Showing posts with label Free Will. Show all posts
Showing posts with label Free Will. Show all posts

11 March 2016

Freewill is Back on the Menu

“There is now no safer occupation than talking bad science to philosophers, except talking bad philosophy to scientists.”

- Mary Midgley.

I don't find freewill a particularly interesting problem, but it does come up from time to time. Because it is essential to Buddhist ethics, I've ended up writing about it a few times despite my reluctance, mainly to try to counter what I see as a pernicious trend to Determinism amongst Buddhists influenced by Advaita Vedanta. My essays on the subject include: Do We Have Freewill?(6 Feb 2015), A Sutta on Freewill Experience and Free Will in Early Buddhism (4 Apr 2014). The problem has become a cause célèbre amongst scientists since the 1980s when Benjamin Libet (1916-2007) first published results of his experiments on the so-called "readiness potential". The correct interpretation of these experiments has always been hotly disputed, though Determinists seldom cite any dissenting voices when they reference this material (so readers might be unaware of the controversy). We have seen a number of physicists in recent years citing Libet in the media and in books for the general public as "proving" that there is no free will. But as Peter Clarke observes:
Despite the fame of the Libet experiment and its frequent acceptance in popular and semi-popular writings, it has been the subject of intense controversy. Indeed, most specialists in the philosophy of free will who have addressed the Libet claim have rejected it. (2013)
Philosophers and physicists seem to fall out in public quite often these days. See, for example, Goldhill (2016) for a typical complaint about facile arguments from prominent physicists on the subject of philosophy. Physicists seem to take a perverse delight in dismissing philosophy out of hand, but often show their deep ignorance in the process. Stephen Hawking infamously declared philosophy dead. Paraphrasing Goldhill, this is a very stupid thing for a very smart person to say. Which just goes to show that smart people do make mistakes and do say stupid things. We can't just abdicate the responsibility for evaluating what people say, even when they are experts. As Richard Feynman said, "science is the belief in the ignorance of experts". And if we are not in a position to judge, then we can always seek out those who are and get their opinion. Unfortunately, physicists are often seen as authorities and thus their views on philosophy are widely taken seriously, even when they are out of their depth and saying stupid things. 

In this essay, sparked by a blog by Deric's MindBlog, I will outline Libet's findings, explore some responses from other scholars, and look at the philosophical implications. We might not be able to put to rest the wailing of "there is no free will" by Determinists, but we can at least give them something to chew on for a while.

Libet's Experiments

Libet was investigating the phenomenon of readiness potential (RP). The RP is a slowish build up of electrical potential in the brain, measured at the scalp over the motor cortex by an electroencephalograph (EEG). It occurs a second or more before people make voluntary movements. As the name suggests, this build up of electric potential was assumed to be the brain "getting ready" to initiate a movement. Libet was interested in the timing of the RP and the decision to move.

Libet's classic experiments (Libet 1985, Libet et al. 1983) asked people to make a simple movement, usually flexing their hand or wrist. The subjects were instructed to move whenever they felt like it (within a 20 sec window). At the same time they observed the position of a spot moving in a circle on an oscilloscope screen and reported the position of the spot when they felt the "urge to move". What he found was that there was a delay of some 200 milliseconds (ms) between becoming consciously aware of an urge to move and the actual movement. However, the readiness potential began to build up 350-500 ms earlier.

Note, this is a very short-range phenomenon. The voltage measured on the scalp is in the order of a few micro-volts (10-6 V). The amplitude drops off sharply. Another few centimetres from the scalp and the electrical activity would be undetectable (so no, this is not a mechanism for telepathy!). Indeed, one of the drawbacks of EEG for measuring brain activity is that it doesn't detect electrical activity below the cortex layer. The technique is also poor at localising the activity - multiple electrodes and sophisticated analysis of the activity can improve this, but EEG is still a pretty blunt instrument. The technique is famous for the early discovery that the activity in the cortex occurs in waves.

Libet controversially interpreted the initiation of the readiness potential as the "decision" to move, the point where the brain unconsciously began preparing to move. Becoming conscious of an "urge to move" came significantly later, and then, finally, the action itself was initiated, the whole process taking almost half a second. In this interpretation, the experience of willing our hand to move comes quite a long time after the brain has decided to move. In other words, the experience of willing our hand to move is a secondary feature in the process. Hence, freewill, interpreted as contra-causal freewill, is not what initiates a voluntary movement.

Contra-causal Freewill

I was alerted to idea of contra-causal freewill by reading Patricia Churchland's book Touching A Nerve: The Self as Brain (2013) in which she argues against it. Contra-causal freewill is the idea that we have an abstract capacity called "will" that is like the executive branch of the abstract capacity of "reason". To be considered contra-causally free, this capacity to make decisions or initiate actions must be free from any influences other than itself. Specifically, emotions, motivations, desires, goals, and knowledge must be excluded as potential influences. If any of these influence our decision making then, in this view, our will is not free. Thus for Libet, if some unconscious part of the brain is making the decision to move and then placing the idea in our conscious minds, then even though our brain is still making a decision, it does not count as free will because it is not based on the abstract reasoning capacity.

Any long time readers of this blog will know that this definition of free will is suspect at best. There are two main problems with it. Firstly, the definition makes an egregious mistake in considering reason to be an abstract capacity. I follow Lakoff and Johnson in taking reason to be a function of an embodied mind. Reasoning specifically uses metaphors grounded in our experience of the world to enable abstraction. It is not that we have an abstract capacity for reason, but that we have an embodied capacity for abstraction. Research by Antonio and Hannah Damasio (amongst others) has shown that emotions are involved in all decision making. As I have explained it, emotions tell us how salient any fact is to our decision making process (see Facts and Feelings, 25 may 2012). There is simply no plausible way for contra-causal freewill to operate. Secondly, the definition involved legacy understandings of how reasoning works. I also follow Mercier and Sperber in seeing reasoning as an argumentative capacity. It is well known that individuals are generally very poor at reasoning tasks. Most of us do not to make rational decisions and when we try to, we almost inevitably fall into cognitive bias or logical fallacy. On individual tests of reasoning, we frequently score worse than random guessing. The long lists of cognitive biases and common logical fallacies that we are prone to bear stark witness to this. Reasoning is not activated until we are assessing someone else's argument or retrospectively justifying our own actions (see An Argumentative Theory of Reason, 10 May 2013). 

As Churchland has said, contra-causal freewill is not particularly interesting. Even if the experiment showed that we do not have contra-causal freewill, this would still not be interesting as the concept is a legacy of a bygone era that has no place in modern discussions about the mind or morality. Clearly, many scientists are poorly informed on developments in philosophy because they think philosophy is worthless. They cannot be relied on, in general, to be guides to the philosophy of freewill. This is an important caveat when considering this question. My suspicion is that the criticism cuts both ways. That as much as philosophers like to comment on science, they are often too poorly informed about it to be trustworthy guides to it. All too many philosophers in this field seem to be ontological dualists who do not believe that studying neurons can tell us how the mind works, for example. Sorting out whether or not any individual commentator on this issue makes sense is really quite difficult. No doubt I am also a poor guide to this issue. However, other scholars have been trying to reproduce Libet's experiments and assessing his interpretation of the results since it was published. And we can turn to them to get some balance.

Re-evaluating Libet.

As I say, the interpretation of Libet's experiments has been the subject of intense controversy since they were first published. It seems, from my outsider point-of-view, as though physicists have lined up to say that they prove that human beings are Deterministic and that there is no freewill. But even casually reading around this subject we see that philosophers have lined up to deny that Libet tells us anything about freewill.

For example, what exactly is a voluntary movement? Peter Clarke (2013), for example, cites the example of a tennis player serving a ball. The decision to serve may be voluntary and the movement of the arm might even be partially under conscious control, but the myriad movements that coordinate the whole body as it moves and balances to support the motion of serving are almost entirely unconscious. The motions that direct the ball to the precise location on the opposite player's court are mostly not under the direct control of the player. Tennis players have developed a kind of reflex that allows them to serve accurately at speeds that does not require conscious thought. Indeed, in many sports, we know that thinking too much about key muscle movements is counter-productive. So, is serving the ball a voluntary act? I get this playing the guitar. I train my fingers to find and pluck the notes I want so that I don't have to think about them and this enables me to sing at the same time. If I was consciously seeking out notes on the fretboard and dredging up lyrics and all the other components of articulation and delivery, I could not play the simplest tune, let alone something as complex as, say, the Beatles' tune Blackbird. The assumption that a decision to act cannot occur without being conscious of it is deeply problematic. In playing a tune like Blackbird, I initiate hundreds of actions with no consciousness of doing so because my attention is usually elsewhere. So this question is far from trivial and it ought to make us pause before considering what it means for an experimental subject to make a "voluntary movement". Even if Libet relies on a single movement, how do we know that this is representative, or that the experiment is able to isolate that movement from everything else that is going on in a conscious subject?

