167 Comments

User's avatar
Timothy M.'s avatar

Some of the block-quote formatting seems off here. E.g. this is inside the quote: "The natural modern way to present this information would be as a list or a table, but an abstracted list of items undistinguished from each other is hard to memorize."

Scott Alexander's avatar

Thanks, I fixed that one, did you see others?

Neike Taika-Tessaro's avatar

The paragraph "A begat B, B begat C, C begat D" seems to have the same issue, the text starting with "This is more abstract than Homer’s list" seems to be meant for outside the quote.

Neike Taika-Tessaro's avatar

Also, the quote Timothy reported seems to have gotten worse instead of better, unfortunately. The image after the quote is now also in the quote, when what I believe was intended was that "The natural modern way to present" and after is outside the quote.

Taymon A. Beal's avatar

1. The code snippet ("main( ) {") is screwed up. To fix this, type three backticks (```) in a blank line in the editor, then press Enter; this creates a code block that you can put the code into.

2. In the line "it’s between the website and God whether those spaces show up \ \ wide or \ \ wide", the two pairs of backslashes each have only one space between them; I assume they're supposed to be different. To fix this, you need to use non-breaking spaces. I don't seem to be allowed to use non-breaking spaces in Substack *comments*, but they work in Substack *Notes* (I checked) so presumably they'll work in posts too. You can copy-paste the character from the first sentence here (it's in the shaded box): https://en.wikipedia.org/wiki/Non-breaking_space

Oleg Roderick's avatar

I read the part about oral speakers connecting sentences with and, and, and. And immediately had a flashback to experience of teaching introductory college mathematics. This is how stereotypical bad students write; they don't believe that an operator is an entity in itself, just a connector to keep the flow of solution going, it's meaning is approximately "and now this happens". Example: 2x-5=3=2x=8=4

I could build a quick argument "naive mathematics is an oral tradition" based on this.

Leppi's avatar

This example is interesting, because it seems to be incorrectly set up, while still solving for x correctly. (if we assume 2x - 5 = 3 -> 2x = 8 -> x = 4). So the student must presumably understand how to solve for x, while not understanding how to properly set up equations? Is this really common?

Oleg Roderick's avatar

It is extremely common. 30% of the class would write like that (in 2003-2009). I compared my experiences against other teaching assistants / early career instructors.

Students would get angry at criticism. I got the right answer, didn't I?

If you give them a harder problem, where "connect everything with =" gets them hopelessly confused. They would get angry at the problem. Just like subjects of study in the post got angry at silly questions about color of bears.

Aeolian321's avatar

I don't think they were getting angry, I think it's "the other place might as well not exist" levels of "I don't trust your syllogism."

All taro is purple. Jill has taro, and she is grinding it. What color is inside her molcajete?

This is the same syllogism (give or take the minor simplification), but it is more concrete.

Awelotta's avatar

I do math sort of like this if I have to use a keyboard. I guess more-so treating it like the calculator ='s button, eg 2 * 3 = 6 * 5 = 30 instead of 2 * 3 * 5 = 30, whereas your example seems different, since it really is being used just as a separator.

Huh. I guess someone learning a foreign language might do something similar, like mapping 而 onto "and" and then that interpretation being fossilized.

Hastings's avatar

The other unpronouncable literacy than code is math. My personal blog and research notes can only really take place on platforms with LaTeX support. It almost feels like the author of this review is grasping for math notation in their quest to invent a new, non-oral writing.

i.e., I use Franklin.jl as my blog generator, and to use the example from the post, could easily blog $$ \overline{\text{best}} ^ {80\%} $$ .

TGGP's avatar

Computer science is just applied math. Lisp began as a mathematical formalism that was then coded into a program later.

Taymon A. Beal's avatar

Substack does actually support LaTeX in posts (but not comments).

Jeffrey Soreff's avatar

>The other unpronouncable literacy than code is math.

Very much agreed! I was rather relieved, when reading the post, to _finally_ see

>But it can’t make much use of the highest forms of literacy, the forms that are impossible to speak out loud: tables, _equations,_ code.

[emphasis added]

I'd emphasize the importance of being able to choose the _order_ in which one pays attention to the components of an equation or expression, unlike the fixed serial order in speech.

If I see a definite integral, I'll look at the most complicated part of it first - _usually_ the integrand, but it might be the upper or lower limits or the differential if those are more complex.

One other class of literary (in a broad sense) texts that bears mentioning is circuit schematics. Again, the ability to shift focus and revisit information is crucial.

Vitor's avatar

This review confused me a lot. It's not clear at all which parts are the reviewer's own arguments, what's the content of the book, and most importantly, on how strong of a footing each of the claims is.

Right at the start we have a claim that for oral cultures, words are alive. Because the world isn't full of labels or something. I sort of catch a vibe here, but there's also a high chance it's nonsense.

The review goes on to talk about code as a form of writing. There's something there, but it doesn't translate cleanly at all. In particular, there's repeated confusion about code as an information encoding technology similar to writing, and the result of running such code on a computer, which are very different things.

The entire thing is full of this type of confusion. At the end, the "reviewer" then goes on a rant about how writing can be improved visually, which... I guess? But I can't bring myself to care.

AV's avatar

Agreed. I would even go one step farther and say that the arguments about making text more interesting seem counter to the points about literacy that they described in the beginning. Expressing ideas like "yelling" through typography is exactly the kind of oral-first writing that they decry in the previous section of the review.

Geran Kostecki's avatar

Yeah, was thinking the same - feels like the final form of those suggestions is just... memes

Hedonic Escalator's avatar

Agreed that the ending was confusing and contradictory. The reason why "text with extra visual elements" has remained relegated to gimmick fiction is because it is inconvenient and not very useful.

TTAR's avatar

If the quality of this review is reflective of the quality of Ong's ideas, it saved me a lot of time wasted on reading Ong's ideas.

Deiseach's avatar

I'm not going to blame Ong since I don't think the book was more than an excuse for the reviewer to exercise their hobby horses.

WSCFriedman's avatar

That is very fair.

TGGP's avatar

> Substack, in particular, pisses me off because it forces authors to write in the most strict plaintext I’ve ever seen. IT DOESN’T EVEN ALLOW UNDERLINES!

My usual complaint is that it doesn't allow "a href" tags, so instead you just have to dump a URL, but that's a bit of an annoyance as well. You could write <u>text</u> to get across the intent to underline (whereas using that format for links would likely mangle the url people click on and misdirect them).

> As long as you have a strong vision, an LLM can guide you through the process of publishing a typographically complex essay on your own website.

I used to have a Tripod site, but I primarily used it to host text documents, and now that Tripod has expired I can't even do that.

> When Darwin first wrote down his theory of evolution, he didn’t do it in words. He did it like this:

That seems like an example of your solitary thinking quadrant, because when he actually published Origin of the Species he used words (and it was printed, rather than being in handwriting only he could read).

Steve Sailer's avatar

A lot of software is improving fast these days, such as Tesla's Full Self Driving (Supervised) code.

But then there's Substack ...

Taymon A. Beal's avatar

These show very strong signs of being deliberate product decisions; it's not a technical problem.

Steve Sailer's avatar

Do you think the single most obvious shortcoming in Substack -- that you can't paste a table of numbers (so you have to use a screenshot as a work-around) -- is a high level decision made by the Substack marketing department after researching the tastes of Substack subscribers: "Our customer hate data tables!"?

Could be ...

Taymon A. Beal's avatar

I was referring primarily to the decisions to (1) not allow arbitrary HTML, and (2) not allow any formatting at all in comments. Lack of table support is admittedly more baffling, especially since they support a lot of things way more niche than tables.

Steve Sailer's avatar

For my Substack, I charge $100 annually or $10 per month. I'd like to be able to offer a four-month $36 subscription (largely because Stripe charges so much for processing payments monthly), but ... No Can Do.

I wonder if this is just because Substack appears to be generally lazy about writing additional code? Or did Stripe negotiate with Substack a ban on subscription lengths longer than one month and shorter than twelve months to maintain their lucrative haul on monthly billing?

Note: on the whole, I very much like both Substack and Stripe. But I wish they were getting better faster.

Viliam's avatar
7hEdited

Perhaps we could collect our wishes, and create a document for anyone who would like to enter the market and compete with Substack.

For example, I sometimes write about programming, so the text includes code samples. When writing code, 80 characters per line feels like a claustrophobic minimum, but in a Substack post you get about 50? I am not sure about the exact number, but it's often not enough to write things like "private MySpecialGadgetFactoryBuilder mySpecialGadgetFactoryBuilder = new MySpecialGadgetFactoryBuilder();". So I would need a paragraph style that is "code" not only in the sense of "monospace font", but also allows me to write 80 characters in a line.

