Many conlangs contain tense, some contain aspect as well or instead, and much of the recommended reading for conlangers assumes that you know what tense and aspect are and what the difference between them is. What actually is the difference, and how would a word or morpheme be classified as a marker of one or the other?
When discussing tense, aspect, and mood, it's important to distinguish a given language's grammatical markers from the abstract concepts being described. Thus, linguists use the words temporal reference and aspectual reference to describe the abstract ideas being described, while the words tense and aspect are reserved for when such reference is marked grammatically.
The simplified explanation for tense and aspect usually given is that tense describes when an event occurs and aspect describes the internal temporal structure of an event. However, this doesn't give a ton of easy answers to "is this grammatical form marking tense, aspect, or both?" Formal semanticists actually have a relatively simple framework for how temporal reference and aspectual reference are defined, the basic concepts of which can be helpful when designing your conlang's own tense/aspect system.
The Neo-Reichenbachian Model
Imagine a timeline:
Earlier Later <-------------------------------------------->
Now, imagine there are two spans on that timeline (the fact that they're spans is important -- nothing is truly instantaneous). The time when you make an utterance is called the utterance time (or UT), and the time that you're talking about is called the topic time (or TT).
Temporal reference describes the relationship between the TT and UT. Past temporal reference is when the TT precedes the UT; future temporal reference when the TT follows the UT; and present temporal reference when they overlap.
Past: <--------------------------------------------> |--TT--| |--UT--| Future: <--------------------------------------------> |--UT--| |--TT--| Present: |--UT--| <--------------------------------------------> |-------TT-------|
This seems pretty intuitive -- if the section of the timeline you're talking about precedes the time you make the utterance, you'll use the past tense; if it follows it, you'll use the future; and if they overlap, you'll use the present. Simple, right?
But what of aspect? Well, aspectual reference isn't as transparent at that. To describe aspectual reference, we'll need to pick out a third span that timeline -- the time when the event being described actually happened. We'll call this the eventuality time (or ET).
"But wait!" you might object. "Isn't that the same as the topic time? If I'm describing something that happened, surely the time I was talking about is the same as the time that it happened." However, there is a difference, and it's exactly this difference that gives rise to aspectual reference. Consider the English past perfect in a sentence like the following:
Yesterday he had gone to the store.
The English past tense includes past temporal reference, and we've established that this means that the topic time is before the utterance time -- something which is explicitly confirmed by the temporal adverb "yesterday" giving us a more specific topic time. But when does the him-going-to-the-store happen? Probably not yesterday, right? Well, this is because perfect† aspectual reference is when the eventuality time precedes the topic time -- aspectual reference describes the relationship between the eventuality time and the topic time. It for this reason that the past perfect is used by English speakers for the past-in-past.
Past Perfect: <--------------------------------------------> |--ET--| |--TT--| |--UT--|
But you can also have the present perfect, right? Well, let's combine our descriptions of the present tense and the perfect aspect and see what we come up with.
Present Perfect: |--UT--| <--------------------------------------------> |--ET--| |-------TT-------|
Here, the topic time and utterance time overlap, which is why you can't say something like "Yesterday he has gone to the store." However, the eventuality time precedes the topic time, so you're describing an event that happened in the past in relation to the present. And indeed, this is how we see the present perfect used in English.
This is far from the most interesting thing aspectual reference can do, however. Let's discuss the popular aspectual references, perfective and imperfect. For those who don't know, perfective and perfect are two different things. Perfective is often described as simply viewing an event as a point on a timeline, and imperfect as viewing it as a span. How do we define this as a relationship between eventuality time and topic time?
Well, a simple way people do it (ignoring a few other related thorny semantic problems that few here will care about) is by defining the perfective as when the eventuality time is contained within the topic time and the imperfect as when the topic time is contained within the eventuality time.
Past Perfective: |--ET--| <--------------------------------------------> |-------TT-------| |--UT--| Past Imperfect: |--TT--| <--------------------------------------------> |-------ET-------| |--UT--|
This results, at least formally, in the same sort of perspective shift we try to intuitively describe as changing the event from a "point" to a "span".
Note that tense and aspect vary between languages a lot, and even related languages can't be counted upon to have the same semantics here. After all, the French passé composé, while very similar in form to the English present perfect, does not share the same semantics (it's more of a perfective). Additionally, while a given language's system of tense and aspect will often combine a tense and an aspect in a certain inflection (such as the perfective past tenses in many European languages), this doesn't mean the two are wed cross-linguistically.
