Chapter 6: Atoms and aggregation, the material end¶

First separate the two meanings of aggregation, or this chapter will confuse two matters. Fascicle 197 of the Mahāvibhāṣā was forced to do so when discussing contact. Calling eye-contact through body-contact "contact through the aggregation of three" presents no problem: faculty, object, and consciousness arise simultaneously. Calling mental contact the same thing is troublesome. The mental faculty is past, the object may be future, and consciousness is present: the three do not occupy the same compartment of time at all. The treatise answers by dividing aggregation into two meanings: coarising and inseparability, and accomplishing one undertaking together without conflict. Contact in the five sense consciousnesses has both meanings; contact in mental consciousness has only the latter. The Venerable Ghoṣaka states it more categorically: it is called aggregation because faculty, object, and consciousness accomplish one undertaking together, not because they coarise and are inseparable.¹

That sentence pinpoints the joint in Sarvāstivāda's theory of matter. What Sarvāstivāda wants is never several atoms piled into a new entity, but several entities accomplishing one undertaking together. Yet here lies the trouble too. When it discusses matter, Sarvāstivāda uses precisely the meaning of aggregation that Ghoṣaka rejects: coarising and inseparability. Atoms do not accomplish one undertaking together. As the end of Chapter 5 recorded, atoms of resistant derived matter cannot be coexistent causes of one another because they do not belong to one fruit.² The same word takes the latter meaning in discussions of contact, the former in discussions of aggregates. Whenever the two Chinese characters 和合, "aggregation," appear in the rest of this essay, we must first ask which meaning is intended.

The previous chapter handed on three questions: what an atom is, how several atoms count as being in one place, and whether what aggregates is the same thing as the atoms gathered together. On the strength of that sentence in fascicle 197, the third answer should be straightforward: no separate thing, just several things accomplishing one undertaking together. Yet the material end cannot give this straightforward answer. This chapter examines why.

Why Sarvāstivāda has only eleven kinds of matter. Chapter 2 gave the list: five faculties, five objects, and nonmanifest material form. The reason is that these eleven can be analyzed down to atoms and no further; they do not dissolve under analysis and so count as really existing under Chapter 1's criterion.¹ Looking back at the Southern tradition's nineteen kinds of matter, what has been cut? The three kinds of mutable matter are states, not constituents; the four kinds of characteristic matter are provided for elsewhere; delimiting matter, the space element, is included in the visible-form base; sex matter goes into the body faculty; nutritive matter into smell, taste, and touch; and the four great elements directly into the tangible-object base. The rule for cutting is the same throughout: what vanishes under analysis is not established.

Yet the same criterion yields opposite results in the two directions. When excluding downward, it is rigorous: whatever can be dissolved is out. When adding upward, however, it establishes nonmanifest material form, something neither visible nor tangible and not on the route of decomposition at all. Chapter 4 has already explained how the need for this thing led to its being posited. Here it is enough to record the asymmetry: the criterion is rigorous when excluding others' entries, and generous when preserving the school's own.

Within the eleven kinds of matter, the four great elements are what produce, and the other ten entries are what is produced. A ready-made list gives their relations. The verse in fascicle 7 of the Kośa says: "Great elements are two causes for great elements; five kinds for derived matter. Derived matter is three kinds for derived matter; only one cause for great elements." In relation to one another, the great elements are only coexistent and homogeneous causes. For derived matter, they can be five causes: productive, supporting, establishing, sustaining, and nourishing. Yet these five are merely distinctions within the efficient cause: it is called productive because something arises from it; supporting, as one relies on a teacher and the like; establishing, as a wall holds a painting; sustaining, because continuity is unbroken; and nourishing, because of growth.² Spread out, the five names are arising, change, support, abiding, and growth. They sound like a theory of physics, yet not one leaves the compartment of the efficient cause. Chapter 5's conclusion is tested here again: the six causes and four conditions merely rearrange the same heap of dharmas by role, and this remains so when the rearrangement reaches matter. What actually explains why things take the form they do is the statement behind the names, "the great elements produce matter." The five names do not undertake that work.

