Abhidhamma Interpretations of “Persons” (puggala): with Particular Reference to the Aṅguttara Nikāya

The Abhidhamma and the Aṅguttara Nikāya For most Buddhist schools, the Buddhist Canon comprises the three Pit˙ akas (Collections), namely the Vinaya-, the Su¯tra- and the Abhidharma-Pit˙ aka. The Abhidharma (P. Abhidhamma) is generally regarded as later than the Su¯tra (P. Sutta) material of the Nikāyas and Āgamas.1 It is widely held that the Abhidharma presents analysis and interpretation of the Buddha’s teachings in such a way that it mostly concerns dharmas (P. dhamma), which are seen as constituents of reality or ultimate constituent elements that really exist; by contrast, persons (P. puggala, Skt.pudgala) do not really exist since a person is analyzed by Abhidharma into dharmas, which are all there is
The Aṅguttara Nikāya is included in the Sutta Pit˙ aka of the Pali Canon, but it can be shown to be closely related to the Abhidhamma. Its framework is based on a numerical scheme according to which each successive section (nipāta) deals with sets of terms one number more than those dealt with in the preceding section. This feature is also characteristic of some Abhidharma works, such as the Puggalapaññatti of the Therava¯dins and the Saṃgītiparyāya 集異門足論 (Ji yimen zu lun, T 1536) of the Sarva¯stiva¯dins. Besides, the Aṅguttara Nikāya’s method of arranging sets of items according to certain principles is similar to that used in constructing the mātṛkās (P. mātikā), or comprehensive lists of the fundamental doctrinal items, as found in many Abhidharma texts. There are different explanations for the historical origin and development of Abhidharma texts. As Cox (1995, p. 8) points out, most Western scholars contend that Abhidharma treatises evolved from the practice of formulating matrices or categorizing lists (mātṛkā/mātikā) of all topics of the teaching arranged according to both numerical and qualitative criteria.3 Bronkhorst (1985, p. 307) notes that the development of such lists had run its course well before the final redaction of the Su¯tra Pit˙ aka. He says that the later tradition which ascribes Abhidharma to S´a¯riputra already finds expression in the Saṅgīti Sūtra and Daśottara Sūtra of the Dīrgha Āgama (equivalent to the Saṅgīti Sutta and Dasuttara Sutta of the Dīgha Nikāya), and thus we have “another indication that the Abhidharma-Vibhan˙ga existed prior to the completion of the Su¯trapit˙ aka” (ibid., p. 316). Frauwallner (1995, pp. 4, 7) also counts these two sūtras as pertaining to the earliest Abhidharma. Bronkhorst (1985, p. 316) says: “[T]he early existence of some kind of Abhidharma would explain the peculiar shape of the Su¯trapit˙ aka, or rather of two sections of it, the Saṃyuktāgama / P. Saṃyutta Nikāya and the Ekottarāgama / P. Aṅguttara Nikāya.” Therefore he asserts the influence of early Abhidharma on the sūtras (ibid., p. 317) and states: “Ma¯tr ˙ ka¯s, and even one or more Abhidharma works, were in existence well before the completion of the Su¯trapit˙ aka.” (ibid., p. 318) Accordingly, the peculiar structure and other features of the Aṅguttara Nikāya can be explained by the influence of early Abhidhamma on this Nikāya. Even some suttas in the Nikāyas were actually Abhidhamma as Bucknell and Stuart-Fox (1993, pp. 27–28) suggest.

Wogihara (1935, front-matter 1–3) takes a different viewpoint on the relation between the Aṅguttara Nikāya and Abhidhamma. He suggests that the Saṅgīti Sutta  and Dasuttara Sutta of the Dīgha Nikāya may have served as precursors of the Aṅguttara Nikāya, and that these two suttas, composed of numerical lists, were later expanded and transformed into many suttas of the Aṅguttara Nikāya (ibid., frontmatter 1–2).4 Since these two suttas of the Dīgha Nikāya are attributed to Sa¯riputta (S´a¯riputra), the first person to preach Abhidhamma according to the tradition, they can be regarded as forerunners of Abhidhamma (ibid., front-matter 2–3). By inference, Wogihara (ibid., front-matter 3) maintains that it is no exaggeration to say that the Aṅguttara Nikāya is the source of Abhidhamma. Similarly, Reat (1996, p. 25) thinks that the Saṅgīti Sutta and Dasuttara Sutta prefigure the organization of the Aṅguttara Nikāya and Saṃyutta Nikāya, which in turn foreshadow the Abhidhamma literature.5 Similarly, Cousins (1983, pp. 3–4) suggests that many of the lists in the Saṅgīti Sutta “must derive from suttas found only in the Anguttaranika¯ya”, and points out that this sutta “is used as the basis for one of the seven canonical abhidharma works of the Sarvāstivāda”, namely the Saṃgītiparyāya. 6 These two views are not diametrically opposite. Accepting the two possibilities, Cox (1995, pp. 9–10) states: [T]he Aṅguttaranikāya adopts a numerical arrangement in which entire sūtras are classified according to the number of items represented by their primary topic. … Even though these modes of organization may indicate antecedents in the sūtra of structural procedures that were to fully develop in the Abhidharma literature, … they may also reflect, in individual cases, the influence of Abhidharma upon the sūtra collections themselves. For the final redaction of the sūtras certainly does not entirely precede but also overlaps the composition of Abhidharma works. This paper is mainly concerned with the influence of early Abhidhamma on the Aṅguttara Nikāya and early traces of the Abhidhamma material or protoAbhidhamma found in this Nikāya, but I shall first illustrate a case in which a sutta of the Aṅguttara Nikāya provides a list of persons which evolves into items of mātikā and new doctrines in various Abhidharma works. This case also shows that in the course of such evolution, this sutta’s list of persons was subject to “depersonalization”, which is typical of some Abhidharma texts. As noted by Bodhi (2012, p. 58), the Aṅguttara Nikāya is distinguished among the four Nikāyas by its interest in defining and describing types of persons. He says (ibid., p. 25): The An˙guttara … abounds in different ways of classifying people … and it gives primacy to their qualities, their struggles for happiness and meaning, their aspirations and attainments. The An˙guttara thus became the inspiration and a major source for one of the books of the Abhidhamma Pit˙ aka, the Puggalapan˜n˜atti. Concrete evidence for this has been provided by Morris (1883, p. x), who points out that nearly the whole of the third, fourth and fifth sections (Tayo Puggala¯, Catta¯ro Puggala¯, Pan˜ca Puggala¯) of the Puggalapaññatti are found in the corresponding sections (Tika-nipa¯ta, etc.) of the Aṅguttara Nikāya. 7Below I will demonstrate that the Aṅguttara Nikāya, besides being a chief source for the Puggalapañatti, provided several points of reference to other Abhidhamma texts as well


