This past fortnight has been rather tumultuous in Australian libraries. Like much of the world, we've been getting some lessons about raw power and resource allocation, and mostly weren't ready to hear it.
Clarivate chooses violence
In November 2021 Clarivate completed the purchase of ProQuest for USD$5.1B, which should have been a giant flashing red light to the library profession that ProQuest was likely to be gutted and radically overhauled in the near future.
Well here we are.
Last Tuesday morning Australian time, ProQuest's Vice President of Product Management sent an email to customers announcing a radical change to their business model. It appears that this was also the exact moment that most Clarivate staff and their publishing company partners found out about the new arrangements, due to kick in from 1 November.
Until now ProQuest has offered a mix of ebook subscription packages, "perpetual" ebook licenses for a one-off fee, and also a platform for purchasing physical books. That's all ending. Siobhan Haimé has provided an excellent summary of the announcement and its consequences for libraries, and CAUL has provided an Australasian library perspective.
Anglosphere suddenly remembers the Library of Congress is a US Government entity
The other big library news from last week was the Library of Congress updating their terms to rename the Gulf of Mexico and Denali to Gulf of America and Mount McKinley respectively. The US Government of course has the sovereign right to refer to physical features in the landscape using whatever names they deem fit. It seemed to come as a surprise, however, to some in the non-US Anglosphere library world that the Library of Congress would update their terms to align with the United States Geographic Names Information System. As friend of Information Flaneur Alissa McCulloch pointed out in The Process is the Point, whilst LoC's Program for Cooperative Cataloguing uses language that sounds like it's a democratic international collaboration, ultimately the Library of Congress decides what is and is not a standard term in Library of Congress Subject Headings.
The problem here is that LCSH is the de facto standard for controlled subject terms in libraries across the Anglosphere. If the US President decides that the official US name for Australia is now "The US Federal Territory of Kangarooland", then under current arrangements Australian library catalogues would immediately start using this term, at least for newly added items. Clearly this is suboptimal.
Infrastructure
A few days ago Robin Berjon published a piece called Digital Sovereignty exploring the titular concept and also usefully defining "digital infrastructure". Whilst Berjon generally seems to have in mind larger or more multi-purpose digital infrastructures, his outline here is a really useful framing when considering the different platforms, standards and entities that make up the library and academic publishing ecosystem.
Berjon helps us to think more clearly about "infrastructural power" – the real power able to be exercised by whomever controls the infrastructure, regardless of any rules, norms or understandings. He outlines an infrastructural good test derived from Frischmann and Rahman, which is a rule of thumb for determining whether a system or platform has become an "infrastructural good":
- Hard to replace: The component in question features high sunk costs, high barriers to entry, or increasing returns to scale notably from network effects.
- Highly diverse downstream uses
- Vulnerability: The good is necessary for participation in some social activity (personal, business, public, etc.) and there are negative repercussions from restricted access to it.
If you're a librarian and you're sweating right now, it's probably because that sounds awfully like a lot of our vendor platforms. Sure, if your library doesn't like a journal subscription deal that SpringerNature Group is offering you could always just walk away. But you'd have to be the sort of university where your researchers don't really care whether or not they have access to read Nature or Cell. It's why public libraries end up with AI slop in their ebook catalogues when they would never have selected it for their collection. It's also why even though we know that the companies academic libraries work with flagrantly abuse the privacy and safety of library users and this conflicts with core values of the profession, it's been extraordinarily difficult to effectively create counter-power. Drive-by commentators often demand that we "just stop doing business with" these companies, but Berjon's point about infrastructural goods is that it's extraordinarily difficult to do that, especially alone.
Monopolies are not standards
Perhaps as a start, we could recognise that monopolies and open standards are not the same thing. Admittedly this can sometimes get confusing because things that look like standards sometimes aren't.
Some incredibly useful library standards are OpenURL, OAI-PMH, MARC, and ARK. None of these are controversy-free, which we'll come to in a moment, but they are at least collectively owned open standards that anyone can use without charge.
On the other hand, there are things used widely in libraries that feel like open standards but aren't quite. As mentioned above, Library of Congress Subject Headings (LCSH) are a de facto standard in that they are widely used throughout the Anglophone world, but they're not a truly open standard since ultimately LCSH is a privately-controlled thesaurus of subject headings for use by the US Library of Congress, which they happen to share with the rest of the world. The Dewey Decimal System is the one thing many people think they know about libraries, but OCLC aggressively reminds everyone that despite it being 135 years old you have to pay them a tax if you want to use it. The International Standard Book Number is technically a standard, but if you want to use one you have to go through your "national registration agency". As I outlined in a post two years ago, this results in a monopoly in each country that can independently decide whether they can be bothered innovating.
Even the "real" standards aren't problem-free, because libraries and their parent organisations haven't been serious about controlling or directing them. Notoriously OCLC – the extortion outfit cosplaying as a collective non-profit – was appointed as the "maintenance and registration agency" for OpenURL by NISO in 2006, but 16 years later unilaterally claimed that this core function of every academic library catalogue was merely an "experimental research project" that they couldn't be bothered maintaining any more.
Counter-power
Where we might draw some hope and focus some thinking is in Berjon's second feature of infrastructural goods: Highly diverse downstream uses. This sounds like a positive feature of infrastructure – because it is! But...
