Understanding perfective aspect

[Curlyjimsam]

I would suggest that from a conlanging perspective the most important thing, probably, is how individual languages divide up their aspectual categories (which if you want inspiration is probably best gleaned from detailed grammars of particular languages). It's entirely possible that a universal underlying semantic notion of (im)perfectivity simply doesn't exist, and every language has a slightly different way of dividing up the conceptual space relating to time.

[Vardelm]

In fact, boundedness I would say is the number one characteristic of perfective. Bounded and unbounded might almost be better terms than perfective and imperfective. Perfective/bounded means the event in question is conceptualized as having a beginning, middle and end – i.e. bounded in time, a whole event. Imperfective/unbounded means you're not thinking of it that way. You're thinking of it as a) part of an event, something ongoing, most likely a part of the middle of the event, perhaps while something else was happening; b) as a state, a description of a general truth/state; or c) as something habitual, something that happens again and again. With the habitual sense, it's really a sequence of perfective events, but the focus isn't on them as bounded events but the fact that it keeps happening, almost like a segmented stative, so again, a description of a general truth/state.

I have struggled with perfective vs imperfective as well. I think I mostly have it down, but there are still times with a bit of confusion. Seeing these labels of "bounded" vs "unbounded" is really helpful IMO, and jives with how I understand perfective vs imperfective. Sometimes re-labeling can be useful, such as calling the perfect "retrospective" to mirror the prospective.

Semantically, some verbs are ill suited to perfective aspect and others ill suited to imperfective aspect. For example, semelfactive verb (i.e. a verb that refers to an instantaneous action, something that takes seconds or less) is generally not used as imperfective except maybe in the habitual sense. So "kick," "slap," "break," "let out a cough" will most likely always or nearly always be perfective. In English, you can imperfectivize these but this implies rapid, ongoing repetition. "I'm kicking him" means that I have kicked him several times, I'm still kicking him, and I haven't stopped yet.

I think it's useful to consider why, when learning this. Semelfactives are so short (basically instantaneous) that considering their "internal structure" as an imperfective doesn't seem practically useful.

Probably "perfect/imperfect" should have been left to the grammarians, and new terms invented to describe the ideal aspects. This is how alignment was handled... imagine trying to explaining ergativity if we had to use "nominative" to mean both "nominative case" and "semantic agent".

And even on that topic, there's still lots of confusion! I find that the terms used do well for describing the morphology, and perhaps the syntax, but start to break down when you get more into semantics where you start dealing with adjacent areas like voice. Maybe that's because terms for voice don't suitably cover all of the phenomena you see. Just look at the confusion over what constitutes a "passive" in nominative vs. ergative vs. Phillipine alignments.

[Richard W]

Saying that different languages "use the imperfective for different meanings" is odd though: specific languages' traditions generally use the terms "imperfect_m/perfect_m" instead, maybe those are what you meant? Imperfective/perfective are mostly used in general linguistics, in their pure semantic sense. (Chinese languages are one of the few language groups where "perfective_(m)" is a normal term...)

Imperfective_m v. perfective_m is fairly standard for Russian and Semitic morphology, and in other languages the related tense labelling is very varied. However, where the difference maps to morphology, I get the impression that the boundary between the two varies from language to language. Certainly English seems not to comply with the European norm when it comes to the choice of past tenses, but I don't think it's just a case of English v. the rest. As to the terminology, I would also use it for the basis to the difference between present system imperative and the aorist imperative in Greek, ancient and modern.

[bradrn]

I would suggest that from a conlanging perspective the most important thing, probably, is how individual languages divide up their aspectual categories (which if you want inspiration is probably best gleaned from detailed grammars of particular languages). It's entirely possible that a universal underlying semantic notion of (im)perfectivity simply doesn't exist, and every language has a slightly different way of dividing up the conceptual space relating to time.

