Comment by vgel | Hacker News Reader

Comment by vgel | Hacker News Reader

Parent Full thread

Eh, you're right if you treat it as just clipping "morphological". But it doesn't sound like it to most people, they think "English and Japanese are in the same category of lacking a future tense" in some non-extremely-limited way. When really English is far closer to a language with morphological future tense than it is to Japanese.

Here's what I mean:

1. Go to this link: http://erg.delph-in.net/logon .

2. Type in "I will go to the store" and check "mrs" in the output list.

3. Click "Analyze"

4. In the MRS (right pane), under "go_v_1", hover over the "e3" of the "ARG0 e3" line.

5. You'll see "TENSE fut" in the feature box.

That's the result of a symbolic, machine grammar, not a machine-learning system. A rules-based system was able to determine that the verb is future tense (based on a syntactic rule that unifies in the future tense based on the presence of the auxiliary).

If aliens came by and zapped every English speaker's brain so that we all started using "I gowill to the store and buywill some fruits" instead of an auxiliary, that would be a simple adjustment to that grammar (delete the old syntax rule, add a will-lex rule to strip the suffix and apply the same feature to the verb).

If, then, another set of aliens came and zapped our brains again, deleting all memory of the concept of "will X" from our heads, that would be different. In the presence of the sentence "I go to the store tomorrow" or "I write code next year", there's no way for any symbolic grammar to handle all the possible ways you can contextually refer to the future. That's solely the domain of ML systems (and all their faults), at least as far as I know.

-------

By analogy, the way people talk about English future tense is like if people said "Python doesn't have closures with more than one statement", and when questioned "Well inline function definitions in Python have closure", they responded "Well of course I meant 'Python doesn't have expression closures with more than one statement'". Like sure, it's true and even a valid clipping to make in a certain context, but as it spreads as a meme people who don't know the context get the wrong idea.

> When really English is far closer to a language with morphological future tense than it is to Japanese.

I really don’t agree, as a speaker of both languages.

There’s a lot of western-exceptionalism invested in the notion that Japanese somehow has an exceptional paucity of ability to express future events that’s categorically different than English, when in reality the situations are quite quite close to one another.

It’s not that English doesn’t have a future tense in a casual/non-strictly-morphological sense of “tense” — it’s that by that casual definition really Japanese is in almost the same boat, and we should stop playing this “invent arbitrary ways to claim the inscrutable Japanese mind speaks without the ability to use a future tense but English is different because I Say So” game.

I don't have a strong knowledge of Japanese (year in college as part of linguistics degree, wife speaks it), so it's possible I'm mistaken about the nature of it grammatically. But my point isn't to make some Sapir-Whorf / Hopi Time-esque argument about its pragmatics, just to argue a point about the grammatical structure (whether the future tense is grammaticalized or not). That point might not hit exactly with Japanese if there's more grammaticalization of the future tense in Japanese than I thought, but my point is regardless to compare a language like English that has a grammaticalized future tense via auxiliary verb to languages that don't grammaticalize future tense.

To be clear, I completely disagree with the Sapir-Whorf hypothesis & don't think there's anything "weird" about Japanese, the past / nonpast split as opposed to past / present / future is quite common cross-linguistically[1]. The idea of whether something is grammaticalized is separate than whether a language can express it -- AFAIK the current theory is that all languages express just about the same set of concepts, just certain languages evolve grammaticalizations of tenses, aspects, topic, politeness, etc. and others represent them "ad hoc" via constructs like "do tomorrow" etc. Hopefully that makes sense. I find the "Mysterious Japanese mind" trope orientalist and stupid.

[1]: And there are lots of systems, like multiple levels of past remoteness, too: https://wals.info/feature/66A#2/26.7/149.2

> but my point is regardless to compare a language like English that has a grammaticalized future tense via auxiliary verb to languages that don't grammaticalize future tense.

These humongous walls of text sound rather like "and therefore I choose to define what English does as 'grammaticalize' and what Japanese does as 'not grammaticalize'".

> I find the "Mysterious Japanese mind" trope orientalist and stupid.

Then maybe have another think about why you're coming off as if that's exactly what you're doing.

There's nothing inherently "western-exceptional" in recognising that Indo-European languages are structured fundamentally differently than Japanese. That's just a basic fact of linguistic typology.

Obviously, nobody is claiming that Japanese has no way of talking about the future, that's the kind of nonsense non-linguists like to read into purely grammatical differences. It's still a fact that English has a grammatical future (actually more than one), even if it's not inflectional and not strictly required in all situations, and Japanese does not.

For example, "tomorrow, it's cold" would generally sound weird in English, the most common way to express this would be "tomorrow, it's gonna be cold". There's nothing weird about 明日は寒い, by contrast.

Reply on news.ycombinator.com