A contemporary reading for an ancient, exact diagnosis.
What this concept is actually saying
The two characters that do all the work: 不如理 + 分別
The term is composed of two parts, and both matter. If you collapse either, the diagnosis loses its sharpness.
- 分別 (vikalpa) — the classical term is not "opinion," not "bias" in the casual sense, not even "judgment." In the Yogācāra framework (per the *Yogācārabhūmi-śāstra*, 瑜伽師地論), *vikalpa* names a discriminating cognition that divides a phenomenon into marks, properties, subjects, objects — the moment thought takes a presented datum and parses it into "this is X, this belongs to Y, this came from Z." It is the cognitive activity of conceptually carving up what appears. - 如理 (yoniśas) — literally "according to reason / according to the right principle." The classical commentaries use this term to indicate that the principle in question is not arbitrary logic but the principle that fits with how things actually are — the way of seeing that aligns with the structure of phenomena themselves (諸法的實際). When cognition is *如理*, it tracks reality; when it is *不如理*, it cuts against the grain of how things actually unfold.
So 不如理分別 is not a synonym for "thinking badly," "having a wrong opinion," or "making a logical mistake." It names a specific kind of discriminating cognition — one that *mis-cuts* phenomena by overlaying them with fabricated coordinates (a self where none can be found, a possessor where there is only process, an author where there is only conditional arising), and that does so in a way that runs counter to the actual structure of those phenomena.
The classical content of the "contrary" — what gets mis-cut
The *Yogācārabhūmi-śāstra* lists several distinct territories where this mis-cutting occurs. They are not interchangeable; each names a specific direction of distortion:
1. On past and future formations (過去、未來諸行) — discriminating "I once was / once was not" or "I shall be / shall not be." The classical text asks: was there a self in the past? Will there be one in the future? These are the boundary questions of existence and non-existence, projected onto time.
2. On inner and outer continuity (內外相續) — discriminating "this is enemy / friend / neutral." This is not a feeling; it is a *classifying cognition* that takes the stream of inner and outer phenomena and sorts them into these three affective-evaluative categories.
3. On action and its fruit (業與異熟) — discriminating an *agent* (作者) behind the action, and a *receiver* (受者) of the ripening. The mis-cut here is to insert a stable "doer" into a chain of action, and a stable "experiencer" into a chain of consequence.
4. On cause (因) — discriminating cause wrongly. Two main distortions: (a) denying cause altogether (無因論), and (b) positing an unequal / improper cause (不平等因): Īśvara (自在), prakṛti (世性), puruṣa (士夫), or some intermediary (中間). The phrase 不平等因 is the classical one — "unequal cause," meaning a cause that is not the same order as the effect, that does not actually produce what is attributed to it.
5. On the frontiers of self (前後際) — discriminating "what is my self? what is my self like?" and asking "from where has this sentient being come? where will it go after death?" This is the canonical 邊見 (view of extremes) territory — eternalism and nihilism about the self.
In all five, the *mechanism* is the same and classical: 增益 — the cognitive activity of *adding* what is not there. The mind imputes a self (我), a belonging-to-self (我所), an author, a receiver, an underlying ground. It is not so much *finding* something false as *imposing* something onto the flow of phenomena.
The grammatical family in the *Yogācārabhūmi*
One careful point that is easy to miss. The text uses a *family* of related terms, and they are not interchangeable:
| Term | What it foregrounds | |---|---| | 不如理分別 (ayoniśo-vikalpa) | The discriminating cognition itself | | 不如理作意 (ayoniśo-manasikāra) | The mental attention / application of mind | | 不如理猶豫 (ayoniśo-kaṅkṣā) | The wavering, the unresolved doubt | | 邪思惟 (viparyāsa-cintana) | The thinking that follows the four inverted views (of Buddha, Dharma, etc.) | | 非理作意 / 不如正理作意思惟 | The deliberately misdirected mental application | | 不如實知 (ayathābhūta-jñāna) | The failure to know things as they actually are |
The classical text does *not* collapse these into one. 不如理分別 stresses the discriminating *act*; 不如實知 stresses the epistemic *result* (a non-knowing). The text explicitly says one who cannot *as-it-is-know* what counts as 如理 and what counts as 不如理 cannot exhaust the āsravas (漏). This is the diagnostic hinge.
