[jacobolus]

> Listening comprehension, IMHO is the last skill to develop in learning a foreign language. [...] The first phase of language learning is mostly theory. Mostly vocabulary and grammar. The second phase is mostly reading, reinforcing the theory, forming a good understanding of how the language is used. Additionally writing things, chatting and the like. Third phase is immersion with speaking and listening.

To clarify: This is the way skills develop when people extensively study but barely learn a foreign language, the way typically happens in classrooms. It is a cruel method with poor results.

All of the actual learning of the language per se happens from listening to (or reading) comprehensible input, which should start from day 1 (yes this takes considerable effort for teachers to implement). Front-loading explicit study of grammar is a total waste of time. Memorizing atomized vocabulary words is also relatively ineffective.

https://en.wikipedia.org/wiki/Input_hypothesis

[EliRivers]

Front-loading explicit study of grammar is a total waste of time.

I am far more competent in written Japanese than spoken, both reading and writing it, essentially through understanding grammar. "Total waste of time" seems a harsh appraisal.

[jacobolus]

If you take two novices and dump one into arbitrarily extensive grammar lessons for 2 years, and the other into spending a couple hours per day listening (i.e. actively focusing attention on trying to understand what is being said) to the language for 2 years, at the end the first person won’t speak the language and the second person will.

The first person is going to learn something, but it’s not an efficient use of their time if the goal is language fluency.

If you want to learn to read, then by far the most effective use of time is to practice reading, with whatever minimal bootstrapping is necessary up front to start reading very basic material.

This is not just speculation. There has been a ton of scholarly research on this topic.

[tasogare]

>spending a couple hours per day listening [...] to the language for 2 years, at the end the first person won’t speak the language and the second person will.

It doesn’t work like that. For the student to make any progress he need to be able to understand most of the things in the speech (this is called comprehensible input, and the theory is both applicable to listening and reading). That’s why there are people living in foreign country for years yet cannot speak or understand the language at all.

Also it’s kinda stupid to say that language learning must be either grammar/voc learning or listening: it needs to be both. Classes are needed because they make portion of listening material comprehensible, which is impossible by listening only.

[jacobolus]

> (this is called comprehensible input, and the theory is both applicable to listening and reading)

Notice the link I put in my first comment.

Yes, the more comprehensible the early exposure, the faster the student will advance at the beginning. If content is too advanced learning will be much slower.

It would be great if more language teachers and curricula put more significant effort into developing hiqh-quality materials aimed at being largely comprehensible to complete beginners. Many extant materials and courses do a very poor job at this.

> there are people living in foreign country for years yet cannot speak or understand the language at all.

Learning a language takes consistent deliberate effort.

> kinda stupid to say that language learning must be either grammar/voc learning or listening: it needs to be both

Grammar lessons have low marginal value for language learners, and are by no means necessary. Many people learn a foreign language without any formal grammar instruction whatsoever.

[bayesian_horse]

Yes, formal grammar instruction is optional. But it does accelerate things.

I mean, it may depend on your goals. If you are focused on communicating, and don't actually intend on using the language efficiently and competently, then maybe formal grammar is a waste of time.

Otherwise you need to have this information. You have a choice to get it in a systematic way or to decipher it on your own.

[jacobolus]

By the way, here’s Krashen (the comprehensible input guy):

http://www.sdkrashen.com/content/articles/teaching_grammar_w...

> Research on the relationship between formal grammar instruction and performance on measures of writing ability is very consistent: There is no relationship between grammar study and writing (Krashen, 1984). Perhaps the most convincing research is that of Elley, Barham, Lamb and Wyllie (1976). After a three year study comparing the effects of traditional grammar, transformational grammar and no grammar on high school students in New Zealand, they concluded that " ... English grammar, whether traditional or transformational, has virtually no influence on the language growth of typical secondary students" (pp. 17-18).

> In addition, research is equally consistent in showing that writing ability and reading are related: Those who read more, write better (Krashen, 1993a). The reform school boys in Fader's Hooked on Books study who read self-selected paperback books for two years outperformed comparison boys on writing fluency, writing complexity, and reading, as well as on measures self-esteem and attitude towards school (Fader, 1976).

> It is well-established that one can become an excellent writer with very little formal instruction in grammar, and those who do often give reading the credit for their writing ability: "I wanted to write and I did not even know the English language. I bought English grammars and found them dull. I felt I was getting a better sense of the language from novels than from grammars" (Wright, 1966, p. 275).

> Finally, our ability to consciously learn the rules of grammar is very limited. Linguists have told us that they have not yet succeeded in describing the rules of language, and anyone who has studied linguistics will attest to the complexity of the rules linguists have described. Studies in second language acquisition show that even experienced students have an incomplete knowledge of the rules they are taught, do not remember the rules well, and have difficulty applying them (Krashen, 1993b, Alderson, Clapham, and Steel, 1997).

