Hide table of contents

Prior more descriptive title: "Why effective Altruism relies on metaphysical Free Will, and why we should care to research it (if it exists)- Philipp Bienert & Ansgar Kamratowski'

Any case in which a subject matter is not fully expounded upon will be denoted with a "*", at the end of the given paragraph, with further information regarding future expected research developments in the given subject. In addition, you can find additional information on our current research and goals in the closing remarks and our contact information. 


A common conception within the scientifically minded community is the assumption of determinism. Determinism is the belief that any event is dependent on an antecedent cause. There was no alternative to what could have emerged from such a cause, this principle transfers to consciously acting agents, making Free Will an illusion because you could not have acted otherwise, you were determined to act. 

*Note that this discounts compatibilism, a philosophy that states that determinism is compatible with Free Will. We presumably will contend with the philosophy of compatibilism in our future paper. Still, for the sake of this essay, we will deny the possible truth of compatibilism for simplicity's sake.

Determinism is seemingly plausible considering the conservation of momentum, the laws of cause and effect, and other observations of the deterministic nature of macroscopic physical processes. 

*The irrelevance of quantum mechanics both to Free Will and in part to macroscopic, even though it calls into question physical determinism, will not be discussed here. 

In our universe, this would equate to a regress towards the beginning of our universe, which necessarily caused the ways things are in the present (Big Bang), with no possible "alternate" future realities. I.e. one and only one future of events could unfold. 

We were determined to join EA. We were determined to be un/successful in changing the world. "We" as such would not be the ultimate dictators of our wins and losses, and how we influence the world might be "up to us" causally speaking, but we could only influence it in one, and only one way. You might be going Vegan, but you have no choice going Vegan.

On the other hand, Free Will describes the agents' ability (conscious entities) to break free from these causal restrictions and make operational decisions, including alternatives. 

The following will dive into the logical nature of powerlessness in the case of determinism and how Free Will is necessary to our shared core concepts. 

Lastly, this post will focus on the metaphysical implications of "actual" Free Will and, most importantly, why we should care about its research and application to EA concepts.

Whilst many neurologists lean towards the position of determinism, the Free Will debate, as far as we can acknowledge, is NOT off the table. However, support for the certainty of determinism is heavily pioneered by neuroscientist Sam Harris in the public sphere.

Why Free Will should matter to utilitarians and any type of practical ethicists

Alternative possibilities and optimisation of ethical values

We will assume utilitarianism as a guiding ethical philosophy in the following description. 

For any other moral philosophy, you can exchange the optimisation as described:

The Kantian equivalent: optimisation of categorical imperative intentions. 

e.g. minimisation of lying

The Devine command theory equivalent: optimisation of the deity's desired moral preference.

 e.g. being more like Jesus

Effective Altruism (a broadly agreed-upon definition) is about the maximisation of wellbeing through careful scientific evaluation. This viewpoint presupposes that an optimisation(min/maximisation of utility) is possible.

From a mathematical perspective, optimisation is max/minimising a given output using variables in an equation by manipulating variable ranges. 

For instance, if we consider the consumption of animal products and how it relates to suffering, the variable is the amount we consume, the range is the amounts we can consume(upper and lower limits of consumption), and the suffering related to consumption, is the function that relates consumption to the level of created suffering.

Let us make an oversimplified assumption that there is a linear relationship between Kg of animal product consumption x and suffering s, expressed as the function 

Now at any given time 

n being a whole number, someone decides the amount of animal products they consume. 

let us furthermore only look at the time frame 

To optimise (minimise suffering), we have to minimise the values of the given t events, i.e. the consumption of animal products at any given time 

If we now presume determinism. Then we accept the premise that the event(and its associated utility) at t1 entails the event at t2(and its utility)and so on so that no alternative events could arise. 

In this case, any event is the brain state (which the person was determined to have) that causes its associated action—the amount of animal products eaten creates the given utility. 

The integer ascribing a value to suffering always was caused by the previous event(brain state), with the initial t1 brain state caused by an, for example, environmental stimulus.

Let us assume that the person goes linearly vegan from an initial consumption 10Kg at t1 to 1Kg at t10. 

Now, suppose previous events determine all brain states. In that case, all the amounts of APC(animal product consumption) are determined because if not, then the previous brain state could have caused the alternative consumption- the contraposition to the deterministic causal chain would be false. 

The false Contrapositon (indeterminism):

Not nine at t2 entails not ten at t1

Is false Because



(for example) Eight at t2 can entail ten at t1, and nine at t2 can entail ten at t1, is true.


More than one thing (alternative possibility) could occur from a prior event.


