[GavinMcG]

Furthermore:

> You know tens of thousands of words instantly, on sight. How did you learn to do that?

> It happens through a process called "orthographic mapping." This occurs when you pay attention to the details of a written word and link the word's pronunciation and meaning with its sequence of letters.

> ...

> That requires an awareness of the speech sounds in words and an understanding of how those sounds are represented by letters. In other words, you need phonics skills.

> ...

> when children don't have good phonics skills, the process is different.

> "They sample from the letters because they're not good at sounding them out," said David Kilpatrick, a psychology professor at SUNY Cortland and the author of a book about preventing reading difficulties.

> ...

> "The three-cueing system is the way poor readers read," said Kilpatrick.

> And if teachers use the system to teach reading, Kilpatrick says they're not just teaching children the habits of poor readers, they are actually impeding the orthographic mapping process.

> "The minute you ask them just to pay attention to the first letter or look at the picture, look at the context, you're drawing their attention away from the very thing that they need to interact with in order for them to read the word [and] remember the word," Kilpatrick said.

[throw333542]

This idea is fascinating and makes me wonder if languages like Chinese are naturally learned this way, since that’s really the only way to learn it.

[jdietrich]

Chinese schools teach using phonics - specifically, using a phonetic transcription system based on the Latin script called pinyin. Thanks to the exact correspondence between spelling and Standard Chinese pronunciation, this system can be rapidly mastered. Once students have a working command of pinyin, they are taught to associate traditional Chinese characters (hanzi) with the pinyin words that they already recognise.

The introduction of pinyin revolutionised Chinese language teaching and had a near-miraculous effect on literacy rates; As a second-language user of Mandarin Chinese, I am indescribably grateful for it.

https://en.wikipedia.org/wiki/Pinyin

https://sci-hub.tw/https://doi.org/10.1080/09500780508668693

[AlchemistCamp]

> The introduction of pinyin revolutionised Chinese language teaching and had a near-miraculous effect on literacy rates;

This is inaccurate. Taiwan continued using Zhuyin, the native syllabary China used prior to the 60s, and Taiwan's literacy rates have never lagged.

Pinyin and Zhuyin have a one-to-one mapping at the syllable level, but Zhuyin is more accurate at the single character scale. E.g., the sounds for both ㄜ and ㄝ are written as an e in Pinyin and ㄨ and ㄩ are both always written as u. Worse still, the ㄩ sound is also sometimes represented as ü in Pinyin (when preceded by an l).

Pinyin requires more context (and rules) to distinguish which vowel sound is being represented.

Zhuyin isn't perfect either. It's definitely a better representation of single consonant or vowel sounds but it still has its own warts around diphthongs and triphthongs.

Pinyin is definitely a flawed phonetic system, but it's usable and widespread. That's it's real strength, along with being less alien to non-Chinese speakers.

[fiblye]

Zhuyin is also relatively new, at around 100 years old. Historically, the closest thing Chinese had to phonics was fanqie, which was a comparatively obtuse way of representing sounds with other hanzi.[1] Basically, any hint of a true phonetic system probably helped literacy rates during the 1900s.

[1] https://en.wikipedia.org/wiki/Fanqie

[jdietrich]

>Pinyin is definitely a flawed phonetic system, but it's usable and widespread. That's it's real strength

I don't disagree. There's nothing particularly special about pinyin, but it was revolutionary precisely because it rapidly became ubiquitous in Mandarin instruction.

I don't think I can reasonably address the issue of zhuyin vs tongyong vs hanyu without stepping into a political minefield, but I will say that there's a world of difference between achieving full literacy on a small and prosperous island versus achieving full literacy in a vast nation with substantial poverty and under-development.

[uranusjr]

I argue that China would still reach the current literacy rate had they continued to use Zhuyin. The switch from Zhuyin to Pinyin was entirely political, and the two are basically the same to L1 learners. The significance of Pinyin is that it drastically lowers the initial learning barrier for L2+ learners from Europe and Americas (which has trade-offs, but that’s another whole topic).

[SuperNinKenDo]

I think Pinyin has had a bigger impact than Zhuyin simply by its use in China and in the broader Huaren communities. That said, Zhuyin is also a recent innovation, and so the point about having some kind of phonemic/phonic/phonetic system of spelling as your basis for literacy probably still stands, at least with whatever strength it has, I think the other poster might be overstating it to some degree, although it's definitely a godsend to have in Chinese now, whether it be Hanyu Pinyin, Zhuyin, or whatever.

[tgb]

I think you're overstating the complexities of pinyin, even if Zhuyin is better. The context pinyin requires is extremely limited, being exactly only the rest of the syllable. It's simple if one just thinks of it as a set of initials/finals and not as letters. The whole thing can be learned in a day, so there's not much to be gained in an improved system.