Also, I can currently choose to mark some words as monospace, and I can choose to make some words be a hyperlink, but for some reason I can't have both. (And this was even possible in the previous version of Substack a few years ago, my older blog posts used it, but then they removed the functionality.) I would like to use monospace for the names of Java classes, and hyperlinks to point to the respective file in the GitLab repository, but I can't.

Would it be so hard to allow the author to center a line? This is just an esthetic preference, but I love separating parts of text with a centered "*" or "***" symbol.

The captions of images are difficult to work with, for some reason I cannot figure out. I do not remember the exact details, but copying the caption from one image to another is almost impossible. If I make a screenshot, put it in the post, and write a caption, and later I decide to replace it with a better screenshot, it is usually easiest to just type the caption again.

...and these are just things that I would classify as bugfixes or trivial enhancements. Then there is a separate category of nontrivial new features I would like to see, for example:

Authors or blog owners being able to appoint other users as moderators of their blog. A moderator can delete comments or block users, but their decisions can be reviewed and reverted by the blog owner. (ACX would need this like a salt; the comment sections are insanely large, and I think most of us would prefer Scott to prioritize writing new content over doing maintenance work.)

Support of multiple languages. Currently, the only thing you can do is start another blog in another language, using a different URL, and maybe add some manual cross-links. Some of us are multi-lingual, could we get some native support for that, please? For example, a page that exists in multiple languages could have the same url, perhaps ending with something like "?lang=en", and buttons on the top to switch between the available languages. (And maybe another button to create an AI-translated version.) All language versions could have a shared comment section, with comments automatically translated to the reader's language, giving the reader an option to specify which languages (yes, plural) should be viewed in original.

moonshadow's avatar

Arbitrary HTML from pseudonymous internet randoms with no skin in the game is never allowed, anywhere, because it's a great gaping security hole ready to abuse.

A small subset of HTML restricted to specific tags with specific attributes is still allowed some places that have either taken heed of the "parse, don't verify" dictum in their engine or have not yet been punished hard enough for failing to keep the tag set small and failing to use sufficiently robust mechanisms for parsing it. However, at that point, most actual users prefer a less scary approach to formatting, like markdown or a WYSIWYG text entry box with styling buttons, which is also much less risky; so these days no-one bothers.

Viliam's avatar
7hEdited

I suspect the high-level decisions are more like this: "Our most profitable customers just want a system that allows them to spam their subscribers with political opinions, conspiracy theories, AI-generated essays on mindfulness, etc. Some features -- such as bold text, italic text, pictures, and embedded videos -- are nice to have, but our typical customer prefers to keep things simple. And then there are a few nerds who want things such as tables, but they generate like 0.1% of the profit, and most of them will end up using a screenshot of a table, and if you gave them tables they would probably start demanding something else anyway, such as interactive animations or scripting, but we are not in the business of making a website for nerds."

Plus I expect the typical corporation setup, where the IT developers are nerds, but their managers are not, and the managers just cannot imagine why data tables could be so important to someone, so they will assign it a priority much lower than e.g. changing the position of a "subscribe" button.

DrMcleod's avatar

The appalling performance of the entire site on mobile browsers surely can't be deliberate.

Taymon A. Beal's avatar

If by "a href" you mean hyperlinks whose text is something other than the URL being linked to, Substack *authors* can do that just fine; there are a bunch in the above post. Substack *commenters* are restricted to plain text; I suspect this is because they're not trusted not to use the power of formatting obnoxiously.

(Neither authors nor commenters can use underlining; the reasons for this are rooted in an old philosophical debate about the semantic meaning of annotated text in the context of HTML.)

FYI there are lots of Tripod-like things still around; Neocities is pretty user-friendly.

TGGP's avatar

Yes, I'm referring to commenting. I'm able to use links in other kinds of comment sections, but not Substack.

Thanks for the tip regarding Neocities.

Taymon A. Beal's avatar

Huh, which comment sections let you do this? In my experience it's a fairly small minority. The only ones I can think of off the top of my head are LessWrong and Scott Aaronson's blog.

TGGP's avatar

All Wordpress blogs, such as my own or Greg Cochran's. Marginal Revolution is another notable one, as is Andrew Gelman's. Disqus no longer lets me copy/paste markup to get rendered that way (annoying), but I can use the toolbar to add links (though not if I stay on the Discussion page after clicking one of their email links, which can in turn link to the specific page but I don't always bother if it's the AV Club where the comment section won't load and bring you there by default).

Mark Roulo's avatar

Note that the philosophical debate here spills over to web page Accessibility desire/requirements and for commercial sites can get the Americans with Disabilities Act involved ...

Steve Sailer's avatar

The logic games portion of the LSAT got cancelled on the (at least ostensible) grounds that it's unfair to blind test takers. Does the ADA explain why Substack doesn't allow you to paste a data table?

Mark Roulo's avatar

Background: I spent a few months investigating the taxonomy of web site data and control flow organization as part of trying to understand for my own satisfaction why the web seems to be getting blander. But I don't work for Substack and I am not a professional web developer.

So ... Substack doesn't support tables because Substack thinks of itself as a providing subscription based mailing list support. Basically, the authors do the equivalent of writing an e-mail newsletter to their mailing list and Substack will handle the rest: blasting out the e-mails, posting a copy to the Substack web site, managing billing, etc.

Tables render poorly/inconsistently in e-mail.

Since Substack thinks it is helping send out e-mails, it intentionally provides no table support. Yahoo mail doesn't support tables on e-mail composition, either, so there is some reason to believe that this answer is correct-ish.

Taymon A. Beal's avatar

Substack provides lots of features that are a lot less supported in email than tables. And tables in HTML email do mostly work.

Performative Bafflement's avatar

Your best bet for tables in a Substack post is using datawrapper, although you have to create a free account there to do it. They render fine.

If it's small enough, you can just screenshot out of excel or a book too, I've done that plenty of times.

AnthonyCV's avatar

As it turns out, when you give most people the power to do the kinds of formatting you're describing, most of it isn't Tufte. It's Geocities. Which is fine, but I can see how the platform evolution happened.

Catherine Proctor's avatar

'For example, Hesiod, when writing about justice, could not write about it abstractly; '

Why is this being held up as an example of Orality if Hesiod is using the written word?

Candide III's avatar

Scholars apparently differ on whether Hesiod himself wrote down his poems or if they existed in oral form and got written down later. Anyway, Ancient Greek culture did not switch from oral to written overnight. It took centuries. In Thucydides we still see a largely oral culture even at the higher social strata. Private libraries only started to appear in V BC.

gordianus's avatar

The archaic Greek poets (the lyric poets,* the authors of the now-lost "Epic Cycle", perhaps Hesiod & the author(s) of the Homeric Hymns) composed poetry with a lot of the techniques used by oral poets, & following oral tradition in their form & subject matter, even though they were literate & therefore had more time to consider their words.

*We have almost no fully-oral lyric poetry or (AFAIK) direct references to it in ancient Greece, but comparative evidence suggests that it was a significant part of the Indo-European poetic tradition & was in some ways more conservative than Homeric epic.

Catherine Proctor's avatar

Oral: What things I adjudicated not from messengers other than you or myself to hear. Written: I did not think it fit that I should hear of this from messengers. Changes: The written version adds abstract concepts of ‘it’ and ‘this’. The verb becomes ‘thinking’ not ‘hearing’.

In fact both quotes contain 'hear', the analogous verb to 'think' in the second quote is 'adjudicate' in the first. And are 'it' and 'this' really abstract concepts? I don't think these analyses are very coherent.

吉田's avatar
2dEdited

This one rang a bell. Sam Kriss’ “Reading Is Magic”?

Taymon A. Beal's avatar

That piece explicitly cites Ong and is based on his ideas.

吉田's avatar

Don’t forget Luria

Steve Sailer's avatar

I've always sympathized with the peasants interviewed by Professor Luria in the 1930s: I suspect some of them reasoned: "Here's some apparatchik sent thousands of kilometers by Stalin to interrogate us with some damn fool questions about bears. I don't know what the Party Line in Moscow is at the moment about the color of bears, so I'll play dumb. They won't shoot you for being dumb, right?"

Steve Sailer's avatar

It would be interesting to study literacy vs. orality in evidence used to condemn Communist Party members in the Moscow show trials of 1937.

If they had you in writing expressing something good about Trotsky or any of the umpteen other forbidden things, I presume they could just show you wrote it and then shoot you.

