Please help me create Japanese _sentence_ diagrams for beginners

[NOTE: This top post was completely rewritten on 9/13/2022 10/1/2022! Early replies in the thread probably make little sense any more.]

Americans of a certain age might remember diagramming sentences while learning English grammar in elementary and middle-school. (I was one of those annoying nerds who loved it!)

This thread aims to develop a set of rules for how to diagram Japanese sentences. As far as I know, this has never been attempted before.

While still far from complete, we’ve made good progress. I’ve certainly learned a lot in the process, and I’m more convinced than ever that:

  • Diagrams like these are helpful for discussing and analyzing the syntax and structure of Japanese grammar.

  • We will eventually work out all the complexities and come up with a system that should work for pretty much any Japanese sentence.

The current state of the system is documented at . This site is evolving rapidly so expect frequent changes. (The commit log might be helpful to see what’s changed).

Please use this thread for any comments or suggestions.


Old version

I’m currently driving everyone nuts in the daily senryu thread by trying to create Japanese sentence diagrams to illustrate how my brain parses recent poems (mostly incorrectly!).

This thread aims to develop a set of rules for how to diagram Japanese sentences. As far as I know, this has never been attempted before. We are still learning as we go, so any and all help is still very much appreciated, but we’ve made much progress.

It’s already very clear that sentence diagrams aid in explaining Japanese grammatical concepts to beginners (including “beginners” like myself who have been speaking Japanese for years).

Sentence diagrams are completely jargon free. They only contain the individual words from the original Japanese sentence, as well as lines and various visual symbols to show how the various parts of a sentence relate to each other.

Just as in primary school, creating a sentence diagram can help you to parse a sentence as an aid to understanding. The also make it easier to communicate to others anything that might be confusing. Comparing and contrasting different diagrams can be particularly useful.

A sample diagram

Let’s start with a very simple example for the sentence, 本を読みました (“[I] read a book”).

We’d diagram it this way:

We put the original Japanese at the top of the diagram, followed (optionally) by an English transliteration and any notes regarding special nuances.

Transliteration vs. translation

By transliteration, we mean attempting to mimic the structure of the original while still keeping it somewhat readable (if awkward) in English. We mimic the structure, but not the order (both languages order things differently).

In this case, we’d normally “translate” the Japanese as “I read a book”, but this isn’t precisely correct.

The sentence might be about what the author just did, but it could also be talking about someone else. Maybe their spouse (“He read a book” or “She read a book”). Maybe a neighbor (“They read a book”). It might even mean “It read a book” (perhaps it’s in a story about a pet chimpanzee learning to read). It depends on context.

The zero pronoun

Japanese frequently uses a so-called “zero pronoun”. We’ve indicated the zero pronoun as @ in the transliteration.

The zero pronoun acts like “I”, “he”, “she”, “they”, “it”, etc. all rolled into one magic pronoun. Which one it means depends on context.

Further, and even more strangely for Westerners, the zero pronoun is never pronounced out loud nor written down at all! It’s only conceptually there. Magic.

The reason is that, just like English, every sentence must have a subject. It’s just that Japanese often “implies” the subject using the zero pronoun.

In Japanese, “something else goes here” would usually be written as ◯ , as in ◯が読んだ . We’ve used @ to make typing a little easier. Anything can go in the circle. We surround the @ in square brackets to indicate that it isn’t literally present in the original sentence, but this isn’t necessary if you are able to type ◯ .

By the way, it’s helpful to actually say まる out loud rather than “zero pronoun” if you need to read a transliteration aloud! In other words, I’d read this transliteration out loud as “‘Maru’ read a book”.[1]


So how do we know whether it was the author or somebody (or something else) that did the reading? It depends on the topic at hand (the surrounding context).

Unfortunately, like many sentences, this example doesn’t explicitly provide a topic for us. The meaning depends entirely on what other sentences came before, or what else is going when the sentence was spoken/written.

Topics provide context. They aren’t always explicitly stated. We will show how to diagram a sentence with an explicitly stated topic later, but this example sentence simply doesn’t have one: we are forced to guess who was doing the reading.

