Writing As Thinking

I refer to my previous blog post, My Current Understanding of the Copyright Law. I think that the most important objection to it is that 'writing is thinking'. I've seen a few articles on the internet about this, and it needs to be addressed. It's a passive-aggressive way of saying that not writing is not thinking, and as someone who thinks sixteen hours a day, I find that naturally offensive. 

I'm thinking of this article in particular. The idea is that "sentences push back, and it is in that pushing that much of the intellectual work gets done." "As the old E.M. Forster quote goes: ‘How can I know what I think till I see what I say?’" In this account, the "thinking" that writing is, is a dialectical process between the ideas in your head, and what the sentences you create formulate them as. Hence, it takes several drafts of writing to work out what you are thinking thoroughly and consistently. 

This is not a controversial thing to say about writing, and I agree with it. Unfortunately, it doesn't work as an argument against using an LLM. That would only work if all you were doing was prompting the LLM and then accepting whatever it output without thinking about it. I don't know whether anyone does this, but I know I do not. It doesn't matter whether you wrote the sentences yourself or got an LLM to generate them, because in either case, if you actually do re-read them, they push back either way. It's just that if you are prompt-engineering, then you are reprompting and pushing back on what the LLM has said. 

In any case, it fails to convince me in any form that using an LLM constitutes not thinking. You can unthinkingly use an LLM, of course — but you can also unthinkingly write. So, whether or not your text is well thought-out really has nothing whatsoever to do with whether you used an LLM to generate it. I said it before and I'll say it again, that there is only one real reason to prefer writing your own text. It means that you can copyright it incontestably.