But if they didn't have written proof of unorthodoxy, well, they had to beat you into confessing you had said it, which took time and trouble. Of course, they enjoyed beating people before shooting them, so maybe I'm exaggerating the distinction. But still, I figure, the likelihood of getting executed by the secret police in 1937 correlated positively with how many political opinions you had put into writing.

Similarly, the permanence of social media contributed to Cancel Culture during the Great Awokening. If director James Gunn had just made a few pedophilia jokes to friends orally, or if he'd written them in private letters to friends, it all would have blown over. But because he put it in universally available digital writing, he was in big trouble for awhile.

Taymon A. Beal's avatar

I've heard arguments along those lines before. I'm willing to credit them so long as other people are willing to grant that there's an important difference between not being *interested* in using the skill being tested here and not being *able* to. Unfortunately concerningly many people want to use this as a stalking horse for epistemic relativism.

Steve Sailer's avatar

Right. I don't know what the universal implications are, just that it's pretty funny that everybody naively assumes that Dr. Luria's findings in Stalin's Soviet Union in the 1930s are universally applicable. After all, why would anybody play dumb in Stalin's paranoid dystopia?

Similarly, I can sympathize with Dr. Luria, just when things are getting tense in Moscow, vanishing to Central Asia to ask peasants about the color of bears, and then publishing his findings decades after Stalin's death: talk about maintaining a low profile!

Aeolian321's avatar

The use of demonic possession as an explanatory method for explaining wrongdoing is a far better argument for epistemic relativism.

Or we could discuss the Hezbollah pager incident, that's another one good for showing "those who cannot think two steps into the future."

Deiseach's avatar

Yeah, I've taken that view myself. "Hang on, what if white polar bears is some kind of coded question about the Whites? Yeah, that was a while back, but if this guy is trying to sniff out regressive attitudes and covert sympathies for the Tsarists in order to get us all shot, I'm going to pretend I have no idea what even a bear is".

Steve Sailer's avatar

Right. Using the W-word "White" in 1930s Soviet Union could mean you are a Wrecker.

Deiseach's avatar

I think that in Stalin's Russia, it wasn't a question of "Am I paranoid?" but rather "Am I paranoid enough?"

John Stathatos's avatar

"They won't shoot you for being dumb, right?" sounds somewhat reasonable, but don't count on it to save your skin. Many hundreds of thousands died in Stalin's purges because they once said a kind word to somebody Stalin disliked, because grandad was the wrong kind of peasant, because they complained about the fact that local shoe store's latest shipment consisted of left boots only, or that old favourite, "looking disgruntled in the face of authority".

Donald's avatar

If you try to do abstract logical arguments, but you aren't trained in knowing which moves are valid logic, you will soon end up with all sorts of nonsense.

These people probably don't need to use abstract logic a lot in their daily lives.

So a general policy of not engaging with elaborate logical arguments will serve these people well.

ilya187's avatar

> I don't know what the Party Line in Moscow is at the moment about the color of bears, so I'll play dumb.

If that were the case, then literate peasants would have played dumb just as much as illiterate ones, possibly more. But according to Luria, they did not.

Jeffrey Soreff's avatar

Agreed! I'd remembered reading a comment along these lines before. I'm guessing that you made the same good point in a previous comment section at some point?

Steve Sailer's avatar

There's a statue or carving of Hammurabi in the U.S. Supreme Court building celebrating him for being the earliest known implementer of written law. The implication is that written law is more just and fair than oral law.

But, I suspect that Hammurabi found written law to be helpful in his being more authoritarian. He didn't have to get into arguments with anybody about what exactly the law is and whose memory is right. Like the busdriver in the Simpsons meme, Hammurabi could just tap the sign.

I recall visiting the Department of Motor Vehicles once and asking a DMV lady if I could do X. She didn't say anything, she just tapped a written sign pinned up in her cube that said: "No, you canNOT do X."

And that was that.

Candide III's avatar

On the other hand, as a matter of revealed preference (Twelve Tables) Roman plebeians found written law to be helpful in getting out from under patricians' thumb, who had been treating law as personal religious property. So it cuts both ways sometimes!

Anomony's avatar

Written law is a key advancement in the concept of equality under the law, as that requires people to agree on what the law is

WSCFriedman's avatar

> We could instead have ended up in a place where every text looked a little like House of Leaves. Maybe we still can.

That sounds horrific and dreadful. Last thing I'd want.

Kevin Barry's avatar

This is why emoticons are low key vital to modern digital casual communications

Hilarius Bookbinder's avatar

“No one feels a deep internal drive to read long texts. It takes a lot of effort for us to start reading.“ That’s just false. I’ve read many long books that I was eager to read just because I knew I would really enjoy them, and longer meant more enjoyment.

WSCFriedman's avatar

Oh yes, I gladly dive into A Song of Ice and Fire or Lord of the Rings every time it's been long enough that I can reread them. The longer the story the more room the author has to work in.

Aeolian321's avatar

You mean the more faaat an author has to roll in. Short stories are often the best for rating authors because they're the hardest to write.

ilzolende's avatar

Most people reading books want to pick the books which produce the best reading experience for them, not the best means of grading authors on technical skill.

Aeolian321's avatar

I'd rate authors based on ideas, the heart and soul of science fiction, not based on technical skill. And that's better done in short stories. Niven's are crystallic.

WSCFriedman's avatar

I mean the longer I'll be delighting in the experience of a secondary world, which is what I read books for. I don't care about how hard or easy it was to write, I care about the spread of the epic and how good a job the author did making it feel real.

Aeolian321's avatar

Suggesting Kulthea, if you're looking for a world to play in. : - )

QuintusQuark's avatar

Even if you love reading as an adult, learning to read takes a lot of effort. Babies are born with specialized neurological systems for learning language and a drive to start doing so early in childhood, but reading typically needs to be explicitly taught. Groups of deaf children exposed to a collection of signs have rapidly invented fully grammatical languages, while new written languages without outside influence can take hundreds of years to develop.

FeepingCreature's avatar

The problem with the text visualizations you gave is that you can't take them with you. You can read them, but you can't copypaste them, you can't quote them, meaning you can't talk about them.

Maybe we need Unicode Vector. A standardized, universal format for embedding graphical context in text, optimally combined with placement hints for adjacent words.

Rana Dexsin's avatar

The technical problem is mainly an interop/fragmentation thing, I think—which you kind of said, but I don't think “new format” would be the best result, because we already have SVG that can be embedded in HTML, and it seems likely that you'd base it on that. The earlier Windows and Macintosh universes had OLE and whatever-it-was-on-Mac-that-I-don't-remember-the-name-of that would allow relatively seamless rich clipboard operations compared to what we have now, but it's harder to establish this on a diverse Internet.

On the flip side, in a way that I find an interesting juxtaposition between visuality and a sort of maybe-neo-orality (I don't think it quite matches up but it seems to be part of the same cultural flow?), people in many corners of the Internet do regularly quote graphs. Because they take screenshots. Of everything. Including normal text. Because that always preserves the thing they see with all of its layout and formatting, and nothing else always preserves it. (And then, I presume, the visually impaired users who might've wanted to be part of the conversation either pull out increasingly advanced OCR tools or cry.)

ilzolende's avatar

You also can't necessarily render them well in an RSS reader. And if you colorize text, it's going to get discarded or warped when the reader wants dark mode and you haven't implemented it. Firefox Reader View will just discard it. Some sites just invert things: I was rereading the Sequences a while back and noticed text about "the redness of red" being rendered in cyan, which was tremendously amusing. Some sites try to only invert luminance and not hue, but this still has issues.

avalancheGenesis's avatar

Numerous distracting formatting errors were already flagged and (being attempted to be) fixed - a mark against written culture. Do wish this review didn't repeat the previous-years habit of not clearly stating the title of the reviewee + the full name of the author early on. I can autocomplete to [Walter] Ong because it's a fairly distinctive name and I have vague memories of him being associated with Marshall McLuhan or Crunk & White, but it's annoying to have to tab out and double-check this for myself. Inclusion of, let's call it various mixed media was creative, I did not have "Gwern reference" on my ACX Book Review Bingo card. Overall I think it made for a decent *essay* on the topic of orality and literacy, but the nominal book being reviewed was a third wheel at best. Just one citation among many. This I unfortunately cannot condone in the ACX Book Review Contest, even if I'd have given a "like" to this post as a normal post otherwise. (And, no, I do not want every text to look like HoL, I'd rather give up reading...)

