Past, present, and future

« previous post | next post »

A couple of weeks ago, I wrote a post about the future:  "Mirai".

The ensuing discussion was quite animated, touching upon the nuances and implications of words for the future in many different languages.  I concluded by saying that I would write a separate post about past, present, and future:  here it is.

The first thing I feel compelled to say is that a language can have a concept of past, present, and future (though I'm not sure that all languages do) without having past, present, and future tense in the language itself (Japanese is one such language).  Furthermore, a language can lack grammatical tense for itself, but still be able to talk about grammatical tense in other languages (here again, Japanese is one such language).

I still remember when I was learning Mandarin nearly half a century ago that some teachers and authorities argued vehemently that Chinese "lacks tense", while others declared that "of course Chinese has tense".  For some, the debate still rages, but for many others, more sensible ways of looking at the question of verbal time have been developed.

It seems that Japanese grammarians agree that the language lacks tense, but has aspect.  Modern grammarians divide the Japanese verb into ru ル and ta タ aspects, the incomplete and the complete (from the speaker's perspective).

In Japanese, the grammatical tenses past, present, and future are known as kakokei, genzaikei, and miraikei 過去形・現在形・未来形.  However, these are not used for Japanese; they refer explicitly to the grammar (verb [etc.] forms) of English and other languages that do have tense.

Thus, modern Japanese verbs are marked for aspect, not tense.  Perfective and imperfective aspect forms are often call "present" and "past," but misunderstandings and misuses arise when people take these labels too literally.

You don't get discussions of Japanese verb morphology until Motoori Haruniwa in the early 19th century.  He did talk about transitive, intransitive, and causative, but I don't know if he also dealt with what we now call perfective and imperfective.  Anyway, since there aren't any well-known similar texts earlier, I doubt there were old precedents for these usages.

If you look at the Wikipedia entry on the subject, you’ll see that Wikipedia correctly eschews the term “tense” in describing Japanese, and characterizes the system as using aspect instead.  Thus, as is stated there, you have the “imperfective”, or “non-past”, vs. the “perfective” forms of the verb.

On the other hand, there is also the complex “native” system of describing Japanese.  Though that system is the traditional Japanese system and looks plenty exotic, it dates from either Meiji (1868-1912) or (at the earliest) late Edo (1603-1868), and incorporates features from Western grammatical analysis.  In any case, the only term from that system that might be of interest in our current discussion is kako 過去 ("past").  We've been talking about mirai 未来 ("future"), but that’s not a grammatical category in Japanese. If anything, that’s covered by the category “non-past”. If we want to talk about how the Japanese express "present" — that is, something equivalent to English “is VERBing” ― that’s a more complex subject.  The Japanese equivalent in the modern colloquial is VERB STEM-te iru.

In writing about English, the following clip from an oft-googled webpage is probably as reliable as any:

Jisei 時制  The Tenses

mirai jisei 未来時制 Future simple ("She will sleep soon.")
mirai shinkōkei 未来進行形 Future continuous ("She will be sleeping at 11 PM.")
mirai kanryōkei 未来完了形 Future perfect ("At 7 AM she will have slept for 8 hours.")
mirai kanryō shinkōkei 未来完了進行形 Future perfect continuous ("By 5 AM she will have been sleeping for 6 hours.")

genzai jisei 現在時制 Present simple ("She sleeps well.")
genzai shinkōkei 現在進行形 Present continuous ("She is sleeping right now.")
genzai kanryōkei 現在完了形 Present perfect ("She has slept well since she was a child.")
genzai kanryō shinkōkei 現在完了進行形 Present perfect continuous ("Tonight she has been sleeping soundly for two hours.")

kako jisei 過去時制 Past simple ("She slept for ten hours last night.")
kako shinkōkei 過去進行形 Past continuous ("She was sleeping when her husband came home.")
kako kanryō 過去完了 Past perfect ("This morning he said she had slept all night long.")
kako kanryō shinkōkei 過去完了進行形 Past perfect continuous ("She had been sleeping when the alarm clock rang.")

