Language Resources
"Desire." You have to have a reason to want to study Chinese if you actually wanna stick with it in the long term. this is obvious of course, but more specifically you have to have some motivating content to 'CONSUME' to keep you at it, basically anything is fine–political news, literature, history, chinese cosplay pornography, biaoqing memes, wuxia movies, mao, whatever. find something or ideally multiple somethings and stick with it.
—neettelemachus
this is a list of language resources i have personally relied on and found helpful, or at least scoped out and hope to use more later. for reading/translation languages, it's readily possible to self-study, but for a language where the goal is speaking, i wouldn't consider it profitable to spend time self-studying without past history or future plan to do in-person coursework.
mandarin
the only mandarin textbook series i've particularly liked is the NTNU MTC A Course in Contemporary Chinese 當代中文課程. it has six volumes, and i've worked through vols. 2–~5. keep in mind that this teaches taiwan-style standard mandarin (aka guóyǔ 國語), unlike materials that follow a mainland standard (pǔtōnghuà 普通話). that means traditional and not simplified characters, no érhuà, fewer neutral tones, and a variety of minor differences of vocabulary and tone choice. as someone with historical interests, i favor learning to write and read in traditional and only read in simplified, but putting both on your flashcards is trivially easy. for this series i use the physical books, but i know there are pdfs floating around online.
an online resource i relied on heavily when i was first self-studying is the allsetlearning chinese pronunciation/grammar wiki (simplified characters). there is also an anki deck containing the grammar examples from that site, which is excellent for drilling grammar.
literary sinitic
Paul F. Rouzer, A New Practical Primer of Literary Chinese, Harvard East Asian Monographs 276 (Cambridge, Mass: Harvard University Asia Center : Distributed by Harvard University Press, 2007).
Morgan Rouzer's textbook is serving me well. as she explains, classical chinese—Warring States period writing—is upstream of all later forms of literary chinese. but those other forms often consist of specialized genre codes (buddhist, medical, etc.) which must be studied separately.
i have created an anki deck of vocab flashcards for this book. there already existed a deck of hanzi-to-english vocab cards for Morgan Rouzer's Primer, but it didn't cover all the lessons. i followed that model and recreated this more complete set from scratch, though it's regex work and not without errors. there are cards for both recognition and recall, and senses of each term are split up.
Paul W. Kroll, A Student’s Dictionary of Classical and Medieval Chinese: Revised Edition, Handbook of Oriental Studies. Section 4 China 30 (Leiden ; Boston, MA: Brill, 2017).
Kroll's dictionary has made working with literary sinitic massively easier. you can buy it in pleco, and there also exists a pdf version. after getting it in pleco, i discovered that for licensing reasons, paid dictionaries don't work with pleco's card export system. again using regex, i made my own deck, again separates and and numbers different senses of words. i haven't uploaded this anywhere, but if you need a copy, do reach out. because it contains the entire dictionary in order, the best way to use it is to first suspend all cards, then manually search for and unsuspend the ones you want to learn.
other canonical dictionaries i've found useful
- the Digital Dictionary of Buddhism run by Charles Muller
- Hucker's A Dictionary of Official Titles in Imperial China: a little old (uses Wade-Giles), but as i understand it, the english translations provided here are standard.
- Le Grand Ricci: chinese to french, the ultimate fallback. especially notable for its coverage of medical/botanical terms.
John Kieschnick and Simon Wiles, A Primer in Chinese Buddhist Writings, 3 vols., accessed January 3, 2025, https://religiousstudies.stanford.edu/primer-chinese-buddhist-writings.
i have not used this, but a friend has, and it seems to be the best introduction to reading buddhist texts.
CUP How to Read Chinese Literature series
this series consists firstly of anthologies of translations and essays covering different genres of writing (poetry, prose, drama). i've learned a great deal from the poetry one in particular. each is accompanied also by a literary chinese course text; i haven't worked through any in detail yet, but hope to do the prose one at some point.
Zong-Qi Cai and Cui Jie, How to Read Chinese Poetry: Workbook (Columbia University Press, 2012).
Jie Cui, Yu cai Liu, and Zongqi Cai, How to Read Chinese Prose in Chinese: A Course in Classical Chinese, How to Read Chinese Literature (New York: Columbia University Press, 2022).
