[louhike]

Just a nitpick but the particle “ha”/“は” is not really used to set the subject but the topic or context. The topic or context can often be the subject but it’s an important subtlety. The particle to set the subject is “ga”/“が”.

[Tor3]

A very important distinction which textbooks often fail to explain clearly. The subject is always and only が. Even if there's no が in the visible sentence.

[shiomiru]

Not sure if I understand this invisible が idea. Where is it in「私は牛肉を食べます」?

[Tor3]

It's called the zero-が and is before the object, in this case. Note that you don't need 私は for that sentence to be a fully valid sentence. It does not serve an actual grammatical function. The subject happens to be 私, and it is the unspoken subject. But every valid Japanese sentence which is not just an exclamation needs a subject, a "が"-marked subject, and if you can't see it you have a zero-が. For this sentence you would only use ”が” if you need to emphasize the subject, i.e. 私が牛肉を食べます - "it was I who..", or "I, in fact, .." - otherwise it's implicit, and not emphasized.

But the topic marker does not mark the subject. Any sentence with a topic can remove the topic, and the sentence is still a fully valid grammatical sentence.

For more about this, check out Cure Dolly's excellent videos about the zero-が on Youtube.

Edit: Various typos

[wst_]

It's implicit, depending on context. Quite often it's "myself" when omitted and without any context. If you'd like to stick to programming examples, it would be default value. Even if not specified, it's always there. Typically you use が when expressing some new information but follow up conversation uses は. There are few other rules, too. The は vs が is a subject of never ending confusion among Japanese students.

[shiomiru]

Indeed, my interpretation is also that the subject can be expressed by any of は, が, or the context. I am having trouble understanding the "implicit が" analysis because it seems to me that in many cases adding a が would make the sentence nonsensical (like in the example.) So it just feels like retrofitting が into a role that it really doesn't have. Then again, I'm not an expert, so I may very well be wrong about this.

[Tor3]

That's why it's called a zero-が if there's no "が", because every valid Japanese sentence needs a subject. It may sound like nitpicking, but it's important to clarify that it is not correct to say that "the subject can be expressed by any of は、が, or the context". Thinking about it that way only makes it difficult to understand where the particles should be used, and indeed the difference between は and が. Either the subject is explicit and marked with ”が" and only ”が", or it's implicit from the context. If it's unclear what the context is, you can clarify the context by introducing a topic with ”は", but that does not in any way make は itself a subject marker.

[shiomiru]

> It may sound like nitpicking... [snip]

You're entirely right, what I said there was incorrect (or sloppy at best.) Thank you for the correction.

> That's why it's called a zero-が if there's no "が", because every valid Japanese sentence needs a subject.

This is the part which I find unhelpful. You could say "every valid English sentence needs a subject and a verb", and you'd be having a hard time finding a sentence in e.g. the U.S. constitution that doesn't explicitly contain a subject and a verb.

But in Japanese, not every sentence needs to specify the subject. Every sentence still has a subject, because the definition of a sentence is "a subject and a predicate", but that applies to every human language.

The problem then is that the "zero-が" then feels like bolting a potentially ungrammatical (or at the very least unnatural) construct onto every sentence to explain the simple fact that yes, indeed, Japanese sentences do have a subject, but it's often unspecified. It also misses the important point that が, by clarifying the subject, also has the role of emphasizing it, which changes the meaning of the sentence.

To go on a bit of a tangent, Hungarian has a similar problem. It also frequently omits the subject, and specifying it also changes the meaning of the sentence:

* Lefekszem. =I'm going to sleep. (The subject is already specified by the conjugation.)

* Én lefekszem. =It's specifically me, who is going to sleep. (For example, you are asked if you're going to the after-party, but in contrast to others, you're tired and want to sleep.)

You could add an "implicit én" to the first sentence, to show that yes, it indeed does have a subject. Maybe it's helpful for people whose native language can't just omit the subject. But it subtly changes the meaning, so ultimately I don't think it's very good analysis, but rather "translating into your native language."

As I said, I'm not an expert, but I hope what I'm saying makes sense.

[ghc]

I recall the explanation of an excellent little Japanese grammar book that made it click for me: 私は puts the emphasis on what you, as opposed to everyone else is doing, and the closest english equivalent is "As for me,".

If you translate it that way, 「私は牛肉を食べます」means, in English: "As for me, I will eat beef."

Notice now that both "me" and "I" are present in the translation. So adding が doesn't in fact change the sentence at all!

[Tor3]

Yes, that's how e.g. Human Japanese (PC, Android etc) defines how the topic marker should be understood, and it definitely helps.

[karatinversion]

If the topic and subject are the same, just は is used, but there are many sentences where the two differ, eg. 「映画館は人が多かった」- “there were many people in the cinema”, where “people” is the subject and “cinema” the topic. Or 「ここはパンが売れる」- “they sell bread here”, where the topic is “here” and “bread” is the subject.

[louhike]

That may seem strange but that would be 「私が」私は牛乳を食べます

[shiomiru]

That's interesting. It indeed seems unnatural because I've never heard anybody say it like that - my impression is that you generally can't just use both particles with the same subject for the same sentence (or at least not like this.) But maybe that's just because of my inexperience. Do you have a real-life example of this pattern being used?

(It's also strange because I typoed 牛肉 in my earlier comment, and milk is difficult to eat. Sorry for the confusion.)

[Tor3]

Almost - you would write it 私は「私が」牛乳を食べます, and obviously 私は doesn't serve a grammatical purpose, it only has an indicative function.

[thrdbndndn]

I'm amused by the transition from beef to milk.

[shiomiru]

My bad. I blame autocomplete :)

[louhike]

Yes, thanks for the correction tor3!

[wodenokoto]

There’s nothing to understand. It’s not a thing.

[kijin]

Just another nitpick: "は" is pronounced as "wa" not "ha" when it is used as the topic marker, as in the well-known greeting "Konnichiwa".

[Nihilartikel]

I've always been amused that a literal translation of Konnichiwa could be something like: So, about today...

Kono(this) nichi(day) wa(is the topic)

[Izkata]

Likewise konbanwa is the same thing with "tonight" / "this evening"

[kijin]

English speakers also often drop the "good" in "good morning" and "good night", reducing the phrase to a single noun that refers roughly to the current time.

[midoridensha]

>English speakers also often drop the "good" in "good morning" and "good night", reducing the phrase to a single noun that refers roughly to the current time.

If I'm at work, I usually prefer to say "morning" without the "good": if it were a good morning, I wouldn't be at work...

[iopq]

it's short for

今日はご機嫌いかがですか？

if you're a rich girl at a private school you can just say ご機嫌よう

[louhike]

Yes but the romanji is still “ha”.

[layer8]

The correct word is “romaji” [0], and in the standard Hepburn romanization, “は” as a particle is written “wa” [1].

[0] https://en.wiktionary.org/wiki/ローマ字

[1] https://en.wikipedia.org/wiki/Hepburn_romanization#Particles

[hinkley]

Okay guys, that's enough 'well actually' for one day. Go touch some grass.