[AlchemistCamp]

I regularly meet people who have been using it for months and still keep using the wrong vowels.

It's not insurmountable by any means. However, if Pinyin came in on a code review, I'd have a lot of comments on it :D

[starfallg]

Hong Kong has been teaching Cantonese Chinese without using a romanisation system to do so and its literacy rate never lagged.

Pinyin may help second language learners greatly, but was never the cause of the literacy spike. It's the rollout of the education system to villages that caused it. If they taught Chinese characters without Pinyin, they would most likely have decreased literacy rates to a similar extent also.

[SuperNinKenDo]

I agree that counter evidence to the claim, but one thing to keep in mind is that Cantonese is a local language, which many speak at home and with close friends and family from birth (or at least some closely related dialect), while Mandarin is often not. So if you concede that something like Pinyin might help L2 learners, it's reasonable to think it might be less necessary, but still helpful in an L1 setting, which Mandarin sits somewhere between, and Cantonese is well in to L1. It will be interesting to see if something like Jyutping takes off in the native Cantonese speaking community at some point or not. It's also a different case because spoken and written Cantonese are such different languages, although I don't want to overstate what effect, if any, that might have.

[krackers]

That's interesting, because people always say that furigana is a "crutch" when learning Japanese vocabulary, but it would seem the be the natural correspondence of what you mentioned (namely first learning spoken japanese with furigana and then later associating the words you know with the kanji). Maybe the larger number of homophones (with lack of tones to distinguish between them) makes this less effective?

[musicale]

Furigana seems like a great idea to me and extremely useful. I expect that native learners use it and benefit from it.

Also I expect the onyomi/Chinese readings of many kanji might fall into phonetic groups as they do in the hanzi. (Note also that kanji and hanzi are basically the same word with some phonetic shifts.) Apparently onyomi actually means "sound reading."

[livueta]

Yeah, 音読み - literally sound-reading.

There definitely is a phonetic element to it. Since most/many complex words (jukugo) you'll encounter are formed from the same ~2k kanji, you often end up quite literally sounding it out. As a simple-ish example, 音読み itself: 音 is read 'on', 'in' or 'son' in descending order of frequency. 読's usual onyomi is 'doku', but the み ('mi') at the end clues you in that this is a on-kun jukugo, where the second kanji uses the kunyomi reading because... nevermind. Anyway, kunyomi is... 'yomi', so, onyomi.

It's plenty complicated in practice with all kinds of stuff messing you up (multiple onyomi, nonstandard readings of all kinds, rendaku, etc.) but fundamentally you're operating phonetically, at least when you first encounter the written form of a word you probably have heard before.

On furigana: it's useful but also absolutely can be a crutch. Although you're generally going to want to acquire jukugo vocab from various sources and not just sound everything out after learning common kanji, that process of phonetically processing a compound word is a great workout for your fundamental kanji knowledge. On top of that, for whatever reason, I find it makes me feel like I've comprehended the text when I actually haven't, like the kid in the article with the dog story, though in this case it's lack of understanding through phonetic over, rather than under, emphasis. At least personally, it's easy to rip through the furigana reading and have my brain be like "yep, just turned all of those symbols into sounds, job done." That's probably part of why many resources, like https://www3.nhk.or.jp/news/easy/, allow you to quickly toggle them. Open a story and hit the 漢字の読み方を消す button.

[soraminazuki]

> Furigana seems like a great idea to me and extremely useful. I expect that native learners use it and benefit from it.

So true, especially when you encounter a proper noun that includes kanjis. It's almost impossible to guess how to read them, even for native speakers, because the usual rules surrounding how to read kanjis are largely ignored when it comes to proper nouns.

[knolax]

Pinyin is only phonetically regular if you don't actually speak Chinese, and therefore learn all your pronounciations from Pinyin. Even for speakers of Mandarin in Beijing it lacks a competent ability to indicate a merged 兒 at the end of a character. In fact many people forget Pinyin by the time they enter highschool.

Whatever success attributed to Pinyin has less to do with Pinyin and more to do with the social changes that coincided with it's introduction. Bopomofo, a more phonetically regular and easier to use system was used for decades before Pinyin and is still used in Taiwan today.

In addition, there are many adults alive today in their 60s and over who never learned any phonetic system and can read and write just fine.

People in Hong Kong also don't seem to learn any romanizatiom system for Cantonese pronounciations in school and can obviously read and write just fine.

[oefrha]

> In fact many people forget Pinyin by the time they enter highschool.

Since most Chinese people type Chinese using Pinyin-based input methods (not counting voice recognition and handwriting recognition as typing here, which are also pretty widely used these days), this claim is absolutely, undoubtedly, 100% false.

Edit: Somehow read “many” as “most”. Since “many” can mean anything from one thousand people out of 1.4 billion to all of them, it’s not possible to refute the claim. But the idea should be clear: forgetting Pinyin upon entering high school is not common.