The atom's definition is a string of negations. The original wording in fascicle 136 is: "It cannot be severed, cut, broken, destroyed, or pierced; cannot be taken up, discarded, ridden, trodden on, grasped, or pulled; is neither long nor short, neither square nor round, neither straight nor crooked, neither high nor low; has no finer parts and cannot be analyzed; cannot be seen, heard, smelled, tasted, or touched."¹ Eight instances of 不 ("not") and seven of 非 ("not"). Only the final half-sentence in the passage is positive: "therefore the atom is said to be the finest matter." This book holds that the issue is not literary style. The atom is the endpoint of a procedure of decomposition, and the endpoint cannot itself be explained by decomposition. One can say only what it is not. Every property that can be stated positively invites a further question: to which part of it does that property belong? Yet the atom is allowed no parts.

Next comes the sequence of sevenfold increases. Seven atoms make one fine particle; seven fine particles make one copper particle. Successive sevenfold increases yield a water particle, hare-hair particle, sheep-hair particle, cow-hair particle, mote in a sunbeam, then a nit, a louse, and a barley grain. Seven barley grains make one finger-joint, and twenty-four finger-joints make one cubit.¹ Fascicle 12 of the Kośa gathers the same sequence into a verse, stating that from the atom at the beginning to the finger-joint at the end, each successive unit is seven times the previous one.² The fine particle in this sequence is the finest grade of matter apprehended by the eye and eye-consciousness, and only three kinds of eye can see it: the divine eye, the eye of a wheel-turning king, and the eye of a bodhisattva in his final existence.¹ In other words, counting upward from the atom, the second level already lies beyond ordinary human sight. Yet from the fine particle, it takes another ten or so levels to reach a finger-joint. These numbers give the impression that Sarvāstivāda has a quantitative understanding of matter. In fact, it has merely multiplied by seven several times.

In passing, let us check a frequently cited account precisely. Seven atoms form an atomic cluster; the usual explanation puts one atom at the center and six in the six directions. Yet the two Chinese characters 六方, "six directions," never occur in the 200 fascicles of the Mahāvibhāṣā, thirty fascicles of the Kośa, or eighty fascicles of the Nyāyānusāra. The Vibhāṣā's sequence of sevenfold increases converts quantities; it does not draw an arrangement. The closest account of arrangement in the three works is a sentence in fascicle 34 of the Nyāyānusāra: "In collected aggregates of matter with extreme resistance, shaped atoms are arranged all around. Held together by the shape of such an aggregate, boundaries and interstices become discernible."³ "Arranged all around" is indeed an arrangement, but specifies no number. The diagram with one center and six directions therefore does not come from these three works. Its exact source still awaits separate verification.

That atoms cannot exist alone is stated explicitly. The verse in fascicle 4 of the Kośa reads: "An atomic cluster in the desire realm, without sound and without faculties, has eight constituents; with the body faculty, nine; with another faculty, ten." The commentary explains that the name "atomic cluster" is established for the finest material aggregate, to show that none is finer. In the desire realm, where sound and faculties are absent, eight constituents coarise, not one fewer: the four great elements and four derived kinds of matter (visible form, smell, taste, and touch). A cluster with the body faculty has nine constituents; one with an additional faculty has ten. If sound is present, add one constituent to each.² Sarvāstivāda thus has two minima: the atom, the theoretically indivisible point; and the eight-constituent atomic cluster, the smallest practical lump.

Another sentence about the atomic cluster deserves to be singled out. The treatise says that within a cluster, whatever has heightened efficacy is clearly apprehended, but the others' substances are not absent. It is like feeling the touch of a needle point combined with a rod, or tasting the flavor of salt mixed with roasted flour.² When the needle point pricks, one feels only sharpness, yet the rod's touch is there too. When salt is mixed into roasted flour, one tastes only saltiness, yet the flour's flavor is there too. This is the sentence in Sarvāstivāda's whole theory of matter that most resembles something drawn from experience, and almost the only one. It aims to explain a crucial point: eight constituents coarising does not mean all eight are apprehended. Only the one with predominant efficacy is apprehended.

The two minima have different origins, a point Chapter 1 left unspoken. The atom is reached by analysis: analyze all the way down until no further analysis is possible. The coarising of eight constituents is not reached by analysis. It is an added restriction, preventing an atom from becoming a little ball that can sit there alone. The first lower limit is the result of analysis; the second is a limit imposed on analysis. The same account contains both an endpoint reached by analysis and a rule forbidding that endpoint from being taken out and used alone. We will encounter this two-level structure again.

