100,000 questions

Wed Oct 30 2019 20:48:08 GMT+1100 (Australian Eastern Daylight Time)

Normally I like to think for a long time and do some research before I publish a blog post. Which is kind of ironic given the post that has definitely been read by the most people was something I bashed out in about 20 minutes. I'm breaking my usual practice partially because I'm feeling guilty about not having posted anything this month for GLAM Blog Club, but also because I noticed this week that some of the smartest Australian library bloggers I know are suffering from perfectionism and leaving their thoughts inside their heads instead of sharing them. So I suppose this is a way of leading by example. All of which is to say that the following is a collection of thoughts I haven't spent a lot of time turning over in my head. They're liable to change, but I hope to publish an update when that happens, assuming they've changed in an interesting way.

I've read quite a bit about AI, machine learning and how they relate to libraries in the last couple of months. Mostly because it's very relevant to my interests, and partially because I'm doing an excessive amount of research in preparation for a possible future blog post (see above). Something noted by many authors - including Safiya Umoja Noble in Algorithms of oppression and Matthew Reidsma in Masked by trust - is the habit of Google to describe what their eponymous search engine does as "providing answers" rather than "providing search results". Reidsma in particular comments on the audacious hubris of describing whatever information is returned by an algorithm processing a string of keywords as an "answer". I'm certainly not shy about engaging in a bit of Silicon Valley bashing when the mood takes me, but that's not really what I want to explore here. Assuming that whatever the machine spits back at you is an "answer" is obviously foolhardy. But I've been thinking for a while that there are things that librarians can learn from Google (another partially-drafted blog post) - but that the things we think we've learned from them are actually the wrong things.

One aspect of this is the idea that people come to libraries seeking answers. Of course, in a public library context we know that different people (or the same people at different times) come to libraries for different reasons and with different needs. But what I'm specifically thinking about is more typically aligned with academic libraries - what might traditionally be termed people with an "information need" or a "reference question". Basically anyone doing research, broadly defined. Despite the fact that we call this a reference question, the point of the library in this context is usually not actually to provide answers. What they're primarily doing is generating more questions. Students researching an assignment topic, professors investigating an area of research, and retirees researching their family history are all really exploring the space of all possible questions they could be asking in order to decide which ones are the most important questions to ask. A nanotechnology researcher is unlikely to find an 'answer' they can pop into a scientific paper, in the library. But they may well be directed towards new research questions. A poet won't find the answer to how to write their next poem - but they may be inspired to ask different and new questions about the world, or about how poems could be constructed. A history student might find some answers to specific historical questions, but they are more likely to seeking guidance on what questions to ask of the texts they are reading, and what answers are missing. The answers are found elsewhere: in the lab, in the archive, or in the scholar or poet's own head.

I think this is an important thing for us to consider when we think about what libraries, and library discovery and search systems, are for. As Reidsma notes: Librarians spent a lot of energy fighting Google, declaring that they could never be as good as us. And now we spend a lot of energy telling people that our systems are just as good as Google. There's a Neil Gaiman quote you almost certainly know - you've probably seen it on a poster in a library, or perhaps emblazened across a tote bag:

Google can bring you back 100,000 answers. A librarian can bring you back the right one.
Neil Gaiman

I've always hated this quote. It's snarky, and a false equivalence, and in many cases simply incorrect. But more and more I'm wondering whether it actually is worse than that. What if the biggest problem with the Google comparison is that libraries aren't actually in the "answers" business at all? The other quote you probably know on this topic is Roy Tennant's quip, "Librarians like to search, everyone else likes to find." (though this has probably been re-circulated out of context). The pushback on this I've tended to see is librarians saying of course they "like to find" just as much as library users. But what if it's not true that nobody else likes to search? This is what research is, after all - an endless search for truth. Some people say that all of life is a search for something - truth, joy, beauty, meaning. And we all know what people who think they already have all the answers are like.

I'm not sure any of this is quite right - like others, I'm still searching. But perhaps I'd be happier with Neil Gaiman's statement if it was more like this:

Google can bring you back 100,000 answers. A librarian can help you to find the right questions.