[knolax]

You forget that a good number of professionals do not use phonetic input methods, and use component based ones like CangJie, Boshiamy, WuBi, Four Corners etc. Plus even phonetic input methods have "fuzzy modes". If you ever watch an experienced typer use a phonetic method you'll see that often they only enter the first letter of each character.

The fact that any number of people forget Pinyin at all demonstrates that it can't be compared to English phonics. You can't forget phonics, you can forget Pinyin.

[oefrha]

Sure, you can forget Pinyin, that doesn’t mean people do forget Pinyin, which you have provided no evidence for and doesn’t match observed reality. No convenience features in Pinyin-based input methods will help you if you actually forget the whole system. In fact, I personally know elderly people who were educated in 1930s and 40s who later learned Pinyin because it’s easier to type that way.

And yes, I’m aware of the non-phonetic input methods, and that professional typists could type faster in them. But the learning curve is way steeper (compared to basically none).

[tgb]

I'm learning Chinese and I think you've avoided the question. The crux of the matter is how to learn the thousands of character-to-pinyin/meanings mappings. It doesn't fit at all into the phonics description given in this article because you can't sound out a character. And the way I'm learning is straight memorization which this article makes sounds bad, but what's the alternative in Chinese?

And this article says that a second grade reader can learn a new word after being exposed to it just a couple of times, but I'm not having that level of success to new characters I read in my passages.

[gizmo686]

I think the idea is to use Pinyin to bootstrap knowledge of hanzi.

A) This would allow you to use written language to teach hanzi and

B) This would allow you to have children read text that includes words they don't know how to read, by providing those words in pinyin.

I'm not sure how it actually works in Chinese, but as a adult second language learner of Japanese, having a phonetic system I can use to transcribe the logoriphic system is quite helpful.

[yorwba]

> B) This would allow you to have children read text that includes words they don't know how to read, by providing those words in pinyin.

Indeed. Here's a page from a textbook for second-graders: https://i.imgur.com/g1UlUZP.jpg

It's a story teaching the meaning of the proverb 揠苗助长 (pulling sprouts to help them grow). Some of the characters are annotated with their pinyin above, e.g. 揠 yà or 筋疲 jīn pí. This text in particular is a bit weird in that it only uses 揠 yà in the title and replaces it with 拔 bá in the story, so the reader would have to infer that 揠 yà and 拔 bá are synonyms. I think 拔苗助长 is actually the more common way to write the proverb.

[tasogare]

> It doesn't fit at all into the phonics description given in this article because you can't sound out a character.

You are wrong, it absolutely does. 放方芳房肪舫防紡 are all pronounced fang and contains 方. This stems from the evolution of the script, where a character was used to write various words with the same or close pronunciation. Then those characters where extended with other parts to reduce polysemy. This is similar the determiner of other logographic scripts (hieroglyphic, cuneiform) except the determiner is embedded in the character themselves instead of being separate ones.

While this phenomenon is less reguglar than in an alphabetic script (not that English has a straight graphemes/phonemes mapping either...) because of the script evolution , it’s happening nonetheless and is quite useful. So it’s not true to say that you can’t guess the sound of a character, because in a lot of case it’s possible to get an approximation.

[fiblye]

Not sure why you're downvoted. You're absolutely right.

I'm only a year into learning Chinese, but I can often guess the sound of a new character just by recalling components of similar characters. The tone takes memorization, but the sound oftentimes has hints. Especially if it's a complex, uncommon character--almost always the sound derives from one of the common radicals it uses.

It's not as intuitive as, say, pinyin, but it's helpful and straight up taught by any decent Chinese textbook.

[tgb]

I do this guessing while trying to read an unknown charger but this article made me question the usefulness of that and is exactly why I'm unsure what the process in Chinese should be. Doing this looks more like the cueing process of picking a word with the right starting letter that matches the rest of the sentence and moving on. I certainly don't learn the word by seeing it a couple times in text.

[brownbat]

> this article says that a second grade reader can learn a new word after being exposed to it just a couple of times, but I'm not having that level of success to new characters I read in my passages.

This tracks with some self reports for the most advanced L2 language learners though, that your ability to learn new words itself improves while you are learning new words. Maybe just requires a ton of exposure.

MattVsJapan talks about drilling vocab early on, then once he hit a certain level, he could often just look up a new word once and know it almost indefinitely, more like with English.

2nd grade is, what, 8 years of immersion? Exposure to something like 100-250,000 sentences, with practice or exposure for all available hours every day? It's a lot to catch up to.

Even so, with 2nd graders, we're not talking like "perspicacity" or "abrogate." Some words and some concepts will be harder than others.

[throw333542]

Wait, did I misread the article?

