Tenses

In grammar, tense can mean two things^{(Sarkar, 1998:92–93)}:

1. A temporal reference found in a predicate—past, present, future.
2. The morphology of a word required to express a temporal reference—the conjugation of a verb to past, present, and future tenses.

If we go by the second definition, neither English nor Japanese have a future tense, since there's no verb form that exclusively expresses a future temporal reference.

• In English
• In Japanese
  • Nonpast
  • Future Statives
    • Futurate Statives
    • Eventivizers
  • Aspectual Markers
• Discrepancies
  • Discourse-Level Futurates
  • Progressive Futurates
  • Pluractionality Ambiguity
  • Present Perfect vs. Past Perfective
    • Perfect and Temporal Adverbs
  • Sequence of Tenses
    • Absolute Tenses
    • Reported Speech
    • Tensed Adjectives
    • Past of Future
    • Qualified Temporal Adverbials
    • Relative Tense of Tenseless
  • Momentaneous Observations
    • Qualified Futures
    • Momentaneous Habituals
    • Habitual vs. Stage Topicalization
  • Narrative Present
  • Imperatives
  • Performative Verb Reportative Futurity
  • Event Cancellation
  • Counterfactuals
• References

In English

English has a present tense "do" and a past tense "did."

The word "does" means the same thing as "do," except it's used when the subject is in third person singular. Observe:

• I do.
• You do.
• *He do. (wrong.)
• He does. (right.)
• They do.

Therefore, "does" is of the same tense as "do"—it's used instead of "do" for reasons unrelated to temporal reference, so, tense-wise, it's the same thing.

In order to express a future temporal reference in English, we use the auxiliary "will."

• I will do.
• He will do.

Above, the word "do" is in present tense morphologically, but the predicate "will do" is in the future as far as temporal reference is concerned.

It's worth noting that "do" is sometimes said to be a nonpast tense^{(Moens, 1987:10)}, rather than a present tense. This is due to "do" being available in futurates such as below:

• I do it tomorrow.

However, given that the word "do" can't express a future temporal reference by itself, it's simpler to say that "do" is only present tense, and the future temporal reference of the sentence above is provided by the adverb "tomorrow" instead.

Such sentence structured is called a "futurate."

Prince ms. 1973 uses the term futurate for present-tense sentences that can occur with future time adverbials.^{(Goodman, 1973:76)}

Although futurates are defined by the use of the present tense, there are reasons to believe they're not exclusive to the present tense. Observe the sentences below:

• *I will did it tomorrow.
  • Ungrammatical at syntax level.
• #I did it tomorrow.
  • Grammatical, but infelicitous.

It doesn't make sense to say I did something tomorrow, since that would place tomorrow in the past, and tomorrow is in the future.

However, that's only if you have a puny human understanding of the space-time continuum. A sufficiently intelligent life-form transcends the concept of linear time.

• Context: we time traveled to fix your mistake.
• Okay, we went one week back in time. When did you make the mistake?
• I did it tomorrow.

We can understand from the above that words like "tomorrow" and "later" refer to the time on calendar and clock, while the tense used in speech refers to the time as experienced by the speaker.

Without time travel, these two things would always match.

Source: japanesewithanime.com (CC BY-SA 4.0)

The past tense often has terminating implicatures, some of which are known as "lifetime effects." For example:

• John was a teacher.

The sentence above can mean three different things, depending on how we limit the period of time John was a teacher.

• In life, John was a teacher, which means right now we're after John's life ended, i.e. John is dead.
• Previously, John was a teacher, which means John is no longer a teacher right now.
• At the time we're talking about, e.g. in January of last year, John was a teacher. That doesn't mean John stopped being a teacher in February, or that John is dead, it just means that, in January, yes, he was indeed a teacher at that time.

Not all clauses have tenses.

In phrases such as "I have done" and "I am doing," the verbs "have" and "am" are tensed, they're in present tense, but the participles "done" and "doing" are tenseless.

Since they're tenseless, they don't run into temporal conflicts when used with either tense: you can say "I am doing" and "I was doing," and "doing" won't have a problem with the tense of "am" and "was."

Grammatical tense is separate from grammatical aspect.

For example, given appropriate context, it's possible to have all sorts of aspects in all sorts of tenses:

Aspect	Present tense	Past tense
Habitual	I watch anime.	I watched anime.
Iterative	I watch two anime.	I watched two anime.
Perfective	I watch an anime.	I watched an anime.
Perfect	I have watched an anime.	I had watched an anime.
Progressive	I am watching an anime.	I was watching an anime.

Note: the table above is merely illustrative, and different contexts will have different aspects for the same sentence. For example, if someone asks "what did you do yesterday?" and you answer "I watched anime," that would be the iterative aspect, not the habitual aspect.

Grammatical tense and aspect aren't the only categories that exist in grammar. There is also mood, or modality, for example.

This article will focus mostly on tense and temporal reference, which is honestly an extremely complicated topic. There are a lot of weird terms that you probably won't understand at first, but there will be examples and drawings to help make sense of it.

In Japanese

Japanese has a nonpast tense, suru する, "do," "does," "will do," and a past tense, shita した, "did."

They're known by terms based on the tenses that they express, which sounds good at first glance, but it's actually an unholy mess:

• hikakokei
  非過去形
  Nonpast form, which is sometimes used in the past, because why not.
• kakokei
  過去形
  Past form, which sometimes used in the present, because why not.

They're also known by terms based on their morphology, which is an even unholier, messier, bloody, cursed mess:

• ru-kei
  ル形
  ru-form, because it ends in ~ru ～る, except not really, as ending in ~ru applies mostly only to ichidan verbs.
  • Also known as ~u form in English, because it ends in ~u in most godan verbs, except that's not quite right, either, because i-adjectives end in ~i ～い in the ru-form, and na-adjectives end in the da だ copula in the ru-form when it's also the "predicative form," shuushikei 終止形, but end in the na な copula in the "attributive form," rentaikei 連体形, while nouns in the attributive end in the no の copula, though this only applies to a subset of no-adjectives.
    
    • Not to be confused with the "u-form," u-kei ウ形, in Japanese, which is known as the volitional form in English, and it's called so because it actually ends in the ~u ～う jodoushi 助動詞 attached to the mizenkei 未然形, like in shiyou しよう, "let's do."
  • Also known as hitakei 非タ形, "non-ta-form," which makes fewer assumptions about the morphology, but sounds kinda broad.
• ta-kei
  タ形
  ta-form, because it ends in ~ta ～た, except not really, because the ~ta ～た jodoushi becomes ~da ～だ in certain verbs due to renjoudaku 連声濁.

It doesn't matter which one the two name pairs above you choose, ultimately, there are only two morphological tenses in Japanese. One expresses the past and ends in ~ta or ~da, while the other does not.

For reference, an example containing both:

Game: Dragon Quest III, Doragon Kuesto Surii ドラゴンクエストＩＩＩ

• *"Neruson-san wo nakama ni
  kuwaeru no ne. wakatta wa.
  ＊「ネルソンさんを　仲間に
  くわえるのね。わかったわ。
  *"Add Nelson to your party, [right]? [I got it.]
  • kuwaeru - nonpast.
• neruson ga nakama ni kuwawatta!
  ネルソンが　仲間に　くわわった！
  Nelson was added to the party!
  • kuwawatta - past of kuwawaru くわわる.

Nonpast

The term "nonpast" includes present and future temporal references, however, the Japanese nonpast form isn't universally nonpast. It discriminates words by their lexical aspect.

• Eventive verbs can be present or future.
• Stative verbs can only be present.

When an eventive verb is in nonpast form, it expresses either:

1. A futurity: an event will occur in the future.
2. A habitual: an event occurs.

Which one exactly depends on context.

An example:

• Tarou wa tabako wo suu
  太郎はタバコを吸う
  Tarou smokes cigars. (present habitual.)
  Tarou will smoke the cigar. (future perfective.)

Statives can't express futurities. This includes lexical statives such as adjectives, stative verbs, and habitual predicates.

Note: Bertinetto (1994:406–413,n19; & Lenci, 2010:14) distinguishes habituals from "attitudinals" and "potentials" in that habituals aren't stative, but attitudinals are. In which case, Japanese habituals are primarily attitudinal, given that the syntax used with them resembles statives.

With statives, the nonpast form is actually just a present form. Observe:

1. Adjectives:
   • neko wa doubutsu da
     猫は動物だ
     Cats are animals. (gnomic present.)
     *Cats will be animals. (no future.)
     (kind-level predicate.)
   • ocha ga atsui
     お茶が熱い
     The tea is hot. (episodic present.)
     *The tea will be hot. (no future.)
     (stage-level predicate.)
2. Stative verbs:
   • Tarou wa {kanji ga yomeru}
     太郎は感じが読める
     {Kanji is readable} is true about Tarou. (double subject construction.)
     Tarou {is able to read kanji}. (potential, gnomic present.)
     *Tarou {will be able to read kanji}. (no future.)
     (individual-level predicate.)
   • Tarou wa sou omou
     太郎はそう思う
     Tarou thinks so. (stative cognitive, gnomic present.)
     *Tarou will think so. (no future.)
     (individual-level predicate.)
3. Habituals, which are stativized eventive verbs, as far as Japanese syntax is concerned:
   • Tarou wa tabako wo suu
     太郎はタバコを吸う
     Tarou smokes cigars. (present habitual.)
     *Tarou will smoke cigars. (no future.)
     (individual-level predicate.)
   • Pengin wa tobanai
     ペンギンは飛ばない
     Penguins don't fly. (present habitual-potential.)
     *Penguins won't fly. (no future.)
     (kind-level predicate.)

Future Statives

There are two ways to use a stative in the future in Japanese: through futurates, and by making statives eventive.

Futurate Statives

Futurates, predicates in the present tense that feature adverbs referring to the future, exist in both English and Japanese. Observe:

• Jon wa Saigo ni Shinu
  ジョンは最後に死ぬ
  John Dies at the End.
  • Here, "dies" is in the present tense, but "at the end" has a future temporal reference, so we imagine that John isn't dead yet, but that he "will die" at the end.

In Japanese, there's basically only one case in which futurates are commonly used to make statives future:

• gakkou wa yasumi da
  学校は休みだ
  The school is resting. (literally.)
  There's no school. (English translation.)
  • yasumu
    休む
    To rest. To stop doing an activity for a while. To take a break.
• ashita gakkou wa yasumi da
  明日学校は休みだ
  Tomorrow, the school is resting. (literally.)
  Tomorrow, there's no school.

Eventivizers

It's possible to convert a stative into an eventive, in which case it gains the ability to express a future temporal reference just like any other eventive.

This is done through the eventivizers naru なる and suru する, which are modified by an adverb, meaning that the stative must be in its adverbial form. For example:

• sekai ga tanoshii
  世界が楽しい
  The world is fun.
  • tanoshii - adjective, therefore stative, therefore only present.
• sekai ga tanoshiku naru
  世界が楽しくなる
  The world will be fun.
  The world will become fun.
  • naru - eventive, therefore it has a future.
• sekai wo tanoshiku suru
  世界を楽しくする
  To make the world be fun.
  To make the world become fun.
  To make the world fun.
  
  • This is missing a causer, who causes the world to become fun.
• Tarou ga sekai wo tanoshiku suru
  太郎が世界を楽しくする
  Tarou will make the world fun.
  • suru - lexical causative eventive, therefore it has a future.

Different words have different adverbial forms. For i-adjectives, as above, it's ~ku ～く, but for other adjectives it's the ni に adverbial copula.

• shizuka da
  静かだ
  [It] is quiet.
• shizuka ni naru
  静になる
  [It] will be quiet.
• shizuka ni suru
  静かにする
  To make [it] quiet.
• shizuka ni shinasai
  静かにしなさい
  Make [it] quiet!
  • This is an imperative, for example: a teacher ordering students to make the classroom quiet, or to make themselves quiet, in other words:
  • damarinasai!
    だまりなさい！
    Stop talking!
    Be silent!

