Why Rules Don't Matter in Language Learning - Steve Kaufmann

:open_book: If we want to speak a language fluently, we first have to memorize a bunch of rules, right? Wrong!

:brain: It turns out that our brains will automatically internalize a language’s underlying patterns when we get enough exposure and practice. So instead of spending all your time memorizing rules, focus on reading and listening. This approach is more natural AND more fun.

To be clear, I am not saying that you shouldn’t use proper grammar in your target language. Syntax certainly DOES matter. What I am saying is that you will naturally pick up the correct patterns with time. As polyglot Barry Farber put it, "You do not have to KNOW grammar to OBEY grammar.”



The fact is that in the video it is clear that “kids” learn by interaction. Now, there is a difference between interaction and input. A big difference if you ask me.
When we learn by interaction there are many more elements in play that we “absorb” from the other person and the environment. Our brain can create way more associations without any effort at all (by absorbing 11M bits/s of information that consciously we are not aware at all). Something that cannot be replicated in front of a computer.

We can compensate the “input” element with software but we are not even close to the capacity of that kind of interaction in matter of time of acquisition, quality and fluency.

Certainly, with AI improving, we can definitely get closer and closer to give to the brain more stimuli. When virtual reality will be a real thing, AI can create different characters with different personalities in different environments and the input given will be much more closer to a real interaction.

However, it won’t be enough as real human interaction is not a machine experience but it will be closer to what the author of the book was writing about.

1 Like

Just because a kid does something doesn’t mean an educated adult should try and emulate them. A child learning their first language is very different from an adult, who already knows three other languages, can already read, can understand complex ideas (such as grammar concepts), has the ability to focus for more than 10 seconds, etc. If people want to copy what a three year old child does, why are they reading in the foreign language at all? A three year old can’t read. The whole idea that adults learners should emulate three year olds as a way to learn their second, fifth, etc. languages is just cherrypicking ideas to suit some theory of who-knows-where.


@nfera I agree, but to be fair, in the video he explains that both kids and adults learn by interaction faster. I mean, he was talking about how the brain forms connections and understand grammar patterns faster by interaction. It was more a “brain/neurons/synapses” thing on language learning.

1 Like

Yes and no.

Over the last couple years, I’ve worked diligently in French, with near exclusive use of comprehensible-input-based techniques. I went all in on immersion and avoided the “traditional methods” that get lambasted by Steve Kaufmann and many others.

I’ve made great progress and it’s been quite interesting all in all.


I have substantial grammar holes. Simply, French has a decent number of verb tenses, a good number of irregular conjugations, a spelling system that isn’t that phonetic.

When I go through level assessment-style tests, I have to guess a lot. Usually my “sounds right” ear for the language will get me on a multiple choice test down to two and often the one right answer, but… it’s not enough. I’m not quite there.

Yes, my brain has “automatically internalized the language’s underlying patterns,” but it’s not yet fully there.

Sure, while Farber points out that “you do not have to know grammar to obey grammar,” there are meaningful corollaries. I don’t know about others’ brains but I think mine likes to know grammar good enough to know how to AVOID much of its complexities.

At this point, what do I do?

I’m now taking a pause from the all-in on comprehensible-input-based learning. My theory is that I can now, at around a B2 level, fill in my B1 and even in some places A2 grammar holes more efficiently with rigorous traditional-style grammar lessons for a while than to wait for my brain to get it all sorted out with hundreds of more hours of listening and hundreds of thousands of more words of reading.

The details about not having to “first” memorize a bunch of rules and “instead of spending ‘all’ your time memorizing rules” are correct. However, to headline that “rules don’t matter” is more click-baity than responsible.

While I love the inspiration of the high-energy advocates of comprehensible input and agree that it is the most important and even game-changing method for self-motivated language learners, right now I’m over on Lingoda a good bit doing traditional grammar lessons.

Is there a way to increase the nuance without dampening the enthusiasm?


@gmeyer I don’t know, in my case my brain doesn’t really internalize patterns so well. BUT it does it with real people interaction (which is not only input btw). That it is why I said there is a big difference.

I agree, same with me. I think that we speed up things by learning those patterns. However, as you pointed out, it is now clear that it is pointless to learn them at all cost from the beginning. In the video, Steve doesn’t deny grammar overall.
I tend to agree that IF we use LingQ as a main method, grammar SHOULD be studied/learnt very well at some point of the process.

Probably something like:

  1. simple overview of language pronunciation and grammar rules/patterns
  2. reading a lot and build vocabulary
  3. reading + listening a lot and build vocabulary plus other unconscious acquisitions
  4. some grammar here and there
  5. writing output
  6. more intense grammar structure
  7. speaking output
  8. I advocate a country full immersion now.

