RSS
The “Flat” Approach To Languages With Tons of Inflection
Another day, another kid named J.R. (different from the last!), all up in my email:
Hey Khatz,
Your method when applied to languages like Chinese and Japanese makes perfect sense but I am trying to learn Korean and Finnish.
My problem is with Finnish. A Finnish word can have up to 14 cases so do I need to make a sentence for each case? If done that way, it seems like I could make it to 10,000 sentences quite easily, but the 10,000 wouldn’t be the same as say 10,000 in Japanese/Korean/Chinese.
Appreciate the feedback.
OK, first of all, I have a secret (the secret, the secret) to tell you. Come closer. Closer. ‘K, here we go:
There are no cases in Finnish.*
Just make whatever sentences you need to make as things come out of your immersion environment. Just treat everything as if it were a different word. Focus on the difference in *meaning*, since that’s what actually counts.
Think about it in English — fundamentally, the difference between “I go to school” and “I went to school” isn’t one of tenses of the verb “to go” or whatever…the two words, “go” and “went”…the two sentences actually have different meanings. In theory, they are mutations of the same word. In practice, they are different words. “He eats the food”, “he ate the food” — these things are different.
Looked at this way, grammatical inflection ceases to be a burden, and instead becomes a tool for expressing oneself more precisely. You go from “Effing sonofa I have to learn all this effing mothereffing B.S.” to “SWEET! I can tell people what I will have done if I were to have been X; the future really is perfect!”.
So don’t think of the depth of variation of a single word. Pretend everything is flat. Treat everything as its own, independent word. Sometimes, it’s just easier this way. In practice, this does mean that every case will eventually be represented in your SRS, but not that you’ll necessarily have to decline or conjugate every single word that inflects — you are after all a human being; you know a pattern when you see it; you don’t need everything declared; you’re a gap-filling, pattern-matching machine. Do as much as you need to “get it”, and no more.
As for number of sentences, I doubt more than 10k will be necessary to reach a high level of proficiency. Remember that the sentences are just a tool/by-product for and of massive exposure to native materials.
For more, check out AntiMoon.com — Tomasz and the crew wrote about learning English, which is closer to what you’re trying to do in terms of certain language features.
Disclaimer: I haven’t actually tried to learn Finnish, but if I were to, this is exactly how I would do it. Starting with a phrasebook, I would just accept the sentences “as is”, and let the patterns present themselves to me over time. In any case the key is always to realize this: learning a language does not require pain, boredom or suffering.
*OK, maybe there are, but only because and as long as people keep saying so. They’re a theoretical construct that’s generally useful for analysis, and generally worth crap-all for praxis.
Read on:
Maya said,
September 29, 2009 @ 12:47
Spot on, Khatz.
I speak Russian at home with my parents, and according to my parents/other Russian adults, I speak correctly/grammatically. But I don’t even know how many cases Russian has, let alone what the cases are and what each one is for. Nor do I know which verbs are irregular.
You don’t need to know any of this stuff. It just doesn’t matter. You don’t need to know that “us” is some fancy inflected form of “we” to be able to know when to say which; likewise, you don’t need to learn cases or verbs seperately in a language to know how to use them. You can pick up anything, given sufficient exposure.
Arro said,
September 29, 2009 @ 15:22
I definitely agree, I passed two years worth of Latin exams with 100% by going with what sounded right and, of course, SRSing the vocab ~ !
BA said,
September 29, 2009 @ 18:29
Finnish has some words which change internally when conjugated for certain cases. Often, ‘k’ will change to ‘j’, but it can also disappear entirely. It is not possible to tell, simply by looking at the word, whether the ‘k’ gradates to ‘j’ or nothing at all. You have to learn each word with a “diabolical k” individually, since there is no pattern.
It’s like learning irregular verbs, but without the added help that irregular verbs are more common.
Turska said,
September 29, 2009 @ 21:30
Dear BA,
I’m a native speaker of Finnish, and I can assure you that for any newly made-up Finnish word, I can always instinctively tell how the word changes (or doesn’t change) when inflected. I think this means that there is a general pattern, from which there are a number of exceptions. There are only a finite number of exceptions. After a few years or so of extensive Finnish input, I’m sure you will have learned all the main exceptions and developed a feeling for “what sounds right”.