The adverbial form of verbs is achieved through the ~you da ～ようだ jodoushi. The ~you da attaches to the verb, and the predicative ~da is conjugated to the adverbial ~ni, turning the suffix into ~you ni ～ように.

• Tarou wa {kanji ga yomeru} you ni naru
  太郎は漢字が読めるようになる
  Tarou will be {able to read kanji}. (future potential.)
• Tarou wa {tabako wo suu} you ni naru
  太郎はタバコを吸うようになる
  Tarou will {smoke cigars}. (future habitual.)
  Tarou will start {smoking cigars}. (inchoative translation.)
• Tarou wa {sou omou} you ni naru
  太郎はそう思うようになる
  Tarou will {think so}. (future cognitive.)
  Tarou will start {thinking so}.
• Pengin wa {tobu} you ni naru
  ペンギンは飛ぶようになる
  Penguins will {fly}. (future habitual-potential.)
  Penguins will become {able to fly}.
  • Note: if we say "penguins fly," that entails "penguins can fly," because if they couldn't fly, they wouldn't fly. Therefore, if we say "penguins will fly," that entails "they will become able to fly."

When the ~naru in ~you ni naru is in nonpast form, it means the stative will become true in the future. If we conjugate ~naru to past form, ~natta ～なった, it means the stative has become true in the past. Observe:

• Tarou wa {tabako wo suu} you ni natta
  太郎はタバコを吸うようになった
  Tarou became in a way [that] {smokes cigars}.
  Tarou started {smoking cigars}.
  
  • {smokes cigars} - present habitual subordinate clause.
  • Became - past perfective matrix clause.

The causative can be used too:

• Tarou wa {tabako wo suwanai} you ni shita
  太郎はタバコを吸わないようにした
  Tarou made {[himself] not smoke cigars}.
  Tarou decided not to smoke cigars.

Using the causative in the negative has certain complexities.

The phrase ~nai you ni suru ～ないようにする is used when the causer indirectly causes the outcome, while ~naku suru ～なくする is used when the causer directly cause the outcome, e.g. when they physically make it happen^{(池上, 2002:1)}.

Syntactically, ~you ni comes after a predicative clause containing a causee subject marked by the ga が particle, which means the causer causes the whole clause to happen, while ~naku will have the causee marked by the accusative wo を particle, which means the causer causes the predicate to happen to the causee.

A couple of examples^{(池上, 2002:11)}:

• watashi wa {musuko ga heya kara denai} you ni shita
  私は息子が部屋から出ないようにした
  I made [it] so [that] {[my] son doesn't leave the room}.
  I made [it] so [that] {[my] son can't leave the room}. (potential entailment.)
  • For example: by locking the door, which would indirectly result in him not leaving the room.
• ?watashi wa musuko wo {heya kara denaku} shita
  私は息子を部屋から出なくした
  (same meaning?)
  • Since whether the son leaves the room or not depends on the son's volition, and you can't directly alter their volition, this doesn't make sense.
• watashi wa te-ashi wo shibatte kare wo {ugokenaku} shita
  私は手足を縛って彼を動けなくした
  I tied [his] arms-and-legs, making him {unable to move}.
• watashi wa te-ashi wo shibatte {kare ga ugokenai} you ni shita
  私は手足を縛って彼が動けないようにした
  I tied [his] arms-and-legs, making [it] so [that] {he can't move}.

Aspectual Markers

Besides nonpast and past, or ru-form and ta-form, Japanese also has the ~te-iru ～ている form and the ~te-aru ～てある form, which are essential to understand how tense-aspect works in Japanese.

However, these two other forms are aspectual markers, not tense markers^{(Sugita, 2009:1, 庵, 2001:76, 近藤, 2018:19–20)}.

It's possible to combine one with the other. Observe:

Aspect Markers vs. Tense Markers

	∅	~ta ～た
∅	suru する	shita した
~te-i~ ～てい～	shite-iru している	shite-ita していた
~te-a~ ～てあ～	shite-aru してある	shite-atta してあった

Morphologically, ~te-aru and ~te-iru can be divided into two parts: ~te ～て, or ~de ～で^(renjoudaku), which is the affix in the te-form of words, and ~aru and ~iru, which are auxiliary verbs, specifically hojo-doushi 補助動詞.

These two auxiliary verbs, ~aru and ~iru, derive from the two main verbs aru ある and iru いる, which are existence verbs, as they're primarily used to say whether something exists or not, and where it exists.

The auxiliary and the main verbs have almost nothing to do with each other. The meaning is different, the usage is different, they're basically completely different things.

However, they're fundamentally related. Or at least I think so.

In summary, the auxiliaries ~aru and ~iru make a verb behave, in certain ways, like the main verbs aru and iru.

Since the main verbs aru and iru already behave like themselves, they can't be conjugated to the ~te-aru and ~te-iru forms.

• *atte-iru, *atte-aru
  あっている, あってある
  (wrong.)
• *ite-iru, *ite-aru
  いている, いてある
  (wrong.)

By the same principle, it's impossible to conjugate ~te-iru to ~te-aru form or vice-versa, or conjugate them to themselves.

• *~te-atte-iru, *~te-atte-aru
  ～てあっている, ～てあってある
  (wrong.)
• *~te-ite-iru, *~te-ite-aru
  ～ていている, ～ていてある
  (wrong.)

The meaning of the existence verbs iru and aru is identical. The only difference is that iru must be used if the subject is animate, like a person or an animal.

For example^{(Sugita, 2009:3)}:

• Mari ga heya ni iru
  マリが部屋にいる
  Mary is in the room.
  • Mary is a person, so iru is used.
• hon ga tsukue no ue ni aru
  本が机の上にある
  The book is on the desk.
  • A book is inanimate, so aru is used.

Animacy means agency, the ability to act on its own. Things that are animate can do things on their own, and they can go to places on their own. Inanimate things can not.

In particular, a car can move, but it can't move on its own, so aru would be used. A vacuuming robot, e.g. a roomba, can move on its own, but it has no real agency, it merely follows a simple program, so aru would be used.

A sufficiently intelligent android, however, may have agency and thoughts of its own, doing things by their own volition, so if you treat it as a person, iru would be used.

This animacy requirement of iru isn't inherited by ~te-iru.

• hon ga moete-iru
  本が燃えている
  The book is burning.
  • This is perfectly valid, even though the book isn't animate.

Existence verbs are incompatible with generic sentences, due to them forcing the actualization of the subject.

To elaborate, observe the sentence below:

• yunikoon wa {tsuno no haeta} uma da
  ユニコーンは角の生えた馬だ
  An unicorn is a horse [that] {grew a horn}.
  Unicorns are horses [with] {horns}.

Unicorns don't exist, and yet we can talk about them. Why is that?

It's because we're talking about the abstract concept of an unicorn, rather than a particular, actual unicorn.

Such sentences which predicate over abstractions of entities are called kind-level predicates. Kind-level predicates can occur with habituals, as habituals are abstractions of events^{(Krifka, 1995:4)}.

Observe:

• hiiroo wa hito wo tasukeru
  ヒーローは人を助ける
  Heroes help people.

The sentence above can be uttered even if heroes don't exist, because we're talking about the concept of a "hero," rather than any particular heroes.

• Context: we live in a world without heroes.
• Heroes don't exist! Heroes help people, and nobody helps anybody in this godforsaken Earth!

Observe how the sentence above isn't contradictory.

Simply saying that "heroes help people" doesn't entail the existence of heroes.

Also, observe that if I said "John watches anime," we understand that John has watched anime at least once.

However, if I say "heroes help people," and heroes don't exist, then the "help" event never happens, not even once.

People can't be helped by non-existential heroes.

By contrast, observe the example below:

• hiiroo wa hito wo tasukete-iru
  ヒーローは人を助けている
  Heroes are helping people.

In the sentence above, we've actualized the event "help." It's no longer a generic habitual. It's actually happening right now.

Consequently, the word "heroes" no longer refers to an abstraction, to a concept. Instead, it refers to a bunch of existing instances of heroes, which we can't determine, but sure enough must exist.

After all, people can't be helped by non-existential heroes. If they ARE being helped RIGHT NOW, then the heroes helping them MUST exist.

Similarly:

• hiiroo wa hito wo tasukete-aru
  ヒーローは人を助けてある
  Heroes have helped the people.

Again, if heroes have done something in the past, then there must have been instances of heroes who must have existed in the past, otherwise they couldn't have participated in the event.

In English, the present progressive and the present perfect actualize events. In Japanese, the ~te-iru and ~te-aru forms do it.

Kind-level predicates are incompatible with actualized events, because actualized events require particularized participants, and kinds aren't particular.

They're incompatible with aru and iru, too, because they express the subject must actually exist somewhere in space. Things that exist physically in space must also exist in time. Kinds, being abstract, are incompatible with spacetime altogether.

The verb sonzai suru 存在する, which also translates to "to exist," doesn't have the same restriction since it has no spatial component:

• yunikoon wa sonzai suru
  ユニコーンは存在する
  Unicorns exist.
  • This is a kind-level predicate, it doesn't refer to any particular unicorn, in any particular place.
• yunikoon wa asoko ni iru
  ユニコーンはあそこにいる
  The unicorn "exists" over there.
  The unicorn is over there.
  
  • This is an individual-level predicate, it requires a particular unicorn to occupy space "over there," and, also, to occupy time.

We can observe that this existential entailment is inherited by ~te-iru.

The ~te-iru form can't be used with kind-level predicates^{(Sugita, 2009:256)}:

• gakusei wa benkyou suru
  学生は勉強する
  Students study. (kind-level predicate, i.e.: in general, students study.)
  The students study. (individual-level predicate, i.e.: these students study.)
• gakusei wa benkyou shite-iru
  学生は勉強している
  #Students are studying. (wrong at kind-level.)
  The students are studying. (okay at individual-level.)

Logically, this happens because ~te-iru, like iru, entails the existence of a particular entity. In the case above, the existence of particular students, even if we can't determine exactly which students we're talking about.

The ~te-iru form has a progressive and a resultative meaning, and which meaning it has depends on the lexical aspect of the word: achievement verbs become resultative, while other Vendlerian categories become progressive^{(Sugita, 2009:23,15n5; Vendler, 1957)}.

• Tarou wa hashitte-iru
  太郎は走っている
  Tarou is running.
  
  • Progressive, because hashiru 走る, "to run," is an activity.
• Tarou wa shinde-iru
  太郎は死んでいる
  Tarou is dead.
  • Resultative, because shinu 死ぬ, "to die," is an achievement.

In either case, what's essentially happening above is that the event is being actualized in the present.

To elaborate: if "Tarou is running," then I'll always be able to say "Tarou ran" afterwards.

However: if "Tarou is dying," there's a chance Tarou doesn't actually die, so I won't always be able to say "Tarou died" afterwards.

This means that, in English, "running" entails the actualization of the "run" event, while "dying" doesn't actualize the "die" event. At best, "has died," the English perfect, would be a present actualization of the "die" event.

In Japanese, ~te-iru always expresses the actualization of events, so shinde-iru doesn't translate to "dying" in English. However, Vendlerian achievements don't have durations, so shinde-iru can't express that the shinu is ongoing.

The result is that shinde-iru means the event is actualized and its effects remain for some reason. In this case, Tarou died, and now Tarou is dead, as result of him dying.

To express "is dying," the auxiliary ~kakeru ～かける is used instead. It means to be about to do something. Unlike ~iru, ~kakeru is eventive and affixes to the ren'youkei 連用形 to form a compound verb.

Since ~kakeru is eventive, that means it normally expresses the ~kakeru event occurs in the future, and we'll need to conjugate it to ~te-iru form to express that ~kakeru is actually occurring in the present.

• Tarou wa shini-kakeru
  太郎は死にかける
  Tarou will be about to die. (futurity.)
• Tarou wa shini-kakete-iru
  太郎は死にかけている
  Tarou is about to die.
  Tarou is dying.