The points above are just a simplification as a general method that I’m thinking about by learning a language from home with LingQ.

1 Like

I studied French at school for five years, hated it, and learnt very little. Then at age 22 I started evening classes in French and 5 years later I moved to Montreal for 2 years, sharing a flat, and working, with francophones. I developed good comprehension of standard French spoken a bit slower than normal, but I could not speak, my grammar was awful, and my vocabulary limited. I never could understand joual.

Fast forward 30 years, and after 15 months of a refresher course then huge amounts of comprehensible input (podcasts), I can understand most native French podcasts with ease, and I can communicate. My spoken French is not advanced, but grammatically (mostly) correct, I can use the commonly used past, future and present tenses and the subjunctive. And I have a decent vocabulary. In short, I have learnt far more from comprehensible input at home, then I did in Montreal.

In my case, immersion did not work. Children and adults are very different:

  1. Children learn through play, using a simplified language, with no real consequences from making errors.
  2. They are more relaxed, with no time constraints.
  3. They will often repeat things, as part of play.
  4. They pick up the grammar naturally, as well as a perfect accent. They seem to be more observant, or adept at seeing patterns.
  5. There are biochemical differences. At least two groups of chemicals are present in greater quantities in kids brains, and those chemicals facilitate learning. Children learn much faster than adults. I saw this when learning ice hockey. The kids advanced 5 times faster than me.

Adults engage in mostly transactional relationships, they are often stressed, they have time constraints, they use a much larger vocabulary, and they are usually impatient. I was unable to pick up the grammar naturally. In fact I have to study grammar first, then when I hear it, I learn it. For me grammar study is part of the comprehensible input process. One colleague got on my nerves by correcting my French. C’est pas grave was wrong, I was to say Ce n’est pas grave. C’est pas la peine was wrong, I was to say Ça ne vaut pas la peine, and yet all the French speakers used the ‘incorrect’ forms. I was also told off for my Quebec accent, I was to speak like the French from France. Unfortunately some people think they are great language teachers, as he did because he was fluent in English and French, when in reality they are completely incompetent and have no understanding of language acquisition. Eventually I spoke only English, as the corrections were driving me mad, confusing me, and destroying confidence.

In my opinion adult learners are totally different from children, their brains are biochemically different, their circumstances are different, and they have to use different methods. As adults we can master a foreign language to a high level, but we have to use our own approach.

As to the suggestion that we can use AI to create language partners, that has merit, in my view. They could create a supportive learning environment akin to that experienced by children. We live in fantastic times, as far as second language acquisition goes, and they improve each year.


@LeifGoodwin It is always interesting to learn new experiences that have a completely different arc than others. I had a completely different learning path when I was in France and I would probably have one even more productive if I were to go today.
I suppose some things come natural and with others we need a more guided method.
I agree as well that kids and adults learn in a different way, however, I was referring specifically on the content of the video and to supposedly the researches cited on those books.

Yep. I never said and I rarely say “c’est pas grave”. I tend to avoid the way how native speakers speak on their slang because I believe that a second language learner should speak clean. I say “Je ne sais pas” and not “scepas” or how you would write it!

That is why I consider writing and grammar important, because we can reinforce them by speaking correctly and remember them for a long time. In that specific case, I was and I’m systematic in remembering that I have to say “ce n’est pas…”, even if I know that I could say it in a different way if I were a native. It is my choice.

Unfortunately I stopped using my languages for 10 years but step by step I’m catching up. We lose the knowledge of a lot of grammar rules without using them.

Another thing: you might have had a bad experience but you might also have let yourself being too influenced by other people judgements. Which is totally understandable but it wasn’t functional to your learning experience.

When I was in France, other people told me that “French people” were bad with foreigners because they didn’t help with the language, they always didn’t understand what others were saying, they didn’t slow down, amongst other things.
I experienced that too but the difference was that I told my friends that this was their country and they didn’t have to change the way they were. I was the one learning their language so I had to make the effort to learn it well. I used their native attitude to my advantage. Today, I would be even better at that. Never had a problem talking with French people, even when they were not understanding me, because I went to France without almost knowing anything about French, and didn’t study it at school either. But I was much older than you when it happened.

So just one fact. For an average adult person, you might need about 1600 hours to reach C1 level in Language. Children get this level in 6-7 years from scratch and full-immersivity? Rules are very effective and required in any language school

I don’t think that Steve is making claims about what is required in any language school, but rather what he has needed (or not needed) to learn a variety of languages–which he has learned without using a language school, I"m pretty sure.


Steve Kaufmann did a video on slang, and why he avoids it and his arguments were convincing. Some very common contractions are fine in casual speech IMO as I don’t wish to sound stilted. It is about context. The French I hear in podcasts is usually more formal, so no dropping the ne from negations. Again, it’s context.

