From the Cheng Weishi Lun to the Impossibility of Cross-Temporal Responsibility: The Problem of the Responsible Subject in Buddhist Karma

WANG MUTI（王穆提）

Abstract

This paper takes the Cheng Weishi Lun (Discourse on the Perfection of Consciousness-only) as its core text. Starting from the theoretical design of karmic retribution, samsara, and subject continuity in Buddhist Yogācāra philosophy, it re-examines a fundamental problem in the Buddhist theory of karma—a problem long considered smoothly resolved by concepts such as "continuity" (santāna), "retaining seeds," and "the infallible maturation of fruits," yet which in reality remains unresolved. The problem is this: if Buddhism insists that all dharmas are non-self (anatman) and that mental states and mental factors arise and cease momentarily, how can the responsible agent who committed good or evil deeds in the past receive the corresponding retribution through a bearer in this life or the next? Put more strongly, if the bearer in the present life is not strictly identical to the responsible agent of the past, can the proposition of "self-agency and self-retribution" (one bears the consequences of one's own actions) still hold as a strict proposition of responsibility?

The Cheng Weishi Lun explicitly states on one hand, "The world and the sacred teachings speak of the self and dharmas, but these are only provisionally established and lack true existence." On the other hand, it secures the continuity of karmic effects and the cycle of death and rebirth under the premise of non-self by stating, "Sentient beings each possess a fundamental consciousness, which continues homogeneously and retains seeds," and "The former ceases and the latter arises, with cause and effect continuing uninterrupted." This reveals its true theoretical mission: to exclude a permanent, substantial self while simultaneously avoiding the annihilation of cause and effect.

This paper argues that Yogācāra Buddhism, particularly the Cheng Weishi Lun, successfully constructs a highly sophisticated theory of causal preservation. Through concepts such as the Ālaya-vijñāna (storehouse consciousness), seeds (bīja), perfuming (vāsanā), manifest action (samudācāra), maturation (vipāka), and the Manas (afflicted mind), Buddhism is able to explain how karmic causes are not lost, how retributions mature infallibly, how personality and habitual tendencies (vāsanā) continue across lifetimes, and how the continuity of death and rebirth remains comprehensible under the premise of non-self. As research by Waldron and Lusthaus points out from the perspectives of intellectual history and philosophical analysis, the most stable theoretical function of the Ālaya-vijñāna lies in acting as the deep mechanism for the preservation of karmic seeds and the generation of experience, rather than as a substantial subject that can directly bear normative responsibility. In other words, Yogācāra is better equipped to answer "where the fruit comes from" and "why it does not mature chaotically," but it cannot necessarily answer "who should bear it" and "why the present bearer is justifiably identical to the past responsible agent."

This paper further points out that the original theoretical function of the Ālaya-vijñāna’s "retaining of seeds" is to prevent causal annihilation, preserve karmic seeds, and establish the directional nature of retribution, not to serve as a true responsible subject. If we rely solely on its original function, it can only secure causality, not responsibility. If it is forcibly required to undertake the function of continuing the responsible subject, the Ālaya-vijñāna would approximate a subtle theory of self. Similarly, the Manas, as a mechanism of self-grasping, can only explain how the sense of subjectivity and attachment to a self arise; it cannot thereby justify a real subject capable of bearing responsibility. Thus, what Yogācāra most stably secures is not strict "self-agency and self-retribution," but rather "karmic maturation within a continuum." Mark Siderits’s analysis of Buddhist personal identity perfectly highlights this: between the structure of continuity and the responsible subject, there remains a normative gap that cannot be ignored.

However, this paper does not stop at an internal critique of Buddhist theory. It further demonstrates that if one strictly demands that the responsible agent and the bearer must be the identical subject, and if this identity is understood such that the responsibility can only be borne by the actual agent at the very moment of the action, then not only is the Buddhist theory of karmic reincarnation difficult to establish, but even delayed responsibilities within a single lifetime—such as repentance, compensation, penal liability, and retribution—will face fundamental difficulties. Consequently, what the Buddhist theory of karma ultimately reveals is not merely a localized dilemma within religious doctrine, but a broader philosophical predicament: Is cross-temporal responsibility possible? Derek Parfit’s reductionist treatment of personal identity and the debates in modern philosophy of personal identity (Locke, Hume, Reid, Butler) allow this problem to be clearly situated within the broader context of the philosophy of responsibility.

The ultimate conclusion of this paper is that what Buddhist Yogācāra—and indeed mainstream Buddhist karma theory—can stably secure is that "prior causes and subsequent effects mature infallibly within a continuum." If one wishes to further secure the strong proposition that "the identical responsible agent personally bears the retribution," one must acknowledge a "thicker" responsible subject, which in turn creates profound tension with the Buddhist doctrine of non-self. Therefore, the true philosophical difficulty of the Buddhist theory of karma lies not merely in how samsara and causality can coexist, but in whether the responsible subject can maintain strict cross-temporal identity under the premise of non-self.

Keywords:Cheng Weishi Lun, Yogācāra Buddhism, Ālaya-vijñāna, Manas, seeds (bīja), perfuming (vāsanā), maturation (vipāka), karmic retribution, self-agency and self-retribution, karmic maturation within a continuum, responsible subject, personal identity, cross-temporal responsibility, theory of non-self (anatman).

Introduction

In Buddhist thought, "self-agency and self-retribution" (自作自受) is perhaps one of the most iconic ethical formulations. It is not merely a religious teaching to encourage good and deter evil; it is also often seen as the most concise and intuitive summary of the Buddhist view of karmic cause and effect: the good and evil deeds committed by sentient beings in the past will inevitably invoke corresponding retributions of suffering or happiness in the future. Good is not done in vain, and evil is not lost. There is no other who bears it in one's stead, nor is there an effect without a cause. From the perspective of religious practice and folk belief, this language almost simultaneously combines an explanatory function and an ethical one: it explains why life's circumstances are the way they are, while also compelling the agent to take responsibility for their own actions. However, this formulation appears solid precisely because it is frequently placed within an educational, prescriptive context, without being subjected to stricter philosophical analysis.

If Buddhism simultaneously maintains that all dharmas are non-self, the five aggregates are not the self, mental states and mental factors arise and cease momentarily, and there is no permanent, unified presiding subject, a deeply fundamental question inevitably arises: Since the past agent of the deed has already ceased, and the bearer of the retribution in this life or the next belongs to a different time and a different stage of life, how can it still be said that the "identical responsible agent" is bearing the consequences? The Cheng Weishi Lun itself attempts to respond precisely to this point: it explicitly refutes a permanent, substantial self, and preserves the infallibility of karmic effects and the unbroken cycle of samsara by positing the continuity of a fundamental consciousness, the retaining of seeds through perfuming, and the concept of "the former ceasing and the latter arising." But is this sufficient to justify that "the present bearer is the past responsible agent himself"? This question is not equivalent to the common doubt of, "I do not remember my past life, so karmic retribution is unfair." It touches upon something much more fundamental: On what grounds can the responsibility for an action cross over time to be borne by a later subject?

The focus of the problem, therefore, is not the presence or absence of memory, but responsibility-identity itself. Buddhism is particularly vulnerable to this question precisely because it rejects a substantial self on one hand, yet must preserve karmic retribution and samsara on the other. Without some structure of forward-flowing continuity, Buddhism cannot explain how good and evil deeds are not lost, how retributions mature infallibly, how cultivation has a direction, and how liberation is possible. Consequently, since the early Abhidharma period, internal Buddhist traditions have been attempting to design a theoretical apparatus that "preserves continuity under the premise of non-self." The Cheng Weishi Lun can be regarded as one of the most mature and precise systems among these attempts. It not only establishes the Ālaya-vijñāna as the basis for retaining seeds, receiving perfuming, maturation, and continuity, but it also uses the Manas to explain how the sense of subjectivity and self-grasping are formed, thus making Buddhism appear capable of maintaining both non-self and the realities of karma and samsara.

However, herein lies the problem: What exactly does the Cheng Weishi Lun secure? If it secures the idea that past actions leave seeds, that seeds mature within a specific continuum of consciousness, and that retributions do not mature chaotically, then it is indeed successful. But if it also wishes to secure the idea that the present bearer is the past responsible agent himself—and that this is not just karmic maturation, but strict "self-agency and self-retribution"—then it faces immense theoretical pressure. This is because the Ālaya-vijñāna was originally a mechanism for continuity, not an ontological personality; the Manas was originally a mechanism for self-grasping, not a justification for a true subject. If they are forced to undertake the latter functions, the theory edges dangerously close to a subtle theory of self, creating an internal tension with the non-self stance Yogācāra originally intended to defend. This is why modern scholarship often interprets the Ālaya-vijñāna as a deep cognitive structure rather than a substantial entity that can directly serve as a responsible subject.

