Determinism vs free will

Did the Buddha’s teachings imply a deterministic universe, where everything is “preordained” (and therefore the future can be foretold) or did it support the concept of free will, or the notion that individuals can shape their destinies?

The answer is: It’s complicated.

From the perspectives of modern science

At the quantum level, it could be argued we are living in a deterministic universe. Quantum states represent probabilistic outcomes, but imply that these probabilistic outcomes all exist more or less simultaneously based on their probability levels. This corresponds to the Sarvāstivādins’ view that “everything exists, past present and future.” Note that the Sarvāstivādins arguably was the most influential Buddhist sect, at one stage dominating Buddhist thinking throughout much of Northern India, spreading across to East Asia and directly influencing Mahāyāna thinking, and Mahāyāna is the dominant form of Buddhism surviving today.

However, at the level of atomic particles, it would seem particles are subject to nondeterministic consequences - ie. the trajectory of an atomic particle seems to follow random paths influenced by entropy. But they are also influenced by existing state and previous states. Can future state be completely predicted?

Could a hypothetical omniscient observer, who is able to observe the quantum states of all matter, be able to infer future state? In other words, does observing current and past state allow the future state to be completely inferred?

The “one electron model of the universe”, first proposed by John Wheeler and popularised by Richard Feynman¹, seem to imply a deterministic universe - everything in the universe is simply the manifestation of a single electron traversing backwards and forwards in time, interacting and combining with instances of itself to form all the other particles that make up matter. And hence the future has already happened, we are simply traversing through time along fixed paths.

And yet this flies against conventional wisdom, and much of “ordinary” science.

Our brains are highly complex collections of neurons and ganglia. The operation of a neuron seems to be stochastically driven - whether it fires or not depends on the level of various chemical substances, and activity of nearby neurons. However, it is also argued that our minds are quantum, they rely on the existence of quantum states and therefore somewhat deterministic. What we regard as consciousness can be likened to a spark of flame traversing through multiple parallel universes governed by quantum states.

Therefore, in a sense, our consciousness may be the ephemeral “entity” that creates, or instantiates, the notion of free will. The universe itself may be deterministic, but our experience of it isn’t.

This is the basis of Mahāyāna’s view that everything is “empty” - ultimately our consciousness is wandering through a constructed universe that does not fundamentally have inherent substance or reality. Our notion of “reality” is completely based on our phenomenal experience of the universe as represented by the “five collections” (khandha) - and these collections are also devoid of inherent substance or reality.

The “real” nature of the universe is something we cannot comprehend. And therefore awakening comes from “perfect understanding” (prajñāpāramitā) - which is based on us understanding and realising “profound emptiness” (śūnyatā) - that there is simply no basis for the subjective and phenomenological collection that we identify as a “self”. Once we realise this, we attain nirvana (extinguishment).

Or, to paraphrase a popular movie,

“All those moments will be lost in time, like tears in rain. Time to die.”

Roy Batty, Blade Runner

Dependent origination

Paṭiccasamuppāda (dependent origination) was the Buddha’s own thought process leading to awakening, and the framework for his phenomenological interpretation of the causes leading to dukkha (suffering). This seems to be arguing for both deterministic and volitional views of our ability to influence the future.

On one hand, the causal links between the links in the dependent origination chain of causes strongly suggests a deterministic view of the future, that future consequences are determined by our past actions and behaviour. This based on a phenomenological and causative view of consciousness - that our consciousness is an aggregation of past subjective experiences and volitional choices, which then determines our future experience of the world.

However, implicit in the above is the notion that we do have volitional choices, we can make decisions today that impact the future. In particular, a decision to cease production of non optimal mental constructions will lead to an avoidance of dukkha in the duture. This strongly suggests the Buddha believes in the notion of free will. Or does it?

The five collections

In Buddha’s second discourse 13S3/1.2.1.7 Anattalakkhaṇasutta, the Buddha seems to be advocating a lack of free will. He points out that our sense of personal identity or self does not seem to have control over our “forms” (including our bodies and the external world), our aperceptions, feelings, mental constructions, consciousness etc.

This seems to imply we have no free will, or does it?

It could be argued the Buddha is pointing out our belief in free will is itself flawed, because we equate free will with the false notion of a self. In reality, we are not able to control the outcomes of the future, any more than we can control our subjective experience of the world through the constructed nature of our consciousness.

But again, this would contradict the notion that we can make a volitional decision, in particular to cease construction of non optimal mental constructions.