ETA: on further consideration I really don't like the conclusion of this one, where the premises simply don't seem supported by the evidence thus far offered. It's the equivalent of suggesting solitary-persistent written culture do a "Pivot To Video", try to compete with the infinite scrolls of the world by apeing that inane mode of passive nonthought, just done with textual gimmicks instead. And, hey, I am a fan of format screws, there are some very cool storytelling possibilities you can unlock even using entirely plaintext! A fun and simple example of "convey through words two different simultaneous conversations happening in realtime", knowledge of plot not needed to appreciate the twist: https://scp-wiki.wikidot.com/counterpoint . Or Zero HP Lovecraft's "Don't Make Me Think", which he sadly deleted, but that was a masterpiece where at least half the meaning in the story was conveyed through hundreds (thousands?) of emoji - but they still need the scaffolding of written words to make sense. I can appreciate that despite agreeing with Ian Bogost that emoji are a degenerate form of smiley, a preliterate corruption turning the clock back on reasoning through abstraction instead of literal representations.

But, no, don't try to out-bullshit a bullshitter. We don't need to destroy literacy to save it. That's like choosing to incept the strong form of Fat Acceptance instead of developing GLP-1 agonists. And I think it's also undersold just how radical a change "put down the phone and touch grass" is, that already cuts The Enemy pretty effectively. After that, all you need is scale...I'm encouraged at the increasing Overton acceptance and prevalence of phone-free schooling, for example. Today's kids might not be alright, but tomorrow's may be.

Taymon A. Beal's avatar

The formatting errors are almost certainly a consequence of Scott's convoluted system for letting people submit reviews and then importing them into Substack (because this isn't a use case Substack is really designed for). It'll likely be better next year, since the shiny new web app for letting people rate reviews can easily be extended to handle submissions.

avalancheGenesis's avatar

Oh, are we getting a New And Improved ACX Contest experience for next year? That's nice, I don't necessarily want to read something with a ton of bells and whistles (this is a blog by and for plaintext lovers), but it's a fair point made about e.g. tables. And having to constantly toggle my contrast controls on and off because some of the formatting (png transparencies?) won't load otherwise, that's a Trivial Inconvenience that does semi-consciously get me annoyed at a review for no real reason.

Taymon A. Beal's avatar

The new app was released this year, at the beginning of preliminary voting. My hope is that having it also handle submissions will make formatting errors less likely (because it can be specifically tested against Substack and then authors will have the chance to check that their reviews look right). You can also use it to read the finalist reviews if you like, since they remain available there after voting closes. That said, Scott is presumably still going to want to post the finalists on Substack, so tables still aren't going to work if Substack doesn't add support for them.

Bean Sprugget (bean)'s avatar

Wouldn't we expect Chinese speakers to be better at abstract thinking if this were the case? That effect seems negligible, as opposed to the benefits of being able to store information. It's also interesting to note that the rise of pinyin, especially for computer input, is a way that Chinese is becoming more "oral".

avalancheGenesis's avatar

I once had a too-smart-to-know-she's-ignorant white college student try to explain to me that people from my own (Chinese) culture think about time different, because we don't have peoper words for it, she read it from Noam Chomsky so of course it's factually true. Maybe there was an abstract truth there I simply failed to grasp, but it seemed just concretely false and I found the whole interaction baffling.

(But then I'm a 3rd-gen twinky who didn't really learn the language. But then, none of the generations before me in my family who did, they never showed a weird understanding of time either. On other topics like gender, yeah, but I think that has a lot more to do with general conservatism rather than features of the Chinese language...)

Kurt's avatar

I can tell you one thing: native Chinese speakers learn arithmetic more easily. Why? Because the way numbers are written in Chinese much more closely matches the base 10 Arabic numeral system we all use for modern arithmetic. Compare with a language like French where 99 is pronounced “quatre-vingt-dix-neuf” or (translated to English) “four twentys, ten, (and) nine.”

In Chinese, 99 is “jiu-shi-jiu” or “nine tens (and) nine.” This pattern is consistent, so that you don’t have oddities like the English “eleven” and “twelve”, you just have “shi-yi” (ten-one) and “shi-er” (ten-two).

avalancheGenesis's avatar

It took me a long time to figure out that English speakers often prefer constructions like "fifteen hundred" for 1500, instead of "one thousand five hundred", and now you've got me thinking it's maybe due to this sort of parallelism. I do have an appreciation for the Chinese manju being written with X number of brushstrokes, at least up to 7 iirc, it makes playing that suit in mahjong much easier. By contrast our Arabic 0-9 numerals seem entirely arbitrary!

I did take French in highschool and found the numbers of the counting to be strange and semi-arbitrary. A lot like doing Roman numerals, actually. Probably part of why they don't usually cover anything more complicated than 1-1000, and maybe decimals to the hundredth for prices.

ilya187's avatar

Same with Russian. Russian words for 11, 12 and 13 are, odinnadtzat' (one-on-ten), dvenadtzat' (two-on-ten) and trinadtzat' (three-on-ten), and the pattern continues. Similarly 20, 30, 50, etc. are two-tens, three-ten, five-tens, etc. Although Russian word for 40 is an extremely weird exception, not at all related to any other Russian numeric words.

Hedonic Escalator's avatar

Has a causal effect of native language on math ability been demonstrated? I've seen this claim made often, but never seen evidence of it. Studies I can easily find are small and not convincing: https://pmc.ncbi.nlm.nih.gov/articles/PMC4341514/

Irrelevant personal anecdote: I was raised bilingual and still mentally consult the times table in Chinese.

Viliam's avatar

There is a weird fetishization of foreign language speakers in American culture. I speak a few languages fluently all my life (which is actually normal in the part of the world where I live), and I am still waiting for the promised superpowers to manifest.

Yeah, Slovaks have many words for cheese. So what?

SimulatedKnave's avatar

Having looked it up, the argument seems to be that because Chinese lacks grammatical verb tenses this represents a fundamentally different approach to the concept of time than in Western thought.

Unfortunately the papers on it are written by modern philosophers so parsing actual meaning out is iffy.

Brendan Richardson's avatar

Whoever claimed you can't listen to code has clearly never heard

"HelloWorld.java (fix/metal-true)" by NanowaR of Steel!

moonshadow's avatar

Halfway between that and non-code, we have the obligatory Mili, complete with illustrative video: https://www.youtube.com/watch?v=ESx_hy1n7HA

Deiseach's avatar

This review is terrible. I'm not even sure it's about a book, since we get a fleeting mention of one and then the review charges off in all directions.

And it's not even *entertaining* digressions.

Tristan's avatar
2dEdited

Fun read. It is perhaps relevant here that no written language exists that does not correspond with either spoken language or sign language (according, I think, to the book Empires of the Word, which takes pains to dispel the idea that Chinese or Egyptian hieroglyphics expressed ideas rather than sounds). Code is perhaps an exception, but it is not used for expressing thoughts and ideas between humans. It is, in the sense of interpersonal communication, a written language by analogy only. (Let me know if there are exceptions: I know some jokes have been written in code).

The takeaway from this is that we should be sceptical about the potential for a visual language that is disconnected from the pre-wired centres of communication in our brain, which evolved specifically for tools that enable near instantaneous outwards communication, ie the mouth or hands.

I do think it could be possible to develop visual communication eventually, given the plasticity of our brains, and our ability to attach new peripherals to existing systems (see the book Livewired), but we would need an interface as quick as our speech and hands to make this work, I suspect. This may only become possible once we have computer interfaces, like neurolink, in our brains.

Worley's avatar

> t is perhaps relevant here that no written language exists that does not correspond with either spoken language or sign language

At least at some abstract level that's true, but there are instances of communities of writing where nobody speaks the written language much. The most extreme example was Heian Japan, where the men wrote literature in Chinese but nobody spoke it. (The women wrote literature in Japanese, so it was much better and has survived!) Similarly, Sumerian survived as an administrative language long after it was replaced by Akkadian, and Akkadian survived as an administrative language long after it was replaced by Aramaic.

Tristan's avatar

Ditto Latin in the 14th century France and Spain. But as with Latin it appears men in Heian Japan used Chinese as a complex way of expressing Japanese words. (I had to look this up: I didn’t know this about Japan). They would know the Japanese word certain symbols were used to reference, or would use the name of certain characters to spell out Japanese words using the rebus principle. They would also use numbered subscripts to help reorder words to Japanese grammar. They were not reading text and understanding it in some purely abstract way as ideas - they were sounding out words in their heads, as MRI and EEG studies show we all do when we read.

