Comment by bravura
5 years ago
"[Hunter S. Thompson] chose, rather than writing original copy, to re-type books like The Great Gatsby and a lot of Norman Mailer, the Naked and the Dead, a lot of Hemingway. He would sit down there on an old type-writer and type every word of those books and he said, 'I just wanna feel what it feels like to write that we'll.'"
HST: "If you type out somebody's work, you learn a lot about it. Amazingly it's like music. And from typing out parts of Faulkner, Hemingway, Fitzgerald - these were writers that were very big in my life and the lives of the people around me - so yea I wanted to learn from the best I guess."
http://brianjohnspencer.blogspot.com/2014/06/hunter-s-thomps...
This was explained by HST in one of his letters, which was collected in the excellent three book collection of letters he sent and received to his friends. Including many famous writers.
HST was always great in small rapid outputs of writing, which is captured well in his letters (similar to how his collection of articles are his most popular works, but these deserve a similar look).
He obviously had some sort of ADD and later on combined with a long series of drug/alcohol addictions, so it makes sense he was better in short blurbs. Even his most famous novel Fear and Loathing in Las Vegas has a feeling of multiple long spurts on a typewriter.
Which is always how he wrote. Always also at the very last minute of the magazine due date and/or because he was running out of money and needed the next advance.
I believe this is common in creative fields. Long periods of meh and spurts of greatness.
Anyway the book series is here, usually called the Gonzo Letters:
https://www.goodreads.com/series/64386-the-fear-and-loathing...
Only the 2nd one has a Wikipedia page for some reason (the 3rd one came out in 2014) but the first one (The Proud Highway: The Saga of a Desperate Southern Gentleman 1955–1967) as a young writer who is often desperate and broke was most interesting IMO, even though his life or writing wasn’t yet as it would become famous for (but definitely still as wild) https://en.wikipedia.org/wiki/Fear_and_Loathing_in_America
Inspiration is lumpy, so I've adopted the Fieldstone method [1] (articulated by the prolific Gerald Weinberg).
In my implementation, I collect little thoughts (shower thoughts, observations, good turns of phrases, etc.) into a single continuous Google Doc. I revisit it often and guided by my current emotions and interests, try to coalesce like-ideas and rewrite them into a large idea. Some ideas eventually snowball into something substantial.
Unless you're a columnist with a deadline (with innate talent driven by adrenalin), everybody knows how difficult it is to write an essay from scratch. However if you've been collecting ideas, and have been developing and coalescing and rewriting them over and over again (often for years), the essay almost writes itself.
[1] https://www.skmurphy.com/blog/2016/05/04/weinberg-on-writing...
I have a similar workflow, but for a different reason. BJ Novak described in an interview how the ideation part of the process is totally different from the productive part. This spoke to me, and I've been approaching them separately ever since. When they get too inflated I either get writers/coders block or absolute spaghetti, so it's easy to see when I've grown less disciplined with the division.
1 reply →
I find this to be true with computer programming/work for me. I’m productive in bursts.
I'm productive in bursts as well. I have the feeling that in between bursts things are still computing in the background so it looks like you're not doing anything but the brain is still churning. This type of bursty productivity is very hard to account for in a corporate environment where one's productivity should be accounted for on a daily basis and laid down in cookie cutter time slots. I often wonder whether I have ADHD since I can hyperfocus when Im in a productive burst. Unfortunately/fortunately I don't tick other boxes so I'm in a limbo with the diagnosis
28 replies →
Still waiting for the third volume to actually drop? Can't find it available and all the amazon reviews are 5-star complaints about the delayed release :(
Oh weird, I wasn't aware they still haven't release it yet. It's been 19 years since the 2nd volume. That's really odd. They didn't fulfill the preorders in 2014 either https://www.amazon.com/gp/product/0684873176/
I thought making it through the other two 700+ pages of letters was quite an achievement. I was planning to read the 3rd one at some later point in my life.
I guess I'll still have to wait ... the title was really great too "The Mutineer: Rants, Ravings, and Missives from the Mountaintop, 1977-2005".
You need the long periods of meh for the bursts of great to happen. The meh puts in the ground work for the good.