It's been noted that, in ergative verb pairs, which are formed by a transitive and an intransitive verb, there's a tendency for the intransitive verb to be interpreted as resultative, that is, as an achievement^{(Matsuzaki, 2001:145-146, citing Kindaichi 1950, Yoshikawa 1976, Okuda 1978b, Jacobsen 1982a, 1992, Takezawa 1991, Tsujimura 1996, Ogihara 1998, Shirai 1998, 2000)}.

For example:

• The verbs tokasu 溶かす and tokeru 溶ける form an ergative verb pair for the "melt" event.
• ike no koori wo tokashite-iru
  池の氷を溶かしている
  To be melting the pond's ice.
• ike no koori ga tokete-iru
  池の氷が溶けている
  The pond's ice is melted.

In verbs that express a change of state such as above, from frozen to unfrozen, i.e. melted, the ~te-iru form can also be understood as progressive even with the intransitive verb. In such case, ~tsutsu aru ～つつある makes the progressive explicit^{(庵, 2001:80n4)}.

• ike no koori ga tokete-iru
  池の氷が溶けている
  The pond's ice is melted. (resultative.)
  The pond's ice is melting. (progressive change of state.)
• ike no koori ga toke-tsutsu aru
  池の氷が溶けつつある
  The pond's ice is gradually becoming melted. (literally.)
  The pond's ice is melting. (progressive.)

More evidence of the existential relationship can be found in iteratives.

Every time you have a number of occurrences of an event, or a period of time through which an event is assume to have occurred multiple times, you have an "iterative"^{(Bertinetto & Lenci, 2010:4–6)}.

While a habitual is generic, an iterative is not, given that: for something to occur a number of times or through a period of time, it must have actually occurred. Of course, if we're talking about the future, it doesn't matter, but if we're talking about the past, it does matter. Observe:

• Tarou wa ashita kara koko de hataraku
  太郎は明日からここで働く
  Tarou will work here since tomorrow. (literally.)
  Tarou will work here starting tomorrow.
  • This is habitual.
• *Tarou wa kinou kara koko de hataraku
  太郎は昨日からここで働く
  Intended: ?Tarou works here since yesterday.
  • This doesn't work because the hataraku would create a generic habitual, and if Tarou has worked since yesterday, he has already worked, so the event has already been actualized.
• Tarou wa kinou kara koko de hataraite-iru
  太郎は昨日からここで働いている
  Tarou has been working here since yesterday.
  • This works because ~te-iru expresses an iterative.

Above, we see that if we say an event has already started in the past, we're forced to use the ~te-iru form because the event is actualized. The same wouldn't happen with existence verbs, because the verb itself already does the actualization.

For example^{(adapted from 庵, 2001:81n5)}:

• kono isu wa kinou kara koko ni aru
  この椅子は昨日からここにある
  This chair has been here since yesterday.

The ~te-aru form always expresses a resultative meaning. It has two functions: intransitivizing and non-intransitivizing. Which one is it partially depend on whether we have a transitive or intransitive verb^{(近藤, 2018:6–7)}.

Simply put, the intransitivizing usage of ~te-aru takes a transitive verb and creates a sentence that looks like if ~te-iru had an unaccusative verb counterpart. Observe:

• Tarou wa tori-niku wo yaite-iru
  太郎は鶏肉を焼いている
  Tarou is cooking chicken-meat.
  • yaku
    焼く
    To cook. (transitive activity.)
• tori-niku wa yakete-iru
  鶏肉は焼けている
  The chicken-meat is cooked.
  • yakeru
    焼ける
    To cook. (intransitive accomplishment.)
  • The verb yakeru is the unaccusative counterpart of yaku.
  • Together, they form an ergative verb pair.
• Tarou wa keeki wo tsukutte-iru
  太郎はケーキを作っている
  Tarou is making a cake.
• keeki wa tsukutte-aru
  ケーキは作ってある
  The cake is made.

As you can see above, ~te-aru replaces ~te-iru in similar way to how yakeru replaces yaku.

It's worth noting that ~te-aru is far less common than ~te-iru for multiple reasons.

In order to use ~te-aru, we must be describing a resultant state that can be observed, and it must have been caused by an animate agent, who has agency, who had the intention of doing the event that resulted in the state^{(Sugita, 2009:92, citing Takahashi, 1976; Soga,1983; Matsumoto, 1990; Harasawa, 1994; Kageyama, 1996)}.

That is, we can't take an event such as "the fire burned the house," paraphrase it to "the house is burned," and use ~te-aru with it, because a fire doesn't have agency.

However, if someone had plans to burn the house, and they burned the house, then you can use ~te-aru, because the person has agency. Then would mean: it is done, the house is burned.

Similarly^{(Sugita, 2009:92)}:

• ki ga taoshite-aru
  気が倒してある
  A tree is fallen.
  • This sentence can't mean a strong wind toppled the tree. It can only mean that someone with the intention to topple the tree actually did it. They toppled the tree. It is done.

The non-intransitivizing ~te-aru is used in a similar way, except that subject is the agent.

For example^{(近藤, 2018:7)}:

• okaasan ga karee wo tsukutte-aru
  お母さんがカレーを作ってある
  Mom has made curry.
• boku wa kinou juubun ni nete-aru
  僕は昨日十分に寝てある
  I have slept plenty yesterday.

Sentences such as the above are used when the subject has done something that they intended to do, such that the resulting situation is somehow relevant in the present.

In particular, the first sentence implies that curry still exists right now, while the second implies that I'm feeling refreshed right now from having slept plenty yesterday. The effect of the event must last until the present for ~te-aru to be used.

Also note that ~te-oku ～ておく is a similar auxiliary used when doing things in advance.

• okaasan ga karee wo tsukutte-oita
  お母さんがカレーを作っておいた
  Mom made curry [in advance, so that we could eat it now].

Although the sentence above is similar, it has a difference: it's still valid if mom made curry and we already ate the curry, because ~te-oku only means something was "done in advance," it doesn't mean the resultant state is relevant in the present.

If we ate the curry already, the curry doesn't exist anymore, so we can't use tsukutte-aru with it^{[How ~てある and ~ておいた differs? - japanese.stackexchange.com]}, but we can still use tsukutte-oita, and we could use the past of ~te-aru, tsukutte-atta 作ってあった.

My guess is that ~te-iru and ~te-aru, besides deriving actualization, also derive the animacy requirement in the form of agency.

The ~te-iru can be progressive because actively "doing" something requires agency, while the ~te-aru is always resultative because ending up in a state isn't active, but passive. When mom is making curry, she's acting, but when curry is made, the curry doesn't act.

Of course, the fact that ~te-iru has a resultative meaning with achievement verbs and ~te-aru has a non-intransitivizing usage with agent subjects complicate things.

I assume these only occur because achievements lack duration so they wouldn't make sense otherwise, and the non-intransitivizing ~te-aru is simply the intransitivizing ~te-aru with a causer.

Compare ~te-iru and both ~te-aru with the ergative pair naru-suru that we've seen previously:

• Tarou wa heya wo souji shite-iru
  太郎は部屋を掃除している
  Tarou is cleaning the room.
  • Tarou has agency in this cleaning event.
• heya wa souji shite-aru
  部屋は掃除してある
  The room is cleaned.
  • The room doesn't have agency over the cleaning event.
  • It simply ended up in the state of cleaned after the event affected it.
• Tarou wa heya wo souji shite-aru
  太郎は部屋を掃除してある
  Tarou has made the room be cleaned. (literally?)
  Tarou has cleaned the room.
  • Tarou causes the resultant state of the room.
• heya wa kirei ni natta
  部屋は綺麗になった
  The room became pretty. (literally.)
  The room became clean.
  • The room changes.
• Tarou wa heya wo kirei ni shita
  太郎は部屋を綺麗にした
  Tarou made the room clean.
  • Tarou causes the change of the change.

Discrepancies

There are several discrepancies in how English and Japanese tense systems work, which cause a lot of problem for learners of either language.

Discourse-Level Futurates

Japanese allows for futurates to form using temporal references that aren't uttered in sentences but found at discourse level.

Consequently, in certain contexts, a stative verb looks like it has a future tense, even though it shouldn't have one.

For example^{(Ogihara, 1995:4)}:

• ashita kite-kudasai.
  Tarou ga koko ni imasu.
  明日来てください。
  太郎がここにいます。
  Please come [visit us] tomorrow.
  Tarou will be here.^{(translation by Ogihara.)}

In the sentence above, we have what looks like the existence verb iru いる, conjugated to its masu ます form, imasu います, displaying a future temporal reference.

The verb iru いる is stative. It can mean "something is somewhere" right now, but it can't mean "something will be somewhere" in the future.

The reason why iru いる expresses a future temporal reference in the example above is because the adverb ashita 明日 in the previous sentence sets a discourse-level temporal reference which iru いる implicitly makes use of in order to form a futurate.

Problematically, the same thing doesn't work in English.

For example, we can say:

• ?Tomorrow, Tarou is here.

That sounds weird, but it's understandable, and a valid futurate.

However, we can't say:

• *Come tomorrow. Tarou is here.

The temporal reference "tomorrow" and the verb "is" are in separate sentences in the example above, consequently, it's impossible to interpret it as:

• Tarou isn't here right now, but come tomorrow, Tarou will be here tomorrow.

We can at best interpret it as:

• Come tomorrow. Tarou is here right now, and he's probably still going to be here tomorrow.
  • I have no idea why I'm telling you this, though, because if you have business with Tarou, he's already here right now, so you don't really have to come tomorrow, you can just talk to him right now.
    • Or maybe you only come here because you want to talk to Tarou, and I want you to come here, so I'm using Tarou as bait to make you come here more often, but why would I be doing this? Do I want to see you often or something?
      • Wait... am I in some cliché love-triangle romcom???

Progressive Futurates

English allows the progressive to be used with a futurate, as well as the simple present.

For example^{(Copley, 2009:26)}.

• The Red Sox are playing the Yankees tomorrow.
• The Red Sox play the Yankees tomorrow.

Since the progressive in Japanese is expressed through the ~te-iru form, it would make sense to think that the ~te-iru form can be used in similar fashion, however, that would be incorrect^{(Sugita, 2009:24)}.

The discrepancy is that English progressive futurates are in the present.

The easiest way to understand this is to consider when you use a progressive futurate to assert that you're really going to do something. You've put your mind on it.

• I'll do this tomorrow.
• What do you mean by "no, you won't"?
• I AM doing this tomorrow.

Above, although we say "I am doing this," I'm not actually doing it right now, and I'm not talking about a period of time tomorrow in which I'll be doing it, either. The "I'm doing" progressive seems to add a nuance of present certainty, rather than change the temporal structure of the future event.

In Japanese, the ~te-iru futurate is simply a ~te-iru in the future. It talks about the future, and puts a progressive in it, or a resultative with achievements, as we've previously seen.

For example^{(adapted from Sugita, 2009:24n9)}:

• ashita Marii ga hashitte-iru
  明日マリーが走っている
  Tomorrow, Mary is running.
  • This sentence means the same thing as "Mary was running," except in the future: "Mary will be running." There will be sometime tomorrow in which Mary is running is actualized.
  • It doesn't mean, "Mary IS running tomorrow," as if we're absolutely certain that "Mary will run" tomorrow.

Pluractionality Ambiguity

One huge problem about understanding English and Japanese tenses is that, ironically, these two languages have too few morphological tenses. Too few conjugations.

Fewer conjugations make it easier to memorize them all. After all, they're few.

However, it makes it very difficult to tell what a word conjugated to a certain form actually means without context, because a single conjugation ends up having multiple different meanings.

It becomes ambiguous, for example, if "watched," mita 観た, is supposed to be perfective—you only watched something once—or imperfective, specifically, habitual—you used to watch something.

In the case of English, we can almost tell these three aspects apart through plurality and definiteness (a, an, the):

1. I watched the movie.
   Perfective, as it refers to a single event, and a single movie: "the" movie.
2. I watched movies.
   Habitual, because "movies" is a bare plural. It could be iterative in some contexts, though.
3. I watched the movies.
   Iterative, because "the movies" refers to particular movies, presumably, so it can't be habitual.
4. I watched two movies.
   Iterative, because "two movies" also refers to particular movies.