Yingde Guo, How to Read Chinese Drama in Chinese: A Language Companion, How to Read Chinese Literature (New York: Columbia University Press, 2023).
hokkien
Bernhard Führer and Xiufang Yang, Southern Hokkien: an introduction, 3 vols. (Guo li Tai wan da xue chu ban zhong xin, 2014). audio POJ 台閩字 eng
this textbook definitely seems the most logical and comprehensive, as well as the least tai-specific. it does take median Taiwan usage as default, but tries to make you aware of Quanzhou/Zhangzhou/Amoy differences in pronunciation, and sometimes flags vocabulary differences from SEA. POJ romanization is the primary form of writing taught, and the characters it does provide are chosen based on putative philological considerations, rather than historical usage or MOE standard—not good to employ unreflectively in communication. the audio for the textbook is available here, and NTU has a public video course in mandarin, by one of the authors, covering the first 6 (of 12) lessons.
other textbooks
Hsiao Hui-ru 蕭惠茹, Practical Taiwanese Conversations 《實用臺語會話》
Hsiao has taught Hokkien to foreigners for close to three decades. the dialogues and audio in this are very useful. however, it uses a very idiosyncratic romanization system with no tones (she just has students manually write them in during class), and no characters.
Lau Seng-hian 劉承賢, Superb Taiwanese Primer audio tâilô 台閩字 官白 english
i tried to take Lau's class, and was very unimpressed. the textbook is shallow and badly organized, and i can't recommend it. for example: it uses the ROC MOE's recommended characters, and does not note which terms are Taiwan-specific loanwords (e.g. lìn-gòo "apple" < jpn 林檎 りんご). it reserves explicit teaching of phonology and tone sandhi for later chapters, and instead dedicates an early chapter entirely to vocabulary for fruit. unlike Southern Hokkien: An Introduction, it includes audio recordings of all dialogues, which i do find helpful, but they are read in a terribly robotic voice.
Taiwanese Made Easier audio POJ 官白 eng
- available freely online, the audio i found helpful for phonics practice.
- the squaracters used in this course should not be relied on; they transcribe using borrowed Mandarin words rather than attempting to accurately represent the corresponding Hokkien.
hokkien/min discord eng
this server is a pretty nice environment, with mostly SEA hokkien heritage people.
dictionaries
Dictionary of Frequently-Used Taiwan Minnan 教育部臺灣台語常用詞辭典 tâilô 台閩字 官白
the MOE sutian is the most convenient reliable dictionary, though the definitions are obviously in mandarin.
wiktionary POJ 台閩字 eng
when it works, wiktionary is the most convenient source for english definitions. unfortunately, it also has serious issues. it does badly distinguishing readings between dialects or literary/colloquial. the project is also very mando-centric, so lots of definitions are not explicitly marked "mandarin," but aren't necessarily relevant to any other sinitic language. you sometimes have to rely on context clues, like whether a hokkien example sentence is provided, to determine what it's telling you.
Zhou Changji, ed., 闽南方言大词典 Minnan Fangyan Da Cidian, Di 1 ban (Fuzhou Shi: Fujian ren min chu ban she, 2006).
the main mainland dictionary. there's no fully searchable copy, but if you google, you can find a version in mdict format as well as the pdf on anna's archive. the mdict version allows you to input a term in characters and be shown an image of the relevant page, and also has pronunciation audio.
iTaigi 愛台語 tâilô 台閩字 官白
a crowdsourced dictionary of nan/cmn corresponding terms.
chhoetaigi 台語詞典 POJ tâilô 台閩字 官白 eng
aggregate of nan/eng, nan/cmn, nan/jpn dictionaries, including an old version of the MOE sutian and itaigi. both of the missionary english dictionaries included (Embree and Maryknoll) give only romanization, not characters, which is very frustrating.
penang hokkien dictionary 台閩字 eng
Penang Hokkien–English dictionary by a guy who uses an idiosyncratic romanization
hokkienhub POJ 台閩字 eng
a janky Philippine Hokkien–English webapp dictionary
pronunciation, romanization and writing
text to speech and script converter audio tâilô 台閩字
attempts to convert characters to romanization and vice versa, though it tends to choke on vernacular characters. most importantly, can generate machine speech for a phrase.
taibun transliterator tâilô 台閩字
alternative site where you provide characters and receive romanization
tâijī chhân 台字田 POJ 台閩字
provide a romanized word and receive recommended characters based on historical usage (as opposed to MOE prescriptions).
i personally find many of the MOE's character choices ridiculous, arrogant, and unnecessary, so this site is a lifesaver. unfortunately, to a degree it remains parochially "taioanese," and so sometimes will recommend things like katakana that are obviously unacceptable if what you're interested in is hokkien as a language that was written on both sides of the strait and across the south sea.
tone drill audio POJ tâilô 台閩字
listen to audio, repeat back and try to correctly guess the romanization
syllable soundboard audio tâilô
the only www pronunciation reference of this type i've found so far.
