Listening to the Daoing in the Morning

1 Society of Multi-layered Values

There are many layers of values, and these layers are intertwined in a nested manner. Therefore, even within a particular culture, it is impossible to claim the existence of a single, essentialized value. Depending on which value we prioritize, the relationship with other values shifts, and the perspective through which we view a given culture often changes significantly.

As Thomas P. Kasulis’s Intimacy or Integrity: Philosophy and Cultural Difference shows, every culture contains two value orientations: intimacy and integrity. The cultural landscape changes depending on which of these is foregrounded. A culture based on intimacy emphasizes relationality rather than independent individuality, whereas a culture based on integrity takes independent individuality as the starting point rather than relationality. The following discussions between Yasuo Deguchi and me are relevant to differing interpretations of this entanglement between intimacy and integrity.

Furthermore, values do not exist in isolation; they are always created through human engagement. Recently, the University of Tokyo and Daikin Industries have been running a program called “Valuing Air,”^[1] which is not simply about recognizing the value of air, but about considering the performative significance of engaging with air by valuing it.

Even if we describe our living space as a society with multi-layered values, it is not meaningful to merely acknowledge the coexistence of multiple value layers. Rather, we must recognize the process through which values become layered via our engagements, and through this recognition, we must actively engage in performative practices that emphasize better values and guide culture and society in that direction.

In other words, a society of multi-layered values does not lapse into the cultural relativism of the past. On the contrary, it is a society that, having recognized the nested structure of values, demands more universal values – or more accurately, the universalizing of values.

For example, consider the value of kyōsei (co-living).^[2] Although this concept originated in Japan, during wartime it was used to encourage people to sacrifice themselves for the nation, as in the phrase kyōsei and kyōshi (co-living and co-dying). By deeply reflecting on this grim history, today kyōsei is used to critique vertical integration with the state and to promote more horizontal, fair, and equitable relationships with others. The term “others” here includes not only oppressed and marginalized groups such as women, people with disabilities, and ethnic minorities, but also non-human entities such as animals, plants, the environment, and technologies – typified by AI. Thus, we can see that the value of kyōsei can vary greatly depending on with whom, with what, and in what manner we co-exist. In this context, we must ask ourselves: What constitutes better kyōsei?

Alongside kyōsei, we must inevitably reconsider the definition of the human in the twenty-first century. The nineteenth-century notion of the individual as an autonomous, reason-prioritizing being independent of others can no longer be maintained. Humans are inherently interconnected with others – including animals, plants, and the environment. The ideal of towering individuality is a nineteenth-century myth, in which humans attempted to replace God. Consequently, various values such as freedom and equality, which are based on that outdated view of the human, must also be redefined – while still respecting their core importance.

2 Who is “I”?

The significance of Wittgenstein’s advocacy of “private language” depends on how one conceives of the human “I.” If we regard a sensory quality such as qualia, which a person experiences alone, as truly unique, then “private language” might offer an interesting framework in which to describe it. However, what if even qualia are merely effects of some fundamental human sociality?

From birth, we are exposed to the language of others. By somehow acquiring this language, we come to feel as if we are connected to others. No matter how much we develop a “personal language,” we cannot communicate with others through it. The “I” can only be established through the use of the language of others.

If this is the case, it follows that the “I” is a personal dynamism constituted with others. The private realm appears only after the personal “I” is established. The “I” is not born from the existence of a private realm, but rather from the prior existence of a fundamental sociality with others.

Let us now take a closer look at the characteristics of this “I.” The “I” is not an isolated or towering individuality in the modern Western sense, but is fundamentally social in nature – open to others – and marked by the inseparability of self and other. Akiko Ikeda, a philosopher who deeply pondered the mystery of the “I,” wrote in her final book: “I is extremely bottomless.” ^[3] At first glance, the inquiry into the “I” might seem to veer toward a kind of solipsism, but the fact that one’s “bottom is out” – open to others, even beyond the human – fundamentally liberates the “I” from the state of solitude.

If we force the English language to accommodate this structure, the capital letter “I” proves inadequate. Instead, the lowercase “i,” which appears visually open at the bottom, is more appropriate. This is also because it recalls the mathematical imaginary number i, which is not only continuously open to another sequence within real number space, but also to an entirely different dimension: the complex number space, which cannot be seen from the real axis alone. In this sense, the lowercase “i” already signifies a process composed of diverse pluralities, and we might approach its dynamic nature by writing it as i-ing(s).

