So at the moment I'm in library school, which means I'm spending a lot of time and energy thinking about things like cataloging and classification schemes. And because of the way my schedule was in the last semester on Wednesdays I was working for the University library learning Sudoc all morning and working for a specialty research library generating bootleg Library of Congress call numbers all afternoon. Sometimes, my supervisor at the research library and I will get to talking about call numbers and how they work and what they do and why, and we generally come to a few conclusions:
1) While it's nice that we can use call numbers to assume things about the resources (books, usually, and I will likely slip into just calling them books) they refer to, the numbers cannot and do not actually tell you that sort of thing
2) The PointTM of call numbers is to Put The Book In The Right Spot And Nothing Else.
And I had Sudoc rattling around in my brain and I was losing my mind a bit from generating call numbers and I had a Thought. So when I caught the bus home, I texted my friend:
So obviously a library call number can be anything for any reason, some libraries use accession numbers cause theyre fucking crazy. This means that theres no guarantee that a library call number contains information. However, most libraries use something like dewey or sudoc or loc or author last names. The question then becomes do these call numbers contain information. I posit first that the answer is yes, but not because of any intrinsic quality of the call number or even call number generation, and in fact may mean no because its less information within the call number and more information associated with the call number. This is because I further posit that information has direction of some sort, in the sort of way where I can use metaphors to say that information leaves traces behind that can allow us to reconstruct what information existed from sources that do not contain information nor communicate it
to which she responded that she had no idea what I was talking about and that it sounded like a lot of handwaving was happening. Which is fair, to be honest.
A few weekends later, after I had the chance to think about it more and add in a bunch of potentially nonsense about The Exception Proves the Rule in specifically archaeological contexts and decide that everyone is crazy for hating on Sudoc, we managed to get together and discuss it in person for a minute. And in this discussion we hit on a couple of major points:
1) I don't actually know what I'm talking about quite yet.
2) It kind of sounds like I'm claiming that words don't actually have meaning.
3) It kind of sounds like I'm claiming that the meaning of information comes from some sort of Plato's Costco situation.
4) Not entirely relevant to the situation at hand, I need to write an entire paper on how I feel about paradoxes and why I think thinking that paradoxes are bad is bullshit.
So I still don't know what I'm talking about so I decided to harass an actual philosopher about it because clearly that's the logical next step. And in that conversation we came to a couple of interesting conclusions.
In short, @everyone I've spoken to about this: none of you are helpful. Also at some point someone needs to let me finish a damn sentence and maybe I'll be able to think about things. I'm sort of teasing. But only sort of.
Anyway, I still have the problem of I'm pretty sure I'm thinking about something here, and I don't know what the fuck it is, so I'm going to write about it. So if you're interested in watching my descent into madness, stick around, I guess. Here are my goals:
1) I'm gonna read the Quine and figure it out.
2) I need to figure out what I was going for in that text, whether it's right or not.
3) I need to figure out whether I'm right, or sort of right, or not right at all.
4) I'm going to be honest, at some point this will likely include a contrived excuse to explain Sudoc in excruciating detail.
So I'm going to be honest with you all. When I wrote that text, I was anxious, caffeinated, half stoned on data entry and fighting cataloging software, riding a bus full of nicotine smoke and undergrads, and preparing to drive four to five hours depending on traffic home for Thanksgiving. The odds of me thinking or saying anything actually coherent were low. And the current time is 1:40am and I haven't eaten real food in uhhhhh... a day and a half? and haven't slept in like 20 hours. So the odds of me saying anything coherent now are still low. But I'm going to take this text sentence by sentence and attempt to explain what I think, a month and change later, I was trying to say.
So obviously a library call number can be anything for any reason, some libraries use accession numbers cause theyre fucking crazy.
This is straightforward. At least to me. Library call numbers can in fact be anything, for any reason. Usually library call numbers are generated via dense lists of rules based on the sort of library you want to set up. Library of Congress classification and the Dewey Decimal System, both popular in American libraries, have rules based on the subject of the book. Sudoc on the other hand is more interested in the publishing agency and the type of content being published as opposed to things like subjects. But I could assign them call numbers based on page count. Or cover color. Or vibes. Or anything at all. You can do whatever you want forever.
