Cold Character Reading is very simple in theory: get some oral language into the students’ heads, then have them directly jump to reading that same language (but a story or text they have never seen or heard before, not something they already know the meaning of) using a non-phonetic writing system like Chinese characters, or a system that is not obvious to them, like Russian or Arabic script.

But there are some conditions that must be met for CCR to succeed. Many teachers are trying it, but some are missing out on some of these conditions, and like any method, if some of the necessary conditions are missing, things don’t go as well as they might. A few of the problems that have been coming up lately:

Texts are repetitive, comprehensible, and short.

Two out of three. Cold Character Reading (CCR) texts are repetitive, and they are 100% comprehensible in terms of language (at least at the beginning; but even later on, there really should not be anything added in writing that is not recognizable in speech, unless it is a “side note” such as the snarky comments included in illustrations that are intended to entertain and are clearly set apart so as not to cause anxiety or reduce comprehension of the text).

But these readings must not be short. The reason is that CCR is based on flooding the eyes with visual input that corresponds to oral input the brain has already grasped (can comprehend easily), if not acquired (can correctly output easily).

A too-short CCR text is in invitation to failure. The length of these texts is a big difference between using CCR and using traditional textbook readings (how many of the first “reading passages” in a Chinese textbook for zero-Chinese beginners is 400 characters long?) and it is one of the main principles that makes the technique work. Students need to see the same characters over and over and over to get used to accepting that information visually rather than in speech.

“But I can just use a short reading and have my students read it several times.”

How many times? How will you keep interest high? How will you avoid the brain remembering the content, rather than really reading (getting new and largely unexpected information from the written form)?

It’s not difficult to make a CCR text long, because the people writing them should already know how to circle in their sleep, unconsciously and endlessly if they want to. I’ll say it again: CCR is visual circling. Yes, a written story does not usually involve questions (but it is a wonderful idea to add some, because how else will your students learn to read the written forms of the question words?) And for all the mode-istas out there, if there are questions, there is likely to be “interpersonal” stuff going on, which adds a layer of buzzword acceptability to the process. They read (interpretive) an account of two people communicating with each other (interpersonal). So that’s either “compound interpersonal interpretive language” or just plain stalking. You pick. 😉

CCR texts tend to use lots of related statements or details about something to get those words out there more. He goes to France. He goes to France on Monday. He goes to France with his friend, because his friend also wants to buy a car. Good style in the target language? No, it’s rather juvenile. But that’s what novice students are: language babies. They do not have to read authentic texts right away. We are training them to read authentic texts and to simultaneously be able to process written language at the paragraph level, something that traditional ACTFL Novice-type reading tasks totally fail to do. In effect, we are jumping our readers to a pseudo-Intermediate level of reading, because with the exception of including all possible topics, they are handling that level of language (connected text, transition words, etc.)

The teacher leads the reading, and students join in when they can. Sometimes they won’t join in at all, so the teacher just leads, or they only join in at the very end.

If your students are not reading confidently with you by the fourth or fifth sentence you read in a well-constructed CCR text, you should not continue reading. That situation requires you to really, really think about how much and how good the oral input that is supporting this reading task was. If the answer is “Hmmm…well, maybe not as good as I’d like” or “Probably not enough repetition for them to immediately understand the language”, then the right thing to do is to find an excuse, stop reading, and give more oral input. Trying to do CCR based on incomplete immediate comprehension of the oral language underpinning it will fail, and it will show kids that reading is hard. That is absolutely not what we want.

This is why having solid basic TPRS skills is so crucially important. I sometimes joke with new Chinese TPRS teachers and call the method the “beat-it-to-death method”. From the teacher’s perspective, this should pretty much be the case. In a situation involving artificial input (as opposed to you-are-three-and-have-caretakers-and-an-immersive-environment-and-nothing-better-to-do situation of natural input) we have to achieve high repetitions of language in a dense manner (dense comprehensible input).

CCR relies entirely on “the Voice”. The Voice is that tiny, very limited but very solid instinctual knowledge of what is correct in the target language. It requires a lot of repetition to build. If output comes when the water spills over the top of a bucket naturally because it has been filled, then the “Voice” is that sound, rising in pitch, that you hear when you’re filling your Thermos up and it gets pretty close to the top but it isn’t spilling yet. You know it’s really close to overflowing, but it isn’t quite there yet. But it’s able to make some noise anyway.

It is certainly possible to read Chinese characters without speaking Chinese or even being able to understand spoken Chinese very well. But that involves feats of straight memorization, and usually a lot of decoding. It’s possible to read character texts without this fluency because reading allows time in most cases (with the exception of real-world reading tasks that involve immediately getting the meaning, like reading a highway exit sign). But this is not the kind of reading skill we want our students to focus on.

Since we focus on acquisition, and literacy is not really an acquired skill or an integral part of natural language, we need to both maximize acquisition of the actual language and also achieve maximum literacy skills. So relying on a solidly acquired or at least easily comprehended mass of oral language (and I do mean easily comprehended — flashlight-in-the-eyes-in-the-middle-of-the-night easily comprehended, not yeah-I-can-probably-figure-this-out-then-I’ll-look-at-my-teacher-for-confirmation-oops-I’ll-modify-that-now comprehended) is the best way to simultaneously maximize language and literacy for languages where literacy is a substantially different task than simply recognizing frozen language sounds on paper.