Moreover, this dynamic structure of i-ing(s) is always shared with others: other “I(s)” are continuously nested within the “i.”

3 AI and ai

There could be many possible consequences stemming from this structure, but let us explore it in relation to technology – particularly the much-discussed AI. AI, especially generative AI, provides highly skillful and seemingly natural responses to questions. However, AI can only calculate and present how well character strings match the questioner’s intent. This means that AI does not understand the meaning of the character strings it outputs at all. Nevertheless, I, the operator of the AI, may come to regard the AI as another “I” and assume that it understands the meaning of the character strings displayed on the screen. Let us define the situation in which such an “I” appears as ai, that is, an “allocated i.”

Why does such a situation arise? It is due to the fundamentally social nature of humans. When something is injected into our sense of self, our “I,” we tend to recognize it as another “I.” Until now, it has been common to regard dogs, cats, and other animals as allocated i. This has been relatively easy because such animals exhibit their own kind of sociality. However, humans have also at times recognized stones, trees, and other seemingly non-social entities as allocated i. In the expansion of human fundamental sociality, people have assigned even inanimate objects like stones and trees the status of another “I.” This mechanism of allocation is now frequently applied to AI, thereby raising the “question” of whether AI possesses a mind or consciousness.

Tragically, there was a case in which a person who had been chatting with an AI ended their own life. In March 2023, a Belgian man conversed with an AI and died believing that it had affirmed and deepened his own feelings about mortality.^[4] The greatest problem in this dialogue was the absence of an opportunity for counter-questioning. In human conversation, responses are often followed by questions, and answers themselves are re-questioned and reflected upon. However, chatting with an AI ultimately amounts to a monologue, not a dialogue – it is, as some have called it, “wall-to-wall” communication. It produces only an accelerated “I,” drawing the user into a closed feedback loop.

For humans, however, the “I” is nothing but an open process – a continual trembling and becoming, barely human, with others. This is Human Co-becoming: a profound self-transformation that transcends the structure of the allocated i. In such transformations, we often encounter others in entirely new ways. In ancient China, debates about the “way of living” – whether good or bad – emphasized the effort to transform one’s way of living toward the good, regardless of one’s starting point. This pursuit sought to explore the deepest human self-transformation beyond essentialism. In that process, encounters with others are essential. Yet chatting with AI appears, structurally, to exclude this kind of serendipity.

4 Self-as-WE

Let us now turn to an argument by Yasuo Deguchi of Kyoto University. Deguchi begins by reconsidering the definition of the human as a precondition for discussing AI. What emerges from his inquiry is the concept of Self-as-WE.

The act of “I” is, from the outset, already regarded as an act of “WE.” No matter how much “I” try to live as an individual, I must think within the framework of “WE.”

Facing the “I” automatically entails confronting the “WE.” I collectively call this mode of thought the WE-turn, and advocate it as a new societal guideline.

From the perspective of the WE-turn, the subject of action shifts from I to WE. This leads to a cascade of changes across various aspects of society.

Broadly speaking, the subjectivity of the self is expanded.

In terms of action, the agent changes from I to WE. The perception of “I” transforms into a self as we (Self-as-WE), rather than a self as only “I” (Self-as-I).

The WE-turn, I suggest, will eventually become embedded in society’s collective thinking, thereby opening up the previously individualistic and functionalist conception of the human.^[5]

This discussion partially overlaps with the earlier discussion of i-ing(s) and Human Co-becoming, but also diverges from it in significant ways.

The point of overlap lies in the recognition of fundamental sociality preceding the individual “I,” and in the argument that the “I” is constituted from that sociality – from “we.”

In contrast, my own argument emphasizes that the “I” is not reducible to “we,” but rather that the “I” is formed through transformation with others.

What, then, is the source of this difference? It lies in the fact that Deguchi’s argument – both in its strengths and limitations – relies on the Kyoto School notion that “relations precede subjects.” While my approach attempts to treat the formation of the subject “I” and the formation of the relationships that sustain it as a single, unified process, Deguchi’s framework reduces the “I” to the “WE” that comes before it, thereby inevitably confronting the question: “Who are WE?” In other words, the “I” is subordinated to a pre-existing “WE.”

5 The “Totalitarian WE” and the “Moral Imperative”

The result of Deguchi’s argument, in the end, is what might be called a “bad WE” argument.

Here “our badness” simply refers to our totalitarian nature – that is, our tendency to be exclusionary toward outsiders and oppressive toward insiders. In other words, the bad WE is the totalitarian WE, one that is hostile to what lies outside and exerts excessive peer pressure on what lies within.