Japanese doesn't have plurals or articles like a, an, and the, consequently, iterative, habitual, and perfective aspects are ambiguous as far as the past tense is concerned.

Japanese does however distinguish between habitual and iterative aspects. The iterative aspect always has the ~te-iru ～ている form, in the past, ~te-ita ～ていた. Unfortunately, this same form is ambiguous with the progressive and perfect:

• watashi wa eiga wo mita
  私は映画を観た
  I watched the movie. (perfective.)
  I watched movies. (habitual.)
  
  • Note: generally, this sentence would be understood as perfective, and it would require some specific context to be understood as habitual, like someone asking you watch you used to do.
  • To force a habitual reading, an adverb like "often," yoku よく, is necessary.
    • watashi wa yoku anime wo mita
      私はよくアニメを観た
      I often watched anime.
• watashi wa eiga wo mite-ita
  私は映画を観ていた
  I watched the movies. (iterative.)
  I was watching movies. (progressive.)
  I had been watching movies. (perfect progressive.)

For comparison, Brazilian Portuguese has more tenses, so it has less ambiguity, because a specific verb form will be used for each specific time-aspect:

Time-Aspects vs. Tenses

Time-aspect	Brazilian Portuguese	English	Japanese
Future perfective	Assistirei	Watch	miru 観る
Present habitual	Assisto
Past habitual	Assistia	Watched	mita 観た
Past perfective	Assisti

Of course, Brazilian Portuguese has its ambiguities, too. I doubt that any language would have no ambiguity at all. It just happens that English and Japanese have too much ambiguity by comparison.

Present Perfect vs. Past Perfective

One troublesome mix up that happens is using the present perfect in place of past perfective or vice-versa.

The perfect vs. perfective thing is important, but ultimately it's not the part that really matters.

The problem is that we have a past tense and a nonpast tense, so the PRESENT perfect and the PAST perfective are going to be different morphological tenses.

Observe:

• Tarou wa hon wo san-satsu kaita
  太郎は本を３冊書いた
  Tarou wrote three books.
  • Past perfective.
• Tarou wa hon wo san-satsu kaite-iru
  太郎は本を３冊書いている
  Tarou has written three books.
  • Present perfect.

Observe that these two sentences describe the exact same facts.

1. If Tarou wrote three books, then Tarou has written three books.
2. If Tarou has written three books, then Tarou wrote three books.

Why do we have two different ways, in two different languages, to say the exact same thing? The reason is as follows:

1. The perfective is used when an event occurred in the past.
2. The perfect is used when an event occurred in the past, and it's somehow relevant in the present.

Thus, whether the perfective or perfect is used depends not on how the event happened, but on whether we're simply telling facts that occurred in the past, or mentioning facts that are relevant in the present.

In Japanese, when dealing with changes of state, expressing the resultant state takes priority over the fact a change occurred in the past. Consequently, the nonpast ~te-iru form tends to be used instead of the past form. For example^{(庵, 2001:81–82)}.

• a', sara ga warete-iru
  あっ、皿が割れている
  Ah, the dish is broken. (resultative translation.)
  Ah, the dish has broken. (perfect translation.)
• #a', sara ga wareta
  あっ、皿が割れた
  Ah, the dish broke.
  • This only makes sense if we're retelling events that occurred in the past, rather than talking about how the dish is right now.

Perfect and Temporal Adverbs

Problematically, in American English—and to a lesser extent in British English, too—the present perfect isn't used when you have a past adverbial providing the temporal reference, e.g. "yesterday," in which case the past perfective is used instead^{(Klein, 1992:unnumbered p.1, Hundt & Smith, 2009:45)}.

For example:

• I've seen the movie.
• I saw the movie yesterday.
• *I've seen the movie yesterday.

Presumably, this happens because "have" is present tense, so it isn't used with a past temporal reference such as "yesterday," which would mean the cases where it's acceptable work by modifying the subordinate clause instead. Observe:

• *Yesterday, I have {seen the movie}.
  • This is always wrong.
• *I have {seen the movie} yesterday.
  • Ungrammatical because "yesterday" can't modify "have."
    • Assuming we can't time travel.
• I have {seen the movie yesterday}.
  (same utterance as above.)
  
  • Grammatical if "yesterday" modifies "seen" instead.
  • Note that "seen the movie" isn't in present tense. The word "seen" is tenseless, so the temporal "yesterday" can modify it without running into the temporal problem it ran into when it modified "have."

By contrast, the Japanese ~te-iru is strongly tied with iteratives, so it's used even when you have a temporal reference, in which case you would use the past perfective in English.

For example^{(adapted from 庵, 2001:83, who marked oddity (?) in Japanese. The ungrammatical (*) marking in English follows Klein, 1992.)}:

• chichi wa wakai koro takusan asonde-iru.
  dakara, wakai mono no koudou ni rikai ga aru.
  父は若いころたくさん遊んでいる。
  だから、若い者の行動に理解がある。
  *[My] father has played a lot in [his] youth.
  Because [of that], [he] understands the behavior of young people.
• ?chichi wa wakari koro takusan asonda.
  dakara wakai mono ni rikai ga aru.
  父は若いころたくさん遊んだ。
  だから、若い者の行動に理解がある。
  [My] father played a lot in [his] youth.
  Because [of that], [he] understands the behavior of young people.

In Japanese, the naturalness of sentences above ranks like this:

1. asonde-iru (present.)
2. ?asonda (past.)

In English, we have the inverse tense-wise:

1. My father played a lot in his youth, so he understands. (past.)
2. *My father has played a lot in his youth, so he understands. (present.)

In colloquial English the auxiliary "have" is sometimes deleted, without changing the aspect. This is a different phenomenon. For comparison, a verb in which the past participle is distinct:

• I saw it, so I know. (past perfective.)
• I have seen it, so I know. (present perfect.)
• I seen it, so I know. (present perfect with "have" deleted.)

The ~te-iru perfect is also used in cases where a single event occurred at a single particular time, which strictly speaking wouldn't be iterative, since iterative means there are multiple occurrences.

For example^{(庵, 2001:79)}:

• han'nin wa mikka-mae ni kono mise de udon wo tabete-iru
  犯人は３日前にこの店でうどんを食べている。
  *The culprit has eaten Udon in this store three days ago.
  The culprit ate Udon in this store three days ago.
• kono hashi wa go-nen mae ni kowarete-iru
  この橋は５年前に壊れている。
  *This bridge has been broken five years ago.
  This bridge was broken five years ago.

In English, when you have a condition, the perfect is typically found in the protasis (antecedent), and it would be unusual to use it in the apodosis (consequent), given that we normally say:

• I've done this, so I can do that.

Not:

• I will do that, so I've done this.

For reference, an example of the latter^{(近藤, 2018:25)}:

• kenkou-shinda ga aru node
  健康診断があるので
  [I] have a physical-examination, so...
• kinou wa ku-ji made ni gohan wo tabete-iru
  昨日は９時までにご飯を食べている
  *[I] have eaten dinner at nine o'clock yesterday.
  [I] ate dinner at nine o'clock yesterday.

It's worth noting that ~te-iru ～ている isn't the only Japanese construction that translates to the English perfect. The ~te-aru form and ~koto ga aru ～ことがある can also translate to the perfect.

• Tarou wa hon wo kaite-aru
  太郎は本を書いてある
  Tarou has written the book.
  Tarou has made the book be written.
  
  • Tarou wa hon wo kaite-iru
    太郎は本を書いている
    Tarou is writing a book.
• Tarou wa {hon wo kaita} koto ga aru
  太郎は本を書いたことがある
  Tarou has {written a book} before.
  • ~koto ga aru is used to ask whether someone "has ever done something before."

We'll also see that, in more complex sentences, the perfect translation is the only one that makes sense tense-wise in English while still being functionally synonymous with the Japanese sentence.

Sequence of Tenses

In a complex sentence with multiple clauses where each clause has a tensed verb, we can describe the relationship between the the clauses' tenses in two ways:

1. Simultaneous.
   In which both events occur simultaneously.
   
   • I saw a man who was smoking.
   • Meaning: the man was smoking when I saw him.
   • Me "seeing" him and him "smoking" happened at the same time.
2. Shifted.
   In which the subordinate tense is relative to the matrix event.
   • I saw a man who was smoking.
   • Meaning: the man was smoking BEFORE I saw him, but he wasn't smoking when I saw him.
   • By the time I saw him, he already had stopped smoking.
   • Him "smoking" occurs in the past of me "seeing" him.

English and Japanese express the above differently^{(Ogihara, 1995:5–8)}.

In English, the past tense is always used when the event occurs before the sentence was uttered, and the present tense is always used when the event occurs at the same time as the sentence was uttered.

In other words, I only say "the man is smoking" if he's smoking as I say this. If he was smoking before I said this, it doesn't matter when he was smoking exactly, the word is always "was."

In Japanese, if the past tense is used in the subordinate clause, we can infer its event occurs before the matrix event. If the nonpast tense is used in the subordinate clause, it occurs at the same time or in the future relative to the matrix.

For example^{(Ogihara, 1995:7)}:

1. Jon wa {tabako wo sutte-iru} otoko ni atta
   ジョンはタバコを吸っている男に会った
   John met a man [who] {was smoking a cigar [when he met him]}. (simultaneous.)
2. Jon wa {tabako wo sutte-ita} otoko ni atta
   ジョンはタバコを吸っていた男に会った
   John met a man [who] {was smoking a cigar [before he met him]}. (shifted.)

Above, we have relative clauses, which are subordinate clauses, qualifying the noun otoko. The tense of the relative clause, ~te-iru and ~te-ita, is relative to the matrix event. When we have ~te-iru, it happens at the same time, when we have ~te-ita, it's shifted further to the past.

In other words, in English, the subordinate morphological tense depends on the utterance time, while in Japanese it normally depends on the matrix time.

This difference is sometimes inaccurately described as English having "absolute tense," while Japanese would have "relative tense."

English has both absolute and relative tenses^{(Declerck, 1988:514)}, and Japanese, too, has both absolute and relative tenses^{(岩崎, 1998:47)}.

Absolute Tenses

The term "absolute tense" refers to a temporal reference relative to utterance time^{(Declerck, 1988:513)}:

Comrie (1986) calls this the 'absolute deixis hypothesis', because it treats ['was'] as an absolute tense form, i.e. as a tense which relates a situation (i.e. event, state, etc.) directly to the moment of speaking.

This is kind of confusing by itself, since "absolute tense" isn't really absolute. It's RELATIVE to utterance time. There's no such thing as a truly absolute tense. Every tensed assertion we make assumes some point of time as reference.

The "absolute tense" assumes utterance time as point of reference, while the so-called "relative tense" would assume as time of reference the time of occurrence of another event that's expressed in the same sentence.

In any case, there are reasons to believe that the English past tense isn't always absolute.

For instance, we NORMALLY interpret "I saw a man who was smoking" simultaneously, which means "was smoking" is normally interpreted as being relative to the "I saw" event that's expressed in the same sentence.

If "was smoking" was always absolute tense, then I'd expect absolute confusion as to whether the man was smoking before, while, or after I saw him, since these three times can all occur before utterance time.

To elaborate, observe the statements below:

1. I saw a man.
2. After I saw him, he started smoking.
3. The man was smoking after I saw him.
4. I saw a man who was smoking.

Above, the man was smoking before I said this, but after I saw him. The "smoking" event occurs in the future compared to the "saw" event, but in the past compared to the utterance time.

This is a case where the two tenses in a single sentence have nothing to do with each other.

In other words, they're parallel: "saw" is past of utterance time and "was smoking" is past of utterance time. They're both absolute tenses.

We don't infer when "was smoking" happened from when "saw" happened. We infer both from when the sentence was uttered.

This parallel interpretation requires context or some temporal adverb to make any sense, so I would say it's not, strictly speaking, about the relationship between two tenses in a sentence, but the lack of such relationship in a sentence.

