Please help me create Japanese _sentence_ diagrams for beginners

No worries. I’ve thick skin, and you clearly aren’t alone in your criticism.

I’ve not seen much defense from anyone but myself for the points I’m making in quite a while. That doesn’t make me wrong, but it can be a little disheartening at times.

To be clear, I’d like the end result to be beginner friendly. That doesn’t mean the process of creating something new is necessarily beginner friendly (that’s why the thread was an appeal for assistance from people more knowledgeable than myself).

Constructive criticism is always welcome.

“Maybe if they looked like this” or “What if we tried that” or “Here’s an example that invalidates the current hypothesis, here’s an alternative hypothesis that still works” are all great.

Criticisms like “I don’t understand this at all”, “That’s wrong”, or “Better left to experts” don’t really help.

In particular, I would welcome any corrections to my translation of the parts-of-speech diagram above (I think it’s pretty close). This is helping me to back up and re-think some parts of the current diagramming “rules”.

I won’t be deterred, though, regardless.

Honestly, I find “that’s best left to experts” more motivating than demotivating. I’d not have learned electronics, machining, woodworking, semiconductor design, coding, the game of go, or any number of other things without challenging myself and diving in as a beginner — including Japanese.

My preference, FWIW, is usually not to follow or participate in threads I disagree with, but to each their own.

1 Like

I understand what you mean, but I think there’s a difference between learning something and creating a resource (at least how I see it). Different types and level of knowledge needed. I consider myself fluent in english, but my competency for creating english learning resourses is probably dire. Portuguese is my native language, but I’d probably be the worst teacher and need to learn (or go to an uni about it lol) grammatical terms for the portuguese language.

Especially the bit about how native speakers think about that language: it’s a rather complex linguistic field, one that (at least in Brazil lol) you can have entire classes for. It just jumped to me as really big task, one you need a lot of technical knowledge of. Maybe only experts was an exaggeration, but it’s both a language and sociology field that research needs to be done, and done properly.

A “public” can’t be just for you, so one needs to be honest on how much you’re able to teach and understand what you’re teaching, if that makes sense. Especially if later on it will be then condensed for beginner learners.

If you plan on making it more beginner friendly at the end, then that’s fine. There was just no indication on if the current diagrams were close to final products (or how the final product should look) or not, so I had to make an assumption.

I think this is a fun project, and my only qualms really are on it being considered a proper beginners resource. Using it for your own learning (both in its construction and as a finished product later down) seems reasonable. Then again, these are just my own thoughts, because if I were in your place, I’d like to hear what others thing overall. If you analyze it and decide its not worth criticism to chase, that’s ok too, it’s just how these things roll.

Sorry if nothing makes sense I haven’t had my meds yet so I’m groggy -_-

2 Likes

Heh. No worries.

Apologies if these ever seemed “fully baked”. They definitely aren’t “ready for prime-time” as we say in the states.

I’m more confident than ever that I’m on track for something that will be useful for beginners, though.

They are definitely helping me understand more about the language, even in their current state. If they help me, they can help others. They also help me to communicate my ideas about the language (right OR wrong).

1 Like

FWIW, I think this is where most of the benefit will be – for you and others who contribute to work through the process of figuring out how various sentences fit together and what the different parts do. I’m less sure the end result will be something useful for others who weren’t part of its construction, but I could be wrong.

3 Likes

Hey, if you’re confident about it, then go ahead. My thoughts are just that, thoughts from what I’ve seen of this thread. You and everyone else are the ones working on it, so you know more about the status and the finish line about the project than me.

If it helps, even if it being a successful beginner resource fails, you yourself learned something and created a resource you can use as you go. Making your own guides are key to understanding a language, so like I said on my original post: have at it imo. It’ll be worthwhile for that for sure

4 Likes

We had a grammatically interesting poem today in the senryu thread:

「いつ買った」? 前からあったと シラを切る

I took a stab at diagramming it, and thought it worthwhile to walk through my thought process in gory detail, step by step.