This paper attempts to follow this path, unfolding an argument that progressively moves from an internal theoretical critique of Buddhism to general philosophy of responsibility. The core argument of this paper is that the greatest success of Yogācāra lies in preserving the causal continuity of karmic retribution; yet, it ultimately fails to stably justify the strict identity of the responsible subject. Put more strongly, if we further accept a stricter premise—that true assumption of responsibility can only be borne by the actual agent at the very moment of the action—then not only is Buddhist karmic reincarnation difficult to establish, but all delayed responsibilities within a single lifetime will also face fundamental difficulties. Thus, the problem of Buddhist karma is ultimately pushed toward a deeper philosophical question: Is cross-temporal responsibility possible? Derek Parfit's reductionist treatment of personal identity, as well as Mark Siderits’s modern reconstruction of Buddhist personality theory, allow this problem to be more clearly situated within the broader debates of personal identity and the philosophy of responsibility.

This paper proceeds along three simultaneous threads: conceptual analysis, internal critique, and comparative philosophy. "Conceptual analysis" refers to distinguishing between concepts that are often conflated: causal generation versus responsibility attribution, the sense of subjectivity versus the responsible subject, and karmic maturation within a continuum versus self-agency and self-retribution. "Internal critique" means not using external theology or modern philosophy to preemptively pressure Buddhism, but rather examining within Yogācāra's own theoretical context whether it can accomplish the task it set out to do. Finally, "comparative philosophy" involves juxtaposing this problem with Western philosophies of personal identity—particularly Locke, Hume, Parfit, Reid, and Butler—to judge whether the Buddhist problem is a uniquely religious dilemma or a structural predicament that all theories of responsibility must face.

Chapter 1: Research Questions, Methodology, and Definition of Terms

I. Research Questions

This paper revolves around four progressively advancing research questions. First, what exactly is the Buddhist theory of karma claiming? Is it saying "actions produce consequences," or is it saying "the agent themselves will bear the consequences of their actions"? While these appear similar, they are not identical. The Cheng Weishi Lun's explanations of the Ālaya-vijñāna, seeds, perfuming, and maturation clearly point more directly to the former. However, the didactic language of "self-agency and self-retribution" commonly used in Buddhism aligns closer to the latter. Second, what does the Cheng Weishi Lun actually secure? Does it truly justify that "present bearer = past responsible agent," or does it merely justify that "karmic causes are not lost, retributions mature infallibly, and continuity is unbroken"? Third, if Yogācāra is insufficient, are "thicker" subject models—such as the Pudgala, soul theories, or divine judgment—sufficient to solve the problem? Or do they merely push the problem further back? Fourth, if we strictly demand that "responsible agent = bearer," does this mean that not only reincarnation and karma, but also delayed responsibilities within a single lifetime become problematic? If so, then what we are discussing is no longer just Buddhism, but rather: Is cross-temporal responsibility possible?

II. Research Methodology

(1) Conceptual Analysis

This paper first distinguishes four sets of concepts: 1. Causal generation vs. responsibility attribution; 2. Sense of subjectivity vs. responsible subject; 3. Karmic maturation within a continuum vs. self-agency and self-retribution; 4. Generative relations vs. normative relations. These distinctions are the prerequisites for the entire argument. If they are not separated first, all discussions easily fall into the conflation of "because there is an association, therefore it is the same person," or "because the fruit matures here, therefore the responsibility belongs here." Such conflations are precisely the pitfalls the Buddhist theory of karma encounters when subjected to philosophical analysis, and they represent the theoretical leaps that thinkers like Siderits and Parfit repeatedly try to avoid.

(2) Internal Critique

Internal critique means that rather than preemptively evaluating Yogācāra using external soul theories or modern philosophy of personality, we first ask: What problem was the Ālaya-vijñāna's "retaining of seeds" designed to solve? What was the Manas's "grasping of self" meant to explain? Is Yogācāra's own goal to preserve causality, or to preserve responsibility? Only by clarifying what it originally intended to do can we determine where it falls short. Based on the literal structure of the Cheng Weishi Lun, the author's most direct concern is indeed answering questions like, "If there is no substantial self, who creates karma? Who receives the fruit?" and addressing "the former ceasing and the latter arising, with cause and effect continuing." Its main task is to explain the preservation of karmic effects and the unbroken cycle of samsara, not to explicitly establish a normative theory of responsibility in the modern sense. Therefore, this paper evaluates whether Yogācāra has been overextended by later "strong responsibility" language, judging it by its original theoretical goals.

(3) Comparative Philosophy

Although centered on the Cheng Weishi Lun, this paper does not limit the issue to within Buddhism. Once the problem is expanded to "Is cross-temporal responsibility possible?", it must engage with Western philosophy of personal identity. Locke, Hume, Parfit, Reid, and Butler are not used to suppress Buddhism, but to measure under which theory of personality Buddhism is most defensible. What theories are closest to this paper's critique? Is the problem of the responsible subject unique to Buddhism, or a universal difficulty? Here, Locke's memory theory, Hume's bundle theory, Parfit's reductionism, and Reid/Butler's critiques of memory theory provide varying comparative frameworks to test exactly where the conceptual commitments of Buddhist karma fall.

III. Definition of Core Terms

This paper defines several core terms as follows:

First, the "responsible agent" (責任者) refers to the subject who actually commits a good or evil act and is normatively regarded as the source of that act. The focus is not merely on the causal source, but on the subject who can be held normatively accountable.

Second, the "bearer" (承擔者) refers to the subject who subsequently endures a consequence, retribution, punishment, compensation, repentance, or the maturation of pain/pleasure. The bearer is not necessarily identical to the responsible agent; the very question this paper investigates is how the responsible agent can be said to equal the bearer.

Third, "responsibility-identity" (責任同一性) does not mean that the earlier and later states are merely "associated." It means that the subsequent bearer is, in a sufficiently strong sense, the past responsible agent themselves, and can thus be justifiably viewed as the continuation of their responsibility.

Fourth, "karmic maturation within a continuum" (相續果熟) means that prior causes and subsequent effects mature within an unbroken continuum; the emphasis is on causes not being lost and retributions not maturing chaotically.

Fifth, "self-agency and self-retribution" (自作自受) is retained as a strong proposition meaning: the original agent themselves subsequently bears the consequences of their own actions. This distinction is crucial to prevent any loosely related consequence from being counted as "self-received," which obscures the true theoretical difficulty of responsibility attribution.

IV. Core Hypothesis and Tentative Conclusion

This paper does not pre-assume Buddhism's failure. Instead, it posits a hypothesis to be tested: The Cheng Weishi Lun may be sufficient to secure "karmic maturation within a continuum," but it is insufficient to secure strict "self-agency and self-retribution." If this hypothesis holds, we arrive at two layers of conclusions. The first is internal to Buddhism: What Yogācāra truly and stably secures is the preservation of causality, the non-confusion of retribution, and unbroken continuity, rather than the strict identity between the responsible agent and the bearer. The second is generally philosophical: If we accept a stricter requirement—that only the bearing of consequences by the identical subject at the moment of action constitutes true responsibility—then not only does reincarnational karma fall, but delayed responsibilities within a single lifetime are also shaken. This leads to a more radical premise: the impossibility of cross-temporal responsibility.

Chapter 2: Two Models of Buddhist Karmic Language: "Self-Agency and Self-Retribution" vs. "Karmic Maturation within a Continuum"

When Buddhism discusses karma, it often uses the same set of linguistic terms to describe two different things, leading to discussions that superficially appear to be debating the same issue but are actually talking past one another. These two things are: how the consequences of an action appear in the future, and whether the future bearer is the past responsible agent himself. The former is a matter of causality; the latter is a matter of responsibility. The former addresses "how it happens"; the latter addresses "on what grounds it is borne by this subject." If not separated, Buddhist karmic theory easily slips into didactic language, directly equating "the fruit matures here" with "you reap what you sow," missing the most crucial step of justification in between.

Therefore, this paper distinguishes between two models. The first is the strict responsibility model, namely the "self-agency and self-retribution" model. This model understands karma as: the original agent themselves later bears the consequences of their actions. It requires at least three things: First, a sufficiently strong identity between the agent and the bearer. It is not enough to have a causal link, memory fragments, or a continuum of consciousness; one must be able to say that the present bearer is justifiably the past responsible agent himself. Second, the bearing must possess normativity. It is not simply that consequences occur within a stream; this subject must be deemed the rightful bearer of that action. Third, this model inherently carries a sense of justice or cosmic retribution. If someone is not the original agent but bears another's consequences, under this model, it is an injustice. Thus, "self-agency and self-retribution" here is not an educational metaphor, but an extremely strong proposition of responsibility.