The second framework: the tenth vikalpa taxonomy of the *Mahāyāna-saṃgraha*
In the *Mahāyāna-saṃgraha-śāstra* (攝大乘論) as explained by Paramārtha (真諦譯), 分別 is divided into ten kinds. 不如理分別 is the seventh. The text explicitly says:
「七、不如理分別……是前分別，聽聞非正法為因……謂正法外人，非正法類分別。」
This shifts the emphasis in a productive way. Here 不如理分別 is not defined by its content alone (self, agent, wrong cause), but by its *genesis*: it is the discriminating cognition that arises *from hearing non-correct dharma* — that is, from hearing teachings that are not aligned with the right principle. The text then locates its practitioners: the 正法外 (those outside the correct dharma), the ninety-six kinds of non-Buddhist āstika/āstika-traditions (九十六種外道), situated outside the triad of 正聞 / 正思 / 正修 (correct listening, correct contemplation, correct cultivation).
This is significant because it changes the question. In the *Yogācārabhūmi*, the question is "what kind of cognition is this?" In the *Mahāyāna-saṃgraha*, the question is also "where does it come from, and whose stream does it inhabit?" The same diagnosis is being made; the second text adds the *sociology* of how wrong discrimination gets carried.
Note carefully: in the *Mahāyāna-saṃgraha*, 不如理分別 (the seventh) and 不如理思惟種類 (which belongs to the ninth, 決判執分別, with 身見 as its root and the sixty-two views) are *not* the same term. They are different entries in the same taxonomy. Conflating them is a classical misreading. The ninth is defined by *unreasonable thought* (不正思惟) as cause; the seventh by *hearing non-correct dharma* as cause.
The causal role: the precursor to āśraya (取) and the staining of consciousness (識染污)
The *Yogācārabhūmi* is explicit on the position of this cognition in the chain of defilement:
「諸取以不如理分別為先，繼而對我、我所的邪境界起取，配合緣自相境界之取及俱有依，使諸識轉起並受染污。」
The chain runs: 不如理分別 → 對我我所邪境界起取 → 緣自相境界之取 + 俱有依 → 諸識轉起受染污.
So 不如理分別 is *the precursor* — the wrong discrimination that *opens the door* to appropriation (取著), which then drags consciousness into staining (識染污). It is not the staining itself; it is the antecedent mis-cognition that *makes* the staining possible.
A walk-through in contemporary life
The classical list — past/future self, enemy/friend/neutral, agent/receiver, unequal cause, the frontier questions of self — maps onto daily contemporary life with an almost uncomfortable precision. Let me walk one extended scenario through each layer, then say where the common mapping errors are.
The scenario: a senior engineer, late thirties, just told her role is being eliminated
A company-wide restructuring is announced. She has seven years at the firm. Her team is dissolved. She is given ninety days and a small severance. She does not, initially, grieve. She begins to *think*.
On past and future formations — "I once was / shall be"
Within hours her cognition begins to project a self across time. "I was the lead on the platform migration — that is who I was." "I will be fine — I always land on my feet." "I was never appreciated here." "I will be unemployable at my age."
These are not feelings; they are *vikalpas*. Each one *imposes* a stable self on a continuous process — a process that has no "lead engineer" essence, no "I always land on my feet" essence, no "unappreciated person" essence. The classical text's words "我曾有、曾無 / 當有、當無" are precisely this: a self imputed into the past and projected into the future, where the only actual data is a series of conditions and a continuing flow.
Where people map this wrongly: they call this "narrative." It is not narrative as such; narrative is a linguistic activity. This is the cognition of *granting a self time-stamps* — saying "this self existed yesterday, that self will exist tomorrow." The classical diagnosis is sharper than "you are telling a story about yourself."
