Chinese from scratch – a 1260 hour work Programme optimising your result.

Mandarin Ducks, Beijing Zoo Français : Canard ...
If you want to learn Mandarin you can’t duck out of the time it takes – you can only optimise it. Here we see some mandarin ducks optimising their time on a lake.

Reader Jarad Mayers wrote the following very good question:

I want to learn Mandarin. I am not sure how to go about it. This is the very first language I am attempting to learn. I have not done anything yet. I am on very tight budget and currently not employed. I tried to access the free material on Mandarin ( )but it is no longer accessible . I was wondering if I could use your experince and if possible sort of outline the steps I need to follow.

BTW, I am not sure where to post my question. I am sorry if this the wrong place for posting it.


I’ve prepared the answer as a table – it is a whole programme to 80% of Chinese that you’d need to get your degree, read newspapers, live an everyday life in China. The rest after that comes down to vocabulary building for which I’d recommend the goldlisting of dictionaries or of bilingual literature. You could spend four times as much time getting from 80% Chinese to 100% Chinese (ask Vilf “the Gilf” Pareto, he’ll tell you why, or might have done, until 1923 – now you’ll have to look up what he thought in order to know why, or simply accept it).

Real Chinese philologists like Victor Berrjod might give you other useful sources better than the ones I have listed. All of the ones I have listed are available on Amazon. The audio courses are expensive so it will pay you to shop around a bit.

1260 hours, based on the number of years/days in Daniel and the Book of Revelation, also happens to be a typical year’s work in a modern Western company. This means that the below work programme can be:

  • done in a year if you treat it as a full time job – you’ll need to plan that one out carefully in order to avoid time-clashing in Goldlist method, namely coming within 14 days of the last time you listed the word in the previous Distillation or Headlist.
  • done over two years if treated as a half-time job
  • done over three or more years if done in your spare time.

Don’t expect to learn Chinese meaningfully any quicker than that.

If you are sure that you want to get seriously into Chinese at an academic level, from my discussions with Victor “Hobbylinguist” Berrjod and concluding from certain things outlined by Mike “Glossika” Campbell. I’d recommend Goldlisting the Mandarin and also the Cantonese pronunciations of the characters in the character only Goldlist and the Words Goldlist.

Please note that in this Programme, quite apart from giving yourself 160 hours worth of work on audio only before even touching the Gold-List method, there are actually four separate Goldlists that you need and I suggest that at least at Bronze Book level you keep them quite separate. They are:

  • 1) pinyin only for words and simple phrases, example sentences from grammar
  • 2) character (one by one) including the elements of Hoenig’s explanations (alternatively Wu Gaofeng or James Heisig with Timothy W. Richardson, certainly Heisig rules the roost if it comes to Japanese and the whole method even Hoenig is using is Heisig’s really). You need several lines per character in the Headlist my opinion. One for stroke order, another for the primitives “story”, another for the pinyin sound and the meaning of the character. At this point in time you are not trying to combine them into words, that comes later.
  • 3) words, based on one to four of the characters, but most commonly two. You might want to split this into different groups depending on tone pattern, eg Groups 11,12,13,14,15, 21,22,23,24,25, 31,32,33,34,35, and 41,42,43,44, 45 exist, with tone 5 referring to the reduced tone or “closed fist” to follow the idea of Dr Goodman.
  • 4) whole practice sentences – you need practice writing but keep them in a separate Goldlist so as to not mess up the monitoring of progress by numbers in the other three. When distilling these you are choosing to let go forward the sentences you wouldn’t be able to write out from memory. You only need to listen to the Pimsleur or MT again when doing the Headlist and not for the subsequent distillations.In the first of the three Goldlists above you might consider using green, blue, red and black ink and pencil for the closed fist to drive home Dr Goodman’s way of teaching tones, which is sans pareil. You may also derive a benefit from making the enforcing movement with your free hand while writing the word with your pen hand.

    Anyhow, here comes the Programme:


    Feedback welcome – rather I should say obligatory, except for the fact I have now mechanism of making it so!

    And thanks very much for the fine question, which with luck will have kicked off some useful debate as well as helped other prospective learners.