In contrast, what mainstream Buddhism, and Yogācāra in particular, can stably maintain is the other model: the causal maturation model, namely the "karmic maturation within a continuum" model. The focus of this model is not "the same 'self' receives retribution," but rather: karma is not lost, retributions do not mature chaotically, continuity is uninterrupted, and there is a directional generative relationship between the preceding and the succeeding. It is much looser than "self-agency and self-retribution." It does not require a "thick" subject; it only requires that this fruit does not arise without a cause, and that it does not randomly jump to a completely unrelated subject, but rather matures within this continuum due to a prior cause. Consequently, it is much more compatible with systems like Abhidharma, Yogācāra, and Madhyamaka that refuse to acknowledge a permanent, unified subject.

The reason Buddhism often uses the language of the former to express the content of the latter is not hard to understand: "Karmic maturation within a continuum" is too abstract and unsuited for general teaching, whereas "self-agency and self-retribution" is concise, impactful, and easily fosters moral vigilance. Thus, religious language operates like this: you create karma in the past, the karmic seeds are not lost, the fruit matures later, and so it is said to you, "you reap what you sow." The problem is: moving from "the fruit matures here" to "you yourself are bearing it" lacks the most vital argumentative step—why does the maturation of fruit in this continuum equal the responsibility belonging to this subject? If this leap is not bridged, then what Buddhism truly stably secures is not "self-agency and self-retribution," but merely "karmic maturation within a continuum."

Therefore, all subsequent chapters of this paper are built upon the following judgment: What Yogācāra and mainstream Buddhist theories can stably explain is "karmic maturation within a continuum," not "self-agency and self-retribution" in the strict sense. This is not to say Buddhism cannot use the phrase "self-agency and self-retribution"; as didactic language, it is perfectly permissible. But if treated as a strict philosophical proposition, it faces immense pressure.

Chapter 3: The Theoretical Goals of the Cheng Weishi Lun and the Original Function of the Ālaya-vijñāna's "Retaining of Seeds"

To fairly critique Yogācāra, one must first understand that the Ālaya-vijñāna was not invented out of thin air. It was proposed to respond to a set of sharp theoretical pressures internal to Buddhism: If all dharmas arise and cease momentarily, how is past karma not lost? If there is no substantial self, how do death and rebirth continue? Without a preservation mechanism, how can retributions avoid maturing chaotically? Without some latent continuation, how can habitual tendencies, preferences, and personality traits extend across lifetimes? The Ālaya-vijñāna was introduced under these exact pressures. It is not a philosophical ornament, but an intensely functional theoretical device. Waldron's research emphasizes that the concept of the Ālaya-vijñāna formed within the historical context of Indian Buddhism's long-term engagement with karma, habitual tendencies (vāsanā), and the deep structures of consciousness, not to establish a hidden soul.

Under the premise of momentary arising and ceasing, the biggest problem is: Since past actions have already ceased, on what grounds do future retributions arise? Yogācāra's answer is: Although manifest actions (samudācāra) cease, their force is not entirely eradicated; rather, they perfume to form seeds, which are retained by the Ālaya-vijñāna. Therefore, the primary function of the Ālaya-vijñāna is to prevent causal annihilation. This is its original, most fundamental function. The Cheng Weishi Lun's response to the non-Buddhist challenge of "If there is no substantial self, who creates karma and who receives the fruit?" unfolds exactly in this direction. It does not prove "there is a thick subject acting and receiving," but explains that "mental states and mental factors, by the force of causes and conditions, continue uninterrupted; thus, creating karma and receiving fruits do not violate reason."

However, merely saying "seeds are preserved" is not enough. Yogācāra must further explain: Why doesn't A create karma and B receive the fruit? Why is there a structural continuation of personality and habits? Why do retributions not mature randomly? Therefore, the Ālaya-vijñāna also serves as a locus of continuity, ensuring that seeds perfumed by a specific stream of consciousness await conditions to mature only within that same stream. In other words, it secures not just "there is a cause and an effect," but "this cause generates this effect within this stream." This allows Yogācāra to secure the directional nature of retribution much better than crude karma theories. Dan Lusthaus, analyzing the Cheng Weishi Lun, stresses that Yogācāra does not arbitrarily throw all experiences into a mystical container, but meticulously explains how different levels of consciousness construct the experiential world and the directionality of life's continuation.

Buddhism can neither acknowledge an unchanging self nor admit total annihilation. Consequently, the Ālaya-vijñāna simultaneously assumes the roles of the locus of continuity between past and present lives, the root of total maturation (vipāka), and the basis for both defiled and pure dharmas. This points to its third major function: establishing a non-annihilated continuum of life under the premise of non-self. Therefore, the original theoretical positioning of the Ālaya-vijñāna is as a mechanism of continuity, not as a personality subject. Most notably, although the Ālaya-vijñāna looks very much like some deep substratum, Yogācāra goes to great lengths to prevent it from becoming a true "self." Thus, it is defined as: momentarily arising and ceasing, neither eternal nor singular, receiving perfuming and retaining seeds, fluctuating and mutating, and capable of transformation (āśraya-parāvṛtti). Its design principle is clear: it must be stable enough to preserve causality, yet transient enough to avoid being a permanent self.

This foreshadows the subsequent problem: Since the Ālaya-vijñāna was originally a mechanism and not a subject, can it actually be used to justify a responsible subject? This paper's critique of the Ālaya-vijñāna does not deny its success as a theory of causality; rather, it denies its ability to consequentially secure responsibility-identity. In other words, interpreting the Ālaya-vijñāna's "retaining of seeds" as a substantial container misunderstands Yogācāra; it is more akin to a structure of functional preservation and conditioned maturation. Precisely because of this, it excels at explaining "where the fruit comes from," but it does not necessarily secure "why it is you yourself who must bear it."

Chapter 4: An Internal Critique of the Cheng Weishi Lun: Why the Ālaya-vijñāna is Sufficient to Secure Causality but Insufficient to Secure Responsibility

If we understand the Cheng Weishi Lun fairly, the Ālaya-vijñāna answers at least these questions: how karma is not lost, how retributions do not mature chaotically, how habitual tendencies continue, and how death and rebirth remain unbroken. Evaluated purely as a theory of causal preservation, it is highly successful. However, the difficulty raised by this paper is not "how the fruit arrives," but "who ought to bear it." These two questions cannot be conflated. Through retaining seeds, the Ālaya-vijñāna explains that: past manifest actions cease but leave seeds; seeds are preserved within a specific stream of consciousness; awaiting conditions, they mature and manifest as effects; the mature fruit does not randomly jump to a completely unrelated stream. This entire framework is elegant, but it addresses the order of generation, not normative attribution. In other words, it explains "why this effect occurs here," but it cannot directly explain "why this present bearer is exactly the past responsible agent himself."

The critical conceptual cut here is: Identity of a consciousness stream does not equal identity of a moral responsible subject. The Cheng Weishi Lun leans toward saying: past seeds only mature in the identical stream of consciousness that was perfumed, so it is not a different stream randomly receiving retribution, hence the creator and the receiver are not entirely distinct. But this paper's critique points out: identical consciousness streaming merely constitutes continuity in generation; it cannot automatically yield identity in normativity. That is, stream-identity explains that the prior and subsequent states have a causally closed relationship; responsibility-identity demands that the prior and subsequent subjects are normatively the identical person. The former does not automatically deduce the latter. This is exactly why Siderits, when discussing Buddhist non-self and personal identity, particularly emphasizes that moving from reductionist psychological or causal continuity to a strongly defined moral responsible subject requires an irreducible theoretical leap.

From this, a very clear dilemma emerges. Scenario one: The Ālaya-vijñāna acts solely as a mechanism of continuity. In this case, it can stably secure the preservation of karmic causes, the infallibility of retributions, and the continuity of samsara, but it cannot stably secure that the present receiver is the past responsible agent himself. Scenario two: The Ālaya-vijñāna is also required to act as the continuator of the responsible subject. In this case, it must do more than just retain seeds; it must undertake the preservation of subjectivity, the continuation of "selfhood," and the cross-temporal identity of a normative bearer. Once it does this, it becomes very much like a cross-temporal carrier where the "things" belonging to both the before and the after are truly "its" things—essentially becoming a subtle self-entity. This is precisely Yogācāra's most awkward predicament: if it is merely a mechanism, it cannot secure responsibility; if it aims to secure responsibility, it resembles a "self" too closely.

From this, we derive this paper's first layer of internal critique regarding the Cheng Weishi Lun: The Ālaya-vijñāna's "retaining of seeds" is sufficient to justify the preservation of karmic causes and the directional maturation of retributions, but it is insufficient to justify how the present bearer is justifiably identical to the past responsible agent. In other words, it secures the directionality of the karmic cause, not the identity of the responsible subject. If the Cheng Weishi Lun is understood as a theory of causal preservation, it is highly successful; if it is understood as a completed theory of the responsible subject, it fails to stably accomplish this task.