For the sake of reference, let's include it here too:

3. Parallel.
   In which the two temporal references are completely unrelated.
   • This only makes sense with appropriate context, or with a more complex sentence.
   • Context: together with a detective, you've infiltrated a party in search for clues to solve a bizarre crime. The detective asks you:
   • Which one of these people did you see at the crime scene?
   • I saw the man who was smoking.
   • I saw the man who is smoking.
   • Here, "saw" refers to some time in the past, when the speaker was at the crime scene, while "is smoking" refer to something that's happening right now.
   • Let's say the crime happened last week, and the party started 1 hour ago. If I "saw" the man last week, and the man "was smoking" after the party started, that means "was smoking" happened way after I "saw" the man.

Now, let's review the previous example accounting for parallel interpretations:

1. Jon wa {∅ tabako wo sutte-iru} otoko ni atta
   ジョンはタバコを吸っている男に会った
   John met a man [who] {was smoking a cigar [when he met him]}. (simultaneous.)
   John met the man [who] {is smoking a cigar [right now]}. (parallel.)
2. Jon wa {∅ tabako wo sutte-ita} otoko ni atta
   ジョンはタバコを吸っていた男に会った
   John met a man [who] {was smoking a cigar [before he met him]}. (shifted or parallel.)
   John met a man [who] {was smoking a cigar [when he met him]}. (parallel and simultaneous.)
   John met the man [who] {was smoking a cigar [just now]}. (parallel.)

Note that when the subordinate past is parallel, it can occur at ANY time before utterance. This includes occurring at the same time as the matrix event, or before it. Consequently, there's a possibility that sutte-ita and sutte-iru are temporally synonymous. But this is very unlikely.

In chart form:

Source: japanesewithanime.com (CC BY-SA 4.0)

The parallel interpretations, although possible^{(Ogihara, 1995:8,159–160)}, don't make sense without a temporal adverb, or without sufficient context to replace the temporal adverb.

This is important to note because if you hear the sentences above without context, you should assume they're either simultaneous or shifted only, not parallel, or "parallel that happens to be simultaneous."

For example, if we say:

• Jon wa
  {tabako wo sutte-iru} otoko ni
  atta
  ジョンは
  タバコを吸っている男に
  で会った
  John met the man
  [who] {is/was smoking}.

We should by default assume that the man was smoking when John met him, it's simultaneous.

However, maybe if you can see a man smoking right here and now, then there's a possibility that we're talking about that man, and he wasn't smoking when we saw him, and these two events are in parallel.

Of course, in such case, the fact there is such man smoking around would be part of the context, and not of the sentence.

By contrast, if we said^{(Ogihara, 1995:8)}:

• Jon wa
  {ima asoko de tabako wo sutte-iru} otoko ni
  kinou michi de atta
  ジョンは
  今あそこでタバコを吸っている男に
  昨日道で会った
  John
  met on the street yesterday
  the man [who] {is smoking cigars over there right now}.

Then we unambiguously understand that sutte-iru is absolute tense, and these two events are parallel.

In order for us to utter this same sentence, with this same parallel meaning, but without the adverbs ima asoko de, we would need the bizarre situation in which:

1. Everyone in the conversation knows that we're talking about the man over there right now, so we don't need to be explicit about that.
2. We need to be explicit about the fact he's smoking because apparently nobody knows this.

As you may imagine, that's just never gonna happen.

If we're talking about whether it's theoretically possible in a theory of tenses, then, yeah, sure, however, in practice, it's extremely unlikely, because if we're just talking about that man over there, we could literally point to the dude and say:

• Jon wa ano otoko ni atta
  ジョンはあの男に会った
  John met that man.

We wouldn't even need a tensed relative clause.

If it's a bizarre situation in which we can't point to the dude, then, obviously, we'll have to describe him with more words. We would then say "the man who is smoking by the window," for example.

But "by the window" is a place in the present context, so the spatial adverb kind of works as a temporal adverb in this case, as it helps us understand the temporal reference.

Clearly none of these contexts help us say just {tabako wo sutte-iru} otoko.

At best, if we had exactly two men, and one was smoking, and the other wasn't smoking, then, maybe then, it would make sense to disambiguate only by the fact and nothing more than the fact that the one we're talking about is smoking.

Parallel interpretations of relative clauses only make sense when you're talking about the thing in relation to utterance time.

To elaborate, an example^{(Ogihara, 1995:159)}:

• Tarou wa
  {Nooberu-shou wo totta} otoko wo
  sagashita
  太郎は
  ノーベル賞をとった男を
  探した
  Tarou searched for the man [who]
  {won the Nobel prize}.
  • Shifted: Tarou was searching for a Nobel-prize winner.
  • Parallel: you know that man that won a Nobel prize? Tarou was searching for him.

A more common situation in which we have two absolute tenses in a single sentence is when we are talking about entire events that occurred at some time, rather than about people that did things at some time.

For example:

• Context:
  • ano otoko ga tabako wo sutte-ita
    あの男がタバコを吸っていた
    That man was smoking.
• watashi wa
  {ano otoko ga tabako wo sutte-ita} no wo
  mita
  私は
  あの男がタバコを吸っていたのを
  見た
  I saw [it] [when] {that man was smoking}.
  I saw {that man smoking}.

In the sentence above, both clauses are in absolute tense in Japanese, which means this is a parallel interpretation. Again, this only makes sense with appropriate context. We must be already talking about a time in the past in which the man was smoking.

Without context, sentences such as above are still tensed relative to the matrix tense. For example^{(Ogihara, 1995:127)}:

• {Tarou ga machigatte-iru} no wa
  akiraka datta
  太郎が間違っているのは
  明らかだった
  *That {Tarou is mistaken}
  That {Tarou was mistaken}
  was obvious.
  • This is a simultaneous interpretation, because the subordinate has the nonpast ~te-iru.
  • This sentence is "it's obvious that Tarou is mistaken" in the past.
  • Tarou was mistaken WHEN we asserted it was obvious.
• {Jon ga moushide wo uke-ireta} no wa
  akiraka datta
  ジョンが申し出を受け入れたのは
  明らかだった
  That {John accepted the offer}
  was obvious.
  • This is a shifted interpretation, because the subordinate has the past uke-ireta.
  • This sentence is "it's obvious that John accepted the offer" in the past.
  • John accepted the offer BEFORE we asserted it was obvious.

In summary: most of the time a subordinate clause in Japanese will have a relative tense and a shifted or simultaneous interpretation depending on whether it's past or nonpast tense. However, an absolute tense in a subordinate clause is also possible depending on context.

Reported Speech

When reporting what other people said, Japanese uses quotations, which are direct speech, while English prefers indirect speech. This discrepancy, added to English's idiosyncrasies, results in Japanese nonpast translating to past in English. For example^{(Ogihara, 1995:69)}:

• Tarou wa {Hanako ga byouki da} to itta
  太郎は花子が病気だと言った
  Tarou said: "Hanako is sick." (literally.)
  Tarou said that Hanako was sick.
  • Simultaneous: Hanako was sick at the time when Tarou said this.
• Tarou wa {hanako ga byouki datta} to itta
  太郎は花子が病気だったと言った
  Tarou said: "Hanako was sick."
  Tarou said that Hanako had been sick.
  • Shifted: Hanako was sick BEFORE Tarou said this, which implicates she was no longer sick by the time when Tarou said it.

The difference between English and Japanese in sentences such as above is that English uses a subordinate clause to report what someone else has said, while Japanese uses a literal quote with a quoting particle, such as to と or tte って.

When the Japanese quote is in the past, it can't have a simultaneous interpretation. It can only have the shifted interpretation^{(Ogihara, 1995:239)}.

To elaborate: if I say "Hanako WAS sick," that means she was sick in the past, before I said it. Thus, if Tarou said, literally "Hanako WAS sick," that means she was sick before Tarou said it.

That's how Japanese works.

In English, if Tarou says "Hanako IS sick," we could report his speech with:

• Tarou said that Hanako WAS sick.
  • This could mean that Hanako was sick at the time Tarou said this, i.e. simultaneous interpretation.
  • Or it could mean she was sick before he said this, i.e. shifted interpretation.

The only way to unambiguously mean the shifted interpretation is the use of the perfect.

• Tarou said that Hanako had been sick.

The same thing occurs when kato かと is used with an uncertainty:

• Context: I almost die.
• shinda kato omotta!
  死んだかと思った！
  [I] thought: [I] died! (literally.)
  ?I thought that I died!
  I thought that I had died!

Tensed Adjectives

The difference between how tenses relate between matrix and subordinate clauses supports the idea that Japanese adjectives should be analyzed just like verbs, for they also have tenses.

To elaborate, it's normal for translations of Japanese phrases with a noun modified by a tensed adjective to translate to an tenseless adjective word in English:

• atsui ocha
  熱いお茶
  A hot tea.

However, a more literal translation, considering tense and syntax, would be a relative clause:

• {atsui} ocha
  熱いお茶
  A tea [that] {is hot}.

Japanese adjectives have tenses like verbs, even though English adjectives don't. Consequently, things get hard to translate non-literally when the adjective is past tense:

• {atsukatta} ocha
  熱かったお茶
  The tea [that] {was hot}. (as simple as it gets)
  The... uh... were-hot tea? (as complicated as it gets.)
  The ex-hot tea.
  Previously hot tea.
  Once hot tea.

A supposed problem with this analysis is that when the tense of the matrix is past, the adjective being tensed nonpast translates literally ungrammatically to English.

• {atsui} ocha wo nonda
  熱いお茶を飲んだ
  *[I] drank a tea [that] {is hot}.
  [I] drank a tea [that] {was hot}.

But this is just the same thing we've seen previously: the subordinate clause has relative tense, nonpast, atsui, so it occurs simultaneously with the matrix event, which is past, nonda.

We can assert that the ~i ～い copula of i-adjectives is tensed nonpast, just like ~te-iru, then. This makes perfect sense together with what we've learned so far about Japanese grammar, even if it makes no sense in English grammar.

The same thing happens with na-adjectives, except their attributive copula is na な.

• {kirei na} hana wo mita
  綺麗な花を見た
  *[I] saw a flower [that] {is pretty}.
  [I] saw a flower [that] {was pretty}.

As we've seen previously, normally, if we have the past tense in a subordinate, we'll have the shifted interpretation and a relative tense.

This means that a sentence such as:

• {kirei datta} hito ni atta
  綺麗だった人に会った
  [I] met a person [that] {was pretty}.

Would actually mean that at the time I met this person, she was no longer pretty.

• [I] met a person [that] {used to be pretty}.
• ano hito wa kirei datta
  あの人は綺麗だった
  That person was pretty.
  • Implicature: she no longer is pretty.

However, also as we've already seen, there are situations in which a past subordinate can be simultaneous if the tense is absolute. For example:

• {touji binbou datta} watashi wa
  anime no buruu-rei wo kaenakatta
  当時貧乏だった私は
  アニメのブルーレイを買えなかった
  I, [who] {was poor at the time},
  couldn't buy anime blu-rays.
  • kaeru
    買える
    To be able to buy.
    Can buy.
    (potential verb derived from kau 買う, "to buy.")

In the sentence above, we're saying that I was poor at the same I couldn't buy anime blu-rays. This is a parallel-simultaneous interpretation.

It works because touji, "at the time," is understood to be a specific time before the utterance time, i.e. touji is relative to utterance time, and binbou datta occurs at touji time, which means binbou datta occurs relative to utterance time, and, as such, is in absolute tense.

Again, typically this won't be the case, and the tense of a relative clause will be nonpast.

• {yasui} kuruma wo katta
  安い車を買った
  [I] bought a car [that] {was cheap at the time I bought it}.
• {oishii} mono wo tabeta
  美味しいものを食べた
  [I] ate something [that] {was delicious at the time I ate it}.

Past of Future

In Japanese, just like the present can be relative to the time of an event in the past, the past can be relative to the time of an event in the future.

Although this sounds simple in practice, there are actually various differences between English and Japanese that can be found when such thing happens.

For example^{(Ogihara, 1995:159)}:

• {ashita no shiai de katta} hito wa
  kin-medaru wo moraimasu
  明日の試合で勝った人は
  金メダルをもらいます
  The person [that]
  *{won in tomorrow's match},
  will get a gold medal.

