This is how it starts. My unconscious throws up a question. The precise nature of this question is shadowy—it may remain unknown to me for ages, only gradually revealing itself; sometimes it never becomes explicit. But it will always present as some kind of idea—a scene, a story, a voice, a paragraph, a sentence, an image—and in this, there’ll be something I want to write about.

If I’m lucky, how I initially respond will interest me enough that I’ll get hooked in fairly quickly. Generally, my process of writing short fiction is more straightforward than novels. I usually write the first draft in longhand—something I started in the noughties, after working with Julia Cameron’s morning pages exercise from The Artist’s Way. This stops me getting seduced into making things too neat too quickly. With pen and paper, I’m more likely to write fast, without planning; the results often feel more authentic, and at the end of a rewrite I can use them as a benchmark: Is this really what I was after? After the first draft of a story, I do a quick second draft—something I find impossible with a novel. Here, I read through for sense and pace, and add text, again in longhand but on separate sheets, code-naming the inserts and marking where they’ll go in with asterisks. At this stage, I might cross out a word or slightly rephrase something, but I try not to change much. After leaving the expanded version for a day, or sometimes months, I type it up, in some cases adding more bits as I go.

For the bulk of the stories in my collection Shift, this was how things went at the early stages: fast, and relatively painless. However, the writing didn’t end there. Most of the stories I tweaked over the years, often for submissions or competitions. During the 10 months between deciding to put the collection together for New Island and publication this April, I worked on all of them again, adjusting them to a greater or lesser extent. My focus was to get the ensemble to work together as much as ensuring the stories were effective individually. At times, it got very gnarly.

If writing is a conversation between parts of the self about problems which, because they live in the unconscious, are unknown by the conscious self, then it’s not surprising that misunderstandings will arise along the way. Once you’re past a certain skill level, it’s these that manifest as ‘bad writing’. Yukky sentences, purple prose, tinny dialogue, stupid characters, ridiculous plot, unconvincing settings, incomprehensible action—you, like me, will probably have your own bêtes noires. Every writer has to find a way through the crap: you can try writing over it, or skipping past it. You can stop writing, in the hope that the Writing Fairy will come along in the night, wave its wand and make everything better. Or, if you’re a compulsive fixer like me, you’ll try to sort things out on a word level.

Over the years, I’ve had chats with lots of my peers about the crippling experience of doing the problem-page Hokey Cokey: working the same text over and over, taking out a word, putting another one in, shaking them all about. I’m not talking about doing this at copy-edit stage, when the piece is nearly ready to go. I’m talking about doing this at the stage when you still don’t want people to read the thing because it’s not Right, it’s not Finished, and it May Never Be. I hate it when I find myself doing the Hokey-Cokey because I know, fundamentally, there’s some bigger problem lurking underneath the text. Yes, the section I’m fiddling with is horrible. But in truth, it’s only horrible because—

And there’s the rub. Because. Because Why? What is the problem? How do you identify it? And how on earth can you solve it?

John Boyne said once he reads through a print-out of every draft of each work-in-progress before he makes any corrections. That makes sense: it’s what we write for, to be read. But it can be hard to read your own work in the right way, attentively enough to see what’s really the issue in a problem text. If I can, I try to read a tricky draft as if it’s someone else’s: closely, slowly, looking out for the good stuff as well as the crap. I usually read off hard copy and mark things in the margins. I try to be instinctive, avoid getting fixated on individual phrases. I try to be honest. Put question marks anywhere I’m confused, write BORING beside places where I’m losing interest.

Sometimes the reading will immediately throw up ideas for solutions which feel interesting and I’ll start trying these out. Other times the reading will end up with me doing a scrawled-up, crossed-out version of the Hokey-Cokey on the page. That’s a sign that I, and the text, aren’t ready yet for a read, because the problem is in too deep under the words. Or I’ll know in advance there’s too much crap in my head for me to read with any kind of clarity. In those cases, instead of tinkering around with the text, I go straight into analysis.