Agreed that you learn much about a work (I've retyped a few articles and books simply to get a usable copy). But copying is not the same as creating --- you'll see the finished product, and may intuit hints as to how it was created, but what you're not getting is the creation process itself: research (especially for nonfiction), structure, plotting and character (for fiction), editing, rewriting, restructuring and reordering content, additions and deletions.
Writing (and reading) short disconnected bits is fairly easy: nothing interrelates strongly, composition is simple and forgiving. Longer works are complex: they have structure, arcs, need to retain interest, jumps and connections need to be plausible, the whole be consistent (unless inconsistency is used for effect), etc., etc.
And you don't get this knowledge simply by copying out great works.
I see a similar failure-to-grasp in some proposals for hypertext or advanced publishing systems. Reading is inherently serial, in that we follow lines of text on a page. Interactivity --- usually defined by the ability to skip between previously-written segments --- just offers more serial paths.
The value of hypertext and related tools may be far more on the writing process, where vastly more sources can be referenced and cited with greater ease. Some might be incorporated into the final work, but an excess of interconnections and quotations is itself distracting.
I see this as a particular blindness of Ted Nelson's Xanadu project, despite many fascinating and original elements to it.
> And you don't get this knowledge simply by copying out great works.
Well, I think the point is that you may get knowledge, but nt necessarily of the type that lets you make another great work. Doing is an act of learning, whether very small amounts of learning as you do something you've done countless times before, or possibly large amounts of learning as you try something absolutely new, or learning about yourself and what you like, what you don't, what you feel is worthwhile even if you dislike it, etc.
Ultimately, there's something to be learned from everything. Even the book you read that purports to teach you someething that you find worthless for the task should be illustrative of either what things are not useful for you in learning that, or how to structure something in such a way that it's hard for you to take away good meaning from it, and ultimately, some hints about what to look for next time that doesn't have the same problems.
So, copying isn't necessarily creating, but it should be an act of learning, and learning should hopefully help in creating, even if it's something totally divorced from what you were originally copying.
I think we're mostly in agreement: copying is useful. It is not complete.
And any learning process which focuses largely on output or product rather than process, shares this deficiency.
To your other point, serendipitous discovery is very much a thing. I find that it is helped by haviing a conceptual structure or mental model which allows slotting new concepts from unrelated areas into a larger whole.
Recent example: moralising pathologies fallacy, in wildfire:
https://joindiaspora.com/posts/b4bbef90e8c60138513c002590d8e...
2 replies →
> Amazingly it's like music.
Maybe he means like jazz musicians? (I'm one.) Every musician I know has transcribed solos. You pick some solo you like but have no idea what they're doing, transcribe it, and learn to play it along with them. Bits and pieces perhaps will rub off in your own playing. You don't want to sound like them, but it's good to be able to if you want – and for that you need to absorb their style, so your body can just go into that mode, without having to think about it. I'm a piano player but have also transcribed sax, trumpet, bass lines, gospel songs, reggae, funk, .. even taps dripping, babies crying etc etc.
In language learning there's something called the "Input hypothesis" which states that we learn languages primarily by reading and listening (consuming input), rather than production of language.
The reasoning is that when you produce output you are by definition producing something at your own level, so little improvement occurs. When you consume input (and transcription is a great way to do it), you are consuming input of a native speaker - or in the case of Jazz, a master player - which is at a much higher quality level than what you could produce yourself. So input drives the learning process.
The exception is for motor e.g. pronunciation or playing technique, which do benefit from practice.
I don't think it matters what kind of musician. I'm mainly a electronic music producer and I've done exactly the same thing; recreate entire tracks from scratch from some of my favorite artists. It has helped me a lot in understanding every part of music production. Everything from designing every individual instrument to arrangement and finally mastering.
Hey! How do you do that? It is my dream to be able to make something as great as Worakis, Rone, N'to do. What tools do you use to recreate tracks?
1 reply →
Why would you take his general statement (musicians) and turn it into such a specific example (_jazz_ musicians)? You make it sound like jazz musicians are the only ones who would transcribe music to study it.
