The Omission of Memory from the List of Dhammas
When we were discussing the faculties in the List of Dhammas we noted in passing that sati occurs only in “good consciousness” (sobhanacitta). This implies that sati means here first of all sammā-sati, right mindfulness, referring to the four “foundations of mindfulness” (satipaṭṭhāna). The original meaning of sati (Skt smṛṛti) as “memory” is, however, not quite excluded, since it has its place in the definition given in the Dhammasaṅgaṇī, but it stands rather in the background and refers always to “good consciousness.” The question now suggests itself: Why has such an important and frequent mental function as that of memory not been expressly included in the List of Dhammas in its quality as an ethically neutral factor? We cannot suppose that it has simply been forgotten. Against any such explanation stands the fact that this list is too obviously the product of a mind working with the greatest accuracy. The list is undoubtedly the result of careful investigation supported by introspective intuition. Certainly no essential aspects of the subject matter have been overlooked here—though, of course, the list does admit of condensation as well as extensions.
This question of memory as an ethically neutral function was actually raised in the Atthasālinī. Here is the passage in full (p. 249):
In this (unwholesome) consciousness faith, mindfulness, wisdom, and the six qualitative pairs have not been included. And why? There is no faith in an unbelieving mind, therefore that has not been included…. And there is no mindfulness in a mind unguarded by mindfulness, therefore that has not been included. How then, do not adherents of wrong views remember their own deeds? They do. But that is not sati (“mindfulness”). It is merely an unwholesome thought process occurring in that aspect (ten’ākārena akusalacittappavatti). That is why sati is not included (in unwholesome consciousness). But why, then, is wrong mindfulness (micchā-sati) mentioned in the suttas? For the following reasons: because unwholesome aggregates (khandha) are devoid of mindfulness; because it is the opposite of mindfulness; and in order to complete the group of factors of the wrong path (micchā-magga). For these reasons wrong mindfulness is mentioned in an exposition of relative validity (pariyāyena). But in an exposition of absolute validity (nippariyāyena) it has no place.
We cannot say that these explanations are very satisfactory. They still leave unanswered the question why memory has not been included in the list under some other name, such as paṭissati, to distinguish it from sammā-sati.
In the subcommentary to the passage just quoted from the Atthasālinī, we find, however, a hint for a plausible theory about the omission of memory (DhsMṭ 120): “[According to that passage just quoted] wrong mindfulness is explained as the unwholesome aggregates that are devoid of mindfulness and contrary to it. This again should be understood as follows: When reflecting on what was done long ago, for example, in the case of inimical feelings, those unwholesome aggregates are associated with keen perception (paṭusaññā-sampayutta).”
Taking up this suggestion we can assume that ancient Buddhist psychology ascribed the main share in the process of recollection to perception (saññā), regarding it merely as a department of the latter. It should be recalled that saññā belongs to the pentad of sense-contact and to the factors common to all consciousness (sabbacittasādhāraṇa), so that the requirement of universal occurrence as a neutral and general factor is fulfilled. We are supported in our theory by the definition of saññā found in the Atthasālinī (p. 110). There two sets of explanations are supplied, given in the customary categories used for definitions (lakkhaṇa, rasa, etc.). According to the first explanation, the characteristic (lakkhaṇa) of perception, applicable to all cases, is “perceiving” (sañjānana, lit. “cognizing well”); the essential property or function (rasa) is “re-cognizing” (paccabhiññāṇa), said to be applicable only to certain cases, namely, when perception proceeds with the help of a distinctive mark of the object, either fixed to it intentionally (e.g., as by woodcutters to trees) or being a characteristic of the object itself (e.g., a mole in the face of a man). The second explanation is said to apply to all cases of perception. The characteristic is again “perceiving.” The essential property given here is: “making marks as a condition for a repeated perception” (i.e., for recognizing or remembering; punasañjānanapaccayanimittakaraṇa). So we may sum up: perception (saññā) is the taking up,75 the making, and the remembering of the object’s distinctive marks. In this connection it is noteworthy that “mark” or “signal” is also one of the different meanings of the word saññā itself.