1. How many sentences?

1. How many sentences?

First, we must recognize that there are actually two sentences here that need to be diagrammed independently.

It starts with a quoted question, a complete sentence in its own right: i

「いつ買った」?

One might roughly translate this as “When did you buy it?”

Next is a complex sentence containing two clauses: a main clause and a quoted clause:

前からあったと シラを切る

(しら)を切る is a set phrase meaning “to feign ignorance”, so this second sentence means something like “‘It’s been there a while’ — I feigned ignorantly”.

So this poem is the author reporting on a dialog. Since none of the subjects are explicitly included, it can be interpreted at least a couple different ways. For simplicity’s sake, let’s assume a scenario I, uh, might know something about:

A man’s wife walks into his shop and notices a new machine. “When did you buy that?” she asks.

“Oh, I’ve had that for a while” — I reply, feigning ignorance.

2. Sentence #1

2. Sentence #1

First, lets diagram :

いつ買った」?

The logical, semantic meaning is “When did [you] buy [it]?”

Note, however, that no words meaning “you” nor “it” are explicitly present in this sentence. English is much more explicit about subjects and objects, so we almost unconsciously add these two pronouns when translating.

Syntactically, the original sentence literally contains only two words:

  • one word meaning the past-tense of “buy”

  • and another meaning “when”.

Yet I’d argue this is still a complete, grammatically correct Japanese sentence (a single clause).

The “core” of this sentence is just the 動詞(どうし) 買った (“bought”).

To my way of thinking, every clause in any language ALWAYS has both a subject and a predicate. Someone or something performs an action or simply exists. Otherwise, it’s not a standalone thought, it needs something else.

Since I feel every clause must have a subject, I’d diagram it using the zero pronoun to stand in for the logical subject of “you”:

This is still read as just the single word 「買った」since everything else is implied (imagined) and in square brackets. None of the square bracket stuff is actually present in the original sentence, but we include them in the diagram to make all the syntactic parts of the clause expressly visible.

It’s not necessary in this example, but it gives us a place to hang any modifiers that affect the subject.

More importantly it makes it impossible to ignore that this clause could be about me/her/it/they/you/I/whosis/whatsit or any other “stand in” pronoun at all.

In this case, the subject is the zero pronoun: @. They don’t state who’s asking the question, but somebody certainly is.

With our interpretation above, @ is standing in for the personal pronoun “you” here, but it could just as well be standing in for “I” or even “they”, for example. There is no way to know without asking the author which was intended, but “you” is the most likely interpretation.

I’ve also shown か in brackets to indicate that it’s a question (arguably, I should have just included the question mark).

The question isn’t just asking whether I bought, though, it’s asking when:

The word いつ is modifying, or more precisely qualifying the core question. It adds the qualifying concept of “when”.

English pretty much mandates that we provide a subject, but is a little less stringent about objects.

If we were talking about a stock transaction, “When did you buy?” is perfectly grammatical, but in a scenario like this, most of us would append “it” to the end of the question.

I wouldn’t normally diagram an entirely imaginary modifier ([◯を]) to illustrate that there is also an implied object here, but to be completely explicit I’d diagram this entire sentence as follows, even though the imaginary modifier isn’t literally present and showing it doesn’t really add any information:

(I could probably have gotten away with diagramming the implied object as あれ instead of the zero pronoun, but who knows? Maybe this was poem was written by slave owners.)

You’d read this diagram as having a core clause of:「買った?」

Popping down to the next level of specificity, you’d read it as 「いつ買ったか?」

This is a terrible example to introduce someone to diagraming!

But it does make visible just how much is left out and merely implied in the original sentence!

3. Sentence #2

3. Sentence #2

The second sentence is

前からあったと シラを切る

As will (almost?) always be the case, the “core” is the predicate at the end, in this case the single 動詞(どうし)()る」.

