An Adherence-Based Redescription of "Jibungoto-ka" (Treating Matters as One's Own) in Organizations: What AI-Accelerated Knowledge Work Requires of Organizations

Kengo Tomita
Institute of Technology, Shimizu Corporation

English access translation · edition 1.0.0 · September 6, 2026
Translated from the Japanese T4, preprint version 1. The Japanese original remains authoritative; this translation does not replace it or constitute a new paper.
Canonical scholarly source: https://doi.org/10.31235/osf.io/495wg_v1

Abstract — translated from the Japanese

In Japanese organizations, jibungoto-ka is demanded across domains, yet the concept remains underdeveloped. What the failures have in common is the absence of an explanatory language that conveys why an activity is undertaken. Medicine encountered an analogous problem earlier, as a shift from obedience to instructions (compliance) to active participation grounded in an understanding of their meaning (adherence); medication counseling was one implementation of that shift. This article transfers that structure to organizations. It redefines jibungoto-ka as a combination of undertaking—holding an internally endorsed purpose within a relation of giving reasons to specific others—and embodiment, its integration into practice. The relational core of undertaking is non-transferable answerability, which requires reciprocity. Unless the party asking also stands ready to answer, the setting for undertaking does not come into being. The asymmetry between medicine and organizations arises not from their domains but from their pathways of connection to encounters. From this, the article derives a two-pathway model: a forward pathway in which an intention is communicated and undertaken by employees, and a reverse pathway in which employee-originated orientations enter a formal process of consideration and the organization gives reasons for adoption, deferral, or termination. Chronic failure results from pressing only the forward pathway while lacking the reverse one. The more generative AI accelerates the externalization and combination of knowledge, the more undertaking and embodiment become the rate-limiting steps. The prescriptions offered are the design of enabling conditions, including encounters; accountability for stopping decisions; and observation through leading indicators.

Keywords: jibungoto-ka; adherence; answerability; generative AI; organizational change

1. The problem

Japanese source: §1

1.1 An underdeveloped demand: “jibungoto-ka”

Japanese source: §1.1

In Japanese organizations, jibungoto-ka is demanded across domains. In disaster preparedness, it is used in contexts that seek to have residents regard flood risk as their own problem [1]; in local government planning, to have residents undertake sustainability goals as their own concern [2]; in safety management, to have people on site observe rules as their own [3]; and, in organizational change and AI adoption, as a term calling for employees' active participation. Research on psychological ownership and the sense of being personally involved has also accumulated around the term [4] [5]. What those studies measure, however, is a state—a feeling or awareness of ownership. Their object of measurement differs from the relational side of undertaking defined later in this article: the standing to give reasons to specific others. Neighboring fields also have bodies of research. Self-determination theory addresses the conditions under which externally given goals are internalized as autonomous motivation [25], while research on voice and silence has addressed the conditions under which proposals and concerns reach higher levels or are blocked [26] [27]. Yet none of these simultaneously addresses endorsing a purpose as one's own reason for judgment, the standing to answer to specific others, embodiment in practice, and a pathway through which the organization returns reasons in response to employee-originated orientations. At least, an integrated framework that places these four elements together and connects them to measurement and intervention has not been presented in a sufficiently developed form. Even the variation in spelling—自分ごと化, 自分事化, and 自分ゴト化—reflects the concept's immaturity. The demand is strong, but what is being demanded remains unclear.

1.2 An analogous problem through which medicine has already passed

Japanese source: §1.2

The point of reference for this article is a transition that medicine has already undergone. From the 1990s onward, medicine advanced a conceptual shift from passive obedience to instructions (compliance) toward active participation grounded in an understanding of their meaning (adherence), which was formulated internationally in the World Health Organization's 2003 report [6]. One implementation supporting this transition was the explanatory practice of medication counseling: a procedure that explains why a treatment is necessary and, through understanding, connects that explanation to active conduct. Organizational jibungoto-ka implicitly calls for an analogous transition. Employees are expected not simply to obey instructions, but to participate actively on the basis of understanding. In organizations, however, an explanation corresponding to medication counseling—a language that conveys, in structured form, why something is to be done—has not been sufficiently systematized or implemented. In many cases, the instrument available to organizational initiatives extends only to shūchi: one-way notification of a settled policy. Initiatives confined to notification remain on the pre-transition side, using the instruments of compliance to seek obedience.

1.3 A cross-domain structure: the absence of explanatory language

Japanese source: §1.3

The structure in which an absence of explanatory language obstructs uptake has been identified separately in individual domains. Research on procedural justice shows that the amount and content of explanations, and the way they are communicated, affect people's sense of fairness and their acceptance of judgments [8]. In safety culture, policy programs are also working to measure and visualize people's understanding of, and acceptance of the reasons for, initiatives, although the methodology remains under development [9]. The domains are dispersed, but the form of the diagnosis is shared: uptake fails because a language for explaining why has not been systematized. The absence of explanatory language, as used in this article, does not mean that individual explanations are wholly absent. It means that a language for conveying why in structured form has not been sufficiently systematized and implemented within organizational arrangements. This article attempts to fill that cross-domain gap with a single framework.

1.4 Why this redescription is needed now

Japanese source: §1.4

The difficulty of jibungoto-ka is longstanding. Why, then, is this redescription needed now? The reason is generative AI. Knowledge conversion comprises four processes: bringing knowledge out into explicit form (externalization), combining it (combination), sharing it between people (socialization), and allowing it to become ingrained in an individual's body (internalization) [10]. Generative AI directly accelerates externalization and combination. Socialization and internalization depend heavily on human relationships, practice, and bodies, and are less readily substituted or shortened in the same way. As externalization and combination become faster, the rate-limiting step—the slowest step that determines the speed of the whole—moves toward socialization and internalization. In this article's vocabulary, organizations unable to sustain undertaking and embodiment increasingly encounter their bottleneck at that very point. Poor adherence has long existed in organizations, but in the AI era it comes to the foreground as the rate-limiting step determining organizational speed. That is why an old problem now needs to be given a language anew.

