What this concept is actually talking about
The technical definition from the 《瑜伽師地論》 (Yogācārabhūmi-śāstra) is short enough to be misleading:
「謂於一切色等想事，分別色等種種自性所有尋思。如是名為自性分別。」
Translated loosely: "For all objects of apprehension such as form (and other phenomena), the probing inquiry (*vitarka-vicāra*, 尋思) that discriminates their various self-natures (*svabhāva*, 自性). This is called self-nature discrimination."
That single sentence has three moving parts, and they have to be unpacked in order — otherwise the concept collapses into "just thinking about things," which is exactly the misreading to avoid.
1. The object: 色等想事 — "phenomena such as form, as apprehended by perception (想) ." This is not "physical things out there." 色等想事 is a technical compound. 想 (saṃjñā, the mental factor of perception/apprehension) marks that what is in view are *the way phenomena appear to perception* — the taken-up images, qualities, textures of experience. 色 ("form," rūpa) is the first item, but the 等 ("and the rest") explicitly extends to the full range of what can be apprehended: 感受 (vedanā), perceptions themselves, volitional formations (saṃskāra), consciousness (vijñāna), and beyond. So 自性分別 operates on the entire phenomenal field, not only on visible objects.
2. The content of discrimination: 種種自性 — the "various self-natures." This is the load-bearing word 自性 (svabhāva). In the 《瑜伽師地論》 context, "self-nature" here means *what a thing essentially is* — the mark by which it is identified as itself. When perception presents a patch of color, sound, or emotion, 自性分別 is the activity that asks and answers: "This is *form*. This is *sound*. This is *anger*. This is *me*. This is a *cow*." Each item of experience is assigned its essential identity.
3. The activity that does the work: 尋思 — probing inquiry (vitarka + vicāra). 尋 (vitarka) is the initial reaching-toward, the directing of thought at the object; 伺 (vicāra) is the sustained dwelling, the working-over. Together they constitute the "searching, inspecting" character of this discrimination. 自性分別 is not a passive recognition; it is an active *probing* into what things are.
Three pieces together: a perceived phenomenon → the search for its essential nature → the mental act of assigning it that nature. That is 自性分別.
Where it sits in the larger taxonomy. The 《瑜伽師地論》 does not treat 自性分別 in isolation. It lists it as one item in a typology of discriminations (vikalpa, 分別), alongside 因分別 (causal), 位分別 (positional), 門分別 (aspectual), 上品分別 (degree), 顛倒分別 (perverted), 差別分別 (discrimination of differences), 過患分別 (fault-seeing) — and also among 總執分別 (general grasping), 我我所分別 (I / mine discrimination), 愛分別 (pleasant), 非愛分別 (unpleasant), 彼俱相違分別 (neutral). Each name picks out a different *angle* from which the mind discriminates phenomena. 自性分別 picks out the "what-is-it-in-itself" angle.
The crucial structural relationship with 差別分別 (discrimination of differences). The text defines 差別分別 immediately after 自性分別:
「謂即於彼色等想事，謂此有色、謂此無色……由如是等無量品類差別道理，即於自性分別依處，分別種種彼差別義。如是名為差別分別。」
The point is sequential and logically prior. *First*, 自性分別 fixes the self-nature — "this is a cow, this is anger, this is a sound." *Then*, on the basis (依處) of that already-fixed nature, 差別分別 overlays further categorical distinctions: is this cow visible or invisible? Is this anger contaminated or uncontaminated? Is this sound conditioned or unconditioned? Past, future, or present? Good, bad, or morally neutral? 差別分別 cannot operate without having first been given something to operate on — and that something is precisely the self-nature fixed by 自性分別. This is a quietly important structural claim: in the Yogācāra phenomenology of cognition, the act of *essentializing* a phenomenon (this *is* X) is the prior layer that makes all further categorization possible.
The loop: discriminations and their supports/objects. The text then collapses the entire discussion into a two-fold summary:
「此中所說略有二種。一者、分別自性，二者、分別所依、分別所緣事。如是二種，無始世來，應知展轉更互為因。」
These two — the discriminations themselves (分別自性) and *what they rest on / what they aim at* (分別所依、分別所緣事) — have been mutually conditioning each other since beginningless time. Past discriminations produced the present objects and supports; the present objects and supports now produce fresh discriminations; because those discriminations are not understood, they produce still further objects and supports in the future; those future objects will in turn produce future discriminations. This is the classic Yogācāra formulation of the *saṃsāric loop of conceptual cognition* — and 自性分別 sits inside that loop as one of its primary engines. Every act of "fixing what a thing really is" helps build the world that the next act of fixing will inhabit.
