Two notes about his book recommendations and three recommendations of my own.
I’m surprised that Might concludes that Strunk and White’s Elements of Style “is still a good, if not perfect, reference on style” since he cites Geoffrey Pullum who tears Elements of Style apart. People should take a look at Pullum here (PDF link) and here if they want to judge for themselves, but I think that Strunk and White is simply not a good book despite being very popular.
A tip about Joseph Williams’s Style book: buy the first edition, which you can find used very cheaply. Prentice and Hall keeps coming out with new editions, mostly I think to prevent undergrads from buying used copies and cutting into new book sales. (They now charge $55 for a paperback that is substantially the same as the book I taught in the 90s when it was about $12.)
Anne Greene’s Writing Science in Plain English is also terrific, and it’s very affordable. (Greene has a more narrow focus and is more basic than Kane, but don’t let the title put you off. Even if you are not writing about science, the core advice is excellent.)
If you care a lot about writing and you are interested in a less directly practical book, I highly recommend Clear and Simple as the Truth by Francis-Noël Thomas and Mark Turner. They take a philosophical approach to writing and treat “clear writing” as one style among others.
Sometimes I think that allowing scientists to publish datapoints and datasets (with clear protocols and supporting data) would be better than publishing the papers themselves. Papers can be full of subjective or misleading narratives or even not accepted if the quality of the prose is not good. Or flagged for plagiarism due to copy-paste. This to say that there must be at least 12 people in the planet that are not English native writers, whose contributions are potentially lost. Perhaps we should collectively consider some international scientific experiment report form…
Unfortunately, the collection, curation, and organization of data provides ample opportunities for subjective factors to influence interpretation. Science is a human endeavor and a form of communication, so it cannot avoid inheriting the challenges of communicating effectively.
I’m not sure it holds as print news continues collapsing, but my undergrad newswriting class helped with fluff like this.
It’s easy to pick up a fluffy style if you spend years writing to minimum page lengths.
Learning to write ledes and fit complex stories in a tight maximum length was a good counterweight. Most influential of the 12+ writing classes I took in undergrad + masters.
I can’t believe “highly” is not on the list of weasel words. And “significantly”, when just used as an intensifier:
Bad: I find these words significantly more abused in academic writing than everyday speech.
Good: I find these words even more abused in academic writing than everyday speech.
Bad: I find these words significantly more abused in academic writing than everyday speech.
Good: I find these words even more abused in academic writing than everyday speech.
Bad: This optimization is not semantics-preserving! It has introduced the claim that the words are abused in everyday speech, which (even if true) was not present in the input.
I don’t think any style manuals are useful for new writers or PhD students. They’re too abstract and they fail to give you a feel for what to do. You need the basics before you can benefit from them.
It’s much better to pick a few papers you like. They can be your advisor’s or just good papers that are written in a style you enjoy. Before you start writing for the day read a little. Don’t try to emulate them but you’ll get a sense for what they would do in your place. Particularly good when you’re stuck. As a writer it will be useful to build up a portfolio of such work: short stories, papers, books, blogs, etc. that stand out to you.
For grammar, Google docs is surprisingly good at catching a lot of low level mistakes. But it also has trouble with more advanced English constructions.
Grammarly helps with some of these. I’ve found it useful as a non-native English speaker. In particular to avoid passive voice, which is often use to make a text more impersonal in Portuguese (which might be considered a good thing depending on where you’re publishing the text).
Isn’t this, essentially, what Microsoft introduced two decades ago when it added grammar correction to MS Word as part of, what was it, Office XP?
Edit: apparently grammar checking was available in some form since Office ‘97 but the out-of-the-box wavy green underlining of grammatical mistakes as you type together with much expanded and improved grammar heuristics was actually not introduced until Office 12 (aka Office 2007). So just shy of two decades ago.
Is there anything you do to improve fish startup rimes for scripting? I enjoy fish but for small scripts I’ve started writing them in Bourne shell or zsh
LanguageTool is quite nice, even the open source version.
Linting text tends to have a problem: we rarely want to lint just text; it’s usually Markdown, TeX, whatever; so you’ll want to skip code blocks and literals, and some kinds of linting require understanding the structure.
For example; recognizing headers, sentences, etc.
And this is also compounded if you want to report errors nicely, with line/column information (or to do squiggly lines on real time on your editor). For XML-based formats this is relatively doable, and for Markdown and some formats there are nice parsers that will give you position information but… it’s not easy.
The Hemingway App is a similar set of heuristics. It’s an old tool, predating LLMs, and basically just has a bunch of hardcoded things for detecting passive voice and adverbs. Remarkably simple! Also it works great and has a nice UI. (There’s now a Hemingway Plus with more AI, I haven’t tried it.)
Two notes about his book recommendations and three recommendations of my own.
I’m surprised that Might concludes that Strunk and White’s Elements of Style “is still a good, if not perfect, reference on style” since he cites Geoffrey Pullum who tears Elements of Style apart. People should take a look at Pullum here (PDF link) and here if they want to judge for themselves, but I think that Strunk and White is simply not a good book despite being very popular.
A tip about Joseph Williams’s Style book: buy the first edition, which you can find used very cheaply. Prentice and Hall keeps coming out with new editions, mostly I think to prevent undergrads from buying used copies and cutting into new book sales. (They now charge $55 for a paperback that is substantially the same as the book I taught in the 90s when it was about $12.)
Thomas Kane’s New Oxford Guide to Writing offers a ton of helpful practical advice.
Anne Greene’s Writing Science in Plain English is also terrific, and it’s very affordable. (Greene has a more narrow focus and is more basic than Kane, but don’t let the title put you off. Even if you are not writing about science, the core advice is excellent.)
