Max’s 5 Types of Writing — A guide for PhD Students

To help students in our lab write when it comes to their critical deadlines, I’ve regularly organised “Writing Tuesdays” or [insert whatever day is free in my teaching schedule that term], to get together and socially write. This is not collaborative writing, but social writing. Over time, when people tell me they are struggling to write, or struggling to make good progress, I ask them if they know about ‘different types of writing’ . My observation is that people are often trying to do all these at once, and this is why they are struggling. Like types of reading (lets assume you know about deep vs shallow reading), there is different types of writing, and it is much more productive to do one type for each writing session, than try to do them all at once.

1. Text dumping

Text dumping is kind of permission/freedom just to write, unrestricted, unedited, unpolished — to get whatever is in your head out of your brain and on to paper. It doesn't matter how good or garbage it is. My colleague Jeremie calls this type of writing ‘Barfing’ for a good reason — just get it out. One of the biggest barriers I see people facing to producing large amounts of text when it comes to writing their thesis, is that they interrupt progress by doing the four other types of writing (below) at the same time. Let’s get to those in a bit. What text dumping is doing is ‘Externalising’ all your writing, from inside of your head on to paper, where you and everyone else can improve it. You, and other people, cannot improve the writing and make progress while it is in your head. Write whichever bit you feel is motivated to write right now, and dont look back (yet). Just keep writing whatever it is. You can reflect on it and edit it later. There is no consequence to unrestricted brain dumping (it’s a draft!), and the opportunity to improve it later only comes if you can get it into a document. Ultimately, this is a generative type of writing — it creates content.

2. Editing

Editing is completely different to writing, and people’s most common mistake is to write and edit in parallel, everything they are writing, as they write it. In truth, however, it is a lot easier to edit text when you know how it relates to the rest of the document… which you can’t do if it doesn't exist. So you cannot do effective #2 editing until you finish #1 text-dumping-writing. In my opinion, this creates a big issue for PhD students, who often feel like they are ‘supposed to finish a chapter’. Finish is a state that can never be fully completely reached. So.. finished how? The best way to finish a chapter, is to make sure it lines up to every other chapter. So you need to have written every other chapter, in terms of text-dumping at least, to finish a chapter in terms of editing. You can’t easily finish your introduction, until you know the full scope of the thesis that it is introducing. You can’t easily finish a discussion, until you have written the chapters they discuss. But you do not need to fully finish all of those chapters to understand what they will all say. Editing is a refinement type of writing — it typically reduces or improves content, rather than generates large amounts of new content. Like text-dumping, dedicate a session to editing something if that is your next task. Today let’s edit the introduction.

3. Reference Finding

Reference finding is a task of its own. What disrupts writing more than writing a sentence and then going ‘hmm — there was a good reference for that’, and then spending 30mins quick-reading several papers to decide what is best to cite, and whether you cited it right. This interrupts progress, and might be something you edit out or change later anyway with your #2 editing. My top tip when doing #1 text-dumping is just to write [REF] every time you think something needs a reference, and [REF wilson design] when you know you want to later get a known reference later. Don’t find it as you are text-dumping; do Reference Finding at a dedicated time, not too late, in the process. This task involves completely different actions and thought processes to text-dumping and editing, and involves searching, copying, pasting, inserting, etc. Dedicate periodic times to do this, and to fill in a bunch of [REF]s. I try to do this after at least some editing-based writing (#2). This is a refinement type of activity — it doesn't generate new content, but improves what you have.

4. Artefact Generation

One thing you often need is a good figure, or good detail in a nice table. This often involves finding, processing, and organising data. Perhaps configuring created figures in something like excel, or writing more code to improve how the figure looks in something like python or R or matlab. Again — is this writing? Not really. Do people try and do this at the same time as writing? Yes. And it slows you down. It is a different task, and you should save it for a time when you are like ‘right, let’s spend the next two hours making good figures for my paper’. This sort of applies to a range of things, where figures and tables are the most obvious examples, but any kind of appendix or resource or thing to go in, on, or with the paper — is not the same as #1 text-dumping nor #2 editing. This is a generative activity — it’s about constructing new content for your document, but different content to writing.

5. Document Polishing

Often as a finishing task, but separate to text #2 editing, #3 referencing finding, and #4 artefact generation, is document polishing. This is really the job of making sure what you have done conforms to the constraints that you have — does it have to be in a specific template? does it fit in any given limits? This is different because its of separate concern to: is the text good? Here you are deciding if you should or need to cut things. Should you move things to different pages, or resize images? Is the document meeting all the requirements? During this task, people focus more on structure, and if they touch text at all it tends to be structural changes, like joining or splitting paragraphs, adding subheadings, or trimming a word out of a sentence so it still means the same but just fits into the page. I recognise that in a thesis, for example, there may not be page limits or length limits that you are hitting, but there are still finisher-completer tasks that need to be done to make sure the ‘document is finished’ rather than just the writing. This is really important to do separately because there’s no point in trying to make something fit a document’s requirements if the text isn’t feeling finished, and you or someone else is about to add a subsection to the discussion. It is much easier to make document polishing decisions when you know there isn’t going to be anything new coming down the line.

Summary

I hope some people find thinking about writing like this useful. In our writing club, we try to commit to doing a type of writing for that session. Usually text-dumping, because editing and reference finding and artefact generation are all often easier to do in a mindless way in a gap between meetings etc. But good generative writing, making good progress on a paper or a thesis chapter, comes from bigger periods of focused uninterrupted free-wheeling text-dumping, so you can improve it all later.

For more reading, I also recommend Thimbleby’s Write Now! book chapter, which inspired me to do a lot of low-level unimportant writing in my PhD, before having to do the hard and important writing later.

[1] Dix, A. (2008). Externalisation–how writing changes thinking. Interfaces, 76(Autumn 2008), 18–19.