Above, the verb katta in the relative clause translates wrongly to past as "won" in English. The correct translation would be the present "wins."

• Who will get the gold medal?
• The person that wins in tomorrow's match.

Source: japanesewithanime.com (CC BY-SA 4.0)

If we remove the temporal adverb ashita, "tomorrow," which provides us we the future temporal reference, we can end up having a perfectly valid sentence in both English and in Japanese:

• {shiai de katta} hito wa
  kin-medaru wo moraimasu
  試合で勝った人は
  金メダルをもらいます
  The person [that] {won in the match}
  will get a gold medal.
  • This English translation assumes the match ended already, so we're right before the moraimasu event.
  • If it didn't end, we would still get the translation "the person that wins will get a gold medal."

The reason the English "wins" doesn't work in the original sentence is because we have a futurate in English.

• A person wins in the match tomorrow.
• The person gets a gold medal.
• The person {that wins in the match tomorrow} gets a gold medal.

By contrast, we don't have a futurate in Japanese, since we can't use the past tense in futurates:

• #ashita no shiai de dare-ka wa katta
  明日の試合で誰かは勝った
  #In tomorrow's match, someone won. (requires prescience to make sense.)

There are two possibilities.

1. ashita no shiai de is modifying the nonpast verb moraimasu, not the past verb katta.
2. In a subordinate clause, the past tense katta can be used even with a future temporal reference such as ashita when the matrix event takes place after that temporal reference.

As it turns out, 1 is incorrect, and 2 is correct.

For example, if we used two temporal adverbs, one explicit temporal reference for each verb, we'd remove syntactical ambiguity, and we would have a sentence like this:

• {ashita no shiai de katsu} hito wa
  raishuu kin-medaru wo moraimasu
  明日の試合で勝つ人は
  来週金メダルをもらいます
  The person [that] {wins in tomorrow's match}
  will get a gold medal next week.

This can only be interpreted in parallel. We would require a context in which we're talking about the people that will win tomorrow.

We can't interpret the sentence above as katsu having relative tense, because if it had a relative tense, it would have to happen after moraimasu, which means "tomorrow" would be after "next week."

By contrast:

• {ashita no shiai de katta} hito wa
  raishuu kin-medaru wo moraimasu
  明日の試合で勝った人は
  来週金メダルをもらいます
  The person [that] {won in tomorrow's match}
  will get a gold medal next week.

Can be interpreted as having a shifted interpretation. We use the past tense katta because "tomorrow" comes before "next week."

As we've seen previously, words like "tomorrow" can be used with the past tense if we time travel. We aren't time-travelling right now, but we're looking at the timeline from the perspective of "next week," which is what allows us to use the past tense with "tomorrow."

Qualified Temporal Adverbials

Japanese has several nouns that refer to time, like ato 後, "after," mae 前, "before," toki 時, "when," and so on.

Although the translations are adverbs in English, in Japanese they're syntactically nouns, and can be qualified by tensed relative clauses.

This gets a bit more complicated by the fact these qualified nouns can be turned into adverbs. For example:

• {suru} mae da
  する前だ
  [It] is before {[you] do}.
• {suru} mae ni
  する前に
  Before {doing}.

These temporal adverbs follow the same sequence of tenses semantics we've already seen. In particular, they follow the idea that the tense of the subordinate event is reliant on the temporal reference of the matrix event.

This is a bit complicated, so first let's see some examples of how mae 前 and ato 後 work^{(Ogihara, 1995:181–182)}.

First, when both subordinate and matrix events are in the future, we have a sentence like this:

• Tarou wa {{Hanako ni au} mae ni}
  denwa wo suru
  太郎は花子に会う前に
  電話をする
  Tarou, {before {meeting Hanako}},
  will make a call.
  • Tarou will call Hanako before going to see her.

Above, both au and suru are nonpast, and the whole thing is set in the future.

• Tarou will meet Hanako.
• Tarou will call Hanako before he meets her.

The problem happens when we try to say this same thing in the past. Since both of them were in the nonpast for the future, we'd assume that we'll have to change both of them to the past to say it in the past, however, if we do that, we get an ungrammatical sentence:

• *Tarou wa {{Hanako ni atta} mae ni}
  denwa wo shita
  太郎は花子に会った前に
  電話をした
  Intended: Tarou, {before {[he] met Hanako}}, called [her].

The reason why this happens is very simple.

The word atta, "met," is in the past, and we know that this means it must be in the past relative to matrix event, denwa wo shita, so we end up with this order of events:

1. Tarou meets Hanako before...
2. Tarou calls her.

However, what we intended to say was this:

1. Tarou calls Hanako before...
2. Tarou meets her.

In other words, the tenses mean the meeting happened before the calling, but the adverb mae 前 means the calling happened before the meeting. It's contradictory and doesn't make sense.

For this sentence to make sense, the nonpast tense is used.

• Tarou wa {{Hanako ni au} mae ni}
  denwa wo shita
  太郎は花子に会う前に
  電話をした
  Tarou, {before {meeting Hanako}},
  made a call.
  
  • Tarou called Hanako before going to see her.

Above, the nonpast tense of au 会う means that it happens after denwa wo shita 電話をした, which means denwa wo shita can happen "before," mae, the au event.

The exact same thing happens with ato 後, "after," except in reverse.

• Tarou wa {{Hanako ni atta} ato de}
  denwa wo shita
  太郎は花子に会ったあとで
  電話をした
  Tarou, {after {meeting Hanako}},
  made a call.
• *Tarou wa {Hanako ni au} ato de
  denwa wo suru
  太郎は花子に会う後で
  電話をする
  Intended: Tarou, after {[he] meets Hanako}, will call [her].
• Tarou wa {Hanako ni atta} ato de
  denwa wo suru
  太郎は花子に会ったあとで
  電話をする
  Tarou, {after {meeting Hanako}},
  will make a call.

As a chart:

Source: japanesewithanime.com (CC BY-SA 4.0)

With the word toki 時, things get more complicated, or rather, more simpler. It has no restrictions, and depends entirely on the tense of both events. Consequently, it doesn't have the restrictions that mae and ato have.

For example^{(朱, 2010:304)}:

1. {nihon e kuru} toki,
   tomodachi ga kuukou made kite-kureta
   日本へ来るとき、
   友達が空港まで来てくれた
   *When {[I] will come to Japan},
   [my] friend came to the airport for [me].
   • Before I came to Japan, my friend came to the airport for me.
2. {nihon e kita} toki
   tomodachi ga kuukou made kite-kureta
   日本へ来たとき、
   友達が空港まで来てくれた
   When {[I] came to Japan},
   [my] friend came to the airport for [me].
   
   • After I came to Japan, my friend came to the airport for me.

In the examples above, the matrix event "came for," kite-kureta, is in past tense. That means that either way the friend has already come.

The question is whether the friend comes for me before "I come to Japan" or after.

In the with sentence nihon e kita toki, we're saying that "I came to Japan" in the past, before the matrix event, so I was already in Japan, in a Japanese airport, when my friend came meet me.

Meanwhile, the nihon e kuru toki sentence is in nonpast, "I will come to Japan" after the matrix event, after my friend comes meet with me, which means I'm not in Japan when the matrix event occurs. My friend came for me in the airport before I entered the plane to Japan.

The adverbs mae and ato only allow either future or past temporal references. The adverb toki also allows a present temporal reference through the ~te-iru form.

For example^{(朱, 2010:310, excerpted from ２００９年０８月０１日　中日新聞　朝刊オピニオン　７頁)}:

• {heya de sugoshite-iru} toki wa,
  ohanashi wo shitari, toranpu wo shite
  tanoshimimashita.
  部屋で過ごしている時は、
  お話をしたり、トランプをして
  楽しみました。
  *When {[I] am spending time in the room}, [I] enjoyed talking, playing cards.
  I had fun talking, playing cards while spending time in the room.

In the sentence above, "enjoyed," "had fun," tanoshimimashita, occurs in the past. While it occurs, the event "to spend time," sugosu, is going on—it's in the progressive: "to be spending time."

Note that there's only three possible relative tenses: before, while, and after, or relative past, present, and future. But we have four possible conjugations, ~ru, ~ta, ~te-iru, and ~te-ita forms.

Since with eventives ~ru is future, ~ta is past, and ~te-iru is present, there's a question of whether ~te-ita can be used with toki, and if it can, what does it mean.

As it turns out, when ~te-ita is used, it's synonymous with ~te-iru, which means ~te-ita has the parallel-simultaneous interpretation. It doesn't have the shifted interpretation. For example^{(朱, 2010:311 excerpted from ２００９年０７月２３日　中日新聞　朝刊三社　２９頁)}:

• yon-nen hodo mae no dekigoto wo omoi-dashimashita.
  四年ほど前の出来事を思い出しました。
  [I] remembered something that happened around four years ago.
• daidokoro de shokuji no junbi wo shite-ita toki,
  台所で食事の準備をしていた時、
  When [I] was preparing a meal in the kitchen,
• niwa no hou kara otto no koe ga kikoemashita
  庭の方から夫の声が聞こえました。
  [I] heard [my] husband's voice from the garden.

In the example above, junbi wo shite-ita occurs simultaneously with kikoemashita.

If junbi wo shite-ita had a relative past tense, this would be impossible, and the sentence would mean that the speaker had already finished the preparations BEFORE they heard their husband's voice.

Since it means they heard the husband's voice WHILE they were doing the preparations, junbi wo shite-ita must have absolute past tense, and the relationship between subordinate and matrix tenses is parallel..

Relative Tense of Tenseless

Relative tenses express that something occurs before, at the same time, or after the matrix event.

Unsurprisingly, relative tenses still make sense if the matrix event is tenseless.

For example:

• {terebi wo miru} toki wa heya wo akaruku shite hanarete mite-kudasai!
  テレビを見る時は部屋を明るくして離れて見て下さい！
  When {watching TV}, please make the room bright, distantiate [yourself], and watch!
  • When watching TV, keep the room well-lit and don't watch from too close!

In a sentence such as the above, we understand that you have to do all these things BEFORE watching TV, not after. We understand this because miru 見る is in nonpast, so it happens after the long sequence of events in the matrix.

However, none of the events in the matrix is tensed.

The verbs shite, hanarete, and mite are in te-form, which is tenseless and derives its tense from whatever tensed verb comes after it. In this case, however, what comes after these is kudasai, which is an imperative, and therefore tenseless.

This sentence, then, lacks an absolute tense to place it somewhere in time relative to utterance time. Simply put, we have no idea when it happens. All we know is that, if it ever happens, then "watching TV" happens after making the room bright.

Momentaneous Observations

Eventive verbs in nonpast express either habituality or futurity. Problematically, sometimes they're uttered in response a momentaneous realization, which makes it look like they're expressing something different.

For example, Hasegawa & Verschueren (1998:2) list 9 functions for the ta-form (past) and 12 for the ru-form (nonpast), for a total of 21 functions. Among them:

(11) A present psychological state: hara ga tat-U ‘I’M ANGRY.’
(13) An event occurring in front of one’s eyes: a, teppan ga oti-RU ‘Oh, a steel plate IS FALLING down!’ (Suzuki 1965)
(16) A past event: kikizute naranai koto o i-U ne ‘You’VE SAID something I can’t ignore.’

As you can see above, (11) would be a present state, which tatsu 立つ, "to stand," shouldn't be able to express since it's not stative; the verb ochiru 落ちる in (13), likewise, can't express the progressive "is falling," without ~te-iru; and, iu 言う can't express "have said" without ~te-aru.

The sentences above can be analyzed as manifestations of the two functions of the ru-form: habitual and futurity. It gets complicated, however, because of the contexts in which they're used. Specifically:

1. When a habitual is uttered right after someone does something that's expressed by the habitual.
2. When a future event is uttered while the same event can also be said to have already occurred in that same instant.

Qualified Futures

Observations made with the ru-form sometimes use a verb whose action, at first glance, can be thought to be already happening, and thus ~te-iru or past form would make more sense. However, the action is actually implicitly qualified, and the qualified action isn't happening yet.

