The Sentence Method: Heavy on Syntax, Light on Vocab?

[Axon]

Hi all,

I'm currently putting my Indonesian to the test after an eightish month experiment to learn almost entirely through parallel sentences, with only the slightest overview of grammar. Indonesian has never been my main focus during this time but I'd guess I exposed myself to roughly four to five thousand parallel sentences through textbook dialogue, Glossika, sentence mining, and copying examples out of reference materials. I read aloud virtually everything I copied down, often many times. Now that I'm in actual Indonesia and speaking with actual Indonesians I find my active vocabulary pretty limited. Growing by the day, certainly, but often missing very simple words that I know I can recognize and I know appeared in those sentences I reviewed.

This same sort of thing happened with my Mandarin, though I think to a bit of a lesser extent since I wasn't as strict with sticking just to sentences (plus I just put in more hours to Mandarin overall in the same time period). And it fixed itself after a while.

But the bright side of this method is that I have rock-solid syntax. I look at compositions from other learners and I immediately catch word order mistakes (though, very importantly, I don't catch mistakes with other aspects of grammar). Same with Mandarin when I started classes after a year of self-study. So it seems that this method could be very useful when learning languages with syntax that often confuses non-natives (Korean?). Or maybe it works well with very isolating languages like Thai, Mandarin, or Vietnamese.

So, has anybody else followed the parallel sentence method and actually gotten a nice big active vocabulary from it? And do you think the vocabulary disadvantage might be outweighed by a better sense for word order?

[tarvos]

Focusing on one thing always leads to a skewed skill set in my experience. This type of technique works if your syntax is the weak spot in your learning process, but it will do jack shit about not being able to produce tones when that is the real killer.

[neofight78]

To answer the topic of this thread: yes.

I use a combination of sentence and individual word flashcards in my study (amongst other things). It seems to work well for me.

[smallwhite]

... experiment to learn almost entirely through parallel sentences...
... parallel sentences through textbook dialogue, Glossika, sentence mining, and copying examples out of reference materials.

If your exposure of Indonesian is limited to these sentences, then it's only normal your vocabulary in Indonesian is going to be limited to the vocabulary in these sentences. If you had made it a point to cover a broader vocabulary in your sentences, I believe you would have gained a broader vocabulary from them.

I read aloud virtually everything I copied down, often many times.
... often missing very simple words that I know I can recognize and I know appeared in those sentences I reviewed.
... I look at compositions from other learners and... I don't catch mistakes with other [non-word order] aspects of grammar

You had only done input (reading) and no output (speaking, writing, L1->L2 vocab/sentence recall tests) so you didn't fare very well in speaking/recall. That sounds like a normal consequence, too.

Confirms my beliefs, so thanks for reporting the results of your experiment.

And do you think the vocabulary disadvantage might be outweighed by a better sense for word order?

I don't think you have to make a choice. Just pick sentences with more varied vocabulary.

[Keys]

(...)
I don't think you have to make a choice. Just pick sentences with more varied vocabulary.

Seems to me the amount of input vocabulary was not the issue, but the recall of it.

The issue with parallel text is free translations, so simple vocabulary gets translated with different vocabulary in your own language, so you're not always learning vocabulary or not learning it correctly, with that method.

Adding flash cards like neofight78 mentions would be a nice combination, but take a lot longer as you have to look up the real meaning of the words, not the parallel one, and then repeat them.

Would be nice if you had a word for word translation and the parallel text so you could just keep reading the sentences instead of doing boring flash card practice. Then you only would need to study the more rare words by re-reading their (marked) sentences.

Kind of comes down to glossed flash card sentences

[Axon]

"But wait!" I thought, reading these replies. "I read hella sentences! There must be some other reason my active vocabulary is so small!"

So I just looked at some numbers. Glossika Indonesian 1 has roughly 750 unique words spread over something like 1250 sentences (it's 1000 entries but a lot of them have more than one full sentence). A selection of 388 sentences from my own mining activity shows 887 unique words, and most of those are certain to be repeated in the Glossika list. Thinking about the rest of the sentences I've been studying from Clozemaster, textbooks, etc - I'm looking at a total unique word count of almost definitely under 2000.

That's nothing! No wonder! Even if I perfectly retained every word in every sentence I read, I'd still be hopeless at reading native text and following native conversation.

So it's time for a change, and now I know I can't rely only on the sentence method in the future. I like the context that sentence mining gives me and I know it helps my reading speed to read target language sentences all the time. But the simple truth is that I need to know more words.

[ASEAN]

So it's time for a change, and now I know I can't rely only on the sentence method in the future. I like the context that sentence mining gives me and I know it helps my reading speed to read target language sentences all the time. But the simple truth is that I need to know more words.

You can find audio with transcripts over at GLOSS. https://gloss.dliflc.edu/

[Koneho]

Axon do you happen to have the legacy course for Indonesian?