42 thoughts on “Chinese from scratch – a 1260 hour work Programme optimising your result.

  1. Thank you for this excellent guide, sir. It is going to help me clear the fog when it comes to such an important yet difficult undertaking. I have a question:

    The time to complete Pimsleur lessons and Michel Thomas lessons that is mentioned above is more than the total time for the audio in it. For example, each of Pimsleur’s levels is around 15 hours of audio (excluding the culture lessons). Have you written 30 hours for each level because you expect us to have to repeat the lessons, or is repeating the lessons itself a part of the plan?

    Thank you.

    1. I am giving an average of two listens on the Pims and MT, which I would do spaced apart by more than two weeks for precisely the same reasons as why that works on ordinary Goldlisting. The point about audio-frontloading before a GL project is to obtaini and refine one’s “inner voice”. It is not a great idea to write out and learn by writing words which we have no idea of how they sound, and then feel very disappointed with ourselves when we can say that we have goldlisted a large volume of material down to next to nothing and therefore have all the grammar and an impressive vocabulary, and can read fairly comfortably, then turn on the radio or an audiobook and understand far too little. We don’t want to be in that situation and we won’t be, if we have a grounded inner voice which means we can imagine quite accurately the sound of a word as we write it out in the lists or look at it again on review later on, whether we actually say the word aloud (once doesn’t hurt but no mantras or we’ll be back with the short-term memory) or just “hear” it the way you “hear” a character’s voice when you read a novel.

      1. This is, for me, the place where most clearly and concisely I see you expressing the justification for why we should study the pronunciation of a language before trying to learn the vocabulary and grammar of a language. You’ve mentioned a desire to compile your work in a book, and I would urge you to include these points at the beginning. For that matter, I recently realized that the only places where I recall having seen you make this argument are in the comments, not in the introduction or main points to the article above. When I’ve referred the system to others, I’ve been careful to include the links to the relevant comments, but it would be good, I think, to include the instructions for this critical front-loading work at the beginning of the main Goldlist article above. In this age of infinitesimal attention spans (perhaps nowadays better called moments of attention), we can’t expect that the everyday Facebook user who lands here will apply due diligence in ferreting out the explanations. Sure, I sometimes doubt whether someone unwilling to dig through the comments would have the application and diligence to follow the system to a T, but that’s probably a less constructive or charitable attitude to assume.

        1. P.S. I mistakenly referred to the ‘article above’ and the ‘main Goldlist article above’, forgetting that this is the ‘Chinese from Scratch’ page. Without the above, I mean the goldlist-eu article linked to in the top bar.

        2. Thanks for this, and thanks for recommending the system. For sure this book needs to be written this year an d only one piece now is needed to fall into place in order to go ahead with that, namely a model Indonesian Goldlist photographed at each Entstehungsstufe.

  2. And another addition:

    One can start with characters right away if one studies the radicals first. If there is a solid understanding of the radicals the characters will be easy to write and learn. But I would spend some time in repetition (and better calligraphy too) at that moment (not only gold list). Because here repetition is very use full for the feeling of the right stroke order and to get it done fast. One can write characters ten times or even more and start to get faster for the last repetitions for example (just a technical skill like in music, there you need more repetition because its a physical skill). Later only gold list should be enough. Sounds are not so difficult and can be leaned faster than you suggest in my opinion.

  3. I studied Chinese and want to add that in my opinion 1. Stroke order will be automatically right after a while even with new characters in most cases, especially when one spends some hours into calligraphy. This wont take very long. 2. I think to study classical Chinese (mainly from the Zhou period) is far more important than Cantonese and it will help a lot to know the traditional characters, the single characters, pinyin, meaning, history and philosophy of China and therefore get a much deeper contact to the language (better listening skills too, because of this). It is a key (and mandatory to a degree) for the higher education (and too understand a lot in between the lines or idioms etc.) and probable way more important then cantonese, because most of the important canonic literature is written in that way (Laotzi, Dschuangzi, Kongzi and so on). Of coarse cantonese is very helpful, but i would do it only after a doing putonghua and classical Chinese. And maybe Japanese too at this time. But thanks for your method anyway, because it is a great help to heal the stress induced damaged they forced me in during studying and university. After a few years I restarted learning it, because the fun is back now :-).

