The attakari sutta is a bit off an odd duck and the word "attakara" (translated as self doer) as far as I know only occurs there without a direct explanation. The full meaning of a word does not always have to be made up by its components, so it isn't always clear from the pali. You have to look in the context. Quite clearly "atta" here does not mean a self as we usually see it, because that is denied so many times.
So I think it goes very far to conclude from that sutta alone that the Buddha thought there is free will. And we have to look at this sutta in context of all the others, instead of the other way around. There are so many suttas that say volitions are not self, and also "one can't have it of *any aggregate*, 'let it be thus'".
So to me the sutta does not imply a free will. I see the answer lying in sentences such as this:
"“So, brahmin, when there is the element of initiating, initiating beings are clearly discerned; of (among) such beings, this is the self-doer, this, the other-doer. "
As also seen in the notes the word of
is to be interpreted as among
. So the sutta is just saying there are beings among beings. There are individual beings, and of each being its actions are part of that individual being. One beings actions are not part of another being. So that beings have a self or whatever else which has free choice - that isn't said in this sutta.
This could be seen as opposed to people who may think that there is a God who controls it all - or that we are part of his dream or whatever. That would explain the view of the brahmin.
I think that's a fair point, and I came to a somewhat similar conclusion myself after a similar discussion about the apparent contradiction between AN 6.38
and SN 22.59
—the former seeming to affirm the existence of a self-doer (atta-kara
), and the latter stating that the five aggregates are not-self (anatta
), which raised the question, How can there be a self-doer if all of the aggregates are not-self? My answer was as follows:
To begin with, I personally tend to translate attakara as 'individual effort,' which I think helps avoid some of these linguistic dilemmas. Moreover, I understand this self-doing or self agency or individual effort that's mentioned in AN 6.38 to be analogous to kamma
, the intentional actions done via body, speech and mind (AN 6.63
). Kamma is intention (cetana
), and intention is a product of the aggregate of mental formations (sankhara-khandha
). Therefore, being a product or process within one of the aggregates, which themselves are types of processes and not-self, this type of internal decision maker or will-to-do, if you will, has its own requisite conditions and is also not-self, since whatever is conditioned and subject to change can't be said to have an unchanging being or essence.
In other words, SN 22.59 is refuting the idea of a permanent and unchanging self or soul (i.e., an agent free from conditionality, and consequently, aging, illness and death) within and controlling the aggregates, whereas AN 6.38 is refuting the idea that there's no volitional choice available to us whatsoever. They're two extremes the Buddha rejects. As Nyana once so eruditely put it
Desire and attention and volitional choice can all be concomitant causes or effects. Freedom of choice isn't independent of other causes and conditions -- it operates within the same conditioned mind-stream. But it does operate, and it does so in consort with desire and attention, and so on. Hence there is no need for Cartesian notions of free will or Upaniṣadic notions of a permanent, unchanging Self for there to be freedom to choose. In fact, these non-Buddhist systems are not sustainable precisely because of the interdependence of phenomena: i.e. an unchanging agent cannot engage in actions, etc.
As for where the line between volitional and involuntary actions begins, I'd say it's consciousness (vinnana
); although, given the complexities of this/that conditionality
and the inherent interconnectedness of all these internal and external processes, I don't think there's anyway to pinpoint precisely what or where this element of intention is. For one, it's not a thing as much as a process, and this focus on processes over substances is one reason I view Buddhism as a type of process philosophy
Basically, I think this apparent contradiction can be easily resolved by taking the context of the two suttas into consideration, as well as surmising that the Buddha essentially took the position that we, as sentient beings, have functional choice via intention operating within a broader framework of causality that conditions the choices available to us at any given time—thus avoiding the extremes of an unchanging agent (eternalism) and no agency at all (nihilism). Others in the Buddhist community see it this way, as well, e.g., Karin Meyers' paper "Freedom and self-control: Free will in South Asian Buddhism
Both Karunaratna and Nanayakkara claim that Buddhism accepts the freedom to choose between alternative courses of action as an "empirical truth," by which they mean a truth that is observed and accepted in the context of moral action. This claim is consistent with the fact that the Buddha offers copious advice as to what actions persons should take or avoid so as to reduce suffering for themselves and others and also arguably consistent with Buddhist accounts of moral responsibility, but Nanayakkara makes the further claim that freedom of choice is essential to the Buddhist concept of karma (Pali: kamma). He explains that cetana, which he describes as “the (free) will," is "equated with moral action (kamma)" in order to account for the fact that one is only responsible for conscious, deliberate actions. A number of other interpreters have made similar claims. For example, K.N. Jayatilleke explains: By 'free will' in a Buddhist context, it is not meant that there is a will, choice or decision which is unaffected by causal factors that affect it, but that our volitional factor or will, choices or decision, while being conditioned by such factors, are not wholly shaped or strictly determined by them, since there is in man 'an element of initiative ' (arabbha-dhatu) or 'personal action' (purisa-kara) or 'individual action' (atta-kara), which within limits can resist the factors that affect it. If not for this factor of human personality, 'moral responsibility' would be a farce and the forces that impel us to act would be responsible for our actions.
