Chapter 1. Mind, matter, and continuity: The questions Abhidhamma taxonomy seeks to answer¶

Chapter 0 ended by asking just what this analysis seeks to answer. This chapter answers the question.

The three words in the title offer three points of entry. The question about mind is how cognizing and its various functions are classified. The question about matter is how materiality is given in experience and analyzed to its limit. The question about continuity is how the two continue from moment to moment without an unchanging owner. The first two questions are developed in Chapters 2 and 3, while the third runs through Chapters 4, 5, and 6. This chapter first explains the tools shared by all three. It identifies the textual layer they come from and explains why it is worth taking the analysis this far. It also describes the format used to define a dhamma. Once the tools are clear, the lists in the next four chapters will be more than entries to memorize.

The chapter closes with a small example in which a phone vibrates on the table and you turn to look. The same event is told twice, once in the language of the discourses and once in that of Abhidhamma. The difference between the two accounts condenses all the questions this essay will address.

Five textual layers. Chapter 4 of Essay 3 in Part I outlined the line of transmission. Here we need only fill in the textual layers, since every number cited later in this essay must be traceable to the layer from which it comes.

The first layer does not consist of systematic works. The Dhammasaṅgaṇī lists dhammas citta by citta, and the Paṭṭhāna pairs the twenty-four conditions with dhammas in more than 130 combinations. The Kathāvatthu records debates. In the second layer, the Visuddhimagga does not belong to the Abhidhamma Piṭaka, yet it reorganizes that collection's material into the sequence of the three trainings, virtue, concentration, and wisdom. It effectively became the route through which later generations read the treatises. The third layer's compendium compresses the entire system into nine chapters and fewer than fifty pages. It is this essay's principal source and remains the standard introductory text in Theravāda countries today. The fourth layer's commentaries supply definitions omitted by the compendium, and many counts now taken for granted come from this layer. The fifth is where the chart edition used in this essay belongs.¹

The layer cited most often in this essay is the latest, and that should be recorded at the outset. Using it is no shortcut. The compendium is the result of the system's effort to organize itself, showing most clearly the form in which it finally chose to present itself. Precisely because it is the latest, its distance from the first layer is itself material for study. "The old catalog, divided anew" and the paragraphs that follow in this chapter address one such distance.

What the analysis is for. A system that divides consciousness into eighty-nine types, mental factors into fifty-two, and matter into twenty-eight can easily seem needlessly elaborate. Before judging it, ask what it seeks to achieve.

Chapter 3 of Essay 13 already noted that all the early discourses' definitions of the five aggregates are verbs: form is afflicted, feeling feels, perception recognizes, formations construct, and consciousness cognizes. Not one of the five definitions specifies components, location, or material. What is being classified is activity, not parts. The argument for not-self works through this list of verbs, asking of each in turn which can claim authority to govern itself.

Abhidhamma takes over precisely this list of verbs. Its work can be summarized in three questions. First, what should be contemplated? The discourses speak of contemplating body, feelings, mind, and dhammas, but each term in the four establishments of mindfulness requires further division. We need to know where that division should stop. Second, what units should contemplation use? We need to identify its smallest object, whether a whole person, an episode of emotion, or something else. Third, how fine should contemplation be? We need to know what level of detail is sufficient and whether going further becomes futile.

The three questions concern practice, rather than theory. The four establishments of mindfulness require practitioners to identify what is happening now. Identification needs names and names need boundaries. Those boundaries in turn need criteria. The discourse lists suffice for instruction but not for comparison cell by cell. If one is told to contemplate feeling, for instance, are an episode of discomfort and an episode of aversion one item or two? The discourses provide no uniform rule. Abhidhamma completes the list of names, turning it into a standing format for practice. The constituents of any moment of experience can be identified by using that format.

The system is therefore primarily practical in character. This book takes the view that the first criterion for evaluating it should be whether its entries can be used for identification. The identifiable part is its achievement, while what cannot be identified but remains on the register is its burden. The section "Four defining aspects" makes this criterion into a more specific test, and Chapter 7 uses it to settle the account.