Remember, though, that these terms are referring to English tenses, not Japanese — past, present, future: kako(-kei) 過去(形), genzai(-kei) 現在(形), mirai(-kei) 未来(形).  By putting the final syllable -kei 形 ("form; shape; type") of each of these terms within parentheses, I'm indicating that they can be used both to talk about time and about tense in languages that really have it.

With only a perfect, imperfect, and a very easy progressive, (contemporary) Japanese verb tenses are so much simpler.

As grammatical terms, these are modern usages that were created to explain English grammar.  In Japanese, present and future verb forms are the same.  Time is not a major factor in determining the form of a verb.   Whether or not a certain action has been completed in a situation which the speaker is describing is more relevant.

For example:

(Watashi wa) ashita, kare ga kita toki ni mō ichido hanasu tsumori da.  (私は)明日彼が来時にもう一度話すつもりだ。 ("I intend to talk [to him] once more when he comes [here] tomorrow.")

The speaker is talking about his/her intended action (talk) in the future (tomorrow), but the verb in a subordinate clause (when=toki) takes the ta-form (来た/kita, not 来る/kuru).  It has to be ta because when the action of talking (話す/hanasu) takes place, his action of coming (here) has to be completed.


The Japanese grammatical concept of conjugation (katsuyō 活用) doesn't include the factor of time:

mizenkei 未然形 ("imperfective form")

This may not be the best gloss.  Literally, incomplete form; sometimes (pretentiously) called irrealis. Auxiliaries indicating negation, presumption, etc. attach to this stem — a real grab-bag of functions; therefore, a worthwhile generalization is hard to come up with.  In fact, historically, forms involving this stem were the result of a very early resegmentation:  Root+aX –> MZ+X.

ren'yōkei 連用形 ("conjunctive form")

Also called the continuative; the Yale term is infinitive.

shūshikei 終止形 ("predicative form")

Or conclusive.

rentaikei 連体形 ("attributive verb")

Or adnominal.

kateikei 仮定形 ("hypothetical / conditional form")

Traditionally called izenkei 已然形. Sometimes called realis, to contrast with the MZ. If you don't use "subjunctive" for something else (as Frellesvig does), then it is available for this form, and appropriate because, in premodern Japanese, it is used for subordinate clauses similar to Latin cum+subjunctive clauses.

meireikei 命令形 (" imperative form")

In premodern Japanese MZ+mu was the presumptive form, often referring to a future event; but there were other inflected forms that serve a similar function, such as SS/RT-ramu, which added a connotation of personal doubt.  More often than not, premodern -mu connoted "may do X, might do X" rather than predictive or admonitory "will do X," which was typically SS/RT-besi (with adjectival inflections). In the modern language, -mu > -u (nasalized in Middle Japanese), with -au > … > -oo. Verbs for which one would expect -eu or -iu have been restructured in this form: -eyoo, -iyoo. For examples, see Frellesvig's History of the Japanese Language.

[Indented notes by Jim Unger.]

Notes by David Lurie:

Terms like mizenkei 未然形 and izenkei 已然形 were systematized by modern national-language linguists (kokugogakusha 国語学者), but they originate in the Edo period and I suspect that in that context they had no connection to Western grammatical concepts like "tense".   For those who are interested in this, I'm pretty sure it is discussed in George Bedell's late-60's dissertation on early modern and modern Japanese grammatical concepts.

Additional notes by Bob Ramsey:

One can say something like:  iku desyoo 行くでしょう ("[He's] probably going"), using, say, the tentative of the copula.  That construction works with those categories of inference and guessing, just as you say.  But, even so, notice how the interpretation is not so much determined by grammar as by the real world and circumstances.  And by the way, doesn't all of this lack of time-oriented grammatical forms apply in spades to Chinese?   After all, one of the first things you learn in Chinese class is that Chinese doesn't have categories of tense, but rather of aspect.  Also, note that the situation about "future tense" is pretty much the same in English, isn't it?  For example, "I'll go" may be, in practical terms, about something you're going to do in the future, but it's not future tense.  Overtly, it's rather a statement of intent.  Closer perhaps is "I'm going to go", or when you're really making a flat-out statement about future action (w/o intent or any such thing mixed into the semantics), you say, "I'm going" — but that in turn can, in other circumstances, mean you've already set off, or it's something you do habitually.  I don't know about all this tense business.  "Future tense" seems to be an ill-defined category in a lot of languages — including, Japanese, Chinese, Korean, and also English — isn't that so?

