Thread by @lorenschmidt

thoughts on alphabets for the library of babel.

given the constraint of 22 lowercase letters, plus space, period, comma, there is some question about what alphabet the books employ. it is not the normal spanish variant of the latin alphabet, or the variant used for english.

i think it's not entirely clear from the story if the writing system used is a variant of the latin alphabet at all.

one possibility is that it's an entirely nonhuman culture, or a non-terrestrial human culture, which would suggest a nonterrestrial alphabet.

the hollywood / AAA games approach is to appropriate arabic or devanagari or maya script or egyptian heiroglyphs, and mash it up into an "alien language".

doom eternal:

which is both crude and objectionable, given the history of othering in genre fiction. (see conlangs in lord of the rings or game of thrones.)

or you can make up glyphs not based on terrestrial languages.

i'm not sure how i feel about this. i have tended away from it, because it feels so similar to the above. it is still a sort of exoticization. it kind of feels like a reskinned version.

or maybe you can do something more abstract? here are two pure representations of binary numbers. top row is 16 bits per glyph, bottom is 8. these feel like scripts, and they fit the combinatorial theme of the story.

that's one option. but back to terrestrial scripts...

specific terrestrial languages are mentioned in the story. there are also specific terrestrial books and traditions of thought mentioned. but it's entirely possible that these were supplied by a translator rendering an alien text approachable to human audiences.

but Borges has a eurocentric bent, and i think it's entirely possible that he imagined this other universe as being imbued with western cultural artifacts. it is, after all, a device reflecting on human experience, not a self-consistent reality.

because of that, i think it's possible he imagined the world to be built around some sort of reduced form of the latin alphabet. if you do the standard 27 letter latin alphabet used for spanish, you get something like

(prototype)

possibly he intended a somewhat more phonetic variant, with soft c collapsed into s, hard c collapsed into k, and "kh" for che.

i am a poor spanish scholar, and i would want to consult with someone, but i could imagine a series of such reductions landing us at 22 letters +3

so one plausible approach would be a reduced latin alphabet.

but one other possibility, which i think is very appealing, is an abjad, rather than an alphabet. this might actually be a better fit for the underlying ideas of the library. the library is a collision between subjective human experience and pure data, all possible ideas.

abjads are writing systems with only consonants. many modern abjads (such as arabic) have been supplemented with diacritical marks to make the vowels explicit. but not all of them are that way.

completely vowel free abjads are interesting in that they have an additional subjective layer to reading them. for instance the english word "head" might be rendered as "hd", which could be "hid" or "had" as well as "head".

so you might end up with pages of dialogue which look more like this. this is the standard spanish 27 letters minus vowels, plus comma, period, space, which comes out to 25 even.

if you don't mind deviating from the specified character count, there is one additional idea which i find interesting, and that is continuous script, i.e. writing systems with no or reduced whitespace.

en.wikipedia.org/wiki/Scriptio_…

some modern writing systems, such as written thai, are this way, along with older versions of written arabic, greek, latin, etc.

if you were to employ such a system, you'd end up further emphasizing the interpretive or subjective.

(credit: Marcin Konsek, wikimedia commons)

if you retrofit modern latin script, you might end up with something like this. this drops you down to 22 characters, which doesn't match the description anymore. but you could do other scripts.

and i'm not personally interested in verisimilitude. i think the spirit is right.

emphasizing subjectivity is one reason i find this interesting. but there's also a really interesting thread in the history of computation which is extremely relevant, and which also leads here.

Ramon Llull was a prescientific mystic and mathematician. he had what i would describe as a "pattern matching" sensibility, in that his exploration of ideas seems to emphasize likeness or resemblance rather than evidence based reasoning or deriving things axiomatically.

one of Llull's avenues of research was an attempt to find the atoms of thought. he believed that if you constructed the right system, you could procedurally generate all possible knowledge.

a correct system would recreate all known fields of thought, and suggest new ones.

and if you backtrack from LLull, you arrive at an older combinatorial system, the zairja, which seems to have inspired much of his combinatorial work.

the zairja was a divination system with natural language output. en.wikipedia.org/wiki/Zairja

this was before digital computers by hundreds of years, but a series of calendrical and textual inputs were run through an algorithm which produced tables of consonantal arabic, written without punctuation or whitespace. these were then interpreted as a cryptic prophecy in verse.

i like that this requires a poetic sensibility and firmly anchors it in the subjective. there is no sense of objective truth, of meaning outside of subjectivity.

in this context, written language is data and interpreting it is a heuristic. it does not pretend to be anything else

this seems an approach which could connect to the underlying ideas in "the library" more deeply than the original conception.

and i wonder if it might make sense to have localized versions of this for different scripts? that would make it less eurocentric.