An important criticism of the Libet experiment is that it is very difficult to judge when one experiences the "urge to move". Clarke (2013) did the experiment himself and commented "When I try this, I find it very hard to judge the precise time when I decided to move my finger / wrist." Clarke describes studies on the reliability of the subjective timing of events which have shown it to be very imprecise. Additionally, the experiment involves an attention shift from the movement to the timing that "may have introduced temporal mismatches between the felt experience of will and the perceived position of the clock hand." Attempts to eliminate this mismatch have shown that the RP occurred before the "urge to move" only in about two thirds of subjects. To try to improve accuracy, the experiment was performed using functional magnetic resonance imaging (fMRI) and this also showed that "the activity of prefrontal and parietal cortex was correlated with the decision [to move] with 60% prediction accuracy, up to 10 sec before the conscious decision." These findings seem to say that there cannot be a causal relationship between the RP and the "urge to move", else it would occur every time. In which case Libet appears to have misinterpreted what the RP is.

Clarke (2013) lists a number of other published criticisms of the Libet experiments. Firstly, Libet takes the RP to represent a decision to move; i.e., he believes that there is a causal connection between the beginning of the RP and the action, if and when it comes. However, it appears that there is no neural connection between the areas that build up potential in the RP (the sensory motor-cortex) and the parts of the brain associated with decision making (in the parietal lobe). Thus, in addition to there being a disconnect in practice, there is no obvious mechanism for the RP to cause the urge to move, either.

Secondly, if the RP was the cause of the movement, then we would expect a strong correlation between the timing of the onset of RP and the timing of the urge to move. But this correlation does not occur. Experiments to test this seem to rule out the RP as cause of the urge to move, though not of the movement itself. So, at the very least, the mechanism proposed by Libet has a missing link. And that link might play an active role in the process (i.e., might be directly causal).

Alfred Mele offers a third criticism of the Libet experimental set up, noted by Clarke (2013). In Libet's experiments, the data was only stored when a movement is initiated. Libet collected no data on what happened if the subject decided not to move. This vitiates his finding because it's entirely possible to prepare to move, which would presumably initiate an RP, and then not experience the urge to move. If an RP can exist and not give rise to an urge to move, then RP may not related to the urge to move at all. This possibility ought to have been excluded, but was not. A variation on the experiment by Trevena and Miller did collect data on cases where the decision was made not to move. The RP was the same whether they moved or not. And this suggests that RP does exist without giving rise to the urge to move, or an action, which undermines Libet's conclusion that RP represents a decision to move. 

A fourth weakness was pointed out by Hermann, et al. They set the experiment up as a decision to press one of two buttons in response to a stimulus. They also found the RP appearing before the urge to move occurred, but the RP occurred even before the stimulus appeared and thus is unlikely to have been related to a decision about which button to press. Again, the evidence points away from a coupling of RP and the urge to move. In fact, Schurger, et al (2012) showed the decision to move occurs very late in the course of the RP, not at the initiation of it.

Libet himself argued that his interpretation showed that, although freewill in the sense of consciously initiating actions was ruled out, we still had the option of inhibiting actions between the initiation of the RP (what he called the decision to move) and the urge to move. Some people called this "free-won't". This might be an interesting thread to follow up, except that considering the various critiques of Libet's experiment and interpretation, it seems that treating the initiation of the RP as the decision point makes no sense.

The real nail in the coffin, however, was published in Feb 2016 (just a couple of weeks ago as I write this). Libet was focussed on spontaneous voluntary movements (SVM) and it turns out that these are rather different in their underlying dynamic than movements initiated in response to a stimulus. Citing from Deric Brown's blog:
"A new generation of experiments is now suggesting that brain activity preceding spontaneous voluntary movements (SVMs) 'may reflect the ebb and flow of background neuronal noise, rather than the outcome of a specific neural event corresponding to a ‘decision’ to initiate movement... [Several studies] have converged in showing that bounded-integration processes, which involve the accumulation of noisy evidence until a decision threshold is reached, offer a coherent and plausible explanation for the apparent pre-movement build-up of neuronal activity.'" (Shurger et al. 2016)
So what looks like a build up of "readiness potential" is, in fact, happening because of anticipating having to make a decision at some point (and in the experiment the subject is explicitly primed to do so). The actual decision is reached when background neural activity reaches a peak:
"In particular, when actions are initiated spontaneously, rather than in response to a sensory cue, the process of integration to bound is dominated by ongoing stochastic fluctuations in neural activity that influence the precise moment at which the decision threshold is reached. ... This, in turn, gives the natural but erroneous impression of a goal-directed brain process corresponding to the ‘cerebral initiation of a spontaneous voluntary act’"
In other words, if we look again at the graph of "readiness potential" the decision to move comes at the peak of neuronal activity not at the onset of the RP. The RP is an accumulation of more or less random neuronal activity. This would explain some of the contradictory results mentioned above.

But, crucially, what this suggests to me is that the urge to move precedes the decision to move. The urge to move may, in fact, be an important factor in the decision to move. So it seems that Libet's interpretation of his experiment was flawed in these various ways and that freewill is back on the menu.


I discovered this information because I happen to read a number of neuroscience blogs and Twitter feeds and one of them happened to mention this new article by Shurger et al. (2016) which drew me into the subject anew. But there was always debate. Over what freewill means. Over what the readiness potential represents. Over the causal relationship between the readiness potential and the urge to move; or between the RP and the actual movement. Over Libet's experimental methods.

This experiment is so often presented with a one-sided interpretation, with no mention of the mass of contradictory evidence that make Libet's interpretation look doubtful. There is no mention of the intense debate that has ensued. Any reader could be forgiven for thinking that it was an open and shut case or that Libet had definitively shown that freewill could not exist. But this was never the case. The interpretation of the experiment could never be considered unequivocal proof of anything. The weakness of Libet's experimental design and the many contra-indications for Libet's interpretation of the readiness potential as a decision or even as causal, ought to have been given more prominence in the discussion of freewill.

Most scientists are aware of the problem with "proving" an hypothesis anyway. As Karl Popper observed, an hypothesis can really only be disproved or a conjecture refuted. A scientific theory may make more or less accurate predictions. For example, the Higgs Boson has not in fact been proved to exist. However, the theory (The Standard Model of Particle Physics) did predict a particle would be found in a certain energy range and such a particle was found in the Large Hadron Collider. So the theory survives another test, and we now try to test other predictions that it makes. The theory could fail at any point, and many scientists hope that it does, because that would make their work far more interesting and open up the field to new discoveries. The failure of the Standard Model would initiate a golden age of inquiry into the nature of the universe. Scientists are frankly bored by the idea that everything has been discovered. Which is the opposite of how they are sometimes portrayed and the opposite of religious approaches to knowledge. 

When so-called scientists give a biased presentation of an issue, citing only the evidence for their interpretation and avoiding even mentioning that there is considerable evidence against it, then that is a kind of fraud. Scientists committing such fraud ought to be censured by their peers. False statements ought to be retracted. And I think in this area of freewill many scientists are guilty of this kind of fraud. And many laypeople have repeated the fraudulent claims and perpetuated a falsehood.

In this the public have been extremely badly served by lazy journalists who have simply failed to report the experimental evidence. Whether this also amounts to fraud depends on your point of view. I see the primary function of journalism as being entertainment. Entertainers are always allowed some "poetic licence" to deceive us about facts if the version of events they present is more entertaining than reality. Hollywood films almost always distort history because the real story is often boring. Science journalists are a mixed bag and you never know which kind of story you are reading, but these days I just assume, with very few exceptions, that if a journalist is writing they are seeking to entertain rather than inform.

There will be those who cite this case as showing that the scientific method is broken. That in overturning a previous interpretation of the data science has proved that it cannot be trusted. To my mind, it says completely the opposite. This is science in action. This is the scientific process at work. The overturning of previous interpretations is part and parcel of embracing science. What we think we knew today is quite likely to be overturned tomorrow. For the religieux seeking certain knowledge and believing that they have found it, this seems anathema. That knowledge could be transient and contingent makes it seem untrustworthy. Religion is predicated on the idea of absolute knowledge, from which comes certainty, and relaxation, as all mysteries are resolved in the long run. But that is an impossible fantasy. In the real world, things are messy. Knowledge is never absolute. There is always the possibility of being wrong.

