This post is for not just recognizing kanas and kanjis but writing them as well.
Some rules of thumb in case we are interested in learning how to write the kana characters:
- The general direction of drawing a character is left-to-right and top-to-bottom
- The general order of drawing the strokes is drawing either the top-most or left-most stroke first; once that stroke is drawn, then the next stroke is the next top-most or left-most stroke left, and so on. The exceptions are the hiraganas せ (se), や (ya), and よ (yo), and the katakana ヒ (hi)
- If a simple horizontal line cuts a simple vertical line, the horizontal line is drawn first, from left-to-right; the only exceptions are the horizontal lines in the hiragana も (mo). However, if we look at も carefully, it is not even an exception because there is no ‘simple’ horizontal line that cuts a simple vertical line; instead, も is a single stroke that starts at the top and, if we were to draw the whole stoke, it would cut itself twice. In contrast, the katakana モ is formed by three separate strokes, so the horizontal strokes are drawn first, as expected:
- All one-stroke curves are drawn top-to-bottom. We can write the arm of ヒ (hi) either top-to-bottom or bottom-to-top; both are accepted. The exceptions are the bottom strokes of the katakanas シ (shi) and ン (n), but we can understand these if we consider that the order and direction of the strokes are simply following the flow of the character; the same happens with ツ (tsu) and ソ (so):
mincho vs gothic
The beginning and end of a character stroke might or might not be preserved in a computer font; this is similar to the way roman fonts work. There are two big groups of fonts based on the roman alphabet: the serif fonts have small decorations and high detail, and work best when used on paper, e.g., the Times New Roman font:
serif: the quick brown fox jumps over the lazy dog
while the sans-serif fonts (meaning ‘without-serif’) are plain, with low detail, and work best on computer screens, e.g., the Helvetica and Arial fonts:
sans-serif: the quick brown fox jumps over the lazy dog
Indeed, the font in a random book or newspaper is likely to be a serif font, while the font in a random website or app is likely to be a sans-serif font. It all comes down to resolution: serif fonts, in general, are more legible than sans-serif ones, but for many years a computer screen just didn’t have the resolution needed to display correctly the details of a serif font; with the advent of screens with ever smaller pixels, this reason might no longer apply.
In Japanese, the equivalent of serif is called mincho, which, again, is a high-detail font that works best on paper:
while the Japanese equivalent of sans-serif is called gothic, a low-detail font that works best on a screen [hayataki-masaharu]:
We can think of sans-serif fonts as serif fonts that have been stripped of their detail; likewise, gothic fonts are like mincho fonts that have been stripped of their detail. In particular, the detail of the beginning and end of a stroke that we find in mincho fonts is removed to achieve the plain looking gothic font, e.g., we remove the traditional detail from both the beginning and end of a mincho stroke when we turn it into a gothic stroke.
There is yet another issue with the mincho fonts: it is common to decorate the end of a rightward horizontal stroke with a triangle. This always happens with horizontal strokes of kanjis, but sometimes it happens with horizontal strokes of kanas too:
mincho: 一 二 三 子供 王 石 手 天
In the image of the Japanese newspaper, on the right, the attention-grabbing headlines, which they want to be legible from afar, where high resolution is not an option, might or might not be in gothic, but the text of the article is always in mincho.
Although at first sight it might appear that to write a character correctly we have to memorize the way in which each stroke of that character ends, in reality stroke ends tend to follow patterns. Next we discuss some of these patterns.
ending a stroke
Although writing beautiful-looking characters has a high artistic value, its practical value is in decline. Japanese write kanas and kanjis as little as we use the latin alphabet. Sometimes we scribe notes, or write down supermarket shopping lists, or write a check, but by-and-large the use of email, instant messaging and social media make the need to write neatly an ever shrinking priority. It might even be the case that in the future dictation and personal assistants like Siri, Alexa and similar might even do away with some of the writing that we still do now, reducing even further the need to write well or at all. It is already difficult to find people that even know how to write in cursive, which was appreciated and common just fifty years ago.
Traditionally, we write kanas and kanjis with ink and brush [takumi]. Each character is made up of a number of strokes, which we draw in a given order. The transitions of the brush between strokes leave traces that become part of the character itself; hence, a character with the correct strokes but with the strokes written in an arbitrary order, becomes difficult to recognize. Still, presently, the original traditional stroke order of a character is not necessarily the same as the stroke order of the same character used in Japan, mainland China, Taiwan, or Hong-Kong [WP].
In spite of the variations in the order in which we write the strokes of a character depending on the region of the world in which we are, the strokes themselves are still drawn in the same direction, i.e., with remarkable few exceptions, a given stroke always starts and ends in the same way. Hiragana and katakana, which are used exclusively in Japan, were both derived from the chinese characters so, naturally, their strokes and the way in which we draw them is the same as that of the original stroke in the original chinese character.
Each stroke has a beginning and an end. If we imagine holding a brush, we can start a stroke slowly, with a fade-in, or abruptly, from a single point; we cannot emulate these beginnings of a brush stroke with a pencil or pen. However, we can emulate the endings of a brush stroke with any writing instrument. A stroke ends in one of three ways:
- tome (とめ): a dead stop
- harai (はらい): a fade-away stop
- hane (はね): a feathered flick in the opposite direction of the main stroke
We came up with the following ad-hoc rules to determine the type of end of a stroke; they appear to capture the feeling of how strokes should end that a person proficient at writting them has; these rules work quite well with both kanas and kanjis.
