Irregular Spanish Verbs: 86 Forms Do 80% of the Work

By Daniel Herrera, Spanish teacher and curriculum designer

Updated September 7, 2026 · Daniel Herrera

Every Spanish course sells you the same promise. Learn the fifteen big irregular verbs and the language opens up. It is good advice, badly measured. Nobody tells you what fifteen verbs actually costs you in memory.

So we counted. We took 1,141,520 running words of Spanish prose and measured every form of those fifteen verbs that appears in it. The answer is 553 forms. That is what the promise really asks of you.

The good news is much better than the bad news. Those 553 forms are not equally worth learning. A small handful does nearly all the work, and we can name them.

How did we measure this?

We built a corpus of ten Spanish-language books from Project Gutenberg. Ten texts, 1,141,520 running words, 57,603 distinct word forms. It is the same corpus we built for our flashcard study last week, so the cleaning has already been checked.

For the verb forms themselves we did not write our own conjugations. We used the curated conjugation templates in the open-source verbecc library, which are built on the long-established Verbiste database. Any verb the library has to guess at was thrown out. Every form counted here comes from a hand-maintained table.

Every number below is a floor, on purpose. This corpus is not grammatically tagged, so a word cannot be labelled a verb with certainty. If a form is also a noun, an adjective or a function word, we dropped it. Vino is both “he came” and “wine”, so vino is gone. Era is both “was” and “era”. Gone. That is the same choice Fratini and colleagues made on the CREA corpus. It can only push our counts down, never up. The real figures are higher than the ones we print.

How many forms do the 15 irregular Spanish verbs actually have?

553 in this corpus, counted conservatively. Spread across fifteen verbs, that is an average of nearly 37 distinct faces each.

The paradigm tables are bigger still. Each of these verbs has between 52 and 62 forms on paper once you include every tense, person and mood. Real books simply do not use all of them.

Together those fifteen verbs account for 54,642 words of the corpus. That is 4.79% of everything on the page, or roughly one word in twenty-one. For fifteen verbs out of the thousands in Spanish, that is an enormous return. The lists are right that these verbs matter.

They are just wrong about the unit. You do not learn a verb. You meet a form.

Which irregular Spanish verb forms should you learn first?

This is where the measurement pays for itself. We ranked all 553 forms by how often they actually appear, then added them up.

Twenty-four forms buy you half. Not twenty-four verbs. Twenty-four individual words, the kind you can write on a single sheet of paper.

Here are the top twelve, in order, with how much ground you have covered by the time you reach each one.

Look at what that list is made of. Third person singular, past imperfect, present. It is almost entirely narration. Dijo alone is 2,603 of them, because books are full of people saying things.

Notice what is missing too. No vosotros. No future subjunctive. Nothing in the second person plural at all. Those forms exist, and this corpus barely touches them.

The practical version. Learn each verb’s five commonest forms and stop. Across the fifteen verbs, those top five carry an average of 71.4% of everything that verb does. The other thirty-odd forms split the remaining quarter between them. Learn them later, from reading, when you actually meet one.

Why do ser and ir look like the same verb?

Because in the past tense they are. Ser and ir share 21 forms outright in their paradigms. Sixteen of those 21 turn up in our corpus, 2,767 times between them.

Fue a la casa is “he went to the house”. Fue una casa grande is “it was a big house”. Same word. Only the sentence tells you which.

This is a rare piece of good news in Spanish grammar. Two of your fifteen verbs hand you an entire tense for the price of one. Learn the fui, fuiste, fue set once and you have covered both verbs.

It is also why our count for ser looks low at 16 forms. We could not safely assign those shared tokens to either verb, so we assigned them to neither. Ser is far more common than our table admits.

Is it really irregularity that makes these verbs so common?

No. And this is the part that surprised us.

The obvious story is that irregular verbs are common because they are irregular. Old, worn-down, high-mileage words resist tidying up. English works exactly like this. Irregular English verbs really are more frequent than regular ones.

Spanish does not. In 2014 Viviana Fratini, Joana Acha and Itziar Laka ran the first corpus study to test it. They used CREA, the Royal Spanish Academy’s reference corpus of 154,212,661 words. They built a list of 13,947 verb forms, 10,284 irregular and 3,663 regular, and counted every one.

At first the effect looks real. Irregular forms came out 2.6 times more frequent on average, a mean of 589 against 219 for regular forms. Then they looked at the spread. The standard deviation for irregulars was 11,997 against 1,481 for regulars. A handful of monsters was carrying the whole result.

So they removed the outliers and ran it again. The correlation between frequency and irregularity vanished (r = .058, p = .18). Across the low, middle and high frequency bands it was never there at all. Their conclusion is blunt: for most Spanish verbs, frequency and regularity are independent variables.

One correlation did survive, though, and it is a much better guide.

Short wins, every single step of the way. The mean frequency of a two-letter verb form in our corpus is 814.4. At three letters it is 109.2. At four, 37.1. It keeps falling without a single reversal, all the way down to 1.8 at fourteen letters. The correlation between length and frequency is r = -0.282 across 17,833 forms.

Fratini and colleagues found the same thing on a corpus 135 times the size, and their length effect held up after the outliers were removed, which the irregularity effect did not.

So here is the honest rule. The famous verbs are not common because they are irregular. They are common because they are short, and they are short because they are old and worn down by use. Irregularity is a side effect of the same wear. It came along for the ride.

This is useful, not just interesting. When you are deciding whether a form is worth a flashcard, its length is a better clue than its conjugation class. Es, hay, fue, dio are tiny and everywhere. Hubiéremos is long, and you will go years without meeting it.

What should you actually do with this?

What this study does not show

Our corpus is literary and it is old. These are nineteenth and early twentieth century books, plus Don Quijote, which is far older. That shapes the result. It is why dijo ranks so high and why vosotros barely shows.

It will not tell you what a taxi driver in Bogota says this afternoon. Spoken Spanish uses more present tense, more questions, more second person. If you want contemporary written Spanish, the CREA study is the better guide, and it points the same way on the length finding.

Our counts are also deliberately low. Dropping every noun homograph cost us real verb tokens. We would rather undercount than guess.

What survives both limits is the shape of the thing. A few forms do most of the work, and short forms win. That held in our 1.1 million words and in CREA’s 154 million.

Frequently asked questions

Keep going

• Ser vs estar, the one distinction that causes the most trouble, explained with a single rule.
• Imperfect vs preterite, which is where most of the forms in our top twelve actually live.
• Gustar and the seven verbs that work backwards, a much smaller memory problem than it looks.
• Anki vs Quizlet vs Memrise, measured on the same corpus, if you are about to build a deck from this list.
• The subjunctive, where quisiera and fuera come from.
• How to read books in Spanish as a beginner, which is how you learn the other 467 forms.

Sources

• Fratini, V., Acha, J. and Laka, I. “Frequency and morphological irregularity are independent variables. Evidence from a corpus study of Spanish verbs.” Corpus Linguistics and Linguistic Theory. Full text (PDF).
• CREA, Corpus de referencia del espanol actual, Real Academia Espanola, 154,212,661 words.
• Conjugation tables from the verbecc library, built on the Verbiste conjugation database. Curated templates only.
• Corpus texts from Project Gutenberg, ten Spanish-language works, 1,141,520 running words.