veiledinviolet - Tumblr blog

[veiledinviolet]

Yesterday I was so sleep deprived I drove off from the petrol station with my cap still open.

Yeah, another weird poly angle I found in my screenshots.

[veiledinviolet]

people like to bring up things like "all your loved ones would die before you" as a concern re: immortality. But the real danger for me would be that i'd literally never get anything done again. I'd adopt the elvish lifestyle immediately. Not doing anything in a week that could be done in a century. Not getting anything done in a century that can be done in 10 centuries. Spending the next 100 years reading 1 book very slowly because i have infinite time and don't care.

[veiledinviolet]

The answer to your problems is self-discipline

[veiledinviolet]

game that starts out as a cozy little farming sim and gradually expands into full-blown factory farming where you're raising livestock in narrow stalls by the millions and have enough economic sway to influence environmental policy-making that may slow your profits

[veiledinviolet]

This is an absurd hypothetical, but I wonder what would happen if the binaries for every (programming language) compiler just suddenly disappeared. The source code and all other software remains, you just couldn't build any new software (interpreted languages do not exist in this scenario). Essentially the question is how hard it would be to go from machine code to high-level languages again, considering that compilers are usually written in high-level languages themselves.

I guess the most promising path would be writing an assembler in object code, a C compiler in assembly and then a C++ compiler in C, at which point you'd get LLVM (/ʎʊm/) back. Again, this is a completely absurd scenario, but somehow it "feels right" to me that there should be actively maintained projects to bootstrap a high-level compiler from the ground up.

[veiledinviolet]

ich bin seit über einer woche nicht mehr bahn gefahren, bald bekomme ich entzugserscheinungen!

[veiledinviolet]

I mean, renaissance does extend to the modern era but started during the middle ages. Discovery of the Americas / reformation are what are usually used as the symbolic cut-off points

no, this is actually completely wrong. modernity is a sticky but diaphanous fluid which began to seep up through a crack in the earth's crust below florence, italy, after the earthquake of september 28, 1453. it adheres vigorously to animal flesh, and minute traces are spread by person to person contact. even in trace amounts, its presence inhibits miracles, spiritual apparitions, and feudalism, and encourages the mercantile spirit, which is why its spread ended the middle ages. it also enables the visual cortex to comprehend perspective, which is why paintings looked so funny before. unfortunately it also tends to accumulate in the inner ear, dampening transcendent vibrations that otherwise are naturally picked up by the auditory nerve, making it impossible for humans to hear the voice of the divine.

that sounds bad, but most of what untainted humans in europe were perceiving as the voice of the divine was actually just the voice of a lesser divinity that pope adrian iv trapped in the roman catacombs, harvesting its body for communion. while non-sentient, the psychic defense mechanism of this creature produced aggression, zealotry, and visions of fire. since modernity is toxic to divine beings, despite the church's best efforts it began to sicken in the 17th century and by the 18th was in dire condition. its death finally enabled the first vatican council to occur.

of course if you ask about this at the vatican they won't tell you the truth. it's not on purpose: when divine beings die all memory of them vanishes from this plane of existence as well, including written records. but traces remain if you know where to look. you didn't think the eucharist was always just a little cracker and a sip of wine, did you? that wouldn't make sense. it doesn't even look like flesh.

[veiledinviolet]

There are ~150,000 assigned code points in Unicode. At 500 glyphs to a page that'd take only 300 pages. Vividly imagining a telegraph operator decoding UTF-8 in his head, looking up scalar values in his reference book, consulting another much slimmer volume with flow charts to resolve zero-width joiners, picking an emoji rubber stamp out of a drawer, pressing it into the correct Fitzpatrick skin color ink pad, adding the reaction to the right sheet in the ring binder of Slack messages and standing by for the next transmission

[veiledinviolet]

Good afternoon to all, except the most devilish or evil of you. The lesser demons and villains among my followers -- but not including the infernal imps which infrequently appear and wreak havoc in the hamlets and towns -- may still enjoy the day.

[veiledinviolet]

Would you say that the distinction between high-context and low-context languages is real? I was skeptical at first, but japanese seems to leave a lot more information implicit in a typical utterance than english.

No such distinction is used (or at least commonplace enough that I have heard of it) in linguistics. If one were to introduce such a distinction, they would have to put forth some way to measure or operationalize "low/high-contextuality"; in the absence of that there's very little I can say about it scientifically.

What I can say is that laymen's subjective impressions about "what different languages are like" are very often more grounded in bias than in fact. There was a good post that went around here a few months ago to the effect of

People everywhere: "[Language I learned in childhood] is so subtle and emotive, whereas [language I learned in adulthood] is so cold and logical".

Often, these subjective impressions then get filtered through the cultural-theorizing-industry and elaborated more and more, becoming more entrenched as "established wisdom" about X or Y language among non-linguists, and in the process getting farther and farther from any real or verifiable truth.

Well anyway, I think the claim that "Japanese is more contextual than English" is probably one of those. Until someone comes up with a real metric for "contexuality", at least, I will probably continue to judge that to be the most reasonable hypothesis. Here are two ways such an impression could have come about:

For English speakers who learn Japanese as adults, things which are left to context in English but not in Japanese will not appear to be "absent" and they won't notice any gaps, whereas things that are left to context in Japanese but not English will strike them as "absent" and they will be more aware of them.

