How not to write badly (4): Do not sprawl

I'm cheating a bit because this is more of a "how to write" than how not to but I think it just about fits.

I'm going to share my method to fix two problems writers often have: loose structure and writing too much. I'm aiming this at assignment writing but the same principles actually apply to writing fiction, playwriting, any form of writing that requires structure.

What do I mean by loose structure? It's easiest to say what is desirable and you can think about how you might stray from this. A great assignment is driven by ideas. It pops from idea to idea without losing its drive. You can readily follow its thesis and the writing has thrust.

The same principle applies to writing, say, a novel. You should have narrative drive. It's what makes a reader want to keep turning the page. If you waffle, you start to lose the reader.

What do I mean by writing too much? Well, for assignments we generally mean you can't stick to the word count. And for a novel, you write too much filler. Or not enough words. Writing too little is just the same problem from the other way round. Assignments with decent word counts can seem daunting. If you ask someone to write 3000 words on a particular thing, they might flinch from the job, procrastinate, see it as a mountain that's hard to climb. How to fix that?

Here's my method. I use a tree.

The first thing to do is to think about what your assignment says in the simplest of terms. You are going to explain it to a child. In fact, your goal in most formal writing is to write at a level that would be comprehensible to a bright 10-year-old. If my Zenita cannot understand what you're saying, you are writing badly. You aren't Heidegger. You are a student trying to get a good grade. Teachers don't want to wade through shit any more than the rest of us do.

So you need the star idea. I'm going to help you grasp this by using an example from an essay I might have written for my degree.

STAR: Generative semantics was a competitor to TG grammar that was discarded because it could not resolve some utterance types

So I have done a literature review on generative semantics, an approach that briefly flamed and then died away. My thesis is that it died away because it could not explain some utterances (sentences, basically) in English. This idea will drive my whole assignment.

I did this approach for a friend. It took a while to elicit what the star idea was because it's probably easier for me to summarise concepts in plain English than it is for her. That's because my brain works that way. But you can do it quite easily. Just remember you are trying to explain it to a child. Write keywords and think about how they link together. Her assignment is a literature review on hyperlexia. She tells me that it was once considered a savant skill and is now considered a learning disability because it shows a cognitive deficit.

So:

STAR: Hyperlexia used to be considered a savant skill but is now seen as a cognitive deficit.

Now, notice straight away that both star ideas have concepts that a child can't understand. The words are meaningless. These are things we must explain.


And what if you were writing a novel? STAR: Young man meets a convict and it changes his life.

These ideas provide drive for your work. You will avoid the common mistake of "narrating" an assignment, where you begin then start to tell the reader everything you know. "And then... And then... And then..." This lacks the thrust you want your writing to have and also is very hard to structure properly (because it's unstructured).

So I have my star idea. Next I create five or six heads. Doesn't have to be any particular number but generally five or six is about right. These can and usually should take the form of questions. They are the questions a child might ask you when you express the star idea.

HEAD: What was generative semantics?
HEAD: How did it differ from TG grammar?
HEAD: Which utterance types did it fail to resolve?
HEAD: How did TG grammar prove superior?
HEAD: Was it correct to discard generative semantics?

So now you answer the questions set out in the heads. Generally, a bullet list of six points is enough here.

SUB: A semantics-based approach
SUB: contrasting with syntax-based approaches
SUB: developed by whoever developed it
SUB: which abandoned the tree structure and strict rules-based approach of TG because
SUB: it aimed to more closely mirror actual cognitive processes
SUB: which was necessary because the previous approach was thought over formal

So for my essay I would have 30 subheads. I have 3000 words. 200 for the intro and conclusion, leaves 2800. Roughly 90 per subhead and a spare 100 to use as my bank.

Now I write each subhead in turn. I have 90 words and should try to write that many for each one. If I write too many, I must borrow. Because I have a bank, I can borrow from that first. If I don't, I must borrow from another subhead. You probably have a feel for which subheads can stand to be a bit light, but if your structure is tight enough, you should avoid too much borrowing. Because you have to pay back. If you had a limit of 90 words for each sub and write 100 for one and cannot cut, you must pay it back from another by only writing 80.

Now you have made your 3000-word assignment into 30 90-word assignments. A much easier task, right?

And you can readily answer another question: how many references should I include? You should include at least one for each subhead. These ideas don't belong to you. You're simply bringing them together. So provide support for each one with a citation. You don't need more than one but you should have at least one.

Can you see how this would work for a novel?

The star idea is the hook, the story in a sentence. The heads are what we call crisis points: big turning points in the action that drive the story. The subheads are action points: the things that happen, the "scenes" if you like. If you were aiming at 80,000 words and had 36 action points, that's about 2300 words for each action point, or about sevenish pages. You can probably break the action points down a bit further by analysing them a bit. Novels aren't quite like assignments after all. The point of this method is that we analyse each idea down to its atoms: the ideas that cannot be analysed further. In an assignment, that is usually the ideas that support the five or six elements of the central thesis. In a novel, we can conceive of plot as: concept > crisis points > action points > subscenes that build action points.

posted by Dr Zen @ 10:28 am