Sutta 3.25 in the Aṅguttara Nikāya (hereafter AN 3.25) expounds the following three kinds of persons (translation mostly by Bodhi 2012, pp. 219f.): (1) What is the person whose mind is like an open sore? Some person is prone to anger and easily exasperated. … (2) What is the person whose mind is like lightning (vijjūpama-citto puggalo)? Some person understands as it really is: “This is suffering” … “This is the way leading to the cessation of suffering.” Just as, in the dense darkness of night, a man with good sight can see forms by a flash of lightning, so too some person understands as it really is: “This is suffering” … (3) What is the person whose mind is like a diamond (vajirūpama-citto puggalo)? With the destruction of the taints, some person … dwells having attained the taintless liberation of mind, liberation by wisdom. Just as there is nothing that a diamond cannot cut, whether gem or stone, so too, with the destruction of the taints, some person …8 This sutta outlines the spiritual progression from an ordinary state to liberation. These three types of persons are listed in the mātikā of the Puggalapaññatti (p. 4) and this Therava¯din Abhidhamma text (p. 30) repeats, almost verbatim, part of AN 3.25 to serve as its exposition of the three items. As Law (1933, p. 48) suggests, that the Puggalapaññatti and the Vibhaṅga show close affinity with the sutta material indicates that they are probably the earliest of the seven canonical Abhidhamma treatises in Pali. Rhys Davids (1903, p. 188), Norman (1983, p. 102), Mizuno (1997, p. 262) and Willemen et al. (1998, p. 13) also regard the Puggalapaññatti as the earliest Pali Abhidhamma text, while Bronkhorst (1985, pp. 309f.) and Cox (1992,p. 156) indentify the Vibhaṅga as reflecting the earliest stage.9 The Puggalapaññatti, faithfully following AN 3.25, does not at all “interpret” the three items in this sutta of the Aṅguttara Nikāya. An interpretation is found in another Therava¯din Abhidhamma text, the Dhammasaṅgaṇi. Two of the above three items are alluded to in the Dhammasaṅgaṇi. The second dyad among the 42 dyads in the suttanta-mātikā of this text (p. 7) is as follows: (1) lightning-like states (vijjūpamā dhammā); (2) diamond-like states (vajirūpamā dhammā). This pair of “states” (dhamma) instead of “persons” (puggala) is apparently derived from the last two of the three persons in AN 3.25. The Dhammasaṅgaṇi explains “lightning-like states” as “insight into the three lower noble paths”10 and “diamondlike states” as “insight into the highest path of arahantship”.11 This interpretation seems to conform to the purport of the above sutta as far as it makes a distinction between the two items in terms of spiritual level with diamond-like states being higher than lightning-like states. While AN 3.25 has no parallel in the extant Chinese Āgamas (two of which belong to the Sarva¯stiva¯da, see below), this sutta apparently has a Sarva¯stiva¯da counterpart which is interpreted by the Abhidharma of this school. The Saṃgītiparyāya (T 1536), among the oldest of the seven canonical Sarva¯stiva¯din Abhidharma works,12 refers to the above three persons in AN 3.25 as three minds (三心). In its chapter on sets of three dharmas, these three types of mind are expounded in detail. Below are only some main points. (1) Why is that mind called “like an open sore”? Because when that mind contacts adverse circumstances, it produces a wide variety of defilements. (2) Why is that mind called “like lightning”? Because that mind attains the fruit of non-return (anāgāmin) and it can shine for a while but soon goes out. (3) Why is that mind called “like a diamond”? Because that mind attains the fruit of one-beyond-training (aśaikṣa/asekha, i.e. the arhat/arahant) and there are no fetters (saṃyojana) and so on that it cannot destroy.13 Similar to “diamond-like states” in the Dhammasaṅgaṇi, “the mind like a diamond” in the Saṃgītiparyāya is associated with the arahant, the fully liberated person without any taints or fetters. This corresponds well with the above description of the “person whose mind is like a diamond” in AN 3.25: “with the destruction of the taints, some person … dwells having attained the taintless liberation of mind liberation by wisdom”. However, “the mind like lightning” in the Saṃgītiparyāya does not match the “lightning-like states” of the Dhammasaṅgaṇi in that the former refers to the fruit of non-return while the latter relates to the three lower noble paths. It should be noted here that what is described as a “person” in sutta 3.25 is replaced by “dhamma” (state) and “mind” respectively in the two Abhidharma books, the Dhammasaṅgaṇi and the Saṃgītiparyāya. This suggests purposeful modification of the sutta/sūtra terminology in line with the Abhidharma tendency to “depersonalize”, which however does not necessarily entail the exclusion of persons from the Abhidharma exposition

There was still further development in the Abhidharma from the foregoing idea in the Aṅguttara Nikāya. As Frauwallner (1995, pp. 177f.) notes, the “person whose mind is like a diamond” in this Nikāya was developed by the Sarva¯stiva¯din Abhidharma in such a way that it led to “felicitous invention” (p. 178) of the term vajropama-samādhi, “diamond-like concentration”, which is used to designate the final “immediate path” (無間道, 無礙道, ānantarya-mārga) to liberation, wherein the last remaining defilements (anuśaya) are eliminated and thus all defilements are eradicated. The “diamond-like concentration” in this sense is found in several Abhidharma texts, including the Jñānaprasthāna 發智論 (Fazhi lun, T 1544),14 one of the canonical Abhidharma works of the Sarva¯stiva¯dins, and some post-canonical Abhidharma texts such as the *Abhidharmahṛdaya 阿毘曇心論 (Apitan xin lun, T 1550)15 and the Abhidharmakośabhāṣya. 16 In sum, sutta 3.25 of the Aṅguttara Nikāya and its Sarva¯stiva¯da counterpart provided source materials for the architects of the Abhidharma of different schools at different times to construct and elaborate their philosophical systems. Lamotte (1988, p. 184) aptly remarks that “the Abhidhamma abounds in repetitions, rectifications, reclassifications and explanations which give it the character of an unfinished work still in the process of elaboration”.


The mātikā of the Puggalapaññatti (p. 8) contains a set of four persons: (1) the unshaken ascetic (2) the red-lotus ascetic (3) the white-lotus ascetic (4) the delicate ascetic among ascetics

This set of four persons is also listed in the Saṅgīti Sutta of the Dīgha Nikāya (no. 33 at DN III 233) and four suttas of the Aṅguttara Nikāya (4.87–90 at AN II 86–91). Therefore, this set of four ascetics, which the Aṅguttara commentary (Mp III 113) refers to as a mātikā, probably belongs to a very old mātikā that already existed before the completion of the Nikāyas. The Saṅgīti Sutta (DN 33) mentions these four ascetics without any explanation. In contrast, a series of four suttas (4.87–90) in the Aṅguttara Nikāya defines each of them in four different ways, but three of them (4.87, 4.89, 4.90) agree partly on the referents of these four ascetics. In sutta 4.87 the four kinds of persons are defined as follows (abridged)17: (1) the unshaken ascetic: a monk is a trainee (sekha) practising the way who dwells aspiring for the unsurpassed security from bondage. (2) the white-lotus ascetic: with the destruction of the taints, a monk … dwells having attained the taintless liberation of mind, liberation by wisdom; yet he does not dwell having touched personally18 the eight emancipations. (3) the red-lotus ascetic: with the destruction of the taints, a monk … dwells having attained the taintless liberation of mind, liberation by wisdom; and he dwells having touched personally the eight emancipations. (4) the delicate ascetic among ascetics: a monk usually uses a robe that has been specifically offered to him … he usually eats almsfood that has been specifically offered to him … lodging … medicines … His fellow monks … usually behave toward him in agreeable ways … With the destruction of the taints, he … dwells having attained the taintless liberation of mind, liberation by wisdom. If one could rightly say of anyone: “He is a delicate ascetic among ascetics”, it is precisely of me [the Buddha referring to himself] that one might say this.19 Let us first look at the definitions of the second and third types of persons. As Bodhi (2012, pp. 57f.) points out, these two definitions introduce a distinction between two kinds of arahants. The white-lotus ascetic refers to the arahant “liberated by wisdom” while the red-lotus ascetic refers to the arahant “liberated in both respects”.20 These two kinds of arahants will be examined in the next section of this paper. The first type of ascetic is said to be a trainee (sekha) aspiring for the unsurpassed security from bondage, so this represents a practitioner who is not yet liberated. The fourth type of ascetic refers to the Buddha himself, a perfectly liberated person who enjoys more comfort and respect than other monks. The set of four ascetics as defined in this way shows a sense of hierarchy. Sutta 4.89 (AN II 89f.) defines the unshaken ascetic as possessing right view, right intention, right speech, right action, right livelihood, right effort, right mindfulness and right concentration, which constitute the Noble Eightfold Path. By contrast, the white-lotus ascetic and the red-lotus ascetic are not only endowed with the Noble Eightfold Path, but also with right knowledge and right liberation. This implies that these two types of ascetics are liberated arahants, while the unshaken ascetic is not yet liberated. Similar to sutta 4.87, sutta 4.89 describes the white-lotus ascetic as one who “does not dwell having touched personally the eight emancipations”, and the red-lotus ascetic as one who “dwells having touched personally the eight emancipations”. Accordingly, the white-lotus ascetic and the red-lotus ascetic as defined in sutta 4.89 also refer respectively to the arahant “liberated by wisdom” and the arahant “liberated in both respects” as explained in the Pali commentary.21 For the fourth type of person, the delicate ascetic among ascetics, sutta 4.89 gives the same definition as that in sutta 4.87, i.e. the Buddha. Sutta 4.90 (AN II 90f.) defines the unshaken ascetic as “a trainee who has not attained his mind’s ideal and dwells aspiring for the unsurpassed security from bondage”, which is almost identical to the definition of the unshaken ascetic in sutta 4.87. The white-lotus ascetic and the red-lotus ascetic are both depicted as a monk who dwells contemplating rise and fall in the five aggregates subject to clinging, but the former “does not dwell having touched personally the eight emancipations” while the latter “dwells having touched personally the eight emancipations”. The definitions of these two ascetics make no mention of their liberation, and the Pali commentary glosses both of them as trainees (sekha) with the distinction that the white-lotus ascetic produces no jhānas whereas the red-lotus ascetic attains the eight emancipations.22 In this sutta, the delicate ascetic among ascetics is again defined in the same way as in sutta 4.87. While the above three suttas appear to define the four kinds of persons in a somewhat similar way, sutta 4.88 defines them in a totally different fashion as follows (in brief): (1) the unshaken ascetic: a stream-enterer (sotāpanna). (2) the white-lotus ascetic: a once-returner (sakadāgāmin).