It’s reasonable to guess the author without any further context, so guessing “I” or “me” as the subject is perfectly reasonable. If you want to be accurate, though, the subject is ◯ (the zero pronoun).

Subjects and predicates (complete sentences/clauses)

Every complete sentence has a subject and a predicate, though. EVERY SENTENCE HAS A SUBJECT AND A PREDICATE.

The subject states who or what performed an action, and a predicate indicates what action is occurring.

In this first example, the subject isn’t stated, so it must be ◯, the zero pronoun. Again, think of it as “I/he/she/it/they” all rolled into one. The subject is a stand-in to represent something, or “Maru” the magic all-in-one-bucket pronoun.

The predicate tells us what action is occurring. In this case, Maru is reading. The predicate for this sentence is the 動詞(どうし) (“verb” in English)「読みました」.

Diagramming 101

Finally we can discuss the actual diagram!

First we must decide the “core” of the sentence: What is the core subject and what is the core predicate?

As discussed, ◯ must be the subject. There isn’t anything else indicating an action, so 読みました must be the “core” predicate in this sentence. Maru is reading. That’s the core of this sentence.

We start our diagram by creating a horizontal line with the primary subject on the left and the predicate on the right. The subject is separated from the predicate with a vertical line extending above and below the horizontal.

Normally, the particle が identifies a subject, so it’s placed immediately to the right of the vertical line. In this case, the が isn’t explicitly present in the original, so we put it in square brackets. In some sentences, the particle might be も or something else instead of が.

Notice how everything on the horizontal line is completely readable: ◯が読みました. This will always be the case, even with much more complex sentences.

Some sentences will have multiple clauses and thus multiple corresponding horizontal lines. Every clause on a horizontal line must always remain complete, readable, grammatical, and understandable clauses in their own right.


When we studied English grammar in primary school, we learned that the subject performs the action indicated in the predicate, while the action is done to an object.

You may have read that English typically uses SVO (subject verb object) ordering while Japanese is SOV (subject object verb) — “I throw a ball” vs. 「私がボールを投げる」.

This makes life difficult for simultaneous translators, but is almost irrelevant when it comes to diagramming Japanese!

In English, the position within a sentence identifies any object, but Japanese objects are identified by the particle を.

This means that objects do not need to reside on the horizontal line as they would in an English sentence diagram. They “hang below” the predicate instead, with the particle を attached to the “stem”, indicating in this case that it’s a book that’s being read.

Modifiers (particles)

Japanese grammar is actually pretty logical for the most part, and the particle system is awesome. Particles usually tell you exactly what a word is doing.

We’ve already covered を in this example. It tells you that (ほん) is an object (目的(もくてき)()). Since it comes right before the 動詞(どうし) 「読みました」, we know it’s a book that’s being read.

Other particles and related modifiers work similarly:

  • で usually indicates something is either a means or a location.

  • の usually indicates possession.

  • に usually indicates either time or a destination.

  • と usually indicates a co-participant.

  • は usually indicates a topic (we will cover this more later).

  • から usually indicates a starting point/time/date.

  • まで indicates indicates an ending point/time/date.

All of these things modify something else. We diagram them by hanging the below what they modify, in the same order as they appear in the original sentence.

A more complicated example

Let’s diagram 「(わたし)友達(ともだ)電車(でんしゃ)(なか)面白(おもしろ)(ほん)()みました」

At first glance that seems a lot more complicated, but the core of the sentence is exactly the same! It’s still our old friend Maru reading!

Let’s examine this piece by piece:

  1. Notice that we still have just one horizontal line, one clause, the core. The subject is still ◯, and the predicate is still 読みました. The core clause is still ◯が読みました. Because we have an explicit topic this time, though, we would read the core clause as 「私はよみました」— skipping over the imaginary zero pronoun and imaginary が. Everything on a left to right line should be read in order. The imaginary stuff is still there conceptually, but we don’t need to read it aloud now because we have the topic to tell us what’s going on.

  2. This time, though, we also have a topic indicated by は. The sentence starts with 私は so we know that unless a new topic is introduced, everything that follows has a context of 私, or “me”. English doesn’t often introduce topics like this so it’s a bit awkward, but the most common way to transliterate 私は is “as for me”.

