Reading Your Own Year

Try it with last week's entry and you'll see why. You read four lines about Tuesday, and nothing happens, because you were there on Tuesday. The note tells you what you already have in your head, only worse, and with fewer details. The obvious conclusion is that old entries aren't worth reading, and most people quietly reach it within a month.

Now read the same kind of note from fourteen months ago. Something different occurs, and it isn't nostalgia. You're reading a person who doesn't know what happens next — and you do. That asymmetry is where the entire payoff of keeping a journal lives, and it's on a delay of about a year.

Why the reading half gets skipped

Partly because nothing in the format asks for it. A notebook has no prompt that says come back. A digital archive shows you a list of dates, which is the least appetising object in software. And partly because the payoff is invisible from the front: on day nine of keeping a journal, the idea that this will be interesting in a year is a promise, and promises lose to sleep.

But the main reason is that people try it too early, get nothing, and file the whole idea under "doesn't work for me." That's a timing error being read as a verdict.

The mechanism is dull and reliable. An entry is only interesting once you've forgotten enough of the surrounding context that the note has something left to tell you. When you still hold the memory, the note is redundant. When the memory has thinned out, the note becomes the memory — and the thinner your recollection, the more work the entry does.

An entry is worth exactly as much as you've forgotten. Which means the best thing you can do with a journal, most of the time, is leave it alone.

The distance threshold

There's no precise number, and anyone offering one is guessing. But the pattern in our own notebooks, and in what readers write to us about theirs, is fairly consistent.

The practical version: don't reread anything from the last three months. Nothing bad will happen if you do, it's just a poor use of the material. Set the whole recent stretch aside, the way you'd let bread cool, and go further back than feels reasonable.

The best single interval, if you want one, is the same week last year. Same season, same light, same weather roughly — so the comparison has something to sit on. What you were worrying about last March next to what you're worrying about this March is a strange and informative thing to see, and neither entry was written with the other in mind.

How to set up a reading session

The one real risk here is volume. People open the archive and start at the beginning, read forty entries in one sitting, and finish flat and vaguely unwell — not because anything in there was terrible, but because forty evenings compressed into twenty minutes is simply too much of a person at once.

1. 1

   Pick a window, not a pile

   One month, or one stretch of a couple of weeks. Choose it by date, before you start reading, so you're not deciding as you go — deciding as you go turns into scrolling, and scrolling turns into an hour.

2. 2

   Read it once, at normal speed

   No annotating, no correcting, no wincing at the phrasing. You're not editing a document; you're visiting a stretch of time. If you find yourself wanting to rewrite an entry, that's the signal to stop for tonight.

3. 3

   Write one line about the window afterwards

   Not a summary. One observation — "kept mentioning the upstairs neighbour," "no entries at all for the last ten days of the month." One line, in tonight's entry, dated today. Over a few years these become the most useful pages you have, because they're the only ones written with any distance.

What's actually in there

Not insight. That's the usual expectation and it's the wrong one — insight is what people go looking for and then feel cheated by not finding.

What's in there is texture, and four kinds of it show up reliably.

Repeated words. You'll find some word you used constantly for six weeks and never since. Not a symbol of anything. Just the vocabulary of a particular period, which is a surprisingly exact way of dating it.

Subjects that never appear. More on this in The Things You Write Around, but you'll notice it here first: a whole month with no mention of the thing that dominated that month. The absence is legible, and only from a distance.

Weather and light. The most underrated content in any journal. "First evening with the window open" does more reconstruction work than a paragraph about how you were feeling, because the weather brings back the room, and the room brings back the rest.

The ordinary Tuesdays. The dramatic entries are the ones you'd have remembered anyway. It's the unremarkable ones — nothing happened, went to the shop, cold again — that give you back what your life actually consisted of, which is mostly Tuesdays.

When the year isn't ready to be read

Some stretches shouldn't be opened yet, and this is the part that usually goes unsaid in pieces about journaling.

You'll be moving along a month at a comfortable distance and hit a page that puts you straight back into a week you had no intention of revisiting tonight. Reading old entries isn't automatically pleasant, and there's no reason to expect it to be. The honest response is to close the archive. Not to push through it, not to sit with it for the sake of some benefit — the writing wasn't a treatment and the reading isn't one either. It's a record, and you're allowed to not want to read your own records this evening.

Skip that period entirely. Skip it for years, if that's what you want. A journal you can only partly read is still doing everything a journal does; the pages you're not opening are not a debt, and there is no version of this where you owe your notebook a full reading.

The reason three lines a night is worth the ninety seconds isn't the three lines. It's that ninety seconds a night, kept up unevenly and with holes in it, eventually produces something you can read a year of your life out of. The writing is the cheap part. The reading is where you get paid.

The evening is warm here.