An accession number, for those interested, is the number assigned to a book at its purchase, usually in chronological order. The research library I work for assigns accession numbers as (year).(three digit code that ascends chronologically by donation event/person/budget line).(three digit code that ascends chronologically by item being touched). Because my project involves cataloging random stuff I find on the shelf that isn't in the system, I assign a lot of 2024.020.0XX type accession numbers, because the year is 2024, "I specifically while doing this project found it on the shelf and can't find it in the catalog" is 020, and I'm counting the books as I go.
This means that theres no guarantee that a library call number contains information.
And this is the first part where everything falls apart. What is information? I think the assumption I was making is that, because most libraries use something like LoC, which is based on subjects (at least in theory), you can determine the subject of the book by call number alone. For example, a book with a LoC call number starting with QA is probably about Mathematics. Or a fiction book with the call number Patterson is probably written by someone with the last name Patterson. Or suchlike. However, you can't do that with things like accession numbers. The only thing an accession number might tell you is the order that the cataloger touched the books in. So, to answer the original question, I think what I meant by information here is some fact about the contents/subject of the book beyond the fact that it belongs on some particular space on the shelf.
However, most libraries use something like dewey or sudoc or loc or author last names.
This is just a list of classification schemes. Dewey is short for the Dewey Decimal System, which is used mostly for nonfiction books and assigns a three digit subject code to the book, then a potentially infinitely long subject clarification containing location, date, and specialization information. Sudoc is short for Superintendent of Documents and is a system published by the USA government for the purpose of their Federal Depository Library program. It groups items by department of publication as opposed to subject. LoC is short for Library of Congress Classification, which is another subject based system ran by the American Library of Congress. And Author Last Names are a common way of doing fiction books otherwise organized by genre in public libraries. There are other ways of doing fiction--notably LoC organizes them by original language in the P section, and Dewey throws them all in the 700s and 800s.
The question then becomes do these call numbers contain information.
Friend. Buddy. Pal. What the fuck do you mean by information? Based on the last time I used information, I think I am again talking about some fact about the subject of the book, or really anything that isn't where the book belongs on the shelf. I think I am asking whether a call number based on a rigid, useful, rule based scheme can tell you about the subject of a book.
I posit first that the answer is yes, but not because of any intrinsic quality of the call number or even call number generation, and in fact may mean no because its less information within the call number and more information associated with the call number.
Well that's a fun and straightforward sentence. Said many of the people I text on a regular basis.
I think I need to go back to the last point and discuss what I mean by "contain" as well, because I don't think this makes sense without that. I think what I'm really answering in this sentence is whether an isolated call number can tell you subject information. If I just said CS71 .H205 D3 2021 vol. 1 to you, would you be able to tell me anything about the book? I think (think) what I'm trying to say here is that it really looks like you can tell me something about the book with that call number, but not because of the call number itself. You know things (or at least I know things; I don't know what you know about bootleg LoC) about the book because you know the rules of call number generation. It's CS71 because it's a Family History. It's H205 because it's written about a family with a specific H last name. It's D3 because it's written by someone with a name starting with D. It's 2021 because it was published in the year 2021CE. It's vol.1 because it's the first volume in a series. You don't know that it's a family history because it's a CS71; you know it's a CS71 because it's a family history. If you don't have that context, the call number is meaningless.
This is because I further posit that information has direction of some sort, in the sort of way where I can use metaphors to say that information leaves traces behind that can allow us to reconstruct what information existed from sources that do not contain information nor communicate it
Is it bad that I think this sentence is extremely obvious?