1.5 Contributions of this article

Japanese source: §1.5

This article makes five contributions. First, it provides a conceptual framework that integrates scattered, localized studies of jibungoto-ka by transferring the structure of medical adherence. Second, it redefines jibungoto-ka as a combination of undertaking and embodiment, placing answerability at the relational core of undertaking. Third, it identifies reciprocity—the asking party also taking the standing of one who answers—as a condition for answerability to be established in organizations. Fourth, it shows that the asymmetry separating medicine and organizations arises not from properties unique to their domains but from pathways of encounter, and presents a two-pathway model of organizational adherence: the process by which jibungoto-ka comes into being when expressed in this article's framework. Fifth, it derives operational principles: the design of enabling conditions, including encounters; accountability for stopping decisions; and observation through leading indicators.

2. Conceptual framework: adherence and the structure of undertaking

Japanese source: §2

2.1 The core of medical adherence: the locus of legitimacy

Japanese source: §2.1

The shift from “compliance” to “adherence” in medicine has not been a question of the degree of obedience—how faithfully the patient follows instructions. Its core lies in a movement in the locus of legitimacy. The term compliance presupposed an arrangement in which the authority to determine the correct treatment lay with the physician, and the patient followed the physician's instructions. Adherence replaced this with an arrangement locating legitimacy in agreement between the patient and healthcare professionals. When the World Health Organization's definition refers to behavior corresponding to agreed recommendations, the key is the agreement that precedes correspondence, not the rate of correspondence [6]. The patient understands what the treatment means and participates in it as a matter of their own judgment. This is the substance of the transition.

One implementation that supported this transition as practical work rather than exhortation was the explanatory practice of medication counseling. Medication counseling explains the mechanism by which a medicine works on this body, the risk of not taking it, and the effects to be expected if it is taken. What these have in common is that each conveys a “why,” rather than issuing an order to “comply.” In situations where explanation alone has difficulty securing the actual taking of medication—an act that only the patient can perform—medicine has used designs such as directly observed treatment (DOTS). Three features can be abstracted from the design of DOTS: the impossibility of proxy performance, direct on-the-spot confirmation that medication is taken, and the bundling of several elements rather than a single intervention. This three-part division is not the World Health Organization's official classification; it is an abstraction made by this article for transfer to organizations. These features will be used in the subsequent discussion.

The scope of the transfer should be limited at the outset. This article transfers to organizations not the entirety of medical adherence but three structures: a shift in the locus of legitimacy, the implementation of explanation that conveys why, and a design that protects processes that cannot be performed by a substitute. Patients and employees do not face the same urgency of interests, contracts governing their relationships, or freedom to refuse. What is analogous is the structure of the transition, not the conditions in which the parties find themselves. This difference itself is addressed directly in Sections 4.1 and 4.2. A further limitation is necessary. Medicine is cited as a prior example not because of its level of success in achieving adherence. Medication nonadherence remains an unresolved problem in medicine, and empirical research also shows that providing explanation or information alone does not change medication-taking behavior [11]. Medicine is ahead in having shifted its framing of the problem from obedience to agreement, and in having established a research cycle comprising scales for measuring adherence [7], analysis of influential factors, and evaluation of interventions. What this article imports is that transition and that research cycle.

2.2 Translation through Mekiki: importing two distinctions

Japanese source: §2.2

There have been scattered attempts to bring medical adherence theory into organizations, but without a common vocabulary the analogy has remained a superficial metaphor. As a translation layer, this article imports two distinctions from the Mekiki framework [12] and its measurement axis [13].

The first is the separation between the substrate of expertise (specification) and the barriers to converting it into output. Performing a task requires an accumulation of domain-specific judgment: what determines which procedure to apply in which situation. This is the substrate. It is not the documented procedure itself but the person's judgment that determines how to use the procedure. Converting the substrate into an actual artifact involves other barriers: accessing information, understanding it as a structure, and allowing it to become ingrained in practice. What generative AI has dramatically reduced is mainly these barriers on the conversion side, not the substrate itself [12]. The second distinction is between factual recognition (Sein) and value judgment (Sollen). Describing what is the case and judging what ought to be done belong to logically distinct layers; the latter cannot be derived from the former. AI surpassing humans in factual description and value judgment remaining on the human side are separate questions [13].

With this vocabulary, what medication counseling was doing can be rewritten in structural terms. Explaining mechanisms hands the patient the substrate behind the act of treatment; explaining risks and effects shows that the judgment whether to take a medicine concerns the patient's own Sollen. Medication counseling is an explanation that articulates and hands to the patient the substrate and the locus of judgment. An organizational explanation of “why this work is necessary” can have the same structure. It can, but at present that explanatory language has not been sufficiently systematized on the organizational side.

2.3 The central proposition: what is jibungoto-ka?

Japanese source: §2.3

The central proposition of this article can now be stated. Jibungoto-ka is neither the state of knowing management's intention nor the state of complying with it. It is a combination of undertaking and embodiment.

The internal side of undertaking is to endorse an externally given intention as one's own value judgment (Sollen), that is, as one's own reason for judgment. This article calls the direction of that value judgment an orientation. When management communicates “why this target,” an employee who has heard and understood the explanation still remains at the stage of knowing management's Sollen. An internal change first occurs when the employee endorses it as their own Sollen. The important point is that the origin of an intention and undertaking are separate matters. Even if a goal was given, once I undertake it, I become the one who answers for it. Its external origin has no bearing on my becoming answerable through undertaking. Embodiment means that the undertaken orientation becomes ingrained in practice and established in the body that carries out procedures. This does not occur by listening to an explanation; it occurs only through the person's actually doing it. In the terms of the previous section, it is the process in which the person crosses the innermost conversion barrier, the barrier to integration into practice; in the terms of Section 1.4, it corresponds to internalization.

Undertaking and embodiment do not constitute jibungoto-ka if either is present alone. Undertaking an orientation without bodily action remains merely verbal; bodily action without undertaking the orientation is simply acting because one was told to. When both are present, management's intention begins to live in employees' conduct.