The four ultimate dhammas. Abhidhamma gathers all dhammas into four classes: consciousness, mental factors, matter, and Nibbāna. Consciousness is the activity of cognizing. Mental factors are the various functions that arise and cease together with consciousness and take the same object. Matter comprises twenty-eight material phenomena, and Nibbāna is the unconditioned. The four classes total eighty-two dhammas: one consciousness, divided by function into eighty-nine or 121; fifty-two mental factors; twenty-eight kinds of matter; and one Nibbāna.²

Why these four? The division into the first three is inherited from the discourses. They distinguish mentality from materiality, and further divide mentality into consciousness and the other mental dhammas. The fourth class is added, and conspicuously so. Nibbāna is unconditioned, while the other three are conditioned. Placing the unconditioned alongside the conditioned in a single table amounts to declaring that the table is not merely a catalog of phenomena but a catalog of everything that really exists. Chapter 6 confronts the cost of this move. For now, we can leave the question open. Why should a catalog prepared for contemplation reserve a cell for something that belongs to no aggregate, neither arises nor ceases, and is associated with no consciousness?

We can compare the numbers here, with eighty-two in Theravāda and seventy-five in Sarvāstivāda. The lists differ, as do their principles of division. Sarvāstivāda separately posits fourteen conditioned factors dissociated from mind, such as acquisition, nonacquisition, the life faculty, group homogeneity, and arising, abiding, alteration, and cessation. It is an entire category absent from Theravāda.³ The next essay develops the difference. Here we should note that Sarvāstivāda is the point of comparison whenever the present essay speaks of Theravāda's restraint.

Ultimate dhammas and conceptual dhammas. The distinction is the foundation of the entire undertaking. An ultimate dhamma is identified by its own characteristic. A conceptual dhamma is a designation, something that language constructs on the basis of dhammas. People, vehicles, houses, directions, time, and shapes all fall on the conceptual side. Only ultimate dhammas are objects of insight.

The distinction first imposes a crucial discipline upon the system. To say that a vehicle is not an ultimate dhamma is to say that no real existent corresponds to the name "vehicle." What can be identified is only a collection of material dhammas and the mind's act of designation upon that collection. The same holds for a person and for the self. The whole system thus gains an internal defense. Whatever can be discussed only as a whole, and cannot be identified separately at any moment, is excluded from the table. The first instance of the restraint discussed in Chapter 7 is here.

The danger lies here too. Once the distinction is established, the question of the status of an ultimate dhamma itself arises. Is it an identified quality, or a thing established in its own right? "Two readings of own-nature" develops the question. Here we can note a piece of etymological corroboration. One traditional gloss of dhamma is "that which is supported by conditions." It is what is supported, not what supports, and the word's form places it in the passive position. On this basis, Karunadasa argues that the dhammas of Abhidhamma were never conceived as substrates bearing properties.⁴ This book accepts that reading of the etymology, while also holding that etymology cannot constrain usage. "The old catalog, divided anew" and the following paragraphs examine how usage took another path.

Four defining aspects. To define a dhamma is to fill four columns: characteristic, function, manifestation, and proximate cause. Its characteristic is what distinguishes it from every other dhamma. Its function is what it accomplishes or the role it fulfills. Its manifestation is how it appears in contemplation, or the result it leaves. Its proximate cause is the nearest condition necessary for its arising.

For contact, the four columns read as follows. Characteristic: contacting the object. Function: bringing object and consciousness into collision. Manifestation: arising through the conjunction of base, object, and consciousness, and bringing about the arising of feeling, perception, and formations. Proximate cause: an object appearing at one of the doors.⁵ Once the four columns are filled, the name "contact" has boundaries. It refers to the encounter between an object and consciousness at any door, and has nothing to do with tactile sensation in the psychological sense. The benefit is immediately apparent, since an abstract term is reduced to four questions that can be checked.

The format of four columns itself has a history. Bhikkhu Bodhi notes that it was taken from a pair of ancient exegetical works, not invented by Abhidhamma. Abhidhamma institutionalized it as the uniform format for the whole table.⁶ A discourse may explain a word differently on different occasions. Once the four columns have been filled, however, the word has a fixed place in the whole table.

This book uses the four columns as its criterion for evaluating the whole system. Where all four can be filled, the account is phenomenological. It concerns what can be identified in experience, by a method others can check. Where all four cannot be filled but are filled nonetheless, the account is metaphysical. Entries are supplied to keep the table tidy, and what is supplied cannot be checked against experience. The criterion does not decide in advance which cell belongs in which category. It offers a way to investigate, which the following chapters apply to each class in turn.