Now that you mention it, I feel that this is a very important point, and one that’s been confusing me until you explicitly pointed it out: that languages can have radically different conceptions of when something is ‘imperfective’ vs when something is ‘perfective’. However, I don’t think that this at all invalidates the existence of an underlying semantic notion of (im)perfectivity. I tend to think about these things as being prototype-based: there are prototypically imperfective events, and prototypically perfective events, but languages divide up the space between those prototypes in different ways. (Personally, I’ve always thought that this is true for a surprisingly large number of grammatical phenomena.)

(As for finding inspiration by reading grammars: I’ve already tried that, and all of the grammars I’ve seen so far — even those which supposedly focus on aspect — are regrettably light on the details I need.)

_______

Anyway, after thinking for a while about what everyone has said, here’s what I have so far:

• Semantically, a prototypical ‘imperfective’ event is one which is seen as being extended in time, for instance by continuing before and after the point of reference (progressive/continuous), or being performed repeatedly over a long duration (habitual).
• Semantically, a prototypical ‘perfective’ event is one which is seen as a single whole, ignoring any temporal details of the event other than its time of occurrence. In particular, a semantically perfective event can only be in the present tense if it is punctual; this is because a present non-punctual event must extend into the past and future, and so cannot be seen as a single whole while staying entirely inside the present.
• Languages generally contrast ‘perfective’ and ‘imperfective’ forms of verbs. When describing prototypical situations, their usage matches the prototypical semantic cases described above, but languages differ in how they divide up non-prototypical situations.
• Semantically, stative propositions are neither perfective nor imperfective. But languages tend to use the unmarked (or least marked) aspectual form when talking about these.

Does this seem correct? If so, then that clears up most of my problems around aspect. (I still have a couple of questions about bits I don’t understand, but before asking them I’d first like to know whether I’m understanding the basics correctly.)

[Vardelm]

Does this seem correct?

Yes, but I defer to those that are more experter than I is.

My only quibble/question would be:

• Languages generally contrast ‘perfective’ and ‘imperfective’ forms of verbs.

This might depend on how expansive "generally" is. My perception is that a lot, if not the majority, of languages have a perfective vs imperfective distinction. However, it seems like there are many where this is not the case. I don't know what the percentage is either way.

[Richard W]

[In fact, boundedness I would say is the number one characteristic of perfective. Bounded and unbounded might almost be better terms than perfective and imperfective. Perfective/bounded means the event in question is conceptualized as having a beginning, middle and end – i.e. bounded in time, a whole event. Imperfective/unbounded means you're not thinking of it that way. You're thinking of it as a) part of an event, something ongoing, most likely a part of the middle of the event, perhaps while something else was happening; b) as a state, a description of a general truth/state; or c) as something habitual, something that happens again and again. With the habitual sense, it's really a sequence of perfective events, but the focus isn't on them as bounded events but the fact that it keeps happening, almost like a segmented stative, so again, a description of a general truth/state.

How do you square this with the gnomic aorist of Ancient Greek?

[Richard W]

Semantically, some verbs are ill suited to perfective aspect and others ill suited to imperfective aspect. For example, semelfactive verb (i.e. a verb that refers to an instantaneous action, something that takes seconds or less) is generally not used as imperfective except maybe in the habitual sense. So "kick," "slap," "break," "let out a cough" will most likely always or nearly always be perfective. In English, you can imperfectivize these but this implies rapid, ongoing repetition. "I'm kicking him" means that I have kicked him several times, I'm still kicking him, and I haven't stopped yet. In many other languages, the verbs semantically refer to a single action so this kind of meaning has to be achieved with some other mechanism than a simple imperfective aspect, maybe a pluractional verbal number, frequentative derivation etc.

In English, the progressive tense is commonly used for any plain verb in a principal clause with future reference, as in "I'm signing the contract tomorrow". It's less common if the modal verb "will" or "shall" is used. (I probably haven't quite got the conditions right.) I don't know whether futurity stretches the time reference so as to make it imperfective - I think that's special pleading.

[Richard W]

Semantically, stative propositions are neither perfective nor imperfective. But languages tend to use the unmarked (or least marked) aspectual form when talking about these.

