Writing Characters Beautifully

When I was in elementary and junior high school, I learned calligraphy from my father, who was a kanji calligrapher. Although I was in a science-track class in high school, that family environment led me to enter the Faculty of Letters at Keio University.

After entering university, I asked my father to introduce me to a teacher I knew. He told me, "Your lines have been soft since you were a child, so how about kana calligraphy?" and introduced me to Ichiro Higashiyama.

At first, I didn't have a strong interest in kana calligraphy and couldn't even read it, but I studied under Master Higashiyama for 38 years until he passed away in 2013.

Your father was watching you very closely, wasn't he?

I don't know if I was suited for kana calligraphy myself. I still don't even know if I like it (laughs). Even so, I've been able to continue for a long time. I can write for as long as I want, so I suppose I was suited for it.

The depth of kana calligraphy is limitless; there are flavors and lines that I can't produce now but might be able to write 10 years from now. Personally, I hope to age gracefully in my art.

"Aging gracefully"... that's wonderful.

On the other hand, I do feel my age. There are things I can't write no matter how hard I try. In the eyes of the world, I'm at an age called an old man, but in the calligraphy world, I'm still just starting. I'm aiming to reach my peak between my mid-70s and around 80. Many of my predecessors left behind great works in their 70s.

I was originally a company employee, but I wanted to acquire some kind of skill. One of the things I tried was traditional Western calligraphy. I started without really knowing what I was doing, but the more I did it, the more I realized it was a world without end. More than anything, I was amazed that there was something I could become so absorbed in. Besides the beauty of the letters, I felt the joy of becoming part of a long history, so I opened a studio and classroom to pass on the technique and culture.

Studying history is also important in calligraphy. I met a Swiss teacher named Muriel Gaggini, who taught me the importance of systematically learning the historical transitions of letters.

Is learning both penmanship and history considered authentic calligraphy?

That is the ideal. The shapes of Roman letters have changed every 100 to 200 years, and behind that are historical events such as wars and changes in lifestyle. While being influenced by these, some parts are passed down and others change. To understand one script, it is important to study the scripts of the periods before and after it. I also tell my students not to be satisfied with learning just one script, but to deepen their understanding within the flow of history.

Are there trends in the scripts and documents that remain today?

Many documents are related to religion or politics. There are also academic ones. Those related to religion are the most common, and the majority are manuscripts of the Bible. Since it takes a tremendous number of people and time to create a single book, our predecessors must have written them down by the light of flickering candles or dim windows, with a passionate desire to pass them on to future generations.

My specialty is robotics. I've loved Doraemon since I was a child, and I vaguely felt a dream in the idea of science and technology creating the future. I went from Keio's affiliated schools to the Faculty of Science and Technology and pursued a path in research. Specializing in data robotics, in 2012 I developed the "Motion Copy System," which reproduces handwritten characters with high precision.

In an era of population decline and labor shortages, I am conducting research day and night to create robots that can replace human hands in fields like production, as well as medical and nursing care—just like writing characters.

However, it is quite difficult for a robot to reproduce human movements. Current technology is good at programming and repeating fixed movements, but it is poor at controlling force. Even with the "Motion Copy System," we struggled greatly to reproduce brush pressure.

Human movement consists of two types: "movement in space" and "the degree of force when touching something." Writing characters is a movement that mixes these two. It was said that the delicate movements of calligraphy were impossible for robots, but we continued our research and, around 2006, were able to clarify some of the fundamental theory that makes these two compatible.

While I was thinking about how to explain this to the general public in an easy-to-understand way, someone suggested that I try reproducing calligraphy using robot technology. That person's father-in-law was Master Juho Sado, who served as a counselor for the Yomiuri Shofukai. Through that connection, Master Sado was the first to take up the brush with the first prototype we developed, and we were able to program those movements [Figure 1].

So you memorized the movement of writing while holding the arm.