Religieux seem very uneasy with the idea that they might be wrong. Buddhists, in particular, seem to find this concept deeply troubling. Scientists, by contrast, embrace uncertainty and the principle that all knowledge may be overturned by a better explanation. Science progresses by testing ideas to destruction. This attitude of contingency with respect to knowledge of the world is, in fact, far more in keeping with Buddhist ideology. Most Buddhists appear to believe that the world can be understood in absolute terms, that the Buddha was omniscient in this sense, and that the Dharma is an expression of this absolute knowledge, i.e., that it represents absolute truth. They further believe that we can come to this absolute knowledge through introspection and believe that we cannot come to knowledge through examining the world. I have been told by a colleague, for example, that "no amount of study of the brain will ever tell us anything about the mind". Which is just Cartesian Dualism, as far as I can see, and thus a thesis that has already been soundly refuted.

However, despite having cast considerable doubt on the Libet interpretation, this is not the end of the story. There are other arguments against free will that are much more difficult to tackle than Libet's and his Determinist fans, for example, the argument by Sabine Hossenfelder on the Backreaction blog. I don't necessarily agree that arguments from fundamental laws eliminate the possibility of unexpected emergent properties that are indistinguishable from free will, but she still makes a strong argument for anyone who acknowledges the laws of physics. And so the arguments will go on. But as religieux we do need to be wary of pursuing a conjecture only because it supports our doctrine. Freewill is interesting because without it Buddhist ethics would be meaningless. If we seek only to bolster our view, rather than to seek the truth, then the possibility of being wrong is excluded and we are unlikely to accept that we have been wrong when the evidence becomes unequivocal. I see this happening in the area of the afterlife, for example. Intelligent people must always hold to the possibility of being wrong. But intelligent people are also the most reluctant to reconsider their considered views. Intelligent religious people are the worst. 



Alexander, P., et al. (2016) Readiness potentials driven by non-motoric processes. Consciousness and Cognition, 39: 38–47. doi:10.1016/j.concog.2015.11.011

Churchland, P. S. (2013) Touching A Nerve: The Self as Brain. W. W. Norton & Co.

Clarke, P. G. H. (2013). The Libet Experiment and its Implications for Conscious Will. Faraday Paper No. 17. Faraday Institute for Science and Religion.

Fried, I., Mukamel, R. & Kreiman, G. ‘Internally generated preactivation of single neurons in human medial frontal cortex predicts volition’, Neuron (2011) 69: 548-562.

Goldhill, O. (2016). Why are so many smart people such idiots about philosophy? Quartz. March 05,

Libet, B. (1985). Unconscious cerebral initiative and the role of conscious will in voluntary
action. Behavioral and Brain Sciences, 8: 529-566.

Libet, B., Gleason, C.A., Wright, E.W. & Pearl, D. (1983). Time of unconscious intention to
act in relation to onset of cerebral activity (Readiness-Potential), Brain, 106: 623-42.

Schurger, A., Sitt, J.D. & Dehaene, S. (2012) An accumulator model for spontaneous neural activity prior to self-iniated movement. Proceedings of the National Academy of Sciences. USA doi:10.1073/pnas.1210467109

Schurger, A. et al. (2016) Neural Antecedents of Spontaneous Voluntary Movement: A New Perspective. Trends in Cognitive Sciences, 20 (2), 77 - 79.

Additional reading

Fischborn, Marcelo (2016) Libet-style experiments, neuroscience, and libertarian free will. Philosophical Psychology. 1(9) doi: 10.1080/09515089.2016.1141399
"The general result is that neuroscience and psychology could in principle undermine libertarian free will, but that Libet-style experiments have not done that so far."

06 February 2015

Do We Have Freewill?

In the latter half of the 20th century a series of pioneering experiments by Benjamin Libet, a neuroscientist at the University of California in San Francisco, demonstrated a rather startling phenomenon. Libet was able to show that a conscious decision to flex one's wrist was preceded by brain activity which prepared to make the movement. It appeared that we decide unconsciously to make the move, the brain prepares to send the signal to move, and only then do we become conscious of having made a decision. This experiment and others like it have been interpreted by many as showing that freewill is "an illusion". In this essay I explore this argument and outline an important counter-argument by Patricia S. Churchland, Professor Emerita of Philosophy at UC, San Diego. I also look briefly at the determinist argument that some physicists profess. Freewill is not a particularly interesting problem, but since a lot of people talk about it, this is my two cents worth.

Is My Unconscious Part of 'Me'?

The first assumption to look at in the claims based on Libet is the idea that unconscious mental activity is somehow excluded from the freewill debate, even though it occurs in the same brain. But if my unconscious mental activity is not 'mine' then whose is it? The conclusion seems to be that when a decision is made unconsciously, even though it is our brain that makes the decision, that the decision does not count as freewill. Churchland sees this as a manifestation of matter/spirit dualism that separates out reason as a function of spirit. As I explain in my essay on this metaphor, having associated reason with spirit (arguing that reason itself is the essence of being human) it is entailed in the metaphor to then see reason as  "good" and the unconscious as more closely related to matter and therefore "bad". Additionally, reason appears to be under our control and the unconscious is not. Indeed part of the power of the Libet results is that it shows that reason is not under "our" control at all. It begins to look like a byproduct or an afterthought. However the general view of reason is in desperate need of an overhaul. 

I've gone over this material many times now: Damasio and others have shown that all decisions involve weighting of information via emotional resonances. In making a decision we defer to our emotions and find reasons afterwards (See Facts and Feelings). The practical demonstration of this is found in the advertising industry which, since the 1920s and the interventions of Edward Bernays, has appealed to desires rather than to reason when selling products and ideas. Bernays was able to apply his uncle Sigmund Freud's ideas to changing views. Most famously he convinced women to break the social taboo on women smoking by linking cigarettes with suffragettes. He did this by paying debutantes to pose smoking cigarettes during a parade, and alerting the press so they published the pictures under headlines touting cancer-sticks as "torches of freedom" and thus doomed several generations of women to horrible deaths from cancer and emphysema. (See Culture Wars, or The Society Pages) Sometimes taboos are good! In addition I've repeated cited the argument by Mercier & Sperber that in fact individuals are terrible at reasoning (An Argumentative Theory of Reason). We almost always fall into bias or fallacy when trying to reason on our own. They argue that this is not the case in small groups where different ideas can be kicked around and the group reasons collectively. Small groups are much better at reasoning. 

So it appears that the idea that conscious reasoning is what defines humans is long past it's use-by date. Any theory which even implicitly relies on this definition of reason ought to be discounted. Human beings make use of a range of faculties, including emotions and unconscious processes to make all decisions. Nor is it true to say that sapience is restricted to humans. We have now documented self-awareness and tool making in a number of species. Somehow the antiquated idea about reason being our highest and defining faculty still seems to be invoked, but we ought to be very wary of this. 

What Kind of Free Will are we Talking About?

Patricia Churchland makes a very important distinction about who means what by "freewill". Most philosophers and many scientists use freewill as a shorthand for "contracausal freewill". This is the kind of freewill described by Immanuel Kant. Churchland says contracausal freewill means that:
"... your decisions are not caused by anything at all—not by your goals, emotions, motives, knowledge, or whatever. Somehow, according to this idea, your will (whatever that is) creates a decision by reason (whatever that is)." (2013: 179; emphasis in the original)
When some scientist says, on the basis of Libet, that we have no freewill, this is what they appear to mean. They are arguing that we have no contracausal freewill, because conscious reason comes into play late in the decision process. Apart from the fact that this definition of freewill is counterintuitive and seems unlikely to non-philosophers, we've already undermined some of the key assumptions involved in it. As discussed above, Churchland sees that entailed in this view is the idea of a non-physical soul. By disconnecting the decision making process from our bodily processes (like emotions) and assigning it to "pure" reason, those who use this definition seem to be subscribing to a matter/spirit dualism in which reason is a function of spirit not of body. 

The more commonsense variety of freewill is less well defined partly because, like many commonsense definitions, we use it efficiently without fussing over the meaning. To make us more comfortable with the fuzziness of the definition Churchland invokes George Lakoff's ideas about categories being defined by relatedness to a prototype. In this view freewill is not an all or nothing proposition, but some actions are more free than others. Some acts are more typical of freewill than others. And people are somewhat free to choose which actions most represent freedom, since categories are what we impose on experience to help organise it. Most people intuitively understand that sometimes we have more choice than others, or that sometimes people are compelled to chose one option even though in theory they have a choice. This recognition of degrees of freedom seems vital to any sensible theory of how we make choices, especially moral choices. 