    Note that it’s important to think “me” when you see 私は. Subjects, things performing actions, are indicated by が, so only 私が should be thought of as “I”. Topics only supply context, they don’t perform actions.

  3. Note that while we still have a subject of ◯ on the horizontal, our transliteration reads “[I] read a book” instead of “[@] read a book” . That’s because this time we have context, the explicit topic of “me”. We know the correct pronoun in English this time is “I”.

  4. Next is our first modifier. We know we are reading with a co-participant (a friend) because of the particle と. We could actually legitimately create a simpler version of the sentence that read ◯が友達と読みました and it would still be completely grammatical. The 友達と part just provides more information about the reading.

  5. Next comes something interesting: stacked modifiers. If we look ahead a little bit, we can see we are reading “inside” (of something) because of the particle で. ◯が中で読みました.

  6. What are we inside of? A train! It’s the train’s inside we are talking about. ◯が電車の中で読みました. Note that modifiers always hang below what they are modifying, regardless of where they appear in the original sentence!

  7. Finally our last modifier, the 形容詞(けいようし) (い-adjective)「面白い」. Note that 形容詞 modify nouns directly, without a particle on the “stem”.

And there you have it! The diagram should make it clear that at it’s heart it’s just about someone reading a book. It also makes it visually clear what parts are modifying more.


  • All the interesting bits in this diagram

  • More “interesting sentence” examples

  • More verb inflections.

  • Explain the dictionary dot notation for verbs in the predicate

  • Explain だ・です as a copula vs “hanging” as a politeness nuance for a sentence ending with a 形容詞.

  1. Most IMEs on a computer allow you to type the ◯ by typing maru then scrolling/selecting either the kanji or the particular circle you might want. ↩︎


I think, in English, you do an backslash-like line for copular sentences. Like this:

Since the diagrams aren’t fully geared toward word order, it shouldn’t matter that the です comes at the end. Just drop it where the English “is” is.

On the other hand, you could create a more Japanese-centric diagram style, which is probably what I’d do.


Perfect! Thank you.

I think I’m going to try creating diagrams for the tofugu grammar article examples. Hopefully, that effort will help me develop some consistent rules and processes. (Memories of an old boss: “Hope is not a strategy!”)


After following along on the Senryu thread, I think this is a brilliant idea. :slight_smile:

I spent a bit of time Googling around thinking that somebody had to have come up with something like this already, but I didn’t really find anything like what I think you’re after. So I think we just start with first principles and simple sentences then work our way up.

I think we have to separate out the は topic somehow, especially when we get to long sentences that reference the topic but never explicitly mention it again. And that brings up the question of how we deal with conjunctions and 連用形, but we can save that for later.

So, simple sentence:

My suggested start as far as rules:

  1. Topic on it’s own dashed line
  2. Subject-Object-Verb order separated by vertical lines. I don’t think we need to distinguish the copula in any special way. In English it gets conflated because we use being-verbs for both existence and as a copula. In Japanese it’s clear when it’s a copula.
  3. zero-が subject set off by parentheses (). I think it’s better to leave the zero-が as implied but that’s probably a topic for discussion. Is the zero-が always known from implication or are we inadvertently introducing new information by being explicit?
  4. Modifiers branch off from the word they modify

I suppose that’s enough to maybe get the discussion started. :+1:


Hooray! I’m so happy someone else likes the idea. I expect I’m going to waste oodles of yet more precious time on this…


Awesome. Topics (は connected) are precisely the bit I’ve been struggling with.

I really like your idea of two vertical lines, but after @rfindley’s reply, I’m wondering if topics (は) and subjects (が) should be separated by a full-length vertical that extends above and below, with は or が to the right of the line as appropriate, while a copula gets the slanted line as suggested:

(I’m unsure if they should slant forward or backward — I suppose it depends on what else we come up with.)

This seems to “visually parse” better to me. Topics and subjects do “feel” similar, so representing them similarly visually makes sense to me.