The image that was in my head as I was forming this extremely obvious sentence was of like a Thing that contained information and a Person that was trying to learn that information, and the path that information was taking from the Thing to the Person. And it went through the call number. It's like I was saying for the last sentence; you don't know a book is a family history because it's a CS71; you know a book is a CS71 because it's a family history. The family history association is a trace left behind in the call number that you can interpret only if you know the rules that the research library uses to generate its call numbers. This is also where the Exception Proves the Rule relation comes in. Imagine it's several hundred years in the future and you're trying to determine if it was legal to sell alcohol in the state of Minnesota. You find a law on the books that says something to the effect of "Thou shalt not sell alcohol on Sundays, you Godless Heathen." From this law, via the Exception Proves the Rule, you can say that yes, you could sell alcohol in the state of Minnesota, because they had to specifically prohibit you from doing so on Sundays. So while the information does not exist within the law, or the call number, you can use outside clues and reasoning to guess at that information from the law or call number.
To summarize, I think a meant a few things:
1) The only thing a call number can actually tell you is where a book goes on the shelf
2) Anything you think a call number can tell you about a book outside of where that book goes on the shelf is not coming from the call number, but instead some sort of Holmesian deduction based on external knowledge of some sort.
3) This Holmesian deduction works because information originates in the thing that proves it (the subject of the book, in this example) and then can move through things that have nothing to do with it (the call numbers) to be imparted upon a learner/observer/idiot.
And now I have a couple of questions for the class (me):
1) Why is this important?
2) Are the premises I've based this on in the library sphere dependent on the research library's conception of call numbers? Are those conceptions right?
a) What about that K0.1 we found in the University library? What does that have to say about all this nonsense?
b) What about Ranganathan's Philosophy of Library Classification's University call number being "wrong"? How does that affect things?
c) What about other classification schemes? Does this all fall apart if I apply it heavily to accessions, or Sudoc, or author last names, or something?
i) There's something to be said about how fiction at the public library I work for is only mostly by author last names, and sometimes by anthology titles or suchlike.
ii) Also what about those children's book subject codes? The Great Thomas the Tank Engine "Go" Versus "Character" Debate would like to enter the chat.
3) Do I dare get Ranganathan's Philosophy of Library Classification involved? How does how he talks about call numbers play into this?
a) For everyone's sake let's hope the answer to this is no, who cares.
b) Okay but the stuff I was talking about with my friend that entered into words don't mean things is sort of based on this, and how classification schemes can maybe be considered an artificial language that requires translation and all of that, and that would be super cool to argue about!
1/2/25: I finished reading Ontological Relativity at the university library job today. Nobody made fun of me for it but I suspect I'm going to get an irritated text from my friend because I was texting her a play-by-play. I do actually have thoughts now, but I feel like it would be best for everyone if I got the most egregious sassiness out of the way now.
Number of times I called Quine a bitch: Too many
Number of times I called Quine a motherfucker: Also too many
Favorite annotation: That'd be a toss-up between "holy mother of fuck get a hobby" and "i have a lot of patience for you, sir"
How much I regret the choices I made that brought me to this point: ... okay look
Okay, with that out of the way, let's actually talk about this motherfucker. I'm going to be honest; I am once again dangerously sleep deprived and low on food and even if I wasn't I am absolutely lacking the required grounding to make this make sense (what the fuck is a Godel number and why does half this paper involve the ontology of math). But I think I've got the shape of it, sort of.
So as far as I can tell, this paper comes in three parts:
PART THE FIRST: Where does language come from and how does it work? Or, I'm going to beat Quine with a stick.
PART THE SECOND: Let's get specific about linguistic relativity! Or, actually, Quine might be onto something and this is interesting.
PART THE THIRD: Let's apply all this to ontology! Or, this stopped being about language and therefore stopped being relevant to this project so I stopped caring. I'm sure it's fascinating. I feel like at one point I wrote a different paper about it? I dunno.
Okay I found the paper, and while it is about Quine and existence it's not about this paper specifically. Fortunately, my general thesis was "Quine is wrong about language in lots of fun and sexy ways" so we all know I experienced 0 character development in the last 4 years.
Quine is stealing a model of language acquisition and function from John Dewey, who, just so we're clear, is a different Dewey than the one responsible for the Dewey decimal system. The idea is that you learn a language from the behavior of the people around you, which is fair, and that means that language has no inherent meaning beyond how people use it, which is a dangerous claim that I don't like. And I can absolutely see how he got there. It's not like people are tapping into a fifth-dimension meaning archive to get the True and Accurate Meanings for each word they use ever. Words are used by people, defined by people, and evolve from the force of people. In fact, my complaint is born much of the fact that Quine died in the year 2000 and therefore did not experience the hell that is the internet.
