I lost most of a year to a filing system.
Not to the system itself — to its absence, and then to my elaborate attempts to build one after the fact. I had the notes. I had read the things. What I did not have was any record of why I had made the choices I had made eighteen months earlier, and so I spent months reconstructing my own reasoning from fragments, badly, under deadline. Every one of those months was avoidable. I did not avoid them because I had never seriously considered that a person of my competence would need infrastructure. Infrastructure was for people who struggled.
That belief cost me a year, and I meet it constantly in other academics, so let me take it apart properly. Here are the five things people say when I suggest they need scaffolding for their research, and why each of them is wrong.
“That’s just being organised.”
No. Being organised is a trait, and treating it as a trait is precisely the error. What I am describing is a set of specific external structures that do cognitive work you would otherwise have to do internally, and the difference matters because traits can’t be installed and structures can.
Take the most basic one: stop highlighting and start extracting. A highlighted PDF is a record that you were once awake near a document. An extraction — the claim, the evidence offered, the method, one line on what it’s good for in your project, in your own words — is a record you can actually reason over eighteen months later. That is not tidiness. It is the difference between a literature review you write and a literature review you re-read eighty PDFs to write.
Or the decision log: date, choice, alternatives, reason. Four lines. It exists because your future self will not remember why you dropped the second case, and reconstructing your own argument under revision pressure is the single most demoralising task in academic work. Nobody’s personality prevents them from writing four lines.
“I’ve published fine without it.”
You have. On borrowed structure.
School gave you a syllabus. A master’s gave you deadlines somebody else set. Coursework gave you a reading list that defined its own scope. You never had to generate pace, scope or checkpoints from your own head, because institutions generated them for you, and you mistook their output for yours. That’s not a criticism. It’s just what an education looks like.
Then you got a project with one deadline in four years, no reading list, and no bell at the end of the hour. Left alone, minds treat everything before the final six months as not-yet-urgent, and then treat the final six months as an emergency. This isn’t weakness of will; it’s the predictable result of removing every external time signal and expecting the internal ones to scale. They don’t. Nobody’s do.
So manufacture what’s missing. Invented checkpoints, with slack built in on purpose — a schedule with no buffer fails on its first bad week rather than its tenth. A visible map of the whole project somewhere you’ll actually look, so the connection between today’s paragraph and the eventual book stays perceptible. Most of what people describe as losing motivation in year three is losing the ability to see where they are.
“But learning styles are debunked.”
Correct, and I’m glad you know it, and it’s irrelevant to the argument.
The meshing hypothesis — that people learn better when material is delivered in their preferred modality — has been tested and failed; Pashler and colleagues’ 2008 review is still the reference, and nothing since has rescued it. Visual, auditory and kinesthetic learners are a folk taxonomy. If you find a piece of writing on neurodiversity in education that organises itself around VARK, distrust the rest of it.
But the practices I’m describing don’t come from that literature and don’t need it. They come from something duller and far better supported: working memory is small, it’s roughly the same size in you as in everyone else, cognitive load is a real ceiling on complex reasoning, and the reliable move is to get material out of your head onto a surface where you can manipulate it without holding it. Every technique here is a corollary of that one fact. None of them requires you to have a type.
This includes the writing templates you’re also about to object to. A planning frame — question, claim, evidence, what would falsify it, what the reader must already believe — followed by a drafting frame that converts each item into a paragraph. The objection is that templates produce formulaic prose. They don’t, because they govern planning rather than sentences. What they solve is the actual reason writing sessions stall, which is not how do I say this but what am I doing right now. The blank page is expensive because it forces planning and composing to compete for the same scarce resource. Separate them and it gets cheaper. Use the frame only on the days you can’t start, if you like. That’s most of what it’s for.
“Isn’t it a bit off to take accommodations and use them as productivity tips?”
It would be, done carelessly, and the careless version is worth naming so you can avoid it: we’re all a bit neurodivergent. That sentence flattens a clinical reality into a personality quirk and quietly guts the case for formal accommodation. Don’t say it. Don’t think it.
The defensible claim is narrower. Neurodivergent researchers built explicit external systems for reading, drafting, sequencing and recovering attention because the mismatch between their processing and the academic default showed up in their first semester rather than in year five. They had no option to coast. The rest of us coasted on raw executive function until the work outgrew it, and then had no vocabulary and no infrastructure, because we’d never been made to build any.
What follows is that constraints which are permanent for some people are intermittent for everyone — under deadline, under grief, under illness, at the end of a bad semester, in the third year. Systems designed to work when executive function is unavailable simply work across a wider range of conditions than systems designed to work when it’s abundant. Architecture settled this argument decades ago: the ramp gets used by the delivery driver, and nobody concludes the ramp was unnecessary.
Two obligations come with that. State the debt rather than laundering it. And keep the formal supports fully in place regardless of who else finds the tools useful.
“I don’t have time to build systems.”
You don’t have time not to, but that’s a glib answer, so here’s the real one: the version of this that fails is the weekend redesign.
People read something like this, spend a Saturday constructing an elaborate personal knowledge architecture, use it for eleven days, and abandon it. Then they conclude that systems don’t work for them. What actually happened is that they built a system requiring high executive function to maintain, at a moment when they had it, for use at moments when they won’t. The failure is a fairly exact illustration of the problem the systems are meant to solve.
Pick one thing. Externalise a decision log for a month. Or write the first sentence of tomorrow’s session at the end of today’s, so you’re handing your future self a running start instead of a cold engine — the initiation cost is where most stalled projects actually die, and specifying when and where a task will happen raises the odds of it happening more than any amount of resolve does. Or send a three-line written summary after your next supervision meeting, which eliminates an entire category of failure in which a decision made out loud is held in two imperfect memories and reconstructed differently by each party a fortnight later.
One. Then check whether anything moved.
The reason I lost a year is that I thought scaffolding was a confession of inadequacy rather than a description of how difficult work gets done. It took an embarrassingly long time and a badly-reconstructed argument to change my mind. You can have that conclusion for free, on a Monday, without the year.
If you’re already stuck in the middle of something and the reconstruction has started, that’s harder to unpick alone, and it’s a good deal of what I do in coaching. But most of this you can start by yourself, today, at a cost of about four lines.