The Vibhāṣā itself has no uniform answer to whether atoms are visible. Fascicle 132 records two accounts. The Venerable Ghoṣaka says that atoms should be called visible because they are objects of the wisdom-eye. The Abhidharma masters say they should be called invisible because neither the fleshly eye nor the divine eye can see them. The treatise's author adds in passing that this question does not concern the wisdom-eye, thereby siding with the latter.¹ The Kośa is more direct: for the five consciousnesses, it is only through the accumulation of many atoms that something becomes a support or object; for this reason, atoms are also said to have an invisible substance.² Fascicle 13 carries the point through: no one atom serves as support and one atom as object for the arising of the five consciousnesses, because eye-consciousness and the other four depend on aggregates and take aggregates as objects, depend on resistant things and take resistant things as objects, depend on aggregations and take aggregations as objects.¹

Yet the same passage in fascicle 13 also says that there is a blue atom, and atoms of such shapes as length, though they cannot be apprehended by eye-consciousness. The reason is carefully framed: if one atom were not blue, a gathering of many atoms should not be blue either; the same applies to length and other shapes.¹ This infers the properties of an individual atom backward from those of the aggregate, on the principle that addition cannot produce a property absent to begin with. Sarvāstivāda thus admits that blue must already be in the atom; otherwise, no number piled together could produce blue. There is nothing wrong with the principle itself. The trouble is that it applies to Sarvāstivāda too: what atoms lack should also be absent when they aggregate. From which individual atom, then, do an aggregate's interstices, boundaries, and visibility come?

Interestingly, the Vibhāṣā knows how to describe arrangements. It gives minute accounts of how atoms of the faculties are placed in the body. Atoms of the eye faculty spread laterally over the black pupil like coriander flowers; another account has them arranged one behind another, like fine needles visible in the water of an autumn pond. The ear faculty coils inside the ear opening like rolled birch bark; the nose faculty lies in the nostrils like a pair of fingernails; the tongue faculty is like a half-moon; and the body faculty extends in sequence throughout the body's interior and exterior.¹ Fascicle 2 of the Kośa records much the same, with an additional account of eye-faculty atoms stacked like pellets.² Where a picture is called for, there is a picture, and a credible one.

Yet where a picture is most needed, there is none. None of the three works gives the arrangement by which several atoms form an eight-constituent atomic cluster. This book holds that this is no oversight. The placement of the faculties can be inferred backward from experience: the range of sight and the shapes of the ear and tongue are there to see. The atomic cluster's arrangement cannot be inferred this way. And once drawn, it would have to answer a question: do two neighboring atoms touch or not? That is precisely the question Sarvāstivāda is least willing to answer.

The passage on the body faculty adds an odd statement in passing. The treatise says that if all atoms of the body faculty produced body-consciousness, the body would fall apart.¹ The same fascicle records the Venerable Vasumitra giving four conditions for failing to see visible form: too near, too far, too fine, and obstructed.¹ Taken together, the two statements show considerable attention to the limits of the senses. This even includes the thought that without such limits the body would fall apart. At the level of atoms, such considerations are entirely absent. Not one limit placed on the atom comes from experience.

Now to the dispute over touching. Fascicle 132 lays out the dilemma cleanly: do atoms touch one another? If they do, how do they avoid becoming one, or becoming things with parts? If they do not, they should scatter when struck, or there should be no sound. The treatise rules that atoms do not touch one another. If they touched, the contact would have to be over the whole or a part. Whole contact incurs the fault of their becoming one substance; partial contact, the fault of their having parts. Yet atoms have no finer parts.¹

Figure terminology. “Function” means activity (作用). “Critics” in the legend refers to the treatise’s adjudicators (评家).

The criterion in Chapter 1 shows why this reasoning is necessary. If atoms touch, they have spatial parts; if spatial parts, they can be divided; if divided, they vanish under analysis; if they vanish under analysis, they do not really exist. The criterion is transferred to matter unchanged, step for step. Nontouching is therefore an inference from that criterion, hardly a claim about physics. The inference is clean. Precisely because it is clean, it is later refuted at the same point.

