1. Forum
  2. >
  3. Topic: Duolingo
  4. >
  5. future-in-the-past and condit…


future-in-the-past and conditional across languages

In English forms of the conditional "I would do [something], if [some condition]" coincide with "future-in-the-past": "He said he would do it" (at the time he said it, his doing it was as yet a future event). Namely, we use "would" for both.

As far as I know (which might well be imperfect knowledge), the forms used in these two instances also coincide in Dutch as well as the more commonly-studied Romance languages. Anybody have any linguistic insight as to why this might be? It doesn't seem a priori necessary, or even likely. I would think it was a common Indo-European phenomenon perhaps, but Latin I think would have used different forms in the two instances, although I'm not sure (and would be happy to be corrected on the point), and in Russian, and I presume Slavic languages in general, it doesn't apply at all.

I'm curious in what other languages this coincidence of forms does or does not exist, as well, so feel free to post about other languages you are familiar with.

January 5, 2018



In Japanese, conditionals are formed by either conjugation, a particle, or both. Which one you use carries a slightly different connotation.

水を飲むと、元気になります。 If you drink water, you'll become healthy (natural consequence).

みんなが行くなら、私も行く。 If everyone goes, I'll go too (given this context, consequence).

今行ったら、間に合う。 If you go now, you'll be on time (consequence can only be realized if condition is met).

おいしければ、買います。 If it's tasty, I'll buy it (emphasis on condition).

For "He said he would do it," I think it would just be a quotation. 彼はすると言っていました。


This seems like it would allow for a nice range of subtle expressions.


These examples are great for what they are, but none of them are the "future in the past" that OP is referring to. An example would be "If everyone had gone, I would have gone too", or "He thought (past) he would go (future of that past) at 10:00, but his car had broken down".


This is a very interesting topic. I think that the situation in the Romance languages is a bit more nuanced, at least when it comes to the translation of your example sentence. But the situation in English is (arguably) similar, so I am going to use English examples first:

(1) He said (that) he will do it. (2) He said (that) he would do it.

I think that only sentence (1) rigidly refers to a moment in time that is yet to come, even at the moment of utterance. As of yet, the person we are talking about in (1) has neither fulfilled nor failed to fulfil his promise, thread, etc. In (2), the situation is less clear, or so it seems to me. It could be that, at the moment of utterance, the person referred to has already acted on his promise, etc. Compare the modified sentences:

(1a) He said he will do it, but he did not do it. (2a) He said he would do it, but he did not do it.

In my humble opinion, (2a) sounds preferable and more logical than (1a). In the case of (1a), I feel that I could just flippantly reply "Well, he still has plenty of time to do it", whereas in (2a) it seems to be a shared assumption that the right moment for performing the action (say, some deadline) has already lapsed.

Perhaps the more obvious point is that (1) sounds more affirmative than (2). I take the latter (i.e. the conditional) to be somewhat hedging. Again, a slight alternation of the original sentences might illustrate this:

(1b) He said he will do it, but I know that he won't. (2b) He said he would do it, but I know that he won't.

To my ears, (2b) sounds preferable. But these are very subtle points, and our intuitions may differ. In any case, if I am right, this might explain why there are, at least when it comes to indirect speech, two different ways of expressing a future-in-the-past in e.g. French (and Italian as well):

(3) Il a dit qu'il le fera. (4) Il a dit qu'il le ferait.

(4) expresses more reservation and doubt than (3), and so the use of the conditional here would not be a mere historical coincidence, as the conditional is often used to distance oneself from a claim, for example in polite requests.

The future-in-the-past is expressed in a similar way in German, though the conditional there does not count as a proper mood, but rather as a semantic function fulfilled by the conjunctive (which is used abundantly, for instance, whenever speech is reported indirectly, in what Classicists call "oratio obliqua"). The example would be translated as follows: "Er sagte, er würde es machen" or "Er sagte, dass er es machen würde".


Very interesting hypothesis!

To make sure I'm understanding, when you said, "I think that only sentence (1) rigidly refers to a moment in time that is yet to come, even at the moment of utterance," the utterance you're referring to is when "he" said, "I'll do it"?

If we create a time line with event 1 being his saying "I'll do it" and event 2 being the present time when event 1 is being recounted: "He said he would/will do it", then I think grammatical use of "He said he will do it" is restricted to situations where "it" has not been done as of the time of event 2 (from this follows that sentence 1a is actually ungrammatical; the second part would have to be in present perfect). Do you know if that's the same condition that applies to "Il a dit qu'il le fera"? (Feel free, of course, to dispute that I have identified the correct English condition.)


You put it much more clearly than I did. That is the point I wanted to make with (1a) and (2a).

