Log: ryanheise tries to solve Japanese grammar

Continue or start your personal language log here, including logs for challenge participants
User avatar
ryanheise
Green Belt
Posts: 294
Joined: Tue Jun 04, 2019 3:13 pm
Location: Australia
Languages: English (N), Japanese (beginner)
x 911
Contact:

Log: ryanheise tries to solve Japanese grammar

Postby ryanheise » Thu Mar 12, 2020 3:59 pm

In this log, I plan to journal my experience trying to solve the puzzle of Japanese grammar. My approach with Japanese has been to get lots of input and notice the patterns, rather than read textbooks or grammar books that teach the patterns. The process of noticing patterns, however, is a mysterious one, and the successful learners and advocates of this approach, for better or for worse, tend not to offer much in the way of advice on "how" to notice.

There may be something to the idea that "noticing patterns" is what our brains naturally do when you put examples in front of it. This is what Edward de Bono's central thesis was: that the brain is by its nature a pattern seeking machine -- and that creative and lateral thinking arises by breaking the default pattern seeking behaviour and throwing randomness and provocation into the mix.

Our best machine learning models, including neural networks (which are inspired by the way our own brains' biological neural networks function), work by exposure to huge sets of examples, and learning is inevitable from the nature of the system.

But I think you can also get better at noticing patterns. I've always loved puzzle solving and always thought that having a guidebook on how to solve the puzzle defeated the purpose and joy of the exercise, and that you can get better at solving puzzles merely by "doing". Language can be enjoyable in the same way, but I understand that many language learners would prefer to have the guide book, and there is not one right way to go about it.

This journal will hopefully serve two purposes. One is to keep a record of every mysterious language feature that has ever puzzled me, the example sentences where I encountered them, and my thoughts about them, so that when I encounter another sentence in the future with a similar grammatical feature or word or colocation, I'll be able to go back and compare the two sentences and connect some dots. The other purpose is to share my experiences with others who may be interested in how I am going about noticing.

Where am I currently? I feel that I have a pretty solid intuition for the majority of Japanese grammar, but there are still a lot of basic things I haven't noticed yet or have only partial intuitions for, like the different situations in which different conditionals are used, and all of the distinct meanings of certain homonyms (particularly those that have a grammatical function). Aside from keeping this journal, my two other strategies which will hopefully complement this are to increase my daily input so that I can get exposed to more examples, and to just generally have a high tolerance threshold for things that I won't necessarily understand until weeks or months later when enough examples come in.

(Note: I am keeping a separate log linked in the footer for language learning experiments.)
8 x

User avatar
devilyoudont
Green Belt
Posts: 467
Joined: Tue Jun 26, 2018 1:34 am
Location: Philadelphia
Languages: EN (N), EO (C), JA (B), ES (a mess), KO (dabbling)
Language Log: https://forum.language-learners.org/vie ... 15&t=12245
x 1287

Re: Log: ryanheise tries to solve Japanese grammar

Postby devilyoudont » Fri Mar 13, 2020 10:16 pm

Are you interested in explanations for things you list here, or would you rather just catalogue them and eventually notice the pattern yourself?
1 x
Joyo Kanji: 1320 / 2136 Japanese Intermediate 2: 4654 / 12000

User avatar
ryanheise
Green Belt
Posts: 294
Joined: Tue Jun 04, 2019 3:13 pm
Location: Australia
Languages: English (N), Japanese (beginner)
x 911
Contact:

Re: Log: ryanheise tries to solve Japanese grammar

Postby ryanheise » Sat Mar 14, 2020 1:15 am

That's a really good question because I wouldn't want to close myself off to help from others. Though, my goal is to figure things out from just examples(*), and with examples being my only intended source of input, I would still happily welcome more examples if you or others would like to suggest some.

I will probably post one update per week collecting the noteworthy sentences encountered during that week and a snapshot of my current thoughts and hypotheses about them. My hypotheses are bound to be wrong many times, but the idea is to self-correct by continually re-evaluating as new examples come in.