(3) the red-lotus ascetic: one of spontaneous rebirth, due to attain final Nirvana there without returning from that world (opapātiko hoti tattha-parinibbāyī anāvattidhammo tasmā lokā), i.e. a non-returner (anāgāmin). (4) the delicate ascetic among ascetics: with the destruction of the taints, a monk … dwells having attained the taintless liberation of mind, liberation by wisdom; in other words, an arahant.23 This sutta corresponds to sūtra 7 in chapter 28 of the Ekottarika Āgama (hereafter EA¯ 28.7) of probably Maha¯sa¯m ˙ ghika origin.24 This sūtra defines four kinds of ascetics in somewhat different terminology and sequence: (1) the ascetic like a yellow-blue-flower: a stream-enterer. (2) the white-lotus ascetic: a once-returner. (3) the delicate ascetic: a non-returner. (4) the delicate ascetic among the delicate [sic]: a person has destroyed the taints, attaining the taintless liberation of mind, liberation by wisdom; in other words, an arahant.25 Thus, sutta 4.88 of the AN, likewise EA¯ 28.7, expounds this set of four persons in terms of the four fruits by the stock description of these four spiritual types that recurs in the Nikāyas.26 Only the fourth kind, the delicate ascetic among ascetics, represents a liberated arahant. The white-lotus ascetic is not distinguished from the red-lotus ascetic by his lack of the eight emancipations, and neither of the two ascetics refers to an arahant. Such a significant deviation from the definitions in the other three suttas of the Aṅguttara Nikāya is surprising since all these four suttas are presented in the form of the Buddha’s discourses delivered to his disciple monks. It is inconceivable that the Buddha should have defined these four types of persons in such contradictory ways that his disciples would have been confused.27 Even the three “similar” suttas also diverge considerably in several aspects as discussed above.28 The only plausible explanation for this anomaly is that these diverse definitions were formulated by different people instead of just one man, the Buddha. Who were they? At least some of them were probably Abhidhamma composers or Abhidharmists as will be elucidated below. Just as sutta 4.88 of the Aṅguttara Nikāya (hereafter AN 4.88), the Puggalapaññatti (p. 63) also explains this set of four ascetics in terms of the four fruits. Even the four definitions in this Abhidhamma text are almost identical in wording to those in AN 4.88, but curiously the definition of the white-lotus ascetic and the definition of the red-lotus ascetic are exchanged; or put more accurately, the places of the two terms “white-lotus ascetic” and “red-lotus ascetic” are swapped. This divergence should be viewed against the broader background of textual compilation. When the Nikāyas and Abhidhamma were compiled, there appeared to be confusions not only in defining the four kinds of ascetics but also in arranging the sequence of these four persons. In the Saṅgīti Sutta (DN 33) the sequence is: 1. the unshaken ascetic 2. the red-lotus ascetic 3. the white-lotus ascetic 4. the delicate ascetic29 In the four suttas of the Aṅguttara Nikāya, however, the sequence is: 1. the unshaken ascetic 2. the white-lotus ascetic 3. the red-lotus ascetic 4. the delicate ascetic among ascetics It should be noted, in addition, that the Aṅguttara Nikāya version of the fourth person contains the expression “among ascetics” (samaṇesu), which is absent from the DN 33 version. The Puggalapaññatti draws heavily on the Aṅguttara Nikāya as pointed out by the scholars mentioned above. Just like the Aṅguttara Nikāya, the Puggalapaññatti has the expression “among ascetics” for the fourth type of ascetic in the mātikā (p. 8) and its exposition (p. 63). However, the sequence of these four types in the Puggalapaññatti follows that in DN 33, i.e. “red-lotus ascetic” before “white-lotus ascetic”. While following the sequence of the terms in DN 33, the Puggalapaññatti seemingly copies the definitions of the four terms from AN 4.88 in exactly the same order as the four definitions appear in this sutta. Consequently, the definition of the white-lotus ascetic (second ascetic) in AN 4.88 is placed under the term “red-lotus ascetic” (second ascetic) in the Puggalapaññatti, and the definition of the red-lotus ascetic (third ascetic) in AN 4.88 is placed under the term “white-lotus ascetic” (third ascetic) in the Puggalapaññatti. Therefore the definitions of these two kinds of ascetics in one text are reversed in the other. It is difficult to assert which of the two versions is original, but considering the general assumption that Abhidhamma texts are founded on suttas, the Puggalapaññatti version may be a secondary development from the above two Nikāyas. Alternatively, it is possible that this Puggalapaññatti version originated from certain ancient Sutta literature or perhaps more likely Abhidhamma literature, which then evolved in different directions into the Saṅgīti Sutta (DN 33) and sutta 4.88 of the Aṅguttara Nikāya as well as EA¯ 28.7. This may also explain why the Puggalapaññatti version bears the different characteristics of both suttas extant today. Let us now return to the point that even the Aṅguttara Nikāya itself describes diverse ways of defining the four types of ascetics in a series of four suttas. Only the version of sutta 4.88 is adopted by the Puggalapaññatti presumably because the author of this shortest canonical Abhidhamma work wanted to give a succinct and standard definition of these four kinds of persons. In contrast, the compilers of the Aṅguttara Nikāya faithfully preserved four versions of the definitions and allocated them to four suttas in a series, and apparently admitted the existence of disagreement on how to define or interpret the four types of ascetics. In this connection, we may venture to suggest that at least some versions of these definitions in the Aṅguttara Nikāya were interpolated into this “Sutta literature” as a kind of “Abhidhamma”, even if the possibility cannot be excluded that a certain version/versions might contain the Buddha’s own interpretation. It was not uncommon for Buddhists to disagree about Abhidhamma even during the Buddha’s lifetime according to the Kinti Sutta of the Majjhima Nikāya (no. 103). In this sutta the Buddha is depicted as saying: “While you are training in concord, with mutual appreciation, without disputing, two monks might make different assertions about the Abhidhamma.”30 These four suttas in the Aṅguttara Nikāya provide us with such different assertions about the Abhidhamma. It should be noted that “Abhidhamma” in this sense cannot be equated with the Abhidhamma Pit˙ aka, but nevertheless, as Kimura (1968, pp. 31–33) and Willemen et al. (1998, pp. 12–13) suggest, such discussions or debates on the Abhidharma among the Buddha’s disciples can be seen as the inception of the Abhidharma literature. When Watanabe (1983, p. 37) raises the question: “What is then the original (or simple) form of the Abhidhamma Pit˙ aka?”, he supplies an answer by drawing our attention to the following two points regarding “the [Buddha’s] disciples’ attempts at the elementary philosophical study of dhammas”: (1) defining dhammas, (2) arranging dhammas in numerical order. Defining the items and arranging them in numerical sequence are exactly the issues involved in the four kinds of ascetics discussed above.