I think that's part of a general process of sometimes avoiding redundancy, like verbs with noun subjects not being marked for subject (Welsh and Persian spring to mind on that one), or no plural marking on nouns when numbers are used (Turkish and some English spring to mind).

[Curlyjimsam]

I would suggest that from a conlanging perspective the most important thing, probably, is how individual languages divide up their aspectual categories (which if you want inspiration is probably best gleaned from detailed grammars of particular languages). It's entirely possible that a universal underlying semantic notion of (im)perfectivity simply doesn't exist, and every language has a slightly different way of dividing up the conceptual space relating to time.

Now that you mention it, I feel that this is a very important point, and one that’s been confusing me until you explicitly pointed it out: that languages can have radically different conceptions of when something is ‘imperfective’ vs when something is ‘perfective’. However, I don’t think that this at all invalidates the existence of an underlying semantic notion of (im)perfectivity. I tend to think about these things as being prototype-based: there are prototypically imperfective events, and prototypically perfective events, but languages divide up the space between those prototypes in different ways. (Personally, I’ve always thought that this is true for a surprisingly large number of grammatical phenomena.)

(As for finding inspiration by reading grammars: I’ve already tried that, and all of the grammars I’ve seen so far — even those which supposedly focus on aspect — are regrettably light on the details I need.)

Yeah, I don't want to say perfectivity doesn't reduce to some universal underlying notion - I think it probably does - but we don't know for sure one way or the other, and in any case this is a theoretical question with minimal bearing on conlanging. (Though I realise now this is not the conlanging section of the forum.)

That's a shame about the grammars, though not terribly surprising. Perhaps people could suggest some that are better in this regard?

[bradrn]

Does this seem correct?

Yes, but I defer to those that are more experter than I is.

Thank you!

My only quibble/question would be:

• Languages generally contrast ‘perfective’ and ‘imperfective’ forms of verbs.

This might depend on how expansive "generally" is. My perception is that a lot, if not the majority, of languages have a perfective vs imperfective distinction. However, it seems like there are many where this is not the case. I don't know what the percentage is either way.

That’s pretty much exactly what I meant by “generally”.

Semantically, some verbs are ill suited to perfective aspect and others ill suited to imperfective aspect. For example, semelfactive verb (i.e. a verb that refers to an instantaneous action, something that takes seconds or less) is generally not used as imperfective except maybe in the habitual sense. So "kick," "slap," "break," "let out a cough" will most likely always or nearly always be perfective. In English, you can imperfectivize these but this implies rapid, ongoing repetition. "I'm kicking him" means that I have kicked him several times, I'm still kicking him, and I haven't stopped yet. In many other languages, the verbs semantically refer to a single action so this kind of meaning has to be achieved with some other mechanism than a simple imperfective aspect, maybe a pluractional verbal number, frequentative derivation etc.

In English, the progressive tense is commonly used for any plain verb in a principal clause with future reference, as in "I'm signing the contract tomorrow". It's less common if the modal verb "will" or "shall" is used. (I probably haven't quite got the conditions right.) I don't know whether futurity stretches the time reference so as to make it imperfective - I think that's special pleading.

That’s pretty interesting. (And resolves my confusion about ‘I’m not getting rid of ⟨q⟩’, a sentence I posted earlier.) Does anyone else have any idea as to how imperfective got extended from ‘event seen as being extended in time’ to ‘future event’?

Yeah, I don't want to say perfectivity doesn't reduce to some universal underlying notion - I think it probably does - but we don't know for sure one way or the other, and in any case this is a theoretical question with minimal bearing on conlanging.

I’m not too sure about that. If I want to make a realistic conlang, and I want it to have a perfective/imperfective contrast, then I’ll need to know what commonalities (if any) those concepts have between languages. If they do reduce to underlying notions, then I’ll have to define the semantics of that contrast by basing it on those underlying notions — but if they don’t, then I have a lot more freedom to do what I want with that contrast.

That's a shame about the grammars, though not terribly surprising. Perhaps people could suggest some that are better in this regard?