Aeolian321's avatar

Keith (C++'s mascot) disagrees with you. Languages do exist even if they aren't written, and some of them do not correspond to either spoken or sign language (ask an artist how they encode the world around them? how they dream?)

Tristan's avatar

Our eyes are capable of encoding many times more detailed information than language is capable of expressing. Paintings are not language in the literal sense, however, and dreams certainly are not. I know a word like language can be used to mean many things, but I and i believe the author are using it to refer to the thing that a book is doing when it is conveying declarative information and ideas to a reader. Yes, pictures can in some ways convey much more information, but in other important ways much less: compare the complexity of ideas that artists could convey before written language was invented compared to what we can convey in books. It’s really a very different kind of tool.

It’s interesting to reflect on why written language is so much more powerful in conveying ideas than pictures are. I know of one children’s book that is only pictures with no words, but it is only about 20 pages long. No one has written the equivalent of Harry Potter as just a series of pictures, and when I’m arguing is that the history of written language suggests, we should be sceptical that adding visuals to words will greatly increase the kind and quality of information that can be conveyed. A diagram or illustration here or there can do wonders, but we already do that.

Aeolian321's avatar

Language is symbols. I've seen paintings that are ... worldincoherent. An outside wall put in an indoor place. This indicates that the painter's internal model ...makes that possible. It seem odd to someone with a fuller visual model.

One can express language as pictures. Chinese and Japanese and Korean do this, albeit their current pictures are highly stylized and complex.

People have directed the equivalent of Harry Potter as just a series of pictures, although the screams do actually add something (watch "The Tribe" and see...).

Comics show that the "inbetweens" of pictures can be interpolated, we don't actually need to see them.

Tristan's avatar
8hEdited

It’s important to be precise about the differences between visual communication in general and written language specifically. Yes we can express internal world models with paintings, but there are many things language can do that paintings cannot. Just try painting “I think therefore I am,” and see if someone will get the point. The author I cited above (Empires of the Word) points out that no language has ever been deciphered without first knowing what the language sounds like. Chinese, Japanese, and certainly Korean (which is an alphabet) do not express abstract ideas. (edit: sorry, you obviously mean korean before the language reform). They express sounds, and often express their meaning via the rebus principle (like “I can see you” with can spelled as a can), which relies on knowing the sound. Chinese is interesting case because different dialects can read the same words while producing different mental sounds. I don’t know how the author would handle this criticism. I suspect it has something to do with how written languages can tolerate quite a bit of diversion from the original sounds (as in English), but you have to make up the gap with considerable education.

Kurt's avatar
1dEdited

> Code is perhaps an exception, but it is not used for expressing thoughts and ideas between humans

The latter is false. Code is used to express ideas all the time. This phenomenon is everywhere in STEM higher education, once students have been expected to have completed their first introduction to coding class. Coding is just an offshoot of mathematics, the older form of “written language without a spoken analogue” used to convey ideas.

Tristan's avatar

Hm, it would be interesting to check what I happening in people’s audio cortex when they read code. Studies have shown that for written language, people do sound the words out in their head when read (based both on MRI/ EEG evidence, but also people take longer to read words that take longer to say). I wonder if people are sounding out code as they read it? Or if something totally different going on?

Ok I looked it up and the answer is it’s actually very different: Broka’s area hardly lights up. I’ll have to process this and adjust my mental model. I’m not sure if this changes the argument about written languages used as written languages - perhaps it’s a different category of thought and expression - but perhaps those categories could be combined in a useful way? I’ll need to reflect

Aeolian321's avatar

Can you cite some japanese or chinese sources on this? Or a source where the "takes longer to say" still has the same number of letters?

Tristan's avatar

For the "takes longer to say," check: https://link.springer.com/article/10.3758/BF03205520

I was relying on a book on the wars over how writing is taught, but the phenomenon appears to be well documented. (Non-scientific language teachers liked to encourage readers not to sound out in their head because they thought it was counterproductive, when actually the opposite turns out to be true).

For the first question, I'm assuming you're referring to the other comment about Japanese writers writing with Chinese characters. I was relying on AI because as I said I'm not familiar with that topic. There appear to be a variety of good sources online that detail how the system worked.

Aeolian321's avatar

It was a reference to "ease of understanding" based on character complexity, versus "ease of pronunciation", but the source you're citing is well good enough, thanks bunches!

Geran Kostecki's avatar

"No one feels a deep internal drive to read long texts."

30,000 words about Fatima say you picked precisely the wrong audience to make that argument lol

Worley's avatar

About halfway through I realized that the language of this column felt considerably more oral than most columns. I did manage to measure that the average sentence had 16 words, whereas the last book review had 21 words per sentence and noticeably more dependent clauses.

Judith Stove's avatar

Just about 'honesty is the best policy': I only realised in recent years that part of the point of this adage is that, for several centuries, one of the meanings of 'policy' was: 'A device, a contrivance, an expedient; a strategem, a trick; obsolete,' c. 1406-1849' (OED, q.v. 'policy,' sense 1.3). The adage, then, was conveying that honesty was a productive form of 'trick.' It thus has a lot of irony built-in, which we tend to lose these days, when it just seems like a trite and obvious saying.

In the age of the podcast, we have returned, to some extent, to the times in which people would flock to hear a known preacher or other public speaker; or, if you prefer, to the decades in which people crowded around the radio. Our faculty of listening has redeveloped, I would say. After listening, just as the 18th-century sermon fans might have sought out the printed version, we can follow up by reading the (often vilely botched) podcast transcript: maybe the oral and the literal are still twined together.

Kenny Easwaran's avatar

I actually thought “honesty is the best policy” is clearly a product of *orality*, because it’s got that rhythm and slant rhyme to make it memorable, even though it doesn’t exactly include anything concrete.

Awelotta's avatar

I'm skeptical of the claim that oral cultures prefer "and" more than literate cultures. Isn't it moreso a function of genre (which is maybe what's being claimed, but it's not totally clear), ie that epics and oral poetry have a specific style? And also that translators would probably borrow stylistic practices between different languages (speculation).

Checking Wikipedia, illiteracy rate was 80% in PRC in 1950s, but I don't particularly have a sense that Chinese especially uses "and" compared to English...

Actually there are sometimes where I would expect a conjunction in English, like "要吃就吃" "[If you] want to eat, just eat". So maybe there is some merit to this idea; though Chinese grammar is also a lot more isolating and just doesn't mark a lot of stuff redundantly the way Indo-European languages do (eg gender and inflection, which are pretty redundant).

Candide III's avatar

Gender and inflection are somewhat redundant only in complete sentences. First, they permit a much freer word order that can be used for emphasis. English and Literary Chinese word orders are much stricter. Also, in re oral culture, a lot of people don't speak in complete sentences. They speak in fragments and/or drop/replace words which are obvious from the context. In languages that retain more Indo-European structure, such as Slavic languages, conversations can go surprisingly far with what is basically indexicals, often expressed with derivatives of the three or four PIE expletive roots, declined and conjugated and pre-post-positioned as appropriate.

Awelotta's avatar

> Havelock, a mentor to Ong who wrote about similar topics, in his book The Muse Learns to Write, provides examples of how sentences in the play Oedipus changed from its oral version to its written one

I'm confused what this means. The play Oedipus Rex (the original presumably?) has distinct oral and written versions? Is this talking about adaptations into written versions? Or is the original manuscript is the written one, in which case, how'd he get the oral version?

Kenny Easwaran's avatar

That did confuse me too.

Awelotta's avatar

Communal-persistent should be "productive", strictly speaking, if you disregard the other factors/arguments for why it is bad

Aeolian321's avatar

From this article, I have concluded that oral culture was quite a bit more literate than written culture. Literacy is the art of common allusion, after all, the common culture and substrate that one can depend on the next schmuck knowing.

... yes, I am continuing an argument from an Open Thread. OP brought up literacy first.

Julia D.'s avatar

Muhly and commelina (from the spelling bee word list) are words I use at least annually. They're both common plants in North America. Most garden centers sell muhly grass. Commelina, aka dayflower, is a wildflower/weed.

People in oral cultures probably knew more plant names than people nowadays.

Yug Gnirob's avatar

I assume eclaircissement means "stop eating that eclair."

Sui Juris's avatar

I use ‘Chaldee’ quite often.

4Denthusiast's avatar

The place names seemed like odd examples of words an illiterate culture wouldn't have too. Surely the people around Cupar and the Uaupés would have words for the places whether they were written or not.