My main analytical tool, after notes on the page, is journaling, a technique which I (again) adapted from Julia Cameron’s morning pages. Orna Ross offers something similar called F-R-E-E Writing, and I’m sure there are lots of other related techniques around. The main rules that work for me are: Write longhand, don’t plan, don’t edit, just let it all come out. I usually start with either a rant (I hate this stupid story), a moan (Oh god, I’m so tired from writing) or a question. Rants and moans are great for venting and can sometimes help me see problems in a new light. But more often than not, it’s the questions that I find most useful:

  • Why is this sentence/para/section horrible?
  • What’s wrong with this story?
  • Why do I hate it?
  • What do I hate most?
  • What isn’t working in it?

I’ll recognise a good question because it snags on my attention and gets me writing freely in response. In my answers, it can take a few goes to get the rubbish out of the way. I hate this story because…

  • I’m a terrible writer.
  • It’s a terrible story.
  • I’m tired.
  • I should be writing something else.

But sooner or later I’ll start getting richer, more specific, information. This story is horrible because…

  • I hate my character/narrator.
  • The whole thing feels false.
  • That passage is super-uninteresting.
  • The action in that sequence is boring.
  • The character doing that thing makes me feel morally uneasy
  • I’ve no idea what’s happening in that scene.
  • I haven’t a clue what happened the moment just before the scene started.
  • I don’t know who’s actually telling the story.

Now the deeper issues, the ones beneath the words, are starting to become apparent. Nearly always, I’ll realise I’m in trouble because I don’t know something. Sometimes it’s a fact—if I’m writing about a woman in the 1970s with cancer but I don’t know anything about 1970s cancer treatment, this might be why ‘the whole thing feels false’. It can be irritating to admit I’ve a shortfall of necessary information. But once I face up to it, it’s straightforward to address. Do research.

In other cases, the gap in my knowledge lies in the text itself. There’s a question I haven’t asked about a scene. An aspect of character, setting, plot I’ve assumed. A crucial piece of action I haven’t bothered writing down, or about, because I was in such a rush to get the book written. These issues can take longer to work through than a factual gap, because they’re pushing up against my own limitations, the porous walls protecting my unconscious from the day-to-day front part of my brain. I’ll need to ask more questions to winkle out the underlying roots and the answers often prompt more questions, which can feel neurotic—and, at times, endless.

I’ve had to learn how to frame questions so they’re useful for analysis. Why questions (e.g., Why don’t I know who’s telling the story?) can easily generate crap: Because I’m terrible, stupid, a bad writer/this is a terrible story/Stop asking, I just don’t know etc. But if I ask a What question (e.g., What’s making me not know who the storyteller is?—or, even more precisely, What in the story is stopping me see who the storyteller is?) I’ll, eventually find a useful answer.

  • I can’t see who’s telling the story because too many people are.

At which point I go, aaah.

When I mentor other writers who are struggling with problems, I often encourage them not to panic, or go for the most immediate fix. Instead I suggest they welcome the problem, take their time to explore possible actions to address it and sit with these before making a choice. If a single clear way forward through a problem doesn’t suggest itself immediately, I’ll try brainstorming. I’ll ask myself something like What can I do about this?, then jot down as many answers as I can. For the problem above, these might include:

  1. Choose one character to tell the story.
  2. Find out who the too many different story-tellers in the story are.
  3. See if it’s possible to have many of them—?? but not too many—tell the story?

It won’t take much time to see that Option 1 is a definite fix, Option 2 is work that might help find a fix but isn’t a fix in itself, and Option 3 is a potential fix—but one which the multitude of scribbled question marks suggest isn’t straightforward and needs to be exploded more. As soon as I’ve jotted down the list, I’ll have an immediate feel, a physical twinge in my gut, for which approach seems the most interesting. With that, I’ll usually experience another sensation—a lead weight on my shoulders or in my stomach—about which is going to take the most work. Generally most interesting coincides with most work, which is both a bit of a pain and an indication that this is the option I’ll probably go for. If I’m still unsure, I’ll do, again by hand, a sort of SWOT analysis using a matrix.