Another way some have put it for Japanese is to say that verbs there don't have a miraikei 未来形 ("future form") as such.  Instead,  futurity is expressed in what may be called the suiryō 推量 ("guess; conjecture; infer") or suisatsu 推察 ("guess; surmise; conjecture") form.  Yet this statement is only partially correct, since you can also "guess" about present and past situations, thus strictly speaking you can't say suiryō 推量 = "future".  If one is looking for a Japanese miraikei 未来形 which corresponds to English miraikei 未来形, I think we have to say that there is none.  Time expressions like ashita 明日 ("tomorrow"), raishū 来週 ("next week"), rainen 来年 ("next year") denote futurity, viz., the tIme of the events. i.e., the time when the events will take place.

ashita, ame ga furu 明日、雨が降る。The speaker is quite certain that it will rain tomorrow.
ashita, ame ga furu darō (suiryō) 明日、雨が降るだろう(推量)。The speaker is guessing that it will rain tomorrow.

kinō, ame ga futta 昨日、雨が降った。The speaker is quite certain that it rained yesterday.
kinō, ame ga futta darō  昨日、雨が降っただろう。The speaker is guessing that it rained yesterday.

ima, ame ga futte iru 今、雨が降っている。The speaker is quite certain that it is raining now.
ima, ame ga futte iru darō 今、雨が降っているだろう。The speaker is guessing that it is raining now.

Now (!), what is fascinating is that the basic terms for past, present, and future — Japanese kako, genzai, mirai / Chinese guòqù, xiànzài, wèilái 過去, 現在, 未来 — were introduced to Sinitic vocabulary via Buddhism, which came to China beginning around the first century A.D.

In Sanskrit, the past, present, and future (there are major Buddhas for each of these three ages) are referred to as atītādhvan, pratyutpannādhvan, and anāgatādhvan.

Notes from Ron Davidson:

Buddhists tend to use these three words, and they are all the same kind of compound, a karmadhārya compound, composed of a noun (adhvan – time) and a past-passive participle operating as an adjective; atīta meaning gone, pratyutpanna meaning arisen, and anāgata meaning not yet come into being. The connotations for the future are rather general, “not yet happened” kind of thing, and in Buddhist use almost always tied to either a prophecy (vyākaraṇa) about a person — in the future he’ll be the Buddha XYX, etc. — or employed in an appeal for the benefit of those in the future — please speak about XYZ.

Notes from Bob Goldman

These three terms are all nominal compounds, the second element of with is adhvan, the basic meaning of which is “road,” “way", “path,” “journey”, etc. But it is used by the Buddhists and the Jains in the sense of time.

atīta is the past participle of the root i, “to go” with the prefix ati, meaning “beyond", "surpassing”, and so on. The form itself means "past", "gone away", "gone beyond", or even "passed away", as in dead.  It is commonly used, e.g., in the Buddhist Jātaka tales at the beginning, to introduce the "story of the past (life of the Buddha)”, and is used nominally, as when these tales begin by saying “atīte” — “in the past", “long ago”.

pratyutpanna is the past participle of the root pad and the prefix ut, utpanna meaning “arisen”, and the prefix prati — “near”, “towards”, “against"; it can mean several things, including “present”, “existing now”, “ready”.

anāgata is the past participle of the root gam, “to go”, with the prefix ā, “back” and the negative prefix an, “not”, yielding the sense “not (yet) come”, i.e., the future.

The above are standard words for time; they are not the words used for grammatical tenses in Sanskrit.  When intellectuals want to indicate tense, they would use Paninian technical vocabulary.

The other words that operate formally as tense markers are bhūta for past, vartamāna for present, and  bhaviṣyat for future (derived from verbs having the sense of "to be"), although these terms are sometimes employed disconnected from grammatical discourse, as in the Bhaviṣyapurāṇa, the Purana of the Future.