That's right. We turned the master's movements into data and automatically reproduced them with a robot. Feeling encouraged by this, we were able to get cooperation from Takeda Souun, Shishu, and Shoko Kanazawa for the second prototype.

By accumulating a lot of this data, will robots be able to predict and reproduce other characters by famous calligraphers?

That's where we expect a lot from AI in the future. Our research is first about faithful reproduction. Completely reproducing brush pressure was a world-first attempt.

Applying this technology, what kind of things will be possible in the future?

One would be the preservation and transmission of skills. By digitizing the penmanship of calligraphers, it becomes possible to reproduce it for future generations. Since writing speed and brush pressure are also quantified, specific movements can be analyzed.

In addition, I think it could be used for training if the differences between a teacher's and a student's characters can be compared using data in practice and instructional settings.

For example, is it possible for a student to hold the robot arm and experience the movements it reproduces?

Yes, that is also possible.

Then, will it eventually be possible for me to experience Mr. Iwai's brushwork and reproduce his characters?

It is likely feasible. Even if Mr. Iwai himself is not there, we can create a situation where he can guide you by taking your hand.

Why did you think of making a robot write characters?

The main reason I wanted to try was that it was initially said that calligraphy was the one thing a robot couldn't do.

This technology was featured on a TV program, and people who saw it asked if it could be applied to support stroke or brain injury patients with higher brain dysfunction.

One symptom of higher brain dysfunction is a state where a person can read kanji but cannot write the same character even with a model in front of them. This is called "agraphia," and while rehabilitation methods have not yet been established, we received inquiries about whether our technology could be incorporated into rehab. That said, there are still many things robots cannot do, and I want to continue improving and enhancing the technology in the future.

My interest is endless regarding how delicate the movements that can be reproduced will become in the future.

As you say, this is a common challenge in the robotics industry. The arm for writing characters only has three motors corresponding to length, width, and height. Human hand and arm joints number nearly thirty. It is a challenge to make it possible to move each part individually and increase the degree of freedom to a human level.

After all, the human hand is the most superior. It is quite difficult to reproduce the movement of using tools with a robot. By the way, Mr. Iwai, how are you conscious of using your body when you take up the brush?

When writing characters, I sometimes write at a desk and sometimes on the floor, so the way I balance myself varies.

For example, I am sometimes asked to write as a demonstration in front of people. Once, I wrote on a piece of paper about the size of a tatami mat in front of about 300 people. At that time, a kanji calligraphy teacher and I each wrote, and the way we placed our left hands and balanced ourselves was completely different. There are also cases where posture changes as abdominal and back muscles weaken due to aging.

You never take the exact same action when writing characters repeatedly. There is a technique called "gyakuhitsu" (reverse brush) where you write with the axis of the brush and the tip of the bristles reversed. In this, there is a moment at the end where you relax your strength slightly and the brush itself writes using the elasticity of the hair. It might be quite difficult to have a robot make that choice while writing.

But we have high expectations, don't we?

Yes, I think it's interesting.

Do you have any resistance to a robot reproducing it?

I have nothing but expectations.

To be a bit difficult, the elasticity of the brush differs depending on the type of hair, and the way the ink bleeds or the brush catches differs depending on the type of paper.

In tennis terms, it might be similar to how the ball bounces, the way you hit, and the equipment differ between clay courts and hard courts. Calligraphy also has brushwork that corresponds to the brush and paper. I think it would take a tremendous amount of effort to input such data.

That is exactly the challenge. Currently, there is still a lack of information, and to reproduce the same character, it must be the same hanshi paper and the same brush.

The road ahead is long and exciting. The reason I have expectations for such technology is that we will all be gone eventually, so we think about how to leave behind our current knowledge and skills.

For example, a quill pen is just a feather plucked from a bird and cut, but it is incredibly delicate. I can take the hands of the students in front of me and convey subtle adjustments of force, but that won't last forever. Also, even if I am confident now in distinguishing between the top and bottom of every single blue line that divides graph paper into 1mm increments, I won't be able to write with the same precision when I reach an age where I can no longer focus. I'm always thinking about what I should do.

