Is it really not helpful to start speaking from the beginning?

Many well known polyglots, including Stephen Krashen say, that it is not helpful to start speaking the language you learn right from the beginning. But so many other polyglots say, that it would be useful. What are your thoughts?


“Different strokes for different folks”. I’m not sure of Krashen’s precise words on that, so I can speak to it. However, I personally think it could be helpful. Early on it might just not be super useful. The main problem is that you’re likely not going to understand anything coming back to you your limited vocabulary is not going to produce much in the way of a long or interesting conversation.

So many may prefer to wait until they get some amount of vocabulary under their belt through reading and listening. There’s a better chance for a better conversation.

I’ve personally taken mostly a waiting approach. I have used the target language on occasion, mostly in brief interactions or speaking to myself. Mostly for a variety of factors…feeling a lack of anything interesting to say (even happens a lot in my own language!), self-consciousness, lack of understanding the speakers. etc.

If you want to speak, by all means speak! If you have someone you can interact with, or speak to yourself (I do this mostly), go for it imo. There are some that will say if you speak to early you’ll hammer in mistakes too. I disagree…at least if you’re doing more listening and reading. We don’t get grammar or sounds correct when we grow up and we eventually figure it out so I find this argument to be a bit specious.


Speaking gave me short term results/effects. Waiting and then speaking gave me long term results/effects


Really, you can do what you want. It depends on your goals, etc. If you are going to Greece next week, you might want to learn how to say a few, important sentences. But if you aren’t in a rush, then I choose to wait, simply because you don’t have much to say, when your vocabulary is super limited. If the opportunity pops up, then sure. But I’m not going out of my way to book a tutoring session, when I only know 10 words.

Here’s a similar discussion: When To Move To Speaking With Native Speakers And Tutors?...


Depends on a number of things, including to what extent speaking early motivates to continue with the language and your overall goals, for example…
… short term learning of a little bit - by all means, start speaking early.
… longer term / more serious + immediately starting to speak in the target language? I would at least be careful with that. Why?

Natural process of acquiring languages: Children listen first before the start speaking - so this seems to be a natural way. That doesn’t necessarily mean that grown-ups need to do the exact same thing, but we deviate from the natural learning process at our own peril.

Speaking early - after having only rather little exposure to the sound of a language - appears to create an ‘accent’. How shall we properly speak a language when we have not had enough exposure? Wouldn’t we then train to get an accent and perhaps even train, not to notice subtle differentiations in the target language?

What if we want the best of both worlds? Getting familiar how to pronounce the words reasonable well + start to speak relatively early?
A lot of repetition of rather limited content could help. The limitation of the content allows for more repetitions, so many in fact, that we naturally tend to start speaking some words - even though we are overall still at an early/beginner stage. This usually requires significantly more repetitions than necessary to simply understand every word of the content.*

[*We can outsource some of the work by ‘listening passive’, as had been suggested by the German Vera Birkenbihl decades ago:
For this method the sound needs to be loud enough that we could understand it (if we would be actively listening). This allows the brain to pick up the already known and thus comprehensible imput and strengthen the corresponding neuronal connections. The idea is not to actively listen (as usually associated with ‘passive listening’) but to keep it in the background while doing something else, perhaps even watch TV.]


Steve Kaufman has said that, in order to be fluent, you have to speak.

Based on my own understanding, it seems to me speaking will make you better at speaking, but it will not make you better at the language.

That is to say, you will not get better at German by speaking German. You will get better at speaking German. So if your German is at a level where you are capable of holding conversations, speaking is pretty useful. If your German is not at a level where you are capable of holding conversations, speaking is probably not very useful.


If it makes you uncomfortable, and you don’t like to do it, don’t do it. I personally think it’s a waste of time since you really need to build a solid model of the language in your head before running the program. Think of it like installing a brand new operating system on your brain. Nothing will work very well without the necessary software to run things. The system is mostly built through listening and reading with very occassional grammar look ups for reference.


I’m at a pretty advanced level of input in my Spanish but still a baby in output after nearly three years of immersion and LingQ.
I totally get some of the issues of outputting early but I wish I would’ve began outputting much earlier.
If I could go back, I would begin after maybe 3 months of intensive immersion and write a sentence a day. Eventually several and then a paragraph a day. Then get a tutor sooner and find a community such as discord maybe 1.5 years in. I waited two years before virtually any input and it’s very frustrating.

