Where the thinking happens

Writing is not transcription

There is a common model of writing in which you have the ideas first and then write them down. It is wrong, and it matters here more than usual.

What actually happens is that ideas are half-formed until you try to put them in order, and the attempt is what finishes them. You discover the argument does not work in the third paragraph. You find that two points you thought were separate are the same point. You realise the thing you assumed was background is the actual claim. None of that was available before you wrote; it happened in the writing.

This is why I know what I want to say, the model just puts it into words is a mistaken description of what has occurred. If a model produced the words, it did the ordering, and the ordering is where the discoveries live.

The stages, and what each one costs

Writing an essay has roughly six stages. They are not equally valuable, and knowing which is which decides everything about where help is safe.

1. Understanding the question. Yours. Getting this wrong invalidates everything downstream, and it is the most common cause of a bad mark.
2. Reading and note-taking. Yours, per Module 5.
3. Finding the claim. Yours, absolutely. This is the essay.
4. Ordering the argument. Yours. This is where the thinking finishes.
5. Drafting sentences. Mixed. The argument is fixed by now; this is execution.
6. Editing and polishing. Safest place for help, by a wide margin.

The gradient runs one way. Assistance at stage six costs you almost nothing. Assistance at stage three costs you the essay, even if every word is later rewritten in your own style, because the piece is now a defence of somebody else’s claim.

The outline is the essay

If your outline is yours (the claim, the order of moves, the evidence chosen for each) then the piece is yours, and everything downstream can be assisted without dishonesty or loss.

Which makes the practical test simple. Look at your outline. Can you say, for each move, why it is there and why it comes at that point? If yes, proceed. If it came from a model and you are reconstructing the reasons afterwards, you are editing an argument you did not make, and you will feel the difference in the third paragraph when you cannot defend a transition.

Writing to find out what you think

A practical technique that predates all of this and works better than anything else when you are stuck. Write badly, on purpose, fast, with no structure. Two hundred words on what you think the answer is, ungrammatical, contradicting itself.

Then read it. Somewhere in there is a sentence that is more definite than the rest. That is usually your claim, and you did not have it ten minutes ago.

This works because it lowers the cost of a bad sentence to nothing. Most writing paralysis is the requirement that the first sentence be good. A model solves that by supplying a good first sentence, which removes the paralysis and the discovery together. Free-writing solves it by removing the requirement, which keeps both.

Talking as a route in

If you think better out loud, dictate. Every phone transcribes free, and many assistants take voice. Talk for five minutes about what you think the answer is, get the transcript, and edit it into an outline yourself.

This is a good use because the content is yours. The machine transcribed; it did not compose. The distinction to hold: it is doing the typing, not the thinking.

The specific loss

When students hand over stages three and four, the essays that come back share a signature. They are competent, well-organised, and say nothing. They land in the middle band and stay there. Markers describe them the same way every time: no argument, descriptive, does not engage with the question.

That is not a detection problem. It is the accurate description of a piece whose claim was the average of what has been written on the topic, because that is what the process produces. And it is why the students who use these tools well are conspicuously careful about which stage they use them at.