From Mr. Katsura's technology, I don't feel a sense of crisis that we will be replaced. There's a sense of security (laughs).

That's true. I'm really looking forward to it.

I'm glad. Today, I was worried about what I would do if the reaction was that AI would take away people's jobs even in the field of art.

Not at all. I really have nothing but expectations.

I think that once you acquire the minimum techniques by learning calligraphy, you should then refine your individual sense and sensibility. It's probably the same with painting. You repeat copying and learn color usage and touch, and then you acquire originality, right?

In the case of calligraphy, there is a strong impression of always looking at a model while writing, but to be able to write characters that are uniquely your own, you have to learn many things. I have high expectations for robots to serve as a bridge to that.

In Mr. Iwai's story about gyakuhitsu, it was impressive that what is written differs depending on the choice of brush and paper. Does calligraphy also have an aspect where the body dialogues with the brush and paper?

We use Western paper, but we pay very close attention to distortion and the degree of bleeding. Even when using the same paper and tools, the influence of the climate is significant.

We use not only paper but also parchment. Since parchment is animal skin, individual differences are huge. It's a part of an animal's body itself. As for quill pens, there's the flexibility, the degree of the curve, the oiliness of the waterfowl. It's different every time depending on the combination of tools, so I have no choice but to change them from time to time. To write accurately, fine adjustments are necessary from the moment you touch it.

When you say writing accurately, what is it accurate in relation to?

The alphabet has very few strokes compared to kanji. There are only twenty-some simple shapes, and many of them are similar. Therefore, the clarity of lines and the accuracy of shapes become important. This is one of the points that gets through to the reader.

I feel every day that the history of letters is the history of the human heart, and that things were written down because there was a desire to preserve them. Precisely because the alphabet itself doesn't have meaning, I want to value "writing to convey" by making full use of text and impression [Figure 2].

Even a little bleeding or the thickness of a joint can greatly change the impression. With wide pens for old scripts, you can't write thicker than the width of the pen, so the difference in thickness can only be created by writing thin parts thinly. But that difference is exactly what elegance is.

The ability to see is also important. Especially with traditional calligraphy, it's said that it looks the same no matter who writes it, but to me, not only the characters of famous calligraphers but also those of my students all look different. The extension of lines, the unquantifiable differences in joints and flicks, the understanding and choice of script. No two people are the same. But since there are still few people with such eyes, I want to pass on the skill of seeing along with the skill of writing.

Mr. Iwai, you are also teaching kana calligraphy, and I imagine that considerable training is required to decipher the works. Do you also teach your students so that they can read them?

I teach them so they can read. But, of course, they won't be able to read suddenly. I trained for many years too. After all, it starts with imitation.

When I was in my 20s, a teacher told me, "Three years for the use of ink, a lifetime for the line." I've become quite proficient with ink, but the line is still a work in progress. Recently, I've finally started to feel in my own way that I'm becoming able to write negative space as well.

"Writing the margins"—what exactly does that feel like?

One of the elements that makes something look beautiful is the resonance or lingering afterglow in the margins, but that is extremely subjective. If someone says it doesn't look beautiful, that's the end of it. However, through my own repeated trial and error, I feel I've reached a level where I can create beautiful margins by writing in a certain way.

How to create margins is also important in calligraphy, but I am still learning.

In the Heian period, literacy was likely limited to the aristocracy, but the phrase "scenery of ink distribution" (sumitsugi no keshiki) appears in Heian literature. Also, in "The Tale of Genji," there is an expression "mixing dark and light to confuse," which used to be translated as "so that no one knows who wrote it," but recently it is translated as "adding shading and contrast."

The existence of the expression "mixing dark and light to confuse" makes me wonder—was there something like a textbook for writing styles even in the Heian period?

No, it seems it was just an expression found within "The Tale of Genji." People of that time likely found the shading of ink to be beautiful.

