JimmySeal > I know you ain't dissing my wife regarding her wisdom, or lack of, but Pauline refined the so called rule when she wrote,

"That's really helpful. I could never sort it out before. I looked up the stroke order for some kanji. The rule seems to apply when the lower primitive touches/starts at the vertical line or the combination "by ones side"+"elbow" ("by my side" written horizontal-vertical) is an exception (the only I found)."

The stroke order of 雄 (おす) frame 743 doesn't contradict the combined wisdom of my wife and Pauline. Is it a woman thing? Or in this case a masculine one?

synewave Wrote:p.s. How do you quote what others have said so that you get those lovely big speech marks?

You can either click the quote link on anybody's post or if you've gotta do it manually, just follow the instructions here.


I hadn't thoroughly read Pauline's post. I see now. I'll keep my eyes peeled for any other kanji that use that primitive.

JimmySeal Wrote:

synewave Wrote:p.s. How do you quote what others have said so that you get those lovely big speech marks?

You can either click the quote link on anybody's post or if you've gotta do it manually, just follow the instructions here.

Thanks for that!

CharleyGarrett Wrote:Lots of kanji dictionaries can show the order, but not the direction.

I have also noticed this. Annoying isn't it? I think stroke order diagrams of any kind have to take this into account. Canon word tanks have good animations. Also, if you look at the strokes carefully then one end is thicker than the other, usually giving you a clue as to the direction.

With the 毛布 thing. If you want to learn a kanji then there's nothing like connecting it with something concrete. You won't forget it now.

wrightak Wrote:With the 毛布 thing. If you want to learn a kanji then there's nothing like connecting it with something concrete. You won't forget it now.

How true that is. Last winter, my friend who was working on RTK1 asked me "How do you say lamp oil?" and I knew he had 灯油 in his head because it can be seen on all the fuel stations. I told him "touyu" and he said, "I'm not sure why, but now I'll never forget that." It was because rather than connecting random Japanese syllables to an English word, he was connecting them to kanji that he knew and those to pre-existing spots in his memory. So all in one shot he learned two on-yomi and one compound, thanks to this book and he'll surely not forget any of them easily.

Still, it's important not to get wrapped up in readings and compounds before finishing RTK1. If you must supplement RTK1, there's stuff you can learn without kanji.

If you can find it, this book gives the official explanation for writing Kanji as issued by The Ministry of Education in the Appendix - Section H:

The Study of Kanji
A handbook of Japanese Characters
by Michael Pye
Hokuseido Press

ISBN 0-89346-232-2

synewave Wrote:My wife's explanation is that for 左 we write the horizontal stroke first and vertical second because the following stroke is horizontal. But for 右 it's vertical first followed by horizontal because the next stroke is vertical.

Although this thread's a little old, I'd like to add my voice to those thanking synewave's wife for this great explanation.

I offer one slight tweak that made it more mnemonic for me, in case anyone else might benefit. In practice, when I would try to mentally apply the rule as stated above, I tended to get confused in planning the strokes - I was trying to hold 3 alternating strokes in my head at once, to figure out how to start.

My spin on this rule is:

"Begin the first primitive the same way you're going to begin the second primitive".

In 左, the "craft" primitive starts with a horizontal stroke, so that's how I want to begin the "by my side" primitive. Same idea, but at least for me, I can apply the rule more easily on-the-fly if I think of it this way.

I noticed that 石 doesn't follow this rule, and I assume it's because the top-to-bottom rule supersedes it.

There is a rule for the cliff radical, and dwhitman has been the closest so far: strokes should alternate directions with the enclosed kanji's first stroke.

Go ahead and pretend to hold a brush. Swing it back and forth as if you were writing the kanji below. It's more natural to alternate directions than to do the same twice in a row. dwhitman's summary,

Quote:"Begin the first primitive the same way you're going to begin the second primitive"

may be the easier way to remember it though.