Determinism, therefore, entails that all APC values for the given times 

would be constant values as they would necessarily follow from the brain states at t1 and could not be different to what they are. All contrapositions would hold. 

The sum of events will give us one value if we look at the summation of utilities. The summation of utility(suffering) values is NOT minimisable and therefore not optimisable. 

The utility function in this example would be equal to the consumption function. 

The effective optimisation of wellbeing through time depends on evaluating possible future scenarios(plural) that necessitate optimisations based on possible reason-expected outcomes. 

A real-world example of this would be reason informed decisions on reducing Co2 emissions concerning future excess deaths. As you can see, projections show alternatives that decision variables of Co2eq can manipulate. Emissions. If such variables are just illusions of decisions, then the outcomes of our choices are, in truth, not maximisable or minimisable. Effective Altruism would become a hollow shell of people who are determined to care. 

For the optimisation to then produce an effective outcome in terms of temporal utility, the following must be true:

At a given time 

(up to t10, frame of reference), the agent, i.e. their brain states 

, can cause alternative values to occur at 

(we discount here for long-term future planning in the form of 

 As in the climate scenario). By that, breaking the initial contraposition.

Not an event(singular) after tn entails not tn

Is false Because


There is one or another event after tn. Therefore not an event(singular) does not imply no agent. 

"not x implies not y is false since not x could also imply y."

Adding the "Principle of agent variability":

The utility at 

is changeable by the agent at 

With this principle, the following would/could be obtained, assuming, for example, that the individual at t1 can decide between the value of 8Kg or 9kg at t1 so that 8Kg or 9Kg would obtain at t2(the additional assumption is that the consumption (if it obtains), would lead to the initial linear descent after t2).

The summation of temporal events could yield an integer sum between 55 and 54.

The function becomes minimisable/optimisable by that value through the freely willed action at t1 to t2.

The peculiarity of this example is that you have to assume that you are the person deciding or influencing the decision at t1 to t2. Assuming the agent's role or restricting their freedom with the freedom you just applied onto theirs (you being part of the same conceptual causal chain summation), by "making them" choose 8 rather than 9, with your freedom to choose.

You are the decider of the variable. 

The principle would state that only in instances in which an agent can metaphysically influence reality in terms of possible alternative states EA principles can apply because otherwise, no optimisation could occur. 

We should discount random chance(quantum mechanical events) that are outside of human volition since these are not subject to the rationale of optimising agents, except if these states are accessible to us in forming alternative outcomes.

We will spare the reader the explication of this principle in terms of notation. However, please let us know if there is increased demand or individual interest.

Importance of Free Will research and potential benefits

Suppose you accept the argumentation from the paragraph above. In that case, the following will attempt to convince you that the investigation of metaphysical Free Will, through neurological research, would benefit the human corpus in general and could shape how we as EA community allocate our resources effectively. 

Considering that the principle of alternative possibility is key to any optimisation, let us expand on how we can improve the utilitarian quality of such hypothetical Free Will circumstances. 

Free Will is entropy defying

Firstly one will have to understand how an agent could facilitate alternatives. 

*The theoretical framework of such possible facilitations through agents is currently drafted and improved by me and my colleague. 

A simplified notion of such facilitation is through the agents' interaction with their neurological substrate. Assuming that we agree that if sentient volition exists, it only affects the neural substrate, and consequently, it indirectly affects its surroundings(e.g. moving an arm through signaling). The agent can manipulate the firing of action potentials within the brain to facilitate alternate paths of thinking or even direct suppression or encouragement of physical movement, which was not to be expected without the agent's intervention. 

*Bienert and I hold the current viewpoint that such action facilitation could be described through neural network representations(weights, bias, inputs and outputs). The agent manipulates or introduces its weights and biases into the neural network to manipulate outcomes. We will not explicate this here further.

Though will provide a conceptually simplified illustration in the following

Illustration of synaptic transmission of stimuli

The agent, as such, would be able to significantly(causing a neural firing or counteracting a neural firing) manipulate synaptic transmission at any point within the axon through the manipulation of action potentials or manipulation of graded potentials before the transmission of stimuli through the release, of neurotransmitters, the opening of ion channels, or by directly influencing graded potentials. -enabling or countering movements through manipulating neurons. 

Such a manipulation, we assume, can occur without violating energy conservation; however, we contend that manipulation that is subject to freedom of the will and actual alternative possibilities arising from agents necessitates violating the law of conservation of energy.

*For a more detailed analysis of the view, keep yourself updated about our research by reaching out to my email.