We attempt accurate rinsho (copying masterpieces or famous calligraphic works). Through rinsho, we sense the breath of the person who wrote it, the rhythm and breathing of their brushwork, and re-experience it. We are learning that way now, but I suspect these aesthetic sensibilities haven't changed much since the Heian period.

Is it possible to study things written in the Heian period by looking at them directly?

Yes, you can. At KeMCo (Keio Museum Commons) on the Mita Campus, you can see ancient books and Chinese classics from the Heian period. You can really sense something like the breath of the people from those items.

When I touch such things, my imagination expands, thinking about the angle at which they must have held the brush. When I actually try it myself, I begin to understand the physical sensations to some extent, such as the angle of the brush and the placement of the paper. There is much to be discovered by trying to reproduce it—like noticing the rhythm of the brushwork is a bit hurried. Isn't there something similar in Western calligraphy?

I understand. Looking at a manuscript from over 1,000 years ago, I might laugh and think, "What a bold person this was."

When I watch my students, even though they are supposed to be looking at and practicing the same thing, differences emerge. Those differences may eventually become their individuality.

When they first start learning, are they instructed to suppress their individuality?

What appears first is usually a habit rather than individuality. I believe individuality is something that appears after a foundation is built through imitation. It might be similar to how the world of academia starts with properly reading previous research papers.

However, in the world of kana calligraphy, it is said that almost all new expressions had been exhausted by around the 1100s. It was invented around the 900s and blossomed over the following 200 years. Even people in the Kamakura period were already doing rinsho and imitating, and that continues to the present day [Figure 3].

In an attempt to be individualistic, some people try to do something different from others without understanding the meaning of being able to write a single line beautifully. If it's just being different and not their true individuality, a day may come when it becomes painful to continue writing those characters. I want them to search for their true individuality without rushing.

By valuing the basics and gradually improving quality, the hand remembers the lines as they repeat several scripts. By the time the lines are mastered, the analysis of scripts becomes faster and understanding deepens. Then, gradually, subtle differences emerge, such as the overall look becoming a bit more flamboyant or having a more solid feel. When striving for and practicing the same thing, those differences appear more significantly. I believe that is what individuality is.

So the basics are the most important after all. By the way, people often talk about having good or bad handwriting.

In calligraphy, "skillful writing" and "good writing" are different, but "good writing" is probably not something that can be taught. For example, the writing of a monk is partly written through their character. Even if you try to imitate it, it's not easily done.

That's true. I believe individuality is not something you create, but something that exists. It doesn't come out immediately. Since the time I can directly instruct each person is only a few hours a week, how sincerely they face the characters during the rest of their time also has a major impact.

It's like muscle training.

Yes, exactly like muscle memory. I feel that the more sincerely a person faces it, the sooner their individuality appears.

By the way, by combining AI with your technology, Professor Katsura, is there a possibility that robots will be able to write characters with individuality in the future?

We are still at the stage of struggling just to reproduce it. From a microscopic perspective, all elements such as paper, brush, and ink are comprehensively involved, so reproducing such physical phenomena requires very advanced technology.

Whether it's "good" or "bad" is also a kind of flavoring, and arrangements like changing thickness or strength are likely possible. However, creating something from 0 to 1 can only be done by humans. As I was listening to you just now, I thought that truly original parts can probably only be produced by humans. However, in the process of training leading up to that, robots might become able to objectively look at the calligraphy they have written and get feedback.

I am very interested in how robots will grow.

Copied things are still just data on speed and force; the intentions, such as why force was applied in this way or why the ink was reapplied here, have not yet been turned into data. Without pursuing those causes, it will be difficult to bring them closer to humans.

Mr. Iwai, I heard you studied under Master Ichiro Higashiyama for over thirty years, but even after decades of learning, your writing never becomes identical to your master's, does it? That's the interesting thing about characters written by people.

