How to Memorize the Braille Alphabet

You do not have to memorize 26 symbols. The braille alphabet is ten cells, a to j, and three rules that produce the other sixteen. What makes them stick is not rereading the chart but producing each cell from memory, in short sessions, with the answer checked straight away.

This article does not explain what to learn (that is in how to learn braille) but how it sticks: which technique, in what order, and which test shows that it has stuck. We give no timeline: a number would contradict our own article on how long it takes, and the value is in the technique.

Key points

• The real memory load is ten cells, not 26. Ten cells and three rules cover the whole alphabet. We computed it over the same braille tables our translator uses.
• Testing yourself means producing, not recognizing. In the reference meta-analysis on the testing effect, cued recall tests (you get the letter, you produce the cell) scored g = 0.61, against g = 0.29 for recognition (Rowland, 2014, 159 effect sizes from 61 studies).
• Without immediate feedback, self-testing is worth half. With feedback, g = 0.73; without, g = 0.39. And in the worst case (testing yourself on something you cannot yet recall, with nobody correcting you), the effect falls to g = 0.03: zero.
• Rereading is the lowest-rated technique of the ten reviewed by Dunlosky and colleagues (2013), while practice testing and spacing are the only two rated "high utility".
• Two questions place any cell in its series: is dot 3 raised? Is dot 6 raised? The only letter with dot 6 and without dot 3 is w.
• The third series does not stop at z. The same rule applied to f to j gives the grade 2 signs for and, for, of, the and with.

Why brute memorization fails

Almost everyone's default method is to open the braille alphabet chart and look at it until it "sticks". It fails for two reasons, and both are measured.

First: looking at the chart is not studying, it is recognizing. With the answer in view, the cell for q always looks familiar. That familiarity is easily mistaken for knowing it, and it vanishes as soon as you take the chart away. In Dunlosky and colleagues' review of ten study techniques, rereading was rated low utility (along with highlighting), and the only two rated high utility were practice testing and distributed practice.

Second: retrieval from memory is what consolidates. The reference meta-analysis pooled 159 effect sizes from 61 studies and found that testing beats restudying with an average effect of g = 0.50 (confidence interval 0.42 to 0.58). Its moderators are exactly what a braille learner needs:

All figures are from Rowland (2014), Psychological Bulletin.

That gives the three method decisions in this article: produce instead of recognize, correct on the spot, and do not test yourself cold on a block you have not studied yet.

The other half, spacing practice, is covered with its meta-analyses in how long it takes to learn braille: there you will find the figures on distributed practice and, above all, the nuance that for skills (and braille ends up being one) very long gaps reduce the benefit. In practice: short sessions, close together.

What you really have to memorize: ten cells

Here is the count that shrinks the task, computed over the same braille tables that run our translator and our alphabet page:

Ten cells by heart, three rules. That is the real size of the problem. The first two rules and their logic are explained in how to learn braille; here is what is rarely said.

Rule 2 keeps going after z

Most guides present the third series as five letters: u, v, x, y, z = a, b, c, d, e plus dots 3 and 6. True, and incomplete. The same rule applied to the next five letters of the first series produces five of the most common signs in English braille, the one-cell contractions of grade 2:

Checked with liblouis, the standard braille translator. So the gap after z is not empty: it is where English braille keeps its five strongest contractions. You do not need them to learn the alphabet, but when you move on to braille contractions, you will already know where they come from.

w is j plus dot 6

w is always presented as the exception to memorize on its own, because Louis Braille's French alphabet barely used it. But there is a handle: w ⠺ is j ⠚ plus dot 6. It is a memory aid, not a rule of the system, and it turns the one loose letter into a variation of one you already know.

The order of work, session by session

This is teaching judgment, not a measured result: no trial compares study sequences for braille, and we say so below. What is supported by data is the shape of each session (produce, correct, stop early), not the order of the blocks.

There is no deadline. There is no session seven by which you must have the alphabet. What decides when you move on is the test in the next section, not the calendar.

The blocks, in order

1. The dot numbers. Five minutes. Everything else uses them as vocabulary.
2. a to e. Five cells. The smallest block, on purpose: it is where the habit of producing gets installed.
3. f to j. Now there are ten, and those ten are practically all the raw memory you will spend.
4. The dot 3 rule. Do not memorize k to t: work out each one and check. The exercise is not remembering p, it is applying the rule to f quickly.
5. The dots 3 and 6 rule, with u v x y z.
6. w, as j plus dot 6.
7. The capital and numeric indicators, and texts that mix them. They are covered in common mistakes and in braille numbers.

A new block does not replace the earlier ones: every session mixes old and new, and the old part is done without the chart.

The shape of a session

The exact minutes are our proposal, not a result. What does come from the data is that the session starts by producing, that correction is available on the spot, and that you do not test yourself on what you have not studied yet.

And the rule above all others: better not to skip a day than to do a long session to make up for it.

How to know you "know it"

This is where almost everyone fools themselves, not from dishonesty but because of how familiarity feels.

Recognizing is not knowing. If you look at the cell for q and think "yes, that's q", that is recognition, the weakest test there is: g = 0.29 against 0.61 for producing the answer. The same goes for an app that gives you four options: choosing from four is a different, easier task than producing the answer from scratch.