2.4 The core of undertaking: answerability

Japanese source: §2.4

At the center of undertaking sits a relation independent of ability: answerability, the standing in which one must answer for a judgment to others entitled to question it. This is not a matter of how high or low the judging person's ability is. It is a relational fact concerning who answers for the judgment, and it cannot be transferred to someone else in the way an ability can.

This non-transferability becomes clear when the familiar institution of contract is separated into two layers. The contents of a contract—what is to be performed and what follows from breach—and the obligation to explain its performance are institutionally specified, guaranteed by third parties, and, where necessary, reassigned to a different party. This article calls that institutional layer accountability. By contrast, the act itself of entering into and undertaking the contract can be performed only by the person concerned. Once it has been undertaken, that person answers for the judgment; this is called answerability. What can be delegated to AI or third parties is the work and record-keeping, while accountability remains with the institutionally designated party or is reassigned. Answerability cannot be transferred away from the person who undertook the judgment. What this article places at the core of jibungoto-ka is the latter: the non-transferable standing of undertaking Note 1. The relation between the internal change described in Section 2.3—endorsing an intention as one's own value judgment—and this answerability should now be made explicit. The two are analytically distinct. The former is the internal side, endorsing a purpose as one's own reason for judgment. The latter is the relational standing of giving reasons for that judgment to specific others. Undertaking, as used in this article, means the state in which the former is held within the relation of the latter. An obligation to respond without internal change is an entry into the hollowing-out discussed in Section 5.1; internal change without anyone to answer to remains a private resolve. Neither alone establishes undertaking in an organization.

This standing is second-personal in that it is directed toward specific others entitled to ask, but it also requires a reciprocal structure. Building on Darwall's second-person standpoint [14], Matsumoto [15] argues that unless the asking party treats the other as an equal subject of judgment and also takes the position of one who answers, the engagement is no longer second-personal. Requiring a response only from the other while not responding oneself is merely a one-sided demand borrowing the form of a relationship. The setting in which answerability can stand is established only on reciprocity. This is where the article locates the condition under which undertaking is possible in organizations. If management requires undertaking from employees, management must also take the answering position in relation to employees' judgments. What this reciprocity requires of organizational design is developed in Sections 4 and 5.

Table 1 sets out the correspondences between medicine and organizations in the vocabulary developed so far.

Table 1. Correspondences between medical adherence and organizational jibungoto-ka

Note: The institutional and relational layers are presented together in one row, but this distinction is not simply one item alongside the other perspectives. It is an analytical axis running through the entire table. Answerability is the relational and non-transferable core of undertaking; the remaining perspectives describe its conditions of establishment and its manifestations.

3. Case analysis

Japanese source: §3

The framework developed so far will now be applied in depth to a single case. The aim is not to enumerate cases, but to demonstrate the procedure for mapping a phenomenon into this article's vocabulary. The leading indicators in Section 5 (Table 2) open a route to verification; this section provides a demonstration of application. The example primarily illustrates the embodiment component of the definition and does not empirically establish undertaking or the two-pathway model as a whole.

Case: manual measurement with ammonia detector tubes

This was a project in which the author was involved. At one site, a young staff member was asked to measure ammonia concentrations once a week using manual detector tubes. The initial purpose was data collection, nothing more than acquiring facts (Sein). An unintended by-product nevertheless emerged. The staff member read the color change in a detector tube, converted it into a numerical value, and connected it to their sensory experience on site. Through repetition of this sequence, the correspondence among color, number, and bodily perception became established in the staff member’s body. The construction manager subsequently recognized that the work had had an educational effect.

Read in the vocabulary of adherence, the case takes the following form. Had the measurements been replaced by automatic sensors, the data would have been obtained in the same way, but the same sensory and embodied learning would likely have been less readily produced. Unintentionally retaining the innermost conversion barrier—the barrier to integration into practice—turned the act of measurement itself into a pathway for accumulating the substrate. With this article's framework, the effect could have been intended at the design stage. It would have been possible to explain, “The purpose of this measurement is actually the accumulation of expertise through embodiment,” indicating the reasons for that orientation in advance and preparing the preconditions for undertaking.

What this case shows is that, with an explanatory language, the effect could have been intended at the design stage. The absence of explanatory language does not mean that effects do not occur: they can arise by chance. It means, however, that the organization cannot possess the effect as a reproducible design. What has been demonstrated here is not an exhaustive collection of cases but a procedure for mapping a phenomenon into the article's vocabulary. Readers should be able to recall corresponding experiences in their own organizations; repeating this mapping for those experiences is precisely how the framework is to be used Note 2.

4. Discussion: the path dependence of asymmetry and the two-pathway model

Japanese source: §4

4.1 Why is adherence more readily established in medicine and difficult in organizations?

Japanese source: §4.1

The preceding case illustrated the coexistence of repeated manual measurement and embodied learning, together with the possibility of intending the effect at the design stage if an explanatory language had been available. Why, then, were the explanatory and observational circuits supporting adherence established earlier in medicine, while they are chronically absent in organizations?

The first explanation concerns the difference in the locus of purpose at the point when treatment or work begins. In medicine, in many cases the patient has already internalized the intention “I want to get well,” and healthcare professionals can treat their task as connecting therapeutic action to that intention. In companies, by contrast, goals descend from higher in the hierarchy. Numerical targets such as sales and market share take the outward form of facts (Sein), but the choice that they should be pursued is a value judgment (Sollen). Moreover, that judgment belongs to management; a corresponding intrinsic intention does not automatically exist within employees. Employees do not necessarily begin with an equivalent of the patient's “I want to get well.” At root, jibungoto-ka is difficult in organizations because purposes come from outside.

4.2 The asymmetry lies in the pathway

Japanese source: §4.2

This explanation still treats medicine as special, however. Is medicine special because it is the domain of medicine? This article argues otherwise. The asymmetry is not domain-specific but pathway-specific.