> [...]

> I am proposing, in other words, a two-step procedure: 1. Students first acquire (absorb subconsciously) a great deal of grammatical competence through reading. 2. Students are taught to use a grammar handbook to increase their grammatical accuracy further, using consciously learned rules. The grammar handbook can be introduced in junior high school or high school. If a great deal of reading has been done, and continues to be done, the grammar handbook will need to be used only occasionally.

[bayesian_horse]

Good luck learning romanic or slavic grammar from "comprehensible input".

And the thing Krashen seems to be talking about is actually phase 2 in my model. The students may have taken the long way towards basic knowledge in English, but in order to advance, they need to spend time reading text to assimilate the grammar and vocabulary more.

[comex]

The essay you linked is about teaching English grammar to native English speakers. I looked it up, and Krashen has written about second-language acquisition as well, but you're citing the wrong thing.

[jacobolus]

You’re right. It was elsewhere in this discussion tree that we were discussing grammar instruction in a student’s native language. Feel free to mentally move it, if that helps.

[jaco8]

Well , there are 2 different kinds of learners. There are those who need grammar and those who do not. I can remember vocabulary but never remembered the grammar of the 4 languages I learnt over the years.

[bayesian_horse]

To use a language efficiently and proficiently, you need to know all the grammar. Regardless if you have learned them systematically or if you have deciphered them by yourself. The latter might be somewhat easier if the foreign language is close to your own language, which means you already got that information. I've found Indoeuropean languages much easier to learn than others.

Sometimes even the definition of "grammar" is confusing. If you study Indoeuropean languages, they are so similar that "studying" grammar is mostly getting the conjugation and declination right. Sometimes you only have to switch some endings between languages. With other languages, the entire system of morphology is different, because for example they have more "persons" or whatever.

Additionally, English or Chinese may seem to have easier grammar, but then the complexity is hidden in the vocabulary or usage patterns.

[bayesian_horse]

I haven't seen any research that says passive listening without instruction has any effect at all.

You may have some success in a language related to your native language. Good luck with something truly foreign...

[EliRivers]

with whatever minimal bootstrapping is necessary up front to start reading very basic material

That's how I started reading. Basic grammar, and some vocab. Vocab I didn't know was simple to look up; grammar I didn't know rendered the entire sentence incomprehensible and looking up the individual words did very little to help. If I knew the grammar but not the vocab, thirty seconds to comprehension. If I knew the vocab but not the grammar, order of magnitude more time to understand, often requiring assistance from someone else. Knowing the grammar felt like doing it on easy mode.

Not knowing the grammar rendered it a waste of time. THAT was the "total waste of time".

It was basically how it worked. "Here is some more reading to do, in order to comprehend it, here is some grammar that you will need." Literally front-loaded. My experience of effective learning is basically the complete opposite of what you advocated; no front-loading of grammar rendered it extraordinarily inefficient.

I dread to think how long even the basic sentence structure of "topic - comment" in Japanese would have taken me to realise if I'd had to learn it by just listening to people use it; like I'm meant to be some kind of linguistic detective. Telling me that before hearing the sentence spoken rendered it SO much more understandable, right from the start. I was able to start making valid, meaningful sentences in Japanese within sixty seconds. Topic - comment. Here's some nouns, here are some words that are like adjectives, here's how you mark the "topic" which you could think of as similar to the "subject" in English grammar, but go easy on that because the grammar is different, here's the copula, off you go.

Great work, that was a nice sixty seconds of comprehensible and correct Japanese, none of which you had ever heard before - you constructed it all yourself. Pretty good for someone who's been studying the language less than ten minutes. Here's the copula in the past tense, off you go with that.

My God, I genuinely flinch to think about how long that would have taken me without any front-loading of grammar. If I was meant to just guess how the language worked based on listening to it. What a waste of time that would be.

I reckon someone could listen to Japanese for a very long time before realising that some of the things that are kind of like adjectives can change depending on tense, but for some of them the copula changes. A native English speaker would have to really be good to spot that quickly. What a waste of time that would be compared to simply learning that piece of grammar.

[ximeng]

I self-studied Chinese mostly using Skritter (skritter.com) for spaced repetition of vocabulary and found it effective. The problem with learning via listening is the availability of suitable material.

I estimate you need around 20k of common words, characters and phrases for understanding a large variety of vocabulary for everyday use (e.g. watching television, reading newspapers, daily conversation, business discussions, books, songs, poetry, etc.). This takes a considerable amount of time to commit to memory, so doing it efficiently is important - at 2000 hours a year, this is two years of memorising 5 items an hour, which is ambitious. While learning, it is necessary to take into account the fact that the same word might be used in different ways in different contexts.