I thought it said that memorization is good?

> Reading scientists have known for decades that the hallmark of being a skilled reader is the ability to instantly and accurately recognize words.33 If you're a skilled reader, your brain has gotten so good at reading words that you process the word "chair" faster than you process a picture of a chair.34 You know tens of thousands of words instantly, on sight. How did you learn to do that?

> It happens through a process called "orthographic mapping."35 This occurs when you pay attention to the details of a written word and link the word's pronunciation and meaning with its sequence of letters.

I thought this means people recognize the entire word, rather than the letters of the word. Just like how you recognize a Chinese character.

Then you learn to map the Chinese character with the pronunciation.

[tgb]

The article claims that bad readers memorize a handful of high frequency words and rely on those. I read the parts you quote as saying that those words are not memorized as a chunk and I don't see how a similar process can be done on Chinese characters. I kind of doubt that they are read by decomposing into radicals.

[solinent]

I think if you can learn one per hour every day you'll be doing pretty well. If you're in school you can easily learn a few every week.

It's also not rote memorization, rote memorization is mapping text->text not text->meaning.

[cwkoss]

Maybe we should be teaching kids to read the phonetic alphabet before we teach them the english alphabet.

[_nalply]

Perhaps you should consider that characters are not images but consist of distinct groups of strokes. When I learned about 20 words suddenly I realized that some of them have sub-words embedded into them.

So it's a bit more complicated.

[Izkata]

My favorite example of this, because it looks kinda silly and yet makes perfect sense (and appears to be the same between Chinese and Japanese, where I first ran across it):

木 - tree

森 - forest

[canjobear]

林 is also forest

[ithkuil]

One is natural the other one is man made (similar to many other languages discriminating between woods and forest)

[est]

Native speaker here. Other comments are not wrong but not complete.

There are several phases of Chinese (the text) learning

1. Mapping your daily voices to phonics and to the characters

2. Expand your vocabulary with phonic tools like PinYin

And the most important part:

3. Learn new words and meanings without any involvements of phonics at all.

Many native speakers can comprehend the concept of a word at young age and when they have grown up and put those words into speeches they found they have the wrong pronunciation or word order. LOL. Chinese words are ideograms and can be linked to "pictures" without proper pronunciation. Also Chinese is analytic language so people can literally guess its meaning from contexts.

[musicale]

Phonetic alphabets (e.g. pinyin) are used by learners as well as in computer input methods.

Moreover, some 80% of Chinese characters are semantic-phonetic hybrids (though the pronunciation of characters that share a phonetic element may not be identical in current Mandarin, they will often resemble each other closely, perhaps with a tone or vowel shift.) Learning characters in phonetic groups helps to anchor their pronunciation and phonetic elements can help when you encounter an unfamiliar character in context.

[bart_spoon]

Its funny you mention that because I instantly thought of Japanese. I spent a few years there and learned to speak fairly well, but never learned to read much beyond the phonetically based character systems (hiragana and katakana), but never more than a hundred or so of the commonly used kanji. And while kanji are technically associated with specific sounds, there are tens of thousands of them, and they all can have different sounds in different situations. Since returning from Japan, I've been trying to do some studying of kanji with mixed success, and I find that it virtually requires the type of thinking here: trying to determine what the word would be based on context, as well as connecting the symbols in the kanji (which often do have meaning) words.

[prox]

Sanskrit is interesting because all symbols are direct phonetic representations. If you hear a sound, you know how to write it. This is why it was able to be super effective at keeping the texts intact through centuries of passing down texts orally.

[mkl]

Sanskrit has the big advantage for that purpose of being "dead" (not taught as a first language), so its existence is academic and pronunciation is relatively fixed through time; Latin would be similar, and its alphabet, this one we're typing, is a good phonetic representation. The pronunciation of "living" languages changes quite a bit.

Most languages with phonetic writing systems (alphabet, abugida, syllabary, etc., but not pure abjad [1]) have direct phonetic representations. English, French, and a few other European languages are unusual.

[1] https://en.wikipedia.org/wiki/Abjad

[prox]

Besides being dead as a reason for it being fixed, I think it might also reflect the high reliance on the oral transmission of texts (which your excellent examples I hadn’t heard about emphasize)

The name for learning (sacred) texts in Sanskrit is “shruti” , that which is heard. Sanskrit grammar did change over time, at least until Panini wrote down the rules around the 6th century bc.

[prox]

Also, something I just remembered, in vedic rituals there is actually a priest present who checks if the ritual and the sanskrit recitation is correct. This is also a reason we still have access to vedic sanskrit and we can be confident in its transmission accuracy afaik.

[tomjohnneill]

Something I found fascinating about learning Chinese while living in China was knowing exactly what a character meant (because it was written on signs etc.), but having absolutely no idea how it was pronounced.