“Liberated by Wisdom” and “Liberated in Both Respects”

 Let us now move on to the interpretations of “liberated by wisdom” and “liberated in both respects”. First, we should look into the Sutta Pit˙ aka. The Kīṭāgiri Sutta of the Majjhima Nikāya (no. 70) gives a listing of seven spiritual types, among which the highest two are “the person liberated in both respects” and “the person liberated by wisdom”. These are the two kinds of arahants or fully liberated persons found in the Nikāyas.31 The Kīṭāgiri Sutta explains these two kinds of persons as follows: Here some person dwells having touched personally those peaceful emancipations that are formless and transcending forms; and having seen with wisdom, his taints are destroyed. This is called “the person liberated in both respects”. … Here some person does not dwell having touched personally those peaceful emancipations that are formless and transcending forms; and having seen with wisdom, his taints are destroyed. This is called “the person liberated by wisdom”.32 According to the Kīṭāgiri Sutta, the distinction between these two kinds of arahants is this: a person who is “liberated in both respects” experiences the peaceful emancipations that are formless and transcending forms (santā vimokhā atikkamma rūpe āruppā), but a person “liberated by wisdom” does not have such experience of those formless emancipations. Apparently the “formless emancipations” (vimokhā āruppā) refer to the four formless attainments, which transcend the four jhānas in the form sphere (atikkamma rūpe). By implication, a person “liberated by wisdom”, although devoid of the formless attainments, may have experience of the jhānas. Disagreeing with the Kīṭāgiri Sutta, the Puggalapaññatti defines these two persons thus: Here some person dwells having touched personally the eight emancipations; and having seen with wisdom, his taints are destroyed. This is called “the person liberated in both respects”. … Here some person does not dwell having touched personally the eight emancipations; and having seen with wisdom, his taints are destroyed. This is called “the person liberated by wisdom”.3

This Abhidhamma definition deviates from that of the Kīṭāgiri Sutta, which also belongs to the Therava¯da. In this case, surprisingly, the Puggalapaññatti, a Therava¯da Abhidhamma text, does not invoke the Sutta literature of the Therava¯da, but rather closely parallels the following Su¯tra literature of the Sarva¯stiva¯da now extant in Chinese translation. The Sarva¯stiva¯da34 counterpart of the Kīṭāgiri Sutta, sūtra 195 of the Madhyama Āgama (MA¯ 195), states: Some monk dwells having touched personally and attained the eight emancipations; and having seen with wisdom, his taints are destroyed and understood. This is called “the monk liberated in both respects”. … Some monk does not dwell having touched personally or attained the eight emancipations; and having seen with wisdom, his taints are destroyed and understood. This is called “the monk liberated by wisdom”.35 A virtually identical definition is also found in sūtra 936 of the Saṃyukta Ᾱgama (SA¯ 936),36 which is widely ascribed to the Sarva¯stiva¯da37 or perhaps more specifically the Mu¯lasarva¯stiva¯da tradition.38 Therefore, according to the Sarva¯stiva¯da tradition (MA¯ 195 and SA¯ 936), what distinguishes one type of arahant from the other is that the person “liberated in both respects” has experience of the eight emancipations whereas the person “liberated by wisdom” does not have such experience of the eight emancipations. The last five of the eight emancipations are the four formless attainments and the “cessation of perception and feeling” (saññā- vedayita-nirodha) according to many suttas,39 but the identity of the first three is unclear in the suttas. In the Dhammasaṅgaṇi the first three emancipations are connected with the four jhānas.40 Bodhi (2007, p. 69 n. 43) explains them thus: “The first three emancipations are equivalent to the four jhānas, but they deal with the state of jhāna in terms of its objects rather than in terms of its subjective experience.” In sum, the eight emancipations cover all the nine attainments in concentrative meditation (samādhi). Therefore, an arahant “liberated by wisdom” lacks all attainments in concentrative meditation, including even the jhānas. The above passage quoted from the Puggalapaññatti conforms to this Sarva¯stiva¯da Su¯tra tradition that dissociates “liberated by wisdom” from all the nine meditative attainments, and deviates from its own Therava¯da Sutta tradition that an arahant “liberated by wisdom” lacks only the formless attainments and may possess the jhānas. What should we make of this bizarre fact? The answer may be as follows.
When this Abhidhamma work, the Puggalapaññatti, was not yet completed, the issue of “liberated by wisdom” as against “liberated in both respects” was still under debate. A consensus was probably reached regarding “liberated in both respects”, which was seen to denote the “fully-fledged” arahant with experience of all the nine attainments in concentrative meditation. In contrast with such a “fully-fledged” arahant, the one “liberated by wisdom” was seen as incomplete in concentrative attainments. Disagreement arose about the scale of meditative attainments that was expected of the arahant “liberated by wisdom”, and hence the divergent definitions of these two kinds of arahants circulated among the Therava¯dins and Sarva¯stiva¯dins. As Kuan (2013a, pp. 64–68) demonstrates, the seven spiritual types, including these two kinds of arahants, were interpolated into many sūtras as a result of scholastic debates after the Buddha’s death. The seven spiritual types are found twice in the mātikā of the Puggalapaññatti (pp. 3, 10) and expounded in this Abhidhamma text (pp. 14f., 72). In view of these facts, the diverse definitions of the two types of arahants, often subsumed under the seven spiritual types, could have stemmed from Abhidharma debates or discussions mentioned in the previous section. It is likely that the Puggalapaññatti was composed before the Nikāyas had incorporated the Abhidhamma issue of defining the two kinds of arahants. Therefore, without being constrained by a fixed definition in the Therava¯da Sutta literature, the Puggalapaññatti happened to choose the definition in terms of the eight emancipations, and such a definition was also accepted earlier or later by the Sarva¯stiva¯dins. By contrast, after the Puggalapaññatti had been finalized, the compilers of the Majjhima Nikāya decided to incorporate the definition in terms of the formless emancipations while redacting the Kīṭāgiri Sutta. On the other hand, suttas 4.87 and 4.89 of the Aṅguttara Nikāya adopted the definition in terms of the eight emancipations to define the white-lotus ascetic and red-lotus ascetic, which allude respectively to the arahant “liberated by wisdom” and the arahant “liberated in both respects” as indicated above. This suggests the possibility that the definition of the two types of arahants in the Puggalapaññatti could be modelled on the two kinds of ascetics in theAṅguttara Nikāya considering the close relationship between the Puggalapaññatti and the Aṅguttara Nikāya as mentioned above. These divergent interpretations of persons as found in the Majjhima and Aṅguttara Nikāyas represent the outcome of Abhidhamma debates, just as we find in the independent self-avowedly Abhidhamma text, the Puggalapaññatti.