John Stathatos's avatar

Regarding the interesting question of what possible practical use 20th/21st century knowledge might be to a time traveller stranded in the distant past, I remember reading a science fiction novel decades ago about a man suddenly transported to Rome at the time of the Gothic wars (535 to 554 AD). He eventually becomes involved in the struggle to free Italy from the control of Byzantium and the emperor Justinian, but first he must survive and earn a living. His answer is to produce and market a kind of eau de vie - a distillation of brandy from wine, which is simple enough in theory and becomes feasible once he solves the problem of acquiring fine enough tubing to make a working still.

John Stathatos's avatar

I vaguely rembered the author probably being Sprague de Camp, so I asked Perplexity to "identify a science fiction time-travel novel from the late 20th century about a man suddenly transported to Rome at the time of the Gothic wars (535 to 554 AD)". It came up with the answer, but also corrected my dating:

"The novel is Lest Darkness Fall by L. Sprague de Camp. It is an alternate-history/time-travel novel in which American archaeologist Martin Padway is struck by lightning in Rome in 1938 and wakes up in Rome in 535 AD, at the start of the Gothic War. It was first serialized in Unknown in 1939 and published as a book in 1941—so it is an early-20th-century novel rather than late-20th-century, but it is almost certainly the book you mean."

Kenny Easwaran's avatar

I’m skeptical that “ephemerality” is the dimension that does the relevant work in that section. The vast majority of internet writing *is* ephemeral in the same way as a note to a friend in class, or a post-it, or a concert ticket. It’s true that any of these things *could* be saved and embarrass you some day, but most of them are designed to be seen once (or by many people for one day) and then disposed of. Scrolling is basically just as ephemeral as orality.

Kenny Easwaran's avatar

I definitely did not expect the twist at the end of this one, where the person endorses visual writing. I can see that there are things in its favor, but I don’t think it’s competing on the same dimension as the orality/literacy axis.

(Also, I expect most people read more words a day now than most people did a century ago, but they’re reading ephemera that has oral features rather than literate writing.)

Demarquis's avatar

"The purpose of writing is to relieve your brain of remembering so that it can focus on reasoning." This isn't true. It can be true, but it isn't the most common use for writing. Writing is for transmission fidelity - when the thought that you put into someone's head has to be precisely what you intend. Most people never did write much outside of school or work. There was never any need to. The question is how important information fidelity is in the modern world. If people have a need for it, the will communicate accordingly. But that's not what they want - they want to share ephemeral feelings, to participate in the creation of a dynamic, ever changing consensus of reaction to something. For this purpose, memeing is superb.

Gullydwarf's avatar

FWIW - some children books do some of tricks author suggested, e.g. all Geronimo Stilton books have 2-3 words that stand out per page

(Example - https://a.co/d/0b2TRCEs)

Pjohn's avatar
2dEdited

I was with you every step of the way… I agreed with every point… I had thought of most of them for myself… until we came to 17776. From that point onwards, we parted ways!

I was frankly astonished to read praise not of 17776’s story (which I would have understood) but of its format. When I last tried to read it, I just found myself screaming, inside my head, “JUST WRITE THE BLOODY THING DOWN! USING NORMAL WORDS, IN A NORMAL SEQUENCE, ON A NORMAL PAGE! I PROMISE that I will understand them! I PROMISE that mere words are able to command a person's attention, and it doesn't take bullshit JavaScript gimmicks to hold people's attention for more than 30 seconds!”

I have tried to read 17776 several times, and abandoned it in frustration each time. I am certain that, had it been written as a novella, I would have read it through, avidly, more than once.

I still don't understand how we could have been in complete agreement on every single prior point about the history, the values, the current situation, all observations and predictions - but that you can conclude that the solution is things like 17776, tweaking the layout of Joyce's reverse-typeset characters, adding multi-coloured, multi-sized text, etc., and provide about a dozen examples of “improvements” to the written word, all of which I find to be annoying, useless, shallow gimmicks.

I entirely agree that if your goal is to convey information in the absolute most efficient way possible, you probably do want a mixture of graphs, tables, images, branching tree-like structures, and text, that features considerably more of the former four than is traditional - but in all the writing where the point isn't merely to maximise “ink efficiency” at the cost of absolutely every single other thing an author might care about, I want the information conveyed by means of beautiful, elegant prose rather than by means of extra visual stimulation.

I want to read things that are capable of holding my interest, sticking in my mind, and transporting me into the minds of others, despite just being ordinary written words on a page.

I want prose that conveys ideas by its use of language, in all its unfathomable complexity and richness.

I wholly reject the idea that you can't convey spatial things, branching conditionals, tone of voice, levels of epistemic confidence, or anything else, effectively, using but normal prose.

I think adding extra icons, distributing words about the page in complex shapes/structures, animating the words, changing their colours/sizes/orientations, etc. etc. is just another way of impoverishing the art of writing. Adding visual stimulation without adding anything to the actual ideas (and I am sure, in nine times out of ten, actual subtracting from the ideas the better to deliver the stimulation..)

(nb. I think this impoverishment-of-language may be entirely worth it when the goal is to convey information as accurately and efficiently as possible and nothing else, in which case our tables and numerical probability estimates gives us back more than we lose by replacing complex sentences and words expressing uncertainty. I'm talking here about all the writing, from Substack to Shakespeare, where 'maximally efficient transfer of knowledge' isn't the point of the writing.)

I want the concept of unfathomable, inhuman, inhospitable, abyssal depths to come alive inside my mind owing to the genius of Jules Verne’s prose, not to have the deeps demonstrated by (say) putting the text in dark blue and animating it to sink to the bottom of the page, as if communicating it to a small child whose interest would otherwise wander without the stimulation of colour and movement.

I want to feel the wit and charm and elegance of Jane Austen's debutantes and suitors as though I were transported into their lives for a while - even if that means I don't quite grasp all the multi-layered social complexities Austen captures in creating this world the reader may temporarily inhabit - rather than having all the subtle social complexities set-out for me in a big unmistakable table (or some entirely unique type of graph or infographic) that teaches me exactly what is going on but doesn't transport me to anywhere.

I want to revel in experiencing the outrages of P. G. Wodehouse's autocratic aunts and peppery old colonels by means of Wodehouse's inimitable smilies, the eloquently acerbic language these characters employ, and the comical-but-utterly-human ways other characters interact with them. I couldn't experience any of the magic of a Wodehousian idyll by merely seeing their dialogue printed in a larger font.

And the idea that all the things that would be lost by replacing ordinary prose with an animated, multi-coloured, shape/size/position-varying fireworks display could be made-up for, for the benefit of the visually-impaired, by means of AI summaries is just so utterly unpleasant to think about that I can't bring myself even to formulate arguments against it.

I challenge you to try it: find your favourite passage from literature and give any dialogue descriptions of shouting/exclamation/murmuring/whispering a change in font size, add icons and emojis, tabulate any structured information, put the king's text in purple and the bishop's text in red, giving any expressions of confidence or belief or uncertainty numerical values, and so on: see whether it adds to the experience of reading the passage, or detracts from it, or makes no difference. (Then, try removing any original prose that's also communicated by means of one of the new visual features, and have an AI summarise the visual part and integrate its summary back into the text....)

Taymon A. Beal's avatar

Without getting into the broader questions here, it is just empirically the case that lots of people greatly enjoyed 17776, and I don't know of any of them who think it would have been better with a more conventional format.

Pjohn's avatar
1dEdited

I don't deny this! I'm just surprised that I can so wholeheartedly agree with the author on every single prior point but then disagree irreconcilably about this one! De gustibus non est disputandum..

Hedonic Escalator's avatar

I strongly disliked 17776. I've enjoyed other stories that used unconventional formatting, but I think 17776 did a poor job with it past the intro.

SMK's avatar

My reaction here was almost the opposite of the last review: here, the best (only good) parts were about the book. Those parts were mostly quite interesting (if at times I thought they seemed naive or tendentious).

Sadly, most of the review was not about the book, and at times, it was barely coherent. I'm highly skeptical that the author's proposals (to the extent they are well-defined) would at all preserve what is effective in writing -- the precision is really quite key to why it's so important, and many of them seem at least as deadly to precision as oral culture is.

I also disagree that only oral culture is fighting back. Getting kids off smart phones is great for both, obviously. If it's raining, or they're lonely, kids will have to resort to books. (At least as much as they did before -- never were all kids readers.) And almost all kids are lonely much of the time.

Finally, the author complains that nobody is even trying to defend written text. That of course is false. Here is one good example.

https://edwardfeser.blogspot.com/2023/07/loftons-youtube-straw-man.html

(Note: I have no dog in the underlying fight.)