Four tests, from weakest to most honest

The standard we propose: you know a block when, the next day, without reviewing first, in random order and in both directions (letter → cell and cell → letter), it comes without stopping to work it out. The exact threshold is ours, not from a study; what does come from the data is that the advantage of testing grows with the delay (g = 0.69 when the final test is more than a day later, against 0.41 within the same day). That is why the real test is tomorrow's, not the one five minutes from now.

One detail that saves arguments: "stopping to work it out" is fine while learning and wrong as an end state. Applying the dot 3 rule to get to p is exactly what to do at first. But a fluent reader does not calculate: the cell comes straight. Repeated practice is what turns working it out into instant recognition.

A two-question check for what you wrote

This check is ours, computed over the whole alphabet. It tells you, without looking at the chart, whether the cell you just produced can even be the letter you meant. Two questions: is dot 3 raised? and is dot 6 raised?

Read it as a self-correction: if you wrote q and there is a dot 6 in your cell, it is wrong, because q lives in the "yes / no" row. And if you find a cell with dot 6 and no dot 3, you know it is w before thinking about anything else.

A second check of the same kind, verified on all 26 letters: no letter has an empty left column. If what you wrote only has dots 4, 5 or 6, it is not a letter, whatever you meant. The reverse does happen (a, b, k, l have nothing in the right column), so the check only works one way.

The hard spots

If your mistakes always land on the same letters, you are not short of memory: you need to work on one specific group. That is diagnosed letter by letter in common mistakes when learning braille; here is only what to do about it within the method.

• The four mirror pairs (d/f, e/i, h/j, r/w) are not fixed by reviewing the whole alphabet: take them out of the general deck and practise them as pairs, in both directions, until the rule that separates them comes without counting dots.
• w stops being an exception once you use j plus dot 6, and the two-question check isolates it: it is the only letter with dot 6 and no dot 3.
• If your mistakes are random and do not cluster, then it is memory, and the fix is more frequent sessions, not longer ones.

From seeing to touch

It needs saying here, in an article about memorizing, because this is exactly where expectations break: memorizing the code is not reading it with your fingers. Everything above (producing the cell, the check, the three rules) trains the first layer, the information. The second is perception, and studying harder does not speed it up.

If you can see and your goal is to read by touch, what you learned with your eyes does not transfer by itself: you need to train without looking, on real raised dots, from day one. How, and when you do not need to, is in learning braille as a sighted person. The orders of magnitude for each layer are in how long it takes to learn braille.

And one practical consequence for the method: the two-question check is a tool for analysis, not for reading. It helps you get unstuck while memorizing. A fluent reader does not check for dot 6; they perceive the whole cell at once.

Memory aids: the ones we suggest and the one we dropped

Braille memory aids that circulate are usually sentences for remembering the order of the letters, and few survive checking. Our rule: if a memory aid fails in a single case, either the case is stated or the aid is dropped.

The ones we suggest, all checked against the braille tables:

One curiosity, not a rule. The sign for the, ⠮, is the exact negative of e, ⠑: it has exactly the dots e is missing. Among the letters themselves, three pairs work the same way: i ⠊ and z ⠵, j ⠚ and u ⠥, k ⠅ and w ⠺. Nice to notice, and handy as a hook for those six letters, but it is a coincidence of the layout, not a rule you can build on: the other twenty letters have no letter as their negative.

What we could not confirm

And one that deserves saying plainly: we do not know whether this method beats another. We know retrieval practice beats rereading in general, and that the alphabet reduces to ten cells. That combining the two is the fastest way to memorize braille is a reasonable hypothesis, not a result.

A method in seven steps

1. Learn the dot numbers before any letter.
2. Memorize only ten cells, a to j, in two blocks of five.
3. Produce, do not recognize: cover the chart and draw the cell you are asked for.
4. Correct on the spot, looking at the answer right after a miss.
5. Work out the rest with the three rules instead of memorizing it.
6. Test yourself the next day, cold, in both directions. That is the test that counts.
7. Short sessions, every day, rather than long, spaced-out ones.

To go through it as a calendar, see braille for beginners: your first 7 days. The full chart is on the braille alphabet page, and you can check any transcription in the translator. What no screen will do is train your touch.

Frequently asked questions

What is the best way to memorize the braille alphabet?

Retrieve it from memory instead of rereading it. You are given the letter and draw the cell, without the chart in front of you, and check straight afterwards. In research on the testing effect, that format (cued recall with feedback) works best: g = 0.61 against g = 0.29 for recognition.

How many braille letters do you really need to memorize?

Ten cells: the first series, a to j. The other sixteen come from three rules: add dot 3 for k to t, add dots 3 and 6 for u v x y z, and w is j plus dot 6.

Do flashcards work for learning braille?

Yes, if the card asks you to produce the cell rather than pick it from a list, and if you can flip it and see the answer at once. Without immediate feedback, the benefit of self-testing drops by about half, and if you test yourself on material you do not know yet, it disappears.

How do I know I know the braille alphabet?

When, the next day, without reviewing first, in random order and in both directions (letter to cell and cell to letter), the cells come without applying a rule. If you have to recite the series from a to reach p, not yet.

Does memorizing the alphabet help me read braille with my fingers?

Not by itself. Memorizing the code is a layer of information; reading by touch is perception, trained separately, without looking and on real raised dots. They are two different skills and one does not carry the other.

How long does it take to memorize?

We do not give a timeline: it depends so much on the person that any specific number would be made up. What we do say is that short daily sessions work better than long spaced-out ones, and that the signal to move on is the self-test, not the calendar. Details are in how long it takes to learn braille.