Tome ends stop dead at a given point. With a brush, we draw a tome end by backing up the brush a bit into the stroke itself, to create a clean dead stop, and then we lift the brush. In mincho fonts we have:
Tomes in gothic fonts look similar to those in mincho fonts:
In mincho fonts, rightward horizontal strokes, which always end in tome, might have a triangle. The bottom stroke of モ ends in a tome but it is not marked with a triangle because it is not a horizontal stroke; instead, it is a stroke that drops vertically and then veers to the bottom-right corner.
Harai stops fade away, the result of raising the brush as we finish the stroke; similarly, with a pencil or pen, we just flick the tip away.
In gothic fonts, we replace all the harai ends with tome ends without affecting our recognition of the character:
We end a stroke with a hane stops when we flick the writing instrument backwards or sideways, to achieve a feathering effect.
Some hane ends, like the one shown for the hiragana れ, are small and we replace them with tome ends in gothic fonts; removing some large hanes, like the one in the hiragana は shown, depends on the font; some large hanes have become an important feature of the character and we cannnot remove them even in gothic fonts.
All together now!
As the number of strokes of the character grows, so does the likelihood for it to have stokes with multiple types of endings:
With the previous information, we can probably intuit how the ends of the strokes of these characters are going to look in a gothic font, i.e., all harais become tomes, the tomes are unaffected, and the hanes might or might not remain depending on the font and on how important the hanes are to the identity of the character, e.g., the hane of the katakana オ below was replaced with a tome in this particular gothic font, but actually it is a pretty important feature of the character so other gothic fonts draw it.
tome rules of thumb
Perhaps the easiest stroke end to describe is the tome, because it remains a tome in gothic fonts. There are three directions in which a stroke that ends in tome might take: downwards (close to vertical), bottom-right-wards, and rightwards (close to horizontal). As long as the stroke is a line or a curve going in those directions, its end is always a tome.
These are strokes that travel downwards, either in a straight line or a slight curve:
All strokes towards the bottom-right corner, either diagonal or flat, end in a tome.
All near-horizontal strokes, whether they have some curve to them or not, end in a tome.
These horizontal strokes are the ones that mincho fonts might mark with triangles:
harai rules of thumb
Harai strokes fade away instead of stopping abruptly. The stroke always ends in a harai if its final direction is left-wards or up-wards.
This is a unusual stroke; normally strokes move downwards and/or rightwards, but these ones go upwards and leftwards simultaneously.
hane rules of thumb
Hanes are flicks that don’t follow the direction of the main stroke. Hanes might be small or large, which often correlates with whether we actually need them to recognize the character or not.
Small hanes can have any orientation with respect to the original stroke, because they are usually remnants of the transition between the end of a stroke and the beginning of the next one. Hence, we tend to find these small hanes in intermediate strokes instead of the last stroke.
Below, the first version of the hiragana さ (sa) has a small upwards hane at the end of the first stroke, while the last two versions have a small leftward hane at the end of the second stroke:
さ さ さ
In the following gothic fonts, the hanes of this same character have been omitted completely:
さ さ さ
Finally, in the following mincho fonts, both hanes have also been completely omitted because the three strokes have been merged into one:
さ さ さ
Since the character is equally recognizable whether it has a hane or not, or whether the strokes are joined or not, then, from the functional point of view, the hane is irrelevant.
If the small hane is ending a horizontal stroke but the font turns it into a tome, like in the case of さ above, we might end up finding it in a mincho font drawn with the triangle decoration, as if it had been a horizontal stroke ending in a tome all along.
Other examples of small hanes, whose roles are more visual and artistic than functional, are the hanes of the vertical strokes of these characters:
Above, the first strokes of は, ほ, に and け have small top-right-ward hanes, while the first strokes of わ, れ and ね have small top-left-ward hanes; in both cases, the hane is simply a transition to the second stoke.
In the following gothic font, all the top-right-ward hanes are preserved, but the top-left-ward hanes are omitted:
while in a different gothic font all the hanes are omitted:
There is no argument that these hanes enhance the beauty and the legibility of a character; however, since they are not necessary to recognize the character, is difficult to decide when to write them. It seems sensible that we should write them if we are trying to write well, while we would probably not write them if we are just jotting down something.
Unlike the mostly ‘optional’ small hanes that we just saw, large hanes are part of the character, and the character becomes mostly unrecognizable if we remove them. The first type of these hanes is the hane that flicks to the bottom-left direction. This hane is part of many kana characters:
All other large hanes that we cannot remove, which appear often in kanjis, are hanes that turn at a right angle from the main stroke. Some turn right from a curve:
some turn upwards from a horizontal stroke:
and some turn left from a vertical stroke:
Character strokes can end in either of three possible ends: tome (dead stop), harai (fade-away stop), or hane (flicker stop). With regards to the tome and harai, they each apply to strokes that end in specific directions: strokes towards the left or top are always harai; otherwise, they are always tome:
Hanes that turn a sharp right from the main stroke are often a distinctive part of the character and we cannot remove them; we will find them in all gothic fonts.