I agree grammar is relatively irrelevant from a wider perspective - there's just not that much of it compared to vocabulary, and an intuitive understanding is fine for the most part. However it is also typically taught at a beginner level with basic vocabulary and varied sentence structures, which are useful in themselves (not "a total waste of time"). Understanding the terminology also makes using reference materials easier in some cases.

In practical terms then, carefully graded tuition and listening practice might be the ideal, but in practice this is an expensive route for most people when you are looking at two years of full-time learning. Realistically drilling vocabulary will get you a long way towards understanding native materials (e.g. movies with subtitles), and allows you to self-study by looking up definitions in the native language. It also allows this to happen relatively cheaply, as you can study yourself.

You're not going to get to fluency without a wide vocabulary, and from a vocab / time unit level, drilling vocab is not a bad way to get there.

[flomble]

> Front-loading explicit study of grammar is a total waste of time.

I completely disagree, partly because studying grammar is an effective shortcut to making input more comprehensible. If I memorise the verb conjugations for Spanish verbs, for example, I now understand more information about every sentence (e.g. omitted verb subjects and tense/mood information). It will still be necessary to listen and read extensively in order to internalise the formation and meaning of the verb forms, but it skips a lengthy process of reverse engineering. Ditto for producing your own sentences, which I think is also fundamentally important for learning a language.

> Memorizing atomized vocabulary words is also relatively ineffective.

This is absolutely true; when using flashcards or spaced repetition, it is vastly better to have a full sentence with the target word blanked out compared to only having the definition.

[usaar333]

One problem with comprehensible input methods (like what Rosetta Stone does?) is that it can be extremely boring as an adult and ultimately demotivating.

Studying grammar can be fun - and it feels like an accomplishment to be able to say some pretty complex things in Chinese.

Sure my listening isn't that great.. but on the other hand being limited to a two year old's conversation ability isn't particularly enticing.

Also, fwiw, reading and listening is a way different skill... Thanks to the ability to rewind when reading

[bayesian_horse]

In my opinion it takes years to be able to comprehend random sentences uttered by native speakers, in pretty much any language. It just takes a brain that amount of time to short-cut and organize all the information in an efficient enough manner.

[james_s_tayler]

>Front-loading explicit study of grammar is a total waste of time. Memorizing atomized vocabulary words is also relatively ineffective.

Self studied Japanese to fluency that way. AJATT method but largely concentrated on front loading all the grammar in 4 months then built 10k vocab over the following year. Worked like an absolute charm.

[bayesian_horse]

I've heard the idea that somehow "front-loading" doesn't work. For one thing, it does. I have done it, repeatedly. I have rarely heard people have success without devoting a lot of time on "front-loading".

You have to learn the grammar rules somehow. If you want to do it through assimilation, as supposed to somebody telling you how it works, well, that's possible but takes a lot longer and is more error prone.

I'm not saying you shouldn't do "output" exercises. I'm just saying that in order to learn information, you better receive it somehow. And you can receive that information either in an orderly fashion ("front-loading") or in a random incomplete fashion. You can't guess your way through language learning.

[SpicyLemonZest]

That doesn't really pass the smell test for me. Even native speakers of a language in its native country are explicitly taught grammar and vocabulary in school. I could believe the Wikipedia article's weaker claim, that there's an extra step between instinctively remembering "了 is the perfective particle" and being fluent in the perfective aspect. But the idea that explicit instruction is useless, that consciously knowing grammar won't help you acquire it at all, seems obviously wrong.

[jacobolus]

> Even native speakers of a language in its native country are explicitly taught grammar and vocabulary in school

It is not that there’s no value whatsoever in formal study of grammar. It might come in handy if you want to be a linguist, an editor, a high-level writer, a lawyer, or the like. If students want to take a grammar course in high school or college that seems okay with me.

It just doesn’t teach basic language fluency.

Native speakers don’t start studying grammar until they have had 10+ years of full-time experience with the language. And anecdotally, the students who spend a lot of time reading independently don’t really need the grammar lessons (they already have a subconscious understanding of what is or isn’t grammatical, and the typical school grammar lesson is very slow and obvious for them), and the students who don’t spend any time reading independently and regularly speaking with educated adults would get more value out of instead spending the time reading or listening to someone read. YMMV.

[dragonwriter]

> Native speakers don’t start studying grammar until they have had 10+ years of full-time experience with the language.

Untrue; formal grammar instruction begins not later than first grade in many curricula, which is age 6-7, which would require using the language several years before birth to reach 10+ years prior use. Native speakers begin studying grammar about as soon as they have the intellectual capacity to comprehend the concepts associated with grammar.