Normally, this word means “cut”, but in this set expression, we can think of the core predicate as the English verb “feign”.

But who’s doing the feigning? Once again, it’s our old friend the zero pronoun (whom I’ve taken to calling “Maru” for short [as a nickname for ◯]):

What is Maru feigning? Maru is feigning ignorance:

The と indicates that the stuff at the beginning is an entirely new clause that’s being quoted.

We need a diagram for this new standalone clause:

前からあった

The core of this clause is once again the predicate at the end (“existed” in English) with our friend Maru as the subject:

Screen Shot 2022-09-21 at 3.50.21 PM

To pop down a level of specificity, it’s existed since before:

How do we show connect these two clauses together? We use the と to show that the entire second clause is quoted and provides further context to the “feigning”:

This structure of the entire second sentence effectively becomes

[@が] ◯と ◯を 切る

Where the subject is the zero pronoun (and implied), and the first ◯ contains an entire clause, not just a single word.

4. Putting it all together

So, in it’s entirety, we end up with this monstrosity for this senryu:

I recognize that this isn’t the easiest thing in the world to understand at a glance, but I hope that walking through the process of creating the diagram helps to explain why I think these are so useful.

To be fair, this is a very dense poem! There is a lot going on with three clauses and so many implied pronouns.

While the goal is to make Japanese sentences parsable, not to make pretty, visually-simple diagrams, I do admit that sentence diagrams are probably a more “intermediate” than “beginner” topic (and this specific example is probably at the advanced end of that spectrum).

5. Diagramming documentation

My thoughts on this topic have outgrown a thread-oriented forum.

I needed a way to publish more static (but still revisable) documentation. I’d like to show in detail what I’m currently thinking, rather than having a much-too-long thread showing the genesis of these thoughts.

I also need revision control and more formatting abilities than are feasible on a discourse site.

I’ve actually created a static website expressly to document my thoughts on Japanese sentence diagramming. I’ll publish the link here once I’ve cleaned up a few things. It’s coming along surprisingly nicely.

So it turns out I needed a LOT of words to fully explain my thoughts about Sentence Diagrams:

https://bunzu.doiwalters.com

I’m far from finished (I’ve only documented the very basics so far). I’ve changed a few minor things (like moving the が into the subject rather than next to the vertical line). I have a pretty good idea of how to continue with more complex sentences. There’s still much to cover, though.

Before I continue much further, I thought I’d give anyone still interested (one can dream!) a chance to review my work so far. Fair warning: it’s a lot to read (mea culpa).

A few things are still broken (OpenGraph tags and search come to mind) but it’s at least functional.

Any and all feedback is welcome.

よろしく

Maybe it would be nice to label differently the implied @ if they are several of them and they are repeated, to better keep track of them ? Something like:

「いつ[@1が][@2を]買った」? 前から[@2が]あったと[@1が] シラを切る

Edit: also, maybe a typo, but I’m not sure what’s the 知らぬ next to シラ, is it 白?

1 Like

I just deleted my original reply because I didn’t realize you were responding an earlier example (that I forgot I’d written!).

Your reply appeared immediately after my more recent post (confusing me terribly).

You raised an important point, though, so I want to respond to fully.

You’ve highlighted a mistake I made in my wordy post walking through creating a diagram.

In that post, I diagrammed a clause with two zero pronouns (BAD EXAMPLE, PLEASE IGNORE AFTER THIS):

I now believe that no more than one zero pronoun should ever appear in a single clause.

I further believe that sentence diagrams must never create a modifier out of whole cloth (like the left hanger above where everything is in brackets). The diagram above is an INCORRECTLY DIAGRAMMED SENTENCE.

Since no clause can ever have more than one logical subject, and imaginary modifiers like this are now prohibited, I don’t think the zero pronoun should ever appear more than once in a single clause. So @1, @2, etc. should never be necessary.

