There are three stages of comprehension in any new language. Early on, you barely understand anything — and that’s fine. The goal at that stage isn’t comprehension; it’s letting your brain pick up the patterns. Then comes the long middle stretch where you’re learning words but listening still feels patchy, until one day you notice you understand more than you used to. Where are you right now?
I’d say I’m solidly in that first stage (0–30%) with Russian right now. I’ve only fully mastered 8 of the mini-stories so far, and most of the time I still feel like I “don’t know anything.”
Every once in a while I’ll read some beginner content and catch little broken pieces , “week on home… director of a company… big technological company…” , and for a second it feels like I’m at 60%. Then I realize it’s still just fragments and I’m very much still at the “I don’t understand” stage.
I’ve been thinking about your question: Can we actually learn when we don’t understand? I don’t think so, at least not in the beginning. But we can get more comfortable with the sounds, the rhythm, and the feel of the language. We can get more familiar, more curious, and more motivated. That part definitely happens even when comprehension is basically zero. I’ve heard that once you have a decent base of known words, then listening to stuff you only partially understand becomes really useful.
I’m reading Atomic Habits right now (one of Steve’s recommendations) and plan to move on to some of Krashen’s books once I finish it. Hoping it helps me become a more consistent reader.
Funny side note, while watching the video something randomly clicked in German (a language I haven’t touched seriously in 6 months). There was a sentence I always stumbled over, and suddenly it made total sense. Even without putting in the hours, German keeps slowly improving every time I come back to it. So maybe I’m creeping toward that 30% mark in German.
With Russian I’m just sticking to the mini-stories for lots of repetition like Steve suggests. I sometimes feel the urge to branch out and read more random stuff on LingQ because it boosts my “known words” count faster, but I’m trying to stay disciplined and focus on repetition first.
Anyway, really good video.
I’m surprised about the 60% associated with C1 level. Positively. I’m closer to C1 than I previously thought. 
For what it’s worth I agree with you.
I advise against reading Krashen’s books, or at least make sure you’ve read more balanced books first, They promote his methodology and won’t point out its weaknesses e.g. it’s complete bunkum. There are plenty of university level books that describe the countless theories of SLA (Second Language Acquisition) that are well worth reading. Krashen’s theory is just one of very many theories. There are also discussions in this forum where various people contribute their thoughts and ideas, and some of these are very experienced language learners. I would listen to them and not a university academic who as far as I know does not speak any L2 fluently.
Understandig 60% means you’re at the C1 level? Seriously?
In my opinion, that’s low B1 at best.
It’s common to hear exaggerated claims about how Krashen’s theories are “debunked”, and much like Chomsky detractors, when one looks into them, they mostly agree that he is directionally correct but have some minor disagreements around the margins.
Could you tell me specifically what is ‘correct’? His hypotheses are almost completely wrong. I can gointo specifics if you wish.
This might be of interest:
The guest is a university linguist specialising in vocabulary acquisition. At about 30:30 he states that Krashen has been discredited and mentions one important problem with his ideas.
The Krashen theory is very popular among social media influencers, probably because it helps sell input based language learning products, but not among academic linguists.
Chomsky is too big a diversion, but he and Krashen have a common characteristic which is that both are extremely charismatic and their arguments are often “It’s true because I say so.” The followers of Chomsky are noted for being quite aggressive.
I welcome you to state his issues with Krashen, I do not feel compelled to watch some YouTube video. All previous academic research I have seen has stated that, while Krashen is correct, some explicit grammar instruction is also beneficial.
Their have been lengthy discussions about that matter. Just search for “Krashen” if you want to read the whole criticism.
Leif hasn’t asked you to watch the whole video, just a small section at the end. But to sum it up for you: Krashen claims that there is a difference between acquisition and learning (which are synonyms) and that explicit learning, beeing taught or actively learn by self-teaching has no value. This is discredited. Actively recalling the language, using it in a meaningful way, spaced repetition and getting feedback has proven as highly efficient in boosting language acquisition. Comprehensible input is good for building a foundation, especially vocabulary, but not if it is restricted to passive consumption only.
Btw.: You’ve claimed twice that Krashen is correct, but despite beeing asked didn’t provide any sources for that claim. So maybe you could link some of the academic research you mentioned?
https://remix.berklee.edu/cgi/viewcontent.cgi?article=1008&context=faculty-works
I gave you the timestamp. If you are too lazy to listen, I’m not going to do the work for you.
Could you post links to research that provides evidence in support of Krashen’s theory? I haven’t seen any. I have read quite a few university level textbooks on SLA.
The Acquisition Learning hypothesis states that there are two ways of learning, namely acquiring and learning, which seem to correspond to implicit learning and explicit learning. One takes place incidentally, the other by conscious study be that in a classroom or elsewhere e.g. someone explaining something to you. He states that language can be acquired but it cannot be learnt.
