I. Introduction

Am I freely writing this submission? Are you freely reading this entry? Or are we both on the whims of countless factors beyond our control? As it stands, free will's existence is one of philosophy's perennial problems, and undoubtedly a greatly consequential one. Science has thus far been man's entrusted guide to our toughest questions. Will it do the same for free will?

First, what exactly is free will? Some say it is the ability to act unimpeded externally - Alfred Mele terms this "moderate free will." But others demand more: we only have free will if we are the genuine source of our actions and could have done otherwise. As you will see in the coming paragraphs, how high one sets the bar for free will's definition impacts my answer. I will state now that if we are talking about moderate free will, science has yet to disprove it, but it remains an open question regarding its demanding counterpart.

II. Modest Free Will

Benjamin Libet rocked the philosophy world with his landmark 1983 experiment, supposedly having disproved free will's existence. He had participants flex their wrists while sitting in front of a rapidly rotating clock—all while monitoring their brain activity with EEG electrodes. Whenever participants flexed, the clock halted, and they had to note the precise moment they were consciously aware of their decision to move. After repeated trials, Libet found that unconscious brain activity — represented by EEG signals termed the 'readiness potential' — consistently began about 550 milliseconds before a flexing action, but participants were only consciously aware of their decision to move about 200 milliseconds before. He concluded that the impulse to flex, therefore, did not originate from us. Instead of our conscious intentions dictating our actions, preceding unconscious causation was the real author. Like puppets controlled by an unknown puppeteer, we lack free will.

Although persuasive, Libet's results had notable flaws. To start, what Readiness Potential (RP) means remains debated. In his experiment, Libet assumed that RP represented the unconscious origin of a decision, but this is highly contentious. Schurger et al., 2012 found that RP represents the rising portion of a regular pattern of haphazard ebb and flow of neural activity — neural noise. They postulated that, prior to movement initiation, a neural threshold is crossed when RP occurs, which tips the scale for action initiation when there is nothing else to base a choice on. This saves us from endless indecision, especially for arbitrary tasks. In short, RP is simply the rising arc of neural noise and only appears to be a non-random pattern because of how Libet measured brain activity — an artifactual error. Furthermore, Alexander et al., 2014 found that RP may instead reflect domain-general activities, like the build-up of anticipation, rather than the unconscious genesis of a decision.

But a more fatal flaw in Libet's argument is his implicit assumption that because some of our choices originate unconsciously, all our choices must too. Right away, an assumption like this seems reckless. His participants were tasked to arbitrarily flex their wrists, worlds apart from the complex everyday choices we make that require conscious deliberation, like creating a grocery list or writing a philosophy essay. Haggard et al., 1999 specifically sought to test for conscious deliberation in Libet-styled experiments—go-signal tests. Participants were tasked to respond to a signal with a predesignated action, like clicking a mouse button when a tone arises. This meant that instead of responding to spontaneous urges, participants had to rationalize instructions and consciously anticipate a signal. In this case, the results showed that the intention to act occurred 231 milliseconds prior to action — closely aligning with Libet's findings on when participants became consciously aware of their decision. If this is true, Libet's argument is defeated. Recall that his argument primarily rests on the disparity between the original (unconscious) conception of intention, about 550 milliseconds before an action, and when participants became consciously aware of the decision, about 200 milliseconds before. If it turns out that no such disparity exists, at least for actions that require conscious thinking, Libet's argument fails. Even if we take Libet's results at face-value, it can only be concluded that only some actions originate unconsciously, not all.

III. Its Demanding Sister

Now, I will address the kind of free will that lies at the heart of current philosophical debates — demanding free will. Those who believe in demanding free will, and contend that it is incompatible with causal determinism, are called libertarians (unrelated to libertarianism in political philosophy). Without getting lost in technical complexities, having libertarian free will is generally defined as being the genuine source of one's actions which entails having had the ability to have done otherwise (Kane, 2005; Pereboom, Fischer, and Kane, 2007). Suppose Jones decided coffee over tea for breakfast, if we rewound the clock back to the exact moment Jones decided, assuming the exact antecedent universe, Jones could have genuinely chosen tea instead. His choice was not causally determined by his past — it was indeterminate up till his moment of decision. An essential requisite to libertarian free will, thus, would be the existence of indeterminacy in the universe and, more specifically, that indeterminacy resides locally in brain activity. Science has yet to provide unequivocal proof for this, but there have been promising findings.

The most oft-cited evidence favoring libertarian free will is the existence of randomized spike timings, which makes synaptic transmission in neuron-to-neuron communication probabilistic, not deterministic, in nature (Dayan and Abbott, 2001). In brief, this happens because the Brownian motion (irregular movement) of neurotransmitter particles causes their arrival onto cell channels to be random. This affects the opening and closing of cell channels and, consequently, causes the timing of neuron firing to be random too. Another occurrence supporting indeterminate neural activity is the random behavior of magnesium ion cofactors in NMDA channels; though, as Dr Peter Ulric-Tse points out in The Neural Basis of Free Will, just a single instance of indeterminacy in brain activity is necessary for libertarian free will. And, as Kane notes, indeterminacy need not function as a macro-level disturbance in neural processing for libertarian free will. Just a "sprinkle" of indeterminacy acting as a hindrance in decision-making suffices.

Still, it is important to remember that even if neuroscience treats these processes probabilistically, it does not necessarily mean they are genuinely indeterministic (Balaguer, 2010). What I mean is that for libertarian free will to hold, we must ascertain that indeterminacy exists in the brain ontologically (in physical reality) and not merely epistemologically (how it appears). To know if these processes (synaptic transmission etc.) are "truly random" is an ongoing physics question—concerning particle behavior in quantum mechanics—rather than a neuroscience question.

IV. Conclusion

Indeterminacy's existence remains hotly contested. We do not know for certain if the brain operates indeterminately, but neither if it operates determinately, as many have assumed. Libertarianism is oft-overlooked in contemporary philosophy discussions for its stringent empirical requirement, but I contend that until science provides a definite answer, we ought to have an open mind.

So, has science disproven free will? For modest free will, no. For libertarian free will, I say the jury's still out. For all we know it could be widespread or may very well be an illusion.

Only time will tell.