The two consequences of nontouching each receive a patch. If atoms do not touch, why do they not scatter when struck? The wind element holds them together. But does the wind element not blow things apart? Some wind disperses, as during an eon of destruction; some holds together, as during an eon of formation. If atoms do not touch, how can they make sound? That is precisely what makes sound possible. If they really touched, the substances of two palms brought together should merge; without an intervening gap, how could sound be made?¹ Neither patch is ugly, but neither comes from observation. Both come from the shape of the opponent's question.

The same passage also leaves a fourfold distinction that completes the combinations of touching and nontouching. Sometimes touching causes nontouching, as when aggregated things are dispersing. Sometimes nontouching causes touching, as when dispersed things are aggregating. Sometimes touching causes touching, as when aggregated things aggregate further. Sometimes nontouching causes nontouching, as in the continuation of motes in a sunbeam of the same kind.¹ These four propositions sound like an account of particles gathering and dispersing. In fact, they never step outside the earlier prohibition, because what they call touching and nontouching concerns touching in name alone. The treatise rewrites the whole phenomenon of gathering and dispersing as a list involving no contact.

The same passage juxtaposes alternative accounts from two venerable masters. The Venerable Vasumitra says that if atoms touched one another, they would have to persist into the next moment. The Venerable says that they do not really touch; only in gathering and arising without an interval are they nominally called touching, in accordance with conventional truth.¹ Fascicle 2 of the Kośa reproduces the passage and explicitly endorses it: "This Venerable's view should be welcomed."² Vasubandhu then adds three points of his own, each tighter than the last. If this were not so, there would have to be gaps between atoms; with the space between them empty, who would block their movement and allow them to count as resistant? Again, apart from atoms there is no aggregated matter: touching an aggregate is touching atoms. Again, if atoms are admitted to have spatial parts, they have parts whether they touch or not; if they have no spatial parts, admitting that they touch involves no fault.²

The third point is crucial. If atoms have no spatial parts, touching creates no problem of division at all, and the prohibition on touching has been established to no purpose. If they have spatial parts, they still have parts when they do not touch, and the prohibition is equally pointless. Vasubandhu's point is that the rule Sarvāstivāda establishes to save real existence fails at both ends.

Sarvāstivāda has routes of retreat, but each step back has a price. If atoms are said to lack spatial parts yet have locations, the difference between location and spatial parts needs a separate account. Once that difference is explained, the atom gains another property open to questioning. If the space element is said to separate atoms, the trouble is more direct. The space element is an item included in the visible-form base. This concedes that another kind of matter lies between two atoms, so the aggregate is no longer merely an aggregate of atoms. The Vibhāṣā touches the edges of both retreats and follows neither through.

The five accounts in the dispute over touching, with the price each pays:

Of the five accounts, only the first is inferred from the criterion. The other four address the first account's consequences. And the final column always lands in the same place: the aggregate.

Turn back another page: the Vibhāṣā never adjudicates the question of gaps at all. Fascicle 132 first asks whether there are gaps within aggregates of great elements and the like. It gives two answers. Some say there are gaps, with the space element intermingled within, so things appear inseparable, like a woman concealing herself in a thicket. Others say there are no gaps: the things press against one another and abide without an interval. If there are no gaps, why do they not become one substance? The answer is that having no gaps does not make them one, just as the aggregates, bases, elements, and three times do not become one even though there are no gaps between them. Further, the great elements and the like differ in their own substances and activities, and cannot become one.¹ Two "some say" accounts stand side by side without an adjudication. Gaps and no gaps are opposite descriptions of the same lump of matter. The treatise takes both in, then moves on to something else.

Another question lies beneath this one, and the two answers touch opposite sides of the same matter. On the route with gaps, the gaps contain the space element. Fascicle 75 defines the space element as matter adjacent to resistant matter: resistance means aggregates, such as walls and other material things; what is near these is called matter adjacent to resistant matter.¹ Since the space element is matter, a third kind of matter lies between two atoms. The aggregate is no longer merely an aggregate of atoms. Here is the price of the retreat described above. On the route without gaps, the treatise asks why the things do not become one. It answers that although there are no gaps, they do not become one, and that the great elements and the like cannot become one because their own substances and activities differ.¹ Note the range of the last sentence. The statement that their own substances and activities differ concerns the differences among the four great elements (earth, water, fire, and wind) within a single aggregate. But ask the same question of two earth-element atoms, and the answer gives nothing: their own substances and activities are the same. The treatise reaches the question, but its answer reaches only the half concerning things of different kinds.