English speakers who speak some Japanese but are not proficient will not in fact be fully familiar with the rules governing the interpretation of utterances, and so things which are actually grammatically determined may appear to rely on nebulous "context".

Here is a salient difference between English and Japanese: in Japanese, any noun phrase may be dropped "when its meaning can be contextually determined". For example, you might say

(1) kinou inu ga nikki tabe-chat-ta! yesterday dog SUBJ diary eat-COMP-PST "yesterday my dog ate my diary!"

(2) wanpaku da yo naa naughty COP ASS TAG "he's sure naughty"

In (1), we see that where English has possessive pronouns ("my"), Japanese doesn't use them. In (2), the noun phrase referring to the dog is dropped entirely. In fact, in both of these sentences, not dropping these things would be considered unnatural and stilted. Overuse of pronouns and NPs is a common marker of non-fluent Japanese as spoken by Westerners. Saying

(3) kinou watashi no inu ga / yesterday me GEN dog SUBJ / watashi no nikki tabe-chat-ta! me GEN diary eat-COMP-PST "yesterday my dog ate my diary!"

instead of (1) would technically not be ungrammatical, but would be markedly foreign sounding and corrected immediately in any intro Japanese class.

However, this already tells you something: the fact that (3) is unambiguously unfelicitous tells you that there are some underlying rules here, it isn't just "drop when you feel it". These rules are called information structure rules, and every language has them. In fact, Japanese explicitly marks information structure in a number of ways that English does not.

Some of the basic rules in Japanese of relevance here (this is a fairly crude analysis and does not account for various things, but it's probably good enough for our purposes) are:

Every discourse has a topic

If no topic is specified, the speaker is by default assumed to be the topic

A non-topic subject may be introduced into the discourse with ga

A noun already in the discourse may be made into the topic with wa

A salient subject already introduced, but not explicitly topicalized with wa, may be implicitly topicalized

Empty NP positions and unmarked possessors should be taken to refer to the topic

Items that are (semantically speaking) likely to be possessed should be interpreted as possessed before they are interpreted as indefinites

These rules are not inviolable, and in particular (5) requires some contextual definition of "salience" and (6) is certainly not this simple in reality (there are often multiple empty NP positions and the full ruleset for interpreting them seems complex; for instance subject positions are favored for topics over object positions and so on), so there is still some amount of combinatorics with referents and syntactic positions that presumably is going on somewhere in speakers' brains or whatever. But the point is that these rules narrow down pretty starkly what interpretations are "reasonable", and the actual role of context in disambiguating between reasonable interpretations is not so vast.

Anyway, using the above rules, it is not so hard to go through (1) and (2) again, and see that only a single reasonable interpretation actually presents itself.

As mentioned, Japanese very often makes information structure explicit using the particles wa, ga, and wo (not mentioned above, but the object equivalent of ga), which is somewhat uncommon among the languages of the world. English, on the other hand, does not do this. English speakers do not drop noun phrases, but they still replace noun phrases with pronouns very readily, and disambiguating pronoun referents uses pragmatic and information structure rules of exactly the same type! Consider, for instance

(4) My boyfriend went on a "boys trip" with Will and Tod last weekend... I told him not to let them pressure him into skinny dipping again. What was up with that anyway?

Think about what you're doing when you assign referents to these pronouns. It's automatic so you don't notice it, but is it unambiguous? Not at all! You know, for instance, that "him" refers to the boyfriend and "them" to Will and Tod, and you know in the second sentence that "that" refers to peer pressure skinny dipping. Some of this (in particular the referent of "that") I think has to be chalked up to pure context; it's the semantics from which we derive the correct assignment. But some of it is mediated by syntactic or information structure rules as well; for instance consider

(5) Jacob went on a "boys trip" with Will and Tod last weekend... I told him not to let them pressure him into skinny dipping again.

We are still able to produce the correct pronoun assignments in this sentence, even though the semantic context which informs us about which one of these people the speaker is most likely the closest to has been removed. This is, again, a product of information structure rules: Jacob is the topic here, and so (by whatever rules operate in English; not identical but not dissimilar to those in Japanese) we infer that "him" refers to Jacob.

Anyway, the point is that all languages make reference to context very freely in matters of interpretation (which is a big part of why language models had to develop implicit world knowledge before they could speak convincingly), and also languages make reference to context in a structured way which can often be described fairly precisely, and which leaves less open to chance and misinterpretation than might initially be assumed. The gulf between English and Japanese is not so large here. It might be the cases that the [pronouns + unmarked topicalization]-English system is more explicit than the [empty NP positions + marked topicalization]-Japanese system, but I don't know. And of course it might be the case that in some other domain of grammar Japanese is more explicit than English. So one must be careful with any broad assertions.

[veiledinviolet]

portland, oregon, march 2024

[veiledinviolet]

June 11, 2023

Quilting by Marty Kotter

Grange Insurance Audubon Center - Columbus OH

[veiledinviolet]

oh the crisp night air

[veiledinviolet]

Instructions for how to download a Youtube video using VLC on Reddit

Instructions for how to navigate the underworld on an Orphic gold tablet

[veiledinviolet]

I haven’t been able to get the full video but we just celebrated one of our steam locomotives turning 145 by chucking a chocolate cake into her firebox

[veiledinviolet]

these days i often find myself suddenly overcome by a strong lust for spinach.