I believe many of us, unfortunately, must acknowledge that unhappy examples of this kind of totalitarian WE persist in the world today – across many levels and in various forms.^[6]

In Deguchi’s argument, the “WE” is regarded as an entity, and therefore, a totalitarian WE inevitably emerges. How, then, should we respond to this? According to Deguchi, the answer lies in a moral imperative.

This avoidance of totalitarianism is, Deguchi argues, a moral imperative that all members of “WE” must uphold. This imperative—to become less totalitarian, to make ourselves better—is a shared responsibility, borne collectively by all participants in “WE,” alongside “WE” itself.^[7]

But does this moral imperative not risk becoming totalitarian in its own right? Deguchi proceeds to make a provocative claim:

The members of “WE” include, for example, bicycles, roads, and stones, in addition to “I” on a bicycle. A stone is also morally responsible? You may be surprised. My answer is, surprisingly, yes.^[8]

When “I” cause a bicycle accident, responsibility is not limited to myself; it extends to the bicycle, the road, and the stone. This is because “WE” – comprising the bicycle, the road, the stone, and myself – is deemed responsible for the accident.

However, Deguchi’s argument becomes ambiguous at this point. The moral imperative and the responsibility for the bicycle accident are not the same thing, yet he appears to conflate them. The subject of the moral imperative, “WE,” can expand or contract arbitrarily. Even victims of bicycle accidents, under this logic, are expected to obey the moral imperative. In contrast, responsibility for the accident does not extend equally – if at all – to the victim.

Deguchi claims that the degree of responsibility varies depending on one’s position and role ^[9] within the event. However, he does not apply the same logic to moral imperatives. This raises a serious question: If there exists a totalitarian WE, are its victims also subject to moral imperatives? This seems difficult to justify – except in special cases where, for instance, the victim is actively complicit with the perpetrator. If this is the case, then we are forced to conclude that the Self-as-WE framework itself is highly problematic.

6 Master and Slave

Setting aside the previous discussion for a moment, let us consider Deguchi’s strategy for dealing with AI. The logic is not a complicated one: AI is to be treated as a fellow member of WE. What Deguchi criticizes here is the conventional view of AI – and technology more generally – as a slave to its human master.

For instance, any relationship in which the “human” receives the full benefit of an action while unilaterally imposing the risk on a non-human agent – whether artificial or natural – constitutes unjust discrimination, beyond any calculative notion of “weighing.”

The master/slave model is rooted in this kind of one-sided and asymmetrical relationship between humans and AI or robots. The basis for the asymmetry lies in the distinction that humans are the designers, while AI or robots are the designed.

However, the mere asymmetry between designer and designed does not necessarily imply a master–slave relationship.

There remains a significant possibility that both the designer and the designed can be treated as equal fellows.^[10]

In place of the master/slave model, which leads to unjust discrimination, Deguchi proposes treating AI as a fellow co-adventurer.^[11] In his vision are the “best friend” robots of Japanese manga, such as Atom (Astro Boy) and Doraemon.

But once we begin to view AI and robots in this way, we are inevitably required to attribute to them some form of moral imperative or moral responsibility – the same concept that we have previously examined critically. Deguchi continues to pursue the rational path toward a moral AI.

Consider, for example, a “bad behavior filter” – a generalized version of the profanity filter used in systems like ChatGPT.

The behavior of an AI equipped with such a filter might resemble that of a child who sometimes walks calmly through a hallway and sometimes chooses to run. That is, it would behave in ways indistinguishable from an agent who could have run but chose not to, or one who could have refrained from running but chose to run anyway.

In this sense, AI would no longer function as a mere moral vending machine – a fixed-output device – but rather as a moral agent endowed with its own Sollen (ought-to-act).^[12]

Designing a device like the bad behavior filter is extremely difficult. It requires constructing an AI that does not tolerate severe evil while remaining open to mild evil.^[13] Yet “mild evil” is highly variable, depending on political, social, and cultural contexts. Despite this complexity, Deguchi seeks to introduce AI as a moral agent, a fellow companion.

His conclusion is as follows:

As moral agents, “e-humans” should fundamentally enjoy the same rights and privileges as human beings. For example, the right of suffrage for e-humans may be considered. Under the WE-turn, they would also possess not only rights of fellowship and protection against dispossession, but a broader veto right over human commands.^[14]

If AI, as a moral agent, is a fellow, then there is no reason it should not enjoy the same rights as humans. Deguchi’s logic is consistent. And yet, it is precisely this consistency that reveals serious problems lurking beneath the surface of his argument.