I suspect it might help to have different visual elements for different categories of connecting particles. Adjectives that attach to nouns are trivial, but coming up with rules for things that connect to verbs with に・で・と・を・から etc. might deserve different symbology.

This should be fun!


Yeah, that could work. My only worry is how busy it can get with complicated sentences. But then, if we’re just talking Senryu it’s manageable. So would a regular verb just get a vertical line that stops at the horizontal?

I like how you do it with the modifiers where the の particle is on the slanted line. What about having は・が・を・に at the top of the vertical line?

Edit: Actually, for things like に and と they would be on the slanted line to the modifier. So it would be just は・が・を that get the top of the line treatment.

Haha, aye. :heartpulse:

1 Like

You can have both a topic and subject, so I’d be inclined to use vertical line for subject and maybe a circle straddling the horizontal verb line to represent topic.

—topic–( )–subject–|–verb----

The circle is a good visual representation of ‘context’. It also lets you attach a ‘clause’ topic at any angle, which is handy since they can get pretty big.


I think the short vertical line will come in handy for direct objects or something else we come across later, but for the primary verb it’s unnecessary. I’m kind of thinking of the entire copula (this ≒ that) as one “verb” in prior sentences, but other verbs would look like this:

I like the idea of putting the は or が on top.

You’ve reminded me of an issue I ran into previously. Is it possible to have multiple topics? That’s when I started down the path of hanging topics down like other particle-connected things. If there can only be one topic (pretty sure there can only be one subject) the circle-connection thing might work.

Things fell apart on me with simple copulas using は: (わたし)はRexです.

I need examples of sentences with both a subject and a topic (or multiple topics if they exist).

Yeah, hopefully we can avoid the trap of trying to come up with something that works for any conceivable Japanese sentence. We just need something that works for the more common and simpler cases. I’ve no doubt we’ll run into examples that force us to come up with something new, but KISS principle definitely applies (always).


(Pardon, I’m on a tablet, and switching between J and E keyboards is a pain)

Mary-san ha neko ga imasu.

It can happen, but it’s usually in compound sentences, so you wouldn’t likely have two topics attached to the same circle.


I think it’s possible to have nested topics, but it may not be a scenario we need to worry about for now.

I like that suggestion better than mine.

Ah, ok. I was thinking of the English version where the direct object is on the same line as the verb and only the indirect objects go underneath.


I don’t think I’ve ever seen this kind of diagram in my life (not a native speaker though), but this is super interesting already. Can’t help much, but will keep watch!


Well, I`m embarassed to admit that I learned something today about Japanese grammar that I never realized before!

One difference between topics and subjects is that the topic can stay the same across multiple subjects (one per subordinate clause). That is, one topic, multiple subjects is likely common. Clearly @alo understood this, but I was puzzled by the “floating” topic in his initial diagram — it makes much more sense to me now.

I’ve been playing around with sample sentences from this tofugu article. They had the sentence 毎週日曜日はお父さんがお寿司を作る, to which I decided to add another clause, 楽しみに待つ.

Dotted-lines are also a thing with English sentence diagrams, so it occurred to me that we could do something like this:

That is, connect the topic with a dotted line to the same vertical lines used to mark the が subjects in each sub-clause “under” the same topic.

Thoughts? Does this work? (Not the greatest example sentence in the world, I realize, but I’m getting tired.)

I’m going to call it a night, but I kinda think something like this might actually work. Kinda riffs on @rfindley’s circles.

(By the way: is an easy way to create these diagrams if you create a free account.)

I had a head-smack realization as I was preparing for bed. I think one reason I got so much push-back on my example diagram in the top post is that 鼻ピアス is actually a topic with an implied は! The reason everyone hated the implied 〜がある is that it doesn’t belong, it’s just a regular topic on top of the implied subject 私.