A series of 10 suttas in the Aṅguttara Nikāya (AN 9.42–51) explain various terms by using two opposite modes or methods, namely “with pariyāya” (pariyāyena) and “without pariyāya” (nippariyāyena). The following is relevant information given at PED p. 433, s.v. pariyāya: in Abhidhamma terminology, specifically: pariya¯yena, the mode of teaching in the Suttanta, ad hominem, discursively, applied method, illustrated discourse, figurative language as opposed to the abstract, general statements of Abhidhamma = nippariya¯yena.

As Gombrich (2009, pp. 6) explains, the word pariyāya literally means “way round” and so “indirect route”, but it refers to a “way of putting things”. To sum up, “with pariyāya” (pariyāyena) is a way of putting the subject matter indirectly in figurative language, and therefore this term connotes “in a provisional sense” as rendered by Bodhi (2012, pp. 1319ff.). Then “without pariyāya” (nippariyāyena) can be understood as a way of putting the subject matter directly and non-figuratively, and thus it connotes “in a non-provisional sense” as rendered by Bodhi (2012, pp. 1319ff.). The first sutta (AN 9.42) in the series of ten suttas makes it clear that each of these suttas consists of dialogues between two disciples of the Buddha—A¯ nanda answers questions put by Uda¯yı¯. As Cox (1995, p. 8) points out, this “catechetical style characterized by an exchange of questions and interpretative answers intended to clarify complex or obscure points of doctrine” is seen as the origin of Abhidharma by many Japanese scholars. As the last member of the ninefold division of the Dharma, vedalla represents the genre of questions and answers.41 Dhammajoti (2005, p. 112) notes that in the Aṅguttara Nikāya (III 107), “vedallakathā occurs juxtaposed with abhidhamma-kathā which according to the consensus of scholarly opinion was an important fore-runner of Abhidharma in the later technically developed sense.” In the sutta AN 9.44,42 Uda¯yı¯ asks: “It is said, friend, ‘liberated by wisdom, liberated by wisdom.’ In what way has the Blessed One spoken of one liberated by wisdom?” A¯ nanda answers (abridged): … a monk … dwells having attained the first jhāna, and he understands it with wisdom. To this extent the Blessed One has spoken of one liberated by wisdom in a provisional sense. … a monk … dwells having attained the second jhāna … the third jhāna … the fourth jhāna … the sphere of infinite space … the sphere of infinite consciousness … the sphere of nothingness … the sphere of neitherperception-nor-non-perception, and he understands it with wisdom. To this extent, too, the Blessed One has spoken of one liberated by wisdom in a provisional sense. … a monk … dwells having attained the cessation of perception and feeling, and having seen with wisdom, his taints are destroyed; and he understands it with wisdom. To this extent, friend, the Blessed One has spoken of one liberated by wisdom in a non-provisional sense