But this raises some difficult questions about the nature of volition and the choices we make, and I'm reminded of the causal determinism found in the Abhidhamma. True free will would seem to require an independent agent, and Buddhism effectively denies such an agency. And if our volition is itself conditioned by other factors, then it, too, must ultimately be the result of causally determined processes, meaning that it's not so much an agent as it has the appearance
of being an agent. For example, Dhammanando Bhikkhu once gave me the example of a mosquito biting you on the nose: first you feel annoyed and want to squash it, but then you recall that you're a precept-observing Buddhist and so restrain yourself.
He explained that when this event is described in conventional terms, or according to the Sutta method, it might be said that you had a choice to kill the mosquito or to refrain, and that you chose the latter. But when it's described according to the Abhidhamma method, your abstention from killing wasn't due to choice but to the arising of kusala cetasikas (wholesome mental factors) such as moral shame and fear of wrong-doing (hiri
), and abstinence (virati
), i.e., it was causally determined.
And there are supporting passages like this from the Dhammasangani
What on that occasion is volition (cetana)? The volition, purpose, purposefulness, which is born of contact with the appropriate element of representative intellection - that is the volition that there then is.
And the Atthasalini
Volition is that which co-ordinates, that is, it binds closely (abhisandahati) to itself associated states as objects. This is its characteristic; its function is conation. There is no such thing as volition in the four planes of existence without the characteristic of co-ordinating; all volition has it. But the function of conation is only in moral and immoral states; as regards activity in moral and immoral acts, the remaining associated states play only a restricted part. But volition is exceedingly energetic. It makes double effort, double exertion. Hence the Ancients said: 'Volition is like the nature of a landowner, a cultivator who, taking fifty-five strong men, went down to the fields to reap. He was exceedingly energetic and exceedingly strenuous; he doubled his strength, he doubled his effort, and said, "Take your sickles," and so forth, pointed out the portion to be reaped, offered them drink, food, scent, flowers, etc., and took an equal share of the work.' Volition is like the cultivator; the fifty-five moral states which arise as factors of consciousness are like the fifty-five strong men; like the time of doubling strength, doubling effort by the cultivator is the doubled strength, double effort of volition as regards activity in moral and immoral acts. Thus should conation as its function be understood.
It has directing as manifestation. It arises directing associated states, like the chief disciple, the chief carpenter, etc., who fulfil their own and others' duties... even so, when volition starts work on its object, it sets associated states to do each its own work. For when it puts forth energy, they also put forth energy... It is also evident that it arises by causing associated states to be energetic in such things as recollecting an urgent work and so forth.
Moreover, I think this Abhidhammic position accords well with what Sam Harris writes about the illusion of free will here
(for some reason, all the original links on Harris' site just redirect here
). And even in the Suttas, there are teachings that evidence elements of casual determinism, lending support to the Abhidhammic position. I find AN 11.2
interesting, for example, in that it seems to show how certain wholesome mental factors (kusala cetasikas
) condition certain wholesome qualities and experiences.
However, I wonder whether an agent or self is absolutely necessary for decision making in the first place. A computer can essentially make choices and 'act' based upon imputed information and various algorithms, for example, without a soul or self in the Western philosophical conception. Having a component of our physiology/psychology that does more or less the same isn't all that far fetched. We have the mental capacity to observe, remember, make distinctions, etc., and the part of us responsible for our actions doesn't have to be an independently existing, eternal entity for decisions to be made and their effects/consequences to be analyzed within a moral framework. And if the answer is indeed no, then the entire debate seems to be moot, at least from a practical point of view.