  4. Learning mandarin in three phases makes it easier to understand and learn correct?

    1. I would say so, Henry. But if someone needs a result within a shorter period and is giving themselves over to the study of Mandarin more or less full time, then a different approach might work better. I think this is the ideal approach for someone not in a hurry, who has plenty of other things on but can give a few hours a week (from 2 or 3 through to 10 or 12, but the speed of covering the 1260 hours will vary accordingly) to Chinese. For such people the breaking of the task down into its components makes better sense.

        1. Sorry I do not know why the url appeared. Would the other approach be goldlisting the characters and pinyin at the same time?

          1. Yes, trying to do everything at once. That is more demanding. The distillation rate will be slower. For those with a lot of time to give it might be preferable – for those trying to follow along with a course but using Goldlist to aid memory it may be unavoidable.

            1. Doing everything else seems like it would take more than 20 minutes to do.

    1. I would do about 20 minutes at a time. Pimsleur lessons are half an hour long and this is the weak point in Pimsleur. You cannot ask the subconscious mind to work for half an hour without a break. Even the conscious mind has trouble with that. If you leave adequate breaks, I don’t see why you shouldn’t do five or six such sessions a day if you are hungry for it, and enjoying it.

      1. In one day, could I do five sessions of pimsleur mandarin and five sessions of cantonese mandarin?

          1. You could do ten sessions a day but in general I don’t advise learning related languages simultaneously, but instead getting one of them to at least 10,000 words and then using materials in that first language of the given family to learn the second one.

            1. Cantonese was my first language,but then I never learned more and just spoke english. I can still understand some and speak some. I can learn again quickly. In this case, would it be okay for me to learn Cantonese and Mandarin simultaneously? I would just learn to speak more and learn reading and writing later. Speaking is a bit different between them.

            2. I would just learn to speak more cantonese and learn reading and writing after reading and writing for mandarin. Cantonese was actually my first language up til 1st grade. Is this okay?

  5. For phase 2, I concurrently make the pinyin goldlist for grammar book, TY and phase 1 audios correct?

    1. You can have three lists going for each resource or do them in a single Pinyin list one after another. I would recommend the latter course as it is less messy and easier to keep table on your overall progress.

      1. The Elisabeth Scurfield’s TY Chinese should have the audio with it correct?

  6. I was thinking of goldlisting the 200 most common radicals as a way to understand words. Would that be okay to learn characters? Also how those going from audio to pinyin to character make mandarin easier to learn?

  7. I really like this post, very informative, good book tips, and last but not least, inspirational! It restarted my efforts to learn putonghua from 8 years back. More language guides like this for French, Spanish, etc would be useful for a lot of us. It’s hard to navigate all the English material on the specific languages.

    I met HobbyLinguist at a China-studies BBQ, he showed me the Goldlist method. Afterwards, he gave me lots of useful tips on learning Mandarin, and linked me here.

    How did you use the audio section?
    How much time on average did you spend on this language compared to others?

    Thank you for creating this wonderful resource!

    1. Eivind, nice to hear from you. I hope you stick around despite my tardiness in replying and read some more bits and pieces and watch some of the video resources also. Nice to see that VB’s enthusiasm is turning into new readers for me, and thanks to him for that.

      The audio is just simply that. In the first case when you do a MT course or a Pimsleur course you are not using your hands to write or your eyes to read. You can of course walk around listening to it but by far the best effect is to sit with your eyes closed and your finger on the pause button, so that you can do the answers. And again I would do about 20 minutes of work and then take a break, just like on the Goldlist. I would not seek to be perfect before going on to the next audio lesson. If you get 80% you can move on because most of it repeats anyway. If using Harold Goodman’s course use the hand gestures as he suggests until you get used to the tones.