In the next sutta, AN 9.45, Uda¯yı¯ asks about “liberated in both respects”. A¯ nanda answers in exactly the same way as in the previous sutta except that the phrase “he dwells having touched that sphere personally in whatever way [it is attained]”44 is inserted into each of the statements of the nine spheres or meditative attainments. For example, for the first jhāna and the highest attainment, A¯ nanda states: … a monk … dwells having attained the first jhāna … He dwells having touched that sphere personally in whatever way [it is attained], and he understands it with wisdom. To this extent the Blessed One has spoken of one liberated in both respects in a provisional sense. ……… … a monk … dwells having attained the cessation of perception and feeling, and having seen with wisdom, his taints are destroyed. He dwells having touched that sphere personally in whatever way [it is attained], and he understands it with wisdom. To this extent, friend, the Blessed One has spoken of one liberated in both respects in a non-provisional sense.45 Now we can work out the “non-provisional” or “definitive” interpretation of the two kinds of liberation. According to AN 9.44, if a person attains the highest meditative sphere (the cessation of perception and feeling, which presupposes the eight lower meditative attainments), and having seen with wisdom, his taints are destroyed, then such a person is “one liberated by wisdom” in a definitive sense. According to AN 9.45, if a person attains the highest meditative sphere, and having seen with wisdom, his taints are destroyed, and he dwells having touched that sphere personally, then such a person is “one liberated in both respects” in a definitive sense. In contrast to the “definitive” interpretation, “provisional” interpretations of the two types are distinguished by describing persons as being able to attain some ofthe eight lower meditative spheres (i.e. the fourjhānas and four formless attainments) but not the highest meditative attainment. Therefore, according to the definitive, non-provisional definitions in these two suttas, both “a person liberated by wisdom” and “a person liberated in both respects” must experience the highest meditative attainment (“dwells having attained the cessation of perception and feeling” as stated in both suttas), not to mention the eight lower attainments. If the texts are taken literally, there is only one difference between the two: “a person liberated in both respects” dwells having touched that sphere (the cessation of perception and feeling) personally in whatever way [it is attained] whereas “a person liberated by wisdom” does not. This distinction is puzzling, however, because “to dwell having attained the cessation of perception and feeling” can hardly be distinguished from “to dwell having touched personally the cessation of perception and feeling”. Therefore, “liberated by wisdom”is hardly different from“liberatedin both respects”. Unfortunately, the Pali commentary (Mp IV 206–207) offers no relevant comment regardingthisissue. In any case, according to these Aṅguttara suttas, all the nine meditative attainments are indispensible to both “liberated by wisdom” and “liberated in both respects” in a nonprovisional, definitive sense. This interpretation diverges significantly from those in the Majjhima Nikāya, Madhyama Ᾱgama, Saṃyukta Ᾱgama and Puggalapaññatti, where “liberated by wisdom” is distinct from “liberated in both respects” in lacking either the higher meditative attainments or all nine meditative attainments. In this connection, it is worth noting the following remark by Hamilton (2000:8): Thatit was the spirit ratherthanthe letter ofthemthat matteredis further supported by the style in which the central doctrinal teachings have been preserved in the Nikāyas: they are nearly all given cryptically, open to various interpretations but with no one definitive interpretation attached. It is, indeed, this cryptic equivocal style that has allowed there to be confusion and disagreement about what the teachings mean at all … The later Therava¯da Buddhist Abhidhamma scholars … pronounced the style of the early material to be just ‘a way of putting things’ (pariyāya). And though they then interpreted and explained the teachings in definitively ‘put’ terms (nippariyāyena) in their own texts … In view of this observation,suttas 9.44 and 9.45 of the Aṅguttara Nikāya may be seen to verify the fact that “liberated by wisdom” and “liberated in both respects” in the Sutta literature are usually described cryptically, in ways open to various provisional interpretations but with no one definitive interpretation attached. Therefore, we have the various interpretations of these two kinds of liberation in the Nikāyas and Āgamas. Most of these interpretations are merely oblique and figurative “ways of putting things”, and thus they should not be taken literally46 but require further explication for clarity. As Ronkin (2005, p. 26) states: “[T]he Abhidhamma methods of instructing the teaching… does not need any further explication, because it is couched in non-figurative, definitely put terms (nippariyāya-desanā).” The two suttas of the Aṅguttara Nikāya quoted above obviously serve as Abhidhamma, which purports to explain “liberated in both respects” and “liberated by wisdom” in a definitive, non-figurative way (nippariyāyena), and hence is entitled to label other definitions as just figurative or provisional (pariyāyena). A certain Abhidharmist probably compiled these two suttas in an attempt to provide a final solution to the issue under debate. Apart from these two suttas, the other suttas in this series of ten suttas on pariyāyena and nippariyāyena are apparently all meant to serve this purpose. Let us examine a sutta on another type of person. In AN 9.43, while explaining the “witness-in-person” (kāya-sakkhi),47 A¯ nanda says:
… a monk … dwells having attained the first jhāna … He dwells having touched that sphere personally in whatever way [it is attained]. To this extent the Blessed One has spoken of a witness-in-person in a provisional sense. … the second jhāna… … the sphere of neither-perception-nor-non-perception … in a provisional sense. … a monk … dwells having attained the cessation of perception and feeling, and having seen with wisdom, his taints are destroyed. He dwells having touched that sphere personally in whatever way [it is attained]. To this extent, friend, the Blessed One has spoken of a witness-in-person in a non-provisional sense.48 Therefore, in definitive terms, a witness-in-person is a taintless liberated person, i.e. an arahant. This “non-provisional” (nippariyāya) interpretation deviates from the usual sutta interpretation of this person as inferior to the arahant. According to another sutta also in the Aṅguttara Nikāya (AN 3.21 at I 120), a witness-in-person is one who is practising for arahantship, or a once-returner, or a non-returner, and is never an arahant. Similarly, according to the Kīṭāgiri Sutta (MN 70), the Buddha defines the “witness-in-person” thus: Here some person dwells having touched personally those peaceful emancipations that are formless and transcending forms; and having seen with wisdom, some of his taints are destroyed. This is called “the person who is a witness-in-person”. I say that such a monk still has work to do with diligence.49 Accordingly, a witness-in-person has destroyed only some of his taints, and thus is not a taintless arahant. This is in stark contrast to the definitions of “the person liberated in both respects” and “the person liberated by wisdom”, who are both said to have destroyed taints and have no work to do with diligence (MN I 477–478); in other words, these two persons are arahants. From the Abhidhamma viewpoint as expressed in the “sutta” AN 9.43, the interpretations of witness-in-person in AN 3.21 and MN 70 are just provisional (pariyāyena), and hence should not be taken at face value. According to the serial suttas AN 9.43, AN 9.44 and AN 9.45, the three kinds of persons, namely “witness-in-person”, “one liberated by wisdom” and “one liberated in both respects”, are only nominally different; in effect they all refer to the same type of person, i.e. the liberated arahant. The standpoint represented in these three serial suttas seems to be unique, as Buddhist literature usually distinguishes those three kinds of persons into three hierarchically different types.50 The position taken by these three suttas is surprisingly similar to the following idea in one Maha¯ya¯na text, namely the part of the *Mahāvaipulya-mahāsaṃnipāta Sūtra 大方等大集經 (T 397) translated by Dharmaks˙ ema 曇無讖 (385–433 CE).51 This text identifies the arahant “liberated-by-wisdom” (慧解脫) and the arahant “liberated-in-both-respects” (二分解脫) respectively with “witness-in-person” (身證), and thereby equates the three types of persons (T XIII 159a). This Maha¯ya¯na conception of spiritual types is consonant with the above Abhidhamma hermeneutic approach (nippariyāyena) to the three kinds of persons. These types of persons are just conventionally, nominally designated and do not exist in ultimate reality, as Apple (2004, pp. 261–262) observes: The Prajñāpāramitā literature repeatedly states that while coursing in the practice of prajñāpāramitā – i.e., viewing things through cognizing emptiness, bodhisattvas see the various stages from Stream-enterer up to Buddhahood as being like an illusion. A bodhisattva, if obtaining a result such as Stream-enterer, does not think “I have obtained the result of Stream-enterer.” … Therefore, all stages are ultimately seen as like illusions, like a fictitious person. Such a fictitious and non-essentialist perspective of persons could have led to the foregoing interpretations that blur the boundaries between the various spiritual types as found in the Aṅguttara Nikāya’s Abhidhamma portion and also in the Maha¯ya¯na sūtra quoted above. Since persons, including the spiritual types, do not exist in essence, their definitions are after all just conventionally designated in order to convey significant information concerning the ultimate goal of liberation and how to progress to this goal. The non-essentialist perspective of persons coincides with the “absence of essence (ātman, Self) in persons” (pudgala-nairātmya) as conceived by the preMaha¯ya¯na traditions.52 This exegetical approach to the concept of person (pudgala) was already taken up in the Abhidharma and then followed by the Maha¯ya¯na. It is widely held that the Maha¯ya¯na, including the Prajn˜a¯pa¯ramita¯ (Perfection of Wisdom) literature, criticizes the Abhidharma for its inability to understand the “absence of essence (Self) in dharmas” (dharma-nairātmya); instead the Abhidharmists considered dharmas to be things that really, ultimately exist.53 If the Abhidharmists simply thought of “persons” (pudgala/puggala) as conventional constructs in opposition to dharmas, the ultimately realities, how could there be so many “Abhidharma” expositions of persons (as discussed above) apart from those of dharmas, which alone were perceived as real and should be the proper subject matter for the Abhidharma? Moreover, as demonstrated above, the three items “person whose mind is like a diamond”, etc. in the Aṅguttara Nikāya were, as expected, “depersonalized” in the Abhidhamma texts of two different schools, namely the Dhammasaṅgaṇi and the Saṃgītiparyāya. But why did the Puggalapaññatti, also an Abhidhamma text, fail to depersonalize those same items in the Aṅguttara Nikāya?


