In some languages like Chinese, it isn’t imperative that the tense of the verb is explicitly marked. So if you mean an action that will occur in the future, you can still refer to it in an all-encompassing way with a “present” form.

But in English, for the most part, the tense has to be marked, and it therefore has to agree with the scenario. If it doesn’t, it’s a communicative error.

How can the language-acquisition mechanisms in the mind permit either necessary explicit tense or a lack of explicit tense? Is it theorised that this is a selected parameter of language, like head-first or -last; SVO order; etc.? It reminds me of programming languages which can be strongly- or weakly-typed. That means that the language either does or does not have to state the type, of the variables.

Is there a connection to be drawn there?

1. Is the claim even true?

Or false, or currently unknown? Justify your assessment.

2. How answerable is the question, right now?

On the spectrum of “it’s not really ready to even be correctly formulated”, to “the question is sufficiently ripened, the answer requires further research,” to, “the answer is essentially already known, it’s already answerable.” Justify your assessment.

3. Is this an open research problem in linguistics?

That would imply it’s unknown, which is still a claim towards something. Then, see #2 above: justify why you consider it unknown, and not at the moment settled.

  • 3
    It's just because some languages do have the category of tenses on the verbs, and some don't. Just the same way as English has the category of definiteness on nouns and it has to be marked explicitly on every noun, but Polish, Ukrainian, Russian and other languages have no such category on nouns, and it's up to you whether you mark it in a way or not. Each language has it's own set of grammar categories which are mandatory to be marked.
    – Yellow Sky
    Oct 27, 2021 at 11:52
  • 5
    I’m voting to close this question because as a Why?-question, it cannot be answered. Oct 27, 2021 at 12:24
  • 10
    @jk-ReinstateMonica I disagree, because many syntactic theories (especially the more UG-leaning ones) do specifically try to explain why this happens. I believe more lexicon-oriented syntactic approaches also have their own explanations of this, though I don't know as much about those.
    – Draconis
    Oct 27, 2021 at 15:03
  • 4
    In Chinese, we use extra vocabulary to indicate the tense and aspect. Since Chinese doesn't have morphology, certain adverbs like "xianzai"(now) "mashang"(is going to) and prepositions like "yao" also indicate the future, etc., and for marking the aspect, we use "le","zhe","guo" which is usually attached to the verb phrases or at the end of the sentence. I agree @YellowSky "each language has its own set of grammar categories". Draconis you are correct I think.
    – Yili Xia
    Oct 28, 2021 at 0:03
  • 3
    It is interesting that there is still a lot of debate on whether Chinese clauses have the difference on finite or non-finite because there is no morphology for tense and aspect. CT James Huang2008 believes Chinese do have finite and nonfinite clauses but Xu Liejiong1994 does not think so.
    – Yili Xia
    Oct 28, 2021 at 0:19

4 Answers 4


Ultimately we can't answer why one language grammaticalises tense and why another language doesn't.

But what we can say is that all languages have at least one major verbal grammatical category. Tense is just one option, others include aspect, modality, or evidentiality; together these are called Tense-Aspect-Modality(-Evidentiality). As far as I know, all languages have at least one of these as a grammaticalised verbal feature. One of them will be the prominent category - it will be obligatory on non-infinitive verbs (though not necessarily as an affix). If there is a secondary category then it will be less significant, and may not be fully grammaticalised. English is an example of a Tense-prominent language: Aside from infinitive verbs, all verbs must have either PAST or NONPAST tense in English. English also has aspect - the progressive -ing, but it's not obligatory, and its absence does not always indicate perfective aspect. English also has auxiliary verbs which indicate modality, but it doesn't express the base realis/irrealis distinction.

An excellent short book on this topic is The Prominence of Tense, Aspect and Mood by D.N.S. Bhat. In addition to explaining these categories he also cautions against letting the prominent category of your language (for example, tense in English) hinder your observations of the languages you study which may in fact have a different prominent category.