I have met a lot of people who learnt English as a foreign language, some at the school I went to, some at university and some later on. Those who started before age 12, give or take a year, had native level fluency. Those who started later on in their teens had a slight to obvious accent, and sometimes made grammar mistakes. Those who learnt as adults usually had a poor accent and grammar despite living in England for more than 10 years. This is consistent with the old concept that the so called Language Acquisition Device shuts down or becomes less effective after puberty.

The comprehensible input method clearly can allow adults to gain native or near native level fluency. Steve Kaufmann and other well known polyglots prove this.

What are your thoughts about Dr. Brown’s research regarding adults not losing an ability to learn a language naturally instead gain an ability to hamper it with early speaking and consciously thinking about gramamr rules?

His research negates above mentioned point.

From his experience and observations Brown concluded that, contrary to the [critical period hypothesis] for second language acquisition, where adults have lost the ability that children have to learn languages to a native-like level without apparent effort, adults actually obstruct this ability when learning a new language through using abilities they have gained to consciously practice and think about language.

You quoted selectively to attribute to me something that I do not believe. In the last sentence I wrote:

“The comprehensible input method clearly can allow adults to gain native or near native level fluency. Steve Kaufmann and other well known polyglots prove this.”

(Unfortunately I cannot figure out how to quote from posts.)

The second comment I made is that most people I met who learnt language in their late teens or adulthood have a noticeable accent and make grammatical errors. That is consistent with the critical period theory, and it is hardly surprising that many people supported that hypothesis, even though I and many others believe it to be false. I note that these people all came to England as a child or young adult, and adopted conventional learning methods.

Since you asked me for my views, it is my belief that children and adults learn differently. Children have an innate ability to learn more effectively than adults in a so called natural approach. Recent research has shown that children learn much more quickly due according to the authors to the presence of greater concentrations of a chemical in the brain. I saw this when learning to play ice hockey. The youngsters (~13 years old) progressed far more quickly than I did. Yes I agree with your statement that adults tend to be hampered in their learning if they use traditional techniques. Classroom learning and rote learning of grammar are poor methods, as I am sure we can all agree.

You seem to be suggesting that adults can learn as well as children if they adopt the same approach. I very much doubt that although it is of course impossible to prove as you can’t put an adult into the same learning environment. I will list some reasons I gave in an earlier post. Firstly as stated children do learn far more quickly as shown by research. Secondly, they are exposed to simplified language. Thirdly they are more relaxed, and engage in play, not transactional language. Fourthly they spend a lot of time learning/playing. Although I do not believe the critical period hypothesis, there is clearly a change in the way that an adult learns a second language in comparison to a child.

Steve Kaufmann and others have proven that adults can learn a second language to native or near native fluency. However, they don’t adopt the same approach as children. They all use massive amounts of comprehensible input. They all use spaced repetition or similar tools. Steve K. uses Lingq to repeatedly expose himself to vocabulary. I know Richard Simcott and many others use Anki to get vocabulary to stick in long term memory.


Amen! Though I have less experience I am coming to similar conclusions.

I became determined to learn French last Christmas. For about two months I listened to French songs, studied flash cards and surfed the net for language learning methods.

I was excited by the Comprehensible Input approach Steve Kaufmann recommended. I’ve been doing LingQ for the past six months. I’ve reached ~14000 known words. I finished the first Harry Potter in French yesterday. I’ll reach LingQ I2 within six weeks.

I’m happy with my progress, but after my first tutoring session the big holes in my knowledge of grammar and sentence construction became glaringly obvious, as well as my limited listening comprehension and pronunciation abilities.

So I’m revising my approach. Comp Input is still the main course of my learning, but I’m adding grammar study, writing exercises, listening/pronunciation practice and tutor sessions.


@zoran (or anyone):

But how much time?

I do believe I could eventually become fluent in French with enough input and a long enough time.

However, I’ve had 6 months of 4+ hours/day with LingQ and I have my doubts how practical that would be.

For instance, I am so glad I broke down and made flash cards for avoir (to have), etre (to be), and the endings of the regular indicative tenses. It made such a difference, as opposed to blundering along, hoping the patterns would sink in.

Just highlight the words in the text and a “quote” popup (little shaded box near the text) will pop up. Click that and it will drop it down into the input field where your cursor is. You can highlight other texts and click quote to add more quotes.

R.e. editing. Thank you. Unfortunately that does not work on an iPad or an iPhone. :frowning:

I’m using iPad mini right now and it worked although I use chrome browser on that. Maybe in Safari it doesn’t works?

Okay, I am now in Chrome, and I cannot highlight the text. But I can click on the chain symbol and that copies the test, which can then be pasted into the message. Sorted, thanks. :+1:t2:

1 Like