Excerpt, Journal of Philosophy 4/4/1968, page 188:
The color word 'sepia', to take one of [Wittgenstein's] examples, can certainly be learned by an ordinary process of conditioning, or induction. One need not even be told that sepia is a color and not a shape or a material or an article. True, barring such hints, many lessons may be needed, so as to eliminate wrong generalizations based on shape, material, etc., rather than color, and so as to eliminate wrong notions as to the intended boundary of an indicated example, and so as to delimit the admissible variations of color itself. Like all conditioning, or induction, the process will depend ultimately also on one's own inborn propensity to find one stimulation qualitatively more akin to a second stimulation than to a third; otherwise there can never be any selective reinforcement and extinction of responses. Still, in principle nothing more is needed in learning 'sepia' than in any conditioning or induction.
That's how my professor talks about training AIs.
Like actually. The "many lessons" is just a huge data set, and all the "wrong generalizations" are about things you need to consider because computers are fucking stupid, and the "inborn propensity to find one stimulation qualitatively more akin to a second stimulation than to a third" is a rating of the merits of one type of AI against another.
I'm tempted to claim that Quine is obligated by this behavior model to argue that AI actually understands language. Especially when he gets into all that bullshit about rabbits versus undetached rabbit parts and how there's no real way to tell the difference between the two, even more so when he starts talking about how it doesn't really matter since homophonic translation is occurring between two people speaking the same language. Because functionally, while the human eye sees sepia as a shade of brown, the AI will 'see' it as a range of percentages of mixing red, green, and blue, or a specific set of hex codes, or something. What's the difference between a color composition makeup and a collection of undetached rabbit parts?
Obviously, to me at least, thinking (current, not sf) AI has any ability to understand anything is a Problem, one which I am Not willing to allow to exist in my view of the world.
Is there a way to say that no, AIs don't understand language while also saying no, there isn't a fifth-dimension linguistic archive? Probably. 'Understand' is doing a lot of heavy lifting for me and I'm not really defining it beyond 'it looks like the AI is doing all the bits Quine says you need for proper language usage' and it should be noted that 'proper language usage' does not necessarily equal 'understanding'. By adding some (magic) condition to my 'understanding' or his "learned" we might be able to sidestep my complaint.
That's not to say I don't have any other problems with this model. I'm not convinced that languages can't be individual, for one thing, which is an important pillar on which this argument stands. I'm also not entirely convinced this model holds up to the stress test that is irony and sarcasm, as well as Janus words and shades of meaning--if language is grounded in behavior, why can one word mean the exact opposite thing at different times?
But really whether Quine's right or not is a different paper. What I'm worried about in this rambling blog post pretending to be academic work is whether call numbers contain meaning. And I think, based on Quine's behavioral model of language, the answer is no.
And see the thing is is I might actually like Quine's Great Relative Universe. He talks about it as a coordinate system, where your theory can only be meaningful within the bounds you set--intentionally or not. The example he gives (204) is of some unknown theory involving natural numbers (1, 2, 3, 4, 5...). It doesn't matter for the theory if the function of 2 is taken on by the symbol 103. The only reason that 103 is not behaving like 2 is because everyone who cares about the theory has agreed that 2 behaves like 2 and 103 behaves like 103. And I do like that. It sounds to me like it's at least adjacent to my conception of context as a necessary condition in philosophical arguments. With the way Quine is using it, though, it seems to me like all of the things that make words mean anything are part of the context and not the words. The words mean things only because we agree on a model where they pick out pieces of a different model that we agree probably exists. So call numbers also can't mean anything in and of themselves because any meaning they appear to have is actually a piece of the agreed upon model of classification schemes and library set ups. All the meaning is held in the deferred ostensions.
I feel like there's something to be said here about call numbers not behaving properly under the behavior model as well, making them not a language, but I am having an impossible time phrasing it in ways that make sense and don't character assassinate what Quine is saying so I'm just going to say it Irks me and I'll worry about it later.
