Some time ago, Instagram user jumppingjack posted the above image of a note she left to her mum. She said that her brother secretly added extra strokes to the characters in the note. The result is interesting though: even though extra strokes were added, the note is still readable to most competent Chinese speakers. This phenomenon is very similar to one not too long ago in English, coined “Typoglycemia,” a portmanteau of “typo” and “glycemia” and a pun on “hypoglycaemia,” where as long as the first and the last letter of the word is preserved, the middle can be scrambled and the words are still understandable.
This is an interesting case in what I call persistence of comprehension, where comprehension of words persists despite efforts to thwart it.
Unlike English, which uses the alphabetic system where each letter is a phoneme, or Japanese, a syllabary system where each character is a mora, Chinese uses a logographic system, using “pictures,” or logographs to represent words. So unlike the other two systems where there are things to scramble, it is hard to “scramble” a picture, and scrambling a picture is no different from adding or subtracting strokes from a character, which is what jumppingjack‘s brother did.
Before I go further, let me type out what the note intends to say:
谢谢 🙂 （我明天应该有吃午餐）
Mum, for tomorrow’s lunch because
there aren’t enough people, they have
changed this activity to next Monday.
(But next Saturday afternoon my
company has a lunch event)
Thanks 🙂 (I should be eating lunch tomorrow)
So how does persistence of comprehension occur in Chinese? I shall illustrate some of the characters that are easily understood despite the scrambling and the ones that threw me off (and my friends) the most. (Also, note that the person mis-wrote the character for 期 where he switched the 月 and 其 around, not of her brother’s doing. But the brother added an extra radical as well)
The image above sorts some of the words in the note in order of persistence of comprehensibility from top to bottom, with top being easiest to understand despite scrambling and the bottom being the hardest. The scrambled word is on the left and the proper word is on the right. Note that the bottom four scrambled words are all actual Chinese words, which I will talk about shortly.
The scrambling of the 的 character is one of the easiest to understand, because despite the additional stroke, it still mostly resembles its original character, and does not resemble any other words in the language. The added stroke is a not a radical, a graphical component of a word that is often semantic, unlike the scrambling of the character 明 (tomorrow). Similarly for 因, the added stroke turns the 大 in the 因 into a 太, but on the overall the word is not a real word and mostly resembles its original.
Now we look at the addition of a stroke in 明, turning the 日 (sun) radical, usually used for weather-related words, into a 目 (eye) radical, usually used for vision-related words. The resultant scrambled word is still not a word, but the morphing of a semantically-relevant radical into another makes one pause when reading the sentence. Also, the addition of a stroke to the 月 (moon) component turns it into a 用 (use) character, making comprehension even more difficult.
One step after the 明 character is the 们 character, where not a stroke but an entire 中 (middle) word has been inserted in the middle (haha) of 们. Some of my friends disagree that it is harder than the scrambling of 明, and I’m inclined to agree, and I’d put it as a toss-up between the two. However, I feel that the insertion of an entire word as opposed to a stroke or radical morphs the word enough to the point that it becomes alien enough not to even resemble its original, but does not resemble any other word in Chinese.
Lastly, the last four words, 公，午，伞，and 下, have strokes and/or word components added to them, that they actually resemble other words in the language, 翁 (old man), 牛 (cow), 伞 (umbrella), and 卡 (card). With such resemblance to real words, little wonder people have difficult understanding the words as they read them.
How is it that we are able to understand the note with little difficulty?
In the English “Typoglycemia,” it has been suggested that we identify words not solely by letter position in a word, but by context, shape of the word, and position of word in the sentence. I’m going as far to suggest that in English seeing the individual letters of a scrambled word draws upon our stored memory of the word, further aiding comprehension of a scrambled word. Compare:
- Adcnirocg to rrasceeh at a ptaruilacr ureitnvisy
- Aoincdrg to rcseerh at a plaaicutr uesvtniiy
- Aroindg to rearech at a pluiraacr utrisveiy
Example 1 is classic “typoglycemia” where persistence of comprehension is strong, example 2 removes one non-essential letter from each word, and persistence of comprehension is still relative strong. Example 3 removes what I consider an essential component to the memory of the word, which are usually consonants and not vowels. Take this example:
- I cnt blve u dd tt!