1. Choose one characterFix
Easy for reader to follow
Not so much work
?does this story need to be complex
?might lose a layer
?what about other characters’ viewpoints
?lose complexity/subtlety
?does this story need to be simple/simplifed
2. Find out who storytellers are Prob has to be done anyway
Will give clarity
Not a fix
Bit tedious
Oh god have to do a read-through/analysis/ line-edit AGAIN on the story
Will I hate this story by the end of it?
3. Have many story-tellers?
…but not too many…
Could fit themes/tone of story??
Could resonate much more
If balance is just right could be a really nice fix
Possibly confusing to reader
Possibly irritating
People might lose interest
Unnecessarily complex
Just playing with form for form’s sake??
How do I get the right balance?
How many is NOT too many?
Loads of work. Aargh
???how to make it work (no idea)
May fail

At this point, I’m often tempted to try a quick fix, but it’s always more valuable to do the dogwork first—here, Option 2, analysing the possible story-tellers/POV’s/voices in the story. During this, or perhaps after more brainstorming, depending on how connected I am to the original ideas behind the story (the longhand draft), a flash of insight might come:

4. What if the narrative voice presents as single storyteller but actually it’s many?


5. What if the narrative voice presents as many storytellers/perspectives but actually it’s only one?

And then, if I’m lucky, I’ll have another aha moment.

This process may sound scientific, especially with the matrices and their neat little boxes, but it isn’t. It’s messy, complicated, chaotic and full of doubt. I’ll usually work on a succession of drafts before reaching a resolution. It can take weeks, months sometimes, even on a single story, especially if I’ve decided to explore something intricate on a tonal or narrative level that I’m not sure will work, or that I’m concerned will just be a formal experiment. Often I’ll need time away from the text for the aha to come, like the chemist Kekulé famously did, drifting off to sleep on an omnibus and suddenly ‘seeing’ the atomic gambolling of chemical structure he’d previously laboured to understand. An early indication that I’ve happened on something effective is if I see that there are places in the existing text that easily, almost naturally, give themselves to the new option. It’s like the text itself wants the change to happen and is bending to accommodate it.

As I’ve said earlier, I usually go for the most complex option, which is usually also the one which requires the most work. This is partly out of curiosity and devilment and a certain wilful stubbornness in my nature: who says it can’t be done? But there’s also something less capricious and more essential at work. Maybe the trickiest, most work-intensive resolution is the one which hovers most tantalisingly on the line between the known and the unknown, the one which might, if I’m careful enough, reflect most accurately the shadowy nature of the underlying force which has triggered the writing process. Having said that, there are times when I go for the simplest option, and those, too, can throw up issues as soon as I begin exploring them.

When I began this essay, I’d planned to write about something specific, like point of view, and the choices I made around it to get Shift working as a collection. Every time I started going into the details, I found myself doing the Hokey-Cokey. Eventually I had to admit I had a problem. It didn’t take too much digging to recognise that this was because deep down, I didn’t want to lift the lid so soon on Shift’s engine, talk through the specific acts of deconstruction, reconstruction, rebalance and composition that went into its making. What I’m prepared to say is that all the stories in Shift are different to their first drafts, the ones that got written easily. With some it’s only a paragraph that’s changed; others have kept their original plots but have altered significantly in their tone and underlying meaning.

I did play a lot with perspective/point of view: adjusting, flipping, expanding, contracting both person and tense. Sometimes I did this after journaling or SWOT analysis, sometimes on instinct, sometimes in a flurry of panicky head-fixing. Understanding what choices I was making around who was telling the story—and where, and when—helped a lot in clarifying what sort of story I was telling, or wanted to tell, or was afraid of telling.

So what have I learnt? Firstly, that a collection is different to a novel, and an individual story; it needs to be understood on its own terms, and figuring that out, and what those terms are, can take time. Secondly, the perpetual lesson of writing: that it’s always easy, and always hard. And thirdly, another reminder: that there’s no guarantees the work will land with readers the way I think it might, but as long as I honour whatever’s provoking the writing, that mysterious prompt emanating from the layers of my unconscious, something will come good—even if it’s just me feeling I’ve happened on some new sort of sense, somewhere.