(*) I'm specifically referring to grammar here. I still freely look up word definitions in dictionaries, but what I'm figuring out myself are things like sentence structures, verb conjugations, etc. My goal is also not at all to figure out the formal rule that you would find in a text book, either. I'm just looking for patterns or heuristics based on the examples in my head. These patterns may not perfectly map onto parts of speech and formal grammar rules, etc., but to me it doesn't really matter, as I'm satisfied with whatever patterns give me a functional command of the language when listening and speaking.
2 x

User avatar
tungemål
Green Belt
Posts: 439
Joined: Sat Apr 06, 2019 3:56 pm
Location: Norway
Languages: Norwegian (N)
English, German, Spanish, Japanese, Dutch, Polish
x 720

Re: Log: ryanheise tries to solve Japanese grammar

Postby tungemål » Sun Mar 15, 2020 11:25 am

What is the reason for doing it this way? Is it because it in your opinion is a better way to learn Japanese? (I am not saying that it isn't). Japanese grammar is well understood and documented in abundance in books and on internet. Besides it is very regular. So it seems to be a lot of extra work to figure out things yourself instead of looking up things that you wonder about.
2 x

User avatar
ryanheise
Green Belt
Posts: 294
Joined: Tue Jun 04, 2019 3:13 pm
Location: Australia
Languages: English (N), Japanese (beginner)
x 911
Contact:

Re: Log: ryanheise tries to solve Japanese grammar

Postby ryanheise » Sun Mar 15, 2020 3:38 pm

tungemål wrote:What is the reason for doing it this way? Is it because it in your opinion is a better way to learn Japanese? (I am not saying that it isn't). Japanese grammar is well understood and documented in abundance in books and on internet. Besides it is very regular. So it seems to be a lot of extra work to figure out things yourself instead of looking up things that you wonder about.


It is an interesting question. I would start by saying that if something truly is regular, I will notice it pretty quickly, and I suspect I would be able to formulate the precise grammar rule if you asked me to convert my intuitions into a rule (not that there would be any need to). But as for everything else, i.e. most things, things where there are subtleties and nuances, the point is that I'm not aiming to arrive at the same sorts of strict grammatical classifications that one might find in a grammar book or text book, because the real grammar that people actually speak is nowhere near fully accounted for in these strict classifications. I'm instead after a more fuzzy intuition that tells me what "feels" correct and what does not. A fuzzy grammar that isn't explained by well defined rules, but rather by multiple hypotheses that are given different weights. For example, I actually don't want to know the textbook definition of when you should use が vs は, I'd rather develop a feeling for it through massive numbers of examples. Despite having never learnt an official rule for this, I do feel that I have a pretty good intuition for it, and if I were to try to write down those intuitions in the form of a rule and make things a bit more concrete, it would not be just a single rule, and each alternative rule I could come up with would only capture a fraction of the full picture that my intuition represents.

When dealing with parts of the language like this that have subtleties and nuances, I prefer the fuzzy view of grammar based on intuition drawn from lots of examples because you can react quickly based on intuitions and that is good for fluency.

Also, going back to the machine learning analogy, and I think this is a helpful analogy, for decades, humans have been designing algorithms to do natural language processing, and lots of other things. The human programmer tries to identify all the cases and all the rules and explicitly program them into the system. It is possible to get quite accurate results with this approach, but it is blown out of the water by machine learning which takes massive numbers of examples, and finds patterns based on weights. What it finds is not precise rules, as the human programmer might hope to see, but instead finds fuzzy and overlapping hypotheses all mixed together. The end result of this fuzzy mix is a model of reality that is much MUCH more accurate than anything that someone could have come up with by looking for well-defined, strict rules.

Hopefully my approach will become clear as I start posting. There's nothing amazing about it, I'll just be writing down my evolving hypotheses. Not a strict grammar, but a growing set of fuzzy and overlapping intuitions that will have greater and greater coverage of reality over time.
2 x

User avatar
ryanheise
Green Belt
Posts: 294
Joined: Tue Jun 04, 2019 3:13 pm
Location: Australia
Languages: English (N), Japanese (beginner)
x 911
Contact:

Re: Log: ryanheise tries to solve Japanese grammar

Postby ryanheise » Sat Mar 21, 2020 11:24 am

Figuring out the verb conjugations

Given my approach to Japanese so far, what I've found is that I only output verb conjugations in speech that I'm familiar with because I've heard that verb used many times before in different ways, and I can hear it in my head, or because a particular verb intuitively feels like another verb that I'm familiar with and then I conjugate it in the same way. But there are many verbs that I don't know how to conjugate because I haven't heard them used enough and I haven't taken notice of how their conjugations go.

So I'd like to finally try to map out a verb conjugation table based on my own current familiarity, and use that to identify where the gaps are in my knowledge. I think that will help me to direct my listening attention in the future to look out for examples of those conjugations.