My accent is generally great and I have a lot of Spanish in me to convert into active words and phrases but it’s hard work.
I totally recommend early output but starting small and consistent. Still making immersion the priority.


Honestly, I believe it depends on the person and more importantly the language. I’ve heard the US military spends 2 days on the Russian alphabet before they speak basically at all but it’s like 3wks learning the alphabet for a language like Arabic.

In my own opinion and preference, I am more likely to and want to learn to speak more than read a language at the start. It’s very valuable, probably makes it easier too, to learn it but I don’t feel it’s a requirement to learn reading/writing first.

TL;DR IMHO: If you’re learning anything other than Latin, you should start speaking pretty early.

This is a rather painful subject for me, since I’ve been losing pretty hard in the speaking area. I’d say that over time my views have changed to the opposite: from “Listen, listen, and everything will be fine” to “If you neglect speaking, you’re setting yourself up for failure”.

I understand that this is a bit of an extreme view and I’m exaggerating a bit))
However, nowadays I believe that once you get used to a language, you better start slowly using it.

Arguments for this:

  • You remember things better, because you get experiences and associations
  • You see what you really need and practice it
  • You train your muscles and speech apparatus
  • Most importantly, you become aware of things that are not important for understanding, but are crucial for production.

I used to think that speaking too early would ruin my pronunciation irreparably. I don’t believe that anymore. Again, based on my own experience with both English and German, it is not that difficult to change and improve your pronunciation, and anything, for that matter. People only struggle because of a lack of motivation.

The only thing it takes to change the skill or relearn something - is time and effort.
Trying to learn it “right” from the start is futile - you have to pay your dues anyway - but the sooner you start, the greater the positive synergy effect.

I don’t buy that “learned for life” stuff anymore, our skills and memories are very malleable. What proved it to me was that after my street was renovated, I could hardly remember what it was like before. And people we know, it’s hard to remember what they looked like 5 or 10 years ago - the “new” image replaces the old one.

Our brain doesn’t cherish memories - it replaces them on a whim. Just by remembering, we change and replace them without even noticing.


A lot of immersion people go so far as to say it hurts you to output early but I think that’s rubbish. I just think it’s hard (obviously) to output before you spend a lot of time with a language but some people find it very helpful. If you start speaking then the language may feel more relevant to you so your other skills improve faster for example. But on the other hand if you took a class or a lesson with a tutor and they immediately were trying to get you to output and you hated it and then thought you were a bad language learner as a result that probably isn’t the case. You’re just not a super extrovert and would probably benefit from immersing first.

edit I still kept thinking about this so I’m going to add a bit. If your goal is to be good at outputting (I.E. to talk to people) and you’re basically inputting to be able to output - I think you should start outputting as soon as possible. The sooner you start outputting the sooner you’ll get good at it. Reading will help give you a large vocabulary which you’ll appreciate when speaking to natives but the more you output the better you output has been my overall experience.

But if you’re looking at the big picture over the course of several years and not what you can achieve in 3 or 6 months then I don’t think putting off outputting for a year or two even will really hurt you and you might be happier for it.


2 days isn’t very long in the life of a language learner. If they start outputting after that then I would say they start outputting very early.

I’m hoping my Korean isn’t like this. It might be that I have to suffer the pain of terrible output skills even if I get a really good passive vocabulary. But I think back on my German and I did a lot of input and this had a big impact on my output. I might have talked to myself in my head though…

I’ve been learning Spanish for a couple of months but really keen to start speaking it. There is a local language exchange but feel I don’t know enough to keep a conversation going beyong ‘hola’ and comó está. To make the bridge I download random pictures from the net and describe what I see in the tense I chose. I then compare that to a translator and learn from any mistakes I’ve made. Helps me think about what I’m saying.