I’ve explained my current thinking on the topic in even more detail here: https://bunzu.doiwalters.com/en/basics/#dont-add-imaginary-modifiers

This is a weird colloquial expression. 白を切る or シラを切る means “to feign ignorance”. It has an interesting etymology.

The しら roughly means 知らぬ (pretending ignorance). It’s sometimes written (しら) as an example of ()() (“found” characters chosen for their sound). As explained in the article linked above, the 切る verb comes from other expressions involving speaking in a conspicuous manner.

So シラ alone basically means 知らぬ (pretending ignorance) but is sometimes written 白, while 切る means “conspicuously”. So the whole collquialism 白を切る means to conspicuously feign ignorance.

Whew!

Congrats for the new website btw!! :partying_face:

Americans of a certain age might remember diagramming sentences while learning English grammar in primary school.

Oh god. This made me realize that “sentence diagram” is an actual existing framework! (Reed–Kellogg diagram apparently?). And your entire project is to adapt those to Japanese. I thought you were creating them from scratch… This kind of diagram is not taught in my country so I didn’t notice.

In a previous post, I was wondering why you put the modifiers on a line going toward the bottom-right, which seems so counter-intuitive for a left branching language like Japanese, but now I get it, that’s just how Reed-Kellog sentence diagram are constructed and we can’t change it as we please.

Got it!
By the way, I wonder what are your thought about how to handle stuff like Verbてくれる・あげる? Especially Verbてもらう since the subject of Verb and もらう are different. It seems quite tricky to represent with sentence diagram. Especially the famous final boss もらってあげてくれませんか :smile:

Ah thanks, I forgot to check the etymology! It’s interesting that the ateji 白 is not completely arbitrary but could evoke stuff like 白々しい. And I kind of wonder if ultimately 知る (しる) and 白 (しろ・しら) could share a common etymology anyway, like in proto-Japanese or something. (and some Japanese people have the same question)

1 Like

Ooh! This is a great question and a good sub-project.

I’ve just been tackling things as I come across them in various grammar guides. This will be the next thing on my queue to consider.

I’ve got some ideas,[1] but this “boss” will definitely tickle out any issues with the foundation I’ve created so far.

That’s the first I’d heard of Reed-Kellog (though I see it in the Wikipedia article now), but yes, that’s precisely what I remember from childhood. Thanks for the name!

I wouldn’t say we can’t change anything, though. It just hadn’t occurred to me. I’ve been unable to find anything similar so I’ve pretty much just copied/stolen visual symbology as warranted from the English system.

Despite the grammatical hierarchy starting with what’s at the end of a sentence, Japanese is still written left to right (or top-to-bottom). I kinda like that the modifiers can be placed on the baseline of a clause in the same position they arrive when reading L-to-R. That’s not the case with Reed-Kellog where the order of modifiers isn’t specified. With 助詞(じょし) order of modifiers doesn’t really matter, either.

It’s increasingly obvious that very little of the English system remains, especially since Japanese uses 助詞(じょし) to identify things, so order is so much less important. Much of the English system is tied to the SVO nature of English.

About the only absolutely mandatory part of a Japanese sentence is the 述語(じゅつご) (even then, だ is sometimes implied!). And since ordering is mostly irrelevant, not much of the Reed-Kellog system applies.

In short, hanging things R-to-L might actually be a good idea. I need to ponder.

(What is your native country, by the way?)


  1. I see dotted lines in my future ↩︎

yay! i’ve always thought keeping it separate is a bit weird

are you mixing zero pronouns with zero subjects? you can absolutely have a zero subject and a zero object in the same clause (as your 買った example demonstrates).

also, what about when you have multiple clauses in a sentence?

I don’t think you have to worry too much about て chained verbs, particularly auxiliaries. the whole thing acts as a single verb anyway. with something like Vてもらう・くれる, the “doer” of the action is the same as the doer of the chained verb or it doesn’t make sense…

and with something like 見ていく, if you’re not explicit about the subjects/objects, I don’t think it’s possible to interpret them as different.

