Explaining the Directions 東西南北, part 1: 東西

Long time no see! As you’ve probably noticed, we haven’t posted anything in a while (several months, actually). Well, that’s because there are a lot exciting things coming our way. We should have some really exciting announcements coming up in the next few weeks.

In this post, I’ll explain the origins of 東 and 西. Part 2 will discuss the origins of 南 & 北.

東 【トウ】 “east”

Overview:

Since the following explanation of 東 is rather involved, I’ll start here with a simple overview. Three characters with similar forms, development and meanings: 東 【トウ】, 束 【ソク】 and 橐 【タク】. Their pronunciations were also much more similar thousands of years ago when these characters were created than they are in modern Japanese. Each of these characters were originally pictures of bags tied at both ends. 東 【トウ】 and 束 【ソク】 were even used interchangeably in bronze inscriptions, while 橐 【タク】 is basically 束 with 石 【セキ、シャク、コク】 added as sound component (though that isn’t obvious from modern Japanese). 東 【トウ】 ended up meaning “east” via sound-loan.

The details:

Well, we can’t really talk about 東 without first talking about 束 【ソク】 “bind; bundle” and 橐 【タク】 “a bag.” What does “east” have to do with bags, binding and bundles? We’re about to find out!

According to Chou Fa-Kao [周法高 Zhōu Fǎgāo], 束 is “a picture of a bag that is tied at both ends.”1 Its original meaning was “bind.” So, a picture of a bag that is bound at both ends to represent the meaning bind. Pretty straightforward, right?

Note that in the diagram above, the full lines represent a change in time period, while the broken line represents forms that were contemporary to one another. It was very common for a given character to have several different forms at any given time, similar to how English words would have several different spellings before spelling was standardized.

It is easy to see that the Bronze Inscription [金文 【キンブン】] form (1c) and the Small Seal script [小篆 【ショウテン】] form (1d) are the direct descendants of the Oracle Bone [甲骨文 【コウコツブン】] form (1a), while (1b) is an alternate form. There are actually quite a few other alternate forms, but here, I’m trying to keep it simple and just show the main branch. Note that though 束 looks somewhat like a combination of 木 and 中 in modern Japanese, in reality, it’s not related to either 木 or 中.

In Outlier terms, we say that the similarity between 束 and 木 + 中 is a surface structure similarity. That is to say, the similarity is not related to the meaning and sound of the components involved, rather it’s due basically to a fluke of history. 束 probably ended up looking like 木 either 1) by the fact that the three lines on the top part of characters tended to flatten out, while those on the bottom didn’t, or 2) by way of analogy with 木.

Even a quick glance at the forms (2a) - (2e) shows that they are very similar to (1a) - (1d) above, except that (2a) - (2e) have an extra line or lines through the main body. According to Lín Yìguāng [林義光], the forms for 束 and 東 where interchangeably used in Bronze Inscriptions. As such, he proposed that the two forms are actually merely variants of the same character.

So, here, we have yet another character that has a similar origin, 橐 【タク】 “bag.” The Shang dynasty Oracle Bone forms (3a) and (3b) look very similar to the Oracle Bone forms for 束 and 東. We can see in (3d) that by the time of the Qin dynasty, 石 【セキ、シャク、コク】 “rock” had been added to the character. Though not obvious from a modern perspective, 石 was added as a sound component. Looking at the Japanese pronunciation, 石 【セキ、シャク、コク】 acting as sound component for 橐 【タク】 seems ludicrous, but if we look at Cantonese pronunciations, sehk and tok respectively, we see that they share a common ending, -k2. Japanese simply adds a vowel because Japanese phonology doesn't allow for syllables ending with -k.

The main vowels in both Cantonese and Japanese are different, but it’s helpful to point out that the sounds of all languages change over time—some slower, some more quickly. During Chaucer’s time (1343 to 1400), the words food, good and blood all rhymed (sounding like goad). Then, during Shakespeare’s time (1564 to 1616), they still rhymed, but by that time they all rhymed with how we now say food 3. Nowadays, they have diverged from one another and all sound different.