1 Like

I think it is important to distinguish between pronouncing words and phrases in the target language to get your tongue and mouth used to making certain sounds and having a “conversation” with someone. If one’s target language has similar sounds and grammar to your native language or to one that you already know well, then pronouncing the words you hear out loud will not be that difficult and I would encourage it early. This does not mean you will be able to have a conversation since you probably don’t have enough vocabulary or grammar to do so in any meaningful way. By contrast, if the sounds and grammar are very different from what you already know, then you will probably need a lot more input before you even try to replicate what you hear. I would still try all the same, comparing single words by a native speaker to your own speech. It takes considerable exposure to another language to hear subtle differences between how you think the native speaker is pronouncing something and your own version. (For example, a “d” in Spanish and English looks the same but is not pronounced exactly the same.) No one improves by being silent and thus practicing out loud on your own, is very, very, very valuable in my opinion. When you actually choose to have a conversation with someone is something else entirely. That requires greater vocabulary and grammar, again depending on how different the new language is compared with another you already know and upon how well you handle stress as the initial “conversations” are never easy :slight_smile: !

It depends on many factors, including your personality.
If, for example, you already have foreign friends who would be happy to talk a lot with you and tolerate your mistakes, why not?
For other people who are more shy, it’s probably better to build some confidence concentrating on the input first. For me it’s sometimes difficult to “break the ice” even when speaking in my native language, let alone when I hardly can say several words.
I’d stop worrying about what is “effective” and just do what keeps you engaged in the language learning.


I think that it depends on the language. Since Spanish is so phonetic and consistent (it even has accent marks to help with pronunciation), it makes reading it pretty easy. Yes, one can even practice the different sounds by reading. I have been doing both comprehensible input (listening to anything I can find) and speaking with myself, groups and native speakers. Within 18 months I have been told by my tutor that I am a B2 speaker/listener. However, when I began learning French and Ukrainian, I had to do A LOT of listening first before attempting any speaking. Yes, I think that it truly does depend on the language you are learning. Do what works best for you. Some people are auditory learners and some are visual learners. However, most of us need to learn a language in context with socialization because body language and tone of voice are important aspects of language and communication.

1 Like

Hello ttatoni, it depends on your chat partner. I found a very friendly interlocutor, he is great. Try to find a native speaker without complicated slang.
I’m learning English and helping him with German. I like a good mix of reading, listing and speaking.
It motivates me a lot to speak.

1 Like

your English doesn’t sound like a beginner hehe.

Yes, I’ve thought about this one a lot because I’m trying to speak Latin and between reading and speaking there is a BIG difference. Also, there’s a limited amount of sources of CI for latin.

All I’ve heard Krashen say (in my own words and this isn’t to say he hasn’t said more) about speaking is that it can be stressful, and stress negatively affects how well your brain acquires language. Therefore if it stresses you out when you’re in class for example, you may not be soaking up the teacher’s output as well as you could. Also, I think I’ve seen glancingly mentioned somewhere that the studies say input not output is where you get your highest bang for buck. But generally I think the message is, don’t screw up your CI in order to struggle painfully to speak a few words.

I’ve certainly heard anecdotally from several people (including several answers to this post) that all input and no output does not not a speaker make.

Personally I believe trying to speak ups the rate of my comprehensible input. Remember, the definition of CI is “Brain receives and understands message”. My experience is that I often understand more when I read or listen if I’ve actually tried to use the input. With pure input, a lot of words just never register as individual words. Yes I get the general meaning of a sentence, but individual words and nuance can remain anonymous. And I can’t USE them. For that matter, I’m pretty convinced trying to understand certain points of grammar (organically as I take in CI) does that as well. It organizes it in my mind, so the next time I understand it better when I read or hear it, and boom more CI and more improvement in my mental model.

Also, easier to target what you want to learn by trying to speak. Rather than reading ten novels which might have the phrase ‘see you later’ a total of 2 times.

The human brain is a very complex thing, and our neurology is also quite diverse. I’m AD(H)D for example. That’s part of why I think grammar actually helps me. Otherwise it starts to turn into a big blur. Also I get bored when I feel like I’m missing nuanced bits of sentences. (again more of an issue with latin where CI is scarce)

Point being, I think there is only so much that these language acquisition studies can prove. I think they’ve proven that a grammar-first approach just doesn’t work, and that vast amounts of CI is necessary. But that’s it. And honestly do children stay mute and then suddenly start speaking in full sentences? No. Do adults have to listen to language for as long as kids do in order to be able to spit out a few words? No.

So as so many people have mentioned already, I’d say keep a couple things in mind: CI is GOLD. There’s more than one way to create the conditions for CI (i.e. brain understands message) Stress reduces the level of acquisition. You can understand a word when you read it and still not be able to access it or use it in conversation.

1 Like