The basic present tenses

I'm going to start off with the basic present tenses for now and move onto other conjugations later. For example, with go / 行く, we have 行きます, (polite) 行きません (polite negative) and 行かない (negative). Or in romaji: iku, ikimasu, ikimasen, ikanai. The pattern here is that when adding -masu or -masen, the last u in iku changes to an i. When adding nai, the last u changes to an a. Most verbs work like this, but others don't.

My first intuition in mapping out what I know and what I don't know was to categorise verbs by their last syllable, compare a couple of examples for each of these, and write out how I would conjugate them (to the extent that I have an intuition of them). It turned out when categorising verbs like this, different words in the ーる category could be conjugated in different ways, so the last syllable doesn't completely define a category. But anyway, here is the result of that exercise:

Original/polite/polite native/negative (question mark means a gap in my familiarity)
行く・行きます・行きません・行かない
泳ぐ・泳ぎます・泳ぎません・泳がない
する・します・しません・しない (taken as a special case)
見る・見ます・見ません・見ない
なる・なります。なりません・ならない (different from previous, despite having the same last vowel!)
食べる・食べます・食べません・食べない (again different from previous, but same as 見る)
待つ・?・?・? (* see note below)
会う・会います・会いません・?
選ぶ・選びます・選びません・選ばない
読む・読みます・読みません・読まない
死ぬ・?・?・?
話す・話します・話しません・話さない

I couldn't think of any verbs in some categories off the top of my head, but I'll add them in the future, and of course I'll turn my attention to different conjugations over time.

Now as for patterns, of course する is a special case, but I can also see that 見る does not conjugate like most verbs. Basically, you chop off the whole last syllable and then attach the new ending. Then 食べる also works like that. Every other verb above more or less conjugates in the same way as 行く. To me, this suggests that there are not actually lots of different categories, I can only see three above: ones like 行く, ones like 見る, and then する is a special case (I'm sure there are also other special cases that just don't come to mind right now).

As for the question marks, I don't recall hearing 待つ (or any -つ verb) conjugated in these present tense forms so I'll have to pay attention to that in the future. If it's like 行く, which it probably is, then it's 待ちます, 待ちません, 待たない. If that's the case, I guess it hasn't sunk in because I don't think of つ and ち as related sounds with my English background, although they are related in hiragana. 死ぬ is also uncommon in my listening material! (I might need to watch some Dragon Ball Super) But it's probably conjugated the same again. And as for 合う, I also don't think I've heard it conjugated in the negative enough, so I'll keep an eye out. It's interesting because both 待つ and 合う are quite common, and I do use them myself when speaking, but I guess certain conjugations are just not as common so I don't have them in my mind. For example, I normally hear and use 待つ in this form: 待って...

So, these are the things that I'll be paying attention to over the next week.
3 x

tuckamore
Orange Belt
Posts: 136
Joined: Sat Nov 19, 2016 6:41 pm
Languages: English (N), Japanese (intermediate), French (intermediate), Thai (beginner), Swahili (beginner)
x 287

Re: Log: ryanheise tries to solve Japanese grammar

Postby tuckamore » Mon Mar 30, 2020 10:30 pm

Very impressed with what you've been able to deduce in verb conjugations!

ryanheise wrote:会う・会います・会いません・?
Does knowing that 会う conjugates similarly to 言う help you fill in that question mark?
ryanheise wrote:待つ・?・?・? (* see note below)
ryanheise wrote:then it's 待ちます, 待ちません, 待たない. If that's the case...
Bingo! You may have also heard お待たせしました? Notice how つ morphs into た here, too. 持つ、 立つ、勝つ、建つ、and 打つ are other common つ verbs that, if you know them, may help strengthen your grasp on their conjugation patterns.
ryanheise wrote:死ぬ・?・?・?
This one will be a harder one because you won't have other verbs that end in ぬ to help you build your mental framework with. You may have to apply what you already know from other verbs to 死ぬ。And, no, it isn't irregular. It's just the only verb ending in ぬ in modern use, as far as I understand.
3 x
: 10000 / 10000 10000 Japanese pages
: 16 / 10000 the next 10000 Japanese pages

User avatar
ryanheise
Green Belt
Posts: 294
Joined: Tue Jun 04, 2019 3:13 pm
Location: Australia
Languages: English (N), Japanese (beginner)
x 911
Contact:

Re: Log: ryanheise tries to solve Japanese grammar

Postby ryanheise » Wed Apr 01, 2020 2:12 pm

tuckamore wrote:
ryanheise wrote:会う・会います・会いません・?
Does knowing that 会う conjugates similarly to 言う help you fill in that question mark?