The first problem is that he provides no way to determine whether something has been learnt or acquired. Thus the hypothesis is untestable, which makes it unscientific. It is no more than an act of faith. The second problem is that research has shown that you can learn items of language explicitly and with time they become a natural part of your language. In fact when we are at school we explicitly learn huge amounts of language, such as the names of chemical compounds, which become part of our L1. When someone explains the meaning of a word to us, thatbis explicit learning. According to Krashen, that is not possible.
The Input Hypothesis states that learners progress when they receive messages that they can understand, and that the input should be a bit above their current level. The first problem is that he does not define this increment i.e. should it be 90% comprehensible, or 95% comprehensible? Secondly research has shown that output is an important and essential part of the learning process for our L1. Input alone is not sufficient. Thirdly, how do learners determine the meaning of unknown words and unknown grammar? Generally I look them up, and that’s what I did at school and home. In other words, I studied. Really this hypothesis is better replaced with the statement that graded input is useful, but not sufficient.
The Natural Order Hypothesis states that we learn our L1 in a particular order, which I assume to be the same for everyone. This was based on a tiny amount of research relating to some word endings. He generalised that research without any supporting evidence.
The Effective Filter Hypothesis is a statement of the obvious. If you’re anxious or angry, you’re not going to learn as well because you’re distracted. In fact he could have made this more profound, by pointing out that when our emotions are heightened, we tend to remember better. If we have a great teacher, who enthuses us, and grips our attention, we learn better. If something extraordinary happens, we tend to remember. I still remember what a department secretary said to me in French when she saw me soaking wet. Thatbwas 35 years ago. Research has been done into this effect.
The Reading Hypothesis states that the more we read, the greater our vocabulary. This again is a statement of the obvious, and cannot be attributed to Krashen.
The Monitor Hypothesis seems to say that you should not consciously pay attention tomyour speech. Well we do that all the while.
He plucked theory from thin air without supporting evidence. I adopted his methods, and after 18 months wondered why I was making almost no progress in German. It was only when I adopted other methods, primarly conscious study and output that I started making progress. His theory is worse than useless, it is harmful.
Thanks Obsttorte. I’m tempted to delete my post, but I’ll leave it for the benefit of insomniacs. and those who wish to waste their lives online.
Well, we learn better through repetition.
My point isn’t that Krashen is 100% right in a vacuum. Laymen often mistake minor academic refinements for a total debunking. These disagreements get blown out of proportion because the stakes are small. People start calling things nonsense when the reality is that the theory is just incomplete.
Most people who find value in Krashen today do not treat his 1980s work as an immutable law of physics. They know his definitions like i+1 are unspecific and lack empirical precision. They still recognize that the core of his work regarding the role of comprehensible input is the most powerful driver in language learning.
What are the minor academic refinements you are talking about? Or more precisely, what exactly is the part of Krashens theory that is correct or the core of his work? What of that core can really be addressed to him and is actually proven? Why do you think that those criticising Krashen are laymen, without even responding to the actual criticism brought up (whose legitimacy doesn’t depend on whether it has been brought up by a layman anyway but on the arguments and empirical proofs), while at the same time seeing a legitimation in layman considering his work useful? You keep your responds very vague, strengthening my assumption you are lacking arguments.
The question isn’t whether Krashen is considered useful (or “the most powerful driver in language learning”), but whether it is. If people starting to learn a language base their methology on unproven or even disproven theories or assumptions, they will be wasting a lot of time in an endeavor that will take them a very long time even when using efficient learning methods. And that is a bad thing, imho.
Edited as the example was wrong, my memory has mistaken me. Apologies.
The core of Krashen’s work that stays valid is the necessity of comprehensible input. He didn’t invent the concept of reading or listening, but he was the one who prioritized it as the main mechanism for moving from one level to the next. Modern research doesn’t disprove the need for input. Instead, it refines the theory by showing that input alone is often insufficient if you want to reach high-level accuracy.
The minor academic refinements I’m talking about involve moving from an “Input Only” model to an “Input Plus” model. For example, Richard Schmidt’s Noticing Hypothesis is a refinement. It argues that while input is the raw material, conscious attention to specific forms is what actually lets the brain process them. Similarly, Merrill Swain’s Output Hypothesis and Nina Spada’s work on Form-Focused Instruction are expansions. They show that producing the language and getting feedback helps learners close the gaps that input leaves behind. These aren’t total debunks of Krashen; they’re just making the theory more complete.
I use the term layman because academic circles have mostly moved past the “is Krashen right or wrong” binary. Researchers treat his work as a foundational step that just needed more variables. The “harmful nonsense” label is usually something you see in internet debates where people mistake these necessary academic additions for a total disproof of the original premise. It’s a common issue where people think that because a theory has been built upon or modified, the original core must have been a lie. In most sciences, we don’t say Newton was wrong just because Einstein provided a more complete picture of gravity. We just recognize that the earlier model had limits.
Regarding your point on efficiency, that’s exactly what I’m getting at. If a learner spends years only consuming input and ignores more efficient tools like SRS or grammar study, they actually illustrate my point. Krashen identified the engine (input), but they lacked the steering and the brakes (focused study and feedback). Using efficient methods alongside a high volume of input is how you avoid that waste of time. We can acknowledge that input is the driver without pretending its the only tool we need. I think it’s fair to say that the validity of a claim is different from how its used, but in this case, the two are pretty closely linked.