The utterance I was referring to, however, was the indirect quotation (1) which reports the previously made commitment. Here it is obvious that the fulfilment of the promise, execution of the order, etc. is supposed to take place in what is still the future. In (2), this is not the case. For in this scenario, it might be that the utterer of the original, direct statement ("I'll do it!") has already acted on his promise, or that he has definitely failed to act on it, etc.

The case of French is interesting. Strictly speaking, (3) is incorrect, as it flouts the rules of "concordance des temps": when "Il a dit: 'Je le ferai'" is transformed into indirect speech, you would have to use a conditional ("Il a dit qu'il le ferait"). But in everyday, colloquial French, the use of the "futur simple" is acceptable, and the nuance expressed seems to be similar to that conveyed by the use of "will" in English, namely that the promised action has not yet been performed. (The use of the conditional leaves it open whether it has been performed or not.)

Italian is another fascinating example. Here you would have to use a past conditional: "(Lei/Egli) disse che lo avrebbe fatto" or "(Lei/Egli) ha detto che lo avrebbe fatto". Again, in colloquial speech, the conditional is sometimes replaced by a future: "(Lei/Egli) ha detto che lo fará". Such a use is certainly not a sign of education, but neither is it an egregious mistake. The use of the simple past together with the future would strike me as odd, probably because the simple past tends to be replaced by the compound forms in colloquial speech.

A further side note on the German translation, as what I said could make some German grammar enthusiasts cringe. Strictly speaking, the correct form in both sentences would of course be "werde" (Konjunktiv I) instead of "würde" (Konjunktiv II). But nowadays this tends to sound overly correct, even somewhat poetic. Also, with a "dass"-clause, you could just stick to the future tense: "Er sagte, dass er es machen wird". In contrast to the Romance languages, such a use of the future would be perfectly correct (and not just acceptable in colloquial speech).


I've only skimmed your comments after the first paragraph, but you seem to be missing actual grammatical nuance here. While "future of the past" may not be linguistically technical terms, they are accurate. "Time of utterance" is always the present, that's exactly what "time of utterance" is, by definition. But "He said (that) he would do it" is, indeed, the "future of the past" because he said (in the past) that he would do it (at the time of him saying he would do it, he had not yet done it). Whether or not he actually did "it" by the time of utterance is irrelevant to what OP is asking about.

you say that between: (1b) He said he will do it, but I know that he won't. (2b) He said he would do it, but I know that he won't. you think (2b) is preferable, but these are two different sentences for two different situations. (1b) is perfectly normal. in (1b) we assume that "he" hasn't done "it" yet, and the speaker is asserting that he won't ever, either. if anything, (2b) is a little odd. If what is meant is he said (in the past) that he would do it (future of past), and you want to assert that the time in which he would have done it is past and he didn't do it, then it would be better as "but/and I knew that he wouldn't". As written, (2b) is actually more like (1b), and probably a little less standard, though I can't quite place why. actually, I kind of can, but I also just told my wife that I'm almost done so that she could get on the computer so...¯_(ツ)_/¯


I did not want to say that there is no future-in-the-past, or that the term is somehow a misnomer. I just wanted to show that a future-in-the-past (to use that cumbersome label) can be expressed by various means, and that the connotations slightly differ.

Admittedly, I do not see why the time of utterance should always coincide with the present, though it does in the cases I cite. The time of utterance of "Veni, vidi, vici" as pronounced by Caesar is most certainly not the present (if he ever said such a thing). In my toy example, the point in time where the speaker we are talking about actually said "I'll do it" (=the time of utterance of that promise) also lies in the past.

(2b) sounds odd because of the use of the future "he won't". In fact, there is a semantic tension between "would" and "won't" in this sentence, or so it seems to me. The speaker in question expresses his/her putative knowledge about a future action (or about a lack thereof), although what is supposed to be relevant is only what has happened up until now. Note that this semantic tension could be alleviated by replacing the future with a form of the present or past: (2b) He said he would do it, but I know that he is not doing it (which admittedly sounds slightly odd to my ears) and (2b*) He said he would do it, but I know that he has not done it.


Spanish has different possibilities. The most used works as in English: "Ellos vendrían si pudieran" (They would come if they were able to do it) vs "Ellos dijeron que vendrían." (They said they would come). However, it is possible a construction in periphrastic future: "Ellos dijeron iban a venir" (~ They said they were going to come). And in simple future: "Ellos dijeron que vendrán" (~ They said they will come). Even in present (informal): "Ellos dijeron que vienen" (~ They said they come). And in imperfect preterite (more informal perhaps): "Ellos dijeron que venían" (They said they came).


Great breakdown, thank you!

Would "Ellos dijeron que vendrán" only be used if "they" haven't arrived yet, or is it less strict than (at least my understanding of) it's word-by-word English equivalent?