  • 2
    Very fascinating! I hadn't heard of this "prominent category" theory before, but I also can't think of any languages which lack one.
    – Draconis
    Oct 28, 2021 at 0:17
  • 1
    I wonder, what's the prominent category in Malay/Indonesian?.. It looks like it's ‘voice-and-the-like’, but then it's not tense, not mood, and not aspect. Maybe the “tense, mood, aspect” list isn't complete? Or is voice a kind of evidentiality? Anyhow, even voice in Malay/Indonesian is in the sphere of derivation, word building, not inflection. And in some dialectal varieties parts of speech don't exist at all and no categories whatsoever are marked. How to treat that?
    – Yellow Sky
    Oct 28, 2021 at 0:52
  • 2
    @YellowSky Austronesian languages are generally modality prominent, though I don't know the specifics of Malay/Indonesian. A quick search suggests that Malay has lost its modal morphology, but the category could be expressed through auxiliaries instead? Voice is not typically grouped with TAM(E) categories, but maybe an argument could be made that a language is voice prominent instead of TAM? As to your final question, I've never seen a credible example of a language without parts of speech. Roots may be category-less, but languages still need nominals and verbals at least.
    – curiousdannii
    Oct 28, 2021 at 1:23
  • 1
    What you wrote in your answer is extremely interesting and I'm in no way going to contradict you, I'm just trying to put the whole picture together. Here's an article Valency Classes in Jakartan Indonesian about the dialect I mentioned. It discusses the OP question as applied to that dialect. You'd do a great favor if you have a look at it and share your thoughts about it.
    – Yellow Sky
    Oct 28, 2021 at 2:58
  • @YellowSky Sorry, I really don't know enough to say more. Maybe you have found a counterexample to the TAM dominance!
    – curiousdannii
    Oct 28, 2021 at 3:12

All human languages allow the expression of distinctions in time reference, so there's always a way to describe the situation that one event precedes another. Some languages do this with special grammaticalized markers, perhaps particles or morphemes attached to other words. Similarly: all languages allow the expression of the idea that a thing is small, or that it is big. Some languages do this with a grammaticalized marker – a diminutive or augmentative affix or noun class.

It is a basic property of languages that syntactic representations can have features of some sort, and that combinations of node can requirement "agreement" in those features. Very common examples are person, number, "gender" (including noun class and animacy-marking), definiteness, case, negation, tense). The general finding of linguistic typology is that in principle any semantic property can be grammaticalized and is subject to obligatory marking. However, some things are more prone to grammaticalization than others. Person and number are highly susceptible to grammaticalization, evidentiality is less-so.

Tense is more complicated because tense is traditionally used to indicate formal differences in verb morphology, but also to refer to precedence relations between events. A similar but different distinction is made between tense vs. aspect, which has to do with the "extent" of an event over time. Chinese does have grammatical markers for aspect.

The immediate explanation for why a given language mandates marking of some event-related property is that children learned that system based on the ambient data produced by their elders, who likewise learned a system from their grand-elders and so on. Most of the answer for a specific language is based on historical conservatism. There is a functional consideration that favors some system of marking, that there is functional utility to distinguishing between saying that some event has already happened, or is happening now, or is expected to happen (it's futile to try to prevent a fait accompli). Such communicative-functional considerations also favor the expression of number.

  • Thanks. So you don’t think there has to be a system which generates the grammar of a language based on an underlying parameter but that much of the grammar rules may be an accidental result of semantic expressive needs getting grammaticalised? Oct 27, 2021 at 16:03
  • 3
    There is a "system" underlying grammar learning, but I don't think it has any substantive parameters, as in P&P theory. So for example I don't think that there is a universal label "tense" that must be present but is obligatorily empty in some languages.
    – user6726
    Oct 27, 2021 at 16:16

After studying the Cambridge Grammar of English a little bit, my perspective on this question has changed, in a way that I must say I find pretty elucidating.

In the CGEL - in many grammatical dimensions, such as clause type vs. communicative function - it is heavily stressed how often there is a difference between grammatical form vs. semantic meaning. In fact, we are so used to hearing grammatical forms defined in terms of, or as having, a specific semantic use; that is (apparently) what defines them. For example, the “preterite” in English (we are told) refers to the past tense. This appears to be so widespread that it may be disorienting and hard to grasp when one first considers the idea that the preterite has no meaning in and of itself. The preterite is (according to the CGEL) only a shape, like looking at keys and saying this is a skeleton key, this a bump key, and not (yet) saying anything about where they can be used (the lock on the front door, the lock on the shed, etc.). Later, you define certain rules/structures/clauses, which do have a semantic meaning, and which call for a certain form. But one can observe sentences in which the preterite form means past, present, or future (in the latter, “You said it started tomorrow.”

It seems like meaning controls form more than form controlling meaning, but that’s my thought, not sure if guaranteed.

The parallel with tense is that while all languages can obviously communicate about states of affairs which necessarily happen to have occurred at a certain point in time (past, present, or future), I think (according to the general view on grammar taken by the CGEL’s authors), the question is not about what languages can communicate, but only a question of what time-related grammatical forms the language has. What I found surprising but definitely memorable was the authors’ claim that, actually, English has no future tense. If you look at the grammatical behavior of a word like “will”, the authors’ argue that it behaves like a modal auxiliary (words like can, might). It also seems to have more variegated meaning than is often relayed in grammar lessons. The authors say that “will” is associated with dynamic modality - a way of talking about what someone is able or willing to do, along with words like can.