One step further: a single atom does not count as matter. In discussing the definition of matter, fascicle 1 of the Kośa raises its own objection: it is called matter because of alteration and resistance; if so, an atom should not be called matter, because it has neither. The answer is that no single atom abides in a place by itself; alteration and resistance become possible when many atoms gather.² Resistance means resisting something else. The definition of matter requires one thing to block another, yet a single atom can block nothing. It is called matter only through the gathering of many atoms. Thus an atom's very being a bounded thing is relational, not something it carries of itself. This comes before the question of number. The question is not why two atoms count as two, but why one atom counts as a body at all. The Vibhāṣā cannot answer the former because it has already given away its answer to the latter. Essay 18, Chapter 7 takes up this thread and sets it beside a modern construction of the same form.

In passing, another popular account needs correction. Later writers often say that Sarvāstivāda holds that atoms form a collection (和集), not an aggregation (和合), making "collection" a technical term. Yet the two Chinese characters 和集 occur just once in the 200 fascicles of the Mahāvibhāṣā, in fascicle 71, where it says that in the tangible-object element many atoms collect in one place.¹ This expression occurs not once in the thirty fascicles of the Kośa, and only with the eighty fascicles of the Nyāyānusāra does the count reach 32.^{2 3} By contrast, the two Chinese characters 和合 occur 292, 53, and 261 times in the three works respectively. The technical term belongs to Saṅghabhadra, not the Vibhāṣā. This book holds that it arose in response: Sautrāntika challenges forced Sarvāstivāda to establish a designation distinct from aggregation, and Saṅghabhadra established it. His direct opponent is the Elder, Śrīlāta. The dispute concerns the objects of the five consciousnesses, with a group of blind people as the analogy. Vasubandhu's three additions in fascicle 2 of the Kośa are another branch of the same line of challenge.³ When the Vibhāṣā discussed nontouching, it did not have this term in hand at all, nor feel the need for one.

The weight of the atom itself grows as it comes under challenge. The two Chinese characters 極微, "atom," occur 127 times in the 200 fascicles of the Mahāvibhāṣā, 72 times in the thirty fascicles of the Kośa, and 214 times in the eighty fascicles of the Nyāyānusāra. As occurrences per fascicle, these are 0.6, 2.4, and 2.7: the latter two are roughly four times the Vibhāṣā's density. The atom is not a hot topic in the Vibhāṣā. It is merely a technical term at the endpoint of decomposition, not receiving even one concentrated treatment in 200 fascicles. Only when it needs defending does it become a point of contention.

The occurrences of the three terms in the three treatises, set out together:

The five rows of numbers point to three sides of one matter. In the Vibhāṣā, the atom is merely a technical term at the endpoint of decomposition; it becomes a point of contention only when it needs defending. "Collection" as a technical term comes from Saṅghabhadra; the Vibhāṣā neither had the term in hand nor felt the need for it when discussing nontouching. As for the diagram with one center and six directions, no literal textual basis for it can be found anywhere in the three treatises.

Modern scholarship has already examined this tradition in considerable detail, with dedicated treatments of atomic contact, the distinction between collection and aggregation, and the dispute over the objects of the five consciousnesses. It also explains the setting in which Saṅghabhadra established his account of collection.⁴ This chapter shares those concerns. Its evidence always returns to the three treatises themselves; responsibility for the counts from searches of the complete works and for the judgments drawn from them lies with this book.

The target of refutation is established here; this chapter will not pursue it. Vasubandhu later raises two challenges to atoms in the Viṃśatikā. If an atom joins separately with atoms in the six directions, that one atom must have six parts. If six atoms occupy the same point, then however many gather, they remain only as large as one atom; a great mountain should be the same size as a particle of dust.⁵ These challenges are left for Essay 20. Here it is enough to record the shape of the target. Both point to aggregation, not the atom. Sarvāstivāda says that atoms do not touch in order to leave no room for spatial parts. Vasubandhu asks: how, then, does a lump of things form? Nontouching preserves the atom at the cost of leaving the aggregate unexplained.