"They said they come" can be used in English, but it would be about an ongoing, intermittent "coming" (e.g. "They said they come to meetings every month"). But I assume "Ellos dijeron que vienen" isn't similarly restricted and the Spanish could be used about coming a single time?

I would be tempted to translate "Ellos dijeron que venían" with "They said they were coming," which winds up being perfectly standard English.

For "que vienen" and "que venían," does it matter if "they" have arrived yet?


Yes, "Dijeron que vendrán" is possible if they haven't arrived yet.

And yes, that structure is also possible in Spanish "Ellos dijeron que vienen [a esta playa [todos los veranos]].", but can be used informally also for "They said they would come".

Only "que vienen" is restricted to a non accomplished coming.


Ok, well for Spanish, in some cases the imperfect tense is used to describe a future action in the past. I think that might be what you are talking about, but i'll use Spanish as an example.

Ellos dijeron que iban a reparar el coche. They said that they would repair the car.

Note: "iba" in Spanish can also be used as "Ellos iban a la playa" "They would go to the beach"

Sorry if I sort of just repeated one of the points you made, perhaps I misunderstood the phrasing if I did


I think that's a function of forming a de facto future with "ir a," right? In English we can say "they said they were going to repair the car," which winds up being essentially a word-for-word translation and also doesn't use "would." I guess the common logic is that at the time of the reported statement the "going to" part is actually in the present, just as we'd say "They said they were going to the store." At the time "they" said that, the "going" is in the present, and that simply remains the same when the "going to" is used with a verb.

"Would" for repetitive actions in the past, yeah, that's a separate case, more likely to be a peculiarity of English.


I said I was gonna do it. suggests that it is really just a backshift as you postulated for spanish, in which case the conditional is just an artifact of "would" being the past of "will".

like russian, czech does not backshift, so Řekl, že by to udělal. is just as conditional as Říká, že by to udělal.


Yeah, the coincidence of "would" in the two cases surely comes from "would" originally being both the past tense form and subjunctive form of "will." One way to restate my question would be: what's the reason that Romance languages wind up doing the same thing?


I would have thought that Russian does do it similiarly. Surely он говорил что он бы делает это is equivalent to he he said (that) he would do it? I have been translating lots of conditionals in the French from Russian tree lately and have been struck by how neat the correspondencies are. Although my understanding may be a bit fuzzy given I am going between languages which are not my own. I am no grammar expert finding it tedious but the Russian form feels much like the English one. More so than like the French.


In french you form the conditional by conjugating the verb (into the appropriate form of the conditonal tense). You form the past conditional by using the conditional form of the verb avoir or etre/ to have or to be (depending on the verb) with (before) the past participle of the the verb. So, I guess in french avoir/etre, take the place of would/ would have, but you only need the helping verb, avoir/etre, in the past. In the conditional (future) you wouldn’t use any auxiliery verb, you would just change the verb to it’s condtional form. So it seems like it’s not really the same as in English at all, because in English guess we don’t have a conditional conjugation for verbs? Avoir or etre are the verbs used to congugate all the tenses in french. French is pretty much all about conjugating with avoir or etre, you use those verbs in various forms to form all the other tenses. I just take English grammer for granted, so I actually have no understanding of it, and confuse myself when I think about it.


Hmm, I'm not talking about past conditionals ("would have") though.

In French, I believe my examples would be "Je ferais [qqch], si [une condition exprimée avec l'imparfait]" and "Il a dit qu'il le ferait" where the word-for-word translation of "je ferais" and "il ferait" are naturally "I would do" and "he would do," which is the condition I was referring to. You're right that, linguistically speaking, English doesn't have a conditional conjugation. We just use "would," which for standard conditional statements winds up being the same in meaning as the infinitive+ais/ais/ait/ions/iez/aient conjugations in French.


I never thought about it before (because I don’t think about the english language much) but now that I think about it, it is super weird that English doesn’t have verb conjugations for any future or conditional tenses. There is not way to change a verb to indicate that the action takes place in the future without adding another word. All other languages I can think of have future tenses. Hungarian has verb conjugation for future too (that is as far as I will go to attempt to explain Hungarian grammer though) and every other language I can think of. I wonder, is this weird for people learning English? Because the more I think about it, it is prompting a mini existential crisis. How can I plan for the future if I don’t even have a tense for it? I was wondering if older forms of english had verb conjugations for the future/conditional (i think of conditional as a future tense, as opposed to past, present, though I know technically that is not completely accurate) Would a Bible such as He cometh down the mountain be in the future tense? I tried to look up the old English suffixes of ith eth but was not able to find any solid answers to my queries. Is the lack of a future tense in English weird for learners? If old English had any future verb conjugations than I lobbyith we bringith them back into current use. Any historical linguists here?