Life's walk-through
Let's take a few ordinary contemporary scenes and walk them through the classical definition, stage by stage. This is where the technical vocabulary either earns its keep or reveals itself to be a mirror for things you already half-noticed.
Scene 1: The internal debate about your own character. You're lying in bed after a social event in which you acted cold. The mind starts up. "Why did I do that? What kind of person am I, really? Am I a kind person? Or a selfish one? Or a guarded one? Or just exhausted?" That last move — *what kind of person am I, fundamentally* — is precisely 尋思 directed at the self-nature (自性) of the perceived phenomenon (想事), where the perceived phenomenon here is your own conduct as apprehended by perception. The mind is not merely classifying your behavior (差別分別 — kind vs. unkind, impulsive vs. deliberate). It is trying to fix *what you are*. That act of fixing — "I am a cold person," "I am a kind person," "I am fundamentally broken" — is 自性分別.
Notice the classical point that this is a *probing search* (尋思), not a settled conclusion. The mind usually doesn't rest; it goes back and forth, testing, revising. That's the 尋-伺 structure: initial thrust toward a hypothesis, sustained working-over of it.
Then the 差別分別 layer kicks in on top: is this fundamental nature good or bad? Was it my upbringing or my choice? Past trauma or present weakness? Each of those distinctions presupposes that the nature has already been fixed as "cold" or "kind" or whatever.
Scene 2: Trying to decide what a relationship "really is." A friend has been texting less. The mind starts: "What is this, really? Are we still close? Has something changed? Is she losing interest? Is this just her busy season?" Again, the underlying operation is the search for the self-nature of the relationship-as-perceived. The mind is not just noting facts (less texting — that's the perceptual datum, the 想事); it is asking what *essential kind of thing* this relationship is. Once the mind lands on a self-nature — "she's pulling away," "we're drifting," "this is over" — every subsequent piece of evidence gets re-categorized (差別分別) as confirming or disconfirming that nature.
Scene 3: The self-improvement / "find your true self" industry. Much of contemporary self-help discourse is, structurally, an invitation to 自性分別. "Discover who you really are." "Identify your authentic values." "Stop living according to others' expectations and find your real self." The promised object is an underlying svabhāva. The act is the search (尋思). The phenomenology is precisely the classical one.
Scene 4: Scientific / algorithmic classification. When a machine-learning system is trained to classify images as "cat / not-cat," it is, in a deliberately limited analogy, performing something like 自性分別 at the input layer — assigning a perceived object an essential identity — followed by 差別分別 at the classification head (cat / dog / car / tree). The analogy is useful but should not be pressed: the human case is *searching* and *interpretive*, where the algorithm is statistical pattern-matching. The human mind does not wait for labeled data; it spins its own hypotheses about essences and then lives inside them. This is why the loop (mutual causation between discriminations and their objects) is so much tighter and more self-reinforcing in human cognition than in any classifier.
Where most people map this concept wrong. The common modern mistranslation is "labeling" or "categorizing" — which is mostly 差別分別, not 自性分別. To label something "good" or "bad" presupposes you already know what it *is*. To ask "what is this, really?" — before any label — is closer to 自性分別. The distinction matters because the Buddhist analysis says the *prior* layer (the essentializing) is the deeper source of trouble, not the categorization that comes on top of it.
A second common misreading is to treat 自性分別 as "having opinions." It isn't. It's a *specific mental operation* — the search for the intrinsic nature of apprehended phenomena — distinguished from volitional reactions (愛 / 非愛分別), from self-referential grasping (我我所分別), from blanket-type grasping (總執分別). These can co-occur, but they are not the same operation. Conflating them flattens an entire phenomenology.
Why contemporary people need this concept
Three reasons, ranked by what feels most relevant today.