Yes please! I’d love to read a grammar with a thorough treatment of aspect (including which aspects there are, which situations they are used in, how they interact with tense etc.).

[zompist]

In English, the progressive tense is commonly used for any plain verb in a principal clause with future reference, as in "I'm signing the contract tomorrow". It's less common if the modal verb "will" or "shall" is used. (I probably haven't quite got the conditions right.) I don't know whether futurity stretches the time reference so as to make it imperfective - I think that's special pleading.

That’s pretty interesting. (And resolves my confusion about ‘I’m not getting rid of ⟨q⟩’, a sentence I posted earlier.) Does anyone else have any idea as to how imperfective got extended from ‘event seen as being extended in time’ to ‘future event’?

This is off the top of my head, but it's common for the future to be built out of morphological scraps-- e.g. the English and Romance future built from auxiliary verbs.

Hebrew and English both arguably started with a past / non-past distinction. The Hebrew "non-past" (yiqtol) had a wide range of uses: present, future, habitual past, modal. It developed into the future tense in modern Hebrew.

Pragmatically, you can emphasize how quickly something will happen by saying it's in progress, or even over. Think of the teenager saying "I'm cleaning my room!" to mean they'll start soon, or the employee answering "It's done!" when given a task. So long as humans are euphemistic creatures, other tenses are going to be used to promise to do something, and from there become future tenses.

Hindi shares the use of the progressive to refer to the near future.

I’m not too sure about that. If I want to make a realistic conlang, and I want it to have a perfective/imperfective contrast, then I’ll need to know what commonalities (if any) those concepts have between languages. If they do reduce to underlying notions, then I’ll have to define the semantics of that contrast by basing it on those underlying notions — but if they don’t, then I have a lot more freedom to do what I want with that contrast.

If I read a conlang grammar and it talks about an (im)perfective, I'll expect it more or less includes the things we've talked about ((un)boundedness in time, etc). Otherwise I'd ask why you're calling it the (im)perfective.

But you can add complications or pragmatic differences. My treatment of aspect in Verdurian may be helpful. It probably should be twice as complicated to be naturalistic, but I tried to add pragmatic uses (such as the use of a completive for the future), forms with multiple aspectual meanings, etc.

[bradrn]

Hebrew and English both arguably started with a past / non-past distinction. The Hebrew "non-past" (yiqtol) had a wide range of uses: present, future, habitual past, modal. It developed into the future tense in modern Hebrew.

If the Hebrew non-past developed into the future tense, then where did the present tense come from? As far as I can remember, Hebrew definitely has a present tense (although it’s a bit odd in that it doesn’t agree with the subject).

Pragmatically, you can emphasize how quickly something will happen by saying it's in progress, or even over. Think of the teenager saying "I'm cleaning my room!" to mean they'll start soon

Yes, I’m guilty of this…

My treatment of aspect in Verdurian may be helpful. It probably should be twice as complicated to be naturalistic, but I tried to add pragmatic uses (such as the use of a completive for the future), forms with multiple aspectual meanings, etc.

I’ll have a look at it, thanks! Generally I stongly prefer looking at natlang grammars for this sort of thing, but given that I haven’t yet found anything useful, I’m willing to read anything I can find on this topic.

[KathTheDragon]

If the Hebrew non-past developed into the future tense, then where did the present tense come from? As far as I can remember, Hebrew definitely has a present tense (although it’s a bit odd in that it doesn’t agree with the subject).

A construction involving the active participle, iirc.

[bradrn]

My treatment of aspect in Verdurian may be helpful. It probably should be twice as complicated to be naturalistic, but I tried to add pragmatic uses (such as the use of a completive for the future), forms with multiple aspectual meanings, etc.

I’ve finished reading this now. A couple of points I thought were interesting:

• As far as I’m aware, most elaborate aspectual systems have a perfective vs imperfective contrast at their core, so I found it interesting that Verdurian doesn’t have this contrast (and indeed, all of the aspects seem more or less imperfective).
• The system of aspectual adverbs reminds me of Chinese, which I found surprising given that the rest of Verdurian is more SAE-inspired.
• You may want to think about reconsidering the sample sentence I always ate at the Corona.