If you care a lot about writing and you are interested in a less directly practical book, I highly recommend Clear and Simple as the Truth by Francis-Noël Thomas and Mark Turner. They take a philosophical approach to writing and treat “clear writing” as one style among others.
S&W has the advantage of being really short. S&W’s competition isn’t all these great works, it’s authors never reflecting their style at all.
Short is only beneficial if the advice in the short book is accurate and helpful. I’d argue that S&W is neither.
That said, Anne Greene’s Writing Science in Plain English is also short (128 pages), though not as short as S&W.
Not reflecting on one’s style at all would be far preferable to reading S&W. It is better to receive no advice than bad advice.
If you want to do something similar, I would recommend checking out Vale for a more robust solution than just shell scripts.
I found a vim plugin that did most of this, and then improved it to do more & better … I should get back into it and figure out a good way to do the misused phrases…
My simple version of this (and my attempt to learn some vimscript) was this: https://github.com/jjasghar/vim-wrong-words-oss.
Admittedly the first plugin you found is so much more complete then my little highlighter.
Sometimes I think that allowing scientists to publish datapoints and datasets (with clear protocols and supporting data) would be better than publishing the papers themselves. Papers can be full of subjective or misleading narratives or even not accepted if the quality of the prose is not good. Or flagged for plagiarism due to copy-paste. This to say that there must be at least 12 people in the planet that are not English native writers, whose contributions are potentially lost. Perhaps we should collectively consider some international scientific experiment report form…
Unfortunately, the collection, curation, and organization of data provides ample opportunities for subjective factors to influence interpretation. Science is a human endeavor and a form of communication, so it cannot avoid inheriting the challenges of communicating effectively.
That was the general idea behind the center for open science: https://www.cos.io/
If you are looking for more than just a few bash scripts, I’d highly recommend the vale linter, which includes functionality for write-good.
s/in order to/to/gwould improve the quality of a lot of academic writing.I’m not sure it holds as print news continues collapsing, but my undergrad newswriting class helped with fluff like this.
It’s easy to pick up a fluffy style if you spend years writing to minimum page lengths.
Learning to write ledes and fit complex stories in a tight maximum length was a good counterweight. Most influential of the 12+ writing classes I took in undergrad + masters.
I can’t believe “highly” is not on the list of weasel words. And “significantly”, when just used as an intensifier:
Bad: I find these words significantly more abused in academic writing than everyday speech.
Good: I find these words even more abused in academic writing than everyday speech.
Bad: This optimization is not semantics-preserving! It has introduced the claim that the words are abused in everyday speech, which (even if true) was not present in the input.
1 is a highly composite number.
1 is a even composite number.
I don’t think any style manuals are useful for new writers or PhD students. They’re too abstract and they fail to give you a feel for what to do. You need the basics before you can benefit from them.
It’s much better to pick a few papers you like. They can be your advisor’s or just good papers that are written in a style you enjoy. Before you start writing for the day read a little. Don’t try to emulate them but you’ll get a sense for what they would do in your place. Particularly good when you’re stuck. As a writer it will be useful to build up a portfolio of such work: short stories, papers, books, blogs, etc. that stand out to you.
For grammar, Google docs is surprisingly good at catching a lot of low level mistakes. But it also has trouble with more advanced English constructions.
Grammarly helps with some of these. I’ve found it useful as a non-native English speaker. In particular to avoid passive voice, which is often use to make a text more impersonal in Portuguese (which might be considered a good thing depending on where you’re publishing the text).
Isn’t this, essentially, what Microsoft introduced two decades ago when it added grammar correction to MS Word as part of, what was it, Office XP?
Edit: apparently grammar checking was available in some form since Office ‘97 but the out-of-the-box wavy green underlining of grammatical mistakes as you type together with much expanded and improved grammar heuristics was actually not introduced until Office 12 (aka Office 2007). So just shy of two decades ago.
“Bad: There is very close match between the two semantics.
Better: There is a close match between the two semantics.”
Best: The two semantics match closely.
“These meanings are similar” or “These two things are similar”
I’d argue that the Oxford comma, which the author uses inconsistently, adds precision.
If I were bored I would rewrite them in fish shell!
Is there anything you do to improve fish startup rimes for scripting? I enjoy fish but for small scripts I’ve started writing them in Bourne shell or zsh
I once did it in Python.
I got Claude to port those Bash scripts to a web UI so I could paste code directly into it and see what came out: https://tools.simonwillison.net/writing-style
Claude transcript here: https://gist.github.com/simonw/e9902ed1cbda30f90db8d0d22caa06d2
LanguageTool is quite nice, even the open source version.
Linting text tends to have a problem: we rarely want to lint just text; it’s usually Markdown, TeX, whatever; so you’ll want to skip code blocks and literals, and some kinds of linting require understanding the structure.
For example; recognizing headers, sentences, etc.
And this is also compounded if you want to report errors nicely, with line/column information (or to do squiggly lines on real time on your editor). For XML-based formats this is relatively doable, and for Markdown and some formats there are nice parsers that will give you position information but… it’s not easy.
I wish there was some common tools for this.
The Hemingway App is a similar set of heuristics. It’s an old tool, predating LLMs, and basically just has a bunch of hardcoded things for detecting passive voice and adverbs. Remarkably simple! Also it works great and has a nice UI. (There’s now a Hemingway Plus with more AI, I haven’t tried it.)
Emacs artbollocks mode is pretty much all of this plus support for calculating reading ease. Combine it with flyspell mode and you’re set.