At the end of the section elucidating the text I listed 3 (heavily expanded) questions that I want to answer. Having now read the Quine thing, spoken to another person about this whole mess, and generally had time to think about it, I'm relisting the questions I want answered.
1) What is a language?
2) What is a call number?
3) Are call numbers a form of language?
4) What is meaning/information?
5) Do call numbers contain meaning/information (beyond where a book goes on the shelf)?
a) If yes, how? Why?
b) If no, then why does it look like they do?
Lovely, easy goals. I'm sure I'll be done with them before I leave work in 2 hours.
Let's Talk Language
I'm gonna be honest, friends; I don't think classification systems are a form of language.
I feel like one of the unexamined contextual models Quine was using was a definition of what language is that we all accepted upon reading his paper, like a checklist of some sort that gives us the traits that make something a language. And that seems kind of important. And as much as I'd love to, that definition cannot be "Languages are systems that convey meaning" because I personally will rip the shit out of that. Did you know wind patterns are a form of language? They tell us when it's going to rain next! Absolutely not.
Quine spends a hell of a lot of time talking about translation. So, for a stupid, somewhat problematic start, a language is something that can be translated into and out of English (or anything else we're certain is a language).
The Library of Congress website contains, free for public use if you really hate yourself, manuals on how to generate a call number using their Library of Congress Classification system. We'll call that a translation guide. Most if not all libraries have their own (much shorter and easier to read) manuals on how to generate call numbers. We'll call those translation guides, too. So, using the bootleg LoC I'm most familiar with off the top of my head, you can translate History of the Hanson family written by Sven Andersen in 1982 into CS71 .H2489 A5 1982. And you can obviously translate it back; that's why I'm writing this not-paper. Well, not quite. The back translation will lose the person's first name, and potentially most of their last name; all the call number tells us is that his name is A-N-something, and that specificity is luck of the cutter table.
It is significantly more difficult to find Sudoc manuals online; most are for finding and shelving resources as opposed to generating the numbers because again, why would you read those if you don't hate yourself. But I know the internal wiki of the university library has several sudoc resources; let's call those the translation guide, and then accept that the numbers and letters I'm about to use are completely made up. (It's a contextual theory for the purpose of this interior theory!) I'll be using NC State's library guide as a reference if you want to follow along. I guess that's also a translation guide. So the US Geological Survey's second 1935 handbook in English can be translated into Sudoc as I19.8:935 2. And a back translation follows as well.
So it seems like both LoC and Sudoc are certainly languages. That means we should be able to translate them into and out of each other. And that's brick wall central. LoC is based on subjects while Sudoc is based on publication entities. And while the US Geological Survey probably has a specific cutter number in most library systems, we don't have a subject for the handbook. LoC to Sudoc is even worse; CS71 means nothing to them.
But this isn't really a fair comparison. Both LoC and Sudoc fail much earlier in the process. Neither can handle the sky is blue. It's beyond their scope. Okay, so we'll limit the scope. A language is something that can be translated into and out of another language limited by the content of the "smaller" language.
My mother's knitting patterns can be translated into and out of English. I don't speak knitting pattern, but she certainly does, and those are definitely symbols that she's reading as knit one, purl two or whatever. Are my mother's knitting patterns language? 'Cause that would be fucking hilarious, although now that I've said it I can't think of any good reason why they shouldn't be and I'm mad about it.
But at the same time, you still can't translate LoC into Sudoc, and you definitely can't translate knitting pattern into LoC. I think I need a different definition.
So what the fuck is a language? Quine also seems to be implying that--whatever the fuck form reference takes--language seems to be referring to objects or concepts or something. Take that sentence kindly; you know what I'm trying to say. Let's try that-- Language is something that is used to refer to objects/concepts/whatever.
I think knitting patterns still count as language, since they refer to the concept of stitches. Just a note.