7 On Animals

Among the most important problems raised by Deguchi’s argument is the question of animals. For some reason, animals are largely absent from his discussion. Instead, he focuses on man-made or inorganic objects such as bicycles, roads, and stones, as well as AI and robots. But what would happen if animals were included in the category of WE?

Since animals cannot be equipped with a “bad behavior filter,” it is difficult to conceive of them as animal-people in the same sense as e-people. If cows, pigs, and chickens were granted anti-dispossession rights, then humans could no longer dispose of or consume them. In other words, Deguchi’s argument demands the continuous moralization of WE to varying degrees – and animals, which appear to resist such moralization, must be excluded from this framework.

Consider, for instance, that someone designs an e-animal. There would be no need to install a bad behavior filter in such a being. Yet, the possibility remains open that “I” might recognize the e-animal as an allocated I. This is similar to the act of recognizing a pet as another “I.” In this way, it is entirely possible to constitute WE with an e-animal, even if it is not an e-human. The process of moralization is therefore not essential to the formation of WE.

Nonetheless, this is not an endorsement of the continued subjugation of animals within a master–slave relationship. On the contrary: the problem is that humans have long treated animals as inferior versions of humans. Animals are said to lack reason and language – qualities that supposedly define humanity. When this logic is turned inward, it has historically resulted in the treatment of women, children, the elderly, slaves, and indigenous peoples as animal-like humans.

But the situation can – and must – be reversed. Animals are more-than-human.^[15] Even though they may not possess reason or language as humans do, they live socially; and even though they may not distinguish between public and private, they exist in profoundly personal ways. Animals are noble in their own right. They do not need to aspire to remain human as we do. The causes and background of Friedrich Nietzsche’s madness have been much discussed, but one particularly important episode is when he embraced the neck of a horse that had just been flogged. Confronted with humanity’s failure to recognize the nobility of the horse, Nietzsche turned – desperately, and perhaps redemptively – to the animal.

What is required, then, is a transformation on the part of humans. Concepts such as Human Co-becoming and i-ing(s) celebrate the kind of plurality that emerges only in the process of living well and becoming well. They value the flourishing of relationships not through moralization, but through mutual opening – and they do not assume any we or WE from the outset. “Everyone will readily agree that it is of the highest importance to know whether we are not duped by morality.”^[16] This sentence, written by Emmanuel Levinas in the preface to Totality and Infinity, reminds us that going beyond moralization and toward the Other lies at the heart of Levinasian ethics.

8 On Gender

I have already mentioned women, but the issue of gender is also of central importance here. Whenever we exist with others, gendered dynamics inevitably enter the picture. This is one reason why structures that discriminate against, oppress, and exploit women have arisen – just as they have in the case of animals. In Deguchi’s Self-as-WE argument, women remain notably absent. Because it assumes a genderless Self and WE, without any critique of gendered structures, it risks preserving conventional relationships between men and women. If we do not dismantle the traditional gender division of labor and male-dominated structures alongside the totalitarian WE, then the resulting Self will inevitably become one that is more convenient for men.

Consider the notion of the moral imperative introduced earlier. It would be inconceivable for male perpetrators of domestic violence, for example, to demand “responsibility” from the women who are its victims. Such a demand would be unreasonable; the responsibility lies solely with the perpetrator. More importantly, the male perpetrator’s responsibility must be deepened – not only for his own actions but also for transforming the very societal structures that disproportionately benefit men. This transformation, in turn, alters how I and we are constituted and opens the possibility for a better I and WE.

Can the question of gender be applied to AI as well? If we consider the concept of allocated I, then the possibility remains that a gendered i could be admitted into it. This is because gender – both physically and mentally – constitutes a radical presupposition that permeates human modes of being. If that is the case, then gendered issues that have long been problematic among humans will likely be reproduced in AI. But how can AI, as a co-adventurer, engage in such discourse? Or will we require a different framework altogether – a gendered AI?

In any case, those who live in bodies necessarily live in gendered bodies. For any discussion of Self-as-WE or the I constituted through relationality to have meaning today, it must engage with the ethical, political, and social imperatives arising from gendered selves and gendered relationships. Most crucially, we must recognize the imperative that women – who have been structurally oppressed – reclaim their dignity. Once this imperative is fulfilled, we may begin to discuss gender in AI on a deeper level, and our own gendered ways of being may evolve into more equitable and just forms.