Chapter 5: The Fundamental Difference Between the Manas's Grasping of Self and the Responsible Subject

If the Ālaya-vijñāna is more like a "backend mechanism for continuity," does the Cheng Weishi Lun have another pathway to answer the question of "how the responsible agent equals the bearer"? The most natural candidate is the Manas (the seventh consciousness). The reason is simple: if the Ālaya-vijñāna provides causal preservation and a locus of continuity, then the sense of subjectivity in ordinary sentient beings' experiences—"I am suffering," "I am creating karma," "I am cultivating"—is obviously not directly manifested by the Ālaya-vijñāna. Rather, it is intimately related to the Manas's constant deliberation and its grasping of the Ālaya-vijñāna as a "self." From this angle, one might propose a defensive strategy for Yogācāra: the Ālaya-vijñāna provides causal preservation and the locus of continuity, while the Manas provides the sense of subjectivity and the experience of self-continuity. Combined, they might sufficiently explain why, despite there being no substantial self, sentient beings can still exist as bearers of responsibility.

However, this paper argues that this route also fails. It conflates two levels: how the sense of subjectivity is formed, and how the responsible subject is established. The Manas can only answer the former; it cannot justify the latter. According to the Yogācāra system, the Manas's significance lies in its constant apprehension of the Ālaya-vijñāna, grasping it as a self, and associating it with the four afflictions (kleshas): self-delusion, self-view, self-pride, and self-love. This indicates that the Manas is not designed to establish a true subject, but to explain how the subjective experience of unenlightened beings is constructed—why sentient beings always feel "there is an 'I' experiencing," why they mistakenly grasp a flowing continuum of consciousness as a solid subject, and why the sense of subjectivity remains stubborn despite the Buddhist teaching of non-self. In other words, the Manas's status in Yogācāra is that of a theory of subjective illusion. The more successful the theory of the Manas is, the more it proves that the "self" is merely the product of erroneous identification, rather than the metaphysical foundation of a responsible subject.

To more precisely illustrate why the Manas cannot secure responsibility, it is best to separate the following three layers. The first is the phenomenological layer of the sense of subjectivity: for example, I feel "this is my pain," "I am thinking," or "I have always been me." The Manas excels at handling this layer. The second is the psychological continuity layer of self-identification: for example, I recognize my childhood experiences as "things I did before," I use narrative to string together prior and subsequent experiences into "my life," and I possess a sense of unity regarding my habits, character, and memories. This layer relates closely to memory, habits, and personality structure, which Yogācāra can partially explain via seeds and perfuming. The third, and most difficult, is the normative layer of the responsible subject. It asks not "why do I feel it is me," nor "why do I identify with my past continuity," but rather, "why ought I, normatively, be viewed as the bearer of that action?" Feeling that "I am suffering" does not equate to truly being the original agent; identifying a past experience as "mine" does not equate to normatively having to take responsibility for all its associated consequences. Therefore, the sense of subjectivity, self-identification, and the responsible subject cannot be conflated.

Viewed in this light, what the Manas truly provides is the continuity of a subjective illusion: sentient beings, through constant deliberation, continuously treat a stream of consciousness as "I," thereby forming a psychological, experiential, and narrative sense of unified subjectivity. This is, of course, highly important, for without it we could not even explain "why sentient beings think they are undergoing samsara and receiving retribution." But the problem is: A sense of a subject does not equal the establishment of a responsible subject. Even if the Manas explains why sentient beings say "I am bearing this," it still cannot answer: Why is this "I" normatively the past responsible agent himself? Without completing this step, the Manas still only explains the psychological origin of responsibility language, without justifying its philosophical legitimacy.

There is an even deeper problem here than mere "functional insufficiency." If Yogācāra attempts to use the Manas to supplement the responsible subject, it directly conflicts with its own theoretical goals. Yogācāra inherently states: the Manas grasping at a self is a delusion; self-delusion, self-view, self-pride, and self-love are afflictions; grasping the Ālaya-vijñāna as a self is a misrecognition. If we were to flip this and say, "precisely because the Manas grasps a self, the responsible subject is established," it equates to admitting that an attachment explicitly defined as a delusion serves as the foundation for the attribution of responsibility. This is profoundly unstable within Yogācāra. The better the Manas explains "how the sense of self is constructed," the more it proves that this sense of self is not a true, solid subject capable of grounding responsibility, but an attachment that can be analyzed, dismantled, and eradicated. If the Manas is merely an illusory mechanism, it cannot secure responsibility; if it truly can secure responsibility, it is no longer merely an illusory mechanism. Yogācāra thus falls into a dilemma similar to that of the Ālaya-vijñāna: to secure responsibility, it must strengthen the subject; strengthening the subject undermines the doctrine of non-self and the critique of self-grasping.

Therefore, looking at the Ālaya-vijñāna and the Manas together, what the Cheng Weishi Lun ultimately handles successfully is: how karma is not lost, how retribution is not chaotic, and how the sense of "I" is formed. However, it still fails to stably address how a responsible subject maintains strict cross-temporal identity.

Chapter 6: From the Cheng Weishi Lun to Thick Subject Models: Pudgala, Soul Theory, and Divine Judgment

After the analysis in the previous two chapters, Yogācāra's difficulty seems quite clear: through the Ālaya-vijñāna and the Manas, it successfully secures continuity, karmic maturation, and the sense of self, yet it fails to stably justify the identity of responsibility. Naturally, an alternative approach emerges: if the problem is simply that Yogācāra is reluctant to acknowledge a sufficiently "thick" subject, would switching to a thicker model solve it? The most natural candidate within Buddhism is the Pudgala; external candidates include soul realism in Indian philosophy, the Abrahamic concept of a single soul undergoing judgment, and general persistent-subject theories of personality. Compared to Yogācāra, these theories are obviously much more willing to assert that there is some cross-temporally enduring bearer; that between the agent and the receiver, there is not just a structural correlation, but a thicker subjective continuity.

The reason the Pudgala theory is historically significant in Buddhism is precisely because it more candidly admits that if there is only the flux of dharmas without some posited "person" (pudgala), the subjects of karma, spiritual cultivation, and liberation lose their thickness. Thus, it argues for a Pudgala that is "neither identical to the aggregates, nor separate from them, and is inexpressible" to serve as the bearer of the "person." Superficially, this seems to perfectly address the critique raised in this paper, because the problem is not just whether the fruit matures, but whether there is a true bearer. However, a further problem remains: even if one acknowledges a bearer in the present life, that is not enough; one must still prove that this present bearer is the identical past responsible agent. If the Pudgala is not thick enough, it merely posits a nominal bearer more than Yogācāra does, but still fails to justify that "present bearer = past responsible agent." If the Pudgala is thick enough, it must advocate for some continuous cross-life subject, which dangerously approaches a true quasi-self or an inexpressible self. This is exactly why it was long viewed with suspicion within Buddhism: its explanatory power stems precisely from the risk it takes in reverting to a theory of self.

If the Pudgala remains somewhat ambiguous, soul theory appears much more straightforward. It asserts: there is a continuously existing soul; it is this soul that committed the past actions, and it is this same soul that receives the retribution in the present life; therefore, the responsibility and the bearing naturally belong to the identical subject. The appeal of such theories is strong because they at least do not evade the issue of the "subject." However, this paper's further critique is that even if the identical soul endures across different times, it does not mean the past responsible agent and the present bearer are strictly identical. This is because the one who actually created the karma was the past stage, and the one bearing the fruit is the present stage; even if both belong to the same soul, the "time gap" is not eliminated. Soul theory can explain why it is not a completely different person receiving the fruit, and why it can be said to belong to the identical life subject; yet it may not be able to refute an even sharper inquiry: if true responsibility belongs to the subject at the exact moment the action occurred, why can it be borne by the identical soul at a later time and still be considered, in a strict sense, as being borne by the original agent? In other words, soul theory merely shifts the problem from "no subject" to "a subject with a time gap," without thoroughly eliminating the time gap itself.

The Abrahamic model of one lifetime followed by divine judgment is often seen as more intuitive than reincarnational karma. Unlike reincarnation, which involves memory discontinuities across past, present, and future lives, this model asserts: the identical person acts in this life, is judged by God after death, and the identical soul receives the moral response. This indeed preserves the continuity of responsibility more easily than "acting in a past life, bearing in the present." However, by the strict standards of this paper, it is still insufficient. One can still ask: why is the consequence not borne immediately by the responsible agent at the moment of the action, but instead judged only after death? The theory of divine judgment can answer that the bearer is not someone else, but the identical personality or soul; yet it cannot avoid the fact that the judgment is still delayed, and a time gap still exists between the acting agent and the judged subject. If the standard is that only immediate bearing by the original responsible agent at the moment of action counts as strict responsibility, then divine judgment merely shortens the problem; it does not eliminate it.