One thing that tends not to occur to beginners is the power of searching professionally translated sentences. Seeing how a word or particular grammatical construct was translated by a professional can be incredibly helpful.

The “Tanaka corupus” and the tatoeba project were developed expressly for this purpose.

I’d be careful with the tanaka corpus. it’s anything but professionally translated and it has a lot of issues (the site you linked also warns of this).

1 Like

This is the kind of thing I want to think about: “can you get it and give it to him for me” — three parties involved.

I wondered about this when I wrote it. The もらってあげてくれませんか “boss” clarified to me when it’s necessary. Still pondering.

I’m so glad the website is helping me to express myself better. Much better to have something I can continually refine and with more layout freedom.

Good point. I will update accordingly.

I’m not convinced (in this form) this is actually a thing people use and understand :sweat_smile:

the only references I’ve found to it are either learners asking what it means (with responses from natives ranging from confusion to attempting to break down the individual bits), or people who seem to be making fun of it like so:



but all that aside, going with your interpretation (take then give)[1], there’s still no problem with the subject or を marked object.

  • もらう: subject is the recipient, object is the thing being recieved.
  • あげる: subject is the giver (same as the recipient of もらう), object is the thing being given
  • くれませんか: subject is the doer of the favour (once again, same as the bits above). I’m not sure talking about objects makes sense here given it’s more of an expression than a straight verb…

strikes me that if you use this construction with more than one implicit subject or object, you have deeper problems than mere japanese grammar :stuck_out_tongue:

if you start looking at the indirect objects (marked with に), things might get interesting. I’m not sure introducing ◯に everywhere is helpful though…


  1. they seem to interpret it in the tweet as “take it as a favour to them as a favour to me” ↩︎

2 Likes

True, but examples of how the language can be twisted are ideal for figuring out diagramming rules. The whole point of diagrams is to show either the correct way to interpret something or multiple ways if there is ambiguity. The corner cases are the interesting bits.

It does seem to be a compound sentence with multiple logical subjects and objects, though none of them are explicitly stated.

Yeah, this was my feeling as well after reviewing the prior example where I introduced a wholly missing object. It added confusing complexity with absolutely no payoff.

Translating into English often forces us to add a pronoun as an object, but that’s due to English grammar rules – it shouldn’t be part of diagramming a Japanese sentence unless it’s explicitly present or somehow necessary to explain how the sentence works.

I’m still pondering, but I wonder if the only exception might be when an implied object is shared between different sub-clauses. Likely a rarity but something to consider at least.

Yeah, this gets to the heart of it, which was @Arzar33’s question, too: “how to handle stuff like Verbてくれる・あげる?”

Tofugu has no less than 9 separate articles on 〜て<用言> constructions plus one more on the 〜れる potential form. Not to mention all the Japanese 中学校 content to wade through.

The most basic question is “do て-extensions/modifications belong on the baseline of the clause or should they be visually indicated as some sort of modifier?”

I’m okay with just tacking 助詞(じょし) at the end of a sentence (よ, ね, な) onto the predicate since they don’t do much structurally. But just tacking on long strings of て-connected clauses feels like we are punting. :wink: It feels like they need some sort of visual treatment.

I need to think about it further.

I refer you back to my breakdown - it only has a single subject/object and it stops making sense if you add more. and people are more likely to interpret it as a single clause with multiple auxiliary verbs.

if you apply this consistently, then you shouldn’t do it with the subject either…personally I’d say that the subject and direct object are special enough to include if you’re marking things that are absent (also they can both be は marked if they’re the same as the topic)

so there’s three cases with て:

  1. linked clauses - in this case you have two separate clauses that can exist on their own. each has its own subject.
  2. the chained verb is a 補助動詞 - you have a single clause (which means you have a single subject), with the predicate being modified by the 補助動詞.
  3. it’s a compound word such as 見て取る

