Academish Voice (2022) | Hacker News Reader

Academish Voice (2022)

inkandswitch.notion.site

80 pointsakkartik16 comments

Going off on a tangent here, but there's an inherent tension to "make no claims without support" when support typically refers to citations. The academic literature has become so large that you find support for almost any claim whatsoever, which makes it possible to string together an entire spiel that supposedly builds on existing science but is actually completely unsupported and void of any argumentation. Whenever you read "Method A has been shown to be a great way to investigate B" with no arguments, no summary, only citations, and especially when those citations don't mention page numbers but are meant to support hyperspecific claims, be suspicious.

Citations aren't proofs of the new paper claims. Sources are pointers to sarguments in support of your claims. It's the reader's job to evaluate those sources' arguments, and by citing sources the author is saved the effort of rewriting that part of the argument.

> It's the reader's job to evaluate those sources' arguments, and by citing sources the author is saved the effort of rewriting that part of the argument.

Hold on. I think it's the author and editor's jobs to evaluate the source's arguments before publishing them. The reason to demand accountability and responsibility from journalists is because we have to place a certain amount of trust in them. Trust but verify, sure, but it's not possible—let alone practical—to evaluate the sources used by every article you read. As it turns out, even journal reviewers don't have time to do that for a handful of quarterly articles, even when they are paid to. So, I don't know how lay readers would be expected to do this for every single article they read on a daily basis.

And I am not interested in saving authors the effort of rewriting parts of an argument, I think I want them to do that if it is necessary for their article.

This is so true. I review for conferences and it is not rare that I am forced to read half a dozen papers to verify what authors wrote and I get disappointed frequently because the interpretation was wrong or authors barely read the cited material in the first place.

> "Make no claims without support." "Be precise and specific with your descriptions."

I hope LLM's (which can be used to create grammatically sound but semantically dubious copy for nearly free) will mean these two points will become more relevant for all writing.

>for all writing.

sure, I'm going to have provide a lot more support for my various stories of time travel and gnomes.

As long as the rifle is hanging on the wall (the gnomes can time travel) at the beginning of the story, I won't mind if it goes off at the end. Deus ex machina is only problematic if neither god nor machine has been properly motivated; when motivated by being shown rather than being told, so much the better!

(compare https://uh.edu/fdis/_taylor-dev/readings/tolkien.html )

Edit: note the importance of canon even (especially!) in fiction...

What is the site? I was interested by the page but the site won't let me in without registering.

I saved a copy: https://readwise.io/reader/shared/01hr7znqjwtrj5c8zt0xm32ntw

I think they publish their research here: https://www.inkandswitch.com

Thank you, that's an interesting link.

Same

>Avoid absolutist language. It's not "the main problem", it's "a problem". (Unless, of course, you can defend the claim.)

I've started thinking this advice counterintuitively causes as many problems as it solves. I try to do it regularly (just had to stop myself now, unlike the author), but find that instead of the neutral, self-divorced report of discoveries academic style is designed to foster I just as easily end up with the overall structure still being my own opinion as drafted, but now much harder to read smoothly as it's clogged up by generic self-depreciating hedges and unjustly short summary references to other work I don't necessarily agree with anyway. It's also not how real science and study is actually done; Feynman famously did most of his progress on his work however he liked, and then only moved it to the formalisms of "academic research", so to speak, if he found most of what his hunch was looking for. I've been wondering whether it would be better to try out the other way and see if it can be made better instead of fighting it to find the right compromise of ingenuine humility. Perhaps write your own opinions on your sleeve, but use something like an asterisk to indicate a stylistic hedge you'd apply to something you only mostly know, or haven't completely reviewed to the same detail as the produced info that inspired you to the thesis, so the reader sees just where your head's at and makes up their own mind on whether it affected your conclusion too much, instead of having to guess between lines of alternatives or pull it out through teeth of qualifications. Plus whatever you seriously don't know a bit of and don't want to commit to will stand out all the more. Maybe pick out a less-used unicode punctuation for it, like the tombstone, or perhaps number them in the margins like mathematicians do with equations. As a bonus you could then attach an appendix as a numbered list that focuses as deep as you'd like on alternate possibilities and gaps in knowledge without drawing out the piece itself.

When I was learning music, my teacher insisted on giving just as much mental energy to the rests as to the notes, because even nominally the right melody loses its impact if it's allowed to sloppily over and underrun its rhythm, and doubly so for the harmony. A lot of "informational" style I try to read or write feels like just a long, flat run through the notes and chords its author needs out of an avoidance of expected criticism for expressing anything not buried under neutralizing boilerplate as uninformed, arrogant, or unsupported, but the reality is that sometimes that's how the sausage is made, and we do pupils and fellows trying to understand our work no favors when we put a curtain between them and the man at the machine.

That part made me think of things like the US Army officer’s guide to writing, and the postmortem of the Challenger disaster. The general sense of how communication should be direct, avoid the passive voice, put the “bottom line up front”. I don’t know if that guideline is an attempt to accomplish something different but I don’t think it’s useful in general.

a 10000ft style guide for the overall structure of technical essays and articles.

I've enjoyed how it focuses on the readers intent to quickly capture and find meaning in what you write, by how you organize your writing and how you make the organization obvious.

thanks for sharing

I find the notion.site page far more readable than their inkandswitch.com site. The wavering link underline on inkandswitch.com is a hard no for me.

Reply on news.ycombinator.com