Therefore, while thick subject models are indeed better equipped than Yogācāra to explain "why someone else is not bearing it," and are closer to the intuition of "the same person takes responsibility" found in everyday ethics, they still may not completely eliminate the time-gap problem of "a later stage bearing the responsibility of an earlier stage." This implies that what this paper forces out is not just the "need for a thicker subject," but a more radical question: Is any cross-temporal bearing no longer a strict bearing by the original agent? This provides the entry point for the comparison with Western philosophies of personal identity in the next chapter.

Chapter 7: A Comparison with Western Philosophies of Personal Identity: Who Supports This Critique? Who Can Best Defend Yogācāra?

At this point, the problem is no longer exclusively internal to Buddhism. As we have seen: Yogācāra is too thin to secure responsibility; the Pudgala is thicker but easily approaches a theory of self; soul theory and divine judgment, though thick, still cannot automatically eliminate the time gap. Thus, the true core has shifted: How is a person identical across time? And is this identity sufficient to support responsibility? This is precisely the core question in the Western philosophy of personal identity. Thinkers like Locke, Hume, Parfit, Reid, and Butler did not discuss karma in a Buddhist context, but their analyses of personality, memory, continuity, and responsibility can perfectly help measure: Which theory is mainstream Buddhism most like? Which stance is this paper's critique closest to? What prices must thick and thin subjects respectively pay?

Locke's classic theory of personal identity posits that personal identity does not lie in the soul or the body, but in the traceable continuity of consciousness and memory. A person is the same person not because of a material substratum, but because their consciousness can appropriate past experiences as their own. If this theory is applied to Buddhist karma, the problem immediately becomes acute: if a person in this life does not remember their past life, and cannot consciously claim past actions as "mine," then according to Locke's standard, the present personality is not the past personality, or at least not strictly identical. Consequently, the initial doubt of this paper becomes highly potent: if the present personality cannot consciously appropriate the past action, on what grounds must the present personality bear the fruit of the past? Thus, under Lockean standards of personal identity, the conscious continuum of Yogācāra is insufficient, the conventional designations of Madhyamaka are even more insufficient; and even thick models like Pudgala or soul theory fail to meet Locke's criteria if they lack conscious appropriation.

Hume is often noted for having a high degree of similarity with Buddhism, especially in his opposition to a permanent, substantial self. In Hume's view, introspection reveals no fixed, unchanging "I," only a flowing stream of perceptions, impressions, and ideas; thus, the so-called self is merely a bundle of perceptions. This is strikingly similar to the Buddhist aggregates (skandhas), the doctrine of non-self, and even certain trajectories in Madhyamaka and Yogācāra. In terms of "shattering the permanent self," Hume is almost a natural ally of Buddhism. But precisely because of this, Hume cannot truly help Buddhism solve the problem raised in this paper. If there is only a bundle of perceptions, and if personal identity is merely a psychological habit or imagination, then the thickness of moral responsibility becomes very thin. That is, Hume can help Buddhism critique permanent subjects and soul entities, but he cannot help Buddhism establish why the prior and subsequent bearers can still be viewed as the same person in terms of responsibility. Thus, Hume's utility to Buddhism is: helping to destroy the self, not helping to establish responsibility.

In modern Western discussions of personal identity, Derek Parfit is perhaps one of the most suitable figures to dialogue with Buddhism. Parfit's core thought is that personal identity is not ultimately what matters; what matters is psychological continuity and connectedness. In other words, one does not need a deep, irreducible "self"; many ethical and existential issues we care about can still be handled via weaker continuity. This is very close to mainstream Buddhism, especially Yogācāra. Yogācāra states: there is no need for a substantial self; the continuous structure of seeds, manifestations, and perfuming is sufficient to explain the continuation of life and the maturation of karma. Parfit could almost provide a modernized restatement for Yogācāra: the so-called "self" need not be a deep entity; as long as there is sufficiently strong continuity and connectedness between the before and after, many relations we care about remain valid. Therefore, from a Western philosophical perspective, Parfit is the one best equipped to defend Yogācāra.

However, precisely because of this, Parfit struggles to fully persuade the stance adopted by this paper. The question here is no longer, "Do we need a deep, irreducible self?" but rather: If responsibility requires strict bearing by the original agent, is a weaker continuity truly enough? Parfit would argue that perhaps the strict identity you demand is actually not important. But what this paper insists on is precisely that this demand is the problem itself. Therefore, Parfit does not negate this logic; rather, he asks us to abandon the standard we care about most. If one is unwilling to abandon it, then Parfit has not truly solved the problem, but only provided a path to thin it out.

The critiques of Locke's memory theory by Thomas Reid and Joseph Butler are highly crucial here. They argue that memory cannot constitute personal identity because memory itself already presupposes the existence of an identical subject who is remembering. In other words, saying "I am the same person because I remember" actually presupposes that the "I" exists first, before one can say "I remember." Thus, memory is not the foundation of personal identity, but only one of its symptoms. This critique is very helpful to this paper, as it prevents the problem from being reduced to "whether one has past life memories." This paper has never simply argued "it doesn't work because there is no memory," but rather: even if memory, continuity, and the soul are all present, as long as there is a time gap between the actual agent and the later bearer, it is still not strictly bearing by the original agent. From this angle, Reid/Butler are closer to this paper's direction than Locke. They too believe that relying solely on psychological continuity is insufficient and requires a deeper subjective foundation. Of course, Reid/Butler primarily intended to secure a thicker enduring subject to maintain responsibility and common-sense ethics; but on the point that "weak continuity and memory are not enough," they stand on the same side as this paper.

Synthesizing this comparison reveals: This paper is not satisfied with Lockean memory theory, because the issue is not just "remembering"; it does not accept the Humean/Parfitian strategy that "we do not need to demand a thick subject"; it shares common ground with Reid/Butler in that weak continuity cannot secure responsibility; but it is even more radical than Reid/Butler, because it further demands: even with a continuous subject, one must still answer why the bearing by a later stage counts as bearing by the original agent. In other words, the stance of this paper is no longer just "we need a thicker subject," but pushes to a deeper proposition: As long as responsibility is not immediately borne by the subject at the exact moment the action occurs, any later bearing is no longer, in a strict sense, bearing by the original agent. This is the "Impossibility of Cross-Temporal Responsibility" formally constructed in the next chapter.

Chapter 8: The Formal Construction of the Impossibility of Cross-Temporal Responsibility

After three avenues of examination—internal Buddhist critique, thick subject models, and Western philosophies of personality—it becomes apparent that all theories are actually circling the same problem. The thin subject models of Buddhism (Yogācāra, Madhyamaka) cannot secure a thick responsible subject; thick subject models (Pudgala, soul theory, divine judgment), while better at securing a subject, still retain a "time gap"; weak continuity models (Parfit) secure relatedness but ask us to abandon strict original-agent bearing; memory theory (Locke) captures the issue of appropriation but fails to handle the deeper subjective problem; and thick subject theory (Reid/Butler), while demanding a stronger subject, may not eliminate the ultimate critique of "later stage bearing." The issue is now exceedingly clear: What this paper truly questions is not just Buddhist reincarnational karma, nor even just "whether the subject is thick enough," but rather: Can responsibility be strictly retained across time? If the answer is negative, we arrive at a much broader conclusion than "Buddhist karma is invalid": strict cross-temporal responsibility is impossible.

This argument can first be established in its most direct version:

• P1. If the bearing of a responsibility is to be valid in a strict sense, the bearer must be the very identical responsible agent who actually committed the action.
• P2. Once time passes, the subsequent bearer (whether a later stage in this life, a bearer in the next life, or one judged after death) is not the identical responsible agent of that past moment, but only a later stage, continuator, or associated entity.
• C1. Therefore, the later bearing is no longer "bearing by the responsible agent himself" in a strict sense.
• C2. Thus, strict cross-temporal responsibility cannot be established.

The key to this version is understanding "the responsible agent himself" in extremely strict terms: it is not enough to be related; one must truly be that original agent. Once this premise is accepted, any delayed responsibility is inevitably impacted.

The second version can be constructed from conceptual dismantling:

• P1. Causal generation merely explains that a certain result comes from a certain prior cause.
• P2. Normative bearing demands that a certain subject is justifiably viewed as the original responsible agent of that action.
• P3. The former does not automatically deduce the latter; a result appearing in a certain continuum does not equal the later subject in that continuum being the past responsible agent.
• P4. Buddhist karma theory, soul theory, divine judgment theory, etc., can mostly handle the former, but may not be able to strictly justify the latter.
• C. Consequently, all theories of cross-temporal responsibility face an unbridgeable leap from causal relatedness to responsibility identity.

The advantage of this version is that it doesn't need to initially demand the strongest subject; it first points out the common theoretical sleight-of-hand: jumping from "the fruit comes from here" to "therefore it is you yourself who must bear it."