An agent has to use their free volition to abstain or indulge in manipulating reality by causing alternative states. Since any state change requires energy to occur, and the sentience cannot be entirely dependent upon an external energy source, it must introduce energy of its own into the system. 

If it were only to transform energy to cause an alternate state, the redirection of such energy through the conscious medium would require itself energy. When it uses exterior energy to facilitate redirection, it will subject such energy(if the agent is free) to redirection to facilitate alternative states, ad infinitum. Therefore, such alternative facilitation energy must emerge from the agent and not be "borrowed" from its surroundings. 

Accepting this premise, we can presume that if Free Will exists, it functions as an entropy defying and energy introducing agent. Entropy is defying because it could reduce the disorder of a system and use more mechanical energy in a given time frame than physically possible. 

Such inducing of energy would still need to be subject to alternatives in that the agent can refrain from inducing such energy if we were to call it "Free Will".

Optimising Free Will

The ability to decide between at least one alternative is the minimal requirement for freedom. However, there is a crucial element, one we decided to call the Bayesian foreknowledge or foreknowledge factor.

*The planned paper will dive deeper into the exact hypothesised models of freedom, including coercion from previous stimuli. 

Equipping individuals with the approximate knowledge of what their actions will entail is a powerful component of Free Will. Let us proceed with two examples to clarify this concept. 

The Button

Imagine an individual deciding between two buttons. That individual has the total freedom(equal chance) of pressing either button. They press button 1. 

They just killed two toddlers. Nothing would have happened if the person had decided to push button 2. So clearly, it was not the individuals' fault since they were pushing buttons. 

In that respect, we can recognise that true freedom is when we have the minimum foreknowledge, i.e. Bayesian reasonable foreknowledge/confidence(rarely absolute confidence), of what at least one precedent event entails, both morally and epistemic. 

Free Will Multiplier 

Suppose a person has the foreknowledge that abstaining from a desirable drug(as a free decision) would lead to more freedom due to the absence of future addiction. In that case, that person could multiply their liberty through the Bayesian foreknowledge of unfolding events. Hence a Free Will multiplier.

Using both the principle of the event foreknowledge and the Free Will multiplier, we can directly investigate such processes with the scientific method we all praise within EA. 

Foreknowledge of circumstances can inform proper, freely willed effective decisions, a powerful argument for EA education of principles, methodology, acquired evidence, and the incentive to maximise given freedoms if they lead to a higher likelihood of positive outcomes. 

Societies, individuals, and organisms can make use of the additional energy supply through conscious agency if it is not at large counterfactual(opposed)to the goals of the organism or society. They gain a significant advantage over those who do not foster or utilise freedom in that respect. 

Western colloquial analogies to such beneficial freedoms would be creativity(making use of ones' sentient creative processes), individualism (acting under ones' own volition), political libertarianism(free markets, and higher individual energy investments into their work). 

However, to make assessments of such beneficial freedoms, we would have to neurologically analyse which types of people, conditioning(societal, social, substance-induced), maximise the liberty to introduce energy, which is not counterfactual to the suggested goal orientation. 

The Utilitarian would support any agent who participates in utilitarian goals with their freedom(energetic investment)counterfactual or supportive towards neural correlates or manifesting lifestyles(work and resource consumption). Conversely, they would aim to reduce or counteract freedoms contrary to the wellbeing of the given individual or others. 

Therefore, if we accept EA goals as our guiding principles, we should invest in neurological research and application of outlining the limits of Free Will and its possible manipulation regarding utilitarian goals. This knowledge could aid us from an individual level, increasing freedoms and decreasing freedom, leading to heightened utility, improving education(and its outputs for the individual and society) up to government and non-Profit policies. 

Whether freedom is generally contrary, neutral, or positive for EA goal orientation at large, would unfold in the process of its scientific exploration. 

Research in the area could reveal if the successive erosion, build-up, or reformation of freedom is desirable.

A sneak peek at current research and our interpretations of the findings

Readiness potential

There has been extensive research in the area, starting with the experiments by Libet in the 1980s concerning the readiness potential. For example, in one contemporary study, decisions were predicted up to 10s before entering awareness. 

