How to Start Learning Chinese

Today is the 1st of August, 2021 and I have decided to start learning Chinese - something that I've been putting off for a few years now. I don't speak any Asian languages, so I'm currently an absolute tabula rasa.

I'm going to use the best method of learning a second language that I know of. The following is the description of its adaptation to Chinese.

I'll be posting updates on my progress as frequently as I can.


Part 1: How to Measure Language Proficiency

Whatever ultimate goals you may be pursuing when learning a language, they all imply the desire to achieve some level of proficiency in it. But what does it mean to be proficient in a language?

I believe we can measure our progress with just one simple metric, the number of words/phrases that you know. Suppose you just got started and the only thing you've learnt so far is the hackneyed 你好 (nǐ hǎo). Your proficiency at this stage equals 1. The average native Chinese speaker knows tens of thousands of words and idioms, so the measure of their proficiency equals whatever that number is, e. g. 30,000.

By adopting this metric, we can easily see how fast our progress is. If I'm learning 70 new words and phrases every week then I know that in order to get to level 3,000 I'll have to maintain that rate for 43 weeks (~10 months).

Of course, there is a tiny snag here: how and what exactly to count? Should we treat the common phrases a speaker knows as each being a single entity, or should we break them up into words and count those instead? Obviously, the results can vary drastically depending on how we count. Doesn't that render the whole idea of measuring language proficiency this way useless? And what on earth do we mean by know?

As valid as these objections may be, they are irrelevant in our case because we are not trying to compare Alice's proficiency to Bob's. Rather, what we want to do here is measure the same person's progress. I'd like to know how far I have advanced in Chinese and how fast I am moving ahead. I also want to be able to compare my speed today with my speed last week, last month, last year, etc.

Here's how I understand the know part for the practical purpose of measuring your progress as a self-learner:

You know a phrase if you are capable of casually recalling and saying it in an appropriate situation.

For example, to know the English phrase don't count your chickens before they're hatched is to be able to recall it within a couple of seconds or less and say it in a situation when someone is being overconfident about something that is not yet certain.

Does this all sound a bit wishy-washy? Well, this approach is as fuzzy as any conventional language proficiency framework, like CEFRL for example, which recognises 6 levels: A1, A2, B1, B2, C1, C2. Such systems are used all over the world quite successfully despite their lack of rigour.  They owe their success to the fact that people never need to know the exact size of a speaker's vocabulary. A rough approximation always suffices. Imagine you're about to hire a waitress for your cosy restaurant. During the interview, she produces a language proficiency certificate claiming that her CEFRL level is A2. You needn't be an expert to tell that if a candidate for this particular role has 2 stars out of 6 she is a fair fit. If you were to hire a customer support specialist then perhaps you'd want them to be at least B2.

Speaking of CEFRL, is there a rough estimate of how many words/phrases you're supposed to know on each of its 6 levels? There are estimates in terms of words. E. g., 3,000 words can place you on B1 and if you are a native speaker you probably know at least 30,000 words. I personally don't like using words as a basic unit of account. A good dictionary "knows" hundreds of thousands of words but since it can't say any of them, it doesn't even make A1.

I prefer learning and counting phrases. No real conversation is an exchange of words. We normally exchange full sentences, each of which is either a single, atomic phrase or a combination of such indivisible phrases. In terms of phrases, here are the numbers that I have obtained from my humble teaching experience:


CEFR Level Number of Phrases




Upper Intermediate




It is not a probelm these days to know the number of words and phrases that you've learnt. Any decent flashcard app will tell you how many entries you have added to it, and you can also figure out how many of them you actually remember.

So, the number of phrases that you know is basically the same metric as A1, A2, B1, ... The added benefit is that whenever you learn a new word, phrase or idiom you can see in real time how you ratchet your level up by 1.


Part 2: The Method

The core idea of my method of learning a foreign language is twofold:


1. Repeat after a native speaker
2. Expand vocabulary with flashcards


Learn Like a Child

Ideally, it would be great if we could learn foreign languages like children learn their mother tongues. A child initially doesn't know any vocabulary or grammar but mocks adults and picks up phrases and words at a marvellous rate. By the age of 7, he or she can talk about pretty much anything, watch TV shows, understand songs, and demonstrate other mastery-level skills that we second-language learners can only dream of. Note that most 7-year-olds don't know a single grammatical concept in their native language but speak both fluently and correctly.

One reason for children's remarkable ability to acquire their first language so quickly is that their brains are still much more malleable than in adults. Another is due to the fact that mastering language is a life-or-death question to them, because you can't get anything in this world without language, you can't even socialise with your peers without it. Children are also better at first language acquisition because they are immersed in the native environment for years on a 24/7 basis.