It's a safe bet that it's more common among jazz musicians than among classical musicians, who have access to the sheet music for practically everything they play.
Probably also more common among jazz musicians than rock musicians, who don't emphasize sheet music and music theory as much (as much, it's certainly more prevalent than outsiders would naively think).
And of course, all these categories blur together. There wasn't any reason to take his post and interpret it in absolutist terms.
In regard to typing out novels, HST went on to say that Faulkner (iirc) doesn't have an out-of-place word anywhere. So yeah, he was speaking about the style: putting down words in an interesting order.
I once read an interview with a Japanese novelist in which she said that, each day, she would copy a page or two by hand from a story by an author whose style she admired. She said that it helped her focus on sentence structure, word choice, and other details and that her own writing improved as a result.
One of the authors she mentioned was Yasunari Kawabata, who is widely admired for his writing style in Japanese. I believe she also said that she got the idea to do this from the tradition of shakyō (写経), the copying of Buddhist sutras by hand [0].
I have never retyped literary texts myself, but I used to teach English literature classes to Japanese students in which I would read the texts aloud in class before we discussed them. I found that, over the years, that experience heightened my awareness of writing style and maybe improved my own writing as well. I can see how retyping might have a similar benefit.
[0] https://en.wikipedia.org/wiki/Sutra_copying
> transcribe solos
Do you mean learn other people's songs from sheet music, or do you mean write down, from the sheet music, the same song again on blank sheet paper?
Edit: Or do you mean write down the song via hearing?
Transcribing solos means, you listen to the music and write it down, bit by bit. Someone's solo = their improvisation. If it's simple/slow enough that it's quicker to learn it by playing along, just do that. (I've never needed to transcribe Louis Armstrong, but have just learnt by playing along with him.) If it's very fast (e.g. Bird, Coltrane) you might need to slow it down to transcribe it. You learn how to transcribe it by doing it, it also trains your ear in recognizing melodies, harmonies, chord voicings, rhythms, forms.
You can buy books of solo transcriptions, but I've found that they're absolutely useless. (and always inaccurate) Maybe it's like reading a book vs. typing it out?! hehe. At first glance I couldn't see how typing a book out would help anything, I mean, it's already written down for you! The main element of learning by working out bit by bit what it is, isn't there. But there are elements in common - it's a little like playing a solo you've written down, and you would absorb the word patterns if you typed a whole novel I guess.
Ear training gives you this skill
e.g. https://www.youtube.com/watch?v=KhyyjRcrn84
Transcription is the latter -- create a paper version from a recording. Same as medical transcription is typing up dictated notes.
Reminds me ... in a computer nerd way ... of "python the hard way" (which used to be very open/free, now it has changed)
You would not read or download the lessons. You typed all the python in word by word. I think it really helped the learning process to type it out. It was slow and deliberate, even to mistyping and making mistakes (and fix them).
Yup! To learn music, cooking, martial arts, etc., you start by following someone else to build the muscle memory on doing the skill. As a follow-up, my own research [1] on retyping code showed students earned higher grades and submitted less erroneous code. Replicating technical skills is a common practice technique that seems to have been "lost" in current CS education.
[1] https://dl.acm.org/doi/pdf/10.1145/3373165.3373177
I still have a directory tree around from doing those lessons. Given how picky python is about indent spacing, I agreed somewhat with the philosophy behind those lessons, particular for people new to coding.
The analogy with music is interesting. Superficially, the musical activity which looks most like typing is simply playing an instrument. And sight-reading does feel a bit like typing. But I wonder if he's talking about transcription in that quote. One can sight read complex music without learning anything about it if they happen to just be a good sight reader. But transcription really does force you to pay attention to structural nuances you otherwise wouldn't get just by listening to a piece of music.
I think the idea is that typing is much slower than reading, so one has forced downtime in which it is possible to consider structure and choices[1].