河洛話讀古文 by Alan Jui 居正中
ancient prose texts annotated with literary readings
臺語呼音音樂寶鑑
teaches pronunciation for ancient poems, scriptures and spells, old and popular songs, etc.
media
anime video tâilô_hardsubs 台閩字_hardsubs 官白_hardsubs eng_softsubs
- the MOE has released dubs of several anime, though you need an IP in taiwan to access them. the most notable ones are probably Bocchi the Rock, Barakamon, and Silver Spoon (Gin no Saji). each is available in romanized, squaracter, and Mandarin hardsubbed form, and it's not difficult to find and extract English softsubs from a release of the original Japanese-audio version.
- a couple more dubbed anime are
pili
input methods tâilô 台閩字
for both Mandarin and Hokkien, i think you're best off with RIME. you can use it as an fcitx5 plugin on linux and android. use the config manager rime-plum to install rime-hokkien support for you. there is also a mobile input method called phahtaigi, though it is far less configurable.
fonts 台閩字
i use the late Andrew West's BabelStone Han, which supports a wide range of uncommon squaracters.
existing resource lists
i won't try to mirror existing lists, just draw out what i find most interesting.
食飽未 forums Getting Started & Resource Guide eng
useful stuff here, but i would like it to be noted that the person who wrote the list and runs the forum is an insane pro-Trump transphobe whiteoid ameriKKKan, as well as a vicarious/backseat hoklo nationalist who believes that the integral "taioanese" nation is, to this day, a victim of the completely foreign chinese (all of which is made clear on his twitter).
learntaigi.com Resources for Learning Taiwanese eng
Tâi-gí做伙耍 官白
chaotic but comprehensive set of links to learning materials and media.
漢語學習筆記 by Alan Jui 居正中 官白
links and videos that seem especially concerned with reading literary sinitic aloud
french
Celia Brickman, A Short Course in Reading French (Columbia University Press, 2012).
i used this for a semester on reading french in undergrad and liked it. my french skills have totally rusted, but i might try it again.
software and study methods
what is the appropriate role of software in language learning? how large of a role can it legitimately play before overstepping its limits and becoming a distraction? it's very easy to fall for gimmicky tools that promise to make unnecessary the hard work of learning, or to spend more time experimenting with methods than employing them. i'm going to explain some tools which are complex to set up, but i will only do so to the extent i have personally found them to be worth that effort, because they clear a path for more effectively going about the daily, incremental, rote parts of learning which are most important. among the four basic language skills, listening and reading are input skills, while speaking and writing are output skills. software can systematize and streamline your exposure to spoken and written input, making sure that instead of stagnating among too-easy materials or bouncing off excessively difficult ones, you climb an easy and consistent slope. it cannot teach you to compose beautiful essays, but it can ensure that a collection of vocabulary and grammar rules increasing linearly in size is retained indefinitely at only a fixed cost in daily review time. using these skills in an integrated fashion is another thing software cannot accomplish for you; naturally it is up to you to find opportunities to practice spoken and written communication in real life.
anki
anki is the most potentially life-changing software in existence, a force multiplier on effort-invested that has no skill ceiling. its only "downside" is that you have to actually use it; it's completely on you to set a routine in which you do your flashcards every single day. if you aren't doing your flashcards, you have no one to blame for your unsatisfactory learning progress but yourself. people will rightly tell you that anki isn't everything, and you can't learn a language just by doing vocab memorization on your phone really hard. that's true—but less true than they think, and routine work in anki is certainly the basis for the rest of everything.
you will sometimes see people distinguish two types of vocab
flashcards: recognition and recall.
recognition means input; when you see (or hear) something in your target
languagea term i'm misusing to mean "the language you're
translating your body into."
and must try to understand what it means intuitively or
in your source language.the language you're translating your body from.