Unfortunately, adults have a much less malleable brain structure. They tend to think in their first language, even if they are fluent in a second one. Most adults don't have the luxury of living in the native environment of the second language. Their day jobs consume a lot of time that could be spent on studying were they carefree kids. Yes, our world is far from ideal but the good news is that adults can make use of the same language-learning tricks as children. Moreover, using them produces the best results!

We should just look at how children learn their first language, and learn our second language (Chinese) in the same way. Of course, we should also make up for the fact that we are no children any more.

The first core principle of the method I'm preaching here - repeat after a native speaker - essentially describes the way children learn. The second principle - expand vocabulary with flashcards - accounts for the fact that the user of the method is an adult.


Repeat after a Native Speaker

The first words children say in their mother tongue are simple phrases that they repeat after their parents. The pronunciation is far from perfect at first but it gradually improves with practice as the parents point out the child's mistakes. So, we should learn our first phrases in a second language in a similar fashion, by mocking a native speaker as best we can. We will make mistakes but that's totally ok, since even the best learners (children) make them. Our pronunciation will get better over time, provided we pay careful attention to how accurately we reproduce each phrase recorded by a native speaker.

We can and should use recordings of a native speaker's voice, e. g. sentences like`See you tomorrow' or `Have you seen my shoes?'  or anything else that is neither too short, nor too long. If a phrase is too short (say, 1 word or 1 syllable) then it generally doesn't serve any practical purpose. What's the point in being able to say the word for `seen'  ripped out of context? A full phrase like `Have you seen my shoes?'  makes much more sense and can be of real use to the student. On the other hand, if the phrase is too long it is impracticable to work with. Based on my experience, 3-7 word phrases are optimal for a beginner. More advanced learners can go for longer phrases.

Memorising and repeating a phrase consisting of, say, 5 words takes about the same time in total as working on just one word but, like I said, full phrases will allow you to use them in real-life situations right away.

What phrases to repeat?

This brings us to another important point of the philosophy of the whole method. We don't learn just any phrases, rather we should take them from practical, contextual situations. It is much easier to learn things in realistic context, especially when the learner is directly involved in it. That's exactly how children learn.

As a beginner, the best place where you can get phrases that meet all of the abovementioned requirements is an introductory audio course for complete beginners. A good course will contain:

  • Dialogues recorded by native speakers
  • An English translation of each phrase
  • Some cultural notes
  • (Almost) zero grammar

Intermediate to advanced learners can obtain new phrases from other sources, like YouTube videos, films, audio books, etc. No matter what your level is, as long as you feel that your pronunciation is unsatisfactory, you should learn new phrases mainly or exclusively from audio, not from written texts. Language is first and foremost an oral form of communication. Our brains learnt how to speak and process speech for about a hundred thousand years or more, whereas the first writing systems date back just a few millennia. In a sense, reading and writing are unnatural to us in comparison with oral speech.

A note about grammar

Some teachers raise a seemingly valid objection that since students are adults, they need to compensate for this unfortunate fact with some sort of shortcut, like grammar. Indeed, grammar provides you with general rules as to how to put a sentence together and therefore saves you the trouble of having to acquire phrases sporadically from your learning material. A hell of a shortcut, the only problem is, rules don't always work. Any natural language like English, German, or Chinese tends to have exceptions and weird idioms in the most common expressions. The more common a phrase is the more likely it is the exception to the rule. Since beginners and intermediate students should always learn more common things first, they automatically become the most affected victims of learning from grammar.

Certain grammar books throw a lot of redundant and brain-damaging information at you that is of little practical use. For example, the phrase `Have you seen my shoes?'  can be termed as interrogative present perfect indicative. Hmm... I can't put my finger on it, but something feels wrong about this approach.

I understand that it is difficult to create an introductory course completely free of rules and generalisations but if you look at what's available on the market for the Chinese language you'll quickly realise that the amount of grammar and its complexity varies considerably from course to course. So, we'll be able to choose one with as little grammar as possible.

How to repeat?

First of all, you need to make sure that you fully understand the meaning of the phrase you are about to work on. That's another reason why learning new phrases in context is so important. Even if you have no clue about the meaning of some words in the phrase you can still capture the association between the situation and what you need to say in it. After all, the ability to say the right thing in the right place at the right time is what language proficiency is all about, isn't it?

Familiarise yourself with the phrase by listening to it a few times. Pay attention to the overall intonation without focusing on any individual sounds, even if you can't hear some of them clearly. Try to treat the whole phrase as a single word. And

Please, do NOT put it down in writing!

Then play the recording of the phrase one more time and try to imagine saying it, without actually opening your mouth. Then say it under your breath a few times, playing the recording first and repeating after it. When you feel that you are ready to say it out loud do so. Keep mocking the speaker on the recording until you are satisfied with how accurately you can reproduce it. Just make sure you play the recording before each repetition.