[Kuchigakatai]

In fact, boundedness I would say is the number one characteristic of perfective. Bounded and unbounded might almost be better terms than perfective and imperfective. Perfective/bounded means the event in question is conceptualized as having a beginning, middle and end – i.e. bounded in time, a whole event. Imperfective/unbounded means you're not thinking of it that way.

I have struggled with perfective vs imperfective as well. I think I mostly have it down, but there are still times with a bit of confusion. Seeing these labels of "bounded" vs "unbounded" is really helpful IMO, and jives with how I understand perfective vs imperfective.

When I went through my own struggle to understand the concepts of perfective/imperfective, I also came across this explanation of "bounded" vs. "unbounded", and hated it. It endlessly confused me. Although I think the choice of words was a good chunk of my problem: I think I would've more easily understood the metaphor if "limited" vs. "ongoing" had been used...

Imperfective_m v. perfective_m is fairly standard for Russian and Semitic morphology, and in other languages the related tense labelling is very varied. However, where the difference maps to morphology, I get the impression that the boundary between the two varies from language to language. Certainly English seems not to comply with the European norm when it comes to the choice of past tenses, but I don't think it's just a case of English v. the rest. As to the terminology, I would also use it for the basis to the difference between present system imperative and the aorist imperative in Greek, ancient and modern.

Hah! You're right about Russian. And I should've thought of that, considering I mentioned Comrie's book... Also, perfective_m/imperfective_m aren't normally used in grammars of Biblical Hebrew and Standard Arabic for the TAM forms, so I guess you're talking about general Semitic linguistics (which I know nothing about).

By the way, Joüon's grammar of Biblical Hebrew (and I'm talking about Muraoka's edition) is a bit funny in this regard. First, he uses "future_m" for the imperfect_m, and then, in general, when he's just mentioning a morphological form he often uses a label (so: perfect, inverse future, cohortative, infinitive absolute), but when the difference between morphological categories and semantics matters, then he makes sure to switch to using samples for the morphological forms (so: qatal, wayyiqtol, w-’eqtla, qatol).

(As an aside, I find samples very useful when talking to people who don't know grammatical terminology actually, like when I talk about language stuff with other Spanish native speakers. It's so much easier to use "que ame, haga" instead of trying to get the listener to remember what verbal forms el presente del subjuntivo refers to.)

Yeah, I don't want to say perfectivity doesn't reduce to some universal underlying notion - I think it probably does - but we don't know for sure one way or the other, and in any case this is a theoretical question with minimal bearing on conlanging. (Though I realise now this is not the conlanging section of the forum.)

I think it does in its definition as a semantic concept, only breaking down because of the type of action you're describing, not the individual language. Punctual actions are hardly worth talking about as imperfective, unless, maybe, we're talking about a period of time when the action happens repeatedly (so, an iterative in the imperfective). And I'd say there is an important need to have terms that can work across languages.

(As for finding inspiration by reading grammars: I’ve already tried that, and all of the grammars I’ve seen so far — even those which supposedly focus on aspect — are regrettably light on the details I need.)

That's a shame about the grammars, though not terribly surprising. Perhaps people could suggest some that are better in this regard?

I think grammars tend to do a pretty bad job when providing the many meanings that can be expressed with the various forms or constructions available. They just work with very wide generalities.

(I am reminded of this article-like paper (in French), which is a beautiful rant about how reference grammars of French for the most part fail to adequately talk about the interaction of the functional word depuis (and some three other ones), aktionsart (punctual/semelfactive, activity, accomplishment) and aspect (perfective, imperfective) whether they use fancy terms or not, which means they do end up saying wrong things about how depuis is used.)

I would suggest that from a conlanging perspective the most important thing, probably, is how individual languages divide up their aspectual categories (which if you want inspiration is probably best gleaned from detailed grammars of particular languages). It's entirely possible that a universal underlying semantic notion of (im)perfectivity simply doesn't exist, and every language has a slightly different way of dividing up the conceptual space relating to time.