Actually I don't think any Germanic language has the kind of future tense conjugations you're discussing. All only have past and present. This sort of situation actually isn't uncommon. Many languages specify tense information even less than English. In Guaraní and I think Chinese and maybe other Southeast Asian languages, you can perfectly well say "I go yesterday" since "yesterday" really provides all the time information you need. In Japanese I believe verbs are either past or "non-past," with non-past verbs having either present or future meaning depending on context; if it's "next year," then it's future. Apparently, this is also how it worked in Old English (Anglo-Saxon), although "willen," which would evolve into "will" had begun to make an appearance. Most languages I'm aware of do this to some extent. I think you can translate "The train leaves tomorrow at 8 A.M." using the present tense in Romance languages as well as Russian.

I think modern Greek has wound up quite similar to English; it has strict conjugations for present and past but forms the future with a helping verb ("θα"). Unlike English, I think, this reflects a simplification from earlier forms of Greek that did have complete future conjugations just as modern Romance languages do. If we only refer to spoken French, then French doesn't really have a past tense any more if you think about it; the "suis" in "je suis allé" is present, simply followed by a participle. Catalan has taken the "going to" thing and turned it on its head; "go to + infinitive" is now the standard way to form the past tense. So its "really conjugated" past tense is on the back foot, too.

So, if I understand what I've read correctly, "He cometh down the mountain... tomorrow" would indeed be future, but "He cometh down the mountain" was merely plain "He comes down the mountain" of yesteryear, and the "th" I'm pretty sure marked only third person singular. But if thou wouldst be a tad antiquarian, there are some present tense verb conjugations and pronouns you could employ :)

EDIT: This description of French past tense forms is incomplete. Rectified below, I hope.


Interesting! It is interesting that from the examples provided some languages have present tenses, some have past and present and helping words to form the future, and some have present and future with helping verbs to indicate the past. I wonder if any languages has conjugations for all three: past present and future.


I guess you mean forms all three tenses with auxiliary verbs/particles? I think normally the present is something like the default, but there might well be one that has to include some sort of helping verb for all three. Who knows?

Of course at a certain level, what's an auxiliary verb and what's part of a conjugated form is something of a matter of writing convention. For instance, in Guaraní people separate a lot more grammatical markers, etc. into separate words when writing than the experts say they ought to (or at least the expert whose book I've managed to obtain).


I was actually thinking of the opposite, a conjugation of the verb into past, present (the default) and future, without the need of helping auxiliary verbs. What you say about the separation of verbs and auxiliery verbs being arbitrary makes sense though.


Well, I see your reply in my e-mail, but it's not showing up here. Been happening a lot lately :(

So I guess it tuns out you meant a languages that simply has different forms for past, present, and future directly included in the verb without any auxiliaries. I think the linguistic term for this is "morphological tense," which just means "separate verb form for a tense" ("morphological" means "relating to form"). Apologies for the misunderstanding.

Yeah, many languages have all those forms, I think; Latin did (it could express a good number of other distinctions with such verb forms as well), so the Romance languages also do, even if they're in limited use now. Sorry my mention of French before was incomplete. It forms the past with such a verb for ongoing actions ("used to go") or background to a single action ("was doing so-and-so when..."), but the one for single actions in the past ("I had a burger for lunch today") has been displaced by a compound form in speech and informal writing. Looks like ancient Greek did, too. And Russian, like for French where the nature of the past action matters, in Russian it depends on the type of action in the future as to whether it's a morphological future tense or a compound future form.

Guaraní is sort of an intermediate case. It doesn't have to mark tense, but it can. But if it does, it often just sticks the same particle onto the verb independent of grammatical person (I/you/they/etc). That's, after all, pretty much what English does with "will," but in Guaraní these particles happen to be written as part of the verb. There are reasons for that, but in part it's just because Guaraní's writing system is more attuned to linguistic realities than English's. For example, we pronounce "a" and "the" differently depending on context, but the distinction isn't rendered in writing. In Guaraní, such a thing would be, and it's rendered by writing things as joined words or not.


My answer might have shown up in your email and not here because I wrote it, deleted it and rewrote it several times because I was having trouble expressing myself. Yes, that is what I meant to say. I didn’t know it was called a morphological tense but that is it. I wonder what the philosophical/ cultural implication are, of having morphological vs. Compound tense. I feel like a morphological tense places more emphasise on the importance of the action whereas a compound one more on the subject, in a very subtle way that is more philosophical than lingusitic and cannot be literally translated. I suppose to me compound tenses (especially in english) feel like workarounds, especially for the special future past cases you mentioned, for things we don’t actually have proper ways to express. I take english for granted, but when I actually think about how we describe time, it seems really cumbersome and circuitous (and self centered because the emphasize is on the beingness of the speaker and not the time). I don’t know any Guarani, but from what you describe at least, it seems like they have solved the whole problem nicely. Good discussion, it got me thinking!

Learn a language in just 5 minutes a day. For free.