May 13th, 2025: One and a Half Months of “Dream”

I was one of the students responsible for documenting Liu Yu’s work. I wasn’t the student who constantly supported her research, but I had regular opportunities to meet her at meetings and events. Here, I would like to share a little bit about Liu Yu from my own perspective.

Since early April, there has been something on my mind. Whenever Liu Yu talks about stories or narratives, the word “dream” appears again and again. In Japanese, “dream” translates to yume. But as I kept hearing her use the word “dream,” I began to feel that it might refer to something different from the yume I know. If we take “dream” to mean only “what we see while sleeping,” the conversation doesn’t make sense. Even in one of her workshop emails, Liu Yu described the workshop as “a space to share dreams and stories.” Isn’t it curious to put “stories” and “dreams” side by side like that?

Around that time, I went to a temple with Liu Yu to conduct an interview about ningyō kuyō (doll memorial rituals). It was my first time accompanying her for research, and to my surprise, I went alone with her—which made me nervous. But as I interpreted her questions and noticed, “Oh, that’s what she’s curious about,” my nervousness faded. Liu Yu wanted to learn not only about the ritual itself but also about the very idea of a “doll’s soul.”

When I relayed what the temple staff had told me—“They say ningyō kuyō is performed to remove the soul from the doll”—Liu Yu asked me an unexpected question:

“Is that ‘soul’ something that comes into the doll over time, like life settling in it, or is it the accumulation of people’s emotions?”

I was caught off guard.

“Um…” I frowned noticeably and asked the temple staff,

“When we say a doll has a ‘soul,’ it doesn’t literally mean the doll has life, right?”

“That’s right… How should I put it… Maybe it’s closer to the feeling that when you treasure something, it seems to have a soul.”

“But the idea that the soul is simply people’s emotions doesn’t feel quite right either.”

“Yes, that doesn’t sound exactly right either…”

Even among us Japanese, we were at a loss. Once we tried to put into words how we understand “soul” or what it means for a soul to “depart,” something seemed to slip away. Liu Yu quietly watched us struggle, then spoke:

“In Taiwan, we also have the idea that objects can possess souls, just like in Japan. But we don’t have doll memorial rituals.”

“Really?” The temple staff and I looked at each other. If the cultural belief that objects have souls is the same, then it wouldn’t be strange to have rituals or ceremonies for them. Yet it seems not. How puzzling.

In the end, the temple staff and I explained it like this:

“When someone cherishes a doll, their feelings accumulate within it, making it seem as though the doll has a soul. In the ritual, those feelings are removed from the doll and entrusted to the Buddha.”

We felt a sense of accomplishment, as if we had completed a task. Liu Yu nodded, saying, “I see,” and moved on to her next question. I don’t know if our true intentions were conveyed. Perhaps the words and phrases were delivered, but in reality—who can say?

Feeling uneasy, I realized something within myself: I had a greedy desire, a strong intention, to “make Liu Yu truly understand.”

Before Liu Yu came to Japan, I had already seen her work at the Yebisu International Festival for Art & Alternative Visions. I remember it well because it was such a mysterious piece. It was a video installation based on a flood narrative. On the screen, sentences narrating the story appeared, or single Chinese characters would show up one by one—but they disappeared almost immediately. The text would change so quickly that by the time you blinked, it was already different. The speed of this shifting varied throughout the piece. Alongside the words, there was a narration reading the characters aloud, accompanied by the rhythm of drums or some other percussion. In front of the screen stood several curved objects, and from time to time, images were projected onto them as well.

What struck me as strange about the work was that it did not aim to “convey the story accurately.” If the goal were simply to tell the story, the text could have been displayed slowly enough for the audience to read, and it wouldn’t even need to be a video at all. But instead, seeing the story through moving images placed a different kind of experience into the audience’s hands—something other than straightforward understanding.

Over lunch, I finally told Liu Yu for the first time that I had seen her work at Yebisu. I had long thought I should mention it someday, but since I didn’t have a clear question to ask, I had kept silent. After joining her on a research trip, though, I felt, “Maybe now is the right time to bring it up.”
“I saw your work at the Yebisu Festival,” I said.
“Oh really? Thank you!” she replied.
“That flood video—it was fascinating. The story appeared on the screen, the text came and went. I don’t remember the actual story very well, but I do remember the image of the story.”
Liu Yu smiled and said, “I’m glad to hear your impressions.” Then she added:
“That was the best way I could think of to express the flood narrative.”

Back at Geidai, I realized Liu Yu would soon leave and return to Taiwan. Before that happened, I wanted to look back on her residency together with her.