1. It names the operation underneath a lot of contemporary suffering. Most psychological distress is not, at root, about events. It is about the *interpretation* of events — and the deepest layer of interpretation is the search for what something *essentially is*. "I failed the interview" is an event; "I am a failure" is 自性分別. "My partner raised their voice" is an event; "they are a bad person" is 自性分別. The classical analysis gives a name to that *essentializing step*, which is the step that locks the door behind you. Once a self-nature has been assigned, the world rearranges itself around it, because (as the text says) discriminations and their objects have been mutually conditioning each other since beginningless time. Contemporary cognitive-behavioral therapy has independently rediscovered this: techniques like cognitive defusion and reappraisal target precisely the act of assigning essential natures. The Yogācāra analysis gives that intervention a much older and more precise vocabulary.
2. It dissolves the "find your true self" project — gently. The modern self-help industry sells the search for an authentic svabhāva as the path to wholeness. The Buddhist analysis, in contrast, points out that the search itself — the probing (尋思) for an intrinsic nature behind the appearances (想事) — is one of the engines of *saṃsāra*, not the path out of it. This is not a counsel of despair; it is a reorientation. The alternative is not "stop thinking about yourself" (which is impossible) but learning to see *the act of essentializing* as a conditioned, contingent operation — seeable, modifiable, no longer the silent ground floor of identity.
3. It explains echo chambers, filter bubbles, and ideological closure. When a community fixes the essential nature of a person, group, or concept ("they are *this* kind of person / *this* kind of threat / *this* kind of savior"), every subsequent fact gets categorized (差別分別) as confirming or disconfirming that nature. New evidence that fits is absorbed; new evidence that doesn't fit is either recategorized or dismissed. This is the classical loop — past discriminations producing present objects, present discriminations producing future objects — playing out at the social scale. The classical analysis offers a vocabulary for what is otherwise described only in sociological or psychological terms, and it points at a structural feature: the loop is *self-reinforcing*, not merely habitual.
Common misreadings and clarifications
Misreading 1: "自性分別 means having a fixed idea or prejudice." Close, but off. A prejudice already presupposes an essential nature has been fixed. 自性分別 names the *act of fixing*, not the resulting state. Two people can hold the same fixed idea; one may have arrived at it after probing, the other by rote. The classical analysis cares about the probing itself, because the probing is what generates the world the next round of probing will inhabit.
Misreading 2: "自性分別 is just thinking analytically." No. 尋思 is *searching inquiry*, but it is not limited to logic or science. A poet searching for the right word is engaged in a kind of 自性分別. A child learning to recognize a cat is engaged in a kind of 自性分別. A meditator watching a sensation and asking "what is this?" is engaged in a kind of 自性分別 — though, importantly, the meditator may also be introducing an element of *noticing the probing itself*, which begins to loosen the loop.
Misreading 3: "自性分別 is bad; we should eliminate it." The classical analysis is more subtle. 自性分別 is not, in itself, the thing to be eliminated. It is a description of an operation that *can be seen*, and seeing it clearly is the first step. The texts repeatedly point out that discriminations and their objects have been mutually conditioning each other since beginningless time — which is to say, you did not invent this operation yesterday, and you will not abolish it by willing it away. The work is to *see* the operation: to notice, in the moment, that the mind is searching for the self-nature of a perceived phenomenon, and that this searching is itself a contingent, conditioned activity.
Misreading 4: "自性 and svabhāva are the same word as in 'no-self-nature' (無自性, niḥsvabhāva)." This is a real conceptual hazard. The word svabhāva is used in two different (and ultimately opposed) registers in the Buddhist tradition. In the 《瑜伽師地論》's analysis of cognition, 自性 names the *content* the discriminating mind assigns to phenomena — the apparent intrinsic nature. In the Madhyamaka / Prajñāpāramitā tradition, the very absence of such an intrinsic nature (無自性, *niḥsvabhāva*) is what phenomena actually are. These are not contradictory positions; they are two ends of the same analysis. The first describes how the mind constructs apparent svabhāva; the second describes what is the case once that construction is clearly seen. Holding both together is part of the mature understanding — and is precisely why seeing 自性分別 clearly is taken to be liberating rather than despairing.
A short synthesis, then: 自性分別 is the discriminating mind's probing search for the essential nature of what perception presents to it. It is the *prior layer* beneath all further categorization. It is one of the engines of the *saṃsāric* loop in which discriminations and their objects mutually produce each other. Seeing it clearly does not destroy cognition — it opens the space within which cognition can function without being secretly run by the silent, unexamined demand that things have fixed essences. That, at minimum, is what the Yogācāra analysis is offering.