The third version pushes the problem back to general ethics and law:

• P1. General ethical and legal responsibility mostly rely on cross-temporal bearing: acting today, bearing the consequences tomorrow or years later.
• P2. If it is strictly demanded that the responsible agent must be the identical subject of the moment of action, then any later bearing can only be a bearing by a later stage, not a bearing by the immediate responsible agent himself.
• P3. Therefore, delayed responsibility in general ethics and law also cannot hold up in a strict metaphysical sense.
• C. Thus, the impossibility of cross-temporal responsibility is not just a speculative critique of Buddhist self-retribution, but a fundamental questioning of the general philosophy of responsibility.

This version shows that this stance is not a localized religious critique, but a deeper skeptical tendency regarding responsibility.

The strength of this position lies not in offering an alternative ethics, but in exposing something often overlooked: most theories of responsibility take it for granted that "an identical subject can bear prior responsibilities later," without truly justifying how. The reason Buddhism is particularly prone to having its loopholes exposed is that it stretches time much longer, cuts memory much more sharply, and dilutes the subject much thinner. This forces problems that could otherwise be masked by common sense within a single lifetime to be laid bare. Thus, Buddhist karma acts as a revealing mirror: it exposes not just Buddhism itself, but the hidden premises of the entire philosophy of responsibility. However, this position is also extremely radical. Because to thoroughly accept it is to admit that an adult taking responsibility for mistakes made as a youth is no longer strictly bearing it themselves; repentance and compensation in old age are no longer strictly self-bearing; legal responsibility is more of an institutional arrangement than a metaphysical self-bearing; and Buddhist reincarnational karma is even less likely to hold as strict "self-agency and self-retribution." This means that the greater the power of this stance, the greater its cost.

Chapter 9: Possible Objections and Responses: Is the Impossibility of Cross-Temporal Responsibility Too Strong?

Up to the previous chapter, this paper has pushed the problem into a quite radical proposition: if it is strictly demanded that the responsible agent and the bearer must be the identical subject, and this identity implies that it can only be borne immediately by the subject at the moment of action, then not only is Buddhist reincarnation difficult to establish, but the vast majority of delayed responsibilities within a single life also lose their strict metaphysical foundation. Such a stance will obviously encounter fierce objections. Indeed, if a theory simultaneously shakes Buddhist karma, soul theories, divine judgment, and general legal and moral responsibility, then the theory itself must face the strictest scrutiny. Therefore, the purpose of this chapter is not to weaken the preceding arguments, but to face the most potent objections head-on and see how far the "impossibility of cross-temporal responsibility" can hold its ground.

The first direct objection is: This paper reaches the conclusion that "cross-temporal responsibility is impossible" merely because it defines responsibility too strongly, too strictly, and too metaphysically. If one mandates from the start that only the actual agent at the very moment of the action, bearing it immediately, counts as true responsibility, then naturally all delayed responsibilities will collapse. The problem lies neither in Buddhism nor in theories of responsibility, but in this definition itself, which severs the possibility of responsibility continuing across time. This objection is indeed powerful, as it points out that the core of the debate is not just whether Buddhism answered well, but how we define responsibility. If the responsibility demanded by general ethics, law, and everyday practice never required such "absolute instantaneous self-bearing," but merely a cross-temporally traceable, attributable, and institutionalized bearing, then this paper's critique might merely be a refusal to accept commonly adequate standards of responsibility, rather than a proof that Buddhism or other theories have failed.

However, this paper can respond: This critique does not arbitrarily inflate standards; rather, it probes what exactly is implied behind strong responsibility language like "self-agency and self-retribution." If Buddhism or general responsibility theories wish to use language like "you suffer it yourself," "you must take responsibility for your actions," or "this is what you reaped from what you sowed," the minimum requirement cannot just be "relatedness," but must entail some degree of "selfhood/original agency." What this paper has done is merely rigorously pursue this requirement of selfhood to its logical end. In other words, this paper isn't arbitrarily raising the bar; it is asking theories to be responsible for their own language. If you want to say "self-agency and self-retribution," you must explain what "self" is; if you cannot, you should admit that what you can truly say is perhaps only "karmic maturation within a continuum." Thus, this objection reminds us: the reason this paper's conclusion is radical is precisely because it refuses to let strong responsibility language pass easily when it lacks sufficient theoretical backing.

A second important objection comes from practical philosophy and the philosophy of law. This objection argues: The establishment of responsibility does not necessarily require a fully justifiable metaphysical identical subject. The operation of law and ethics might rely merely on a sufficiently stable practical identity—the identical legal person, the identical social identity, the identical narrative subject, the identical inheritor of commitments. If so, then this paper's strict demand for "responsible agent = bearer" seems excessive. Responsibility, inherently, isn't necessarily metaphysical; it could be: institutional social arrangements, normative succession, the continuity of narrative identity, or self-restraint within practical rationality. The strength of this objection is that it prevents responsibility theories from collapsing alongside metaphysics. Even without a deep identical subject, we can still say: society requires treating prior and subsequent stages as the identical responsible agent, and practical rationality requires treating commitments, compensation, and repentance as extensions of the identical life; thus, responsibility still holds. This direction resonates particularly well with Parfit's reductionism: even without deep identity, what matters is still sufficient to support responsibility and ethical practice.

To this, the paper's response has two layers. Layer one: This paper acknowledges that such practical responsibility can be established. In other words, this paper does not claim that society cannot hold legal responsibility, or that people shouldn't compensate, repent, or honor agreements, nor that Buddhism cannot encourage good and deter evil. Layer two: But this paper denies that these equate to strict metaphysical "self-agency and self-retribution." This is the key. This paper aims to point out that practical responsibility can hold, but its mode of establishment is perhaps not "bearing by the original agent" in a strict metaphysical sense. If so, many theories must revise their language. For instance, to be precise, Buddhism should perhaps say "prior causes and subsequent effects mature within a continuum," rather than "the identical 'I' bears my own actions"; law should perhaps say "institutionally pursuing the one deemed the identical legal person," rather than "strict metaphysical bearing by the identical subject himself." Therefore, while this objection is strong, it does not truly overturn the paper; rather, it supports a more nuanced conclusion: many responsibilities can hold practically, but this does not mean they are justified as strict metaphysical original-agent bearing.

The third objection leans more metaphysical. It argues: The difficulty in this paper arises because it imagines the subject as a series of fragmented moments in time, refusing to acknowledge a whole that extends across time. If one adopts endurantism or perdurantism (four-dimensionalism), the problem is reframed: it is not "a later entity bears for a past entity," but "the identical whole subject unfolds different stages of its own history across time." The advantage of this objection is that it directly tackles the most radical aspect of this paper: it rejects the premise that "the past person and present person are fundamentally not the same," arguing instead that different temporal stages are merely facets of the identical whole; responsibility inherently belongs to the whole, not just to an instantaneous slice.

This line of thought is very strong and aligns with much of common sense. But this paper maintains its critique because it will ask: Is this holistic language truly answering the question, or just rewriting it? Because the core doubt all along is not "whether the past and present belong to the identical whole," but rather: the one who actually acted was the past stage, not the present stage; the one who actually bears it is the present stage, not the past stage. If responsibility demands immediate bearing by the original agent of the action, why does packaging different time slices into a whole eliminate this difference? That is, four-dimensional holism or endurantism changes the metaphysical description, but it may not address the strictest question this paper demands: why does "bearing by a later slice" equal "bearing by the original acting agent"? Without answering this directly, holism might merely be a higher-order restatement rather than a fundamental solution. Therefore, this paper's response is: these are indeed among the most potent thick-subject objections; but they only elevate the "difference between temporal stages" to be handled as "holistic identity"; whether this is sufficient for the strictest demand of original-agent bearing remains debatable.

The fourth objection would argue: Perhaps mainstream Buddhism was never meant to understand karmic retribution as strict "self-agency and self-retribution" anyway. In other words, what Buddhism truly meant to say was never "there is a fixed, identical 'I' receiving retribution in the future," but rather: prior causes and subsequent effects mature within a continuum; causality is not destroyed under non-self; and the purpose of cultivation is to transform the continuum, not to maintain a subject. If so, then the preceding critiques of Buddhism, while valid, only hit "folk Buddhism" or the "strong formulations of Buddhist didactic language," missing the more refined philosophical versions. Because sophisticated Buddhism would inherently admit: what it secures is karmic maturation within a continuum, not strict identity of responsibility. This objection is, to a large extent, exactly what this paper is willing to accept. For one of the true aims of this paper is precisely to force out this conclusion: if Buddhism is to maintain theoretical rigor, it perhaps should abandon the strong language of "self-agency and self-retribution," and replace it with the weaker, yet more precise, expression of "karmic maturation within a continuum." In other words, this objection does not overthrow the paper, but meets it at a higher level: if you want to preserve non-self, you shouldn't use overly strong responsibility language; if you want to use strong responsibility language, you must admit a thick subject; it is difficult to have both.