Another easily confused point needs clarification: Sarvāstivāda's atoms are not the permanent atoms of non-Buddhist schools. Discussing cosmic destruction, fascicle 12 of the Kośa says that at the end of destruction not even an atom remains. It immediately records the opposing case: one group of non-Buddhists holds that atoms are permanent and says that atoms remain even then.² Sarvāstivāda explicitly does not take their side. Like other conditioned dharmas, atoms arise and cease moment by moment. Their minimality belongs only to the spatial dimension; in the temporal dimension they are as short-lived as other dharmas. The same fascicle's refutation of the Vaiśeṣika account of a whole with parts, and fascicle 7's refutation of a single cause, also follow this position.²

Two accounts are thus laid over the atom: momentary cessation and real existence in the three times. Chapter 3 dealt with this pair: substance always exists, while activity arises and ceases; activity establishes the distinctions among the three times.¹ Transferred to atoms, however, it is especially glaring, because an atom alone never acts. The five consciousnesses depend on aggregates and take aggregates as objects; one atom serves as neither support nor object. On what grounds, then, is a single atom judged present? This book's position is: through the aggregate to which it belongs. Activity belongs only to aggregates, yet the boundary separating the present from past and future must be drawn through activity. The smallest unit of real existence must therefore borrow its presentness from the nominally existing lump it composes. This is the sharpest point at the material end: the real must borrow time from the nominal.

Set beside the Southern tradition, the difference lies not in the picture but in what the picture is meant to do. The twenty-eight kinds of matter in Essay 14, Chapter 3 contain no level of atoms. The material group is the smallest unit that actually exists. Eight dhammas coarise and cocease; the group itself does not count as a dhamma, but is merely a designation.⁶ The Southern tradition does not ask whether the material group can be divided further, because further division would go beyond what contemplation can reach. The two schools seek different kinds of minimum. Sarvāstivāda wants a theoretical lower limit, where analysis can go no further; the Southern tradition wants an observational lower limit, where seeing can go no further.

This difference has another, more concrete expression. The Southern tradition must answer how material groups arise. Its answer is that karma, mind, temperature, and nutriment each produce their own groups.⁶ Sarvāstivāda has no corresponding account of the arising of its eight-constituent atomic cluster. It has great elements producing matter, but this concerns what depends on what within a single cluster, not where that cluster comes from. This book holds that the reason lies in the nature of the lower limit. An observational lower limit is encountered in direct contemplation; once encountered, it naturally invites the question of its origin. A theoretical lower limit is inferred, and an inferred thing need have no origin.

Strikingly, the two lists of eight dharmas almost coincide. The Southern tradition's pure octad group consists of the four great elements plus visible form, smell, taste, and nutritive essence. Sarvāstivāda's eight coarising constituents are the four great elements plus visible form, smell, taste, and derived touch.^{6 2} The first seven entries are identical word for word; only the eighth compartment is filled differently. This book holds that this is no accident. In listing the constituents of a minimal lump of matter, the four great elements are unavoidable, and visible form, smell, and taste allocate one item each to eye, nose, and tongue. Who gets the remaining compartment depends on what the tradition wants the list to do. The Southern tradition wants material groups to be produced by four kinds of cause, so it includes nutritive essence. Sarvāstivāda wants body-consciousness to have an object, so it includes derived touch. The same table, filled differently, discloses each school's real concern.

Chapter 1's little example can now be closed. Both schools call a pot nominal, but for different reasons. Sarvāstivāda says that when the pot is broken, cognition of the pot is gone, so the pot is nominal; the remaining visible form, smell, taste, and touch do not dissolve under analysis, so they are real. The Southern tradition says that a pot cannot be entered in the table of ultimate dhammas; what actually exists is a number of material groups. One nominal existence is reached through decomposition, the other fails to gain entry. At the material end, this difference comes fully into view: one lower limit is stipulated, the other encountered.

The two schools' minima, compared item by item:

Both schools can answer six of the seven items. Only for arising is Sarvāstivāda's compartment empty. A theoretical lower limit is inferred, and an inferred thing need have no origin: laid out in a table, that sentence is the empty compartment.