To understand what this means, let's first see an example by Suzuki (鈴木, 1965:10).

• a, teppan ga ochiru!
  あ、鉄板が落ちる！
  Ah, the steel plate will fall! (literally.)
  Ah, the steel plate is falling! (progressive translation.)

The sentence above literally means "will fall," in the future, but is translated to English in the progressive, "is falling." Why is that?

First an foremost, the lexical aspect of ochiru 落ちる would be an achievement. Consequently, ochite-iru 落ちている doesn't mean the progressive "is falling," but the resultative "is fallen."

Typically, ochite-iru is used when something "is on the floor." We could say that ochiru doesn't mean "will fall" in this case, then, but "will hit the floor" instead.

Naturally, it doesn't make sense to say something "will fall" if it's already "falling," but it makes perfect sense to say that something that hasn't "hit the floor" yet "will hit the floor" in the future.

Likewise, if the steel plate has already fallen 20cm mid-air, we could still say "the steel plate WILL FALL 30cm," as it still needs to fall 10cm to go from 20cm to 30cm.

By the way, the ~te-iku ～ていく form and ~te-kuru ～てくる form are used to force a dynamic reading:

• ochite-iku
  落ちていく
  To be falling [away from me].
• ochite-kuru
  落ちてくる
  To be falling [toward me].

Another example^{(尾野, 1998:33, citing 三上:1953, 219)}:

• Context: the straw hat is already mid-air.
• a, (mugi-wara) boushi ga tobu!
  あ、（麦わら）帽子が飛ぶ
  Ah, the (straw) hat will fly!

In this case, again, it has to do with the speaker's conception of the action described by the verb.

It seems that with movement verbs, such as tobu, time alone can't explain why the ru-form is selected, however, once you account for both time and space, it all makes sense.

The speaker might not consider a hat to be actually "flying," tonde-iru 飛んでいる, until it's beyond its grasp. Thus, the hat is not, in fact, flying, according to the speaker's criteria, even though by someone else's criteria it would be flying.

Finally, we have this notable example^{[時制を旅する(8) あっ、バスが来た。ほんとだバスが来る。 - je.at.webry.info, 2010-07-22]}:

• Context: two people are waiting for a bus. One of them sees the bus and says:
• a', basu ga kita
  あっ、バスが来た
  Ah, the bus came.
• honto da, basu ga kuru
  ほんとだ、バスが来る
  [It] is true, the bus will come.

If the two people are in the same place, the bus is in the same place, and the two sentences were uttered at about the same time, then why one person uses the ta-form while the other person uses the ru-form?

The first person refers to "coming" into an area defined by the space that they can see, while the second person refers to "coming" into an area limited by the space in front of them.

• Ah, the bus came [into sight. I can see it over there, far away.]
• It's true, the bus will come [in front of us in a few moments, so we'll still have to wait a bit until we can get inside.]

Source: japanesewithanime.com (CC BY-SA 4.0)

This is also the reason why, in action anime, a character notices an enemy "is coming" and utter instead:

• teki ga kuru!
  敵が来る！
  The enemy will come!
  
  • They haven't come until they're literally in front of the character crossing swords and shooting beams, apparently.

A similar, but inverted case, happens when when realizing something in the present, and using the ta-form to refer to its current state in the past^{(尾野, 1998:33, citing 牟世鍾, 1993)}:

• Context: I search for something.
• a', mitsukatta. koko ni aru.
  あっ、見つかった。ここにある。
  Ah, [I] found [it]. [It] is here.
  • I found it now, and it "is here" now.
• a', mitsukatta. koko ni atta
  あっ、見つかった。ここにあった。
  Ah, [I] found [it]. [It] was here.
  • While I was searching in other places in the past, it "was here."

Momentaneous Habituals

Sometimes a person does X, and the speaker sees that person doing X and utters a habitual sentence featuring X.

Although the grammar is pretty straight-forward, it's a quite weird and confusing thing to happen and to understand.

Habitual sentences generally imply habits, and habits make you think of a somewhat frequent repetition of some act. However, the usage we'll see below happens when the speaker sees the habit-doer do the "habit" ONLY ONCE.

That is, they see them do it ONCE, and assume that's something they're CAPABLE of doing, something that they DO.

For example, the verb iu 言う, "to say," is typically used as an habitual, to note the sort of thing someone is able to say, upon hearing someone say a certain kind of thing once.

• Context: you hear someone say something horrible.
• hidoi koto iu ne
  酷いこと言うね
  You say horrible things, don't you. (literally.)
  You HAVE said something horrible, just now. (perfect translation.)
  You SAID something horrible, just now. (past translation.)

It's merely the lack of an adverb that makes one translate the habitual, "say," as a perfect, "have said," or past, "said." Observe an example with an adverb:

• Context: you hear someone say something good for once.
• tama niwa ii koto iu ne
  たまにはいいこと言うね
  Sometimes you say good things, don't you.

This feature of habituals isn't limited to iu 言う. Notably:

• yaru na!
  やるな！
  [You] do! (literally.)
  • What this actually means: I'm impressed! You did well! Good job!

In case you're wondering why a single word can be construed as an "habitual," it's simply because everything else in the habitual wasn't uttered in the sentence, but inferred by context.

For example, if someone manages to run a single lap very quickly, then the thing they "did," yatta やった, was running the lap very quickly. The word yaru infers the habit: you "do" run laps very quickly! You can do it! You manage this stuff! yaru na!

Manga: World Trigger, ワールドトリガー (Chapter 6, 嵐山隊)

• Context: Kuga Yuuma 空閑遊真 sees someone lying.
• omae...... tsuman'nai uso φ tsuku ne
  おまえ･･････つまんないウソ　つくね
  You...... tell boring lies, [don't you]?
  
  • tsuman'nai is a contraction of tsumaranai つまらない.

Note that this stuff is practically the same as:

• Context: something unbelievable happens.
• shinjirarenai!
  信じられない！
  [I] am not able to believe [it]!
  [I] can't believe [it]!
  
  • shinjiru
    信じる
    To believe. (an ichidan verb.)

Or the most common observation ever made in fighting anime:

Manga: One Punch Man, Wanpanman ワンパンマン (Chapter 10)

• hayai!
  速い！
  [He] is fast!

The only thing that changes is that the sentence has a habitual instead. Tense-wise the structure is the same. Translation-wise they're also awkward to translate to English. Same thing, really.

Habitual vs. Stage Topicalization

The topic and focus of an assertion regards what new information (focus) is being asserted about old information (topic). Although both English and Japanese have topics and focii in their sentences, only Japanese has obvious markers for them.

The wa は particle and tte って particle are topic markers, while the ga が particle has two functions, one which marks a subject focus and the other that can express a sentence focus, they're called "exhaustive listing" and "neutral description" respectively.

These two functions discriminate between stage-level predicates and individual-level predicates. SLPs can have both functions of ga が, while ILPs only have the exhaustive listing function^{(鈴木, 2014)}.

Stage-level predicates assert things that are true about the stage, that is, that are true about here and now. In such predicates, it makes sense to consider the stage itself as the topic. Observe the difference below:

• sora wa aoi
  空は青い
  The sky is blue. (generally.)
  • This is an individual-level predicate.
  • sora is the topic, because we're talking about a generic property of it..
• sora ga aoi
  空が青い
  The sky is blue. (right here and now.)
  • This is a stage-level predicate.
  • The stage is the topic, because we're talking about what we're observing right here and now.

The reason why the ga が particle is used in the sentence above is that the topic is supposed to be the stage, but we don't explicitly refer to "right here and now" when talking about "right here and now."

If we don't say a word for stage, there's no way to mark it with wa は, so nothing gets marked with wa は.

Tense-wise, it's doesn't make sense for a habitual sentence to be part of a SLP.

Habits express that an event has occurred multiple times across a likely long span of time. It's possible to observe a single instance of the event occurring "here and now," but it doesn't make sense for us to observe multiple instances occurring "here and now."

That's simply because the "now" in "here and now" is a single point in time, but to observe multiple occurrences we would need multiple points in time.

Thus, I can't observe that "John reads books" here and now. I can only observe that "John is reading a book" here and now. At best, I can observe that "John is reading two books at once" here and now, or "reading a second book," but that still only counts as observing a single event.

Based on the above, it's reasonable to assume that habituals can't happen with SLPs.

In other words, when you have a simple sentence with a verb in nonpast and the subject is marked by the ga が particle, that won't be a habitual, that will be the future tense.

That's what occurs with tepan ga ochiru, basu ga kuru, etc. Those are all observations made about right here and now giving conclusions about what the speaker thinks is going to be true in the immediate future.

Similarly, if we were too see the following sentences without context, we'd assume the interpretations::

• Tarou wa manga wo yomu
  太郎は漫画を読む
  Tarou reads manga. (habitual.)
  Tarou will read the manga. (future.)
• Tarou ga manga wo yomu
  太郎が本を読む
  Tarou will read the manga. (future)
  #Tarou reads manga. (habitual doesn't make sense.)

Of course, you're always going to have proper context in a conversation. It's unlikely you're going to have a situation in which both "Tarou reads manga" and "Tarou will read the manga" make sense and you'd have trouble trying to disambiguate the meaning.

In practice, this is more useful to know when speaking Japanese. If you have a habitual sentence, you're going to have to mark the habit-doer as the topic. For example:

Manga: Horimiya ホリミヤ (Chapter 1)

• Context: Hori thinks Miyamura is a reserved, introverted person, who doesn't laugh a lot. Upon seeing Miyamura laugh a lot, Hori thinks:
• Miyamura tte kekkou
  warau-n-da...
  宮村って結構笑うんだ･･･
  Miyamura laughs a lot...
  • The tte って topic marker is used in this monologue instead of wa は because the new information contradicts Hori's previous assumptions about the topic.

It's reasonable to assume that you're never going to have the ga が particle marking the subject when you have a habitual.

Except when you have a question, because then ga が marks the focus.

• dare ga kekkou warau?
  誰が結構笑う？
  Who laughs a lot?
  • "Who" is focus.
  • "Laughs a lot" is topic.
• Miyamura ga kekkou warau
  宮村が結構笑う
  It is Miyamura who laughs a lot.
  • "Miyamura" is focus.
  • Topic same as above.

And except when you have a double subject construction in which the large subject is the topic and has been omitted probably because it's the speaker himself (i.e. "I," watashi wa 私は), and the small subject gets predicated by the habitual. For example:

• watashi wa {hara ga tatsu}
  私は腹が立つ
  {The stomach stands} is true about me. (literally.)
  My {stomach stands}.
  I {am angered}.
  • watashi wa - topicalized large subject.
  • hara ga - small subject.
• sore wa {hara ga tatsu}
  それは腹が立つ
  {The stomach stands} is true about that. (literally.)
  That {is angering.}
• ∅ {hara ga tatsu}
  腹が立つ
  [Something] {is angering}.
  • The phrase hara ga tatsu must be about something, in other words, it requires a large subject, which, if not uttered, presumably exists in the context.

And except when habituals occur with the ga が particle when they're uttered in surprise after finding out that someone or something is capable of doing something.

For example^{(尾上, 1982:21, as cited in 尾野, 1998:37)}:

• wa'! ningyou ga ugoku!
  わっ！人形が動く！
  Wah! The doll moves!
• wa'! nezumi ga shaberu!
  わっ！ねずみがしゃべる！
  Wah! The mouse speaks!

Well, that's a lot of exceptions, but I swear they make sense.

After all, in order to assert that "the mouse speaks" in the sense of "the mouse can speak," you only really need to observe it speaking once, so it should be perfectly fine as a stage-level predicate, in which the ga が expresses sentence focus, also known as the "news" ga が.

The basic idea that you can't observe multiple events at stage-level remains.

This is also observed with iteratives.