I know that some students tend to play one phrase, repeat it one time, then proceed to the next phrase, repeat it once, and so on. When they hit the end of the dialogue (or whatever their material is) they return to the first phrase and do this whole cycle again. This is counterproductive because just when the brain starts to adjust to phrase #1 it is forced to switch to phrase #2 which contains different sounds in different order, and it has to re-adjust before any real learning can happen.

Come back to what you learnt yesterday, last week, last month...

Even if you have mastered a phrase to perfection and can say it now it doesn't mean that you should stop practising it altogether. Most likely, the quality of your pronunciation will deteriorate gradually, unless you come back to it from time to time. For example, during each session, you can repeat yesterday's phrases first, then the phrases you learnt 7 days ago, then 30 days ago, or something like that. You repeat those old phrases in the same way you work on the new ones, i. e. you keep playing the phrase and copying it to the best of your ability until you sound as close to the native speaker as possible.

It seems natural to repeat the old phrases before you start learning today's new material but I don't see why you can't do it the other way round, if that suits you better.

See also: My typical session of Chinese explained in detail.



Once you have mastered a phrase by following the method described above you'll see that you have acquired much more. Besides being able to whip out this phrase unthinkingly when you need it, you now possess a pattern. I define a pattern as a phrase in which you can replace its parts and thus use it in similar situations. Our example `Have you seen my shoes?' is a pattern because if you are able to easily blurt out this phrase then it'll take you no time to figure out how to say `Have you seen my hat?' or `Have you seen my umbrella?'. All you have to do is substitute something similar for one of the words.

If you've been learning the language (Chinese in our case) for at least 1 month you must already know a bunch of useful words. From this stage and on, you should practise modelling. This is how it works. You picture a situation similar to the one that the phrase was originally about. For example, if in the original dialogue someone was looking for their shoes and asking their parents whether they'd seen them then you can imagine yourself in a similar situation, say, looking for your umbrella. The more similar and the more realistic your imagined situation is the better.

We should only practise modelling on phrases that we know and can say fluently, without error.

Modelling best works for analytic languages such as Chinese, French, and English, where you generally don't have to change the form of the word you substitute. We are in luck because Chinese doesn't have any word forms at all, so modelling should work perfectly.



I mentioned earlier that our brains are much better at acquiring oral speech than reading and writing. It is possible to learn a language without using its writing system at all. In fact, most children learn their native tongues in this way. The standard of 7-year-olds is somwhere between C1 and C2 and yet many of them don't even know how to write a single letter. So, why shouldn't we do our best and learn like they do?

Starting to learn a new language is a stressful challenge. Some people naïvely think that to learn a foreign language is to get used to pronouncing the same sounds as in your native language in a different order; or that you can get away with using foreign words while maintaining the same sentence structure as in your native language. Real language learning is much harder than that. It is not like switching from eating with your right hand to eating with your left. A more appropriate metaphor would be that of a snake whose brain was somehow transplanted into a dog's body and whose challenge is now to learn how to walk and bark.

One of the biggest impediments that second-language learners face is their first language, which constantly gets in the way and weighs them down. It is impossible to get rid of this factor, since you can't unlearn your mother tongue. So, things are already very complicated by default, before you even take your first steps. That's why the learning process should be as simple as possible for beginners. You can't realistically hope to learn speaking and writing, poetry and slang in parallel. Your time is limited. Your brain hurts. So, I believe that if you are a beginner, it's best to foucs on one thing first, namely, on speaking.

When you can fluently use around a hundred basic phrases you'll be able to easily learn how to write/type them. It'll just take some time to memorise that this word here (that you already know with condfidence) corresponds to this set of characters. On the other hand, if you start reading and writing words when you first encounter them your beginner brain will be unnecessarily distracted from what's really important and difficult (correct pronunciation).

The bottom line is, we should put off reading and writing until A2-B1 (elementary-intermediate). Besides, we are going to learn Chinese. Its writing system is not alphabet-based, which means that there is no connection between oral speech and writing. Learning written Chinese is, in a sense, equivalent to learning a separate new language.


Expand Vocabulary with Flashcards

Children absorb new words and expressions like a sponge. Unfortunately, adults are slow learners compared to them. Flashcards are a wonderful tool that allows you to overcome the lack of children's malleability. The idea is simple yet brilliant. Regardless of whether you have linguistic flair or not, if you repeat a phrase long enough you will retain it.

For our purposes, we are going to use one-way flashcards, where a Chinese phrase is on one side and its English translation on the other, the latter serving as the cue. In other words, we are only going to train ourselves to recall the Chinese phrase given its English meaning. The reverse direction is unnecessary because if you are able to say the Chinese phrase and you know its meaning you will easily recognise it in other people's speech.