Churchland argues that:
"...if contracausal choice is the intended meaning, the claim that free will in that sense is an illusion is only marginally interesting, Because nothing in the law, in child-rearing, or in everyday life depends in any significant way on the idea that free choice requires freedom from all causes." (184)
In other words the freewill that is being denied by philosophers is not very interesting because, being divorced from experience, it's hardly credible anyway. Churchland likens the claim that contracausal freewill is an illusion to announcing that alien abductions are not real. The response is, "So what?", "Who cares?" or "Duh!" Those who deny freewill on the basis of the Libet experiments are not saying anything interesting, though of course at first glance it appears to be a controversial thing to say so the media covers it and the meme gets spread. This whole section of the debate about freewill can safely be shelved with other legacy ideas from philosophy that are no longer relevant. The question is not "Are we free?", but "How free are we now and how free can we be?"

Self Control

Even if there is some doubt about what freewill means, Churchland argues that there is a related concept about which there can be no doubt: self-control. She points out that self-control, the over-riding of impulses to act, takes conscious effort. And in terms of morality, self-control is often just as significant as conscious choice. Morality is very frequently defined in terms of refraining from actions: "thou shalt not..." (in a Christian context) or "I choose to refrain from..." (in a Buddhist setting). Libertarian secularists often complain about religious morality as just being a bunch of rules, but it might be a natural consequence of self-control being a much clearer concept. And although our laws are profoundly influenced by religious models, there has been no significant move away from prohibitive rules even in secular (or nominally secular) countries. 

Most of being a good group member would appear to be inhibiting impulses that go against group norms. Any sociable animal must at times repress selfish impulses in order to benefit the group. Social animals for example prosper by sharing food sources in a way that solitary animals do not. Our motivation for exercising this impulse control vary: fear of reprisal, shame, habit, altruism, and generosity can all come into play. Or we may feel that the "law is an ass" or decide that a small breach of the rules will draw attention to a greater breach (civil disobedience to protest government corruption for example). In other words we can be negatively motivated or positively motivated to follow established norms or to break them.

My reading of Churchland's account of the freewill debate is that for the most part it is poorly framed and thus does not produce interesting results. The reasons for considering contracausal freewill to be the best definition are no longer plausible if they ever were. It serves to confirm that the freewill debate, such as it is, is not particularly interesting. 

Making Moral Judgements

This is not to say that the matter of voluntary actions is unimportant. Social groups operate with norms and rules and when enforcing those norms it's important to know why breaches happened. This is why most legal systems make distinctions of degree in crimes like murder. A murder than is planned months in advance is always seen as a worse crime than one committed in the heat of the moment. A calculated crime is relatively more serious than an impulsive one. This is because consciously breaking the rules is a clear repudiation of those rules. In this case we have serious doubts about the willingness of the person to return to lawfulness. Part of any calculation to commit a crime is usually elaborate planning to avoid detection and punishment. Even if the rule-breaker shows remorse, we have reason to distrust them in the future.

The crime of impulse however is more likely to be understood as a momentary lapse and to be treated more leniently if accompanied by suitable remorse and a willingness to admit fault. Those who plead guilty tend to get lighter punishments. However if someone is prone to repeated crimes of impulse then we tend to treat them like the person who does calculated crimes, because we cannot trust them to keep the rules.

If someone sets out to injure a person and that person inadvertently die then this is less serious than if the assailant intended kill. It might still be considered murder depending on how we judge the risk involved. An attack with a weapon is more likely to kill than a fist-fight for example. This situation can be seen in the light of calculation and impulse also. If someone is killed purely by accident, with no intent to harm, we may still be found culpable for depriving them of life, but the consequences may be still less severe. For example neglecting our duty of care while doing an inherently dangerous activity, like driving a car, is still quite a serious crime. But if we were proceeding with due care and a pedestrian crosses the street without looking causing them to be knocked down and killed then we are not culpable even those someone has died.

On the other hand if we kill someone in the process of defending ourselves or our property we may not be culpable at all as long as the force we used is judged to be proportionate to the threat we faced. Police officers and soldiers are seldom held to be culpable of murder when they kill someone in the line of duty, even though the community may feel they should be held accountable. This is extremely controversial, but in a culture where murder is fairly routine the enforcement of law comes with severe risk. It's unreasonable to expect police to risk their lives when apprehending a suspect. Soldiers are not given carte blanche to kill. Under the modern rules of war, they may not purposefully kill civilians for example, though this is not a universally recognised restriction especially in asymmetric war where one side is far more powerful than the other. Soldiers may not only kill enemy combatants, but will be rewarded for doing so. In the Vietnam War, efficiency guru Alain Enthoven used the "body count" as a measure of how well the war was going (he subsequently was brought in to reorganise the British health service by introducing the "target culture").

People can be found not-guilty of even the most serious crimes if they do not have the ability to understand the consequences of their actions - either permanently or temporarily. We often detain such people purely on safety grounds. In making judgements about the severity of breaches of social norms we have to take many degrees of intentionality and self-control into account.

Thus an all-or-nothing freewill is not a very helpful instrument in thinking about morality. Moral judgements can be very complex indeed and always take in the motivations and the underlying mental and emotional state of the perpetrator (and often the victim as well). Thus contracausal freewill is fully irrelevant to how our laws operate and to how common sense morality operates (as already pointed out by Churchland). 

As an aside, it is interesting that the baby boomer counter-culture seemed to be all about allowing one's impulses free reign. From "free love" to "greed is good", sections of the post-war generations felt the need to stop restraining themselves and let it all hang out (as the saying goes). As it turns out the backlash against this call for loosening of social restraints has been a far more significant social movement. Neolibertarianism was driven primarily by conservative business people. They wanted freedom from government control on their collective ability to do business, and conceived of this within strong social boundaries which restricted what was acceptable behaviour. The irony is that Neolibertarians are often authoritarian control freaks. They saw increasing liberalism and individualism as a threat to their way of life and took steps to take back control. Now, ironically, we struggle to pass laws to curb the excesses of those same business people even in the face of global economic instability and catastrophic climate change. We can now talk openly about sex, and women have a great deal more social equality, but the businessmen own a great deal more of the wealth and have virtual control over governments. The ideology of the world's leaders is that nothing ought to restrain the creation of profits and that abstract markets are more efficient than governments (though every empirical fact shows this to be untrue). Conservative elements in society still allow liberalism to make gains, such as same-sex marriage for example, but only where it has no consequences for the wealth of the wealthy. At the same time the threat of terrorism continues to eat away at civil liberties and individual freedoms. So the disinhibition of the 1960s is a pyrrhic victory.

The question of who is responsible for actions has become obscured to some extent by determinist scientists. The media has shown itself time and again to be highly irresponsible when reporting science. Media companies are in the business of entertainment and so news streams are only secondarily about informing us and are primarily about distraction and sensory stimulation. Scientists with a controversial message are more likely to get the oxygen of media attention than those with the more sober message. However there is still an argument about freewill based on the view that the universe is deterministic. We turn now to this argument.

Are We Deterministic Robots?

The view that being able to frame regularities in the universe in mathematical expressions, means that the universe is therefore deterministic is popular amongst physicists. In a deterministic system if we had perfect knowledge of the starting conditions, the elements, and rules, then we could perfectly describe the behaviour of the system indefinitely far into the future. This kind of Determinism was espoused, for example, by Stephen Hawking in his last book The Grand Design:
"so it seems that we are no more than biological machines and that free will is just an illusion." (32)
Sean Carroll has also expressed the view that we're all machines that think. This argument is related to the one I was exploring with regard to the afterlife. Life is made up of atoms and we understand the behaviour of atoms, so we understand the basis upon which life exists, even if we don't quite understand all the processes of life yet. But whereas the claim about the afterlife was strictly limited to the persistence of information about the person after death as governed by the Second Law of Thermodynamics (Entropy always increases in a closed system), this claim about a deterministic universe is unlimited. The unlimited nature of the claim trips it up.

It is true that we understand the behaviour of atoms at the energy, mass and length scales relevant to living things. But we also have to take into account the nature of complex systems. Even when a complex system is made up of simple elements following simple rules, the behaviour of the system is nondeterministic: we cannot predict it. When a system is made up of complex elements which combine according to complex rules and we get emergent properties at several different levels at once, then that system is decidedly not deterministic. An economy or the weather are not deterministic, not predictable.

As far as life is concerned we don't have perfect knowledge of the starting conditions and nor can we ever gain such knowledge. As far as the universe as a whole this also appears to be true. We can conjecture, but not have perfect knowledge. In fact because of random quantum fluctuations in space-time we can never be entirely sure about the elements in play. And the rules are sufficiently complex that to date no one understands them with anything like perfect knowledge (something acknowledged by Hawking, who goes so far as to say that he doubts we'll ever have a unified set of equations for the universe). The mathematics describing a single sub-atomic particle interacting with all the known fields has yet to be solved: it involves 7 or 9 extra dimensions of space that themselves at so small that they add nothing to the dimensionality we experience.