Master Higashiyama had a policy of not writing models, so the basis of his instruction was rinsho of the classics. He would say, "Even if we study the same classic, what I learn and what you learn will be different, so the results should be different." He was the type of person who found it creepy if people wrote the same way despite the age difference, and he wouldn't even look at things that imitated him.

In that sense, you could say he cultivated my ability to find and eat my own food. He was a teacher who said that if you only use him as a model, you're just a zoo animal, so I was grateful for that policy.

He sounds like a wonderful teacher.

Even so, when Master Higashiyama passed away, I felt like I had lost my support. Until then, he had been choosing my works to be submitted to exhibitions. He was the kind of teacher who would criticize or say, "It's not to my taste," but would still tell me, "Submit that one."

Looking back at his way of seeing things now, I realize he was saying the right things. I can understand what he told me more and more as I get older.

Therefore, I might not be able to understand the things he said in his 70s and 80s until I approach that level myself. I started feeling that way in my 40s, and since then, I've become a bit more obedient toward him (laughs).

In academia, we also start by thoroughly doing the basics, but research inevitably requires originality. That said, something doesn't just come out of a state of absolute nothingness. After mastering the theories passed down from my mentor's generation, I began to add my own arrangements starting around my Doctoral Programs. I'm still halfway through my journey as a researcher, so I can really empathize with what you've said.

What was the teaching method of Muriel Gaggini, whom you studied under, Ms. Nishimura?

I have completely taken over Master Gaggini's teaching method. First, I was told to look at manuscripts and look closely at the models. She would say, "You think you're looking, but you're not, that's why you can't write."

Since the alphabet is made of combinations of simple shapes, we first analyze the characteristics of the script using circles and lines—namely O, I, and L. Once you can write the basic forms of all the letters in a script, you move on to variations and arrangements. This includes techniques for saving space when it's tight, or how to extend letters when there's room, effective image changes, and how to balance the whole. We acquire these along with the history.

Is the "overall balance" similar to the margins Mr. Iwai mentioned?

It might be. Previously, when I submitted work to an exhibition overseas, a local artist told me my work was very Japanese. I had intended to submit a "classic Western" piece exactly like a manuscript, so at the time, I didn't understand what he meant at all. Apparently, I had left wider margins and written more straight than a typical Westerner would choose.

Of course, Westerners also write straight, but there must be points that look Oriental to a Westerner that a Japanese person wouldn't notice. It might be one aspect of my individuality as a calligrapher, but it was a shock because I hadn't intended to express "Japaneseness."

It probably just seeped out naturally.

There might have been something cultivated in me as a Japanese person. If so, I'm happy.

There are certainly similarities to the margins in kana calligraphy. This is an original Japanese culture, but perhaps there is a beauty of margins that only Japanese people can sense. Even in kana calligraphy, individual differences often appear when writing the same thing. For example, even if you try writing the same two lines, a slight difference of a few millimeters in the line spacing can make it look cluttered or lonely.

So subtle differences significantly change the impression given.

There is also line spacing that brings out the best in your own characters. Furthermore, there is a technique for creating space within the characters themselves. There are margins at various levels, and the whole is created by their accumulation. The starting position of the next character can even change depending on the momentum or shape of the line written before it.

I understand that very well.

Kana calligraphy and your Western calligraphy might be quite similar. I feel the types are close.

I'm very honored.

Even if the position of the lines is the same, the line spacing can look wider depending on the shape of the characters. It's a truly subtle difference, but it's a mysterious thing.

This might be quite a tough conversation for you, Professor Katsura (laughs).

It's often said that with the advent of AI, it will surely do everything for us, but today I feel like I've been taught that's not the case after all (laughs). Both of you practice every day and say you are halfway through your respective worlds, and there is difficulty in improving skills within that. I felt that for robots to improve, they must similarly accumulate various processes.

The world of computers has high-speed calculation, so there is the advantage of being able to handle large amounts of data, but I thought that if we don't do the steady work very steadily, we might fall into the trap of extracting only habits rather than individuality.