Section 4.1 described the arrangement at the start of treatment, not the prior history from which the purpose arose. Tracing its formation changes the picture. Medicine appears to be a domain in which purposes are intrinsic because the structure of medical practice connects both parties to bodily encounters with the problem. Patients encounter, bodily, their own symptoms and constraints on their lives. Physicians and pharmacists encounter not an abstract task on a document but a suffering patient before them, bodily. Such encounters promote the crystallization of an orientation toward treatment within each person concerned. The purpose, then, was not intrinsic from the outset; it became internalized through encounter. Misunderstanding this leads to misunderstanding what organizations can do.

Purposes that a person treats as ends in themselves do not arise from nothing. This article posits crystallization from embodied experience as one important pathway of their formation. The hypothesis is not confined to speculation: four kinds of observational sources converge on the same form. First, in the cases collected by Nonaka and Takeuchi as examples of organizational innovation, developers' commitments are formed not in planning meetings but through direct contact with customers' and sites' difficulties [10]. Second, the publicly documented winners of a recent AI-assisted development competition were practitioners who held commitments before the tool appeared: a lawyer confronting delays in housing permits, an engineer responding to his own child, a cardiologist confronting post-consultation care, and a practitioner confronting the evaluation of road investment. For them, AI removed conversion barriers but did not change where their orientations came from [16]. Third, a study of civil engineers' boundary-crossing learning records a process in which a placement at a national research institute took engineers beyond their own organizations; joining surveys of disaster-affected areas and directly encountering peers at work in disaster response led to recognition of their industry's mission [17]. Fourth, the AI-assisted development case treated in the author's previous research also records the sequence in which an orientation formed from experience in the field steered AI as a tool [12]. The four sources share neither methods, subjects, nor periods, but share the sequential form of encounter first and crystallization of orientation afterward. They are not causal proof of this mechanism, but heterogeneous observations consistent with the hypothesis. The author has also formulated an analogous structure from encounter to crystallization as a design hypothesis in a separate paper addressing education [28].

The consequence changes the point at which organizational design acts. Purpose itself cannot be injected [25]. As the preceding sections have shown, undertaking cannot be forced through explanation or command. What employees encounter bodily can nevertheless be designed. What organizations can do is not inject purposes but design enabling conditions, including encounters. Hobbies, often placed in opposition to work, are not an enemy from this perspective. They are reserves of energy directed toward ends in themselves that have not yet been connected to organizational problems. The problem lies not in the existence of hobbies but in the absence of an open pathway directing that energy toward organizational issues.

4.3 A two-pathway model of organizational adherence

Japanese source: §4.3

Viewed from the organizational side, the hypothesis that encounters may promote the crystallization of orientation reveals two pathways of adherence running in different directions.

Pathway 1 runs forward. Management communicates an intention—why this target—and employees undertake it. This is the standard image of jibungoto-ka described in Section 2.3 and the pathway implicitly assumed by many initiatives intended to secure uptake. Yet it is structurally difficult for the reason described in Section 4.1: the purpose is not intrinsic to the employee. Pathway 2 runs in reverse. An orientation crystallized by an employee through encounter enters a formal process of consideration, and a specific authorized party gives reasons for adoption, deferral, or termination. What the organization necessarily undertakes here is not the orientation itself but its own judgment about how to treat that orientation. Where an orientation is adopted, the organization undertakes it as its own purpose, allocates resources, and works toward alignment with social demand. When the organization is said here to undertake or answer, this does not mean that an abstract corporation answers. It means that a specific institutionally designated office-holder or decision-making forum answers. Because the orientation is already intrinsic to the person by the time it enters Pathway 2's process of consideration—it has crystallized from encounter—this pathway can structurally bypass the internalization difficulty of Pathway 1. Pathway 2 is not, however, a universal bypass. Pathway 2 leaves its own design problems: whether the crystallized orientation aligns with social demand, whether resources should be allocated, and who answers if it is stopped. That design is addressed in Section 5.

This two-pathway model explains why initiatives for jibungoto-ka chronically fail. Many unsuccessful initiatives press Pathway 1 strongly, repeatedly calling on employees to “treat it more as your own concern.” The notification discussed in Section 1.2 is the typical institutional form taken by this kind of initiative. The actual binding constraint, however, often lies in the absence of Pathway 2. Research on voice and silence has accumulated knowledge about one part of this: the conditions under which circuits of expression and response are blocked [26] [27]. Encounters have not been designed in the first place; there is no setting in which orientations crystallize; and impediments remain that stop what is beginning to crystallize. The pathway being pushed is not the pathway that is blocked. Raising one's voice therefore does not make it pass through.

The encounters from which Pathway 2 begins are not limited to the organization. It is well known that the origin of 3M's Post-it lay in a developer's encounter outside work: frustration with bookmarks falling from a hymnal during choir activities [10]. Like the hobbies mentioned at the end of Section 4.2, orientations crystallized through encounters outside the organization are among the resources Pathway 2 brings into formal consideration and response. The crystallization row in Table 2 distinguishes whether proposals originate inside or outside work in order to translate this domain of application into observation.

The pair of pathways also translates the reciprocity established in Section 2.4 to the scale of an organization. Through Pathway 1, the organization asks employees to undertake; through Pathway 2, it responds to the orientations employees have crystallized and, where it adopts them, undertakes them in return. Initiatives that press only the forward direction without Pathway 2 take the form of a one-sided relationship: asking without responding. This is the organizational-scale manifestation of the distortion discussed as self-conceit in Section 5.

The two pathways have a prior implementation. The middle-up-down management described by Nonaka and colleagues is a form in which middle managers resolve contradictions between top management's ideals and the realities of the workplace [10]. Reread in this article's vocabulary, middle managers translate higher-level intentions into the language of the workplace and convey them downward (Pathway 1), while picking up orientations crystallized on the ground and conveying them to management (Pathway 2). A person occupying a role was thus mediating both pathways simultaneously. A role-based solution, however, reintroduces dependence on the exceptional middle manager as an individual. The operational principles addressed in Section 5 attempt to shift this mediation from individuals to design. This article does not attempt to replace Nonaka's theory. It attempts to decompose its mediating function into two pathways and translate them into a vocabulary of design and observation.

