link resolver’s often-graceless handling of not-too-infrequent errors.
Many of the other seemingly defective features reported by my colleagues are in fact the case of a system working exactly as intended. The most recent case of this was someone discovering that a search for items from 2019 will obtain records with a date field of, say, “1972–” in the mix. I suggested a workaround of instead sorting the results with the newest ones showing up first, while other librarians commenced a discussion on how to get our vendor to reprogram the product to function according to their desires, ignoring the reality of how new systems may at times require new ways of doing things.
Actually, now that I think about it, the most recent example was from someone who wanted to look through our shelf list. For vinyl.
Anyway, we’ve accordingly had to throttle our hold shelf notification e-mail messages, as generated by our LMS, that were being fired off right when the items are unloaded off the delivery van or checked in by the acquisitions unit. The availability, or immediate lack thereof, for “in process” items—for which robust, in-house generated records have not yet been crafted—to readers has made for some rather exasperated conversations between staffers in various service units and amongst those with different philosophies over the optimal ways of providing timely access to our collections.
You gotta love it any time librarians start quoting Ranganathan at each other. But I’m not a cataloger. I don’t say that to hide behind my admittedly simplistic views on how those with responsibilities in that area should apply their expertise. I likewise think we shouldn’t have an approval plan, or order anything else for that matter, which isn’t a title specifically requested by name from a non-librarian patron, for starters.
Take a look at the flap over what remains the official LCSH of “Illegal Aliens.” Why are we pulling our hair out over this? I don’t mean to say that the term isn’t needlessly derogatory, but to rather point out how Google sure as hell doesn’t label every webpage about unlawful immigration by a sanctioned topic known officially by such an obnoxious title. That way of representing content to users, as far as the web goes, died with the Yahoo! directory.
Moreover, as unfortunate as it is that it still seems like something out of science fiction to declare, “we’ve learned not to fear words,” our reliance on taxonomies with no natural order can oversimplify reality and disregard individual differences. “If you label me, you negate me,” as the saying goes. Using ordinary language of the day has its own pitfalls, but then again, so does limiting a classification system to a discretely constrained and curated list of authoritative terms.
I don’t see why it remains sacrilegious to posit that a combination of full-text indexing, natural language processing, and machine learning might be fully capable of generating everything necessary to describe a publication. You cannot seriously suggest that unless we stick with a system for controlled vocabulary, one that is based on the use of index cards to facilitate the discovery of materials with a single and physical access point, people will be unable to locate the best sources on their topics.
That’s about as valid as the decades-old refrains of how “the open web is not only terrible, but it merely achieves accuracy through volume and you cannot find what you’re after in it either.” Google is a better search engine, maybe not in terms of its sources, yet when you consider the relevancy of the results ranking in comparison to comparable library services, particularly integrated discovery layers, there’s clear evidence of how systems moving beyond a reliance upon human-generated metadata are able to adequately organize and provide access to information.
The perils of tagging works into a compartmentalized set of terms are fast becoming more trouble than their worth. It’s not just deadnaming in name authority records, as we are all a ship of Theseus in some form or another, but the constantly changing ways in which things are described. I also say this while recalling getting stumped, as a new Midwesterner, by being asked “Where’s the bubbler?”
Those bemoaning the latest advent of the singular they as an example of a grammatical convention that should be unchanging, well, thou art mistaken. As a non-dead language, English continues to evolve, meaning that the need to constantly update descriptors to cover an increasing array of historic terms, let alone imposing them in a top-down fashion, might at some point become unsustainable.
I’m sure Naomi Wolf, who suffered from a pretty major and public whoopsie daisy over her misinterpretation regarding the use of the word “executed” and “death recorded” in historical law enforcement documents to mean “prisoner killed” in what were rather apparently only cases of “non-capital punishment being carried out” could testify as to the dangers of working with a set of constantly changing linguistic descriptions.
It also makes me think about how one of the arguments against time travel ever being possible is simply that because it could create impossible situations, such a concept must therefore not be possible. It’s somewhat fatalistic to give up any attempt at making value judgments about naming major subject categories merely because they could always potentially offend someone, but again, considering how other search engines effectively dodge this issue, perhaps there are alternative methods which our MARC-era-based ontologies are unable to take proper advantage of.
The way that you can, in some databases, map your keywords to subject headings has it half right. Instead of forcing people to use a thesaurus of terms before they can even start a search, it obeys the adage to favor recognition over recall in UX by allowing people to use their preferred language. The next step, as exemplified by commercial search engines, would be to map keywords to, well, keywords, or at least hidden categories without making them front-facing search tools.
Put another way, who needs a concordance once you have grep? Now, professional indexers, unsurprisingly, often assert a superiority over their mechanized would-be successors — given current technology, rightfully so at times, but in other ways it can come across as protesting too much. We shouldn’t be so afraid to get rid of traditional methods. After all, disciplines unwilling to embrace innovations have a tendency to go before their time.
Thinking again of the shuttered current periodicals sections in many libraries, it’s strange how we look at waning practices in the periphery of our field, such as typesetting and whatnot, with a reasoned acceptance that others have to change with the times, yet often fail to make the connection that we also need to consider new ways of doing things ourselves (see also: staffing a reference desk, offering one-shot instruction sessions, and perhaps even facilitating discovery).
It’s kind of like the widespread delusion that death is something that only happens to other people. As stated by a character in the latest Neal Stephenson novel, “We are all mortal, and we differ only in the extent to which we ignore that fact.”