      I personally have not spent much time yet on Mandarin. I have done only 5% of what needs to be done by me on this language. But I am intentionally delaying for the sake of getting my Japanese up to scratch first. This will enable me to reduce interference between the Japanese and Chinese ways of using kanjies which can occur if learning them at the same time. In Japanese I am maybe 30% of where I want to be but that takes anyway rather longer than I would expect Chinese to take.

  8. I wonder if you could elaborate on combining Cantonese and Mandarin on the Gold Lists? I am an intermediate Mandarin speaker and plan to study Cantonese starting next year. Should I start including the Cantonese pronunciations on my Mandarin Gold List now, or wait until my focus is on Cantonese?
    Thanks for this great outline!

      1. Up to this point, my studies have been for college courses, so I haven’t kept a Gold List, but now that I am free to structure my studies, this is the method I will use. I know around 5,000 (my estimate, may be lower) words and the corresponding characters. For class I had to memorize the writing of each character, but I will only be working towards character recognition and pronunciation now. My list will have both the pinyin and the simplified (and maybe traditional) characters. I am working on expanding my vocabulary (probably a focus on nouns), so the words won’t be terribly advanced. I am wondering if it would be wise, in your opinion, to go ahead an work in the Cantonese pronunciations as well, or save that for a later time when I can start from the beginning with that set of vocabulary? I have no study experience with Cantonese, except for casual exposure to the language while living in Hong Kong. I am familiar with the traditional characters. Thanks for your input! This method helped me prepare for the GRE exam several years ago 🙂

        1. I’m in the closing stages of a ~4000 character project where I’ve been goldlisting the Cantonese and Mandarin readings as well as the meaning of the characters. When I finish the bronze book in the next month or so, I’ll start a vocabulary project using a source list I found for the HSK. My plan is to run it through the great parser you can find at and use the format as for the current project, namely one line per reading and one for meaning:
          20. 安全 – on¹cyun⁴
          21. 安全 – ānquán
          22. 安全 – safety

          In this case, you can easily deduce the Mandarin reading just from the Cantonese one, though, so I would just put the Canto one. You’ll find that most of the time, you won’t even need to include the Mandarin one unless there is some kind of irregularity going on.

          Before you start, I recommend reading the Wikipedia article about the four tones of Middle Chinese if you don’t know about them already, and you can also check out Mike Campbell’s excellent video about the universal Chinese tone system. I might as well give you the mapping between Canto and Mando while I’m on it; you can skip to the last paragraph if you already know it.

          You may know that Cantonese has 6 phonemic tones. It’s still useful to separate the 入聲 for philological purposes, because although they are not distinct in Cantonese, they are in some other Chinese languages. Mandarin has not preserved any of the 入聲, and only one of the (historically voiced initial) 陽 tone syllables, but Cantonese still distinguishes all of them. This means that it’s easy to predict the Mandarin (or any other Chinese language, as well as Japanese, Korean, and Vietnamese) reading of a character if you know the Cantonese one.

          平 上 去 入
          陰 1 2 3 7/8
          陽 4 5 6 9

          These are the numbers given to Cantonese tones. Note that 7=1, 8=3, and 9=6, except for the presence of an obstruent in the syllable coda, e.g. 入 is jap⁹ or more commonly jap⁶.

          In conservative speech, tone 1 is further divided into a high level tone and a high falling tone. None of my friends from either Hong Kong, Guangzhou or Shenzhen makes this distinction, though. Interestingly, tone 4 (also 平) is sometimes described as low and slightly falling, sometimes as low and level. The difference is very small, in any case, but it does make for an interesting symmetry. The other categories are more symmetrical, though.

          The 上 tones both rise, with 上陰 being higher and 上陽 being lower. The 去 tones are level, with 去陰 being higher (but not as high as 平陰), and 去陽 being lower (but not as low as 平陽). 入陰 is has split into a high tone with short vowels and a mid tone with long vowels, but they are both level (入陰高 and 入陰低). Finally, 入陽 is low.

          That’s not the whole story, though, because Cantonese has a very common changed tone, which can change any tone lower than 平陰 (and obviously 入陰高) into the second, or 上陰, tone. This can even affect 入聲 syllables, like in 擒賊 kam⁴caak⁶*².