A possible misunderstanding should be cleared up here. The four columns are not four definitions, but four aspects of one dhamma. An account with even one column missing does not qualify as a definition. The column for proximate cause in particular binds every dhamma to conditions. Any dhamma for which a proximate cause can be supplied cannot be self-sufficient. Of all the columns in the system, this one is least susceptible to ontologization. Chapter 7 returns to it.

The old catalog, divided anew. Abhidhamma did not abolish the five aggregates. It retained them unchanged in the table and added a correspondence:⁷

The table reads smoothly, yet each row rewrites something. The form aggregate changes from a class of activity into twenty-eight entries. Feeling and perception each contract to a single mental factor. Consciousness spreads out into eighty-nine cittas. The formations aggregate is obtained by subtracting two from fifty-two to give fifty. Nibbāna receives a separate cell outside the aggregates because it belongs to none of them.

The benefits are real, and we should consider them first. The table connects the five aggregates to the workings of the whole system. Once we know that the feeling aggregate is that single mental factor of feeling, we can follow the connections to the four columns defining it. We can also look up the cittas with which it is associated and the stage of a cognitive process at which it occurs. The five aggregates become five regions of the entire table, in place of five terms requiring separate explanation. The convenience for teaching is enormous. The cost lies in grammatical category.

The cost lies in grammatical category.

Constructing, and those fifty items. Consider the formations aggregate. The discourse defines formations with a verb. They construct the conditioned, hence they are called formations. In relation to form they construct formness, and in relation to feeling, perception, formations, and consciousness they construct their respective characters. The Chinese parallel says, "Constructing is the characteristic of the clinging aggregate of formations. What does it construct? It constructs in relation to form."⁸ The defining term is an ongoing event.

Chapter XIV of the Visuddhimagga still observes the definition. Everything with the characteristic of constructing is collectively called the formations aggregate. It is a definition by characteristic, of the same kind as the discourse's. More significant is the way it enumerates the entries citta by citta. Buddhaghosa counts thirty-six constituents of the formations aggregate arising with the first wholesome sense-sphere citta: twenty-seven fixed, four supplementary, and five inconstant. A greed-rooted citta has seventeen constituents of the formations aggregate, while a hatred-rooted citta has eighteen.⁹ Within the same book, the number belonging to the aggregate changes because it changes with the citta. Buddhaghosa never gives a flat list of fifty items.

The number could vary because the list of names had not yet been closed. When the Dhammasaṅgaṇī lists dhammas citta by citta, it leaves an open clause at the end, "or whatever other immaterial dhammas arise at that time in dependence on conditions." Bhikkhu Bodhi notes that the clause presupposes an open universe of mental factors. The commentaries closed it by naming each factor under the heading of "supplementary dhammas."¹⁰ The four supplementary items among Buddhaghosa's thirty-six use the passage that was still open.¹¹

After the list was closed, the compendium fixed the total of mental factors at fifty-two. Only at this point could the formations aggregate first be defined by subtraction. Deduct feeling and perception from fifty-two, and the remaining fifty are the formations aggregate.

For the subtraction to work, the denominator must first be closed. The section's most significant observation concerns the closing of an open list that varied with the citta into a fixed catalog. That was the step by which formations changed from the constructing taking place now into "those fifty items." This book takes the view that the five aggregates were not reified by the announcement of any doctrine. Reification occurred through a cataloging procedure. The procedure itself is unobjectionable, and every step has a reason. Teaching needs definite numbers and comparison needs a fixed number of entries. Meditation instruction must be able to answer students' questions. Yet when the procedure is complete, the grammatical category of "formations" has changed. Originally a nominal form of a verb, it referred to the activity of constructing. Thereafter it is a list, referring to those fifty things. No one announced the change because no one ever made that decision.

Two limits on the claim. For this judgment to hold, two boundaries must first be drawn.

First, defining the formations aggregate by subtraction already occurs in the Dhammasaṅgaṇī. When dhammas are enumerated citta by citta, the formations aggregate is the associated dhammas minus feeling and perception. What changes is the nature of the denominator, since subtraction is present on both sides. An open denominator specific to each citta becomes a closed, standing one. The denominator is the subject of this section.

Second, Theravāda has no conditioned factors dissociated from mind. As noted above, Sarvāstivāda's formations aggregate separately includes acquisition, nonacquisition, the life faculty, group homogeneity, arising, abiding, alteration, cessation, and so on. It adds an entire set of entities dissociated from mind, a step Theravāda did not take. The reification discussed here is therefore a milder kind. Nothing is added, but a table of functions is read as a register of things. Milder does not mean absent. Chapter 7 explains where its cost falls.