We can demonstrate the problem by considering a simple pendulum and then adding complexity. A simple pendulum vibrates in two dimensions, with one end fixed. The behaviour of this pendulum follows a simple law: the period of the vibration for small amplitude (θ << 1) is approximated by:

Where L is the length of the pendulum and g is the acceleration due to gravity. In fact for longer amplitudes the equation is more precisely:

This is complicated, but in fact not difficult to solve to an arbitrary level of accuracy (the factors in the series quickly become vanishingly small). For most large clocks only one or two members of the series are required for sufficient accuracy in calculations.

Intuitively we might think that adding a joint to the pendulum halfway along it's length, in effect a pendulum attached to the end of another pendulum, would complicate matters, but not so much. But in fact a double pendulum's motion is chaotic. Technically if we precisely specify the starting conditions we can predict it's motion, but we can only calculate the next moment, by precisely knowing what has happened from time = 0. For each moment in time the calculation gets longer until it very quickly becomes too difficult a problem for all the computing power inherent in the universe. If we start at an arbitrary time we have almost no chance of calculating what will happen next. A double pendulum is still technically deterministic, because it is theoretically possible to know the starting conditions, the precise details of the system, and the rules that must be followed.

If we conceive of an atom as being connected to other atoms by forces, then a system with two atoms would be like a double pendulum with no fixed end and instead of vibrating in only two dimensions they vibrate in three. The motions of these two atoms are chaotic and far more difficult to predict than a simple double pendulum, i.e. far more difficult than virtually impossible.

Now consider than there are of the order of 10100 atoms in the universe and all of them are connected via forces to each other. And we need to keep in mind that atoms, themselves are in fact systems of smaller particles which are again all interacting with all the other particles, and that fundamentally all that we see as particles and forces are simply vibrations of interacting fields that extend throughout the universe. Conceived of as a pendulum the overall motion of the universe is essentially infinitely complex. Even if we could precisely define the first moment in the history of the universe (something we cannot yet do), then by the second moment the vibrations in the various fields would be impossible to calculate. By the time particles appeared on the scene as an emergent property of the cooling universe, the system is already impossible to predict on the lowest scales. A system like this cannot be considered deterministic, even in theory.

What Kind of Ordered Universe Do We See?

So an obvious question then is, why do we see ordered behaviour at all? The order we see emerging from this 3D pendulum with 10100 moving parts is because of emergent properties when looking at different scales. Order, or quasi-order, appears in chaotic systems. Think of a hurricane. From space it looks like a relatively regular spiral, or a circle, even though at ground-level it can be chaotic. Also the intensity of the forces involved follow inverse square laws, or inverse fourth-power laws. In theory all fields extend throughout the universe, but the effects of forces are typically short range. Gravity is the only force with a very long range and that is mainly because the masses involved in cosmological phenomena are unimaginably large.

The characteristic ordering (or quasi-ordering) we see depends on the scale we adopt. For example 1g of pure carbon contains about 6 x 1023 atoms. In a previous essay I pointed out that if each atom was one millilitre in volume, that gram of carbon would fill the western Mediterranean Sea. The atoms are in motion, but the motions are many orders of magnitude smaller than a human eye can see. When we look at this many atoms, the tiny motions of each atom are cancelled out by other atoms doing the opposite. Each atom is regular in a number of ways: each carbon atom has six protons and six electrons, and either 6, 7, or 8 neutrons (giving 12C, 13C, and 14C), the chemistry of carbon is very predictable and the shape of its molecules known very precisely. But a diamond, a single gigantic molecule of carbon atoms, does not behave like an individual atom. Crystals are macro-structures that exhibit different kinds of regularities than atoms do. Sit two diamonds together and they do not interact, do not behave as a system at all. Carbon macro-molecules have very different properties to individual carbon atoms. A carbon atom is highly reactive and can form millions of compounds. Diamond by contrast is one of the most inert naturally occurring substances.

Steven Hawking wants us to believe that people are just complex machines. But this is not credible either. Perhaps at some absolute level of abstraction this is true, but not in any meaningful sense. The most complex machines we can make are still less complex than a single cell in our body. We are made from atoms, but millions of billions of billions of atoms, following complex rules; built up from another system of simpler components, also following complex rules, itself the visible manifestation of fields. We could not specify all the atoms of a person and predict what was going to happen next without first calculating every vibration in every field in the entire universe from the first moment in time. With all due respect, Hawking might be a good physicist, but he appears to be a poor philosopher. This may be why he also wrongly claims that philosophy is dead. There is nothing deterministic about a human being, which is why philosophy is very much alive (if not entirely well).

Nothing we know about the emergent properties of collections of Septillions of atoms rules out freewill as an emergent property. Nor are consciousness, or for that matter life itself, ruled out as properties of these unimaginably complex systems. We are very far from having plumbed the depths of the complexity of the universe, despite the fact that the elements and the rules governing the system are quite clear. An analogy here is the chess board. There are 32 pieces on 64 squares and the game has clearly defined rules. We can calculate the theoretical number of different games, and the best computers are better than the best humans, and yet not once has a recorded game ever been the same as a previously recorded game. The difference is that our game has 10100 pieces!

So, Do We Have Freewill?

The answer to the freewill question appears to be the one that is ascribed to the Buddha in last week's sutta translation and commentary. We unquestionably have some choice And at the very least we exercise self-control. Perhaps this is why the Buddhist precepts are phrased in terms of refraining from actions? 

The arguments against freewill that have emerged recently in the scientific community are simply poor philosophy. As Mary Midgley (1979) has said:
"There is now no safer occupation than talking bad science to philosophers, except talking bad philosophy to scientists."
That so many scientists are poor philosophers is of course deeply unhelpful. Midgley had Richard Dawkins firmly in her sights in making this comment. She considered his metaphor of the "selfish gene" to be very poor philosophy indeed (as do I). To be fair Dawkins and his followers thought Midgley completely misunderstood what he was getting at. From my point of view, Dawkins' idea is just a Neolibertarian reading of Darwinism. That's not science, it's not even philosophy really; it's ideology. What's more Neolibertarianism is rooted in the Utilitarian philosophy of Jeremy Bentham, which is really rubbish philosophy since it fundamentally misunderstands human beings. Many of these behemoths of popular science are in fact quite poor at philosophy and have created a legacy of poor thinking—especially in the form of unsuitable metaphors—that will continue to haunt intellectuals for many years to come. 

In many ways this debate about freewill is simply silly. It's a legacy of theological debates that were silly to start with. In order to deny freewill one must make a choice. In order to argue against free will, one must make a sustained effort. It's simply not credible. Of course one can choose not to believe in freewill, but that argument is self-defeating. Anti-free will campaigners must argue that they are compelled to believe what they do. This leaves them trying to explain why not everyone is compelled to the same conclusion. If we are not free, then we are apparently not free in a variety of different and conflicting ways. The different conclusions are a powerful argument against determinism if ever there was one. 

Churchland, Patricia S. (2013) Touching A Nerve: The Self as Brain. W. W. Norton & Co. 
Midgley, Mary. (1979) 'Gene-juggling'. Philosophy. 54(210): 439-458.

See also:-
Metzinger, Thomas. (2013) "The myth of cognitive agency: subpersonal thinking as a cyclically recurring loss of mental autonomy." Frontiers of Psychology, 19 December 2013 | doi: 10.3389/fpsyg.2013.00931.

30 January 2015

A Sutta on Freewill

your move...
This is a text I recently stumbled upon. It is quite interesting because it directly addresses the issue of freewill, something I have not come across in a Buddhist text before. The Buddha is seen arguing with Brahmin who denies freewill and argues for a form of determinism or fatalism. This kind of view is popular again today. Advaita Vedanta enthusiasts, such as Gary Weber, also hold that free will is an illusion and that everything is predetermined. We are increasingly seeing the influence of Advaita Vedanta on Buddhists who use the self-enquiry methods of Vedantins. Many scientists are also determinists are well. Therefore knowing how the suttakāro dealt with this assertion is of some interest.