Can “Persons” (puggala) be Subject Matter Appropriate for the Abhidhamma?
The mātikā of the Puggalapaññatti consists of the following six designations or descriptions (cha paññattiyo): 1. designation of aggregates (khandha-paññatti) 2. designation of bases (āyatana-paññatti) 3. designation of elements (dhātu-paññatti) 4. designation of truths (sacca-paññatti) 5. designation of faculties (indriya-paññatti) 6. designation of persons (puggala-paññatti) According to Kimura (1968, pp. 70, 82), this list represents the southern, or Therava¯da, classification of the Abhidhamma. While the last designation is expounded in the Puggalapaññatti, the first five designations relating to dhammas are treated in the Vibhaṅga. 54 As mentioned above, the Puggalapaññatti and the Vibhaṅga are probably the earliest of the seven canonical Abhidhamma works in Pali. It is likely that these two texts were meant to be complementary to each other, and therefore persons and dhammas may have enjoyed equal status at the incipient stage of the Abhidhamma. However, Nyanatiloka (1957, p. 57) questions whether the Puggalapaññatti qualifies as an Abhidhamma text: This smallest of the seven Abhidhamma books appears to be somewhat out of place in the Abhidhamma Pit˙ aka, as shown even by its title “Description of Individuals”. For it is one of the main characteristics of the Abhidhamma that it does not employ conventional concepts like “individual” (puggala), etc., but deals only with ultimates, or realities in the “highest sense” (paramatthadhamma), i.e. the mental and material phenomena, and their classifications into groups (khandha), bases, elements, etc. This treatise, however, in accordance with its subject-matter, is written in the conventional language as used in the Sutta-Pit˙ aka
In a similar vein, Gethin (1998, p. 209) regards the Abhidharma as an attempt to give a systematic and exhaustive account of the world in terms of its constituent physical and mental events, which are known as dharmas, and “ultimately dharmas are all that there is” while “a person is analysed by Abhidharma as consisting of innumerable dharmas”. Nyanaponika (1998, p. 5) also says that in the Abhidharma such sutta terminology as “persons” is replaced by a more precise terminology, which accords with the “impersonal” nature of actuality. Therefore, a clear distinction can be drawn between puggala (person) and dhammas. While dhammas are “realities” in accord with the “impersonal” nature of actuality, a person (puggala) is just a “conventional concept” or a composite built from innumerable constituent elements or dhammas. The conventional terminology “person” is used only in the suttas; it is not employed in the Abhidhamma, which is devoted to the exposition of dhammas. Consequently, there appears a dichotomy between persons and dhammas in relation to the way such terminology is used in the Sutta Pit˙ aka and the Abhidhamma Pit˙ aka. As mentioned above, Nyanatiloka holds that while the Sutta Pit˙ aka employs “conventional” concepts such as “person” (puggala), the Abhidhamma deals only with “ultimates”, or “realities in the highest sense”, his rendering of paramatthadhamma. In other words, dhammas (realities) in the highest sense should be distinguished from puggala (person), which exists only in the conventional sense. In this context, we should note that, as Gethin (1998, pp. 207f.) points out, some of the Buddha’s teachings are said to be expressed in conventional terms (saṃvṛti/ sammuti) while others are expressed in ultimate terms (paramārtha/paramattha), and that according to the later tradition, the sutta/su¯tra collections contain teachings of both kinds whereas the Abhidharma is “an attempt to give a comprehensive statement of the Buddha’s teachings exclusively in ultimate terms”. Thus the dichotomy between persons and dhammas can be seen as a conventional-ultimate dichotomy. This dichotomy, however, seems to contradict the aforementioned fact that the Puggalapaññatti begins with a mātikā composed of the six designations, including both dhammas and persons. In addition, Mizuno (1997, p. 262) points out that many types of persons explicated in the Puggalapaññatti correspond to those discussed in the Saṃgītiparyāya (T 1536),55 a Sarva¯stiva¯din Abhidharma work, and the *Pudgala Varga 人品 (Chapter on Persons) of the *Śāriputra-abhidharma 舍利弗阿毘曇論 (Shelifu apitan lun, T 1548),56 probably belonging to the Dharmaguptaka school.57 Therefore, apart from dhammas, the subject matter of “persons” is indeed among the concerns of the Abhidharma of various schools. If the Puggalapaññatti, along with those pudgala-related parts of the Saṃgītiparyāya and of the *Śāriputra-abhidharma, is regarded as an addition to the “Abhidharma proper”, we can certainly assume that the Abhidharma is concerned solely with dhammas as opposed to persons. But it is not beyond doubt that such person-related portions of the Abhidharma literature can be set aside in this way. Moreover, whether the conception of persons and that of dhammas are categorically different is a crucial question that requires further clarification. The Dhammasaṅgaṇi, a canonical Abhidhamma text of the Therava¯da, does not appear to uphold that dhammas are ultimate realities as against conventional constructs like persons. The title of this text means “compendium of dhammas”.58 The text affirms that “all dhammas are ways of designation”, that “all dhammas are ways of interpretation” and that “all dhammas are ways of expression”.59 While defining paññatti (designation), nirutti (interpretation) and adhivacana (expression), the Dhammasaṅgaṇi in each case enumerates a long list of words as follows: Yā tesaṃ tesaṃ dhammānaṃ saṅkhā (enumeration) samaññā (appellation) paññatti (designation) vohāro (parlance) nāmaṃ (name) nāmakammaṃ (name giving) nāmadheyyaṃ (name assigning) nirutti (interpretation) vyañjanaṃ (wording) abhilāpo (talk)—ime dhammā adhivacanā/nirutti/paññatti. 60 In light of this passage, when the Dhammasaṅgaṇi describes all dhammas as “ways of designation, ways of interpretation and ways of expression”, the text cannot but mean that all dhammas are just conventional (sammuti) constructs rather than ultimate (paramattha) realities.61 The commentary on the Dhammasaṅgaṇi explains paññatti in this context (ways of designation) thus: “Making [one and the same idea]62 known in various ways, such as ‘takka, vitakka, saṅkappa’, is called paññatti (designation)”,63 thereby suggesting that the word paññatti in this context denotes conventional usage of language or concepts. Incidentally, among the above list of words, paññatti, samaññā, vohāra and nirutti already appear together in the Poṭṭhapāda Sutta (DN 9),64 where the Buddha uses these words to indicate that the things he just mentions are merely designations in common use in the world (lokapaññatti, etc.). Kalupahana (1986, p. 340) suggests that by this sutta passage the Buddha intended to take saṃvṛti (P. sammuti, convention) and prajñapti (P. paññatti, designation) as synonyms. Although the above passage in the Dhammasaṅgaṇi perhaps suffices to illustrate the point, it is worthwhile to compare this passage with the following passage in the Śikṣāsamuccaya by S´a¯ntideva (early eighth century CE):

anupalabhyamāneṣu sarvadharmeṣu katamo ’tra buddhaḥ |… śūnyaṃ hi rūpaṃ rūpeṇa yāvad vijñānaṃ || pe || yāvad eva vyavahāramātram etat | nāmadheyamātraṃ saṃketamātraṃ saṃvṛtimātraṃ prajñaptimātraṃ | 66 (Since all dharmas are not to be obtained,67 what is Buddha? … For form is empty of form68 as far as consciousness [is empty of consciousness], and so on [the same is said of the five aggregates]. All this is just parlance, just name assigning, just agreement, just convention, just designation.) Among this list of five Sanskrit words each combined with mātra (just), three words also have Pali counterparts in the above long list of words in the Dhammasaṅgaṇi: vyavahāra = vohāro, nāmadheya = nāmadheyyaṃ, prajñapti = paññatti. These five words, each combined with mātra, are apparently enumerated here as synonyms. Consequently, prajñapti (paññatti, designation) is synonymous with saṃvṛti (sammuti, convention) as well as with the other three words in the fivenumbered list of synonyms in the Śikṣāsamuccaya. 69 It would not be farfetched to speculate that this Maha¯ya¯na text and the Dhammasaṅgaṇi draw on certain Abhidharma material in common, which associates “all dharmas” with “parlance” (vyavahāra/vohāra), “designation” (prajñapti/paññatti), etc. to indicate that dharmas are conceptual constructs in conventional terms (saṃvṛti/sammuti) rather than realities in the ultimate sense (paramārtha/paramattha). Kalupahana (1992, p. 145) articulates a similar view on the Abhidhamma: If the intention of the discourses [i.e. suttas] in analyzing the human personality into five aggregates was merely to indicate the absence of a metaphysical agent (anatta) and not to discover a set of irreducible elements called “ultimate realities,” there seems to be no justification for the various psychological and physical items [i.e. dhammas] listed in the canonical Abhidhamma texts (both in Pali and in Sanskrit) to be considered ultimate realities. He goes on to argue that the Puggalapaññatti, while explaining the conceptions relating to a “person” (puggala), adopts “the contextual analysis of the conceptions of aggregates and so forth”, i.e. dhammas, “in the previous books of the Abhidhamma” (ibid., p. 150). But the matter could be the other way round since the Puggalapaññatti presumably belongs to the earliest stratum of the Pali