People have sometimes carried these two pathways, yet the pathways have not been explicit objects of design. Experienced managers have tacitly managed, through experiential knowledge and without naming them, the themes of this article: orientation and momentum, encounter, and answerability. As AI absorbs externalization and combination, however, the room for relying on tacit operation diminishes. Restating the shift in the rate-limiting step from Section 1.4 on the management side, a structure that could be run on experiential knowledge alone becomes something that will not function unless it is consciously designed.

This consequence also intersects, in the reverse direction, with a classical diagnosis of organizational responsibility. Discussions of the problem of many hands [18] [19] and the responsibility gap [20] have argued that both organizations and AI tend to diffuse the locus of judgment. As more hands become involved, and as judgment moves to machines, it becomes harder to see who answers. This article's framework opens the opposite possibility. Delegation that explicitly distinguishes factual recognition (Sein) and value judgment (Sollen) [13] structurally requires a record, each time, of the boundary between what was passed to the machine and what the person retained. Insofar as that boundary is actually drawn, AI adoption may instead make the locus of judgment visible. This reversal is not a claim about AI adoption in general. Adoption without the explicit decomposition will blur the locus, as the classical diagnosis predicts. The reversal arises only where the design of delegation includes that explicit distinction. Records may also become merely formal, leaving a discrepancy between the recorded respondent and the actual decision maker. Whether visibility has substance can only be checked by comparing the records with practice. This is both a consequence of the framework and a hypothesis to be subjected to observation. The “explicit locus of judgment” row in Table 2 is included to preserve these conditions and consequences in an observable form.

4.4 A limitation of scope and its reversal

Japanese source: §4.4

An important limitation must be introduced here. The problem this article has called “the difficulty of jibungoto-ka” may not be universal. Contract-defined employment, in which the scope of duties is specified in writing in advance—so-called job-based employment—does not explicitly ask for undertaking beyond that scope. Within this structure, jibungoto-ka as undertaking beyond the defined scope is less readily posed as a demand. By delimiting the scope of undertaking, the institution of contract avoids the hierarchical split in intentions. Japanese-style employment, by contrast, leaves the scope of duties ambiguous—membership-based employment—while calling for undertaking beyond that scope [21]. This article's problem is likely to become particularly salient within that Japanese-style structure, which asks for jibungoto-ka without contractually delimiting its scope.

This limitation can also be read in reverse. Contract-defined employment had been able to avoid undertaking because duties could be specified in writing in advance. What generative AI is absorbing is precisely the part of work that can be made explicit in that way. As contract-specifiable work moves to AI, the avoidance strategy of drawing boundaries by contract itself becomes less effective. Judgments and undertaking that cannot be written into a contract come to occupy a larger proportion of work. If so, the problem this article addresses—how to treat undertaking—will also converge upon contract-defined employment. Japanese-style organizations' long experience of this problem is both a weakness and a leading indicator of a problem that many organizations will face. There is another side to this limitation. An analogous failure—expecting valuable capacities as by-products of work without explicitly cultivating them—is also surfacing in education and research Note 3. The diagnosis may therefore extend beyond organizations.

5. Modes of failure and success: hollowing-out and 3M

Japanese source: §5

Undertaking has paired modes of failure and success. They will be considered in turn. These are not separate phenomena, but the same process of undertaking turning in opposite directions under different conditions.

5.1 Modes of failure: hollowing-out and self-conceit

Japanese source: §5.1

Under certain conditions, the demand for jibungoto-ka may become dysfunctional. This article calls this the hollowing-out of jibungoto-ka: the substance of undertaking falls away while the form of the demand continues to circulate. What drives this hollowing-out is a spiral in which value judgments are reinforced within the group without grounds supporting the judgments. This article extends to the group level a phenomenon described at the individual level: the self-reinforcement of confidence without supporting grounds [13].

A typical turn takes the following form. As jibungoto-ka comes to be treated as a value that must not be questioned within the organization, the statement “you are not treating it as your own concern” comes to be received not as a comment on the work but as an attack on loyalty to the organization. People then compete not over the validity of content but over formal behavior signaling loyalty. It becomes a contest in loyalty signals. Rational criticism of the jibungoto-ka policy is equated with betrayal of the organization, and the domain of value judgment is placed beyond discussion. This corresponds to the known structure in which a challenge to a sacred value generates moral outrage and halts rational consideration [24]. Because criticism is construed as betrayal, it strengthens defensiveness, which in turn invites further criticism; learning stops within the resulting cycle.

The decisive point here is the distinction made in Section 2.4. In a hollowed-out organization, a surrogate—an expression of loyalty—occupies the place of answerability. Actually standing answerable and merely appearing to do so become difficult to distinguish from outside. This is why Section 2.4's definition of answerability as a relational fact rather than an individual ability provides a basis for identifying the failure.

There is another mode of failure: the attitude Matsumoto [15] calls self-conceit, a relational distortion in which one requires the other to answer without taking the answering position oneself. When a leader demands jibungoto-ka from subordinates but does not receive their judgments as judgments made by equals, the reciprocity described in Section 2.4 is absent. The very setting in which answerability could stand on the subordinate's side then disappears. Without another who asks about one's judgments and receives one's answers, the standing to answer cannot come into being. Whereas hollowing-out replaces undertaking with a surrogate through group dynamics, self-conceit first destroys the conditions under which undertaking can be established through the structure of the relationship. These failures operate at different levels and complement one another. Where competition in loyalty signals circulates at the group level, the relational posture of asking without answering often removes the recipient of criticism.

5.2 A mode of success: three operations at 3M

Japanese source: §5.2

At the opposite pole are designs that have succeeded in fostering undertaking. As a well-known example, this article rereads a set of 3M's arrangements in its own vocabulary. 3M has been called a paradigm of bottom-up management, but the reading here reverses that picture. The weakness of bottom-up management is its dependence on individuals: success rests on the breakthrough capacity of a person whose orientation has crystallized. 3M's arrangements can be read as a design in which the organization takes over that dependence. If Nonaka's middle manager, considered in Section 4.3, is a mediator who deeply amplifies a few orientations through personal qualities, 3M's institutionalized practices are an apparatus that protects crystallization through a wide aperture without privileging particular people. The two are complementary. This reading does not claim historically that 3M's designers thought in this article's vocabulary; it redescribes the functions of publicly known arrangements from the framework's perspective. The important point is that 3M does not manufacture intentions themselves. As Section 4.2 showed, that cannot be done. What 3M does consists of three operations on the conditions in which intentions find expression.