To sum up:

atītādhvan refers specifically to the "path that has been gone on"
pratyutpannādhvā means "the path that has arisen (and re-arisen)"
anāgatādhvan signifies "the path that has not come yet"

Each has some grammatical resonance in terms of the words often used to indicate past (atītā), present (pratyutpanna), and future (anāgatā).

If you are wondering why some authorities will cite forms like pratyutpannādhvā and others will cite forms like pratyutpannādhvan, both forms are virtually the same.  The first one is the nominative singular and can be treated as either a karmadhāraya compound (the path that has arisen presently) or as a bahuvrīhi compound (the one whose path/or for whom the path has arisen).  The second form is just the stem form and means the same thing.

Jan Nattier has a thought-provoking book entitled Once Upon a Future Time (1991) which is concerned with the narrations of prophecies about the future of the Buddhist faith.

For those who read Chinese and want to consult a scholarly study of how the currently used words for "past", "present", and "future" entered Sinitic vocabulary from Sanskrit as loan translations (fǎngyì 仿譯), I recommend this paper by Zhu Qingzhi:

朱慶之.  2000.   佛經翻譯中的仿譯及其對漢語詞彙的影響,載《中古近代漢語研究》第1輯(浙江大學漢語史研究中心編),上海:上海教育出版社.  頁 247-262。

Lest anyone think that it was impossible for Chinese to think or write about past, present, and future before the arrival of Buddhism, I need only point to the following pre-Buddhist terms:

xī 昔 ("past; former")

wǎng 往 ("past; previous")

nǎng 曩 ("in former times")

jīn 今 ("now; present; today")

lái 來 ("coming")

What is significant, however, is that all of these pre-Buddhist terms for expressing the time when something happened (and there are other ways as well) are monosyllabic, whereas the Buddhist-influenced terms for past, present, and future are all disyllabic.  In fact, the dramatic polysyllabicization of Sinitic vocabulary that began in the first millennium A.D. is intimately linked to the massive translation projects associated with the advent of Buddhism in China.

[Thanks to Ronald Davidson, Robert Goldman, Deven Patel, Dan Boucher, George Cardona, Linda Chance, James Unger, Bob Ramsey, David Lurie, Nathan Hopson, Cecilia Segawa Seigle, Hiroko Sherry, Tomoko Takami, Fangyi Cheng, Rebecca Fu, and Xiuyuan Mi]


  1. Leo E said,

    December 5, 2014 @ 1:50 am

    On the theme of Buddhism bringing in new ways of expressing time into Chinese, possibly as the result of massive translation projects, I was wondering if a similar phenomenon might have occurred in other languages which were repurposed for Buddhist and Sanskritic notions of time. Uyghur/Archaic Turkic does express tense, and after a very cursory look at a few pre-Buddhist sources alongside some sutras and commentaries there seem to be some significant changes in the way time is expressed, but not how tense in the language itself functions. For example, the phrases a) ärtmiš öd [-ki burhanlar]; b) kälmadük öd [-täkilär]; and c) közünür öd [-täkilär] appear in the Uyghur Mañjuśrī-Nāma-Saṃgīti to describe Buddhas of the a) past ("having passed"); b) future ("not having come"); and c) present ("manifesting, appearing") times. Again, this would take much more care to say anything that might correspond to reality, but references to the past in pre-Buddhist Turkic favor simply "ol ödke" ("at that time"), and for the present "bödüke/bu ödüke" ("at this time"). Likewise, the use of the word "bengü" ("eternal") seems to decrease in Buddhist texts (at least in the ones whose lexicons I have here) in favor of the distinctive descriptions of not-quite-eternal time involving billions of kalpas ("kalp" in Uyghur). The word for time, öd, is itself related to the word öt- for "passing through, penetrating, traversing," as in later ötgän zamanda = in a past time.
    Comparing Skt./Uyg. atīta/ärtmiš "past," pratyutpanna/közünür "present," and anāgata/kälmadük "future," the first and last pairs of which are literally the same, there does seem to be a close connection in how they're formed. Not exactly on the topic of the post other than the notion of the spread of Buddhist/Sanskritic expressions of time, but just something it made me curious about. I would love for someone more knowledgeable to point out any naive assumptions.