recall, conversely, is when you're prompted with
something in your native language and must recall the corresponding
fragment in the language you're studying. i belong to the faction that
swears by recall cards for vocabulary. you should be looking at a prompt
in your native language and working to recall the corresponding
pronunciation and written form of the word in the language you want to
learn, not the other way around. a mechanical connexion between an
english and a chinese term does not guarantee you can employ the chinese
term fluently, but it is a base to rely on when you do practice those
output skills. recognition is also important, but it is better to use
sentence cards with natural usage context than to try to recall
individual words.
having accepted all this, the main issue left is: how is one to acquire a sufficient body of example sentences, and ensure that one only studies directly pertinent ones, rather than ones that are excessively easy or difficult? the latter issue is solved by the add-on ankimorphs, which is complicated to initially set up, but incredibly powerful. ankimorphs analyzes the cards you've already studied to see which morphemes ("morphs") you know well, and which you are still learning. it then reprioritizes the cards you've never studied, so that you will only encounter ones which contain morphs you need to work on, or morphs that would be valuable to learn next. the intrinsic strength of SRS systems like anki is that they will always tell you exactly what you need to be reviewing; ankimorphs extends this principle to new territory by also telling you exactly what new information you need to be studying. in my setup, ankimorphs controls the order in which i see new sentence recognition cards, but for vocabulary recall cards, i mostly determine exactly what new vocab i want to study. some of this vocab is what i encounter in class or elsewhere outside anki, and some i create based on the previously-unknown terms that come up on the sentence cards ankimorphs chooses.
one last caveat is that ankimorphs does not know how to distinguish between different languages. if you want to use it with multiple, you will need to create multiple anki profiles and configure them separately, and set them to all sync to the same ankiweb account so that they can share data. this adds a little bit of hassle in switching between profiles, but it's not too bad. remember to set your secondary profiles not to sync media, or you may end up with unnecessarily bloated disk space usage!
when creating a sufficient body of sentence cards for ankimorphs to work with, there are a few sources.
- you may be able to find anki sentence decks for a language just by googling.
- using the movies2anki addon, you can feed in movies/dramas/anime you like and create cards with short video/audio clips. this is the main way to integrate listening practice into anki, rather than only reading.
- i have experimented with feeding novels and other long texts i enjoy to bertalign so that i can create bilingual cards from them. this is excessively complicated, so probably not worth it to you.
- if you read texts in a browser with yomitan, you can send words/sentences you look up to anki.
- as will be discussed next, pleco is a very powerful dictionary app that can help you fill anki with both chinese vocabulary recall cards and example sentence cards.
at lower levels of skill, it's most convenient if your sentence cards have an english translation on the back, but because ankimorphs will only show you cards with up to one unknown term, you may be able to do without, especially as your skills advance—like a native speaker would, simply fill in the meaning from context, or look up the unknown word in the dictionary and add it to your vocabulary recall cards.
pleco integration
pleco is a must-have chinese
dictionary app, which on android supports exporting dictionary
definitions to ankidroid flashcards. setting up the feature is a bit
complicated, but well worth it, and i rely on it daily. once it's
configured, you merely need to tap the +
button on a term
entry to add a vocab card to anki. further, you can add example
sentences as their own cards. my habit is to spam-add several example
sentences at the same time as i create vocab cards, or when i encounter
a term while reviewing and feel uncertain about its usage. ensuring in
this way that you will run into the word's various usages separately as
you study is far
more effective than simply staring at the dictionary entry, hoping
to comprehend and memorize the different usages all at once.
TODO reading texts (yomitan)
addons
i have made use of a few addons that helpfully extend the features of anki and what kinds of material it's suited to learning.
- ankimorphs: described above
- ankimorphs-chinese-jieba: used by ankimorphs for identifying terms in chinese cards. i've found it to work alright for hokkien and literary sinitic as well as mandarin.
- movies2anki: if supplied with a video file and two sets of subtitles (native and target language), creates cards to test your comprehension of individual dialogue lines. great for practicing listening/comprehension based on movies/tv you're already familiar with.
- lyrics/poetry cloze generator: for each line n in a poem, tests whether you know line n+1
- image occlusion enhanced: allows you to black out the labels on a diagram and test whether you know the points indicated; good for learning maps
- leaderboard: syncs your flashcard stats to a public leaderboard so you can compete against your friends as a source of motivation. don't lose your streak! my username is DarkWarrior, and i've created a public group you can join for mandarin learners.
- ankiconnect: makes possible yomitan integration.
- hanzi stats: for mandarin; checks your vocab flashcards against HSK/TOCFL lists for coverage and missing characters