The first operation removes impediments to expression. The 15% rule—a practice said to allow part of working time to be devoted to personally chosen problems without prior justification—and the in-house facilities supporting it remove the costs of starting and the barriers of resources. The second changes the allocation of accountability. Ordinary organizations require justification for starting something: the default lies on the side of not doing it. 3M's practice can be read as placing the default on continuation and shifting the demand for explanation to the side of stopping. In a well-known origin story, a developer continued development against an instruction to stop, and management later turned this into a policy of leaving the right person alone when absorbed in the work and trusting initiative [10]. The manager expressed the view that mistakes made by people were far less serious than the management error of directing them at every step. The design on the stopping side is not laissez-faire—“do not stop”—but an allocation: “whoever stops it must answer.” If an organization intervenes in and stops undertaking, accountability attaches to the intervention. The structure can be read as making failure itself less likely to trigger a demand for an individual to answer.

This second operation translates Section 2.4's reciprocity into institutional form. The organization that asks for undertaking moves to the answering side regarding its own judgment to stop it. Borrowing the movement of “looking back” in the etymology of respect, respicěre, as discussed by Matsumoto [15], it is a design in which the organization looks back, at the level of institutionalized practice, at the employee's gaze of undertaking. Two misunderstandings should be prevented. First, what is protected is not a privilege for individuals with an established record. The Post-it began from a weak adhesive regarded as a failure and the private difficulty of keeping bookmarks in a hymnal; at the point of starting, it had no record of achievement to present [10]. The protected object is not a person but a crystallization that has yet to show results, from its beginning through its early continuation. Second, this is not immunity from answering. It is not that no one need answer, but that the obligation to answer is moved from the initiating side to the stopping side. In the distinction of Section 2.4, accountability here is an allocation at the institutional layer. Institutions cannot create non-transferable answerability, but they can provide the setting in which it stands.

This design has an earlier analogue in medicine. When a pharmacist finds a questionable point in a prescription, the pharmacist must not dispense until it has been checked by inquiry to the prescribing physician (Article 24 of the Pharmacists Act), and the physician receiving the inquiry has a duty to respond appropriately (Article 23(2) of the Rules for Medical Care by Insurance Medical Institutions and Insurance Physicians). The provisions establish both the obligation to ask and the obligation to answer; in medicine, the reciprocity in which the party making a demand also takes an answering position is implemented at the level of law. No law binds organizations in this same form. Yet this legal arrangement is an example of institutionalizing reciprocal reason exchange by pairing obligations on the asking and answering sides. From it, this article posits a hypothesis: an obligation repeated institutionally may eventually support the formation of a practical stance. The institution need not be law. 3M's example suggests that institutionalized practices within a company may also suffice.

The third operation aligns incentives. A rule requires products introduced within the last few years to account for a certain share of sales, and gives management interests aligned with that requirement [10]. Allowing intermediate trials thereby aligns with organizational self-interest rather than depending on managers' virtue.

Taken together, the three operations constitute a single design: protecting the substrate and removing impediments. They avoid killing an orientation crystallized through encounter before it can develop, and remove factors that obstruct its expression—demands for justification, lack of resources, and premature termination. What this design guarantees, however, extends only to the volume of trials and to crystallized orientations not being killed off, not to the success of individual trials. This limit must be stated explicitly, lest successful cases alone be used to claim that the design is universally effective. An orientation crystallized through encounter aligns in advance with social demand to the extent that the person encountered represents a wider population. Patients may represent other patients, and customers may represent a market. Placing accountability on the stopping side is also a device for carrying out this check of alignment in a way that requires an answer.

Finally, it is necessary to articulate why these arrangements must be institutional. Verbal encouragement such as “take on challenges,” “treat it as your own concern,” and “you may go ahead on your own” cannot in principle substitute for institutions, for three reasons. First, initiative itself does not travel through forward transmission. Initiative begun on command is already no longer initiative. What can be conveyed in the forward direction extends only to the reasons for an intention; the expression of initiative is supported by the design of conditions. Second, verbal permission alone does not institutionally fix the relation of who answers for that permission. When a failure occurs, the individual who gave permission can withdraw by saying they do not remember doing so. A junior employee's repeated requests for confirmation before starting are not a lack of motivation; they are rational attempts to verify whether the answerability attached to the permission is real. Third, the manager granting permission is themselves bound by an order of accountability, and verbally protecting a subordinate's trial consumes that manager's personal virtue and courage. The 15% rule is established as an ongoing practice to address all three simultaneously. It does not command initiative; it specifies only the condition that beginning requires no permission. Unlike an individual, an institution cannot say it does not remember giving permission and cannot simply withdraw. Nor does the design depend on the virtue or courage of the protector. The explanatory language supplied by this article must also be distinguished from these words of encouragement. Explanatory language conveys the structure of why something is done and prepares the preconditions for undertaking. What secures expression is the design of conditions: language and institutions are complementary, not substitutes.

5.3 Observation through leading indicators

Japanese source: §5.3

Having distinguished modes of failure and success, organizations can shift innovation management from lagging indicators—sales—to leading ones. What is observed is whose orientation is crystallizing where, and the state of its drive: whether that orientation is actually being acted on and where friction arises.

A boundary must nevertheless be drawn around what may be measured. The objects of measurement are not the contents of values but the conditions of their expression and their behavioral correlates. The internal side of undertaking—whether an intention has really changed—cannot be measured directly. The relational side of answerability, by contrast—the configuration of who asks whom for reasons and who actually answers—is observable. The indicators in Table 2 belong to this relational configuration and to enabling conditions. Table 2 presents the observational framework while retaining that limitation. Measurement design, including each indicator's unit, data collection method, and inter-rater agreement, is outside this article's scope and remains a task for research implementing the framework.