Tantz-Baer's avatar

Thanks for this review, this was one of my favorite books in my PhD years and I really enjoyed reading about it!

And I realized something about the ubiquitous, here orality-advocate quote "Kids need to be kids again": this always translates to "kids should be exactly like we were when we were kids" (touch grass, date, get drunk etc), always going back exactly 1 generation. Nobody reads it as going back multiple generations, "kids should again work on farms and marry their cousin at 16"

Performative Bafflement's avatar

Well, I for one will be quite pleased when my teenage sons decide to pursue the cursus honorum, and will be proud to sponsor their quaestorships and praetorships.

Yug Gnirob's avatar

I have never read House of Leaves. That page does not make me want to read House of Leaves. Is that layout supposed to look like something? It doesn't look like something. Is it just playing with the word direction? But every direction is the same phrase, and if it's going to be the same phrase, why only the two paths? And why is he riding his bike inside the house? That's unseemly.

John Stathatos's avatar

House of Leaves is a novel with at least three parallel narratives, but the main one is a nicely realised horror story which I can recommend. The pissing around with eccentric layouts is post-adolescent showing off, and can be ignored. It would be nice to see it reprinted without all the superfluous nonsense.

Viliam's avatar

> post-adolescent

Adding the word to my list of sophisticated insults.

Ralph's avatar
1dEdited

The typographical quirks serve the specific scene being described. It's not all (or even mostly) like that. You can think of it as the textual equivalent of "switching to the shakey camera" in a chase scene in a horror movie.

You'll have lovecraftian "impossible twisting hallways" described in formatting that "twists", or if someone gets isolated in the dark you might get a thinning out of words such that each is "isolated" on one line.

What's happening in this specific scene is that "his house" has become some spooky infinite extra dimensional space, and the physical structure is "becoming incomprehensible".

You can imagine the parallel paths like someone looking down a hallway and seeing themselves at the end (spatial loop-style).

I think it Compares favorably to Lovecraft describing R'lyeh in "The Call of Cthulhu"

----

"the damp Cyclopean city of slimy green stone—whose geometry, he oddly said, was all wrong [...] He had said that the geometry of the dream-place he saw was abnormal, non-Euclidean, and loathsomely redolent of spheres and dimensions apart from ours."

----

The book has a few layered frame stories (spooky haunted house, obsessed strange academic researching the haunted house, mentally ill person reading the research) that are all, roughly, about "desperately getting lost in something" while trying to "escape something". On some levels it's literal, on others it's like "losing yourself in gambling to escape a shitty marriage".

I like it a lot. IMO, it's not the tightest writing in the world but the gimmicks work for me and the thematic resonances between the layers are cool to see.

Patrick Stuart's avatar

I asked Grok for Tufte-style infographic depicting just how awful I found the closing argument to be but the A.I. refused to show “my desire to see the reviewer rendered down to paste and fed to dogs” as part of the information included. So much for Elons dream.

moonshadow's avatar

> The equivalent of ‘code like you talk’ would be to code in Python, a much more popular language, but sometimes you have to use Lisp

This is very domain dependent. Computer languages are not like real languages: typically, there are around a hundred words to learn, and some very simple grammar rules, and the rest is up to you. What they /do/ do, and the reason we have so many different ones, is very naturally fit the patterns of speech and thought when discussing and solving the particular kinds of problem they are primarily intended for.

Depending on the problem and on your speciality, Lisp /is/ coding like you talk, with Python being unnecessarily indirect and verbose (and for other problems, it is the other way around).

Actual human languages are more generally applicable, and in exchange for this are much more complex.

Candide III's avatar

I don't know if there was anything similar in ancient oral cultures, but in (ex)USSR there used to be an interesting oral culture genre counterpart to the detective mystery. One name for it is 'situation game'. The game master describes a mysterious happening, then the rest of the participants discuss it, ask him questions, and repeat the process until they are sure that they have figured out what had actually happened. They then describe their solution to the game master. If it matches closely enough, they win, if not, the game master wins, and if the participants wish it, describes the actual solution (sometimes people refuse this because it disqualifies them from being an active participant when the same 'situation' is played elsewhere). The game master must answer questions truthfully (according to the fixed plot he knows) but is not obliged to answer fully, include necessary context etc. The mysterious happening, of course, always has a logical, if outre, explanation, or the game would make no sense. This game used to be popular in hiking, mountaineering and spelunking subcultures, where groups of people had time to while away without much in the way of modern amenities.

Crooked Bird's avatar

This is fascinating, thank you.

Tom Craven's avatar

How do emojis not get a mention? 🤔

Bugmaster's avatar

I was also looking for emoji, as well as 1337-sp33k and its modern descendants like "lol", "jk" etc. The latter are especially notable for making their way into oral speech "backwards" from written text.

Freddie deBoer's avatar

Funnily enough the head of my PhD program was one of Walter Ong's students so I guess I'm part of his lineage

magic9mushroom's avatar

This one doesn't have the header saying that it's by an ACX reader who'll remain anonymous until voting's done. Is this intentional (i.e. are you saying it's yours?) or a mistake?

John Stathatos's avatar

In the context of this review, I remembered Jorge Luis Borges' essay The Kenning and looked up my copy. Two passages seem of immediate relevance.

One is "Literature begins with poetry, and poetry with the epic; it is as if before he spoke man sang. Since the origin of literature is spoken, this order may well be attributed to the mnemonic nature of verse. In India, as Winternitz tells us, the legal code was framed in verse so that it might be committed to memory."

And the other, "The themes of the epic are limited; alliteration forced the poet to coin synonyms. Like Greek, the Germanic languages led easily to the formation of compound words. This facility suggested the sought-after synonyms, which, at first, were rarely metaphors. [...] Kenning (the Norse plural form is kenningar) is the name that the Scandinavians gave to these metaphors. In England, poets came to feel that such figures of speech were burdensome, and with time their use began to diminish. Ninth-century elegies include only a few, or only those that were felt as common synonyms, not as metaphors."

Like the Homeric singers, Scandinavian skalds could call upon a wide range of kennings familiar to their audience; amongst the more familiar to today's readers may be "the whale's road", for the sea; "the red swan's meat", for the dead; and "the King of the Greeks" for Jesus, for the sufficient reason, as Borges notes, "that King of the Greeks is one of the titles for the emperor of Byzantium, and that Jesus is surely no less.".

Nikita Krikunov's avatar

This is a good review, albeit very brief (for such a fundamental subject as orality and literacy, it's not long enough). But isn't this a BOOK review contest???

I also agree with everyone who says that the proposals to change writing make it resemble spoken language, not code. House of Leaves-style writing is found in many works: SCP, Homestuck, YouTube animations—all products of the internet age. Graphics don't make speech more like writing because they attempt to present text visually, like a painting, or auditorily, like a speech. Visual and textual perception are significantly different, and auditory perception is orality

For example, in the first sentence I put a word in all caps and used three exclamation marks instead of one; this is a way of conveying intonation

vantablight's avatar

Author needs to read further into the Ong/McLuhan universe since what they are proposing is a return to ornate manuscript media which is distinct from modern repeatable type print and is not necessarily the entrenchment of modern literacy they think of it as.

Tatu Ahponen's avatar

Meandering, boring, badly formatted, couldn't finish.

moonshadow's avatar

"No climax, no point, no meaning."

...but I digress.

gordianus's avatar

The part about Ancient Greek oral poetry is actually something I know a good deal about. Some notes:

- The part about oral language having a limited vocabulary has an interesting extension: plenty of words that fell out of use in common speech remained in use in poetry. (This was, indeed, a substantial part of what distinguished the ancient poetic register from ordinary conversation or prose rhetoric.) But as with ordinary words, they needed to remain in regular use to be preserved. There are even examples of words that fell out of use except in one specific phrase, whose meaning was forgotten & later guessed incorrectly. (The example I know of is μέροπες, whose apparent original use as a local ethnonym is preserved in the geographical catalogue copied in the Homeric Hymn to Apollo (Μῑ́λητός τε Κόως τε πόλις Μερόπων ἀνθρώπων), but in the Iliad the phrase is used generically to mean 'mortal men'. Though the same happens even in modern literate language, when a rare word is repopularized in a single high-profile use: eg. 'wight' or 'nimrod'.)