Chapter 10: How Should the Buddhist Theory of Karma Be Reformulated? From "Self-Agency and Self-Retribution" to "Karmic Maturation Within a Continuum"

From the preceding analyses, we can see: the greatest difficulty in the Buddhist theory of karma is not that it completely fails to explain cause and effect, but that it frequently uses a responsibility language much stronger than its own theory can stably support. That is: theoretically, it is better at explaining continuity, maturation, the preservation of seeds, and infallible effects; linguistically, it often frames things as "self-agency and self-retribution," "you suffer it yourself," and "not done by another." This discrepancy might be acceptable in a pedagogical context, but it becomes problematic in philosophical analysis. Once someone asks—as this paper has done throughout—how the present bearer equals the past responsible agent, if the theory lacks sufficient resources for subjective identity, it must honestly admit: what it truly secures is not strict "self-agency and self-retribution," but merely "karmic maturation within a continuum." Therefore, if the Buddhist theory of karma wants to maintain philosophical precision, it is necessary to reformulate its core propositions.

The most central principle of reformulation is adjusting karma from the strong language of responsibility to the language of continuous generation. That is, changing strong formulations like: "This is what you yourself did before, so now you suffer it yourself," "It is completely the identical 'I' receiving the retribution," or "The past acting agent himself receives the retribution in this life," into: "Past karmic causes mature into retributions of suffering or happiness within the subsequent continuum." Such a formulation has at least several advantages. First, it aligns better with the theoretical capacities of Yogācāra and mainstream Buddhism, since the Ālaya-vijñāna, seeds, perfuming, and maturation were inherently better suited to handle causal preservation, unbroken continuity, and infallible effects, rather than strict personal identity. Second, it avoids smuggling in a thick subject. Directly saying "the original person receives retribution" easily misleads people into thinking Buddhism secretly acknowledges a truly unchanging cross-temporal subject; saying "maturation within a continuum" clearly shows: this is neither soul theory nor eternalism, but causal continuity under the premise of non-self. Third, it more honestly faces the tension of responsibility. This formulation no longer pretends the problem is resolved, but admits: Buddhism can stably explain the generative order, but strict responsibility identity remains a point of tension.

Another key is acknowledging that Buddhism inherently possesses different levels of language. The goal of didactic language is not strict theoretical analysis, but to encourage good, prompt self-reflection, establish causal concepts, and keep people in awe of the consequences of their actions. In such contexts, the language of "self-agency and self-retribution" is highly effective. It is brief, powerful, and rapidly links actions to outcomes. But when the problem enters the theoretical level, one cannot rely solely on pedagogical language. At this point, one must be clear: Is "you suffer it yourself" a didactic strong-saying, or a strict metaphysical proposition? If the former, it should not be treated as a proven philosophical truth; if the latter, one must supply a complete theory of the identical responsible subject. And this is exactly where mainstream Buddhism struggles most. Thus, to avoid confusion, Buddhist karma theory should clearly distinguish: pedagogically, one can say "self-agency and self-retribution"; philosophically, it is more fitting to say "karmic maturation within a continuum."

A further adjustment the Buddhist theory of karma could make is to acknowledge: a consequence maturing in a continuum does not necessarily equal strict responsibility identity. That is, Buddhism does not need to understand all karmic effects as judicial, moral-retributive "accountability pursuits." Some karmic effects can be understood as: past tendencies continuing later, structures shaped by prior actions manifesting subsequently, and certain defiled conditions maturing in the ongoing continuum. In this way, the meaning of karma aligns closer to: the succession of consequences within a continuum, the extension of habits and structures, and the manifestation of the generative order of pain and pleasure—without needing to be fully described as "you, as a person, are being punished." This adjustment helps bring Buddhism closer to the spirit of the non-self doctrine and further avoids direct conflict with strict theories of responsibility.

Such a reformulation does not come without a price. What might be lost is the strong sense of moral intuition and cosmic justice. "Good brings good rewards, evil brings evil rewards" is an extremely powerful language of retribution; changing it to "maturation within a continuum" feels cold, abstract, and lacks that intuitive sense of justice. Folk religious narratives often require clear causal-ethical pictures; "maturation within a continuum" is too philosophical, less direct than "you did this before, now suffer it yourself." By saying "it's not entirely the original person receiving it, just the consequence maturing within the continuum," some might worry it weakens moral vigilance.

However, this reformulation also secures something important: theoretical consistency, a higher degree of alignment with Yogācāra and Madhyamaka, and stronger resistance to external critique. Especially for Yogācāra, the Ālaya-vijñāna, seeds, and perfuming are fundamentally best suited to explain maturation within a continuum. If Buddhism no longer presents itself as a thick responsible subject theory, then critiques like the one in this paper targeting "responsible agent = bearer" can no longer be used to demand it complete a task it originally never promised to accomplish. In other words, this reformulation does not strip Buddhism of everything, but returns it from an overly strong, overly moral-retributive framework to one that actually matches its own theoretical resources.

This also means that if Buddhism wishes to maintain its status as a non-theological, non-soul-based system of thought, it must accept: what it can stably secure is not cosmic retributive justice, but the infallible maturation order of causality under the premise of non-self. This conclusion may not be comforting, but it is likely more honest.

Conclusion

Beginning with the Cheng Weishi Lun, the core text of Buddhist Yogācāra, this paper investigated a fundamental difficulty in the Buddhist theory of karma: under the premises of non-self and momentary arising and ceasing, how can the responsible agent who committed past good and evil deeds receive the fruit through a bearer in this life or the next? If it cannot be justified that the past responsible agent and the present bearer are identical in a sufficiently strong sense, what content is actually left in "self-agency and self-retribution"?

To answer this question, the paper first distinguished two models within Buddhist karmic language: one is the strict responsibility model of "self-agency and self-retribution," requiring the original acting agent themselves to bear the consequences; the other is the causal maturation model of "karmic maturation within a continuum," merely requiring prior causes and subsequent effects to mature within an unbroken continuum. This paper argued that what mainstream Buddhism, and particularly Yogācāra, can truly and stably secure is the latter, not the former.

Secondly, through an internal critique of the Cheng Weishi Lun, this paper pointed out: the original function of the Ālaya-vijñāna's "retaining of seeds" lies in securing causality, preventing chaotic maturation, and preserving the cycle of death and rebirth. What it truly preserves is the directionality of karmic causes, not the identity of the responsible subject. If acting solely as a continuity mechanism, it sufficiently secures causality, but insufficiently secures responsibility; if forcibly required to bear the function of subject continuation, it dangerously approaches a subtle theory of self. Furthermore, the paper analyzed that the Manas's grasping of a self forms the sense of subjectivity and self-attachment, rather than a responsible subject; it can explain "how the sense of self is possible," but cannot justify "how responsibility is identical." If used to secure responsibility, Yogācāra falls into the internal contradiction of justifying a normative subject via a deluded attachment.

Subsequently, the paper pushed the critique towards thick subject models: the Pudgala addresses the bearer directly, but if not thick enough, it still cannot secure responsibility identity; if thick enough, it approaches a self-theory. Soul theory better preserves the subject, yet struggles to eliminate the time gap of "a later stage bearing the responsibility of an earlier stage." Divine judgment merely compresses samsara into one lifetime plus the afterlife, failing to truly eliminate the structure of delayed responsibility. Next, the paper engaged in a comparison with Western philosophies of personal identity: Locke best supports the initial critique of "without appropriation, how does it count as me bearing it?"; Hume aids Buddhism in dismantling the self but cannot help establish thick responsibility; Parfit can best defend Yogācāra, but on the condition of abandoning thick responsibility identity; Reid/Butler approach the demand for a thicker subject but may still fail to resolve the ultimate questioning regarding the time gap itself.

Finally, the paper expanded the problem to the "Impossibility of Cross-Temporal Responsibility." That is, if one strictly demands that the responsible agent must be the identical subject at the moment of action, and that any later bearing not done immediately by the original subject does not count as strict responsibility, then not only Buddhist reincarnational karma, but even delayed responsibilities within a single lifetime face fundamental difficulties. This indicates that what the Buddhist theory of karma exposes is not merely a problem unique to Buddhism, but a deeper predicament that the general philosophy of responsibility must confront.

If summarizing the entire paper in the most condensed and complete manner, one could say: the reason Buddhist Yogācāra, especially the Cheng Weishi Lun, established the Ālaya-vijñāna, seeds, perfuming, and Manas was primarily to preserve causal continuity, infallible retributions, and the generation of a sense of subjectivity under the premise of non-self, not to establish a strict responsible subject. What the Ālaya-vijñāna preserves is the directionality of karmic causes, not the identity of the responsible subject; what the Manas forms is the sense of subjectivity and self-grasping, not a real subject capable of normative responsibility. Therefore, what Yogācāra ultimately can secure is "karmic maturation within a continuum," rather than the strong sense of "self-agency and self-retribution." If we further demand that true responsibility can only be borne by the identical subject at the moment of the action, then not only Buddhist reincarnational karma, but all delayed responsibilities within a single lifetime face difficulties. Thus, the ultimate revelation of the Buddhist karma dilemma is not only the tension between non-self and karma, but a universal and profound philosophical predicament between personal identity and the possibility of cross-temporal responsibility.