Now formally transfer real existence in the three times to atoms. Fascicle 132 explicitly says that great elements producing matter must belong to the same time as what they produce: past great elements produce past matter, present ones present matter, future ones future matter.¹ The weight of this sentence is often underestimated. It means that atoms exist not only in the present: past and future atoms also have substance, each producing the matter of its own time. Something without volume, invisible and intangible, stands solidly existent in all three times.

The sole exception is, precisely, nonmanifest material form. In its first moment, nonmanifest material form arising from manifestation is produced by great elements of the same time, just like resistant matter. From the second moment onward, if it is past or present, it is invariably produced by past great elements; if future, it can be produced by present and future great elements, or even great elements of all three times.¹ Chapter 4 said that nonmanifest material form was made to sustain the persistence of karma. Here the reverse side of that design becomes visible. For it to bear the load, the great elements that produce it must also cross the times. The only one of the eleven kinds of matter not produced by great elements of the same time is also the only one neither visible nor tangible.

Yet the same fascicle immediately gives an almost opposite account. The treatise asks why there is no possession of past and future great elements, and gives three explanations in succession. Great elements have only so much power to be possessed: possession exists only when they appear for a moment, not when they have ceased or have yet to appear. Again, the impressions of great elements are not firm, because their nature is weak and indeterminate. It is like handling an intensely fragrant flower: its scent remains after washing. Handle wood or stone, and nothing remains once one lets go. Again, if one could possess past and future great elements, one destiny would possess the great elements of all five destinies, and one body would be the body of all five. The destruction of the distinctions among destinies would also destroy the basis.¹ Here we meet Chapter 4's acquisition again. Acquisition can attach to wholesome and unwholesome dharmas across all three times, yet cannot attach to the four great elements, because they are too weak and too indeterminate. Dharma-substance exists in all three times, yet no one can possess it. Having substance but being impossible to possess: the material end makes this strange statement a settled conclusion.

There is a subtler misalignment too. Fascicle 138 records the Venerable Left-grasper's definition of the appropriated: a dharma is called appropriated if it has spatial parts, belongs to the category of sentient beings, belongs to the body, is resistant, and can be pulled and pushed. His own gloss says that "having spatial parts" excludes the past and future.¹ In this venerable master's usage, having spatial parts is precisely the mark of the present. Yet atoms must lack spatial parts; otherwise they would be divisible, and if divisible, not real. Set the two statements together: as the final unit of matter, the atom lacks precisely the mark that separates the present from past and future. This book's position is that this is insufficient to establish a contradiction: "spatial parts" need not mean the same thing in both places. But it illuminates the shape of the same trouble. The more Sarvāstivāda wants to call something real, the more it must say that it lacks properties. Yet things without properties are precisely the hardest to place in time.

The step left open in Chapter 1 is now supplied. That chapter said that the criterion presupposes an endpoint to analysis, but does not itself say who determines the endpoint. It assigned that step to this chapter. The answer is now on the page. At the end of its account of atomic measurement, fascicle 136 says: "The very least of those conditioned dharmas are one moment and one atom."¹ This is the tone of stipulation, not argument. The reason an atom is indivisible is not that analysis really cannot proceed when it reaches that point, but that there must be somewhere it cannot proceed. Otherwise analysis would have no endpoint and real existence no place to rest. The dispute over touching shows this most clearly. From beginning to end, it uses not one item of experience, only "if this, then that" inferences. Their major premise is that atoms have no finer parts, the very claim that needs proving. This lower limit is established, not demonstrated. Here Sarvāstivāda's theory of matter reaches bottom.

Over to Chapter 7. With six chapters complete, the first two columns of the table in Essay 14, Chapter 7 are filled. The left column holds the empty places in the Southern tradition's list; the right holds what Sarvāstivāda puts into the same places: the criterion of "all exists," the seventy-five dharmas, real existence in the three times, acquisition and nonmanifest material form, the six causes and four conditions, atoms and aggregation. The next chapter supplies the third column. The material end has already provided a sample: each time something is made to qualify as really existing, another rule must be added, and that rule itself does not submit to the same test. Chapter 7 asks whether six such additions, stacked together, can still stand.

References¶