          To derive the Mandarin tone from a Cantonese one, you simply look at the category:
          平 上 去 入
          陰 1 3 4 ?
          陽 2 ↑/4 ↑ 4/2

          Unfortunately, the 入陰 tones cannot be derived automatically, although you’ll see tendencies of syllables with the same pronunciation in Cantonese having the Mandarin equivalent. The 上陽 that have an obstruent (that is, anything other than (jyutping) m, n, ng, j, w, l) in the syllable onset, will typically be T4 in Mandarin, while the rest will stick to their category and be T3. The same thing goes for the 入陽 characters, except that because Canto T6 regularly turns into Mando T4, the sonorants’ regularizing it results in T4. Obstruent onsets turn into T2. Examples: 市 is si⁵ in Canto and shi⁴ in Mando, but 女 is neoi⁵ in Canto and nü³ in Mando. There are several exceptions, though, but the 入陰 ones are more regular and go like 入 jap⁶ (Canto) ru⁴ (Mando) and 十 sap⁶ (Canto) shi² (Mando).

          Be aware that historically breathy voiced initials, although now voiceless in both languages, may still be aspirated in one but not the other. That is the case with many characters, for example 賊 caak⁶ (Canto) and zei² (Mando), but both are often aspirated, as with 平 ping⁴ (Canto) and ping² (Mando). If you decide to learn Shanghainese or another variant of 吳語, you’ll be able to predict voiced initials from Cantonese low (陽) tones, in any case. For example, 平 is IPA /biŋ/ in Shanghainese, which still distinguishes /pʰ, p, b/.

          In the beginning, you may want to list both readings unless you know the Mando one already, so as to get used to the relationship between them. Soon, it will become second nature, and you’ll easily distill the extra lines away. When it comes to deciding on simplified or traditional, I personally went with traditional because it’s easier to learn how to read simplified variants of characters you already know than the other way around, traditionals are used for Korean hanja, and they also look more like Japanese kanji than most of the simplified ones do. Another important reason is that I prefer the looks of the traditional ones. An added bonus for traditional is that they are the characters used by Cantodict; although you can still search using simplified, all the entries are in traditional.

          Finally, sorry for this long message. I hope it makes sense and helps you out. Please do ask if you have questions!

          1. Thanks Victor for the detailed response! There’s a lot of good stuff in there! I hadn’t thought to separate the meaning and the pronunciation into two separate items on the headlist, but that actually makes a lot of sense. Do you always do it this way, or sometime include the character, pronunciation, and meaning all as one item?

            Thanks for the feedback!

            1. My very first four pages contained all on one line; that was back in December 2010, but the problem with that was that I could remember many half lines, but few whole ones. When I split them, I would occasionally remember too many, but that’s better than too few. I may put everything on one line if I know that the readings of all the characters are already covered, but even then I don’t worry about saving space. If I already know how to pronounce a word just by looking at the characters, I’ll drop the reading altogether and just include the meaning.

              If I know how to read the characters in a word, but there is one that doesn’t get its normal tone (Mando), or it has a changed tone (Canto), then it’s very tempting to put everything on one line, but I’ve decided I’ll spend an extra line in those cases too. They can easily be merged later.

              I’ve found that this approach works wonderfully with words, but not so wonderfully with single characters. I think it’s because single characters are less specific. If I were to do my character project over again, I would use Cantodict to find representative words containing each character instead.

  9. Thank you for referencing my piece on nciku. I am so glad you like the post, and I’m sorry it took me so long to respond. (I like your grid, by the way, and also the idea of learning language via long-term memory and vocabulary.)

  10. Hi,
    This is a nice matrix and easy to follow. This is exactly what I was looking for.
    I really appreciate it.

  11. I’m impressed with how you’ve made this, and in my favourite format; table! I used to think audio courses were overpriced bullshit that wouldn’t teach you anything you couldn’t learn from Wikipedia, but since you and several other accomplished polyglots recommend them, I’ve realized I have to give them a fair chance and try one. It’ll have to wait until I start a new language, though.

Your thoughts welcome, by all mean reply also to other community members!