I think 1 is a separate clause and so gets its own line, 2 is similar enough to 助動詞 to be treated the same way, and 3 is a single word anyway

1 Like

My take is that in 持ってあげる,
あげる is the main verb, and もって describes the verb adverbially. Just like in
歩いていく。 (I will go (“walkingly”)) or
泣いて起きた。(I woke up (“cryingly”))

This is similar to the adverbial participles used in English like:
Sleeping, he missed the meeting. (Additional words make the English more clear "Sleeping late, he missed the meeting)

Edit:
I started writing an edit explaining I don’t think this always applies to 〜て form, but I changed my mind. I think 〜て form always functions at least somewhat adverbially to the main verb at the end of the sentence (or implied verb from an inversion). Whether it’s functioning to show temporal order, logical order, means, manner, etc. these are all standard uses of adverbs and adverbial participles. My take (that’s developing as I write this) is that all 〜て form verbs are basically adverbial participles. Totally willing to back down from this position since I’m developing it as I’m writing, but I can’t think of any exceptions! I think I used to tend to think of 〜て form as merely creating compound sentences like in English, but I’m increasingly becoming convinced that this is simply because we don’t use long detailed adverbial participial phrases in English. The final verb (or implied verb) is still always the most important verb seen in light of the “participial” 〜て form verb.

1 Like

I’m not following:

The logical subject is the same for all three (“same as the recipient of もらう” as you said).

But you’ve used the words “recipient”, “giver”, and “doer” to represent three different syntactical/structural subjects. Diagrams show syntax.

I agree about not introducing an object: there isn’t one here.

I feel strongly that subjects are different. Every clause always has a logical subject, and diagrams should make it visible when it’s implied. Topics are a whole lot easier to reason about with a zero pronoun as the subject in sentences that omit an explicit subject.

Not every predicate has an object, explicit or implied. And adding an implied one because English wants one isn’t a sufficiently good excuse.

As I wrote in the section on not adding imaginary modifiers:

In other words: It’s correct and necessary to add imaginary subjects because every 文節 must have a logical subject, but don’t add imaginary objects or other modifiers to a Japanese sentence diagram. Only diagram the modifiers that are actually present in the original sentence.

I’m leaning this way, too. Which argues for some visual indication of modification rather than just tacking them onto the verb.

I like this. Need to think about it.

くれませんか has no subject or object - it’s pretty unambiguously a 補助動詞 here.

for the other two, there is semantically a single subject and a single object. the verbs are transitive and they don’t make sense without them. syntactically there is no subject or object. that’s why we introduce zero pronouns!

now, the context for my response was the thing around whether you need @1 and @2. if you’re only doing zero subjects, in a sentence like もらってあげてくれませんか you only have one that’s repeated everywhere if you really want to analyse it like that.

the more usual way of interpretting the first two is that あげる and いく are auxilliary verbs that add more meaning to the verb they come after (much like ない, for instance). if we’re interpreting them as proper verbs though, then they’re two linked clauses (see below).

泣いて起きた is an interesting one but mostly because て is so flexible and the meanings of the two verbs confuse things. て linked clauses can be related in a bunch of ways, including “A then B”, “A allowed/led to B” or “A and at the same time B”.

there’s not really anything adverbial happening in these cases.

  • 持ってあげる[1] - held it and then gave it/holding it allowed me to give it
  • 歩いていく - walking allowed/led to me going (bit awkward but you get the point)
  • 泣いて起きた - cried[2] and then woke up/cried and at the same time woke up

  1. I hate that we’re using this as a proper verb here. it’s unnatural and confuses things ↩︎

  2. in a dream? while sleeping? ↩︎

1 Like

Gotta run, but I will note that your final examples would be translated (not transliterated) as

  • give while holding

  • go by walking

  • wake crying

What is that called in English grammar? More relevant - how are they diagrammed in Reed-Kellog?

I’ll research when I’m back.