Four ways of dividing the same material. The chapter on the compendium of categories takes the operation further. It supplies not one correspondence but four: aggregates, bases, elements, and truths. The same ultimate dhammas are divided once by the five aggregates, once by the twelve bases, once by the eighteen elements, and once by the four truths. The twelve bases comprise six internal and six external bases, dividing by the doors through which cognition occurs. The eighteen elements add six consciousness elements to the twelve bases, dividing by cognition's three-sided structure.¹²

The consciousness aggregate is worth examining under these divisions, because it most clearly shows the relation between a division and an entity. Divided by aggregates, consciousness is a single group of eighty-nine cittas. Divided by elements, those same cittas fall into seven consciousness elements. Eye-, ear-, nose-, tongue-, and body-consciousness each comprise two cittas, making ten, the so-called two sets of five sense-consciousnesses. The mind element comprises three: five-door adverting and the two receiving cittas. The mind-consciousness element comprises the remaining seventy-six. Ten plus three plus seventy-six is still eighty-nine.

The same dhammas make the consciousness aggregate when arranged in one group, and the seven consciousness elements when arranged in seven. Which is their true form? The question is already mistaken, since the divisions serve different purposes of contemplation rather than describing an internal structure. Yet when several divisions stand alongside one another in a table, they look like several plans for the same piece of furniture. A plan always points to a definite piece of furniture. This book takes the view that the second source of pressure toward reification lies here. Parallel descriptions themselves suggest an object they all describe.

Three scholarly positions. The judgment above is not a scholarly consensus. The opposing positions must be represented correctly.

Ronkin argues for a shift. In her view, Abhidhamma applies to discrete momentary events the vocabulary used in the discourses to describe continuous processes. Dhammas change from phases of processes into types of events. Both Theravāda's eighty-two types and Sarvāstivāda's seventy-five result from this rewriting.¹³

Karunadasa and Gethin argue for continuity. Karunadasa's argument, cited above, is that a dhamma is supported by conditions and there is no further thing called a dhamma apart from its own-nature. The refinement has not yet introduced a view of existence contrary to early Buddhism. Discussing the treatment of the aggregates in early Abhidhamma, Gethin concludes that the Nikāyas and early Abhidhamma can be regarded as a broadly consistent whole.¹⁴ Gethin's conclusion does not support this section's argument. He stands on the other side of it, and his position must be made clear.

This book's position lies between the two, while its focus differs from both. At the level of doctrine, it accepts continuity. Theravāda Abhidhamma never posited an ontological claim contrary to the early teaching, and it always understood itself as analysis. But change need not occur at the level of doctrine. The argument here concerns a change of format, which requires no announcement and leaves no declaration in any treatise. An open list varying with the citta and a fixed catalog of fifty entries can be entirely consistent doctrinally and entirely different in use. The former can identify things only in the present act of contemplation. The latter can be memorized, tabulated, and counted as the world's components apart from contemplation. Doctrine has not changed, but what can be done has.

The essay will repeatedly return to the two sides of this operation. The same operation is both the system's achievement and its burden. Chapter 7 takes stock of the achievements and reckons the burdens.

Two readings of own-nature. Abhidhamma uses sabhāva to define dhammas as what has own-nature. The term admits of two readings.

The first reads it as an identifiable quality. To say that a dhamma has own-nature means that it has a characteristic of its own by which we can distinguish it from other dhammas in contemplation. The earth element's own-nature is hardness, and contact's is contacting the object. On this reading, own-nature is a term of identification, with no claim about a mode of existence. To say that a dhamma has own-nature is to say that it has a seat in the table that cannot be confused with another entry's.

The second reads it as an essence established in its own right. To say that a dhamma has own-nature means that it is what it is without depending on anything else to be itself. On this reading, own-nature is an ontological term, and this is precisely what Madhyamaka seeks to refute.