Her research, which began with the anthropomorphization of plants, eventually culminated in interviews about the Android Kannon and ningyō kuyō (doll memorial rituals). Plants, dolls, androids—what a surprising connection. I wondered how her research had expanded in such unexpected directions. Following the path of her research through conversation was fascinating.

“Did you start out only interested in mandrakes?” I asked.

“I began looking into mandrakes because I was curious about the relationship between Western people and nature. From the beginning, I was also exploring other forms of anthropomorphism.”
“And then, along the way, you happened to discover that aibo robots were included in doll memorial rituals?”
“Yes, and that led me to the Android Kannon.”

This was the first time I learned that her research had passed through Aibo. That made sense—since Aibo embodies qualities of both dolls and androids, it was natural that her research would extend to androids and doll rituals.

Apparently, Liu Yu had been conducting multiple strands of research simultaneously. She read the Kojiki, investigated the Edo-period figure of the Fuji Musume, and checked articles about Kodaiji Temple, where the Android Kannon is enshrined. At her talk with Anzai Tsuyoshi, she presented her findings neatly arranged in slides, so I hadn’t realized how chaotic her actual research process had been. She laughed, saying that during her stay, people who knew her interests kept recommending things to her, and she simply absorbed everything she could. Indeed, whenever I saw her, she was always thanking someone for lending her a book or a reference. It was a familiar sight.

She also told me that when she first began researching mandrakes, she was doing a residency in Taitung, a region inhabited by indigenous people. This was new to me as well.
“The people in Taitung treat trees as members of the family. They believe their ancestors were born from plants—they have a very strong bond with nature. I’m not sure if I’m expressing this correctly…”

Though it felt somewhat distant from the world I live in, hearing it from Liu Yu made it feel real and continuous with our own reality. It reminded me of her Yebisu Festival work.

I usually spoke with Liu Yu in English, though I’m not particularly good at it. While writing this article, I replayed the recordings of our conversations several times, and honestly, they sounded terrible. I was fumbling through words as if I were blindfolded, cautiously tracing the outlines of objects with my hands. I kept asking her, “Is this what you mean?” Her English, though still English, felt completely different from the way I spoke it. In the end, the only way forward was to draw closer to her in my own words. At one point, I realized that my clumsy manner resembled the way she herself behaved during her fieldwork interviews or when she interacted with students in workshops. And in that moment, the meaning of the word “dream”—why she used it—suddenly became clearer to me.

The word “dream” here is closer to the sense of daydream, or something uncanny, like a mysterious event or being caught in a trick of the fox. It’s not the kind of dream one has while asleep. Yet, I don’t think she used dream merely to describe something “strange.” This part is my own interpretation.

When she says “dream,” I think she emphasizes the way an experience exists. A dream is something that belongs only to the dreamer, impossible to fully explain to others, yet still a complete world of its own. One might describe the events that happened within it, but not the underlying reasons—those remain tied to a subconscious realm that resists language. This way of being surfaced both in the stories she told me about Indigenous peoples and in the work she showed at the Yebisu International Festival for Art & Alternative Visions. For the Indigenous community she described, whether or not their ancestors literally came from plants is not the issue—it is simply true for them. To call it merely “culture” would be reductive; there are countless things in this world that can only be expressed as “that is just how it is.” My own clumsy attempt to explain the ritual of “removing the soul from a doll” belonged to the same category. And these dreamlike experiences, once they take shape in words, unconsciously transform into narratives. In Liu Yu’s work at the Yebisu International Festival for Art & Alternative Visions, what the audience gained as “another kind of experience” was the vivid recognition of the existence of a world different from the one they themselves live in.

Now I think I finally understand why her workshop was called “a space to share dreams and stories.” The deeply subjective “dreams” of each participant were reborn into the shared world only when spoken aloud, when expressed through ordinary language to others. The workshop was, quite literally, a space to share that moment of birth together. That is why, I suspect, she insisted on distinguishing “dreams” from “stories.” Stories are already formed, molded, and framed by a fixed vocabulary. Dreams, on the other hand, sometimes resist words, sometimes shouldn’t be put into words at all—they are unshaped clusters of experience that nonetheless exist.

Liu Yu is extraordinarily kind. That’s something I felt every time I met her. Whenever she sensed that her way of understanding might differ from someone else’s, she made sure to ask again, patiently. It may seem simple, but to hold that same attitude with everyone, always, is incredibly difficult.

Speaking with her constantly made me reflect: when we exchange words, what exactly do those words point to? If two people dream the same dream, can they describe it in the same way? Or would different “tellings” inevitably point to different things? Sometimes I worried: Am I really communicating with her? But the times when she listened to me, laughed, or shared her thoughts about her work—those moments remain deeply with me even now.

Author: Asuka Taniguchi (GA master's student)

Project Coordinator: Qiuyu JIN (GA PhD student)