The title of the sutta is Attakārī Sutta AN 6.38 (AN iii.337-8). The adjective atta-kārin and noun atta-kāra are central to the text so let us first pause to consider what they mean. Attan (atta- in compounds; ātman in Sanskrit) is, of course, the reflexive pronoun 'self, own'. It's not being used here in the sense of a metaphysical self. It is being used in an empirical sense: the experiential self, or, for the finicky, the physical locus of awareness and intention, broadly speaking the body ('body' is one of the meanings of Vedic ātman). Since the text itself provides the argument for this, we'll let it speak first. The other part of the compound is kāra 'act, deed'. Like the world karma it stems from the verbal root √kṛ 'do, make'. So, atta-kāra refers to 'one's own act'. In this type of compound the -kāra can mean 'maker' (literally 'one whose action is...'). So the term suttakāra can mean the one whose act was the creation of the suttas, a 'sutta-maker'. Another term, drawn from the Sāṅkhya school is ahaṃkāra 'I maker'.  Kicca-kāra is doing what ought to be done, doing one's duty.The adjectival form atta-kārin means 'doing one's own action'. The word para used as a pronoun means 'other' and contrasts with attan. If attakāra is one's own action, then parakāra is another's action.

The text begins with the meeting of the Buddha and an unnamed Brahmin who tells the Buddha his view, there's no nidāna beginning 'evaṃ me sutaṃ' or telling us where the encounter takes place, we just dive straight in. The whole Pāḷi text is cited below, with my translation and commentary interspersed.
Atha kho aññataro brāhmaṇo yena bhagavā tenupasaṅkami; upasaṅkamitvā bhagavatā saddhiṃ sammodi. Sammodanīyaṃ kathaṃ sāraṇīyaṃ vītisāretvā ekamantaṃ nisīdi. Ekamantaṃ nisinno kho so brāhmaṇo bhagavantaṃ etad avoca – ‘‘ahañhi, bho gotama, evaṃvādī evaṃdiṭṭhi – ‘natthi attakāro, natthi parakāro’’’ti. 
Just then a certain Brahmin approached the Bhagavan and exchanged polite greetings. Having greeted each other the Brahmin sat down on one side and spoke to the Buddha. "Mr Gotama, my philosophy, my view, is that there is no 'one's own action'; there is no 'another's action'.
Bodhi translates attakāra as 'self-initiative' which I think hints more at free will. I suppose we could say it means acting on one's own accord, or being free to act. Bodhi wants us to think about who is initiating the action. Vedantists say that no one initiates the action. Things just happen. There are hints here of Sāṅkhya darśaṇa. The Sāṅkhya view is that in reality what is most fundamental in us is a passive essence called puruṣa. The active side of experience (prakṛti) is like a distracting illusion that keeps puruṣa involved in the world of matter and away from quiescent perfection (kevala - literally isolation). In order to get back to perfection one has to role back the illusion until prakṛti returns to it's quiescent potential state. Sāṅkhya is very vague on some of the details of this view and many of the questions we'd like to ask don't seem to have answers on Sāṅkhya terms. But this view that there is no such thing as 'one's own action' shares some characteristics with Sāṅkhya. This is apparently news to the Buddha.  
Māhaṃ, brāhmaṇa, evaṃvādiṃ evaṃdiṭṭhiṃ addasaṃ vā assosiṃ vā. Kathañhi nāma sayaṃ abhikkamanto, sayaṃ paṭikkamanto evaṃ vakkhati – ‘natthi attakāro, natthi parakāro' ti!
Brahmin I've never seen or heard of this philosophy, this view. For how indeed does one who comes and goes under his own steam possibly say: there is no 'one's own action'; there is no 'another's action'.
So the Buddha's first reaction to this previously unknown philosophy is to ask how anyone who had just walked up to him, greeted him, sat down on one side, and stated his philosophy (all apparently of his own free will) could possibly believe that he did not do so of his own accord. The commonsense response is that the view cannot make sense of what is happening right now. The Brahmin arrives by himself (sayaṃ abhikkamanto) and he leaves by himself (sayaṃ paṭikkamanto). So the determinist view is at best counter-intuitive.

The Buddha then asks a series of questions:
Taṃ kiṃ maññasi, brāhmaṇa, atthi ārabbhadhātū ti?
Evaṃ, bho.
Ārabbhadhātuyā sati ārabbhavanto sattā paññāyantī ti?
Evaṃ, bho.
Yaṃ kho, brāhmaṇa, ārabbhadhātuyā sati ārabbhavanto sattā paññāyanti, ayaṃ sattānaṃ attakāro ayaṃ parakāro.
Do you think, Brahmin, there is a factor of instigation?
Yes, Sir.
And when there is a factor of instigation, is it evident that beings are instigating?
Yes, Sir. 
So, when there is a factor of instigation and it is evident that beings are instigating, this is the 'one's own action' of beings, this is another's action.
This question is obvious. It stems from what the Buddha said initially. If we see beings instigating actions (ārabbhavanto) then why would we assume that they are not doing their own actions? 'Instigation' is a translation of ārabbha from the verb ā√rabh 'to begin'. Here dhātu is similar to the word dharma in many respects: 'a factor, an element'. 
Taṃ kiṃ maññasi, brāhmaṇa, atthi nikkamadhātu…pe… atthi parakkamadhātu… atthi thāmadhātu… atthi ṭhitidhātu… atthi upakkamadhātū ti?
Evaṃ, bho.
Upakkamadhātuyā sati upakkamavanto sattā paññāyantī ti?
Evaṃ, bho.
Yaṃ kho, brāhmaṇa, upakkamadhātuyā sati upakkamavanto sattā paññāyanti, ayaṃ sattānaṃ attakāro ayaṃ parakāro.
Do you think, Brahmin, there is a factor of going out... a factor of advancing...  a factor of resistance... a factor of endurance... a factor of approaching?
Yes, Sir.
And when these factors are present is it evident that beings are performing them?
Yes, Sir.
So, when there are these factors and it is evident that beings are performing them, this is the 'one's own action' of beings, this is 'another's action'.
Note the CST version of the text here seems to have been abbreviated more than the text that Bhikkhu Bodhi translates in his AN translation (p.902-904). I've followed the text I have, though I rather than using only the first and last members of the list, I've rendered the final question as a collective inquiry about all the actions involved.

The Buddha lists a series of generic actions which beings are seen to perform. And he asks the same question in each case. And, weirdly, the Brahmin answers "yes" in each case. And the Buddha simply points out the obvious: we all make choices all the time and act on intentions all the time. To argue against free will on some abstract principle is bizarre. Presumably the Brahmin thinks that even though we give the appearance of willed actions, that this is an illusion, a la Sāṅkhya or Advaita Vedanta. But the Buddha is far from impressed by this and repeats the phrase above:
Māhaṃ, brāhmaṇa, evaṃvādiṃ evaṃdiṭṭhiṃ addasaṃ vā assosiṃ vā. Kathañhi nāma sayaṃ abhikkamanto sayaṃ paṭikkamanto evaṃ vakkhati – ‘natthi attakāro natthi parakāro’ ti.
Brahmin I've never seen or heard of this philosophy, this view. For how indeed comes and goes under his own steam possible say: there is no 'one's own action'; there is no 'another's action'.
Then the Brahmin, in a predictable change of heart, converts to being a follower of the Buddha:
Abhikkantaṃ, bho gotama…pe… ajjatagge pāṇupetaṃ saraṇaṃ gatan ti!
It is amazing, Mr Gotama... etc... from this day on [I've] gone for refuge for life. 
Again Bhikkhu Bodhi seems to have an unabbreviated text. I translate the text as I have it. Bodhi says that the Brahmin becomes a lay follower. So a determinist is now convinced that we have free will (attakāra) simply be having the obvious situation pointed out to him. Not a very inspiring story - he doesn't even argue. But it shows that free will is a given in early Buddhism.

This word attakāra is in fact quite rare. It occurs in only one other sutta, Jātaka 528 (Mahābodhijātaka) and an Apadāna Story (i.24). The sutta is the Samaññaphala Sutta (DN 2) where this view on self-willed actions is associated with Makkhali Gosāla (DN i.53-55). Makkhali is a determinist, in that he doesn't believe any theory of causation or conditionality, nor does he see the point in religious exercises. He sums up his view as
Seyyathāpi nāma sutta-guḷe khitte nibbeṭhiyamānam eva paleti, evam eva bāle ca paṇḍite ca sandhāvitvā saṃsaritvā dukkhass'antaṃ karissantī ti.
Just as a ball of string that is thrown, will run away always unwinding, even so the fool and the wise running on, circling around, will eventually make an end of suffering. 
So despite being a fatalist, he's also an optimist because he believes that events will play themselves out positively. The ball of string will eventually unravel and the end of dukkha will be reached. It's just that there is nothing we can do to speed the process up and no external power that can come to our rescue. What will be, will be, and it will take as long as it takes. One just has to accept that events will play themselves out for the best. To counteract this we simply point out that one can choose to believe that or not. It's up to the person, because we do in fact have choices.

These days many scientists are also determinists with no teleological bent: "there is no free will; what will be, will be; we have no idea what it will be, except that the entropy of the universe is increasing." Tackling this view is a more difficult problem that I'll try to address in my next essay.