If your students are not giving you a strong choral reading on any sentence you are reading, you should not continue reading. But in this case, I’m talking about the class not sounding “unified” on a single sentence, not the class being consistently unable to read in anything approaching the voices of more than one or two students. We first rule out issues with the underlying language (has there been enough and good enogh oral input?) and then address issues of visual recognition (have the students seen these characters enough yet?) Whenever the response is not strong, that’s an indication that the visual aspect needs more practice. It will happen at the beginning. It will happen when they encounter a new character, even if they successfully guess it pretty quickly based on the Voice. But like everything in TPRS, we are not in a hurry. Unless the reading sounds strong and confident, we stop, compliment the students, maybe do a quick pop-up or ask a question, and then loop right back to tackle that sentence again. If you need to use an excuse, do so (“One more time, and this time let’s have Mickey Mouse read it out loud”) but get them to go through this one sentence again.

What do you do with the “outlyers” — the kids who don’t seem able to read with everyone else?

You give them extra help, yes. But you need to figure out what kind of extra help they need, first. And just as in the above with the whole class, the first order of business is to figure out what the problem really is.

Can the student understand that same language in speech, easily and immediately? If so, the problem is visual recognition of the written forms of that speech. If the student cannot comprehend the same language in speech, the issue is not the written forms (alone), it’s the lack of a firm language base. Some possibilities:

  1. The student has not yet heard the language enough to immediately understand, so even if he can decode individual characters, he loses the meaning of the whole phrase or sentence as he continues along it. (This is what we see with readers taught traditionally. They can bump through reading out loud but can’t make meaning.) Solution: more oral input.
  2. The student has not yet had enough opportunities to recognize that word in unpredictable contexts. Solution: more visual input.
  3. The text the student is being asked to read is too difficult. There are too many words that the student would know in speech but has not encountered previously in written form. This has to do with comprehensibility in a slightly different way than we usually think about it.  Solution: fix the text or use a different one so it is comprehensible.

Comprehensibility of the text is really important. Really, really important. In CCR we are harnessing the Voice to provide massive prediction of what written forms mean. The Voice is also what confirms that the proposed meaning makes sense. If you listen to CCR-taught kids read you will hear some of them actually going through this out loud, especially when they first propose a wrong reading for a word and then stop, look thoughtful, and correct themselves. Quite often they will comment in the native language, “No, that’s not [word], it can’t be…um…[other word]…yeah.”

There are examples of what texts look like at 98% comprehensible, 95% comprehensible and 80% comprehensible here if you read Chinese, and here if you read English. Texts for CCR have to exceed this standard. They need to be 100% comprehensible at the beginning. Remember that there are two ways to make language comprehensible in the TPRS classroom, and reading is no exception: either they have already “gotten it”, or it can be made comprehensible by pausing and pointing.

There shouldn’t be any Pinyin [spelled-out sounds of Chinese substituted for characters] in a CCR text. If the language (oral language) is not comprehensible or cannot be made comprehensible, it’s out of bounds for reading and cannot be in that text.

So some teacher will ask, “How is it that the word ‘because’ is in the very first CCR text the students read after just two or three hours of oral input? Surely they haven’t acquired ‘because’ in that short an amount of oral input?” No, they haven’t acquired it. But they should have gotten enough repetitions of it that it is easy to remind them (which I would do simply by pointing to my logical connectors signs on the wall). In this, reading serves sign-based diffuse input (no one should be circling “because” as an item, because it IS so frequent, or can easily be made so frequent, that it will be input enough naturally. You get it as a “freebie” just by pausing and pointing every time it comes up. This is distinct to the view supporting diffuse input and no purposeful repetition for everything, because function words and logical connectors are so much more semantically-frequent than content words in general).

The Pinyin is right there on the logical connector sign or where you’ve written that word up — just as you would for oral input — on the board. It’s not in the text, because Pinyin doesn’t belong in a written Chinese text. It’s accessible, just as we make the meaning accessible during oral input to students who need to “check in” on it by looking. But we would never say “Target-language target-language because target-language target-language” during oral input, and we shouldn’t present that in writing, either.

CCR texts should use short, simple sentences.

If that’s all you want your kids to be able to handle, sure. But CCR texts should mirror the complexity of language the students can handle in speech. And we should be using compound sentences all the time with our students during oral input. It’s perhaps just a little unnatural to use “although” all the time — but we want them to pick up the logical connectors, because those are the words that dictate the relationships between ideas, regardless of what the topic might be. They are the roadmaps that support comprehension of unfamiliar language because they help to link ideas to known information and, well, logic. So no: bring on the long sentences, the long modifiers, all of that. It’s easier to deal with these things in writing than in pure speech. That having been said, the comprehensibility rule still stands: if the students can’t comprehend the same language when they hear it, they shouldn’t be asked to read it.

CCR is really, really simple. Deceptively so. But if you lose sight of the reasons it works, it’s easy to miss out on some of the conditions that are needed for it to succeed: 99.5-100% comprehensibility, ridiculously plentiful opportunities to recognize the same words in different unexpected contexts, and most importantly of all, a solid Voice underneath it all. With those elements, CCR achieves impressive results and shows high levels of transfer to other tasks, such as recognizing words individually out of context (Riggs, 2016).