But what would call numbers refer to? Just so we're all clear, I am NOT writing another fucking paper on the nature of reference. NO. Go read someone else's. Because here's the thing. It looks, on paper, like the reference should be to the subject, or the publication agency, or whatever. But that's not... right. If I'm out in the wild and see CS71 .H2489 A5 1982 written down in isolation I don't immediately think "oh, family history of someone named Hanson," I think "oh, I'm supposed to find a specific book in the CS71 section of this library." The reference of the call number is to the book itself, not whatever concepts may or may not be involved in the number.
This doesn't seem to make call numbers not language, by the way. If you're keeping track at home.
This probably has issues in it somewhere which will be pointed out to me by my insomniac brain at three am and I'll come on here and ramble out another paragraph, but I actually kind of like this model for weird metaphorical reasons. Cause there's a bit of a names thing going on here. And look, names are complicated and stupid and frustrating and honestly we should just do away with them so I can stop thinking about it, BUT I think there's a parallel between call numbers and pet names. Or human nicknames, actually, but don't take that out of context. Cause, see, if you name your cat Steve, no one is going to assume anything about the cat. But if you name your cat Midnight, people are going to assume you have a black cat. If I name my white cat Midnight, people are going to glitch out about it. But there's nothing in the name Midnight that means it can only be applied to a black cat. You can't stop me from naming my white cat Midnight. In the same vein, if a book call number is 2024.020.015 you're not going to assume anything about it. But if it's DL596 .C12 D45 2021, you (assuming you know what that means) are going to have a couple ideas in mind about the content of the book. But why? There's no law of the universe that says you can only make Books Of The DL596 Subject DL596. I can make fucking Artemis Fowl DL596 if I want. (Which actually raises another question about Quine's Behavior Bullshit: why do people subscribe to other people's linguistic behavior corrections? "Because they do" is a stupid answer, and Calvin would like to have a word with Quine, I think. This aside should have been hover text, but putting links in hover text is mean.) I think most people's argument against this comparison is that there are, actually, rules stopping me from making Artemis Fowl DL596, and those rules are elucidated in those manuals I called translation guides earlier. Which is fair, to an extent. But way back in the list of questions for the class I mentioned a K0.1 we found in the University library. K is supposed to be the subject code for Law, but 0.1 is meaningless. It turns out that 0.1 is (potentially in-house rules only, I'm not sure) shorthand for "holy fuck I'm overwhelmed, not supposed to be working in this subject class, or otherwise incapable of doing this properly right now, so I'm giving it a random number so we can find it and fix it later". And the point of giving it this number is to make sure that all the books that need fixing are together on the shelf. I see two possible solutions to explain this away.
1) 0.1 is ackshewally a linguistic marker with no meaningful content/is too divided into disparate parts to actually be meaningful or whatever the fuck Quine's Japanese example showed/demonstrates that call numbers contain meaning even beyond the subject code!
2) Call numbers can be whatever the fuck you want, and the rules are more of just guidelines anyway.
But I actually think I might be getting off topic. Because while nitpicking this theory to hell and back is fun, the point of this exercise is to figure out how call numbers look like they have meaning outside of "this book goes at this spot on this shelf". And in this set up, that meaning appears to be coming from the fifth-dimensional linguistic archive. And while I don't necessarily have a problem with the fifth-dimensional linguistic archive, I feel like I should explore other solutions so that zombie Quine doesn't come after me.
I could probably keep coming up with definitions of language that get more and more contrived until we're all happy and none of the counterarguments bother me, but instead I want to touch back on the idea I've mentioned offhand a couple of times now: call numbers aren't actually language.
Let's Talk Art
Okay, to start with: yes, I'm aware that language can be considered art, and yes, I'm aware that art can sometimes be considered language. For the context of this paper, I don't care; shut up.
I don't actually know where the "call numbers are a form of language" assumption came from. I don't know if it's an effect of the reading and thinking I've been doing or an actual intuition I've had this whole time, but saying "call numbers are language" is deeply, deeply weird to me. But every person I've spoken to about this has immediately gone the language route, 2/3 of them immediately comparing it to coding languages and the third (who knows almost nothing about coding) asking about natural words, so I assume either I hadn't noticed I was making the assumption and everyone else just went with it, or I said "meaning" and/or "information" in the same sentence as "conveying" and everyone went "ah, yes, the thing languages do."