Both readings find support in the texts, and a chronological line can be traced. Examining the Paṭisambhidāmagga, Ronkin argues that sabhāva was first used there to define dhammas. The ontological reading had found concrete expression in the work of the sixth-century commentator Mahānāma. He called dhammas real existents (sat) and introduced momentariness, whereas the Paṭisambhidāmagga neither accords dhammas any ontological status nor mentions momentariness. Mahānāma himself still oscillated between epistemological and ontological readings, and the latter had not yet become settled in his work.¹⁵ If this line of development holds, the second reading belongs to a later phase of Abhidhamma, not its beginning. Karunadasa's counterreading is also forceful. He notes that if arising, presence, and dissolution were real dhammas, secondary characteristics would have to be posited for them, and so on without end. Abhidhamma itself rejected that route. A system truly prepared to regard dhammas as independent entities would not have stopped here.¹⁶

This chapter does not adjudicate the issue. We need only remember that the term marks the most important fork in the rest of this essay. Madhyamaka attacks the second reading when it confronts own-nature, whereas Theravāda itself may not stand consistently on that side. Much of the restraint Chapter 7 will argue for depends on the first reading. The vacancies Chapter 7 will identify lie mostly in the unguarded stretch between the two.

One cognition, two accounts. The preceding discussion has concerned the framework. To see whether it is useful, we need to watch it run once.

The scene: a phone vibrates on the table, and you turn to look.

First, let us describe the event in the language of the discourses. The contact and resistance encountered are form. The discomfort or anticipation that follows is feeling. Recognizing that the phone is vibrating is perception. The intention to turn your head has taken shape, and that is formations. Cognizing throughout the sequence is consciousness. With the five verbs, the account ends. The discourses have good reason to stop here, since their aim is for practitioners to see on the spot that none of the five is self. Once that is seen, the task is complete.

Now let us describe it in the language of Abhidhamma. The account provides a fixed worksheet, each cell with its name, position, and configuration of mental factors. Consciousness begins in bhavaṅga, the underlying stream of continuity that takes no external object. The vibration strikes the body door. Bhavaṅga first vibrates for two moments, then is cut off. Five-door adverting arises, directing the continuum toward that door. Body-consciousness arises, merely experiencing touch. Receiving follows, taking over the object. Investigating assesses what kind of object it is. Determining makes the choice. Then come seven javanas, the only stage at which kamma is produced. They are followed by two registrations that trail the object and take it up once more. Consciousness subsides into bhavaṅga. The entire cognitive process ends here, having taken seventeen mind-moments. A subsequent mind-door process takes over, linking the preceding series of objects to piece together the concept of a phone.¹⁷

A traditional analogy for the sequence can be read alongside it. A person sleeping under a tree is bhavaṅga, and a mango falling to the ground is the object striking the door. The person waking is adverting, and opening the eyes to look is eye-consciousness. Reaching to pick up the fruit is receiving, and examining it is investigating. Sniffing it is determining, and eating it is javana. Swallowing the aftertaste is registration. The person then returns to sleep, back to bhavaṅga.¹⁸

The second account adds three things to the first. First, it locates kamma at the javana stage. The preceding adverting, fivefold sense consciousness, receiving, and investigating produce no kamma. Only the seven javana moments do so. The distinction has practical consequences as well as classificatory ones. An offense can carry more or less weight depending on which mental factors arise in the javana stage, not on the event itself. Second, it provides a worksheet with no observer. The first account's five verbs can still readily be heard as "I see" and "I think." None of the second account's dozen or more cells is called "the one who sees." Each is the result of the preceding cell and a condition for the next. Third, it gives contemplation a workable unit. When told to contemplate the impermanence of mind, what does one contemplate? The first account cannot answer, while the second points to the arising and ceasing of any cell in the sequence.

The cost is in the second account too. Once you ask which dhammas occur in this cognition, the worksheet becomes a register. Each cell on the worksheet was an event in progress, while each entry in the register can be counted away from the scene. Nominalization occurs here, and it is necessary. Without asking the question, neither teaching nor checking would be possible. Yet as soon as it is asked, the five aggregates and the dhammas have already moved a little from ongoing activities toward things that can be listed. The cataloging procedure described in "Constructing, and those fifty items" is repeated in every student.

The next four chapters supply precisely this register. We should read it with both of this section's insights in mind. It is an extraordinarily detailed worksheet, and a register extremely easy to mistake for a catalog of the world.

On to Chapter 2. Once the four columns for defining a dhamma become a machine for definition, consciousness is the first object it processes thoroughly. Eighty-nine cittas are not eighty-nine minds. They are eighty-nine specifications of the same event, obtained by crossing several axes. The next chapter starts here.

References¶