It isn’t minor academic nitpicking His theory is misleading because it makes blatantly false statements which harm learners. The small parts that are valid are nothing new e.g. input is useful and input should be graded. To tell people that input alone is sufficient is plain wrong. To tell them that conscious study is useless is wrong. It is almost complete junk.
If you watch his lectures, he gives a sales pitch where he presents the old fashioned, rigid, boring way of learning based on rote learning and compares it with his new, modern method. In truth the old method he describes is the grammar translation method, and that was used for teaching dead languages such as ancient greek. Very few teachers used that method. In other words, he created a straw man in order to promote his method. In truth there are countless SLA methodologies, his is but one.
Can you provide links to research that supports his theory?
My personal subjective view is that Krashen is an extremely arrogant man who has done more harm than good to the field of SLA. He came along and presented a theory, based on no solid evidence, and claimed it to be true. Other linguists spend their lives carrying out careful research to further our understanding of SLA. I spent 8 years working in academic research in physics. If you put forward a theory, you expect it to match observations and make testable predictions.
We do say that earlier earth centred theories of cosmology are wrong because they are wrong. We don’t say Newton’s theory of motion and gravity is wrong because to a very good approximation it is correct. Objects do continue in a state of constant motion unless acted on by an external force, for example. The problem came when for example studying the orbit of mercury, small discrepancies were seen. Einstein’s theory accounted for those discrepancies, and made predictions which have been verified. Newton’s theory was shown to be wrong on the very small scale, where quantum mechanics provides a more correct model. In life sciences, theories of evolution prior to Darwin and Wallace were just wrong. The idea that bleeding a patient with leeches was good has been disproven. Many people died due to the attentions of their physicians.
The Krashen theory isn’t a good approximation, it’s junk. People often misrepresent it as the idea that we need a lot of input, and that input should be graded. Those ideas were not developed or first proposed by Krashen. That is a misrepresentation of his theory.
It is clear you have a bone to pick with the man himself. Coming from a physics background, I understand why Krashen’s lack of cold variables is annoying. If you cannot measure i+1 in a lab, it is hard to call it a scientific law. Still, calling it “complete junk” is a massive overcorrection. Claiming his work caused “harm” to the field is utterly silly when you consider that he shifted the entire focus toward meaningful communication. Before him, the standard was often mindless drilling and rote memorization.
Even if Krashen used a straw man to sell his ideas, those ideas provided a necessary reset for the field. Most people are not arguing for Krashen’s 1980s “Input Only” vacuum anymore. We are talking about how that foundation was refined by researchers like Michael Long and Richard Schmidt.
Take Long’s Interaction Hypothesis for example. It shows that input is the raw material, but the actual learning happens during the negotiation of meaning in conversation. That is a testable advancement that keeps the engine of input while adding the steering of social interaction.
This is not a choice between “Krashen’s junk” and “Physics-level proof.” It is about an evolving field. We can admit Krashen’s original model had massive holes without pretending the entire concept of input-based acquisition is a fraud. Dismissing the role of input just because the guy who championed it acted like a salesman is throwing the baby out with the bathwater.
Anyway, I’d rather put my focus on progressing with my target languages. It would be a shame to spend more time with academic post-mortems than actually making progress.
No he didn’t and no it wasn’t. You’re repeating his propaganda. I started to learn French at school 52 years ago, we did not use mindless drilling and rote memorisation. What’s more so called natural methods were developed in the nineteenth century. His ideas were not new. If you do your research into SLA theories, you will see that there are almost as many theories as you’ve had hot dinners. I exaggerate, of course, but there are and were a lot. The grammar translation method is just one such method. To present that as the standard is misleading.
A theory is harmful when it makes blatently false claims that mislead people. We see people such as Stephen Kaufmann making videos which claim that Krashen’s theory is science. Similar videos are all over YouTube. It is better to just say that we don’t know, than to make false claims.
A few years ago I was chatting online with someone who was English, and spoke Dutch, German and French and was often mistaken for a native speaker of those languages. He said that there’s not really much to learning a foreign language. I think he was right. Academic research has its role, but teachers are not stupid. Krashen’s theory was a backward step.
The argumentation is valid, but the premise is wrong. Which is the major point, I think. Using large loads of input hasn’t been a new idea. The principles of his idea can be seen in the post I made above, quoting @LeifGoodwin . These are nothing he came up with, they are taken from a book on that matter. The statements made by Krashen go far beyond the benefits of comprehensible input. Thus anything beyond that exceeds “minor academic refinements”.
That is actually incorrect. As Leif pointed out, rote memorization and drilling has been the approach when dealing with dead languages, so mainly Latin and Ancient Greek, but other languages were taught completely differently. A lot of the language learning methods applied in the early 20th century have made use of speaking, output and an interactive use of the language, where the language was taught by actively using it. That using the language at some point incorporates reading a lot goes without saying. We do the same in our mother tongue.