Horizontal first: 左　友
Vertical first: 右 布 有

Kanji were standardized to almost always follow the same rules. Top to bottom, left to right, alternating strokes, etc. It makes brushwork easier since it naturally preps you for the next stroke. The best example is in the teepee primitive (starting at 1703, 登). Strokes 3, 4, 5 alternate; the two little strokes are not drawn first before the long stroke. Try it out. Not only does it feel more natural to alternate, but it sets you up to then continue the enclosed primitives by pulling you back towards the center of the kanji. Now, quiz some Japanese on it. Many will get it wrong!

CharleyGarrett Wrote:I have Kodansha's Kanji Learner's Dictionary, but y'know there's this thing you mentioned about the direction of the stroke. Sure, it's first or second, but which end of the stroke is the beginning, and which end of the stroke is the ending? Lots of kanji dictionaries can show the order, but not the direction. I've found a few animated stroke order diagrams online, but they're a bit clunky. What do you use to know that sort of thing? For example, the first and third strokes of "compare" 比 are written right to left.

No, only the third stroke of 比 is right-to-left; the first is left to right. Stroke order, and especially stroke direction, is one area where Heisig gets a little negligent sometimes. He never mentions that the difference between 匕 (spoon) and 七 (seven / diced), apart from a small difference in the angle of the horizontal stroke and how far it extends past the vertical, is the direction of that first stroke. Then he compounds the problem by glossing 比 as "two spoons," when it's really seven+spoon.
As for the stroke order of the "by one's side" primative, I don't recall whether Heisig's stories use this imagery consistently, but for the ones where the horizontal stroke comes second, I'm using images of someone holding something in an outstretched (right) hand; the horizontal stroke is the arm. When the vertical stroke comes second, my images put that thing sitting on the ground by the person's side. (of course, the "thing" is whatever follows the "by one's side" primative). This method hasn't given me any problems.
The one that I'm always second-guessing myself on is the direction of the horizontal stroke in the cliff-like primatives: I can remember that it usually goes left-to-right (as in暦), and I know that in the case of "axe" (斤) it goes right-to-left (which Heisig never mentions). But a few kanji like 栃 I always have to double-check. Actually, I'm still not sure about that one as I write this. 励 is left-to-right, but it doesn't have the slope like 栃 does.

Are you saying that the horizontal stroke of spoon is opposite that of seven? What do you use to know that sort of thing?

CharleyGarrett Wrote:Are you saying that the horizontal stroke of spoon is opposite that of seven? What do you use to know that sort of thing?

My teachers said that this was the proper stroke order, but they themselves never actually used it, so I never worried about it.

You can usually tell by the shape of the end of the stroke - if one end comes to a point/trails off, as in the case of the left side of spoon, then that's where the stroke ends (of course a hook is also a giveaway). Some fonts don't show this difference, but you can definitely see it on the study pages here where the characters are nice and big. If you look at 七, the horizontal stroke is the same weight all the way across and has the little bulge at the end - this is where the stroke ends. Also, a diagonal stroke that curves slightly generally goes from top to bottom, whereas one which is perfectly straight goes from left to right. I find that this is usually harder to see though.

This is not always a perfect guide - drops as in the top of 学　seem to be drawn bottom-to-top by this rule - but for these horizontal strokes I've found that you can usually tell the correct direction. Compare the slingshot primitive in 写　vs 考 also, for example.

As a disclaimer, I wasn't taught any of these rules - my college textbook just had little arrows indicating stroke direction. This is all from my own observation.

EDIT: I think as a general rule diagonal strokes go from top to bottom, with the following exceptions: 1) 七, and 2) when a primitive is "squashed" on the left side of the kanji, and a normally horizontal stroke becomes tilted.

EDIT 2: 栃 is left-to-right. For some reason in font sizes 10pt and smaller the stroke becomes tilted, but in larger font sizes it is horizontal, as I believe it is supposed to be.

Sorry for the large image. This is 72-point Mincho font:

As you can see, some of the strokes have a triangular notch above one end of the stroke, as in 七(seven, top right) and the left side of 比. In brush writing, this would be where the calligrapher twists their brush to finish the stroke and therefore signifies the end of the stroke.

Others have a notch below one end of the stroke, like in 匕(spoon, top left) and the right side of 比. This is where the brush jabs downward at the beginning of the stroke, and therefore signifies the beginning of the stroke. This can be seen in the 6th and 7th strokes of 隶, which go in opposite directions.

