Breaking down a page of Japanese

Using the free sample pages from bookwalker, I printed out a few first pages of books I’ve read and have been working on non-exhaustively analyzing them. Lately at work I spend some time each shift out counting customers so we don’t go over capacity, and when it’s slow I have little to do. I can’t read a book, and I can’t do flashcards on my phone, but there is a clipboard out there. (None of my coworkers read Japanese, and a clipboard makes anything look official, right? Not that they would mind either way. :relaxed:)

コンビニ人間 attempt

I started with コンビニ人間. I’ll put the blank page below for easier reading than the marked copy, and also in case anyone wants to do something before seeing my attempt.

First page of コンビニ人間, unaltered

What came to mind at first was just breaking down the sentences into: nouns (including noun phrases), verbs, particles, adverbs, etc. After doing it on paper, I marked up a digital copy as it’s easier to see.

  1. nouns (including noun phrases) in red boxes
    • an orange line marks where the main noun (that is being modified) starts
  2. verbs (including all conjugation bits) and the copula with blue lines
  3. particles with dark purple dots
  4. adverbs with light purple lines
  5. adjectives (not in noun phrases) with yellow lines (I only had one)

I chose not to break down the noun phrases at this point. Marking any one thing in more than one way made the page a real mess.

There were a couple of parts I was confused about! When I read the sentence, I understood what was being said, but when it came down to marking it as one thing or another, my brain faltered.

Here's what I ended up with (first attempt)

Some of my doubts
  1. I decided 整然 was an adverb here, taking と, but I decided 勝手に was an adverb all together (including the particle). I am not sure why.
  2. ため is listed as a noun and an adverb depending on what it’s modifying, and here I think it’s modifying a verb, so I made it an adverb… And the stuff in front of it I guess stands alone? Even though it’s kind of modifying ため, no? (This is one of those things where in plain reading it wouldn’t bother me so long as I understood the sentence, but in analyzing it makes my brain hurt.)
  3. I assume in the grammar point verbと, the と still counts as a particle, right?
  4. Things with the nominalizer の, I marked の as the main noun being modified… I guess?

Those are only the things I have noticed enough to have doubts about. I bet there are things I’m simply messing up or overlooking, so if you see anything, please let me know!

Finally, I thought the most fun part of this page with its many, often long noun phrases was finding the main noun and figuring out what all modified it. Thus, I tried taking out all the modification, including the stray adverbs. I instead put words like ‘this,’ ‘that,’ ‘these,’ ‘those.’ Where I needed a modifier (like ‘my’) to have the sentence make sense in English, I kept it or added it in.

Rough English meaning stripped of modifiers

Convenience stores are full of noises. This noise, and that voice. That shout, and this noise. This noise, that noise, and the other noise. They all mixed together, became “the convenience store noise,” and were (constantly) brushing at my eardrums.

At that noise, I lifted my head. It was because there were a lot of those customers that I reacted, my body moving that way. After confirming that, I returned my gaze to my hands.

While gathering information from these noises, my body lined up those onigiri. At this time, it was onigiri, sandwiches, and salads. On the other side, that Sugawara-san was checking goods with a scanner. I lined up that food in that way. This mentaiko cheese in two rows in the middle, that tuna mayo to the sides in two rows, and that other onigiri on the end. Since speed was of the essence, those rules gave my body directions without me (hardly) using my brain.

A few days later, the final (?) marked-up page:

Note: For the clause ending in ため in column 7, I made the whole thing a noun phrase, with ため as the main noun. The whole thing is acting as an adverb, so I also put a little purple line by ため. Not a perfect solution, but it will do.

Thank you for any feedback! I’m doing this for a little brain exercise during dull moments, but I’d be happy to learn more about how this stuff works.


You planning to dive into the modifying phrases at any point? Murata-sensei does rather love those.


I could, and I’d enjoy it! I think this first page is a particularly great first page, and those noun phrases are a big part of it. But part of what makes this easy enough logistically to be fun is not having to copy it out or manipulate the text at all. I feel like to break down the modifying phrases, I’d have to take them from the page as a whole. Right now rather than doing that logistical work, I’d maybe rather just go to the next first page? :thinking:

1 Like

Wow, that’s amazing :slight_smile:
I don’t have much time rn, but one quick remark: you put “particle” dots next to all three of なので, but I think the で is not the particle but actually the copula だ in て-form, so maybe it qualifies more as a verb according to your marking system?