On inner and outer continuity — "enemy / friend / neutral"
Within days she has sorted the world. The CTO who approved the restructuring: enemy. The recruiter who cold-messages her on LinkedIn: ambiguous, possibly friend. The colleague who didn't reply to her email: enemy. The old mentor who calls: friend. The barista she passes each morning: neutral.
The *Yogācārabhūmi*'s phrase 是內外相續分別怨親中 (discriminating enemy / friend / neutral on inner and outer continuity) is *exactly* this: a discriminating cognition that runs through the stream of what is happening (people, messages, glances, status updates) and tags each node with an affective category. The categories themselves are not errors; the *act of imposing them as the structure of perception* — letting them organize what appears — is the 不如理 of it.
On agent and receiver — "there is a doer, there is one who reaps"
She says "they did this to me." "I did this to myself by staying too long." "I will pay for this." "I deserve better."
Classical: 對業計有作者, 對異熟果計有受者. The act of attribution is a vikalpa that *inserts* a doer into a stream of conditions (the restructuring was a multi-causal process involving markets, boards, contracts, timing — no "they" did it), and *inserts* a receiver into a stream of consequences (the consequences will be lived through changing conditions — no stable "me" receives them). The mis-cut is not the moral judgment of fairness; the mis-cut is the *ontological insertion* of an author where there is only process.
Where people map this wrongly: they treat this as the language of "responsibility" or "victimhood." Both readings are real, and morally important, but the Buddhist diagnosis is at a different level. It is not asking whether she is right or wrong to feel wronged. It is noticing that the very *cognition* "someone did this to someone" is a fabricated architecture laid over conditions that do not contain such an architecture.
On cause — unequal cause, or no cause
Two error-modes here, and she will swing between them.
*No-cause mode:* "It's just bad luck. Random. The market. The universe." — 無因. This denies the specificity of conditions; it treats the event as an uncaused intrusion.
*Unequal-cause mode:* "The CTO ruined this." "The algorithm picked my role." "If only I had stayed closer to the leadership." These posit causes that are not the same order as the effect: a single person or single decision is treated as the ground of a multi-causal outcome. The classical phrase 不平等因 (unequal cause) means a cause that does not actually have the power to produce what is attributed to it — an *Īśvara* in the structure. (In the contemporary register: the CEO-as-Īśvara, the algorithm-as-prakṛti, the single bad-decision-as-ground.)
Where people map this wrongly: they call this "attribution theory" or "locus of control." It is neither. Those are psychological categories. The classical category is *ontological*: the question is whether the mind has *posited a cause in the wrong register* — elevated a contingent event to the status of an underlying ground.
On the frontier questions — "what am I? where from? where to?"
These are the most classical: 我何等 (what kind of thing is my self), 我為何等 (what is my self like), 何處來 (where from), 死後何處去 (where to after death). In her case, reformulated: "Who am I without this job? Where did 'me' come from? Where is 'me' going?"
These questions are not pathological. They are the canonical frontier questions. The 不如理 is not that she asks them; it is that the *vikalpa* answering them — "I am a person whose identity was shattered" — *mis-cuts* by handing her a fabricated self to possess, lose, or rebuild.
**On the *Mahāyāna-saṃgraha* layer — hearing non-correct dharma**
And one more thing. The texts she is "hearing" in this season: the LinkedIn posts of others who've been laid off ("I bounced back, you will too"); the leadership memo ("this is a strategic realignment, we wish our colleagues well"); the algorithmic feed that surfaces either stoicism or grievance in equal measure. None of these are necessarily wrong in fact. But each is a *discourse* — an *other's words* (他言說) — and some of them are not aligned with how things actually are. The classical text says this kind of discrimination arises 因他言說 (from others' speech), and when that speech is 不正法 (not aligned with correct principle), the vikalpa that arises is itself 不如理. The contemporary algorithmic feed is, structurally, an industrial-scale 他引分別 (discrimination induced by others' speech) — and its 不正法 proportion is, structurally, very high.
Why contemporary people need this concept specifically
The "I-and-mine" structure is the most common contemporary affliction, and it is exactly what this concept names
Contemporary life is, in a specific sense, an industrial amplifier of 不如理分別. Three reasons.