I find this refreshing. It seems that the verb “will” indeed often conveys one’s will (in the sense of volition), for example, in the sentence I asked John to fix it but he won’t. Bringing back the connection to Chinese, I am pretty sure the Mandarin word hui similarly can mean roughly “can” or “will”, depending on context.

In the CGEL, it’s stated that English has 2 tense-related forms - present vs. preterite (run vs. ran), and perfect vs. imperfect (run vs. have run).

How I understand this to bear on the question and some of the answers above is that it may not be the case that languages are so strictly time-marking or not - according to the analysis above, English is nowhere near as time-marking as I originally claimed it was; and a language like Chinese has ways of implying time that are actually similar to at least some English constructs. For example, I said that Chinese can use the present form to refer to the semantic future - but this is not actually distinctive to Chinese. I know this to at least occur in Germanic languages such as German and Swedish. Furthermore, it can even occur in English: Tomorrow, we ride. / You didn’t pack? We leave tomorrow!

Although I don’t know Chinese well, one might inquire into whether, since Chinese I think is an “analytic” (high morpheme-to-word ratio) language, that Chinese can’t denote tense through inflection - since it doesn’t have inflection. This is a very trivial characteristic which we should not take to imply “there are no tense markers in Chinese”. English (according to the CGEL) has a tense system marked by a separated word (“have”). In principle, it’s perfectly possible the same is true in Chinese.

This is a satisfying development on my perspective on this question, some time later!


The question basically asks, “Why do human languages have different features from each other, at all?”

I think the hard part of the question comes full circle to I think is meant by “universal grammar” (but I’m not sure).

It’s actually a really simple and common question to formulate:

how much of what is observable in human language use is a priori

(innate - there is something in the human, in the brain, at the beginning, that had a role in the observed patterns in human behavior, ie language, being what they are) - how much is a fortiori - there’s still a reason it got like that, but it doesn’t commit the “incidentally/necessarily” fallacy: it happens to have turned out at that on this specific occasion - which is not to say that it has to be that way, which implies limited ability to say something about constraints on the whole system - rules about how language can or cannot be; in some ways of looking at it, a huge amount of constructed languages are perfectly valid languages are we don’t need to assume the common structures in natural language right now have a strong genetic component. In fact, we might even challenge ourselves to reject the term “natural language” until we can justify that there is some clear reason to distinguish between things that way. Humans can learn Python, and they do, so why is that not a human, or “natural”, language? This is meant more as a thought experiment to see the question more broadly.

The key difference is modality: true vs. necessarily true, (or even “necessarily necessary”).

One really interesting place to look is considering to what extent people’s “natural languages” already are - in part - constructed. Humans consciously manipulate their own language use. Hebrew was rescucitated by linguists, and put into use. Language reform initiatives modify their own orthography, and we call it “natural language”.

It would be very interesting to find that commonalities in the spontaneous, undirected evolution of human language activity (chattering) are common patterns which would likely come about in some stochastic process, for any context where “agents” wanted to communicate. Maybe a game-theoretic, mathematical model of some kind can show why clauses being made of NP-VP so broadly occur in our speech: perhaps it is the simplest way to express essential components of human conceptual reality, including things like:

  • objects
  • agency
  • object persistence (that a static form in time is “one thing”)
  • attributing changes in “processes” (residual objects) to something called “actions” (the thing is still a thing, but it changed (state) in a certain way)

and so on.

I don’t know what the strongest experiments have been against the paradigm of extreme anti-nativism (which I guess is this perspective), but it could be a really useful paradigm to explore, to check nativist assumptions, a contrarian point of view or contrapositive: if we return to a blank slate and take the classic B. F. Skinnerian view that language is a completely socially learned behavior, is there any fatal, death-knell observable phenomenon that absolutely cannot be explained without assuming an innate syntactic structuring mechanism, a “language faculty” - which is non-trivially different merely from the “language of thought”, the structure of human cognition itself? Is oral language common just because the mouth is such a convenient way to communicate, with sound, but human language can just as well be predominantly a sign language or any other sensory modality, if a physical environment, a context, made that a go-to modality? (If humans evolved back into an aquatic species, how would might language behavior change? Might it become haptic, based on physical taps, instead?)

So, why do some languages have this or that feature is basically just the question of universal grammar and linguistic nativism.

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.