I’m really pleased to hear that someone is studying Finnish.
Finns generally show great respect for foreigners who have learned Finnish to a functional level.
Personally, I think that students of Finnish will benefit from occasionally leafing through a competent grammar book. I’d recommend “Finnish: An Essential Grammar” by Fred Karlsson. I have found its Finnish edition useful for learning to see (not just feel) the patterns of my native tongue.
BM said,
September 30, 2009 @ 00:01
Turska:
Interesting you should mention Karlsson, as he himself has said that ‘k’ gradiation has no pattern (you can find it in the book you mention. I don’t have my copy with me, but it’s in there!). It’s not a simple case of a few exceptions either, but many exceptions. Many, many exceptions. I’m not saying it’s impossible to learn which are gradated to ‘j’ and which to nothing (or indeed to analogise), but rather that it isn’t patterned, and is much more difficult to learn by symbiosis than the even most difficult morphology of other languages.
Seth said,
September 30, 2009 @ 01:44
Hi Khatz,
I’ve been doing this a lot of lately, but I find it’s sometimes hard to find certain example sentences for Japanese verbs. Like, I want to learn how to recognize and interpret the causative and passive tense. My dictionary doesn’t have a feature where I can just search for verbs in certain tenses, so it’s hard to find examples. I’ve just been using the ones off of Tae Kim’s grammar guide for now. Any ideas?
Also, when is the “My First Sentence Pack” coming out? I didn’t buy QRG, but I think I’ll buy My First Sentence Pack.
Eldon said,
September 30, 2009 @ 03:13
“I’ve been doing this a lot of lately, but I find it’s sometimes hard to find certain example sentences for Japanese verbs. Like, I want to learn how to recognize and interpret the causative and passive tense. My dictionary doesn’t have a feature where I can just search for verbs in certain tenses, so it’s hard to find examples. I’ve just been using the ones off of Tae Kim’s grammar guide for now. Any ideas?”
You could always just try googling the verbs you want to find in the forms you want to learn (you can find out how to conjugate them easily), and then use sentences you find there. Although really, I think in the true spirit of AJATT, you could just keep a beady eye out for them when such verbs crop up as you read massive reams of text…
Off-topic: who noticed the donate button’s text has changed?
captal said,
September 30, 2009 @ 09:13
Eldon-
This dictionary allows you to search by whatever tense, etc you want- if there is an example in the dictionary it will come up with the sentence it is in. http://www.alc.co.jp/
And… I noticed
Jonathan said,
September 30, 2009 @ 16:41
Eldon:
I noticed it too. I guess I got accustomed to the way it used to look, and always skipped over it (sorry Khatz! I bought the QRG though!), but now that it’s different it jumps out at me. Weird.
Huh… getting used to things through repeated exposure, such that the underlying patterns become second nature. I feel like I’ve read about this somewhere before.
(On another topic: My First Sentence Pack is due to be released today, apparently! Who else is totally stoked for this?)
きのこ said,
September 30, 2009 @ 20:54
I noticed he took the “passion” out of it :p
Greg said,
September 30, 2009 @ 21:20
In reference to what Seth mentioned above, I always get a little nervous that I’ll never be able to find the meaning of a verb in a certain conjugation or form WITHOUT the use of grammar structures. Perhaps it’s like an over attachment to an old girlfriend, but I feel it’s really hard to let go of it because it provides form and function so readily. Granted, I’m studying Chinese right now and there’s just ‘sentence patterns’, however, I want to pursue other languages and this has been a stumbling block for me on just how to go about it, without any formal grammatical training (e.g. class).
Jasmine said,
October 1, 2009 @ 01:09
Aweseome. Aweseome, awesome, awesome. I’ve been following you for nine months now with my study of Finnish, and this is just what I needed at this moment- a reminder that, despite all the emphasis on grammar in the boks, and the supposed “regularity” of the rules, you DON”T actually need an example of every single case in order to use them.
Very timely and extremely helpful!