1. Identity is increasingly narrated. Across the lifespan we are asked to construct, refine, optimize, pivot, and rebrand a self — to render the self into a coherent narrative that can be presented (on résumés, dating profiles, social media, in therapy). Every such construction is an instance of 增益: the mind adding a self where none is given in the data. The classical text's 我曾有、當有 is not an abstract worry; it is the daily cognitive work of *time-stamping a self*.
2. Sorting the world into enemy / friend / neutral has been algorithmically optimized. The classical phrase 內外相續分別怨親中 names a discriminating cognition. Contemporary platforms *industrialize* that discriminating — feeds, recommendation systems, follower counts, block / mute / unfollow mechanics. The mechanism is exactly the classical one: a stream of phenomena appears (a post, a face, a name), and the cognition instantly tags it as for-me / against-me / irrelevant. The classical diagnosis is that this tagging, *when it becomes the structure through which phenomena appear*, is 不如理 — because phenomena do not arrive pre-tagged, and the tagging is an imposition.
3. Causal attribution has become the central anxious activity. Contemporary cognition spends an enormous amount of energy on causal attributions — who caused what, why me, why now, why not them. The classical text's two errors, 無因 (no cause) and 不平等因 (unequal cause), are both very common: either dissolving causation into randomness, or concentrating causation onto a single Īśvara-like figure. Both errors *mis-cut* the actual relational texture of conditions.
The diagnostic precision matters because it specifies the cure
This is not a moral failing. It is not a sign of weakness, lack of resilience, or insufficient therapy. It is a specific cognitive pattern with a specific position in the chain of defilement (it is the *precursor* to 取著 and the *condition* of 識染污). And the classical text says: *whoever does not know 如理 from 不如理 as-it-is cannot exhaust the āsravas.* So the cure cannot be generic — it has to be the cultivation of the *contrary cognition*, which is what the tradition calls 如理分別 or 如實知 (knowing as it is). This is not a vague positive-thinking antidote; it is the mirror-image cognition that the same classical text specifies with the same precision.
Why this concept rather than another
There are many concepts one might reach for here — 煩惱 (affliction), 執著 (clinging), 顛倒 (inversion). Each is real and each is needed. But 不如理分別 is uniquely useful because it operates at the *cognitive* level, *before* clinging takes hold. It names the *first cut*, the *first mis-categorization*, the place where the door to defilement is opened. Working here means working upstream — at the source of the stream rather than at its visible waterfalls.
A brief note on the contemporary brain-science analogy: cognitive science has long distinguished fast automatic categorization (System 1) from slower reflective cognition (System 2). The classical 不如理分別 is not identical to System 1, and the classical 如理分別 is not identical to System 2 — the Buddhist categories are about *fit with reality*, not about speed. But there is genuine convergence on the point that *most of the mis-categorization happens before reflective cognition has a chance to inspect it*. The classical tradition's whole project of 正聞正思正修 is, in part, the training of the slow inspection. The analogy illuminates; it does not prove. Neuroscience has not validated the Buddhist claim that such training *exhausts āsravas* — that claim remains on its own metaphysical and contemplative grounds.
Common misreadings and clarifications
Misreading 1: "This just means 'wrong thinking.'"
It does not. "Wrong thinking" is far too broad. The classical term is specific: it is *discriminating cognition* that runs *contrary to principle* in a particular way — namely, by *imputing* a self, an agent, an underlying cause, or an enemy/friend/neutral structure onto phenomena. A math error is not 不如理分別. A forgotten appointment is not 不如理分別. A false factual belief ("the meeting is at 3") is not 不如理分別 in the classical sense — though it can become so if one then *cognition-constructs* a self who is "the kind of person who gets meeting times wrong." The classical category is *ontological*: it is about the structure of the imputation, not the truth-value of the content.
Misreading 2: "It's a philosophical problem, not a practical one."