Unconscious determinants of free decisions in the human brain - PubMed (nih.gov)

We found that the outcome of a decision can be encoded in brain activity of the prefrontal and parietal cortex up to 10 s before it enters awareness. This delay presumably reflects the operation of a network of high-level control areas that begin to prepare an upcoming decision long before it enters awareness

Whilst these experiments indicate reduced freedoms and coercive mechanisms, they cannot discount freedom based on a statistically significant prediction if this prediction cannot fully predict the agents' actions. An agent, for example, could have an unconscious neurological disposition to push one or the other button that is reflected in the p-value of that prediction. The agent is coerced but not absolutely coerced to perform a particular action so that the ascribed 60percent decoding accuracy would only reflect an observable significant, but not absolute coercion. Subsequently, the agent is partly "robbed" of its freedom due to presumably physical coercion, instrumentally determined through observation. The free volition would vanish at the point of an almost 100% predictive decoding accuracy. In our interpretation, the findings suggest either slightly physically coerced, impeded, circumstantially absent free action or that current decoding accuracies are confined to instrumental limitations of absolute predictive capacities(if determinism is true).

The vertical red line shows the earliest time when the subjects made their decision.

Microsoft Word - NN-BC20093C_SupplementaryMaterial_rev1.doc (springer.com)

Furthermore, there has been evidence of agents being able to veto movement decoding.

Beyond the point of no return: Last-minute changes in human motor performance - PMC (nih.gov)

Having identified this point of no return, Schultze-Kraft et al. (1) argue against the idea that the onset of early movement-related brain signals, e.g., the Bereitschaftspotential (BP) (English: readiness potential) (23) triggers a causal chain that cannot be interrupted. Such an idea, promoted by the interpretation of an experiment performed by Libet et al. (4), was used to deny the existence of a "free will." We would like to congratulate Schultze-Kraft et al. for their significant contribution toward clearing up this false doctrine.

Split-brains consciousness- unknown mechanisms

Other peculiar and intriguing findings are current Split-brain studies, where individuals seemed to perceive tactile sensory information between hemispheres without them being connected via the corpus callosum. 

Whilst we should not rush to conclusions, we remark that these findings might indicate conscious mediation of neural information between brain hemispheres(possible Free Will or energy inducing?) or that the person remains a unitary agent. 

Unified tactile detection and localisation in split-brain patients - PubMed (nih.gov)

in 'split-brain' patients, the corpus callosum has been surgically severed to alleviate medically intractable, severe epilepsy. The classic claim is that after removal of the corpus callosum an object presented in the right visual field will be identified correctly verbally and with the right hand but not with the left hand. When the object is presented in the left visual field the patient verbally states that he saw nothing but nevertheless identifies it accurately with the left hand. This interaction suggests that perception, recognition and responding are separated in the two isolated hemispheres. However, there is now accumulating evidence that this interaction is not absolute. Recently, we (Pinto et al., 2017) showed that accurate detection and location of stimuli anywhere in the visual field could be performed with both hands. In this study, we explored detection and localisation of tactile stimulation on the body. In line with our previous results, we observed that split-brain patients can signal detection and localisation with either hand anywhere on the body (be it the arm or the leg) but they remain unable to match positions touched on both arms or legs simultaneously. These results add to the evidence suggesting that the effects of removal of the corpus callosum may be less severe than sometimes claimed.

Since there is a large body of science concerning theories of consciousness, split-brain patients, and other neurological studies, we are unequipped to make any judgments before analysing such existing theories and empirical data. 


Closing remarks

We believe that Free Will research, aside from its apparent disconnection to everyday Altruism due to its metaphysical complexity, should gain more attention in the future. 

Since there is still a great divide in the philosophical and scientific community, we propose to:

1. Outline a coherent theoretical framework of Free Will and determinism, analyse current research findings and methodologies, and suggest rules for future research methodologies and interpretation of data. 

*The aforementioned is what me and Bienert are currently striving to accomplish, currently drafting our first paper concerning the subject. 

A more detailed account of human free agency, the ethics of Free Will, and Free Will study designs and interpretation are planned to be released by Bienert and me later this year. 

2. Resurface discussions of Free Will to the public sphere and question the predominant beliefs of "scientifically-minded public intellectuals" in determinism on rational and scientific grounds. 

Whilst still lacking the intellectual wit and video production skills, I started to publish "simplified" video content to counter predominant deterministic beliefs on the YT platform "Celestial Altrunaut", aiming to publish more EA related content in the future. 

Proceed in advancing the study of neurology concerning Free Will and consciousness.

If you intend to get in touch concerning inquiries about the paper or the essay in question, collaboration(coauthoring) interests, or financial support for our work, please contact:


Link to my channel: https://www.youtube.com/channel/UCU24RnaLHI_4v0U06XkSoHg

Link to my Patreon: https://www.patreon.com/Ansgar_Kamratowski


Bienert and I additionally co-run a German Animal Rights related channel, "Jung Brutal Vegan".