In addition to writing characters, I have a desire to pass on the history of calligraphy. From that standpoint, I feel I have to say that what is poor is poor (laughs). I believe my role is to keep building, adding, and conveying.

Yes. I suspect every world is probably the same. I want to proceed with robot research that reproduces various skills like manufacturing, but in the world of research, training also takes time. If we pursue why there are things that only humans can do, we will eventually reach questions like what skills are and what is necessary for robots to acquire them.

In your research, I would be happy if you could connect it to something that can objectively explain why the margins of my characters look beautiful, for example.

I also teach lectures at the university, and the first thing I tell the students is that I want to give a class that requires as little sensitivity as possible. This is because Master Higashiyama was the kind of person who taught by saying, "Don't you think it would be more beautiful to put this dot here?" If you listened in silence, he'd say, "I guess you don't think so, which is why you put the dot there." Sarcastic, isn't it? (laughs).

But in a class, you have to be able to give an explanation with objectivity, such as "putting a dot here expands the margin." I don't think "this way is more beautiful" will get through to people with different sensitivities. I hope for that kind of analysis from you, Professor Katsura.

That said, today's theme of "writing beautifully" is a difficult problem.

It really is difficult.

In the world of Western calligraphy, clarity, accuracy, and the beauty of lines are important. We must be careful not to become ambiguous.

While characters and words change, there are always parts that don't change. I believe that writing each line carefully, making clean joints, and writing accurately with all your heart leads to beauty. Because the characters are small and have few strokes, I strive for precision. Naturally, my feelings also get carried into the characters.

In that respect, kana calligraphy might be a slightly different approach from the general public writing beautiful, neat characters. In our case, we write several characters connected together. But "connecting" includes both cases where they are connected by an actual line and cases where the breathing appears connected even if they are separated. It's a sensation like "receiving," and there is a just-right distance in that interval; you could say that way of connecting and the flow of breathing create beauty.

However, even if I try various things and think, "This is interesting," the truth is that almost everything was already done in the Heian period. There are almost no new inventions. Even if there were, such works would be self-indulgent and wouldn't move the viewer.

Therefore, it's better not to do too many unnecessary things and follow tradition. This is because at the root, there may be the DNA of Japanese aesthetic sensibility that hasn't changed since the Heian period. If that aesthetic sensibility is still alive, I don't think modern people need to go out of their way to try to write in a modern style.

It feels like something that has been built up day by day.

I think that is something that naturally seeps out. If you follow classical styles and sincerely try to express your own individuality, I believe that will look the most beautiful. In words, it would be described as elegance and dignity.

I agree.

It is amazing that everything was already explored to its fullest during the Heian period.

It is the same for our calligraphy, and perhaps for your technology as well, Mr. Katsura, but even if we are determined to do something different, most attempts made without knowing the past are things that were already weeded out in the past. I believe that the seeds for one to shine even brighter are found within the old rather than the new, so I think it is important to first sincerely study the things of the past.

I feel that calligraphy is suited for people who enjoy the process of holding a brush or pen and writing something, without seeking immediate results.

Among my students, those who continue steadily and dispassionately tend to become more unique. Their characters become free-spirited and more like themselves.

Recently, opportunities to write with a pen have really decreased; at most, it is when I am correcting students' papers. There are various tools now, but I try to return corrections written by hand as much as possible. This is because I believe there are parts of writing where the person can put their heart into it. Even when submitting documents to the university, I attach handwritten notes, as I feel the human heart manifests in such things.

Robots are also often thought of as being somewhat inorganic, but I thought that if we don't just replicate human actions, but see the underlying theoretical principles through considering historical things as Ms. Nishimura does, then those robots might become capable of movements that have an affinity with humans.

Much like Leonardo da Vinci, from an era when art and engineering were once integrated, academic disciplines gradually became subdivided, and now we have reached a point where we can read and write on things like smartphones. From now on, I want to incorporate human-derived elements into robots in a way that reunites art and engineering.

That sounds wonderful. I look forward to such a future.