Abhidhamma as mentioned above. In any case, Kalupahana is probably right in suggesting that the Abhidhamma treats both puggalas and dhammas as contextdependent, thereby dissolving the dichotomy between persons and dhammas which is linked to the distinction between convention and ultimate realities as elaborated in the commentarial tradition.70 As Ronkin (2010, p. 356) suggests, in the Pali tradition it is the commentaries, including the commentary to the Dhammasaṅgaṇi, that first draw metaphysical conclusions with regard to the reality of dhammas; the commentaries contend that “there is no being or person apart from dhammas” in order to “refute the rival Puggalava¯da position of the reality of the person” (Dhs-a 155, etc.), and that “a dhamma arises as a present, ultimate reality, and its sabhāva attests to its actual existence as such” (Dhs-a 45, etc.).71 As mentioned above, the Maha¯ya¯na criticizes the Abhidharma for its failure to understand the “absence of essence in dharmas”, which were considered by the Abhidharma to be things that really, ultimately exist. This critique is intended primarily for the Sarva¯stiva¯din Abhidharma rather than the Abhidharma in general.72 The Sarva¯stiva¯dins maintain that all (sarva) dharmas exist (asti) as real entities (dravyatas), whether past, present, or future, and are defined by a fixed, intrinsic nature (svabhāva),73 but this ontological position is not accepted by some other schools, including the Therava¯da during the canonical period. Cousins (1983– 1984, pp. 106f.) notes: The Therava¯da does not reify dhammas to anything like the extent found in the Sarva¯stiva¯din abhidharma. … In North India where the Sarva¯stiva¯din abhidharma eventually established a commanding position, the term dharma came to be interpreted as a ‘reality’ and given some kind of ontological status as part of a process of reification of Buddhist terms. … In the South, at least among the Therava¯dins, dhamma retains its older meaning of a less reified, more experiential kind. Ronkin (2005, p. 226) also says: “The Therava¯dins … do not subscribe to the Sarva¯stiva¯da metaphysics: first and foremost they do not hold that a dhamma is a dravyasat [referring back to ‘primary existent’] and do not use the category of sabhāva as an ontological determinant of primary existence—at least not until late into the post-canonical period.”74 In view of the foregoing discussion, the notion of persons and that of dhammas cannot be categorically differentiated from the perspective of the Abhidharma in general. Just like persons, dhammas exist as experiential events rather than ontological entities according to at least the Therava¯din Abhidhamma system of thought as discussed above, although the post-canonical tradition interprets dhammas as ultimate realities and this is followed by many scholars. In this sense there is no big difference between the two categories. Nor is there a wide gap between the Sutta and Abhidhamma literature in terms of how “persons” are treated, despite what the later tradition holds. This has been shown to be corroborated by several facts about the Puggalapaññatti and the Aṅguttara Nikāya. The listings and interpretations of “persons” form a significant portion of the Aṅguttara Nikāya and of the Abhidhamma literature. We have seen above that several suttas on persons (puggala) in the Aṅguttara Nikāya appear to be proto-Abhidhamma. The Aṅguttara Nikāya contains a sizable amount of such Abhidhamma material in an eclectic style, embracing a wide variety of divergent Abhidhamma interpretations. By contrast, the self-avowedly Abhidharma literature of the different schools represents the collections of scholastic thoughts which were standardized depending on the schools or the texts themselves. It is undeniable that the ultimate concern of Buddhism is the welfare of people. The Buddha’s teaching is meant to help or liberate each person as a whole rather than each of the individual dhammas. Even an adequate exposition of Buddhist ethics also entails at least a certain treatment of the person as a whole. The Vijñānakāya, a canonical Abhidharma text of the Sarva¯stiva¯dins, presents a debate between the “Personalist” (pudgalavādin) and the “Voidist” (śūnyatāvādin, follower of the emptiness teaching). The debate is aptly outlined by Cousins (1994, p. 21) thus: The Personalist asks what is the object of loving-kindness. The Voidist replies that it is the five aggregates given the label of ‘being’. The Personalist, not unreasonably, suggests that this is not in harmony with the suttas which recommend loving-kindness towards living beings rather than aggregates. … Therefore, although the Abhidharma adopted the methods that attempt to systematize and preserve the Buddha’s teaching (Dharma) by collecting all possible doctrinal concepts or elements, i.e. dharmas, to be found in his sermons,75 it could not evade the topic of “persons” that is also essential to the Buddha’s teaching. Therefore, in spite of the “depersonalizing” tendencies perceptible in some Abhidharma works, three different schools unanimously devote part of their Abhidharma to the subject of “persons” as discussed above. In his Philosophy and Psychology in the Abhidharma, Guenther (1974, p. 213) comments: “The Path in its four stages is essentially meant to overcome unhealthy attitudes and to produce a certain type of man in whom unhealthy attitudes can no longer operate.” It is therefore appropriate for the Abhidharma to designate and interpret the various spiritual types of persons for the sake of illuminating Buddhism’s aim to transform a person into an arahant and its skilful means in designating spiritual stages which a person can go through to achieve this aim. This is not tantamount to reifying “persons”, which are still treated as conventional constructs no less than dhammas are

According to a widespread opinion based on the tradition, “person” (puggala/ pudgala) is a “conventional concept” used only in the suttas/sūtras; it is not employed in the Abhidhamma/Abhidharma, which interprets the Buddha’s teaching exclusively in terms of “ultimate realities”, i.e. dhammas/dharmas. This results in a dichotomy between persons and dhammas, which can be seen as a conventionalultimate dichotomy. The Abhidharma tendency to “depersonalize” is discernible in how two Abhidharma texts interpret a list of persons in a sutta of the Aṅguttara Nikāya. Although the Abhidharma adopted the methods that attempt to systematize the Buddha’s teaching (Dharma) by collecting all possible doctrinal elements, i.e. dharmas, to be found in his sermons, it could not evade the topic of “persons” that is also essential to the Buddha’s teaching. There are indications that “persons”, just like dhammas, can be subject matter appropriate for the Abhidhamma. Three different schools devote part of their Abhidharma to the subject of “persons”, notably the Therava¯dins’ Puggalapaññatti, which is closely connected with the Aṅguttara Nikāya. The mātikā of the Puggalapaññatti consists of six designations or descriptions (paññatti). The last one is the “designation of persons”, which is expounded in the Puggalapaññatti. The first five designations, which relate to dhammas, are treated in the Vibhaṅga. The Puggalapaññatti and the Vibhaṅga are probably the earliest of the seven canonical Therava¯din Abhidhamma works, and were apparently meant to be complementary to each other. Therefore persons and dhammas may have enjoyed equal status at the incipient stage of the Abhidhamma. That dhammas are reckoned as conventional constructs rather than ultimate realities is implicit in the Dhammasaṅgaṇi, a canonical Abhidhamma text of the Therava¯da. Dhammas are reified by the Sarva¯stiva¯din Abhidharma and the Therava¯din commentaries, but not by the Abhidharma in general. Just like persons, dhammas exist as experiential events rather than ontological entities, at least according to the Therava¯din Abhidhamma. Thus, there is no big difference between the two categories. Nor is there a wide gap between the Sutta and Abhidhamma literature in terms of how “persons” are dealt with, despite what the later tradition holds. This is corroborated by several facts about the Puggalapaññatti and the Aṅguttara Nikāya. Several suttas on persons (puggala) in the Aṅguttara Nikāya appear to be proto-Abhidhamma. The Aṅguttara Nikāya contains a significant amount of such Abhidhamma material in an eclectic style, embracing a wide variety of Abhidhamma interpretations. The divergent definitions of four kinds of ascetics in a series of suttas of the Aṅguttara Nikāya and the Puggalapaññatti represent the outcome of Abhidhamma debates on how to define the items and arrange them in numerical sequence. A series of ten suttas in the Aṅguttara Nikāya bluntly adopt the Abhidhamma hermeneutic approach, i.e. nippariyāyena, whereby three types of persons are rendered just nominally different. Therefore, various kinds of “persons” are seen as conventionally designated rather than really existing in accordance with the non-essentialist perspective of persons in the Abhidhamma. It is no less appropriate for the Abhidhamma to designate and interpret various types of persons than to treat dhammas also as conventional constructs. The Abhidhamma has to accommodate the “conventional” fact that the Buddha’s teaching is meant to help or liberate each person as a whole rather than each of the individual dhammas