Ahhh, I thought of this and wasn’t sure. I mean, I do think ので’s で comes from the copula? But isn’t ので still also a particle (or particles)? I almost marked it as both, but it felt like cheating. :grin:


Hmm, that’s true indeed :thinking:

Probably this would lead to two different interpretations?
Seen as copula: Speed is the most important factor, and … gave directions to my body.
Seen as particle: Because speed is the most important factor, … gave directions to my body.

So yeah, the particle version makes much more sense in that sentence. :+1:

1 Like

Interesting, that is logical! I can see your interpretation. :thinking:


Hot take… there is no such thing as an adverb in Japanese

1 Like

Huh! What would you call ずっと if it’s not an adverb?

I’d call it an adverbial noun but I agree that in the small class of words like ずっと、もっと、ゆっくり、etc, calling them adverbs isn’t a terrible idea. But what I unfortunately see a lot of is calling things like 急に (adverbial nouns) or 優しく (adverbial verbs/i-adjectives) “adverbs” which is a flimsy and confusing model imo. Words like ずっと、もっと, ゆっくり are a relatively small set of nouns that have the power to drop the にorと particles that are used to form adverbial noun phrases (in fact, ずっと、もっと pretty clearly evolved from noun phrases ending in と). I find it simpler to just think of these as nouns that have adverbial English translations. This is mostly just to simplify the structure of my model (which is pretty similar to Cure Dolly’s model) - this way, you really only have to deal with nouns, verbs, i-adjectives, particles, conjunctions, and the copula.

In the end though, there is no objective true grammatical model, only more and less useful ones, and I don’t think the classification of adverb is a particularly useful one for Japanese.

The video i got this perspective from


Thank you for explaining! If it’s labeled 副詞 in Japanese, I’m going to continue calling it an adverb in English, but I can see where you’re coming from.

1 Like

Yes, but, we deal with adverbs all the time in English. They’re words that modify the verb, same as adjectives modify nouns. How does treating them as “nouns that behave funny” simplify things?

And how do you make an adverb in English? Well, a lot of the time, you take an adjective, like “kind”, and add a “ly” on the back - “kindly”. Same as 優しい becomes 優しく. The concept directly correlates.


Yeah, 副詞 is a defined part of speech in Japanese. What’s the point in pretending they don’t exist?


Because if we pretend everything is just nouns with different usages, it makes grammar so easy! Don’t know what word goes here? Use a noun!


I don’t think the model pretends that they don’t exist. Rather, it just redefines the term as an adverbial noun rather than an adverb.

It kinda does. At least for me.

Japanese is a very object oriented language to use a programming term. It’s less “noun” as part of speech and more “noun-thing” as a modifiable concept. If you further subclass those nouns you can build a grammatical model of Japanese that works as well as traditional models and maybe better in some cases for some people.

But again, there’s an intuitive leap that’s necessary and it’s definitely not the only model.

Basically this.

Thinking about a natural language in programming language terms just blew my mind. I’m not sure how useful the metaphor is, but either way I love it. Thanks for this!

‘Confusing’, I can understand, because you seem to prefer having fewer grammatical classes to deal with, but ‘flimsy’? Would you care to explain?

I’d just like you to consider the fact that 副詞(=adverb) is a separate class used in Japanese grammar. You’re free to use what you want, but why in the world would Japanese linguists adopt a foreign concept if it’s useless to their language? It might have been a concept inherited from Chinese grammar, but Chinese and Japanese use different terms for concepts as common as the ‘object’ (賓語 in Chinese; 目的語 in Japanese), so I strongly doubt the concept of adverb would have been kept around for nothing. The monolingual definition I’m looking at for 副詞 even provides sub-classes for it, so the concept has clearly been developed quite a bit by Japanese linguists. Also, if we’re talking about the ‘flimsiness’ of a model in the sense that it falls apart easily/has too many edge cases, consider this:

Japanese also contains words like 全然 and 突然, which can’t be used as anything other than adverbs and adjectives, in part because in the language they were taken from (Chinese, which I speak), they’re adverbs or adjectives and nothing else, which is really exceptional because Chinese words can change grammatical classes just by being put in the right place in a sentence, and without any particles being added. This is especially true for 全然 because of its meaning: I could probably still force 突然 to act as a noun in Chinese because it can function as an adjective, even if it’s more often an adverb. Classifying these things as nouns could tempt one to assume that they can function alone as the subjects or topics of sentences, whereas that doesn’t reflect actual Japanese usage at all.