Flashcards imply the use of the writing system, which we have decided not to touch until we reach A2 or B1 (elementary or intermediate), so we are not going to practise flashcards until then. From our very first steps we will favour full sentences over single words when adding new flashcards. A single word can be put on a flashcard only if you are absolutely sure that its meaning is straightforward and matches the meaning of the corresponding English word. Such words are generally concrete nouns like cat, dog, armchair, etc.

The question of what to use as a writing system for learning is not a trivial one in the case of Chinese. Let's take a look at the reasons why:

  • There are two alternative writing systems in Chinese: traditional and simplified.
  • The phonetic system is totally detached from the writing system.
  • Handwriting and typing are two independent skills that you will have to learn separately.
  • Touch typing is impracticable unless you already know at least a few thousand characters.

We might use Pinyin (a romanisation of Chinese) at least for our first flashcards but I see no point in it because Pinyin is something that you will never need in real-life situations. Besides, typing Pinyin on a keyboard is essentially equivalent to typing characters with the Pinyin input method, so we might as well start learning real characters while spending the same time and effort.

Which set of characters to learn: traditional or simplified?

This was a very painful question for me. On the one hand, traditional covers a broader linguistic landscape as it was China's writing system of the olden days and the writing systems of other languages in the region (like that of Japanese, for example) overlap with it to a large extent. They say that traditional is also somewhat more consistent and switching from traditional to simplified is a bit easier than the other way round. On the other hand, traditional Chinese is used officially only in one country (Taiwan) and two cities (Hong Kong and Macau), whereas simplified is used by over 1 billion people.

Seeing as China is not going to switch back to traditional any time soon, I think it would be only wise to learn simplified.

We'll be working with flashcards using software like Anki or similar. Whenever we encounter a new phrase in our learning material we add it to the app by entering the Chinese phrase along with its English translation. The daily flashcard routine will consist of repeating most recent phrases and phrases added some time ago that we frequently forget. The system will automatically figure out which flashcards to challenge us with every day and adjust their frequencies according to how well we retain the phrases. Flashcards that we can firmly remember will have a gradually decreasing frequency while the frequency of "difficult" flashcards will be increased each time we can't reproduce their Chinese side.

A Single Repetition Instance

Some clarification is in order as to what a repetition is. So, the system shows us the English cue, say, `See you tomorrow'. We are supposed to recall the Chinese side (明天见) and say it out loud. Remember, language is mostly an oral skill, so we should say the phrase, not type it. As we discussed earlier, speaking, writing, and typing Chinese are three separate skills and we'll have to learn each of them separately. Our primary goal with flashcards is to grow our vocabulary of phrases, so we shouldn't mix this activity with less relevant skills, like writing and typing. Of course, this doesn't mean that we can't use another stack of flashcards for practising writing and/or typing.

Method Summary

The method we've described here is aimed at adults learning Chinese. It consists of two elements: repeating after a native speaker and expanding vocabulary with flashcards.

As a total beginner, the student should first focus on their pronunciation and learn basic phrases by copying and mocking the way a native speaker says them. An introductory audio course would best serve this purpose. At this stage, the student is not supposed to take notes or write down anything in Chinese.

When the student reaches A2 or B1 (elementary or intermediate) they can start learning the Chinese writing system. They should also start working with one-way flashcards at this level.

I believe we should learn foreign languages the same way we acquired our mother tongue, i. e. just by repeating after native speakers in contextual situations. However a method for adult learners should take into account that the brain of an adult lacks the malleability we see in children and acquires a second language in a different way. The second element of the method, flashcards, is essentially an instrument that compensates for the fact that adults are no children any more.

What's also worth mentioning here is an obvious point regarding real-world conversational practice. No technique beats actually chatting with native speakers, provided you know the basics of the language. No methods or tricks are required here. You either find a way to talk to native speakers or you don't.


Part 3: First Steps

Without further ado, I'm going to just start learning Chinese using this method.

The first thing I need to do is pick an introductory audio course. I've considered a few:

  • M. Campbell, S. Chen. Chinese Mandarin
  • Pimsleur Chinese (Mandarin) Levels 1-5
  • E. Scurfield, L. Song. Beginner's Mandarin Chinese

I have actually consulted a native Chinese speaker with regard to the quality and `standardness' of the accents in each course and taken into account the requirements I mentioned in this section. The verdict is in favour of Scurfield & Song.

It seems like it's safe to just follow Scurfield & Song's course lesson by lesson because it appears to be very well in line with our method.

So, at this point I have nothing more to say, except that I'd really appreciate your feedback. Wait for updates and 再见.