For me, the most helpful thing I’ve done is increase my vocabulary, using picture flashcards; the Finnish version of learning Kanji
Ramses said,
October 1, 2009 @ 23:39
Sorry for plugging my own site, but back in February I write an article about exactly the same thing. Maybe some people think it’s and interesting read.
http://www.spanish-only.com/2009/02/how-to-ignore-grammar/
NDN said,
October 2, 2009 @ 02:32
Makes a lot of sense. Indeed, in many sentences I would just write the meaning of a verb in it’s infinitive(?) although that verb in the sentence was in a modified(?) form. And then I would struggle every time I reviewed those sentences. Some weeks ago I started applying this “flat” approach and just write the meaning of the verb in the infinitive and in the tense in which it appears in the actual sentence.
Example:
Question:
研ぎ澄まされた五感から生み出せる
Answer:
研ぎ澄まされる 【とぎすまされる】 (v5s) to be sharpened
研ぎ澄ます 【とぎすます】 (v5s) to sharpen
五感 【ごかん】 (n) the five senses
生み出せる 【うみだせる】 (v5s)can produce
生み出す 【うみだす】 (v5s)to produce
And then I just let my brain work.
The Chosen One said,
October 2, 2009 @ 13:04
My guess is that the vast majority of words will only be used in one or two particular cases each (for example: ‘air strike’ is usually used as a noun, but the other day I heard someone say ‘they got air stricken’ and I laughed for about 10 minutes straight because of how funny that sounded…. although grammatically correct, I think, most of all words will only be heard in one particular pattern). Only the most common words will be used in all cases, so in the end, it probably turns out that Finish has just about the same amount of “stuff” to learn as all other languages.
Eldon said,
October 2, 2009 @ 21:29
Captal:
That site rocks, cheers
Ryan said,
October 14, 2009 @ 08:15
You’ll quickly find that the cases in Finnish are by no means terrifying…in fact the only difference between the cases of Finnish and the particles of Japanese are that when you attach the “particle”/case ending stuff sometimes happen to the original word. They’re just as easy to use though, and generally pretty easy to recognize. In fact…the only problem they might cause for people is being able to look a word up in a dictionary, because when you find a word looking like “alusta” you don’t know that it’s probably actually the word “alku” in disguise (it could be other words too though…). A really useful website to get around this problem is http://www2.lingsoft.fi/cgi-bin/fintwol , you can put any word as you find it in Finnish and it tells you what word to look up in the dictionary! It also tells you other things like what case it is in…or tense…or singular or plural, but you don’t really need most of that exactly like Khatz describes!
Conjugation Doesn’t Exist! | Daily Chinese Mnemonic - Beta said,
February 16, 2010 @ 02:00
[...] The “Flat” Approach To Languages With Tons of Inflection “OK, first of all, I have a secret (the secret, the secret) to tell you. Come closer. Closer. ‘K, here we go:” [...]
All Japanese All The Time Dot Com: How to learn Japanese. On your own, having fun and to fluency. » Success Story: I’ve finally figured out this AJATT thing said,
March 21, 2010 @ 08:22
[...] I understood the input hypothesis, but I didn’t understand the implications, specifically on our biology. I figured that was just “We learn better by seeing examples than by trying to use grammar to produce sentences” — but that’s only half way there. I’m realizing that we learn better by seeing examples and not attempting to understand them. [...]
Lindsey said,
June 27, 2010 @ 14:51
Hey, could you alter your site code to stop forcing separate windows or tabs to open when a link is clicked? It ends up taking more RAM than necessary.
Feanaro Surion said,
August 25, 2010 @ 21:30
I’d just like to point out that the difficulty to this approach is that there is only one word in the dictionary, and the 14 cases in both singular and plural all just have to be recognized. That does mean though that every case needs representation in SRS, as you say, but for slightly different reasons in my opinion. The patterns need to be recognized as part of a single word. They exist as part of that word. There’s no getting around it. But unless the recognition is almost instant, the meaning won’t be known properly anyway. The only implication that has upon sentences is that for the ones that you do end up picking, you need to make damn sure that you’re going to understand what’s going on with the word, or with that inflection, because if you don’t, you’re sunk, grammar or no grammar. With enough exposure, the cases sort themselves out. In fact, they really have no meaning without a sentence, but in the contexts of a sentence, it’s easy to see what a word is doing. Even if I have no clue what the word itself means, I have a clue as to what it’s doing. In that way, cases could probably be treated as particles easily enough, as long as the changes to the original word are observed.