To summarize the paper in a single sentence: The Cheng Weishi Lun secured causality, but may not have secured responsibility; Buddhism can stably explain "karmic maturation within a continuum," but struggles to strictly justify "self-agency and self-retribution"; and once this problem is thoroughly expanded, the entire philosophy of responsibility must confront: Is cross-temporal responsibility possible?

Footnotes

[1] The "strict proposition of responsibility" referred to in this paper does not merely mean "good is rewarded with good, and evil with evil" as found in general didactic language, but points to a stronger proposition: the subsequent bearer is normatively the original acting agent themselves. If this point cannot be established, "self-agency and self-retribution" can at best serve as a pedagogical strong-saying, rather than a strict proposition of responsibility.

[2] Regarding the historical formation of the Ālaya-vijñāna and its theoretical context as a mechanism for preserving karmic seeds, see William S. Waldron, The Buddhist Unconscious: The Ālaya-vijñāna in the Context of Indian Buddhist Thought; regarding the philosophical analysis of Yogācāra thought and the Cheng Weishi Lun, see Dan Lusthaus, Buddhist Phenomenology: A Philosophical Investigation of Yogācāra Buddhism and the Ch’eng Wei-shih Lun.

[3] In Fascicle One of the Cheng Weishi Lun, the author systematically refutes the real self of non-Buddhists. When answering "if there is no real self, who can create karma and who receives the fruit?", it explicitly states "mental states and mental factors, by the force of causes and conditions, continue uninterrupted; thus, creating karma and receiving fruits do not violate reason," and further explains how karmic effects are not lost by stating "sentient beings each possess a fundamental consciousness, which continues homogeneously and retains seeds," and "the former ceases and the latter arises, with cause and effect continuing." This indicates that Yogācāra's main concern is the preservation of karmic effects, not directly establishing a thick responsible subject.

[4] The reason this paper uses "responsibility-identity" instead of simply "personal identity" is that this paper cares more about normative bearing rather than just psychological or narrative continuity. This point is closely related to Siderits's modern reconstruction of Buddhist personality theory.

[5] When emphasizing the Ālaya-vijñāna as the "Buddhist unconscious," Waldron pays special attention to its developmental background in the intellectual history of Indian Buddhism; Lusthaus, however, focuses more on how Yogācāra analyzes the experiential world and issues of subjectivity via conscious structures. Both are helpful in understanding why Yogācāra needed to introduce the concept of the Ālaya-vijñāna to explain karma and reincarnation.

[6] The core of the Cheng Weishi Lun's response to "if there is no real self, who can create karma and who receives the fruit?" does not lie in justifying a substantial subject, but in explaining that mental states and mental factors, under the force of causes and conditions, "continue uninterrupted," thus being able to create karma and receive fruit. See relevant passages in Fascicle One.

[7] This dilemma can also be expressed as: Yogācāra is stronger at explaining the "transmissibility of karmic retribution," but weaker at explaining the "attributability of responsibility." The former is a problem of causality and psychological structure, while the latter is a problem of normative philosophy.

[8] Siderits's research is significant because it does not merely restate the Buddhist doctrine of non-self, but tests it within the contemporary philosophical context of personal identity, reductionism, and anti-realism. This allows "whether Buddhism can secure a moral responsible subject in a strong sense" to become a rigorously analyzable problem.

[9] The treatment of "provisionally establishing the self and dharmas" at the beginning of the Cheng Weishi Lun shows Yogācāra has no intention of acknowledging a truly existent self; yet it later explicitly states "the fundamental consciousness continues homogeneously and retains seeds," indicating it indeed requires a stream of consciousness sufficient to carry karmic seeds, habits, and the continuation of life. This is exactly where its theoretical tension lies.

[10] The fundamental depiction of the Manas in the Yogācāra tradition is its constant deliberation, grasping the Ālaya-vijñāna as the self, and associating with the four afflictions. This is the core mechanism Yogācāra uses to explain how the sense of subjectivity is formed.

[11] This paper strictly distinguishes between the "sense of subjectivity" and the "responsible subject," not to deny that the two are correlated in daily life, but to point out: the former cannot automatically constitute the metaphysical foundation of the latter.

[12] If a theory of responsibility is built upon deluded self-grasping, then that theory can only be a linguistic convenience (upaya), and cannot serve as a strict theory of responsibility-identity.

[13] The reason the Pudgala is important in the history of Buddhist thought is not that it necessarily succeeds, but that it sees more honestly than mainstream Buddhism: without some kind of "person," the subject of karma and cultivation becomes too thin.

[14] The shared advantage of soul theory and divine judgment is securing the subject; their shared limitation is the inability to automatically eliminate the time gap.

[15] Therefore, this paper's critique of thick subject models is not "you have a subject, so you are wrong," but "even with a subject, it may not be sufficient to justify strict cross-temporal responsibility."

[16] The focus of Locke's theory of personal identity lies in traceable consciousness and memorial appropriation, making it extremely susceptible to the critique against past-life karma: "If I don't remember it in this life, on what grounds does it count as me bearing it?"

[17] Hume's bundle theory is similar to Buddhism not in solving karma, but in mutually weakening the deep self.

[18] Parfit's importance lies in providing a theoretical framework that maintains some ethical correlation without needing a deep self; but this exactly demands abandoning a thick responsible subject.

[19] Reid and Butler's critique of memory theory is important because it forces responsibility theories to face: if memory is not the foundation, on what exactly does the identity of the subject rely?

References

I. Primary Buddhist Texts and Classical Commentaries

• Dharmapāla et al. (護法等菩薩造), trans. Xuanzang (玄奘譯). Cheng Weishi Lun (成唯識論). Collected in Taishō Shinshū Daizōkyō, Vol. 31, No. 1585.
• Kuījī (窺基撰). Cheng Weishi Lun Shuji (成唯識論述記). Collected in Taishō Shinshū Daizōkyō, Vol. 43, No. 1830.
• Nāgārjuna. Nāgārjuna’s Middle Way: Mūlamadhyamakakārikā. Translated by Mark Siderits and Shōryū Katsura. Somerville, MA: Wisdom Publications, 2013.
• Śāntideva. The Bodhicaryāvatāra. Translated by Kate Crosby and Andrew Skilton. New York: Oxford World’s Classics, 1995.
• Śāntideva. A Guide to the Bodhisattva’s Way of Life (Bodhicaryāvatāra). Translated by Vesna A. Wallace and B. Alan Wallace. Ithaca, NY: Snow Lion, 1997.

II. Modern Research Literature

• Finnigan, Bronwyn. “Madhyamaka Ethics.” In The Oxford Handbook of Buddhist Ethics, edited by Daniel Cozort and James Mark Shields, 162–182. Oxford: Oxford University Press, 2018.
• Lusthaus, Dan. Buddhist Phenomenology: A Philosophical Investigation of Yogācāra Buddhism and the Ch’eng Wei-shih Lun. London: RoutledgeCurzon, 2002.
• Parfit, Derek. Reasons and Persons. Oxford: Clarendon Press, 1984.
• Siderits, Mark. Personal Identity and Buddhist Philosophy: Empty Persons. 2nd ed. London: Routledge, 2015.
• Waldron, William S. The Buddhist Unconscious: The Ālaya-vijñāna in the Context of Indian Buddhist Thought. London: RoutledgeCurzon, 2003.
• Williams, Paul. Altruism and Reality: Studies in the Philosophy of the Bodhicaryāvatāra. Richmond, Surrey: Curzon, 1998.

III. Western Literature on Personal Identity and Responsibility

• Butler, Joseph. The Analogy of Religion, Natural and Revealed, to the Constitution and Course of Nature. To Which Are Added Two Brief Dissertations: I. Of Personal Identity. II. Of the Nature of Virtue. London: J., J., and P. Knapton, 1736.
• Hume, David. A Treatise of Human Nature, Book 1, Part 4, Section 6, “Of Personal Identity.”
• Locke, John. An Essay Concerning Human Understanding. First published 1689/1690. Modern reference edition: London: Penguin, 1998.
• Noonan, Harold W. Personal Identity. 3rd ed. London: Routledge, 2019.
• Perry, John, ed. Personal Identity. 2nd ed. Berkeley: University of California Press, 2008.
• Reid, Thomas. Essays on the Intellectual Powers of Man. First published 1785. Critical edition edited by Derek Brookes. University Park, PA: Pennsylvania State University Press, 2002.