One further observational bias should be stated. The indicators in Table 2 are designed to capture undertaking as it appears in crystallization and challenge. Undertaking, however, is not limited to the crystallization of new orientations. Holding the quality or safety of existing work as one's own reason for judgment and continuing day by day to stand answerable for those judgments—in other words, undertaking maintenance—also satisfies Section 2.4's definition. The standing can be identified by asking who answers for a decision to stop upon noticing an abnormality, or for a decision to keep observing a procedure. It simply does not take a conspicuous form and is less readily captured by indicators such as proposal counts. The absence of an indicator in Table 2 must therefore not be read as the absence of undertaking.

Table 2. Leading indicators of undertaking: candidates

Note: No row in this table measures the change of intention itself—the internal side of undertaking. The relational configuration of answerability, however, is observed through the rows on the quality of stopping and the explicit locus of judgment. The remaining rows capture enabling conditions and behavioral correlates.

6. Conclusion

Japanese source: §6

Jibungoto-ka cannot be produced directly by command or implanted directly through training. Yet it need not be left to chance. Redefining it as a combination of undertaking and embodiment draws a boundary between what organizations can and cannot do. They cannot inject intentions themselves, but they can design encounters in which employees bodily meet those who bear the problems. They do not stop an orientation that is beginning to crystallize without reasons; if they do stop it, the stopping side answers. Knowing that the internal establishment of undertaking cannot be measured directly, they observe the relational configuration of answerability and the conditions of its expression. What this article has presented is these operational principles and the two-pathway model behind them.

What medicine learned over a generation was that, unless people understand meaning and share a purpose, they will not actively participate even in their own treatment. Organizations need this recognition precisely in the AI era. Generative AI accelerates the externalization and combination of knowledge, but the subjects who undertake that knowledge and the settings in which undertaking develops—the socialization settings in which people face one another—are not accelerated in the same way. Rather, as AI takes over externalization and combination, the rate-limiting step shifts toward those settings, and the ability to sustain undertaking and embodiment comes to determine organizational speed.

Finally, the limitation of scope should once again be read in reverse. The problem of demanding jibungoto-ka without contractually delimiting the scope of undertaking is particularly likely to surface in Japanese-style employment. Yet contractual boundaries had been possible because duties could be made explicit in advance, and it is precisely the part that can be made explicit that AI will increasingly absorb. As contract-specifiable work moves away, judgments and undertaking that cannot be written into contracts come to occupy a larger proportion. Japanese-style organizations' long experience of this problem is a weakness, but also means that they have already experienced a problem that many organizations will come to face.

Those who know it are not equal to those who love it; those who love it are not equal to those who delight in it Note 4. What organizational notification can convey extends only as far as knowing. Those who love and delight do not emerge from notification alone. What this article has structured as a two-pathway model resonates with these words. Ever since Shibusawa Eiichi placed the Analects beside the abacus Note 5, Japanese organizations have kept these words at hand. What was missing was not the teaching but a vocabulary for translating it into organizational design. To preserve that experience together with such a vocabulary: that is the purpose of this article.

Original author notes

Note 1

In this article's terminology, accountability (説明責任) is assigned to duties of performance and explanation that are located at the institutional layer and can be transferred or reassigned between roles; answerability (応答可能性) is assigned to the non-transferable standing to respond that belongs to the person who undertook the judgment. The related distinction in measurement theory follows [13]. Darwall [14], however, uses accountability broadly in the cited book, including second-personal relations. This article's distinction is not a reproduction of that book's terminology, but an operational distinction of its own, separating the institutional and relational layers for measurement and design.

Note 2

An analogous example is found in education. During a visit to a university laboratory attended by the author, a student was beginning to judge that research might not suit them. In answering the student's question, the faculty member disclosed uncertainties from their own graduate-school years. This is an observation outside the organization of the structure discussed in Section 5: the party that asks also moves to the answering side.

Note 3

From education comes the argument that generative AI has exposed a failure to teach critical thinking explicitly [22]. In knowledge work, research reports that the exercise of critical thinking under AI use depends on users' self-confidence and their confidence in AI [23].

Note 4

The Analects, Book VI, “Yong Ye.” The Japanese reading follows Osamu Kanaya's annotated translation, Rongo (Iwanami Bunko, 2001).

Note 5

Shibusawa Eiichi, Gendaigoyaku Rongo to Soroban (modern Japanese translation by Atsushi Moriya; Chikuma Shobō, Chikuma Shinsho, 2010). In the chapter on taking an interest in one's work—Chapter 5, “Ideals and Superstition”—Shibusawa himself quotes these words from “Yong Ye.”

References — source-language records

The bibliographic records below are reproduced verbatim in the source language. Reference numbers preserve the original labels, rather than imposing a new order of first citation.

[1] 国土交通省，「水害リスクを自分事化し，流域治水に取り組む主体を増やす流域治水の自分事化検討会」とりまとめ（2023），https://www.mlit.go.jp/report/press/mizukokudo03_hh_001203.html

[2] 陣内雄次，ＳＤＧｓと地方自治体計画の自分事化に関する一考察，共生科学，13，72-83（2022）

[3] 庄司敬子，清水裕美子，堀下智子，武内寛子，安全行動に影響を与える要素の検討，あんけん，Vol.18，JR西日本安全研究所（2025）

[4] 福原康司，自分事化に関する探索的研究：心理的所有（psychological ownership）概念を拠り所にして，専修大学経営研究所定例研究会資料（2022）

[5] 神吉直人，就労者の「仕事に対する当事者意識」概念の検討，追手門経営論集，31（1），1-18（2025），https://doi.org/10.69274/105250601

[6] World Health Organization，Adherence to Long-term Therapies: Evidence for Action，World Health Organization（2003）

[7] 上野治香，山崎喜比古，石川ひろの，日本の慢性疾患患者を対象とした服薬アドヒアランス尺度の信頼性及び妥当性の検討，日本健康教育学会誌，22（1），13-29（2014）