04 April 2014

Experience and Free Will in Early Buddhism

The Four Humours
image via Musings on the 18th C
In a recent exchange in comments on Dhīvan's blog, Two Meanings of Karma, he drew my attention to the Sīvaka Sutta. This sutta says that kamma is only one of eight causes of experience and introduces the term pubbe-kata-hetu "caused by former actions", which is discussed below. Related suttas help to flesh out what is meant by the term and place important limits on the doctrine of karma.

The early Buddhists were critical of the view that everything we experience is a result of past actions because it is a form of determinism that eliminates meaningful moral choices. As such this teaching touches on the subject so dear to Western moral philosophers, i.e. free will. I begin with my translation of the Sīvaka Sutta and a discussion of the main terms and ideas and then contrast it with the Titthāyatanādi Sutta (AN 3.61) with passing reference to the Devadaha Sutta (MN 101).

Sīvaka Sutta SN 36.21 (iv.230)
One time the Bhagavan was staying in Rajgir in the Squirrel Sanctuary Bamboo Grove. Then the ascetic Moḷiya-Sīvako approached the Bhagavan and greeted him. When they have exchanged pleasentaries he sat to one side. And sitting on one side he asked:
Mr Gotama, what would you say to the toilers and priests whose ideology is "whatever a person experiences (paṭisamvedeti), whether pleasant, unpleasant, or neither, is all caused by past actions (pubbekatahetu)"?
[The Bhagavan replied], Sīvaka, some experiences arise from the rising of bile (pitta); [this fact] one can personally know and [it] is considered true by [people in] the world. In this case, Sīvaka, those toilers and priests who claim that "whatever a person experiences is all caused by past actions" are wrong (micchā). They overshoot what is personally known and what is considered true by [people in] the world.
Sīvaka, some experiences arise from the rising of phlegm (semha)...
Sīvaka, some experiences arise from the rising of winds (vāta)...
Sīvaka, some experiences arise from interactions of the humours (sannipātika)...
Sīvaka, some experiences are produced by changes in the season (utu-pariṇāma)...
Sīvaka, some experiences are produced by adverse circumstances (visama-parihāra)...
Sīvaka, some experiences arise from physical injury (opakkamikānipi)...  
Sīvaka, some experiences are also produced by the ripening of actions (kammavipāka); [this fact] one can personally know and [it] is considered true by [people in] the world. In this case, Sīvaka, those toilers and priests who claim that "whatever a person experiences is all caused by past actions" are wrong (micchā). They overshoot what is personally known and what is considered true by [people in] the world.
That said, Moḷiya Sāika said this to the Bhagavan: "Awesome, Mr Gotama, that's awesome. Please remember me as an upāsaka who has gone for refuge for life. 
Bile, phlegm, and wind.
The humours, and the seasons,
Adversity, injury,
And ripening of actions as eighth.

Comments on Sīvaka Sutta

Firstly for enthusiasts of the punctuation problem related to the standard Buddhist sutta opening "evaṃ mayā sutaṃ..." note the opening of this sutta "One time the Bhagavan was staying in Rajgir" (Ekaṃ samayaṃ bhagavā rājagahe viharati). If we care about such things this shows that "evaṃ mayā sutam" is one syntactic unit and "ekaṃ samayam... viharati" is another. In fact to my mind this is the obvious way to read the Pāḷi. A qualifier like ekaṃ samayaṃ is far more likely to appear before a verb or participle than after it. Reading ekaṃ samayaṃ as going with sutaṃ looks like special pleading. If we were going to punctuate evaṃ mayā sutaṃ ekaṃ samayaṃ... we'd mark a new clause with a punctuation mark after sutaṃ. Which is to say that it is not:
"Thus I heard at one time, the Buddha was staying in Rajgir";
"Thus I heard, at one the Buddha was staying in Rajgir".
Secondly the name Moḷiya-Sīvaka is quite interesting. In How Buddhism Began (p.135-164; esp. 151-4) Richard Gombrich proposed that we take Aṅgulimala to be a Śaiva doing extreme antinomian practices. Now the name Sīvaka is probably from siva 'auspicious, fortunate' with the suffix -ka (causing the lengthening of the initial vowel) and thus on face value means 'one who is auspicious'. In which case the name would be synonymous with svāstika. The Pāli is equivalent to Sanskrit śivawhich is also the name of a god: Śiva. Pāli sīvaka might well be Sanskrit śaivaka, "one who belongs to Śiva". Thus we might read Moḷiya-Sīvaka as 'a top-knotted Śiva devotee'. The Dictionary of Pāli Names has nothing to settle it either way, though the name is not common: in addition to our Sīvaka we find just one yakkha (DN iii.205; SN i.211); one physician who was Ānanda in a previous life (J iv.412); and two theras (Thag vs.14 & vss.183-4), one of whom lives in Rajgir

"Experiences" (present participle, not nominal plural) translates paṭisaṃvedeti. Here what one experiences is obviously vedanā and is characterised (as vedanā always is) as sukha, dukkha, or adukkhamasukha. Both paṭisaṃvedeti and vedanā come from the root √vid 'to know'. From this root we also get veda 'the knowledge'; vedanā 'the known', i.e. what becomes known to us, what we actually experience. Also in this passage: vedayita 'felt, experienced'; in the plural 'experiences'; such events are veditabba is 'to be known; knowable'. This cluster of terms is part of what makes "feelings" an unsatisfactory translation of vedanā. What we are talking about is that which we become aware of due to the activity of all our senses, including the mind. "Feelings" is far too narrow. 

Pubbekatahetu is a three-part compound: hetu = cause, kata = past participle of √kṛ 'to do, to make' and pubbe 'before, formerly'. The compound is a bahuvrīhi meaning 'whose cause is what was done before'. There is a related term pubbekatakāraṇa which we find in a commentarial passage on AN 3.61 where it is also glossed as "Experiencing with actions formerly performed as the only condition" (pubbekatakammapaccayeneva paṭisaṃvedeti). 

Visama-parihāra is an odd word. Bodhi translates 'careless behaviour', Thanissaro "adverse behavior", though PED suggests 'being attacked by adversities'. Parihāra is from pari√hṛ 'to attend, shelter, protect; carry about; move around; conceal; set out, take up, propose'. PED takes it to mean 'surrounding' in the figurative sense. Visama is literally 'uneven, unequal, unharmonious'. Figuratively in a moral sense, 'lawless, wrong'; and 'odd, peculiar.' Buddhaghosa glosses: "Produced by adverse circumstances" means carrying a heavy load, pounding cement etc, or snakes, mosquitoes or falling in a pit, etc for one wandering at the wrong time. "Visamaparihārajānīti mahābhāravahanasudhākoṭṭanādito vā avelāya carantassa sappaḍaṃsakūpapātādito vā visamaparihārato jātāni." (SA iii.81) Thanks to Dhīvan for helping me with this passage. I think his translation of visama-parihāra as 'adverse circumstances' is better than either Bodhi or Thanissaro and I have adopted it.

The main point is that the view that everything we experience is a result of past karma is in fact wrong (micchā). I've pointed this out before and drawn attention to the Devadaha Sutta (MN 101) as another text which refutes this view. There it is attributed to Nigaṇṭhas who we usually take to be the Jains.

Titthāyatanādi Sutta

The idea of pubbekatahetu is also criticised in the Titthāyatanādi Sutta AN 3.61 (i.173) where it is one of three sectarian heresies (tīṇimāni titthāyatanāni). Faced with such a claim as "everything that one experiences is due to past actions" the Buddha questions his opponent about the reasons for unethical behaviour (the dasa kusala-kammapatha; known in the Triratna Order as that "the ten precepts").

In some sense this is a question of free will. The idea that everything we experience is due to past action is a form of determinism. The Buddha's critique points out that if we accept a form of determinism then we have no motivation in regard to moral moral choices in the present, and thus the possibility of liberation is lost.
Pubbekataṃ kho pana, bhikkhave, sārato paccāgacchataṃ na hoti chando vā vāyāmo vā idaṃ vā karaṇīyaṃ idaṃ vā akaraṇīyanti. Iti karaṇīyākaraṇīye kho pana saccato thetato anupalabbhiyamāne muṭṭhassatīnaṃ anārakkhānaṃ viharataṃ na hoti paccattaṃ sahadhammiko samaṇavādo
However, bhikkus, for those falling back on former action (pubbekata) as the essence it is not a motivation for, not an effort towards, distinguishing between right and wrong. As a result of right and wrong not being truly and reliably ascertained, there is  dwelling forgetfully and vulnerably. [Former action] is not, on its own, the doctrine of the samaṇa who is Buddhist.
Of the three other translations I consulted (Bodhi, Thanissaro, and Piya Tan) I disagree with all of them as to how to render the last sentence. All translate samaṇavādo in the sense of 'call oneself a samaṇa'.