When ~te-iru ～ている expresses an iterative, an actualized habitual, it can only be interpreted as an ILP, so using the ga が particle would have the focus marking, "exhaustive listing" function. This is a bit tricky, however. Observe the examples below^{(Sugita, 2009:260–263)}:

1. Mari ga mai-nichi baiburu wo yonde-iru
   マリが毎日バイブルを読んでいる
   It is Mary who is reading the bible every day.
   
   • Habitual, subject focus.
2. mura no hitobito ga eiyou-shicchou de dondon shinde-iru
   村の人々が栄養失調でどんどん死んでいる
   The village people are dying from malnutrition one after the other.
   • Progressive, stage topic.

The sentences above, despite being in the same ~te-iru form, display different functions of the ga が particle, which means the first one is an ILP, while the second one is a SLP.

The first sentence expresses that the "read event" occurs multiple times. We can't observe this at stage level, so it must be ILP.

The second sentence appears to express something similar. If lots of people die, that means "die" occurs multiple times, therefore, we shouldn't be able to observe this at stage level either. However, there's a difference.

Everybody dies only once.

Presumably.

The fact that there are multiple occurrences of death doesn't change. What changes is that the second sentence isn't expressing a habitual meaning.

It doesn't express the repetition of a same event with a same subject^{(庵, 2001:82, Sugita, 2009:251–252)}.

It expresses that some people did die, and other people haven't died yet, but will die. These are all facts observable at stage level, so the combination of these facts ends up being observable at stage level as well.

Narrative Present

There are cases when the nonpast form is used for events that occurred in the past as a feature of narration.

The simplest case of this is when the to と particle means "then." For this to happen, the protasis must be in nonpast form, but the apodosis must be in past form.

For example^{(Hasegawa & Verschueren, 1998:4)}:

• mado wo akeru to, tori ga haitte-kita
  窓を開けると鳥が入ってきた
  [I] open the window then: a bird came in. (literally.)
  [I] opened the window then a bird came in. (same tense.)

In the sentence above, akeru translates to English as "opened," as the literal translation sounds too weird.

A more complicated case happens when telling stories in past tense, when the narrator creatively speaks in nonpast tense to give his impression on the facts.

For example^{(Takamura (1990) as cited in Hasegawa & Verschueren, 1998:4)}:

• Hirota-san wa hige no shita kara ha wo dashite waratta
  wariai ni kirei na ha wo motte-iru
  広田さんは髭の下から歯を出して笑った。
  わりあいにきれいな歯を持っている
  Hirota-san showed [his] teeth from under his mustache and smiled.
  ?[He] has rather pretty teeth.
  • The sentence is taking from the novel Sanshirou 三四郎. The text continues like this:
  • Sanshirou wa sono toki kyuu ni natsukashii kokoro-mochi ga shita.
    keredomo sono natsukashisa wa Mineko wo hanarete-iru.
    三四郎はその時急になつかしい心持ちがした。
    けれどもそのなつかしさは美禰子を離れている。
    At that time, Sanshirou suddenly felt nostalgic.
    ?But that nostalgia is apart from Mineko.

According to Takamura, ~te-iru is used in motte-iru instead of ~te-ita because it's a subjective observation made by the narrator, rather than an objective fact.

The sentence "he HAS rather pretty teeth" works as if the narrator is present somewhere in the scene, giving his impressions on what the Hirota-san character looks like.

Source: japanesewithanime.com (CC BY-SA 4.0)

Presumably, this creative narration would only work in present tense. When the past tense is used, the subjectivity is lost, simply because the rest of the story is also in past tense, so the sentence won't look any different from the rest.

English also has narrative present, also called historic present. Alternating tenses like above is highly unusual, but it's not impossible when you consider we're talking about creative works of fiction, and not a typical dialogue on the street.

The same sort of subjectivity also applies to habituals, exactly like the momentaneous observations we've seen previously:

• "oide ni narimasen ka" to kiku to, sensei wa sukoshi warai-nagara, mugon no mama kubi wo yoko ni futta.
  kodomo no you na shosa wo suru.
  「おいでになりませんか」と聞くと、先生は少し笑いながら、無言のまま首を横に振った。
  子供のような所作をする。
  After being asked "Won't you go [out]?" Sensei, smiling a bit, in silence shook [his] head.
  [He] behaves like a child.

Imperatives

In English, imperatives are expressed with the same verb form as a present tense. Observe:

• Robot: what's my purpose?
• Rick: pass the butter.
  • Imperative.
• Robot: *passes the butter*
• Robot: what's my purpose?
• Rick: you pass butter.
  • Habitual.
• Robot: oh my God.

In Japanese, there are several ways to make imperative sentences:

• damare
  黙れ
  Shut up.
  • meireikei 命令形.
• damatte
  黙って
  (same meaning.)
  • te-form.
• damari na
  黙りな
  (same meaning.)
  • na な sentence ending particle.
• damari-nasai
  黙りなさい
  (same meaning.)
  • nasai なさい form.
• damaru na
  黙るな
  Don't shut up.
  • The other na な sentence ending particle.

Besides the above, there are cases in which the nonpast tense and the past tense can be understood imperatively.

For example^{(高木一広, 1999:71)}:

• Context: a teacher orders an student during class.
• motto oogoe de yomu!
  もっと大声で読む！
  Read [it] with a more loud-voice!
  Read [it] louder!

Above, we have a sentence in the indicative mood as far as grammar is concerned. It's certainly not in the imperative mood. However, when the speaker states something that should be done, the listener might interpret it as something they have to do.

Such sentences are indirect imperatives.

A common example with the past form would be:

• matta!
  待った！
  *Waited! (literally.)
  Wait! (indirect imperative.)

Performative Verb Reportative Futurity

A performative verb is the verb of a performative utterance. They're used in the simple present in English, and in the nonpast form in Japanese. Observe:

• ore wa sore wo kyoka suru
  俺はそれを許可する
  I permit that.

Above, we have a speech act: a permission. By uttering the sentence above, I permit someone to do something. Such sentences are called performative utterances, and the verbs expressing the action performance performative verbs.

A performative verb only exists in a performative utterance, and a performative utterance only exists when you're doing something by saying something.

When you use the same verb to report what someone else is doing, the utterance no longer performs any action, so the verb is no longer a performative verb.

• I permit that.
  • By uttering this, I permit that.
  • So, here, "permit" is a performative verb.
• Tarou permits that.
  • By uttering this, I don't make Tarou permit anything. I'm just reporting what Tarou has permitted.
  • So, here, "permit" is NOT a performative verb.

In Japanese, performative verbs are used in the nonpast form. When those same verbs are not performative, they're treated like eventive verbs. Which is different from how English performative verbs are treated in reportative sentences.

Observe:

• Tarou wa sore wo kyoka suru
  太郎はそれを許可する
  *Tarou permits that. (wrong.)
  Tarou will permit that.

In order to obtain the English meaning, the ~te-iru form is required.

• Tarou wa sore wo kyoka shite-iru
  太郎はそれを許可している
  Tarou is permitting that.

Event Cancellation

In Japanese, there are cases in which accomplishments in past tense don't entail their telos was achieved, a phenomenon similar to the imperfective paradox, but way worse than it.

Such sentences are completely infelicitous when translated to English literally.

Observe the example below^{(Sugita, 2009:49)}:

• imouto wo okoshita kedo okinakatta
  妹を起こしたけど起きなかった
  #[I] woke up [my] sister, but [she] didn't wake up. (literally.)
  [I] tried to wake up [my] sister, but [she] didn't wake up. (felicitous translation.)

Above, we have the ergative verb pair okosu 起こす and okiru 起きる.

Normally, if I were to say "I woke up my sister" in English, that would entail its telos: I succeeded, and my sister woke up.

However, the example above creates a contradiction in English if translated literally, because it says my sister didn't woke up, even though I did wake her up.

That makes no sense at all.

Nevertheless, it works in Japanese, which means the Japanese past tense is fundamentally different from English past tense.

Specifically, in English the telos of an accomplishment in the perfective past is an entailment, so it can't be cancelled, while in Japanese it's an implicature, so it can be cancelled^{(Tsugimura, 2003, as cited in Sugita, 2009:50–51,63)}.

Counterfactuals

In English, counterfactuals are expressed through the subjunctive modal "would." In Japanese, the past tense is used in such cases.

For example^{(郡司, 2004:30)}:

• ano toki shinde-itara,
  あのとき死んでい { たら / れば }、
  If [I] had died at that time,
• kono koukei wo me ni suru koto wa dekinakatta.
  この光景を目にすることはできなかった。
  *[I] wasn't able to see this spectacle. (literally.)
  [I] wouldn't be able to see this spectacle. (subjunctive modal translation.)

References

1. Vendler, Z., 1957. Verbs and times. The philosophical review, 66(2), pp.143-160.
2. 鈴木重幸, 1965. 現代日本語の動詞のテンス―言いきりの述語に使われたばあい―. 国立国語研究所 ことばの研究, 2.
3. Goodman, F., 1973. On the semantics of futurate sentences.
4. Moens, M., 1987. Tense, aspect and temporal reference.
5. Declerck, R., 1988. Sequence of tenses in English. Fac. van de Letteren en de Wijsbegeerte, Univ. Campus.
6. Klein, W., 1992. The present perfect puzzle. Language, pp.525-552.
7. Bertinetto, P.M., 1994. Statives, progressives, and habituals: analogies and differences. Linguistics, 32(3), pp.391-424.
8. Krifka, M., Pelletier, F.J., Carlson, G., Ter Meulen, A., Chierchia, G. and Link, G., 1995. Genericity: an introduction.
9. Ogihara, T., 1995. Tense, Attitudes, And Scope.
10. Hasegawa, Y. and Verschueren, I.J., 1998. Tense-aspect controversy revisited: the-TA and-RU forms in Japanese. Pragmatics in, pp.255-40.
11. Sarkar, A., 1998. The conflict between future tense and modality: the case of will in English. University of Pennsylvania Working Papers in Linguistics, 5(2), p.6.
12. 岩崎卓, 1998. 連体修飾節のテンスについて. 日本語科学, 3, pp.47-66.
13. 尾野治彦, 1998. 瞬時的発話における 「ル」 形と 「タ」 形の使い分けについて: 認知のあり方をめぐって. 北海道武蔵女子短期大学紀要, 30, pp.31-74.
14. 高木一広, 1999. 日本語の命令表現に関する覚え書き. 九州工業大学情報工学部紀要 人間科学篇= Bulletin of the Faculty of Computer Science and Systems Engineering, Kyushu Institute of Technology. Human sciences, (12), pp.69-75.
15. Matsuzaki, T., 2001. Verb Meanings and Their Effects on Syntactic Behaviors: A Study with Special Reference to English and Japanese Ergative Pairs (Doctoral dissertation, University of Florida).
16. 庵功雄, 2001. テイル形, テイタ形の意味の捉え方に関する一試案. 一橋大学留学生センター紀要, 4, pp.75-94.
17. 池上素子, 2002. 「~ ようにする」 の意味特徴:「~ ようになる」「~ に/くする」 との比較から. 北海道大学留学生センター紀要, 6, pp.1-20.
18. 郡司隆男, 2004. 日本語のアスペクトと反実仮想. TALKS: Theoretical and applied linguistics at Kobe Shoin, 7, pp.21-34.
19. Copley, B., 2009. The semantics of the future. Routledge.
20. Hundt, M. and Smith, N., 2009. The present perfect in British and American English: Has there been any change, recently. ICAME journal, 33(1), pp.45-64.
21. Sugita, M., 2009. Japanese-TE IRU and-TE ARU: The aspectual implications of the stage-level and individual-level distinction. City University of New York.
22. Bertinetto, P.M. and Lenci, A., 2010. Iterativity vs. habituality (and gnomic imperfectivity). Quaderni del laboratorio di linguistica, 9(1), pp.1-46.
23. 朱薇娜, 2010. トキ節のテンス・アスペクトについて. 言葉と文化, p.303.
24. 鈴木彩香, 2014. ガ格の総記/中立叙述用法と裸名詞句の総称/存在解釈の統一的説明. 言語学論叢 オンライン版, (7).
25. 近藤かをり, 2018. テアル構文の統合的研究 ー主語性, 格配列, および文法化をめぐってー.