The opposite is true. The *Yogācārabhūmi* is explicit that 不如理分別 is the precursor to 取著 and to the staining of consciousness (識染污), and that *whoever does not know 如理 from 不如理 cannot exhaust the āsravas*. This is not abstract philosophy; it is a claim about the working of suffering. The cognition of "I am the kind of person who…" and "they did this to me" are not philosophical curiosities. They are the daily cognitive events through which suffering is generated and reinforced.
Misreading 3: "It's just intellectual error — fix it with better ideas."
The classical texts do not recommend *better ideas* as the cure. They recommend 如實知 (knowing as it is), which in this tradition is not a propositional knowing ("I now believe the right things") but a *mode of cognition* that fits with the structure of phenomena. The cure is not at the level of opinion; it is at the level of *how cognition operates*. This is also why the *Mahāyāna-saṃgraha* locates the issue not just in the content but in the *genesis* — in the stream of what is being heard and from whom — because changing what is being heard (entering 正聞) changes the *kind of cognition that arises*.
Misreading 4: Conflating 不如理分別 with 不如理思惟種類 (or with 不如實知)
These three are not synonyms:
- 不如理分別 — discriminating cognition contrary to principle (the central term). - 不如理思惟種類 — in the *Mahāyāna-saṃgraha* taxonomy, this is the *content* of the ninth vikalpa (決判執分別), rooted in 身見 and correlated with the sixty-two views; its cause is named 不正思惟. It is *not* the same term as 不如理分別, which is the seventh. The classical text is careful to keep them separate. - 不如實知 — failure of knowing-as-it-is; an *epistemic* term about the result of cognition. The classical text explicitly says it is *related* to 不如理分別 but *not identical*.
To record the concept correctly, this distinction must be preserved.
Misreading 5: "It only happens to non-Buddhist contemplatives."
The *Mahāyāna-saṃgraha* does say 不如理分別 is the discriminating cognition of those *outside the correct dharma*, exemplified by the ninety-six non-Buddhist schools. But this is a *sociological* diagnosis about *predominance*, not an absolute boundary. A Buddhist practitioner who hears the correct dharma and *also* hears much else, and whose cognition *also* imputes self / agent / unequal cause in everyday contexts, is operating inside 不如理分別 in those moments. The classical text on 取著 and 識染污 makes no exception. The diagnosis applies wherever the cognitive pattern is present; the tradition simply observes that the *proportion* of such cognition is reduced inside the stream of 正聞正思正修.
Misreading 6: Treating 無分別智 as the goal of just stopping thought
The *Mahāyāna-saṃgraha* does say 無分別智是正道 (non-discriminating cognition is the correct path). But this is *not* a recommendation to suppress thought. It is the claim that the *non-discriminating cognition which fits with reality* (如理無倒) is the path. The background argument of the text: if the objects of discrimination were substantially existent, then discriminating them would not be inverted, and non-discriminating cognition would be inverted — which would be absurd. Since the objects are not substantially existent in that way, the cognition that does not impute such objects is the *fitting* cognition. The point is not "stop thinking"; the point is "stop *imposing* what is not there." This is a precise instruction, and reading it as "mindfulness means blanking your mind" mis-cuts the classical concept in exactly the way the classical concept diagnoses.
A closing note on the precision
What makes this term worth the trouble is its *precision*. It does not say "you have wrong thoughts." It says: *here is the specific way the discriminating cognition mis-cuts* (it imputes self, agent, receiver, unequal cause); *here is its specific content* (the past/future self, the enemy/friend/neutral sorting, the action and its fruit, the cause and its nature, the frontiers of self); *here is its specific causal position* (it is the precursor to appropriation and the staining of consciousness); *here is its specific genesis* (it arises from hearing non-correct dharma); *here is its specific mirror* (the 如理 cognition that knows-as-it-is). The contemporary value of the term is not that it gives us a fancy word for confusion. The contemporary value is that it *gives us a map of confusion* — precise enough that we can locate, in our own stream, the exact place where the mis-cut is happening.
The texts say: those who cannot know 如理 from 不如理 as-it-is cannot exhaust the āsravas. Those who can, can. That sentence — simple, exact, undecorated — is the whole teaching, and the whole practice begins with taking it seriously.