Using Initials in a Journal Without Losing the Thread

2026-09-04 · Yesterself

Somewhere in most journals there is a person the writer will not name. A manager, an ex, a friend in the middle of something, a family member who might one day pick the notebook up. So the name gets cut down to a letter, or swapped for something invented, or replaced with a job title. It is a decision made in about two seconds, and it quietly sets the terms for how readable your own archive will be in ten years.

Here are the systems people actually use, what each one costs, and how to keep a key that makes any of them survivable.

Why you are coding the name at all

Three different problems hide behind the same habit, and they do not want the same solution.

Someone might read it. A partner, a roommate, a curious teenager, a coworker glancing at your screen. A real concern, and initials are a weak answer to it.

Writing the name makes you polite. This is the more common one and people rarely admit it. Type "Sarah" and some part of you starts writing for Sarah. Type "S" and the sentence loosens. The code is not hiding the person from a reader; it is hiding them from your own manners.

Speed. You write about the same six people constantly and you are tired of spelling them out.

Notice which one is yours, because the fix differs. If it is speed, initials are fine. If it is politeness, a code buys you honesty. If it is a genuine reader risk, no naming scheme solves it, and you should be solving it somewhere else.

The failure every system shares

Journals get read by a stranger eventually, and the stranger is you. Not the you of next week, who remembers everything. The you of six years from now, who opens a page about a hard week and finds three sentences about someone called K and no idea which K.

Everything below is really a question about that reader.

Initials

The default, and the most honest of the options. Initials keep the person a person. You still feel the shape of the real name behind the letter, so you write about someone rather than about a character.

They fail in three specific ways:

• Collisions. You know four people whose name starts with J. Within a year your archive has a J who is a coworker, a J who is your cousin, and a J you went on two dates with.
• Drift. Names change. People marry, transition, go by a middle name, start using a nickname. The letter you chose in 2021 may not match how you think of them now.
• Evaporation. Frequent people stay decodable from context. The person who appears once, in one paragraph, on one bad Tuesday, does not.

The repairs are small. Use two letters instead of one for anyone who might collide. Or disambiguate at the point of collision rather than everywhere: J stays J, and the second one becomes J-from-work, permanently. Whatever you pick, do it the first time the collision happens, not the fifth.

Invented codenames

Aliases, animal names, characters from books, the name you almost gave a cat. These solve collisions completely and they create real distance, which is exactly what some entries need.

They fail in a way initials do not. The alias slowly becomes its own person. Give someone the name Wolf and within a month you are writing scenes about Wolf, with Wolf's motives and Wolf's arc, and the writing turns faintly literary. You stop describing what happened and start narrating. That is fun, and it is not the same document.

The second failure is decoding. An invented name has no thread back to the real one, so nothing helps you recover it. Initials at least narrow the field to one letter. Wolf narrows nothing.

If you want a codename anyway, anchor it. Pick something with a link you would rediscover on your own: a shared first syllable, the street they lived on, the thing they always said. Anchored aliases survive. Arbitrary ones do not.

Role labels

The landlord. My manager. The neighbor with the dog. The one from the conference.

These are the most decodable option on the list, because the label carries its own context. Six years later, "the landlord" is still legible in a way that "R" is not. They are good for people who appear rarely, and for people whose only relevance to you is the role.

Two costs. First, roles expire. The landlord stops being your landlord. "My manager" covers four different human beings across five years, and your archive silently merges them. Second, roles flatten. It is easier to be unfair to a role than to a name, and easier to write the same complaint about it forever.

The fix is a date or a letter: "my manager (R)" or "the landlord, 2023." One extra character, and the entry stays true.

The hybrid that usually wins

Most people who have kept a journal for years land in the same place without planning to.

• Real first names for anyone you have no reason to code.
• Initials for the handful of people who need distance, with two letters where collisions exist.
• Role labels for the walk-on parts, stamped with a year.
• Invented names only where the real name is too loud to write, and anchored to something when you use them.

A mixed system is fine. What matters is that one person keeps one code. Consistency inside a single name beats consistency across the scheme.

Keeping a key

The key is what makes all of this work, and it costs about a minute a year.

Keep one, in one place, with the journal itself rather than somewhere clever. A line looks like this: the code, who it is, when you started using it, one detail you would recognize them by. That last item is the part people skip and the part that does the work. "JM — Jenny from the pottery class, started 2024, moved to Denver" will still decode in a decade. "JM — Jenny" will not.

Two rules keep the key honest. Add the entry the moment you invent the code, because you will not do it later. And do not delete retired codes when a person leaves your life; mark them retired and leave them. The whole point is reading old pages.

The obvious objection: if you keep a key, what is the coding for? Mostly it is for you, at the moment of writing, when the letter loosens the sentence. It also blocks casual reading, which is the realistic risk in most homes. It is not a cipher, and it will not hold up against someone determined.

Repairing an archive you already coded

If you are sitting on years of unlabeled letters, do not rewrite the entries. Annotate.

Start the key today with everyone you currently use, then work backward at your own pace. Read old entries in any order and add a line to the key each time you decode something. Where a code has genuinely gone, write "unknown" next to it and move on; that is information too. The goal is not a complete index. It is that the codes you still remember survive the years when you will not.

What a naming scheme is not

Initials are not security. Neither are aliases. If your real worry is that someone will read what you write, the naming scheme is a comfort rather than a lock, and it is worth reading about what actually keeps a journal private instead of asking a single letter to do it.

This matters more in a journal that hands your past back to you. Yesterself mails you one letter each morning drawn from your own earlier writing, so entries you wrote at twenty-eight arrive when you are thirty-four, codes and all. It is also somewhere a key naturally lives: any sentence can be postmarked to a future date, so a key line written today can be scheduled to arrive years from now, when you have forgotten it and need it.

Pick your system tonight. Write the four or five lines of the key. Then go back to the actual entry, which was always the point.

Common questions

Is using initials in a journal enough to keep entries private?

No. Initials mostly stop a casual glance, and anyone who knows your life can usually guess who a single letter refers to. If the real concern is someone reading your journal, handle that with encryption and device security rather than with a naming scheme.

How do I stop forgetting who the initials referred to?

Keep a key in one place, and write four things on each line: the code, the person, the year you started using it, and one detail you would recognize them by. The detail is what makes an old entry decodable a decade later.

Are codenames better than initials?

They solve collisions and create more distance, but they are harder to decode later and they tempt you into writing about a character instead of a person. If you use one, anchor it to something real, like a shared syllable or the street they lived on.

What should I do about years of old entries with codes I cannot decode?

Do not rewrite the entries. Start a key today with the codes you still use, then work backward when you feel like it, adding a line each time you decode something. Where a code is genuinely gone, write unknown beside it and move on.

Should I use one naming system for the whole journal?

You do not have to. Most long-term journals end up mixed: real names for most people, initials for a few, role labels for walk-on parts. What matters is that one person keeps one code, consistently, rather than that the whole scheme matches.