So just for a moment here I want to stop viewing the rules of call number generation like the grammar to some artificial language and start treating them like a recipe for a cake or something. You follow the recipe and then arrive at a creation. That's really how generating call numbers works; you have a list of actions and you apply them in the correct order. How to make Macaroni and Cheese: make pasta. Dump in pot with a bunch of cheese and milk and assorted other nonsense. Cook for three hours. How to make a Bootleg LoC Call Number: Alphabetize by family. Number alphabetically. Massacre some word from the front of the book, ideally author's last name, using a Cutter table. String together. Shelve. And if, at the end of the process you look at your macaroni and say "ah, yes, this dish tells me that it contains evaporated milk by taste alone" everyone around you looks at you like you're fucking insane. Because you are. But looking at a call number and saying "ah, yes, this number tells me that it belongs to a book about Grecian logicians" is apparently fine. And alright, sure, fine, you can determine the recipe of a dish by taste and texture and close examination assuming you have the skills and experience to do so, but "assuming you have the skills and experience to do so" is doing a loooooot of work in that sentence. Not having those skills will not in any way detract from your ability to find use in the art you're examining. If we expand "art" out away from things that had a formula to create them and into whatever stereotypical definition you'd like, there's even less guarantee that whatever message or fact about the world or however we're planning on defining information this time is present in the art itself, as opposed to within the person doing the interpreting. I can't engage with the Good Omens fandom on the internet anymore because the message I got out of the last episode of the second season is fundamentally different than what anyone else got out of it. And I think getting subject information out of call numbers might be like that. By saying that BC28.2 means Grecian logic you are in fact calling on outside knowledge and outside experience, and it's sheer dumb luck that your Library Fandom interpretation is the same as everyone else's. There's nothing inherent in the number that makes you right.
I'm going to be honest; I don't think this digression really got me anywhere either, except one place. The more I'm thinking about this, the more I think this really is about trusting librarians and intent.
Let's Talk Trust
Just so we're clear, I am NOT advocating to not trust librarians. We do good work, and we are generally not particularly nefarious.
But also there are a lot of things that can go wrong in call number generation.
Errors of Librarians Are Human
Generating call numbers involves a lot of attention to detail across a huge collection of numbers that mostly look the same. Like imagine you're looking at
| 2024.001.034 |
CS71 .H254 J66 2014 |
| 1997.025.354 |
CS71 .H579454 L3 |
| 2005.156.001 |
CS71 .H524 T8 1905 |
for several hours a day, only your spreadsheet is way longer (mine are generally between 20 and 30 entries) and the numbers more annoying. At some point you are going to type H253 when you meant H523 and you are absolutely not going to notice. In my bootleg LoC, that's a whole different subject. Now, assuming that the call number in the catalog and the call number on the book are the same, odds are you'll be able to find the book you're looking for, and everything is fine. But at the same time, this number can be considered incorrect. And if the book and the internet DON'T match, then you're in trouble.
Errors of Call Numbers Are Hard
So I know that I've complained a bunch about the pdfs on the Library of Congress website but I doubt all of you are going to click my links and see for yourself how awful they are, so I'm screenshotting part of the Bs (which is the pdf linked) and forcing you to look at it.
Quite honestly, with how awful this bullshit can get, this specific example isn't that bad. I picked it because I've been using Grecian logicians as an example this whole time and why stop now?
So the way this works is that you figure out what your book is about (Grecian logic, for example) and then you find the section in the pdf that tells you what to do with it (28.2). Where thinks get tricky is when your subject is something like "Insanity" which I picked because it's not on the list above "Grecian." Alphabetically, "Insanity" comes between "Inference" and "Insolubilia", so you'd give it a number after 26 .I6 but before 26 .I64. And that's where things get weird. If I'm Head Cataloger and I decide "Insanity" is 26 .I63, then every book about insanity should be cataloged as a 26 .I63. But if my Associate Head Cataloger doesn't get the memo and makes "Insanity" 26 .I62, the library is now inconsistent. Error.