Now that you mention it, I feel that this is a very important point, and one that’s been confusing me until you explicitly pointed it out

I just wanted to mention Richard W mentioned it first (and I reacted to it saying it was a good point). Maybe you missed that part...

[Vardelm]

When I went through my own struggle to understand the concepts of perfective/imperfective, I also came across this explanation of "bounded" vs. "unbounded", and hated it. It endlessly confused me. Although I think the choice of words was a good chunk of my problem: I think I would've more easily understood the metaphor if "limited" vs. "ongoing" had been used...

I've seen this while participating in all kinds of different activities. Different words click with different people. It's important for people to not get hung up on their own very specific, technical definition of a word and be able to think of how else that word might be used or words that could be substituted to get the idea across.

[bradrn]

I would suggest that from a conlanging perspective the most important thing, probably, is how individual languages divide up their aspectual categories (which if you want inspiration is probably best gleaned from detailed grammars of particular languages). It's entirely possible that a universal underlying semantic notion of (im)perfectivity simply doesn't exist, and every language has a slightly different way of dividing up the conceptual space relating to time.

Now that you mention it, I feel that this is a very important point, and one that’s been confusing me until you explicitly pointed it out

I just wanted to mention Richard W mentioned it first (and I reacted to it saying it was a good point). Maybe you missed that part...

Are you talking about this post?

I think it's best to say that the English simple present is protoypically habitual. That's easiest to see with sentences like "I read linguistics" or "I live in Barcelona."

I think it's more useful to note that the English progressive tenses are 'marked', so a verb can have the 'perfective' form simply because making it progressive doesn't add anything to the meaning. This then covers the reluctance of the 'stative' verbs to take a progressive, and eases the fact that whether 'to be good' uses the progressive depends on its meaning. There are a few idioms that require the progressive, and they mess up the test for an English verb being stative.

Thus, if one wants to think of the 'simple' forms in English as 'perfective', defining the 'perfective' as 'not imperfective' makes very good sense. Note also that different languages use the imperfective for different meanings.

If so, I’m not sure it’s saying exactly the same thing. And also, given that this was a point that I missed, I think it was useful for me that Curlyjimsam explicitly said ‘individual languages’ division of imperfective and perfective can vary wildly’.

[bradrn]

I’ve been thinking about this topic a bit more since my last post, and I have two more questions now:

1. We have already established that there exist semantically ‘prototypical’ perfective and imperfective events, which always get marked with the perfective and perfective aspects. But does anyone have any more information about how languages assign aspect to non-prototypical events? (e.g. the stative always gets assigned to the unmarked aspect, if there is one; I’m particularly interested in similar generalisations.)
2. Some languages have aspectual systems which are not based around a perfective/imperfective contrast. (I was reading about one such language earlier today — Northeast Ambae, spoken on Vanuatu.) In that case, how do these sorts of systems tend to be organised, and what sort of aspects do they tend to have instead of a perfective and/or imperfective?

[Vardelm]

In that case, how do these sorts of systems tend to be organised, and what sort of aspects do they tend to have instead of a perfective and/or imperfective?

I've seen a few that use perfect (retrospective), continuous/progressive, and prospective aspects. This is essentially a past, present, & future tense system, but because it's aspect it's relative to the time of the event being discussed, rather than relative to the time of speaking.

[bradrn]

In that case, how do these sorts of systems tend to be organised, and what sort of aspects do they tend to have instead of a perfective and/or imperfective?

I've seen a few that use perfect (retrospective), continuous/progressive, and prospective aspects. This is essentially a past, present, & future tense system, but because it's aspect it's relative to the time of the event being discussed, rather than relative to the time of speaking.

That’s a pretty interesting system! But I don’t think that’s quite what I’m looking for — to me, that seems like more of a variation on a tense system than an aspectual system per se. Sure, it’s not actually tense, but I don’t think calling it aspect is quite right either.