Link to our channel: https://www.youtube.com/channel/UC85qo7ShfHi7ncNIDO74-LQ

Link to our Patreon: https://www.patreon.com/jungbrutalvegan


If you have any ideas on how to acquire research funding concerning this subject, please contact us. Currently, the research is confined to the free time we can offer.

For those interested in my(Kamratowski) current position concerning Free Will (as people frequently ask):

Whilst there is no conclusive research that indicates us having freedom of the will, I find that adopting an agnostic position towards Free Will reaps the benefits of "both sides" by not underestimating causal influences of particular events—at the same time not falling victim to the over-interpretation of causal necessities, when it comes to one's personal life and decisions, as well as considering broader ethical questions and decision making. 

The existence of Free Will would still be subject to addiction, institutional and geographic inequality in different countries, and majority and minority groups.

At least concerning Free Will beliefs, there is evidence to support that belief in Free Will can increase the risk of correspondence bias. 

In which people underestimate environmental influences of a person's behaviour or in/action and attribute it to the person itself, which could be indicative of typical inequality justifying attributions. 

"They could be richer if they wanted to.", "These people are deliberately lazy.", "There is x person who made it out of poverty-poverty is not a thing that is caused by the environment and opportunity.".


All studies demonstrate a positive relationship between the strength of the belief in free will and the correspondence bias. Moreover, in two experimental studies, we showed that weakening participants' belief in free will leads to a reduction of the correspondence bias. Finally, the last study demonstrates that believing in free will predicts prescribed punishment and reward behavior, and that this relation is mediated by the correspondence bias.

On current grounds, I deny compatibilism if the compatibilist notion of determinism is defined both as mental and physical determinism. However, I believe that compatibilists' theories better understand actual libertarian freedom than their determinist or libertarians counterparts. 

I would fall into the category of an agnostic-incompatibilist. 





More posts like this

Sorted by Click to highlight new comments since:

Trying to prove the matter of metaphysical free will in either direction via neuroscience would likely be either practically impossible, or at least very expensive and drawn out. True, perhaps the introduction of artificial general intelligence (AGI) could answer the question satisfactorily. But at that point, such matters as minimising sentient suffering may be as well within short reach. While I may agree that obtaining more data on human behavioural causation is a worthwhile end for other reasons, so far I cannot see the rationality of the proposed or implied need for this data in order to enable or facilitate "freely willed optimisation".

The arguments presented seem to suggest that there must be an agent, of perhaps moral responsibility, to own the choice for, or results of, optimisation.

  • Was there an argument given as to why optimisation requires, or even would benefit from, personal ownership?
  • Do complex systems not have natural tendencies toward certain states?
  • Might the personal narrative be an after-effect, or add-on, to otherwise natural tendencies of the system?

From the perspective of embedded agency, the agent exists as a conceptually convenient cut-out from its broader system. This enables heuristically modelling and selecting between epistemic possibilities, even if there is only one actual possibility. Imposing personal moral responsibility upon that agent is one way of modelling the environment. It has some downsides, however, such as placing the burden of correction on the agent. This may work for trivial behavioural matters, but it breaks down when the causal factors involved are difficult to reach or modify for the agent.

Generally speaking, optimisation occurs naturally and automatically upon updating the world model or self-model to a more overall accurate and efficacious state. That is, when an agent encounters information that brings causal insight about important matters, the agent's behaviour optimises automatically as a result. Such information need not be owned by anyone. I, for example, need not be said to own these words. They, after all, are the result of an unbearably long and complex web of events. Per the butterfly effect, if anything were otherwise in the distant past, everything personally meaningful, including one's genes and conditioning, simply would not be as they are. Indeed, without the assumption of an incorporeal agent, technically the existing agent ceases to be at every instant. The seeming subjective "person" that goes on could easily be explained as the comparison of a mental object of memory with its ever modified self. Obviously the "same" essence appears to persist when comparing the modified memory to itself, perhaps separated merely by iterations of the perceptual memory loop.

On the topic of felt "energy" -- as would drive the sensation of self-determination -- one might note that ego alone is sufficient to provide such "energy", or impetus. This makes perfect sense if we recall that ego drive is the social-symbolic aspect of self-preservation, aka. fear. Hence, ego drive, as amplified, for example, in narcissism, can indeed heighten motivation. But this effect comes at a huge moral cost, in that fear triggers shortening of one's causal inference chains, making one both short-sighted, and self-interested. Thus, if we should intentionally or inadvertently increase ego drive, we may well create more suffering than we relieve. So pushing ideas of metaphysical free will without proper evidence and specificity could easily have net negative societal consequences.

Curated and popular this week
Relevant opportunities