This is going somewhere, I promise. Then again, I'm still getting there faster than fucking Quine, so
Errors of Books Don't Like Sorting
In front of you are two boxes. Box A is labeled "Books about cars, trucks, airplanes, trains, tractors, et cetera." Box B is labeled "Books about characters from movies, television shows, other books, and video games." You are holding a book about Thomas the Tank Engine. You are required to put it in a box. Begin at your leisure.
The majority of the people I've spoken to think it should go in the characters box. The catalogers at the library put it in the locomotives box. Are they wrong? Not really, I guess.
Or, more to the point, in front of you there are 100 boxes, each with labels ranging from "slightly different" to "holy fuck that's a whole new sentence." In your hand are 1000 books. You are 900 books behind schedule. Another 1000 books are coming your way. Put them in the correct box. Physics 101 is easy to sort--there's a box labeled "Physics." But where do you put fucking I don't know, Thucydides? "History of Greece"? "Primary texts related to Greece"? "Greek language"? What about the Iliad? "Poetry"? "Classical literature"? Make the decision FASTER, you are BEHIND.
Incidentally, this is probably how that K0.1 I was talking about showed up.
So what's my point? My point is that we can argue all day and night about whether or not call numbers contain information in a philosophical sense if we want to. Unless we consider the outside mechanics, we're not going to get anywhere. Because at this point I've spoken about this project to 6 different people, plus me, plus the supervisor that started all this. And I'm seeing a pattern in the gut instincts that tells me something interesting.
| 1 |
Friend; uses library |
Yes, it seems like call numbers contain information |
| 2 |
Philosopher; uses library |
Yes, it seems like call numbers contain information |
| 3 |
Library Colleague |
No, it does not seem like call numbers contain information |
| 4 |
Library Classmate 1 |
No, it does not seem like call numbers contain information |
| 5 |
Library Classmate 2 |
No, it does not seem like call numbers contain information |
| 6 |
University Library Supervisor (not The Supervisor) |
REALLY no, it does not seem like call numbers contain information |
| + |
Supervisor who started it all |
REALLY REALLY no, it does not seem like call numbers contain information, are you insane? |
| + |
Me |
... <gestures at this webpage> |
People who use the library are totally willing to believe that call numbers contain information. People who maintain the library absolutely are not willing to believe such nonsense. And I have to wonder if the knowledge of how much (or little) goes into call number generation has something to do with that. Because, well:
Question: If I found a book on the shelf at the research library that was useful, would I look at the books around it to see if any of them are helpful as well?
Answer: No.
Question: Do I trust the call number generation at the research library?
Answer: No. Please don't tell my supervisor. Although there's a good chance she'll agree with me.
Question: Using only the research library as a basis, would I believe that call numbers contain information?
Answer: No.
Question: If I found a book on the shelf at the University library that was useful, would I look at the books around it to see if any of them are helpful as well?
Answer: Absolutely.
Question: Do I trust the call number generation at the University library?
Answer: ... Yes, sure, for the moment.
Question: Using only the University library as a basis, would I believe that call numbers contain information?
Answer: Probably, but I wouldn't like it.
I've listened to too many idiots trying to explain metaphysics to me to really believe that intuition itself is a useful explanation in philosophical discourse, but I think it's a good starting point. Because what this tells me is that somehow my ability to believe that call numbers contain information is linked to my ability to trust that the librarians have done a good job. And I think that tracks (if we return to language for the moment) with English as well. Ignoring all the Quine bullshit of how it got there and if it's there at all, it seems like this paragraph contains information in a way that the famous lorem ipsum copypasta doesn't. And I think this points to the actual question I was trying to ask all those paragraphs ago: where is this information coming from? And I think the answer might be a blend of intent and us as library users making it the fuck up. Because look. Librarians want you to find things. We want you to find what you're looking for, and a bunch of helpful and fun stuff you didn't know you were looking for, and generally make your projects as easy as possible. There is a lot of thought that goes into where a book goes on the shelf. But at the same time, most of that thought happened in the creation of the system and the manuals that explain the system, not in the actual generation of a specific number for a book. And in order to pull that thought back into the specific number you have to interpret it for yourself. The information that appears to be within a call number actually comes from you and your knowledge and your assumptions.