In fact, I hope this answer has given some conlangers the tools to try and invent their own tenses, aspects, or combinations of the two that they haven't themselves encountered cross-linguistically. Be creative!
- Temporal reference is the relationship between the utterance time (when you say something) and the topic time (the time relevant to the discussion at hand).
- Aspectual reference is the relationship between the topic time (see above) and the eventuality time (the time when the event being described actually occured).
- Tense is the grammaticalization of temporal reference.
- Aspect is the grammaticalization of aspectual reference.
- Things called "tenses" in natural languages (as well as in conlangs, to be honest) are often conflations of tense and aspect (and sometimes even mood, but we're not here to talk about that).
†When I refer to perfect aspectual reference, I'm actually more specifically describing what is known in semantics as the existential perfect. Perfect forms in English and many other languages include other types of perfect aspectual reference (universal, resultive, etc.) that differ subtly in their semantics, and other aspectual and temporal baggage as well; this is, however, probably not particularly interesting or relevant to most people here at this moment.
The difference is that tense refers to the time an action (or state or phenomenon) happened:
I was slim.
I am fat.
I will be fatter.
while aspect refers to the way an action (or state, or phenomenon) develops along time:
I eat sushi. (habitual)
I am eating sushi. (continuous)
I used to eat sushi. (habitual)
I ate sushi. (perfect)
I was eating sushi. (continuous)
I will have eaten sushi. (perfect)
I will be eating sushi. (continuous)
I don't know if someone has a complete list of possible aspects, but here are some possibilities:
- The action was/is/will be finished (perfect)
- The action is/was/will be ongoing (continuous)
- The action is/was/will be continually performed (habitual)
- The action is/was/will be going to start (inceptive)
- The action is/was/will be intermitently performed (iterative)
That's the difference. Now comes the confusion - because that is what languages do, they confuse tense and aspect, in incoherent and incomplete systems that fail to either integrate or clearly separate those features.
One first source of confusion is that many languages operate not only with the simple distinction between tenses as outlined above - past, present, future - but also with dependent and independent tenses. For instance, I may say
I used to eat sushi when Susan introduced me to Javanese food.
I am eating sushi while the president threats nuclear war against North Korea.
I will be eating sushi when they bring the wine.
in which one action is past, present or future regarding the time of the speech (Susan introduced me to Javanese food in the past, the president is threatening nuclear war in the present, they will bring wine in the future) while the other (me eating sushi) happens in the present - not the present regarding the time of speech, however, but the present regarding the other action.
And while this may be a source of confusion by itself, it expands to a further level of confusion because the interaction between dependent and independent action often has implications on the aspectuality of each of them. For instance, I say
I eat sushi. - or
I am eating sushi.
it is understood that these are different aspects: in the first sentence, I am saying that I do habitually eat sushi, not that I have a dish of sushi in front of me. In the second sentence, however, I am saying that I am actually eating sushi in the moment I am talking, so the sentences are not synonim to each other.
But if I say
Go there and stop the president from doing bullshit while I eat sushi. - or
Go there and stop the president from doing bullshit while I am eating sushi.
the phrase "eat sushi" loses its habitual aspect, and turn into continuous, and the sentences are effectively synonimous.
The there is the confusion that Sparksbet remarks, between "tense" as a grammatical form, and "tense" as a reference to actual time. As she says, "tenses" as grammatical forms are used to denote a mix of tense references and aspectual references.
Furthermore, while the tense and aspectual references can be several, their grammaticalization, or lack thereof, is much more constrained. For instance, neither English, nor any other Western European language that I know has verb inflections for iterative or inceptive aspects. They instead use either syntactical devices, or morphological derivation - and those often vary weirdly from lexical item to lexical item. Let's take for instance Portuguese:
Eu como sushi (habitual)
Estou comendo sushi (continuous)
Estou beliscando uns sushis (iterative) (where a different verb is used to make the aspectual reference)
Eu bebo uísque (habitual)
Estou bebendo uísque (continuous)
Estou bebericando uísque (iterative) (where a morphological derivation of "beber" is used to make the aspectual reference)
So, probably no natural language has a coherent tense-aspect system; the fact that some tenses seem to imply some aspects when put in relation to others probably means that, for reasons of economy, grammatical "tenses" will be used irregularly, given internal or even external context.
Could it be possible to build a conlang with a complete and coherent tense-aspect system? Perhaps, but I would say such coherence and completeness are only going to survive as long as the language is not effectively used by a human population.