In cases where a stroke is rounded into a teardrop shape, the round part is the end of the stroke and the tapered part is the beginning.

So, let me get this straight. The basis for your statement is your analysis of the characteristics of the computer font? (I have to think for at least a minute about the vast difference between the way the font looks and the way the kanji are actually written by pen, or brush) The statement of a Japanese teacher who doesn't follow their own advice? Arrows in the textbook would probably be cool, unless they're wrong. (Just think for a minute about the errata in RTK). I'm sure that authors and publishers are very careful, but still errors can creep in, and who's necessarily going to notice, especially if they got the stroke order correct.

It just seems to me that there should be somewhere that is the basis for orthodoxy, if you catch my drift. It cannot be "handed down from one nihonjin to the next". You know that some would get it wrong and pass it on to their protoges incorrectly. KKLD doesn't have the arrows. Nelson's doesn't have the arrows. Nor Tuttle's. I may have a little booklet for grades 1 thru 6, but no, it doesn't have arrows either.

Do any of you know of a reference that includes not only stroke order but also direction information?

This site seems to have a fairly complete set of kanji animations, including stroke direction. I don't know of any reference like the one you're asking for, but on the other hand, I've only looked at English kanji dictionaries. I'm sure it is not, as you say, "handed down from one nihonjin to the next"; at the very least books on Japanese calligraphy should have this information.

CharleyGarrett Wrote:Are you saying that the horizontal stroke of spoon is opposite that of seven?

Yes.

CharleyGarrett Wrote:What do you use to know that sort of thing?

Beyond saying, "not Heisig, obviously," you just need a dictionary or learner's guide that shows the stroke direction. One online version is called the Taka database, which is used by my electronic dictionary. Shaydwyrm provided another. JimmySeal's information was, I think, just intended to show how you can make a safe guess about the stroke order when you don't have a reference.
Yes, even dictionaries have mistakes in them; every book has mistakes that often aren't even the author's fault, but it's still a bit surprising how many serious mistakes (including mistakes of omission) have survived into the 4th edition (I don't know if they've all been fixed in the 5th). I don't know whether ignoring the proper stroke order in 七 and 匕 is typical of native Japanese, but I think most writers conform to the proper stroke directions roughly as much as they do to stroke order, and certainly more than they do to stroke count (do any Japanese really use 3 strokes to write 子? It's pretty common to write it in 1...)

CharleyGarrett Wrote:So, let me get this straight. The basis for your statement is your analysis of the characteristics of the computer font?

No, the basis of my statement is my knowledge of how kanji are written with a brush. I used the font to demonstrate how the font reflects correct brushwork, and how the direction of the strokes can be deduced from the font.

Quote:(I have to think for at least a minute about the vast difference between the way the font looks and the way the kanji are actually written by pen, or brush)

Obviously, but for the characteristics I pointed out, the Mincho font accurately shows the types of strokes involved.

I guess I'm just more interested in a reference that I can use to check my assumptions, rather than trying to memorize a rule, which is the point of this thread, isn't it? Mincho gives hints that are frequently right, but there are points in Mincho that are also misleading concerning pen-ji forms. I cannot always be sure which times it's decieving. As far as computer fonts go, if you're not writing the kanji, then the stroke order and direction are pretty superfluous.

Thank you for reminding me of the Taka database. I see that an online kanji dictionary has been implemented from it at http://www.mahou.org/Kanji/

It's still a pain to look up things that way. I guess I could add arrows to my KKLD after looking up the ones I have doubts about....

CharleyGarret,

If you use wwwjdic you can also use the "find kanji in the database" and then click on SOD. Each stroke is drawn rather fast but it gives you the direction. Try it with spoon. It is a bit of a pain if you need to check a lot. There are actually a few places where Heisig didn't give enough information on how to write the kanji. I tried to check when he introduced new primitives towards the end.
On a mac you could also use JEdict, it will give arrows for stroke order, but I noticed one or two mistakes (or inconsitencies with what the wwwjdic SOD gave me). It has a feature where it monitors the clipboard, so with a few keystrokes I can get the writing order.
I'm also grateful to some people who posted notes on peculiar writing orders in their stories on this website.