As for these…

I don’t know about the origins of ずっと and もっと because I can’t find any information on them, but OK, fair enough, it seems likely that they’re some sort of と phrase. However, if you start looking into 〜っ〜り words, I think you’ll start to see a different picture…

Never mind whether or not calling what comes before と a ‘noun’ lines up with Japanese usage (i.e. whether or not they’re actually used as nouns in Japanese). Isn’t calling them nouns misrepresenting their true nature? Let’s leave ゆっくり aside. Take a look at other 〜っ〜り words like こっそり and ひっそり. Notice how they have corresponding non- 〜っ〜り forms like こそこそ and ひそひそ with very similar meanings? These things aren’t ‘noun phrases’. They’re onomatopoeia. They capture a certain impression, and tend to be used only adverbially because they are by nature a description of the manner in which something is happening. Why should they be lumped together with all other nouns? Yes, if you look into Classical Japanese poetry, you can find some support for the ‘noun phrase’ idea, but that only existed because と (the quotation particle) had a function back then that has since narrowed: it was once OK to use [noun]+と in order to create a simile and say that something behaved ‘like [noun]’. However, the difference is this: in modern Japanese, for true nouns, these ‘[noun]+と’ phrases can be converted into ‘[noun]のように’. In the case of ゆっくり(と)and company, however, this is not possible because these words are not nouns in their own right, but impressions.

Ultimately, just as you can say that the standard model used by most Japanese linguists introduces too many classes, I too can say that the ‘everything is a noun’ model (yes, this is a gross oversimplification that shouldn’t be taken literally, but you know what I’m referring to) risks causing enormous amounts of confusion by classifying things with very different grammatical characteristics all as nouns even though they can’t all be used in the same syntactic contexts. The only way out of this confusion is creating subcategories of nouns, or knowing that some of these ‘nouns’ must be followed by a particle, or that they must be considered as action modifiers (i.e. as adverbs). I think that effectively brings us back to the same set of problems that led to the rise of the adverb class. If, however, you feel that what I’ve raised above are not edge cases, or that they are not exceptions that merit consideration, and that your model that relies on understanding these things as nouns is more helpful for you, then I hope that that model will continue to serve you well and not cause you any frustration down the road.

I guess either classification is possible, but it seems that 整然 must take と in order to modify an action (based on the example sentences I can find), so it might be better to treat 整然と as a single unit and classify the entire thing as an adverb. It’s apparently a たる-adjective, so it can’t stand as a noun on its own. 勝手 can function as a noun on its own, so you could either classify it as a noun with a particle (に) that creates an adverbial phrase, or you could do what you did and treat the entire block as an adverb.

In order for the relative clause before it to have a sensible function (i.e. modifying a noun), I would treat ため as a noun. As for how to explain its adverbial function, I would remind myself that ため and ために are generally treated as having the same meaning, and that as such, に, the adverbial marker, has simply been omitted.

My understanding is that it’s technically a 接続助詞, which should translate as… ‘conjunction’, I think? It’s still a type of 助詞 though, so if you consider all 助詞 ‘particles’, there’s no problem with that classification from a technical standpoint.

Sounds reasonable to me. :slight_smile:


But adverbial nouns are a separate concept. Anything that is labeled as 副 in a dictionary is an actual adverb, not an adverbial noun. @Jonapedia already covered my opinion pretty well, but my point is that if you call adverbs “adverbial nouns”, but they act exactly the same as adverbs (i.e. they are used adverbially and can’t stand alone as nouns), all you’ve done is make it harder to differentiate two distinct things that aren’t actually used the same. Finally, I’d like to point to a dictionary entry on 結果, which starts with “副詞的に用いて”. That is actually an adverbial noun.

  1. I am going with your idea of including the と in 整然と, thank you!
  2. Yes, I’m counting all 助詞 as particles.
  3. When I was working on a different first page today, I had another very similar usage of ため that made me think about it again. I knew that ため got hooked up to by other things in the same way a noun does (若いため, 行くため, 静かなため, etc), so it’s easy to reclassify it in my mind as a noun. However, I now think it and the whole thing modifying it are an adverbial phrase, and maybe I want to treat that separately the way I’m doing noun phrases?? :thinking: Nouns and noun phrases together, adverbs and adverbial phrases together, sort of thing.

Also, unrelated to @Jonapedia’s helpful reply, the first page I was working on today was from the first volume of 本好きの下克上 and it is way harder to parse than the first page of コンビニ人間 was. :sweat_smile:


@Jonapedia and @seanblue I’ll try to write exactly what I mean when I get to a laptop because I think the confusion is that I’m not arguing against how it works, because that’s the reality the model has to fit, but rather the terminology and a different way of understanding that reality.

1 Like