Not only the “taking up” but also the “making” and the “remembering” of marks may be relevant to all cases of perception if it is understood as follows: What really happens in a simple act of perception is that some features of the object (sometimes only a single striking one) are selected. The mental note made by that perception is closely associated with those selected features; that is, we attach, as it were, a tag to the object, or make a mark on it as woodcutters do on trees. So far every perception is “a making of marks” (nimittakaraṇa). In order to understand how “remembering” or “recognizing,” too, is implied in every act of perception, we should mention that according to the deeply penetrative analysis of the Abhidhamma the apparently simple act of seeing a rose, for example, is in reality a very complex process composed of different phases, each consisting of numerous smaller combinations of conscious processes (cittavīthi), which again are made up of several single moments of consciousness (cittakkhaṇa) following each other in a definite sequence of diverse functions.76 Among these phases there is one that connects the present perception of a rose with a previous one, and there is another that attaches to the present perception the name “rose,” remembered from previous experience. Not only in relation to similar experiences in a relatively distant past, but also between those infinitesimally brief single phases and successive processes, the connecting function of rudimentary “memory” must be assumed to operate, because each phase and each lesser successive state has to “remember” the previous one—a process called by the later fibhidhammikas “grasping the past” (atīta-gahaṇa). Finally, the individual contributions of all those different perceptual processes have to be remembered and coordinated in order to form the final and complete perception of a rose.
Not only in such microscopic analysis of sense perception but also in every consecutive thought process, for example in reasoning, the phase of “grasping the past” can be observed, as for instance when the parts of an argument are connected, that is, when conclusions are built on premises. If that “grasp” of the past is too weak to be effective, one says that one has “lost the thread.” The way in which one remembers the earlier phases of one’s thought process is likewise through selected marks (nimittakaraṇa) because it is neither possible nor necessary to consider all the minor aspects of a thought. But if the “selection” is too incomplete and overlooks essential features or consequences of the past thought, then a faulty argument built on wrong premises follows.
In these two ways we can understand how “remembering,” that is, connecting with the past, is a function of perception in general. We can now formulate the following definition: saññā is cognition as well as recognition, both being by way of selected marks.
We can summarize our findings as follows:
1. Memory, as we usually understand it, is not mentioned as a separate component of a moment of consciousness because it is not a single mental factor but a complex process.
2. The mental factor that is most important for the arising of memory is perception (saññā = sañjānana), being that kind of elementary cognition (jānana) that proceeds by way of taking up, making, and remembering (i.e., identifying) marks.
3. Apart from what, in common usage, is called “remembering,” the reminiscent function of perception in general operates also: (a) in the imperceptibly brief phases of a complete perceptual process, the sequence of which is based on the connecting function of “grasping the past phases”; (b) in any consecutive train of thoughts where this “grasping of the past” is so habitual, and refers to an event so close to the present, that in normal parlance it is not called “memory,” though it is not essentially different from it.
Another reason for the omission of memory from either the components or the classes of consciousness is this: remembrance means merely the fact that a state of consciousness has objects of the past (atītārammaṇa). But as mentioned already (pp. 34-35), in the Dhammasaṅgaṇī the objective side of the perceptual process is used for the classification of consciousness only in a single instance and refers only to the division into visual objects, etc. The time relation of objects, in particular, does not enter into the classification or analysis of consciousness at all, being irrelevant for that purpose. Still less could the time relation—for example, that of memory—be counted as a separate component of consciousness. In the Dhammasaṅgaṇī the time relation of objects is treated separately in the “triad of things with past objects, etc.” (atītārammaṇa-tika). But the fact that a moment of consciousness has objects of the past does not warrant the inclusion of a separate factor called memory.
As a point of comparison between the Pāli Abhidhamma of the Theravādins and the Abhidhamma of later Buddhist schools, it deserves mentioning that in the lists of dhammas compiled by the Hīnayānist Sarvāstivādins and by the Mahāyānist Vijñānavādins, sati (= smṛti) is given as a neutral factor. It is included there in a group called mahābhūmikā, composed of factors common to all consciousness, corresponding to the category of sabbacittasādhāraṇa in the Theravāda. The fact that smṛti is really intended there as an ethically neutral and not a wholesome factor is also proved by the definition given, in this same connection, in the commentary to the Abhidharmakośa: anubhūtasya asampramo˝a (“the not forgetting of what has been experienced”). This divergence from the list given in the Dhammasaṅgaṇī shows that these old thinkers too had noticed the absence of memory in that list, assuming perhaps that it had been forgotten. But for the reasons given above we think that this omission was not only deliberate but fully justified. In other cases of divergence, too, we have found that, on close examination, the Theravāda’s List of Dhammas is far preferable, being based on a much more mature judgment of psychological facts. But here we are not concerned with any such comparative study of Abhidhamma systems.