And thanks JimmySeal for the comments on the fonts. I thought the printed fonts would be misleading but the observations you gave are good.

another couple of examples of Heisig errors/omissions regarding stroke order or direction:
#1485, "call," 呼 -- Heisig properly notes that this use of the "water lily" primitive (aka #1484, "even") is different from the others in having a hook on the final stroke, but doesn't mention that the first stroke of that primitive is also different, being written right-to-left (and sloping downwards).

the "cornucopia" primitive, first used in #1508 -- Heisig says this is written in two strokes, but I can't find any evidence to back this up -- I checked several online dictionaries as well as my printed Nelson kanji dictionary. I checked all the kanji using that primitive (well, all the ones that are in the same chapter; there might be more later in the book). They are:
碑 卑 収 糾 叫
In four of these five, the "cornucopia" primitive is written in three strokes, not two.
The exception, #1510, "income," 収, is different from the others. The "cornucopia" primitive is written starting with the *right-side stroke*, and is followed by the rest of the primitive written in a single stroke (according to the two online dictionaries I checked, although it's written as a 5-stroke kanji in my printed Nelson). So in this one case, it probably really is a two-stroke primitive, but it's still not written the way Heisig claims.
I noticed the stroke counts given on this site conform with what I've written here, rather than Heisig's claims, so apologies if this has already been discussed at length.

[Katsuo]

snallygaster Wrote:碑 卑 収 糾 叫
In four of these five, the "cornucopia" primitive is written in three strokes, not two.
The exception, #1510, "income," 収, is different from the others.

The stroke count is correct in the 3rd Ed. of RTK1 (I haven't seen the other editions). Heisig also explains about combining two of the strokes when "cornucopia" is on the left, but omits to mention the new stroke order.

On a related topic, 卑 also has a variant where two strokes are combined into one (see no. 2196 "handmaiden" 婢 and no. 2579 "paralysis" 痺). In fact for less common characters this is the norm.

For SOD with direction, there's also the Yamasa dictionary - which is unbearably slow, not to mention incomplete, but offers handwriting samples and animated SOD, neither of which I've found elsewhere.

I wrote more about it here.

FWIW, the Yamasa SOD for compare shows both first and third strokes right to left. The handwritten version is also interesting (where clearly the "third" stroke goes left to right!).

synewave Wrote:My wife's explanation is that for 左 we write the horizontal stroke first and vertical second because the following stroke is horizontal. But for 右 it's vertical first followed by horizontal because the next stroke is vertical. This seems to hold up over other kanji like 有 vertical; horizontal; vertical. And the 'linen' kanji (don't know it's reading!). Her point was, if I get confused about stroke order, particularly when there are both vertical and horizontal lines involved, writing 2 vertical or 2 horizontal strokes should give me a warning that the order is possibly not correct. There are probably examples that contradict what she says but she was speaking generally.

The above is pretty good for checking the stroke order but since I was talking about Shirakawa in the Kanji Etymology Book thread, I thought I'd mention something else I learnt recently.

Apparently the real reason for the different stroke order for 右 and 左 is as follows: The first two strokes of each are pictographs of the hand. In 右, the first two strokes are a pictograph of the right hand. In 左, the first two strokes are a pictograph of the left hand. I'll try and explain so you can see it.

Hold up your left hand, palm open, palm facing towards you. The first stroke goes from the tip of your thumb to the tip of your little finger. The second stroke goes from the tip of your middle finger and down your arm.

Do exactly the same thing with your right hand. The first stroke goes from thumb to little finger as before, and the second stroke goes from middle finger down your arm.
The tilt of your arm in this case won't be the same as it was before.

Those two strokes may look similar for both 右 and 左 but as Heisig mentions, they are in fact different. In both, the second stroke is longer. This makes sense because the second stroke for both is the one that goes down your arm.

So there you have it, after first posting about this over 2 years ago, it's taken me this long to discover the etymological reason. Synewave's rule is great for checking though.