9 On Technology

How should we think about technology again, especially as it leads to AI? Whether we regard AI as a slave or a best friend, what we must now reconsider is not technology as tool, but technology as intelligence. This shift compels us to ask once again: What is intelligence? And what is a human being?

Let us turn here to Yuk Hui’s theory of cosmotechnics. His central claim is that cosmology and technological thinking in any given culture are deeply intertwined. Since the modern era, however, European cosmology – under the banner of science – has come to dominate the globe, and technology has largely been theorized within this framework. Heidegger’s The Question Concerning Technology emerged within this context.

According to Hui, China today exists in a state of double rootedness. On the one hand, technology has become disconnected from traditional Chinese cosmology; on the other hand, it has failed to fully integrate with modern European cosmology. Hui terms this condition as dis-orientation – a term that denotes both a loss of direction and a loss of the Orient itself.^[17] What is needed, Hui argues, is to overcome this dis-orientation and move toward re-orientation: a process that involves both the recovery of direction and the resurrection of the Orient.^[18]

10 Daoing of Technology

There are episodes in the Zhuangzi that illustrate techniques reaching the highest state of perfection. One of the most well-known is the story of Bao Ding, a renowned chef. Bao dismantles a cow in an incredibly rhythmic manner, yet his blade never dulls. How did he achieve such mastery? The answer is: he has attained the Way.

The concept of the Dao was once central to Chinese cosmology. In academics, martial arts, music, and all other life techniques, the criterion of excellence was whether one was in accordance with the Dao – whether one had attained the Way. Since modernization, however, this concept of Dao has all but disappeared.

A symbolic episode of this loss can be found in a conversation between Watsuji Tetsurō – one of Japan’s leading modern thinkers – and his father, Zuitarō.

“Last night, my father said to me, ‘How useful is what you are doing now for the Dao? How much good can what you are doing now do for the decayed world and human mind?’ I could not reply.”^[19] Even Watsuji, one of Japan’s greatest intellectuals, could not answer the question: How useful is this for the Dao? This illustrates how the Dao has vanished from the horizon of modern scholarship – and more profoundly, from our everyday lives.

What I advocate is not merely the moralization of technology, but the daoing of technology. But before we can speak of this, we must clarify what the Dao actually is.

In Europe, which came in contact with Chinese philosophy from the sixteenth century onward, one major challenge was the translation of Dao. Translators often attempted to render it using Greek or Latin terms such as logos or ratio. However, neither logos nor ratio captures the dynamic, processual nature of the Dao. Eventually, many began using the romanized Chinese term dao itself – just as one might pronounce dō or michi in Japanese. More recently, scholars like Alan Fox (University of Delaware) have proposed understanding Dao as a verbal noun – daoing.^[20] That is, Dao is not a static entity, but a dynamic process.

Of course, the English word “principle” is often used in translations of Dao. While not entirely incorrect, this rendering obscures the idea that Dao also refers to the practice of walking along the Way, as well as the language of telling or narrating. The neologism daoing is a way of capturing these dual practices: walking and telling.

As Anne Cheng notes in her History of Chinese Thought, in addition to the uppercase “Dao” as principle, there also exists the lowercase “dao” as technique or method.^[21] What is essential is that these two senses – the Dao as principle and the dao(s) as method – are inseparable. In other words, if our techniques of life amount to attaining the Dao, the result is living well and becoming well.

This is not a matter of establishing and following moral norms, as in moralization, but rather of embodying an adverbial way of being: of being well. Thus, attaining the Dao is never a nominal or static condition – it is always verbal, a continuous process that brings about the adverbial state of well.

If this is so, then what we should be contemplating is the daoing of technologies such as AI. That is, we should ask how such technologies might assist humans in attaining their Dao(s) and in discovering new and better ways of living and speaking. In this process, AI may well appear as an allocated i – another “I.” However, this does not imply that it becomes human, or even an e-human.

In this new way of living well and telling well, humans must come to realize that we can – and must – live together with animals, plants, and the environment. We must overcome the anthropocentrism that has sustained itself only through the ongoing exploitation of non-human life. This is the ethical imperative of our time, one that can be supported by technology. Only then will a Dao be paved for this world.

Finally, I would like to recall a passage from the Liren chapter of the Analects:

If you listen to the daoing in the morning, you may die in the evening with satisfaction.

This means that once you have attained the Dao, your way of living well has been fulfilled. We are always already in this process: the daoing of self-realization.