In English, vowels can be removed quite easily and the comprehension of the word is still possible. This suggests that consonants play a slightly more important part in the reading of words. In that aspect, comprehension of written English has some similarity to comprehension of written Arabic or Hebrew, where typically vowels are not included in the writing (in the way English does anyway). Thus, it is harder to understand “aroindg” as “according” because
- An essential component has been removed (3 syllables, essential components in bold: a-cc-r-d–ng). This might be so because consonant representations are tied up with its phonetic properties. This is why “cc” or has to be removed as opposed to just “c” from “according” for comprehension to fail, because “cc” in the word correlates to the /k/ sound in /əˈkɔː(ɹ)dɪŋ/; even with just one “c” or it is sufficient to clue us in that there might be a /k/ or sound in the scrambled word.
- It is the first in the sequence of essential components, suggesting that perhaps we process essential components sequentially in our head. It could be that when we see the word “according,” we could be drawing upon the idea that “according” has the components “a-cc-d-ng” in that order. Hence removing the first component “cc” impedes comprehension as it cannot give the subsequent components context of what the word might be (compare understanding: “aroindg” (“cc” removed) with “aocrcdg” (“in” removed)).
How does this relate to Chinese? If we can say that we draw upon essential sequences of components in the comprehension of written English, perhaps there is an equivalent of that in the comprehension of Chinese. I believe that in reading Chinese, there is a stored visual memory of what the character looks like in general, and also an idea of what strokes the character should contain (“legal strokes”), and what it should not (“illegal strokes”).
First, we address whether modifying a Chinese character sets off alarm bells to the reader. Adding legal strokes to scrambled characters should stand out less to the reader, causing him to accept the character as a real word visually. We look at the following example where this is demonstrated:
In the note, a floating shuzhe (vertical-bend) stroke is added to the 够 character, and in the Chinese language there is no such occurrence of a floating shuzhe; they are always attached to other strokes, such as in 喝 (with some exceptions, like 断, which may or may not be attached). Being visually alerted that there is something wrong with the character, we immediately visually discount the scrambled 够, and are able to extract the original word. In the example of 他, the pie (leftward-slant) stroke is added on top of the 亻radical, creating a 彳(step) radical, which exists. Thus when reading the scrambled word of 他, it does not jump out at the reader visually as the shuzhe stroke in 够 does, and we are likely to gloss over it and accept it as it appears to us and are less likely to question whether the character is out of place contextually or not.
Next, adding a legal stroke to scrambled words causes more confusion when the stroke turns the original word into a semantically different word. There is extra confusion when the meaning of the new word does not fit in the context of the sentence, especially when the word has been accepted as it is, as explained in the previous paragraph. These can be seen in the following examples:
If my premises are right, in example 1, readers should be able to identify the error most easily and yet still read the sentence in its original context. In example 2, they should gloss over the wrong character, and since it still resembles very much like the original, is not a new or any word at all, persistence of comprehension should still be strong. In example 3, this is where comprehension begins to be thwarted, where 他能够卡去吃午餐 (He is able to card go eat lunch) and 他能够下去吃牛餐 (He is able to go down and eat cow meal) don’t make any sense as the scrambled words have both legal strokes and are real words, and the meaning of the scrambled words are contextually out of place in the sentence.
What I have coined “the persistence of comprehension” is a seemingly little-researched area in English, much less Chinese. I offered the following reasons explaining the persistence of comprehension in English “typoglycemia,” where through the combination of context, length of word, letter position, shape of word, word position in a sentence, and (what I have demonstrated with examples) identifying the letters, which draw upon phonetic representations of the word in our head, we are able to read English.
In the more interesting case of Chinese, which is logographic, I posited that there are legal and illegal strokes which can be added to a character. Legal strokes are less likely to be noticed than illegal ones. If the scrambled word is a real actual word, the effect of having legal strokes masks the fact that the word has been scrambled, and when we read it, the sentence doesn’t make sense because we do not suspect a character has been tampered with.
All in all, more extensive research must be done, than what this blog can provide. I don’t know if I will be able to do so, but if anyone wants to hear my notes on this topic, feel free to reach out to me at ws672[at]nyu[dot]edu.
Note: In the original version of this post, I wrote that jumppingjack was male, when she is female. Corrections have been made.