So the Buddha seems to be implying that even though our notion of self is “constructed”, we do have the agency to “unconstruct”. Awakening is possible, by us consciously making a decision to cease production of non-optimal mental constructions. This would also agree with the common view that as individuals we do possess the power of free will.

Many Buddhists believe the Buddha advocated the concept of “no self” but as some scholars² have pointed out anatta is more correctly interpreted as “not self”.³ The Buddha has not claimed that we have “no self” and therefore “no free will.” The Buddha skirts around answering the question whether we do in fact have some notion of “self” that is beyond the five collections (khnadha), and some have argued that there needs to be a concept of an “individual” possessing “free will” in order for the Buddha’s teachings to be realised. This is a contentious view that has been argued against by many Buddhists who support the “no self” interpretation, but they have avoided answering the question: If there is no self beyond the five collections, then what is the “agent” responsible for ceasing mental constructions, since the Buddha has stated the five collections have no control over production?

The analogy of a Large Language Model

A Large Language Model (LLM) seemingly generates coherent responses and exhibit signs of reasoning ability. In other words, an LLM seem to exhibit traits of consciousness.

But in reality LLMs are neural networks that simply “predict the next token” based on previous tokens fed into it as input, and relying on a large sequence of tokens that have been used to “train” the LLM (creating the model through modifying it’s parameters).

It could be argued that LLMs operate both deterministically and stochastically, just like our brains. It is deterministic in that the next token is completely influenced by the parameters of the LLM and past tokens. And yet there is an element of randomness - the tokens generated by an LLM are influenced by hyper-parameters such as temperature, top-k and top-p.

If we regard an LLM as an “illusion” of consciousness, in the same way that our notion of self is also incorrect, then arguably we are similar to LLMs. However, the Buddha seems to be implying we have the ability to modify our parameters and actively curate our training, in order to create different responses and future outcomes. This implies we have a volitional decision making capability that can be modified.

Conclusion

The Buddha is surprisingly quite aligned with modern science. Because our experience of the world is perceptual, we have no “control” over the external world or “reality.” Therefore the Buddha seems to be implying whether “the real world” is deterministic or not is irrelevant, we only experience it subjectively and perceptually.

What is important is that we can use the “illusion of the self” as an agent of free will or self control to alter our perception of the world. Because we believe in the notion of a “self”, we can actually use that belief to “alter” future perceptions. It is the very impermanence of our notion of the self that allows us the ability to change future outcomes.

If we think of a stylised view of the “self” or consciousness as that spark of flame that is wandering through a perhaps deterministic universe, then we have some agency or “free will” to direct where that “self” wanders to. And ultimately, this is the essence of the Buddha’s teaching.

For further discussion, [@Adam2010] argues perhaps it’s not a question of “free will” but the degree of freedom in achieving awakening (due to the “burden” or kamma from past actions), and perhaps true freedom is not realised until after awakening. This is an interesting approach, but one that I ultimately reject since there nothing in the Buddha’s discourses in the Khandhaka that suggests that there are impediments (from past actions, current circumstances etc.) that inhibit the ability to achieve realisation and awakening, and the concept of “kamma” representing such impediments is not a concept articulated by the Buddha in these discourses (notwithstanding that the concept of “kamma” is extensively explored in the suttas and widely believed by Buddhists).

I will take the same approach as the Buddha by deliberately avoiding an answer to whether there is a “self” or “no self” and therefore the question of “will” is also inconclusive. Clearly, however, “the individual” (no matter how we define such a notion) is capable of ceasing production of non-optimal mental constructions, according to the Buddha, so let’s leave it at that.

Feynman, Richard (1949). “The Theory of Positrons” (PDF). Physical Review. 76 (6): 749–759. Bibcode:1949PhRv…76..749F. doi:10.1103/PhysRev.76.749. S2CID 120117564. ↩
Wynne, A. (2010). The ātman and its negation - A conceptual and chronological analysis of early Buddhist thought. Journal of the International Association of Buddhist Studies, 33(1–2), 103–171. ↩
Norman, K. R. (1997). A Philological Approach To Buddhism: The Bukkyō Dendō Kyōkai Lectures 1994: Vol. V.: Translations such as “without self” and “having no soul” cannot be correct, because the grammar and syntax show that anattā is not a possessive adjective, which it would need to be to have such a meaning. It is a descriptive compound, and if the correct translation for attā is “soul”, then the word would mean that these various things are “not soul”. This, however, cannot be correct, because the Buddha sometimes exhorted his followers to regard these things as parato, i.e. “as other”. We cannot, however, co-relate “as other” and “not soul”. It is clear that the only translation which it is possible to co-relate to “as other” is “not self”. ↩