- What the review calls a "cliche" in oral poetry is more usually called a formula, & it has the additional constraint of having to fit into a meter: thus there were generally several different formulae to describe a particular person or thing, which fit into different-sized metrical gaps, & (because case-endings changed between words) these were often different depending on the case of the name/noun. So, for example, just in the beginning of the Iliad, Apollo is called ἑκηβόλον Ἀπόλλωνα (originally 'who strikes where he wishes', but reinterpreted to 'far-shooting'), ἀργυρότοξε ('[having a] silver bow'), φοῖβος Ἀπόλλων ('bright' or 'radiant'), Ἀπόλλωνος ἑκατηβελέτᾱο ἄνακτος ('the far-shooting lord'), & where the name needs to be extended to fill a whole line, Ἀπόλλωνι ἄνακτι τὸν ἠύκομος τέκε Λητώ ('Lord Apollo, whom beautiful-haired Leto gave birth to').

- The bit about oral poetry having less tolerance for idiosyncratic subjects needs to distinguish between oral *composition* & oral *transmission*. The archaic poets whose writing preserves a good deal of oral technique wrote plenty of poetry whose subjects are of only local interest: Sappho & Alcaeus wrote love poetry, Theognis wrote aphoristic advice to his friends, Pindar wrote praise of specific aristocrats' athletic victories, & Ibycus offered to immortalize the memory of the tyrant Polycrates of Samos by writing poems in praise of him. From comparative evidence it appears that similar genres go back to the prehistoric Indo-European tradition, but -- presumably because their subjects pertained mainly to a particular context -- the versions of them that were composed before the advent of literacy were not preserved. The oral poetry that was preserved long enough for its original composer to be forgotten (mainly that of "Homer", a false identification because the Iliad, Odyssey, & Homeric Hymns seem each to have been composed by different people) is generally either mythological, or about mortal men who did great deeds, mainly in war. It follows that, although the promise of eternal glory to the subjects of poetry was a long tradition (Ibycus: τοῖς μὲν πέδα κάλλεος αἰὲν/καὶ σύ, Πολύκρατες, κλέος ἄφθιτον ἕξεις: "among them [the Greeks] you too, Polycrates, will forever have undying fame for your beauty"), in most cases these pre-literacy praises did not long survive their composition.

- The part about oral thought being less able to represent abstraction is partly true, but only partly. Even one of the particular examples given is not true: the Iliad has Achilles "think about hypothetical soldiers", & whether fighting is worth it in general, in his argument with Agamemnon. But, of course, this is embedded in the context of the story, about particular warriors fighting a particular war, & even if you accept the modern argument that the Iliad's plot is making a critical point about the archaic-Greek warrior ethos in general, it's true that that has to be expressed through a particular example.

ilya187's avatar

> People are reading less than ever

This is both trivially false (before widespread literacy most people did not read anything) and false even if you only look at the modern times:

https://testprepinsight.com/resources/us-book-reading-statistics/

The age cohort which reads LEAST is the people born in 1971-1980. Millennials read more, and Gen Z more still, although not quite as much as people born before 1961.

Scott S's avatar
1dEdited

I've always considered the book part of these book reviews to simply be the surface excuse to dive into one topic or another, so unlike the other commentors I'm unbothered that I still don't know what Walter Ong's main arguments were. I enjoyed the quad chart framing and the overall exploration of its corners. Scrolling delenda est.

And who doesn't love a good call to arms! I'm feeling properly inspired to plop beside Pangur Bán and pen the next exploration of the typographical periphery.

Lucas's avatar

>Code is writing on crack. Unlike regular writing, it basically can’t be spoken or listened to. It is extremely analytical and precise, in a way that writing rarely achieves.

I don't think this is true. It can be spoken (talon is software that allows you to do that), and I’ve listened to it plenty of time. It looks hard when you have no knowledge of it, this is true, and notice, this is true of music notation as well!

Lucas's avatar

I finished the review now. The idea with the quadrant is interesting.

Some arguments are weak. The code I mentioned earlier for example, or saying the literacy camp has nothing on why reading/writing is better. I have one: reading only progress when you are actively reading, same for writing. It is very very hard to be as passive writing or reading as you can be listening or watching.

The review at the end argues for adding more tools so that writing can get back some of the advantage. That makes sense, but why not demonstrate it for that review instead of very small example? The review speaks about programming languages, and let me tell you, everythomg looks great and simple with trivial examples, even Lisp! In the review itself, most images didn't add much.

Also digression, Paul Graham point about Lisp was economy of expression, and today that looks like telling Opus 5.5 “so this sms notification stuff is pretty good, now add email and push notifications too”. That doesn't give you economy of expression during code review or when other people read your code but that part was always about shared knowledge (do you know the cool library that allows you to be super concise?) rather than programming languages themselves.

The speed advantage of reading vs listening is not mentioned, and that's before counting graphs.

PowerPoint is not mentioned but it seems closer to the ideal presented here than substack articles, especially the slides where you can go left and right to progress through the argument, or up and down to go into more details.

To finish on a more positive note, it's full of conversation starters that go in many different directions. I do prefer when a book review is more focused and that's what I expected here. That does make this one ambitious, which I like.

mabel's avatar

I really like a lot of the discussion in the last segment - enough that it kinda nerdsniped me and I spent like 7 hours making a spec for a (mostly)plaintext file format that supports embedding svgs as part of the text, for custom punctuation. It's just a spec for now, and I'm probably not going to build a *platform* out of it (though I'm sure I'll make an html exporter in whatever local app I build), but I figured the author might enjoy hearing how much this got to me. Good stuff!

Enow's avatar

The no-spaces Latin detail is the best part. Writing started as a transcript of a guy talking, and it took centuries before anyone thought a page could hold an argument instead of a speech. Every inscription was basically a very patient voicemail from a king.

Gon-no-suke's avatar

> The equivalent of ‘code like you talk’ would be to code in Python, a much more popular language

This should be Perl, not Python, since one of the reasons for creating Perl was to make a programming language exhibiting aspects of natural languages.

Sholom's avatar

The idea that we can't stop the scroll is lazy nonsense. A couple thousand middle-aged women wearing 50 pounds of lace through mud and sun managed to ban alcohol for the whole country, at a time when most people drank beer for breakfast and split a bottle of brandy with their spouse for dinner every single night.

It was hard, and it ultimately didn't work, but A) We have proof it's possible to ban something popular if you can convince elites that it's bad enough B) It's a lot easier to enforce the regulation and suppression of services requiring massive infrastructure and network architecture and who by definition rely on massive economies of scale to attract users than it is to enforce the ban on a physical product that can be made at any scale, anywhere.

Hedonic Escalator's avatar

"Code is writing on crack. Unlike regular writing, it basically can’t be spoken or listened to."

Pedantic objection, blind programmers typically use screen readers. See discussion on HackerNews, for example: https://news.ycombinator.com/item?id=5352608

But this is more literacy in an auditory modality than it is orality, since you can take in information much faster than natural speaking speed, rewind with perfect fidelity, etc., so your point stands.

Hedonic Escalator's avatar

17776 sucks, by the way. Unbearably tedious.

John Fast's avatar

I primarily identify as a ttrpg Gamemaster. A lot of us run adventures with a basic structure and a lot of improvisation tailored to the audience, meaning the players, the way oral poets recited epics.

Stop the Slop.'s avatar

the ai punctuation mark idea is great. the em dash already got drafted into that job without asking

DrMcleod's avatar

On the other hand, when I recieve emails that are formatted in some fancy way and embedded with images, I usually don't bother to read them at all. I assume that it is just corporate slop trying to sell me something (such as acquiring a positive view of the latest winner of corporate musical chairs in my place of work). Alternatively, raw text from the same source looks like it might actually contain useful information that I should know about.

Enow's avatar

The no-spaces Latin detail stuck with me. Punctuation reads like a favor to a reader who can't ask what you meant. The time-traveler opening is a good way in, because the gift turns out to be the thing we never notice.

Gia Gupta's avatar

I love the idea of writing trying to go beyond just a transcription of speech. I've wondered if you could make written language more beautiful or interesting (similar to rhythm in spoken word), kind of like the glyphs in Ted Chiang's "Story of Your Life" (“The semagrams seemed to be something more than language; they were almost like mandalas.”). Similar to that world, reading would not only have an aesthetic visual appeal but enable us to think in a way that is different than speech (nonlinearly? think at different scales?).

Deiseach's avatar

I like the part about the table of the ships: "This Wikipedia table is both hard to memorize and impossible to say out loud."

*Somebody* never did learning by heart and rote memorisation in school! Or reading texts aloud in school! It's perfectly possible to read that aloud. If what they mean is "Unless you add in 'such-and-such came in so many ships led by so-and-so from these settlements' it won't make sense'", okay but that is **not** the same as "impossible to say out loud".

Ready for more?