[8] 今在慶一朗，手続き的公正要因としての説明責任と鄭重さに対する中心的・周辺的認知処理の影響：裁判での弁護活動を模したコミュニケーション実験，社会心理学研究，31（3），184-192（2016），https://doi.org/10.14966/jssp.31.3_184

[9] 特定非営利活動法人保安力向上センター，令和4年度産業保安等技術基準策定調査研究等事業（産業保安に関連する課題に対する新たな解決アプローチ推進調査）報告書，経済産業省（2023），https://www.meti.go.jp/meti_lib/report/2022FY/000160.pdf

[10] 野中郁次郎，竹内弘高（梅本勝博訳），知識創造企業，東洋経済新報社（1996）

[11] 櫻井秀彦，アドヒアランス研究の意義と現状，薬局薬学，14（2），83-90（2022），https://doi.org/10.32160/yakkyoku.ra.2022-3000

[12] K. Tomita，Domain-Native Development: A Mekiki Framework for AI-Assisted Knowledge Work，SocArXiv（2026），https://doi.org/10.31235/osf.io/cwkav_v1

[13] K. Tomita，Philosophy as Cognitive Assay: Measuring the Delegation Legitimacy Boundary in AI-Assisted Knowledge Work，SocArXiv（2026），https://doi.org/10.31235/osf.io/e9qw5_v2

[14] S. Darwall，The Second-Person Standpoint: Morality, Respect, and Accountability，Harvard University Press（2006）

[15] 松本大理，ダーウォルの二人称的尊敬と相互的尊敬について，山形大学紀要（人文科学），21（1），1-15（2026），https://doi.org/10.15022/0002001699

[16] Anthropic，Meet the Winners of Our Built with Opus 4.6 Claude Code Hackathon（2026），https://claude.com/blog/meet-the-winners-of-our-built-with-opus-4-6-claude-code-hackathon

[17] 宮原史，土木技術者の越境学習：民間企業から国の研究所への派遣に着目して，土木学会論文集，81（1），24-00203（2025），https://doi.org/10.2208/jscejj.24-00203

[18] D. F. Thompson，Moral Responsibility of Public Officials: The Problem of Many Hands，American Political Science Review，74（4），905-916（1980），https://doi.org/10.2307/1954312

[19] I. van de Poel，L. Royakkers，S. D. Zwart，Moral Responsibility and the Problem of Many Hands，Routledge（2015），https://doi.org/10.4324/9781315734217

[20] A. Matthias，The Responsibility Gap: Ascribing Responsibility for the Actions of Learning Automata，Ethics and Information Technology，6（3），175-183（2004），https://doi.org/10.1007/s10676-004-3422-1

[21] 八代充史，メンバーシップ型雇用管理とジョブ型雇用管理：ジョブ型雇用管理は日本に定着するか？，日本労働研究雑誌，755，34-43（2023）

[22] P. Soundar-Shah，A. Kestigian，We Have Never Taught Critical Thinking，Inside Higher Ed（2026）

[23] H.-P. Lee，A. Sarkar，L. Tankelevitch，I. Drosos，S. Rintel，R. Banks，N. Wilson，The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers，Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems，Article 1121，1-22（2025），https://doi.org/10.1145/3706598.3713778

[24] P. E. Tetlock，Thinking the Unthinkable: Sacred Values and Taboo Cognitions，Trends in Cognitive Sciences，7（7），320-324（2003），https://doi.org/10.1016/S1364-6613(03)00135-9

[25] R. M. Ryan，E. L. Deci，Self-Determination Theory and the Facilitation of Intrinsic Motivation, Social Development, and Well-Being，American Psychologist，55（1），68-78（2000），https://doi.org/10.1037/0003-066X.55.1.68

[26] E. W. Morrison，F. J. Milliken，Organizational Silence: A Barrier to Change and Development in a Pluralistic World，Academy of Management Review，25（4），706-725（2000），https://doi.org/10.5465/amr.2000.3707697

[27] E. W. Morrison，Employee Voice and Silence，Annual Review of Organizational Psychology and Organizational Behavior，1，173-197（2014），https://doi.org/10.1146/annurev-orgpsych-031413-091328

[28] K. Tomita，Decomposing Agency, Isolating Answerability: Cultivating What Cannot Be Delegated in AI-Assisted Learning，EdArXiv（2026），https://doi.org/10.35542/osf.io/hvbfe_v2

English abstract supplied with the Japanese paper

Existing English text from the Japanese source, reproduced without rewriting. It is distinct from the translated Japanese abstract at the beginning of this access edition.

In Japanese organizations, jibungoto-ka (treating matters as one's own) is demanded across domains, yet the concept remains underdeveloped. Medicine confronted a similar problem in shifting from compliance to adherence; medication counseling was one practical implementation of this shift. This article transfers the structure to organizations, redefining jibungoto-ka as undertaking—holding an internally endorsed purpose within a relation of answerability—combined with embodiment, its integration into practice. Answerability is a non-transferable standing to give reasons to specific others and requires reciprocity. The medicine–organization asymmetry derives from pathways of encounter, not from domains. We propose two pathways of organizational adherence: a forward pathway through which employees undertake managerial purposes, and a reverse pathway through which employee-originated orientations enter a formal review process, and decision makers give reasons for adoption, deferral, or termination. Chronic failure results from pressing only the forward pathway. As generative AI accelerates externalization and combination, undertaking and embodiment become the rate-limiting steps. We derive three principles: designing enabling conditions including encounters, assigning accountability for stopping decisions, and observing leading indicators.

Edition and source integrity

English access edition 1.0.0, September 6, 2026. Japanese source: T4 preprint v1; local source papers/T4.md from corpus release 3.2.1. This is a translation layer only: the Japanese original and the T1–T5 consolidated corpus are unchanged. Section links lead back to the Japanese reading view. The manifest makes the source edition and alignment inspectable; mechanical completeness checks are not a certification of semantic equivalence.

Japanese original · English HTML · Translation manifest · Citation policy