Samaṇa-vāda is a nominal compound in the nominative. PED sv. vāda has "2. what is said, reputation, attribute, characteristic." PED cites Sn 859 for this reading, and one Jātaka reference. The final pāda of Sn 859 reads tasmā vādesu nejati. The Niddesa glosses vādesu here as ‘criticisms, blame, reproaches, not getting any renown, not being praised’. Though the SnA has "On that account he is not cowed because of criticisms" (taṃ kāraṇā nindāvacanesu na kampati) and K R Norman seems to follow this interpretation in his translation : "therefore he is not agitated in [the midst of] their accusations". (p 107 & 338-9). But the this also fits with the context. It seems to me that PED is probably correct to include this second sense of vāda in SnA, since that is how the commentator understood it, but wrong to attribute it to Sn. It's a commentarial usage not a sutta usage. As far as I can see there is no parallel usage in Sanskrit.

I had some discussion with Dhīvan on this and he pointed out: "Looking at the comm., it’s clear that the translators are following what it says but putting it into clearer English.
na hoti paccattaṃ sahadhammiko samaṇavādo: for you beings or other beings thinking, ‘I am an ascetic’, individually the reasonable characteristic of an ascetic isn’t, doesn’t succeed. For though there are ascetics whose reason is only past action, also there are non-ascetics whose reason is only past action. ‘Reasonable’ (sakāraṇa) means having a reason. (AA ii.272)
So Buddhaghosa’s argument is that the ascetics who claim that what is experienced is caused by past action are not really ascetics because non-ascetics also believe this, it’s not a right view that will get you anywhere, so it’s hardly a good view for a so-called ascetic.
Thus we can see where the other translations are coming from. My feeling, however, is that we should always make a strenuous effort to translate the text and at the very least include it as a footnote, before adopting the commentarial gloss. Buddhaghosa's view is not that of early Buddhism, but that of 5th century Theravāda scholasticism. Sometimes it's helpful and sometimes not. Here I disagree with him. His reading is one that requires us to treat rather too many words as meaning something other than their obvious meaning.

Na hoti samaṇavādo would be an entirely straight forward sentence meaning, 'It is not the doctrine of a samaṇa' [with an emphasis by putting the verb first]. Sahadhammiko is in the same case as, and thus goes with, samaṇavādo. Piya Tan says it is an adverb 'with justice' and Bodhi also translates as an adverb, 'legitimately'. This appears to be based on the commentarial gloss: sakāraṇa (above). However as an adverb it ought to be in the accusative, not the nominative. By contrast paccattaṃ is a neuter accusative used adverbially (individually). So, sahadhammiko simply cannot be an adverb, it can only be in apposition with samaṇavādo. The commentarial gloss is mistaken and misleads those who follow it. Here I take sahadhammiko in the obvious (and dictionary) meaning of 'one who shares a dhamma' or a 'co-religionist', i.e. from our point of view, another Buddhist. And the idea that everything we experience is due to past action is a not the doctrine of a samaṇa who is Buddhist. Indeed, as above, it is the doctrine of a samaṇa who is a follower of Nigaṇṭha Nāgaputta, which is to say, a Jain. So sahadhammiko samaṇavādo must mean 'the samaṇa doctrine which is co-religionist' (as I understand it sahadhammiko specifically qualifies vādo). In more elegant English, "the doctrine of the samaṇa who is Buddhist."

One of the problems here is the switch from plural to singular. Bodhi, for example, translates as though the whole as plural. But I think in "na hoti paccattaṃ sahadhammiko samaṇavādo" we have a completely separate sentence in the singular with an implied 'it' as agent. And the obvious candidate for 'it' is pubbekata 'former action'. Thus the sentence means "[Former action] is not, on its own, the doctrine of the samaṇa who is Buddhist."

The problem here then, is saying, as the Jains do, that experience is based on former actions alone. Notwithstanding this, as I pointed out in my 2009 essay, many Tibetans insist on a pubbekatahetu doctrine. For example Tai Situpa has said :
"Now, this way, everything is karma. Only one thing that is not karma that is the Buddha nature and the enlightenment."
Or Ringu Tulku Rinpoche cited on the Rigpa Wiki:
Strictly speaking, therefore, from a Buddhist point of view, you cannot say that there is anything in our ordinary experience that is not somehow a result of our karma.
These same teachers argue that their view is not deterministic and that a particular calamity cannot be viewed as a punishment for some particular act. However, the Dalai Lama, for example (and I have heard Robina Courtin of the FPMT say the same), believes that the Chinese invasion of Tibet was because the present occupants of Tibet had accumulated bad karma in past lives (See this personal account of a discussion the DL for example).

I happen to think that the early Buddhist view is more coherent and less a result of blind faith in a supernatural force, but the really interesting thing is that once again we see that the metaphysics of a basic Buddhist doctrine changed. I say once again because I have already written about another way that karma changes from being inevitable, if mitigable, to being entirely avoidable through the use of mantras. Over the centuries the doctrine of karma has been modified to suit the needs of Buddhists. Perhaps the two changes are related. After all a hardening of views towards everything being a result of karma would probably make the ability to avoid the consequences of karma seem more attractive. Perhaps the change to everything being the result of karma required a let out so that it was not absolutely deterministic?

Niyama & Naturalism.

The commentarial teaching of the "fivefold restriction" (pañcavidha niyama) is sometimes cited as another example of how karma is not the only type of causation in our lives. This is mainly due to a modernist interpretation promulgated by Ledi Sayadaw and Mrs Rhys Davids in the 1930s, and Sangharakshita's development of their ideas in the 1960s and 2010s. (For the history of the idea see Dhīvan's essays and published article).

In Pāli, we don't have "five niyamas" but one fivefold niyama which is five applications of one principle of conditionality. The doctrine seems to aim at naturalising Buddhists ideas about three subjective or supernatural processes: cognition (citta); the functioning of karma; and the miracles associated with a buddha (dhamma-niyama). This done by likening them to observable processes in nature. So we have bījaniyama which describes rice seeds becoming rice plants and producing rice grains; and utuniyama, the fact that trees flower and fruit together in the appropriate season. These are limitations or restrictions (niyama) on how natural events unfold that can be observed by everyone in nature and they form the model of understanding unseen processes.

To some extent the Buddhist model of cognition is a result of introspection by yogis, but we can only ever observe our own cognitive process and never someone else's (at least this limitation clearly applies in Iron Age India). However Buddhists felt confident in providing a generalised description of cognition all the same. Similarly the process of karma is unseen and supernatural - it operates behind the scenes and cannot be understood in it's specifics. Karma, the idea that good and evil deeds have appropriate consequences for the appropriate person, is an article of faith. The various miracles accompanying the life history of a Buddha are also supernatural and by the time the niyama doctrine is composed they occurred centuries in the past.

The argument is that the limitations on the natural, seen processes of seeds and seasons, apply also to the unseen and supernatural. Clearly the analogy of karma with the process of planting seeds and reaping grain was one that appealed to the Indian mind, because a more literal version of this same analogy became the main Mahāyāna view of karma. The seeds were even provided with a storage pit in the from of the alayavijñāna.

The point about karma here is not that it is only one of many types of conditionality, but another example of the one type. It is a "natural" process characterised by inevitability, by results which are appropriate to the cause, and by ripening in due season.The idea that not everything is a result of karma is fine. As above it is definitely part of the early Buddhist view on karma. It's just that this is not the point of the niyama.


So the basic Buddhist teaching is that experiences are not all dictated by karma. A variety of causes and conditions including health, seasonal changes, and just plain luck can be invoked. The view that everything we experience being the result of karma is specifically criticised as deterministic. Such a view leaves us unable to make moral choices which is why early Buddhism rejects it. Buddhist soteriology requires that we have a measure of freedom to choose between right and wrong. However, according to this basic teaching, karma does determine which realm we will be born in. And one of the characteristics of the manussaloka or human realm, is that humans have sense objects, sense organs, and the potential for sense consciousness. Thus as human beings we experience vedanā every moment of our waking lives. And how we respond to vedanā is karma.


See also previous essays:
See also 'Recent Buddhist Theories of Free Will.' Journal of Buddhist Ethics. [This article, along with its predecessors, explores various attempts to define Buddhist morality as in/compatible with Western ideas of free will. On the whole I think the attempt tells us much more about Western Philosophy and its preoccupations that it does about Buddhism.]

Dhīvan's Essays on Karma

Norman, K R. (2006) The Group of Discourses (Sutta-Nipāta). PTS.