[jacobolus]

There is little if any formal instruction in grammar in reasonable primary schools. All else equal, students who attend primary schools that don’t teach grammar at all end up speaking and writing just as well as students who attend primary schools that try to teach grammar.

When primary schools try to teach grammar it is boring, stressful, and generally unhelpful to the students.

The dominant factor affecting students’ reading comprehension and writing ability is how much time they spend listening and reading, especially to material which is at an appropriate level to slightly stretch their abilities.

If schools want to spend a relatively small amount of time formally teaching grammar to 12–17 year old students there’s nothing inherently wrong with that, but it’s mostly only useful insofar as it attaches names to concepts so that students can have conversations with each-other about what makes communication effective of ineffective, or more explicitly discuss their existing subconscious grammatical knowledge. Formal instruction in grammar (or other kinds of formal analysis) is still no substitute for practice listening and speaking and reading and writing (ideally with effective feedback), which should be the main focus of language arts instruction.

[musicale]

> There is little if any formal instruction in grammar in reasonable primary schools

Maybe we need to bring back grammar school.

[jacobolus]

> Although the term scolae grammaticales was not widely used until the 14th century, the earliest such schools appeared from the sixth century, e.g. the King's School, Canterbury (founded 597) and the King's School, Rochester (604). The schools were attached to cathedrals and monasteries, teaching Latin – the language of the church – to future priests and monks. Other subjects required for religious work were occasionally added, including music and verse (for liturgy), astronomy and mathematics (for the church calendar) and law (for administration).

I am not knowledgeable enough about the topic to say how efficient or helpful medieval Latin schools were at teaching Latin as a second language, but the idea of “grammar” as part of the “trivium” (alongside logic and rhetoric) meant something substantially different than the modern usage of the word.

[bayesian_horse]

I can remember being taught about grammar in first grade. And getting it wrong at first.

[SpicyLemonZest]

But I don't follow the basic assumption that the way native speakers become fluent is particularly effective. No matter how much explicit instruction helps, native speakers couldn't learn basic fluency that way, because there's no way to deliver explicit instruction to a baby.

[yorwba]

Agreed, but learning grammar before you can even understand the examples used to illustrate a rule is putting the cart before the horse.

On the other hand, once you know a few specific instances, it's helpful to explicitly point out that they're governed by the same rule and give an abstract statement of how the rule works. That way you can check yourself without relying on a teacher to point out your mistakes.

[ximeng]

Providing a few specific instances and then talking about the general rules is typically how grammar is taught when acquiring second languages.

[yorwba]

Being shown a few examples is not the same as understanding them. If the first time you see an example of the rule is during the explanation of the rule, you have to deal with too much new information at once. A good explanation should refer mostly to information the learner is already familiar with.

The first time you see an example of the rule, there's no need to explain the rule yet. You can just memorize the example. The same is probably true of the second and third example. Explaining the rule only makes sense once remembering the explanation becomes easier than remembering the set of examples you need to know.

[bayesian_horse]

I've heard toddlers learn about 7 words a week.

When I'm into a new language I can learn 20-50 words a day, sustainably. Short term more.

[jacobolus]

If you define “toddler” as age 1–1.5 then sure.

By the time a child is 2.5 or 3 years old they can be learning dozens of words per day.

The most efficient way to teach new words to kids is to spend as much time as possible reading books with them.

[pas]

Grammar for learning a second language is helpful. Helps you to understand how to map what you want to say, and what you heard/read.

That said, brute force practice work reeeeally well for language. And naturally people will build a kind of internal grammar anyway. But knowing what is what can help with that too.

[topmonk]

> Native speakers don’t start studying grammar until they have had 10+ years of full-time experience with the language.

That's by necessity because you can't explain perfect tense to a newborn.

[nilkn]

I think the statement was just phrased too strongly. The proposition isn’t that one shouldn’t study grammar at all, but rather that one shouldn’t front-load grammar and theory too much.

Speaking from experience, I gained basic conversational fluency in Spanish from English in a few months by spending a ton of time with LingQ (essentially reading and listening to untranslated passages and creating digital flash cards), obsessively practicing pronunciation, and listening to lyrics in Spanish music for hours on end. I focused on grammar to the extent that I had to, but no more than that. For instance, I learned only those conjugations I actually needed for simple speech. I compiled the ones I needed to study based on what I actually encountered in real passages.

Now, by no means did that render me fully fluent in the language. But, case in point, I took two years of Spanish instruction in school with lots of grammar front-loaded and I gained more or less no speaking or listening ability at all. For me there was a dramatic difference in efficiency between the two approaches to language learning.