  2. Chris Kern said,

    December 5, 2014 @ 7:54 am

    Scholars earlier than the 19th century had ways of talking about grammar in relation to the Japanese classics, but they weren't codified. One passage from a 15th century Tale of Genji commentary says:
    "This is speaking of something that happened long ago, so there is no doubt that the 'shi' in 'chuujo nado ni monoshi tamahi shi" is the past 'shi'."
    The term 過去 is used to refer to the function of the し particle, although obviously their framework for such discussions was fairly thin compared to modern linguistics.

  3. Victor Mair said,

    December 5, 2014 @ 12:51 pm

    @Chris Kern

    Thank you very much for digging out that 15th-century occurrence of 過去 with reference to し. This is the same 過去 that I traced back to early Buddhist sources in Chinese. Fascinating!

  4. Chris Kern said,

    December 5, 2014 @ 2:21 pm

    The word 過去 was originally borrowed from Chinese Buddhist writings, of course.

    An even more interesting quote is given in the 日本国語大辞典 entry under 過去, from a book of poetry criticism or perhaps instructional techniques, written some time in the 15th century:
    "The various kinds of し include the past し, the future し, and the present し"
    I'm not sure what the writer means by a "future" or "present" し, but whatever the meaning is here, the words 過去, 未来, and 現在 are being used in their current vernacular meanings of "past", "future", and "present" in reference to grammatical usage. It's particularly interesting to see this three-fold division of time which would seem not to be influenced by any sort of Western grammatical ideas.

  5. julie lee said,

    December 5, 2014 @ 2:35 pm

    Thank you for this wonderful post on Past, Present, Future in Japanese and Chinese.

    Re. Chinese words indicating past, present, and future, I'd like to add two high-frequency words,

    yi 已 , indicating past, perfective
    jiang 將 indicating future, "will, shall, about to"

    I believe they are pre-Buddhist.

  6. Chris Kern said,

    December 5, 2014 @ 2:48 pm

    Incidentally, the reason the note above was written has to do with some strange elements of the early chapters of the Tale of Genji. The comment is for the beginning of Chapter 2, which states that "When Genji was still a captain…" Early commentators were puzzled by this since the explicit past "shi" is typically only used for referring to one's personal past, or in a narrative like this to refer back to things that have already happened in the story. This is indeed how the narratorial "shi" is used in the rest of the story, to refer to a person or event from a previous time period than the current narrative one. So some commentators tried to view this "shi" as a separate emphatic particle (which doesn't work grammatically here).

    This issue was almost abandoned in modern criticism because we aren't so bothered by the idea of a narrator saying something like "When Genji was still a captain" even though we haven't been yet told that he ever became a captain, much less progressed beyond this. However, an alternate theory is that Chapter 2 was composed after many of the other chapters, and so that this was indeed a callback to a previous time in the narrative than the readers had reached. Later the chapters were reordered by someone into chronological sequence. (There are other reasons to believe this than this one particle.)

  7. Victor Mair said,

    December 5, 2014 @ 2:50 pm

    From Xiuyuan Mi:

    I just finished reading the article. The idea of Buddhist influence on Chinese language is completely new to me. Just have a question: would you count phrases like "曩時" as disyllabic? The structure of 曩時 is similar to the example "x+ādhvan" you mention in the post, and this expression exists in 過秦論 (深謀遠慮,行軍用兵之道,非及曩時之士也). I think the author 賈誼 lived in 2nd century BC, though I'm not sure if it's the first instance of using 曩時, and how influential Buddhism was in early Western Han.

  8. Victor Mair said,

    December 5, 2014 @ 6:44 pm

    From Marcel Erdal:

    @Leo E

    "The word for time, öd, is itself related to the word öt- for "passing through, penetrating, traversing," as in later ötgän zamanda = in a past time."

    There are some errors here: Firstly, the word for 'time' is consistently written with ü in Brâhmî writing and there are quite a number of examples. It is true that some modern Turkic languages have ö but the Muslim Middle Turkic instances which Gabain quotes in TT VIII p.44 as evidence for ü are of course nonsensical as Arabic writing did not make the distinction between these vowels. The connection between the word for 'time' and the verb öt- (already made in Gabain's note which I just mentioned) is not convincing either: She proposes the etymology *öt-üt for 'time', but this noun clearly had a /d/ and not a /t/ and the consonants do NOT just alternate (and this /d/ becomes /y/ in most Turkic languages).