Thanks very much for suggesting some more words both this group and also the -つ group. I actually have 言う and 持つ in my full spreadsheet since I tried to think of a few examples for each one, but still couldn't bring to mind the conjugation for あう/ない. It just wasn't frequent enough in my listening (which reminds me that I really should step up the amount of input I get). In the end I ended up searching for those words in databases of example sentences saw how they were conjugated, although I look forward to also noticing it in my listening in the coming weeks, because that's how it's going to sink in.


Week 2 update

I'm almost a week behind on my update, so I'll post one today, and try to post another one at the end of this week.

I have filled in the missing parts of last week's conjugation table, and extended it by one column to include conjugations like 話せる:

Original/polite/polite negative/negative,can (new parts underlined)
行く・行きます・行きません・行かない・行ける
泳ぐ・泳ぎます・泳ぎません・泳がない・泳げる
する・します・しません・しない・できる?
見る・見ます・見ません・見ない・
なる・なります。なりません・ならない・
食べる・食べます・食べません・食べない・
待つ・待ちます待ちません待たない
会う・会います・会いません・合わない
選ぶ・選びます・選びません・選ばない・選べる
読む・読みます・読みません・読まない・読める
死ぬ・死にます死にません死なない死ねる
話す・話します・話しません・話さない・話せる

Again, question marks indicate where I am not sure. It occurred to me that できる is probably the special case form for する since it is a special case after all, and this seems to serve the same function that the -える form serves in other verbs.

Looking at the verbs above where I failed to fill in the last column, my lapses with the つ and う verbs were to be expected since last week's post revealed the exact same weakness in coverage. And when I look at the remaining ones, I can see one very obvious thing which is that they are mostly those verbs that have a different from usual conjugation rule where you chop off the last syllable, namely 見る and 食べる. In the next update, I might restructure the table into 3 groups since it looks like the conjugation patterns happen according to those groups. This makes me think that なる probably conjugates in the regular way and I should probably fill in the last column as なれる.

As for the conjugations of 見る and 食べる, I have a sense of what it might sound like, but I'll have to wait until I hear it again.

I may not fill in all the question marks by the end of this week since there aren't many days left, but for the next update, I will at least try to add some more columns to my table.

Reflection: I would say this has definitely been a worthwhile exercise so far. Having a goal of what to pay attention to does help (targeted) things to sink into my brain more effectively. It allows me to focus attention on parts of the language that I get exposed to less frequently and helps to compensate for that lack of frequency, to an extent. Ideally, I would like to also ramp up my immersion. I'm currently at 1 hour of Japanese exposure a day, and I know I can increase that, by replacing a lot of my media consumption in English by Japanese.
4 x

User avatar
devilyoudont
Green Belt
Posts: 467
Joined: Tue Jun 26, 2018 1:34 am
Location: Philadelphia
Languages: EN (N), EO (C), JA (B), ES (a mess), KO (dabbling)
Language Log: https://forum.language-learners.org/vie ... 15&t=12245
x 1287

Re: Log: ryanheise tries to solve Japanese grammar

Postby devilyoudont » Thu Apr 02, 2020 12:13 am

Listing a few more words ending in う to give you a few more examples to look for as a model for those verbs. Tried to think of words that would be very common, but your mileage may vary.

行う (おこなう)
使う (つかう)
戦う (たたかう)
思う (おもう)
歌う (うたう)
買う (かう)
1 x
Joyo Kanji: 1320 / 2136 Japanese Intermediate 2: 4654 / 12000

User avatar
Gustav Aschenbach
Orange Belt
Posts: 131
Joined: Fri Jun 10, 2016 4:22 pm
Location: Germany
Languages: German (N), English (C1), French (C1), Dutch (beginner), Spanish (beginner), Portuguese (beginner)
x 206

Re: Log: ryanheise tries to solve Japanese grammar

Postby Gustav Aschenbach » Thu Apr 02, 2020 6:52 am

Very interesting log. I learned Japanese grammar years ago with a very different approach, but it helped me a lot to really understand the conjugations (before that I had learned it the traditional way and I had never really mastered it). In order to analyze the morphology, we didn't rely on kana and kanji, but on romaji, which sounds counterproductive first, because we want to study Japanese characters along the way, but makes it actually much clearer. The last two posts in my log are examples for that, maybe it is of interest for you?
2 x


Return to “Language logs”

Who is online

Users browsing this forum: Google [Bot], vegantraveller and 2 guests