In Spanish (and also in Latin, and most of Romance languages, too) the present subjunctive is formed by a kind of swapping the indicative endings between the verb declension classes: the 1st class (-a-) takes the indicative endings from the 2nd class (-e), and the 2nd and 3rd class (-i-) take the indicative endings from the 1st class:

enter image description here

QUESTION: Is it only coincidental, or is there something deeper behind it? Something like "as the subjunctive often expresses uncertainty, let's make the verbs sound weird by deliberately using a wrong set of endings to emphasize that we are not sure"?

It's clear that nowadays the endings are just inherited from the language history, but is there any theory about the origin of this particular coincidence?

  • It's the first time I've seen this, my guess is it has to do with the evolution of latin to spanish, but I don't know...
    – Suriya
    Feb 1, 2016 at 15:14
  • 1
    @PabloSaudiBombsYemen: It works very similarly in Latin, too. So the idea of "swapping the endings" is more or less copied from Latin. Feb 1, 2016 at 15:30
  • 1
    Yes, even I'm a native speaker it's the first time I've noticed this
    – Suriya
    Feb 1, 2016 at 17:20
  • 3
    Well well, now this is an interesting question!. I look forward to the answer.
    – Jose Luis
    Feb 4, 2016 at 15:18
  • 2
    My answer is speculative, so keeping it as comment. As Spanish kept a small number of conjugation types (like most modern Romance languages) it may look like a swap to us. But Latin had five conjugation types. According to this, Latin had just one thematic vowel of choice for the present subjunctive of all regular conjugation types. This vowel happened to be -a- but also happened to prove cacophonic in the 1st conjugation, which also had a thematic -a- of its own. The change to -e- in just that case had euphonic reasons.
    – guillem
    Apr 21, 2016 at 15:28

1 Answer 1


The following is what I have understood from the linked sources without me being a linguist:

PIE had two different moods to express things that are not real: subjunctive (non-factual or depending on a condition) and optative (wish).

Italic languages (among which, Latin) merged both functions into one.

The new Latin subjunctive mood was (normally) constructed using the optative forms, leaving what was the PIE subjunctive to express future now.

For some reason (which I haven't found), the optative thematic vowel -i- was replaced by an -a- in Italo-Celtic languages.

This is speculation, but I feel like all five Latin conjugations would have had an -a- thematic vowel for the subjunctive... If it weren't for the first indicative already having one. The way this was resolved was using -e- for the first conjugation and I guess some standard euphony rule took place there.

Of the five Latin conjugations, we inherited just three (like in Catalan, French and Italian; others, like Romansh and Romanian, kept four).

So the situation that once just looked like a rule with one exception now looks like some sort of vowel swap, but that's just a matter of perspective:

                      12345 Conjugation
Latin:   Indicative:  AEIII (2nd pers. sing. thematic vowel)
         Subjunctive: EAAAA
Spanish: Indicative:  AEE
         Subjunctive: EAA

I must insist, I'm not a linguist and haven't dug deeply in the linked documentation; but, as requested in the above comments, this may serve as a starting point to discuss.

  • 1
    It sounds plausible! The textbooks usually teach you to create subjunctive by swapping the thematic vowels, thus obfuscating the reason why it is like this. Apr 24, 2016 at 16:21
  • Maybe the stem a resisted the shift of optative i to a, resulting in a+i which (as is not uncommon) merged to e. Oct 3, 2022 at 2:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.