If we look at the Old Chinese (OC) reconstructions for these words, 橐 *tak and 石 *dak4, we can see that the initials (the t- and the d-) are pronounced in the same part of the mouth, and therefore are very closely related sounds, as are the main vowels and -k endings. As such, 石 made a rather suitable sound component for 橐. Don’t let the * symbol scare you! It just means that these sounds are reconstructed and have not been attested directly.

Interestingly, there is yet another character with similar origins, though to keep things simple, I’ll leave it out of the analysis: 柬 【カン、ケン】 “card, letter; choice” (in ancient times, it meant the same thing as 束, but had a different pronunciation. In other words, the characters 束 and 柬 represented two different spoken words that had the same or very similar meaning). In Japanese, this component is usually simplified to 東 (for example, in kanji like 練 and 錬, both pronounced 【レン】).

Looking at the OC reconstructions for 東 *tong and 束 *s-tok5, we can see that they both share the same main vowel “o” and an initial “t-.” Though the endings “-ng” and “-k” aren’t exactly the same, they are pronounced at the same place in the mouth (which indicates that they are closely related sounds). So what is up with the “s” at the beginning of 束? The dash in front of the “s” indicates that it is a prefix.

This is similar to English, where we have a series of related words, but differ phonologically because of the addition of prefixes and suffixes. Take the root word “get”: forget, beget, got, gotten, begotten. This series of words have related meanings which are expressed by differing in grammatical affixes (i.e., things which you can attach to root words that express meanings) and changing of the main vowel.

Just like the prefixes for- and be- in the word family for the root word get, the OC *s- prefix is attached to the root word tok. Sharing the same root word (though not necessarily the same affixes) was the main requirement for characters sharing the same sound component. This is one of the reasons we see a significant amount of sound variation in Chinese character sound series.

So why does 東 mean “east” if it’s a picture of a bag tied at both ends? Well, because the word for “east” sounded similar to the word that the character 東 represented, i.e., “bag,” so it was borrowed by way of sound loan to write “east.”

In our dictionary, 東 is explained like this:

西 【セイ、サイ】 “west”

The explanation for 西 is far simpler than the one for 東. There’s an agreement among scholars that 西 is a picture of a bird’s nest. According to the Setsumon Kaiji 説文解字, the first character dictionary that tries to explain character forms (published in AD 121), the connection to “west” is because birds return to their nests at sundown and the sun sets in the west6. This is somewhat confirmed by the fact that the character 棲 【セイ】 “resting place for birds; nest” was also often used to write “west.” However, it’s also possible that it is simply a sound loan (i.e., the word for “west” sounds similar to the word meaning “bird’s nest”).

Be sure to check back for the forthcoming part 2 of this series, where we take a look at the origins of 南 & 北. We also have several very exciting announcements to make soon! So stay tuned!


  1. I’m consulting Chi Hsiu-Sheng [季旭昇 Jì Xùshēng]’s (2014)《說文新證》,page 512。

  2. Many southern Chinese dialects, such as Cantonese, Hakka and Southern Min retain the entering tone [入声 【ニッショウ】] endings: -k, -p, -t and maintain three nasal endings: -ng, -n and -m. In Japanese, these entering tone endings are generally retained in onyomi readings ending with -ku, -ki, -tsu, -chi, and so on.

  3. Source: http://grammar.about.com/od/fh/g/GreatVowelShift.htm

  4. These are simplified versions of Baxter-Sagart Old Chinese reconstructions, version 1.1:

    read as: {char} *{OC} {(OC rhyme}) > ({fǎnqìe spelling}切) {MC}
    橐 *tʰˤak (鐸部) > (他各切) thak
    石 *dAk (鐸部) > (常隻切) dzyek

    I’m using a simplified version of the reconstructions, not because they are problematic, but rather to point out to people not trained in historical linguistics the sound similarities between the two words. Explaining the basics of OC reconstruction is a can of worms I’d rather not open just yet.

  5. As above, these are simplifications of the Baxter-Sagart reconstructions, version 1.1:

    東 *tˤoŋ (東部) < (德紅切) tuwng

    束 *s-tʰok (屋部) < (書玉切) syowk

  6. 《說文》: 「西,鳥在巢上。象形。日在西方而鳥棲,故因以爲東西之西。凡西之屬皆从西。 」


Leave a comment

Please note, comments must be approved before they are published