Compare > "It will even at times be of value to read speeches which are corrupt and faulty in style, but still meet with general admiration thanks to the perversity of modern tastes, and to point out how many expressions in them are inappropriate, obscure, high-flown, grovelling, mean, extravagant or effeminate, although they are not merely praised by the majority of critics, but, worse still, praised just because they are bad."dfan
5 years ago
whymauri
5 years ago
sndwnm
5 years ago
junon
5 years ago
codeisawesome
5 years ago
tsumnia
5 years ago
codeisawesome
5 years ago
mistersquid
5 years ago
tweetle_beetle
5 years ago
ponker
5 years ago
atulatul
5 years ago
taneq
5 years ago
derwiki
5 years ago
moron4hire
5 years ago
http://www.perseus.tufts.edu/hopper/text?doc=Perseus%3Atext%...
and makes a good point about a pedagogical advantage to criticism of works other than one's own:
> "I will venture to say that this particular form of exercise, if diligently pursued, will teach learners more than all the text-books of all the rhetoricians: these are no doubt of very considerable use, but being somewhat general in their scope, it is quite impossible for them to deal with all the special cases that are of almost daily occurrence. The art of war will provide a parallel: it is no doubt based on certain general principles, but it will none the less be far more useful to know the methods employed, whether wisely or the reverse, by individual generals under varying circumstances and conditions of time and place. For there are no subjects in which, as a rule, practice is not more valuable than precept. Is a teacher to declaim to provide a model for his audience, and will not more profit be derived from the reading of Cicero or Demosthenes? Is a pupil to be publicly corrected if he makes a mistake in declaiming, and will it not be more useful, and more agreeable too, to correct some actual speech? For everyone has a preference for hearing the faults of others censured rather than his own."
It was very common in the past for composers to learn the craft by copying out the musical scores of the masters.
This is also why I always type in all the examples from programming books instead of cutting and pasting.
This is how I first learned technical writing. Found a few exceptional papers in my field, sat down, and literally just typed them out.
The downside is that it's easy to 'overfit' and lose your own voice.
For this to actually work you probably need to put the original text away, try to recreate a piece of it from memory, and repeat until you get it close to right. AKA the Benjamin Franklin method.
Just like anyone can trace or copy a picture drawn by a master, copying text won't grant understanding. What actually matters for a creator is the ability to produce content from higher level ideas, eg from whatever mental representation you used to hold the writing or drawing in memory.
> Amazingly it's like music
Programming is the exact same way. If ever I'm in doubt of someone's design and I have an inkling of respect for them, I'll try to design the code from scratch in my head, or sometimes even redesign a toy version myself. This usually surfaces some weird property of the problem that explains their code. You can also go through their thought processes and see exactly how they made decisions along the way. I'd argue you can even infer philosophical viewpoints in some cases, too - all without actually communicating with the author.
I love that about humans. One of the things I really admire about us.
Would be super interesting if this holds true for 'copying out' code as well... it's all the edge case handling that can get boring to re-type out without context / comments...
It does hold out! Here is a link to my research on retyping code [1]. To summarize my findings, students that regularly completed typing exercises earned higher course grades and submitted less erroneous code.
[1] https://dl.acm.org/doi/pdf/10.1145/3373165.3373177
Interesting - thanks!
This is myth and it is wrong.
Writing does not proceed letter by letter, word by word, paragraph by paragraph.
Earlier writing is deleted, moved, revised. Passages are moved earlier or later. Key passages are rewritten.
Etc.
Typing character-by-character an entire novel will likely teach you little about the act of literary creation.
I'm not sure the sources in the parent comment are making the argument that you are countering. Neither suggests to me the claim that using a touch typing website, in our case, replicates the complete process of writing literature.
The second quote in the parent comment is "If you type out somebody's work, you learn a lot about it." - the focus is learning more about the finished product. And I read "I just wanna feel what it feels like to write that well." as about experiencing the quality of the finished product.
Important to note that it was on a manual typewriter. The key is making it slow, so you can observe how the author is doing what they're doing. Reading is very fast ... typing on a manual typewriter is slow enough to give you time to observe.
Here is a Stephen Fry Oxford union video where he talks about trying to write from memory a scene from The Great Gatsby:
https://youtu.be/oSdLfPas8dw?t=546
Sounds a little bit like the process of training a GPT-3 style network...
He required copious amounts of Wild Turkey and cigarettes though
Dunhills, if I remember correctly.