These diagrams are already helping me to understand basic Japanese grammar significantly more than anything I’ve tried.


i remember low key hating diagramming sentences in 7th grade english class (even though i was really good at it and english was one of my strongest subject in school throughout). been since like 1990-91 since i’ve done any of this sort of stuff and i recall at the time thinking ‘ugh what use will this skill ever have??’. how ironic that now so many years later i’m extremely interested in seeing how this works for japanese and possibly being a good tool to break things down for me lol…guess i better apologize to my teacher from back then…


Personal opinion: I never did ‘sentence diagrams’ in school (they seem to be a US thing). I find tree diagrams much much easier to understand because they show you the actual structure of the sentence, and the labels tell you what part is what rather then relying on some weird line type conventions. Everything in this thread is just incomprehensible to me.


just out of interest, why are you separating は・が・を from the clauses they mark? I don’t really know how to read these diagrams, but it feels to me like it breaks from the structure to have them so strongly separated

also, は is a bit special but が・を behave the same way as the other particles. do they deserve to treated differently to preserve SOV?

also also, I’m interested to see how you fit も into this schema given it completely replaces は・が・を

it’s worse than that because one topic can stay the same across multiple sentences. thankfully I don’t think you can have more than one topic per sentence (although you can have multiple things marked with は) :stuck_out_tongue_closed_eyes:

I think @alo’s suggestion of having the topic floating separately without explicitly linking it to the clauses captures the topic-comment sentence structure better. it’s also neater since the everything in the comment links to the topic anyway.

Although it's probably rare enough to not worry about at this time, here are some examples of multiple は within a sentence for various reasons.

This one, the second は looks to be part of the subject clause.



(This one, the second は is within a quote.)


These are perfect! Thanks so much for the examples. The first one clearly shows that topics can nest, which I’d like to make clear as well. “As for me, when it comes to dirty money … “

Very cool. I’d have thought a new topic replaces the old one without concrete examples.

I’ll reply to the others points once I’ve had my coffee.

Thanks for the replies!


you’ve got to be careful because は doesn’t necessarily mark a topic…I’d say all of those examples only have a single topic, except for the last one because it’s quoting another sentence

for instance with the first example, the は in 「噂の君は」 is the topic marker and the ones in 「ヴェネツィアには」 and 「そんな怪談は」 are contrast/emphasis markers (indicating we’re talking about “in venice as opposed to anywhere else” and “that kind of ghost story as opposed to some other kind”).

this is definitely one of the places where japanese grammar gets a bit hairy and it’s why you’ve got to be careful saying things like 「今日はかわいい」(what to you mean I’m not cute on other days? :anguished: )

you can also mark the topic with も, which is inclusive rather than contrasting (so 今日もかわいい is a better shout). it also lets you talk about multiple topics (or maybe it’s better say combine multiple topics?). e.g. as a slightly forced example, 僕も君も腹が空いた


I was only really talking about this example (I didn’t actually read the rest until now :smile:):

あたし ゼッテー(よご)れた(かね) 使(つか)わないのが主義(しゅぎ)なんだ

That seems to be a clear example of nesting to me:

As for me ◯

enter the nested sub-clause in the circle

As for dirty money ◯

now push down to the actual subject clause in this circle

I have a strict rule to never use [it].

I think the dotted line thing may be able to make this clear in a diagram. I’m still working on how to visually represent it clearly.

I think I understand the point you’re making. In particular, your point that も is inclusive while は is contrasting/disambiguating/specifying is a good one.

My (possibly incorrect) understanding of the grammatical term “topic” is that it’s basically context. The over-simplified English loose-equivalent might be “as for ◯, …” where the circle marks a topic.

So I’m not sure I agree that all your examples aren’t topics as I understand them:

As for [being] in Venice, …

The topic is “in Venice”.

As for the rumor about him, …

The topic is “the rumor about him”.

(Please correct me if I’m misreading any of this.)

Fortunately, I’ve talked to the mods, and they’ve marked this thread as non-mandatory for participation.


I’ve written elsewhere about experts like yourself sometimes not being the best teachers. This is a thread for (and by) beginners trying to feel their way forward. Like it or not, over-simplifications are a time-honored and proven form of pedagogy.

You may enjoy this thread that I linked to in the OP much more than this one.


Oh yeah, I don’t mean to say “you shouldn’t do this”; just that I think this thread is going to work more for people like yourself who already happen to be experts in this weird diagram format.