The diversity of users also makes it harder for them to coordinate in order to exert political pressure on the infrastructure operator as their needs and communities will often differ greatly.
This applies to some extent in libraries, but I don't think our "needs and communities differ greatly" – or at least, not greatly enough if we focus on the many things we have in common across the various public, academic, corporate and other "special" libraries. That means we still have a place to think about building collective counter-power.
Dan Goodman called for this in The Transmitter last week, saying that Science must step away from nationally managed infrastructure. Goodman was thinking more about datasets and censorship of inconvenient research publications, but he also helpfully noted the problem of both ownership and effective control of the academic publishing industry. And while it's tempting to roll our eyes and say "I told you so", we'll get further by harnessing the energy of these newly-awakened scholars and researchers to build a new collective reality.
For me, "vulnerability" is the most important aspect to address in the Infrastructural Good Test. Lots of technology is "hard to replace". But the thing that gives companies like RELX, Clarivate or even Overdrive their power is that libraries can't see a way to get out of their relationship with this companies without catastrophic damage to their social license to operate in their given context. Your University Librarian certainly has the power to cancel every contract with Clarivate and Elsevier, but they'd be very unlikely to retain their job for long if they did that arbitrarily, no matter how strenuously they argued that it was the ethical thing to do.
If you explain the academic publishing market to a normal person, they don't believe you because it's so obviously absurd. Making academic libraries less "vulnerable" in this context requires a complex series of moves by academic scholars, university administrators, librarians and probably national governments. It's tied up in ideas about research quality, education as an international market, neoliberal economics, elitism, and great power politics.
I'm not a populist politician so I don't have a simple answer to all this. We're facing the equivalent of trying to rebuild an apartment block and replace its foundations while hundreds of people are living inside it. I will say that begging limited-liability shareholder corporations to do the right thing or trying to make them feel guilty (as if a legal abstraction has emotions) isn't going to work.
America First
To heed Goodman's call we need to reckon not just with the "America first" mentality of the United States government, but also our own. "International cooperation" on systems and standards often rests on an assumption that the bulk of the resources, funding and leadership will come from the United States. Logically this approach makes sense, since the USA is home to over 340 million people who collectively (though certainly by no means equally) hold vast hoards of wealth and enormous industrial and intellectual productive capacity. But as the people of Ukraine are discovering, the United States can't be relied upon indefinitely. Waiting for US institutions to lead everything is problematic in other ways, even – perhaps especially – when they're enthusiastic. American software is constantly trying to enforce American spelling, American terminology, American weights and measures, and (worst of all) American date conventions upon the rest of us. Though Goodman is right to point to the huge market share of the US directing the priorities of corporations, the situation isn't helped by the rest of us not bothering to try building something else. We can complain about lack of budget funding, but budgets are statements of priorities and political intent as much as they are about pure finance. If we want something more decentralised and truly collaborative, we can't wait for America to lead the way.
Speciation and multi-polarity
Australians looking for something about a bloke driving his ute with a tinnie on the back through a bushfire while wearing thongs need to be able to use those terms when searching a library catalogue, and get something sensible back as the result.
We don't have to rebuild everything completely from scratch. Especially when it comes to controlled vocabularies, we might think about techniques akin to speciation by building on top of existing systems with local variations. We can use local extensions that both add new terms and define alternatives. We can incorporate specialist vocabularies over the top of broader ones, and use additional vocabularies when more generic subject headings are problematically based in a particular worldview, not nuanced enough, or simply too broad. And perhaps the era of Linked Open Data replacing systems designed for browsing card catalogues is nearly, finally, upon us? Enabling many names to point to the same thing via an opaque identifier would solve some of these problems.
When it comes to project governance, we can also think about techniques to protect them in ways that don't rely on benign or enlightened stewardship. After an attempt at a hostile takeover of KohaILS , the community deliberately distributed project infrastructure in such a way that no single entity controlled more than one part. The koha-community domain name and website, the git repository, and the bug/feature tracking systems are all managed by different organisations to prevent any one entity from asserting control of the entire project again.
Your nearest exit may be behind you
Some of the infrastructure required already exists, or can be resurrected or repurposed. Australia still retains the Australian National Bibliographic Database, managed by the National Library of Australia and separate from OCLC's WorldCat. There is also the aforementioned Australian extension to LCSH which needs some TLC, but has existed for decades. Whilst these are "nationally managed infrastructure", they're also for localised needs, so a national approach makes sense.
Our predecessors confronted some of the same issues, and built systems and solutions to try to deal with them. We have cooperative organisations like CAUL which presents a single voice to vendors in contract negotiations, and CAVAL which does many things, including custodianship of a shared physical collection and a reciprocal borrowing network separate from the national document delivery system managed by the NLA. The UNIMARC standard is managed by IFLA.
Librarians can't escape the fact that most of our libraries are funded and regulated either directly or indirectly by governments. The idea that we could operate with complete autonomy is a fantasy. But there are plenty of precedents for taking our professional values seriously regardless of the political situation we find ourselves within. The founders of CAVAL created a shared collection agreement that deliberately makes it extremely costly to leave the collective. The Koha community found ways to stash power in multiple places. The librarians who built Australia's LCSH extension realised duplicating the Americans' work wasn't a good use of their time, but made sure Australians didn't have to search for "pickup trucks" or "wild fires". What is our generation going to build and maintain as our own legacy?