Y and En in French: The Rules and the Real Data

Y replaces à + a place or a thing; en replaces de + a thing, or a quantity. That is the whole rule, it takes five minutes, and almost nobody can use either one fluently after five years.

The gap between how simple these two words are and how long they take is the interesting part - so we measured it. Using Lexique 3.83, the French frequency database, and a corpus of 928,000 words of French prose, we counted not just how often y and en appear but what those appearances actually are. The answer explains the difficulty better than any rule does.

What do y and en actually replace?

A preposition plus its object. Which pronoun you need is decided by the verb, not by the meaning.

Two constraints do most of the damage.

People are excluded. Y and en replace things and places, not human beings. Je pense à Marie becomes je pense à elle, never j’y pense; je parle de mon frère becomes je parle de lui. Spoken French blurs this at the edges, but the written standard does not.

En is compulsory where English has nothing at all. English drops the object once it is understood - I have two, I’d like some. French cannot: j’en ai deux, j’en voudrais. This is where learners fail, and they fail by saying nothing rather than by saying something wrong, which means the mistake is almost never corrected by anyone.

How common are y and en really?

On the surface, extremely. En is the 11th commonest word in written French and the 17th in speech.

That ranking says these should be among the first things you master by exposure alone. They are not, and the reason shows up as soon as you stop counting the letters e-n and start counting what they are doing.

We took 928,000 words of French prose and classified every occurrence.

Only about one “en” in six is the pronoun. The rest is the preposition - en France, en 1830, en colère - and the gerund, en disant, en regardant, which is a different structure that happens to be spelled the same. And more than half of every y is il y a, an expression beginners produce correctly in their first month without ever parsing the y inside it.

Translated into reading, per 10,000 words of French:

So why do they never become automatic?

Because the input is six times thinner than the ranking promises, and because there is nothing in English for them to attach to.

This is the mirror image of what we found with the relative pronouns. Dont is hard because it is rare in speech - 402nd - so learners never hear it. Y and en are hard for the opposite-looking reason that turns out to be the same reason: they are everywhere, but the versions of them you need are not. A learner who reads 50,000 words of French meets en about 550 times and the pronoun en fewer than 100.

The second problem is that they map onto nothing. Le, la and lui have English counterparts - it, him, her - so there is an existing slot to drop them into. Y and en have none. English deletes what French pronominalises, so the learner’s instinct is to delete too, and deletion produces a sentence that sounds merely clumsy rather than wrong.

That is why reading, which fixes gender and agreement by sheer repetition, does comparatively little here. Reading works on items that recur relentlessly. On this measurement, these two do not recur nearly as often as their frequency rank suggests.

Where do they go in the sentence?

Before the verb, except in the affirmative imperative. The full order is fixed:

me / te / se / nous / vous → le / la / les → lui / leur → y → en

Y always precedes en. That ordering is what produces the single most useful phrase in the language containing either word:

• Il y en a. - There is some / there are some.
• Je lui en ai parlé. - I spoke to him about it.
• Je n’y vais pas. - I’m not going there.

Three details are worth memorising outright.

The imperative restores the -s. Parle becomes parles-en, va becomes vas-y, pense becomes penses-y - the -s comes back on -er verbs and on aller so the two words link. In the negative it vanishes again: n’en parle pas.

En blocks past-participle agreement. Des lettres ? J’en ai écrit. - invariable, always. This is exactly the opposite of the behaviour of the relative pronoun que, where les lettres que j’ai écrites takes the agreement.

Y is dropped before the future of aller. J’irai demain, not j’y irai - French avoids the collision.

Which verbs decide which pronoun?

The ones that carry a preposition into their meaning. This is the list that matters, because getting the verb right makes the pronoun automatic.

There is a measurable pattern in what follows each pronoun, too. In our corpus the pronoun en sits overwhelmingly in front of avoir, être and faire - j’en ai, il en était, en faire - while the pronoun y clusters with être and aller: j’y suis, j’y vais, allez-y. The two commonest fixed phrases in the whole sample are il y en a and je vous en prie, which is a good argument for learning the frequent combinations as blocks before assembling them from the table.

What is the fastest way to fix them?

Treat them as a production problem, not a comprehension one. You already understand j’en ai deux perfectly; what you do not do is say it.

Our own A1 to C1 curriculum introduces both pronouns in a single unit at A2, with seven separate grammar points covering the substitution, the quantity rule, the imperative and il y en a. The full ordering table - the one ending y → en - does not arrive until the final unit of C1. That span is not an accident: the substitution is A2 material, and the fluency is a C1 outcome.

Three things shorten the gap. Learn the verb with its preposition attached from the first day, so avoir besoin de is never stored as avoir besoin. Drill the four or five combinations that carry most of the traffic - il y en a, j’en ai, vas-y, je m’en vais, je vous en prie - as whole phrases rather than as constructions. And when you answer a question about a quantity, force the pronoun in even if it feels redundant, because the feeling of redundancy is the English talking.

Two honest caveats on the measurement. Our corpus is literary prose, where the gerund runs richer than in conversation, so the pronoun share of en is probably slightly higher in speech. And our pronoun counts are deliberate floors: we counted an en as the pronoun only where the following word could only be a verb.

Where these two sit against everything else you have to learn, level by level, is set out in French grammar by CEFR level - and the vocabulary arithmetic behind the reading argument is in how many French words you need to know.