    Old Turkic including Old Uyghur had a past tense (-dim, -ding, -di etc.) and a present tense. The future is a complicated topic in language in general, being related to epistemic or deontic modality or coming from the present tense, and might get a book-long treatment for Turkic.

  9. Victor Mair said,

    December 5, 2014 @ 6:46 pm

    From Peter B. Golden:

    Marcel has it right and he is the major specialist in Old Turkic. I would add Hung. idő “time, weather” from West Old Turkic *üdeɣ (East Old Turkic *ödäk) < öd, üd discussed at length in A. Róna-Tas, Á. Berta, West Old Turkic. Turkic Loanwords in Hungarian, Turcologica 84 (Wiesbaden; Harrassowitz, 2011)I: 437-439

  10. DMT said,

    December 6, 2014 @ 4:35 am

    Would it be correct to say that 過去, 現在, 未来 were all originally shortened versions of 過去世, 現在世, 未来世? The latter terms seem to be more direct translations of the Sanskrit compounds, since they include 世 (age, era) as a translation of the element "adhvan" (time).

    Thinking about the subsequent history of these words, I can't recall having seen them much in pre-modern Chinese or Japanese sources (except as specifically Buddhist concepts). I'm guessing wildly here, but were they perhaps re-introduced into modern East Asian languages as Meiji philosophical vocabulary and subsequently imported back into Chinese?

  11. Victor Mair said,

    December 6, 2014 @ 11:03 am

    From George Cardona:

    I enjoyed reading your recent post. I also recalled finally where I had seen pratyutpanna used of present time. In the Kaccāyanavyākaraṇa Pāli grammar (ed. L. N. Tiwari, Varanasi 1962) one has 3.1.9: vattamānā paccuppanne ‘A (ending called) vattamānā [= vartamānā (vibhaktiḥ)] (occurs when) present (time is involved).’ Commentary: paccuppanne kāle vattamānā vibrate hoti.

  12. Victor Mair said,

    December 6, 2014 @ 11:06 am

    From Alexander Vovin:

    A couple of marginal notes to your very interesting posting.
    Actually, in Old Japanese there was a past tense in -ki1 (attributive -si), but in Middle Japanese it developed into retrospective, and then disappeared altogether.

    Verbal "bases" in Japanese are fatamorgana created by the inability of native philologists from kokugogaku school to describe the morphemic boundaries that are found inside a syllable. For details see my A REFERENCE GRAMMAR OF CLASSICAL JAPANESE PROSE, Routledge 2003, pp. 165 ff (hopefully you can get a copy that was not misprinted.

    My favorite example illustrating aspect-based nature of modern Japanese:

    People are waiting for the bus on a bus stop. The bus apppeares in the distance. This is described as "A, kita, kita!" in perfective aspect. When it is actually on the bus stop, this is described as kite [i]ru in the progressive aspect (which here has its perfective-progressive function).

  13. Duncan Mak said,

    December 7, 2014 @ 1:13 am

    I started studying Korean recently, and one thing of note is that, compared to Japanese, Korean does seem to have a future tense, the -l goshieyo and -l ggessyo ending.

    I'm still learning so my spelling /understanding might be incorrect, but I think the first form is more purely time-based, while the second form also conveys a sense of volition (similar to how 'will' is used in English).

    I'm sure someone more knowledgeable will be able to write up something more precise here.

  14. flow said,

    December 9, 2014 @ 4:35 pm

    @Chris Kern "particularly interesting to see this three-fold division of time which would seem not to be influenced by any sort of Western grammatical ideas"—certainly true, but then again the linguistic background where the ideas were formulated is an Indoeuropean one, too. i'd guess that the preference for a tripartite tense system (instead of, say, a bipartite aspect system) is a commonality amongst Indoeuropean languages, in which case those Buddhist ideas are certainly 'western' in a broader sense.

RSS feed for comments on this post