If we assume that macro-scale matter like we have in brains only acts in ways that are caused, then would not all actions in human behavior be a result of prior physical events beginning with development from conception? Alternatively, if you look at a brain as a information processing machine, or the mind as algorithmic/functional, then even ignoring material determinism, would not the deterministic character of algorithms lead to determinism in human behavior? Note this second point applies even if some sort of dualism were true. In the face of this logic, how have philosophers defended free will?
Most academic philosophers (around 60%, according to the PhilPapers survey) lean toward compatibilism: the view that determinism (which is what you are getting at, more or less) is compatible with free will.
Think of it this way: even if every decision is determined (in the "determinism" sense of the word), it's still useful to distinguish between what you are doing because someone is pointing a gun to your head, and what you are doing without any such external compulsion. Perhaps this distinction is what we aim at with our talk of free will. Getting clear on it is a worthwhile philosophical and psychological project.
Schiphol's answer is correct, in that you need to first say what kind of free will you're talking about. I'm going to answer your question assuming that you're talking about libertarian free will. It's the "real", pre-philosophical kind of free will that most people think that have. Ted Wrigley's answer is also correct, in that free will seems like something that we obviously have, so that any argument against its existence had better be very convincing.
The issue with arguments like yours, that try to show that free will "obviously" doesn't exist, is that they typically beg the question. You've assumed what you're trying to prove as a premise. You've assumed that given the same situation, antecedent causes will necessitate the same decision. Then you ask, "where can (libertarian) free will fit into the picture?". It can't once you've made that assumption, but people who argue for libertarian free will reject your assumption. There's also good reasons for believing that your assumption is false independently of the free will debate. In quantum mechanics, the collapse of the wave function is indeterministic, and possible outcomes can only be assigned a probability. You might be impressed by quantum mechanics and say that it may introduce randomness into the picture, but randomness is still not the same as a freely chosen action. A true random number generator doesn't choose its output. If our brains make choices randomly, like a sophisticated "random decision generator", then we're still not really any more free than a random number generator.
Your modified argument would then look something like this:
- Every event is either completely determined by antecedent causes, or else involves some random processes.
- If an action is completely determined by antecedent causes, it's not freely chosen.
- If an action involves some random processes, it's not freely chosen (like the random number generator above).
- Therefore, no action is freely chosen.
- If no action is freely chosen, then free will does not exist.
- Therefore, free will does not exist.
Libertarians would say that this argument still begs the question. Depending on what you mean by "random", either premise 1 or premise 3 exclude outright the possibility of free will. Libertarians would say that premise 1 or 3 is exactly what's at issue, so you cannot just assume them. There are good reasons for thinking that there are third possibilities between determinism and complete randomness, such as agent-causal theories which are neither deterministic nor random.
A good introduction to libertarian free will is Robert Kane's A Contemporary Introduction to Free Will, especially chapters 4, 5, and 6.
We believe in free will because — aside from a few people with particular psychological conditions — we experience ourselves as beings capable of making choices and exercising free will. When someone brews a cup of coffee and drinks it, they do not generally think:
That was not an action any agent took, but an event predestined since the beginning of time by the physics of cause and effect.
No, they generally think:
I wanted a cup of coffee, so I chose to brew one and drink it. Yum!
It takes a good bit of effort to convince oneself that free will does not exist, because that conclusion flies against our native experience of the world (which is, as you know, the root of empirical science). And in my (extensive) experience, the arguments that lead towards the rejection of free will are convoluted, abstract, and deeply suppositional, asserting all sorts of untested/untestable premises.
Look — and I'm going to toss this out here in the hopes it helps someone clarify their inner lives — the reason there is such a kerfuffle about the philosophical notion of free will has less to do with free will than it has to do with our attitudes towards religion. A typical point of religious doctrine in the Western (Abrahamic) world is that God gave humanity free will (in the form of a soul), and that we are thus obliged as humans to use our free will to choose right over wrong, moral over immoral, good over bad, etc. Some sects have even gone so far as to undercut the very concept of free will by demanding we 'must choose' to follow a rigidly defined set of 'approved' behaviors on pain of eternal torment. Religions equate being an agent (a being with free will) with being a moral agent (a being who uses free will to choose 'right' behaviors). A lot of people in the secular, technological world — a world of ever-growing possibilities — are frustrated or offended by that sanctimonious religious attitude; they chafe at the religious conception that we must restrict our behaviors to some (to their mind) arbitrary code of morality. But instead of rejecting or reimagining that code as an act of will — a deeply uncomfortable path to choose, since societies are built on systems of moral codes — they philosophically remove the concept of free will (of a soul, of agency) so that they can (paradoxically) do what they want to do by insisting that no one is there doing it.
We have a subjective experience of agency, and even a subjective experience of moral agency: e.g., evaluative self-reflections like guilt, shame, pride, confidence, angst, peace of mind... If we cannot uncouple that experiential sense from the sociopolitical squabbles over religious and secular ideologies, then we knot ourselves up in philosophical loops.
I would say that any definition of free will would require a conscious beeing and that any object considered to posess a free will would have to be considered conscious.
If we follow the opinion that consciousness is strongly emergent (which is a popular opinion) i would say this would allow for actual free will.
If we follow the opinion that consciousness is weakly emergent, actual free will could not exist.
As the question wether consciousness is strongly or weakly emergent is not decided upon and can probably never be answered, the existence of free will can be easily defended.
There are at the moment I'm writing this
five six answers to this question. I'm not going to give you yet a sixth seventh guy's opinion... I think that's a bit useless (and you'll probably get at least a seventh an eighth); rather, I'm going to make an attempt to specifically address your concerns here.
So let's start here.
1. The problem of determinism
All actions in human behaviour are a result of events that happened in the past, brain structure at birth and later self programming of the brain during life through experiences.
This is a premise; one does not have to accept it, and scientifically speaking, there's at least good reason to reject it... specifically, if I were to accept the Copenhagen interpretation of quantum mechanics, then this quite plainly is not in fact true, never mind five people's opinions on free will. The Born Rule in quantum mechanics suggests that the outcome of events is based on a pure probability, resulting from to be simplistic a type of normalized square of the wave function.
There are, however, deterministic interpretations of quantum mechanics, so let's just put this aside. I'll presume that the universe we live in is deterministic.
2. Predeterminism vs Fatalism
If you would upload a brain to a computer and run it it would make the same decisions as the human version.
I think you're confusing determinism with fatalism (no worries; this is incredibly popular, but still wrong).
Consider this nuance. So the universe is deterministic. Now, suppose we simulate the universe, including me, on a computer, and let's say we even do it "perfectly". Now let's consider a very simple task... I'm just going to pick a number between 1 and 10, and the computer will predict which number I pick. Now what happens if the computer predicts that I will pick 5, it predicts I will pick 5 before I pick the number, and I know what number it predicts I will pick? Presumably, the universe is predetermined for me to pick 5, the computer simulation says I will pick 5, the universe is deterministic, and therefore, I have no choice but to pick 5. This is fatalism; I'm forced to pick 5 because it's been "predetermined" by past events, and there's no other thing that can happen. All the computer did was predict what I was forced to do.
There's a flaw here; to see it, ignore the free will question; and toss me out of the picture. We'll put a robot in my place. We'll grant the robot exactly two capabilities: (a) the capability to understand what number our computer predicts, (b) the capability to execute a program. Now we turn the computer over to predict what number this robot will pick. For (b), all we need is a trivial program: "IF computer predicts robot will pick 5, THEN pick 6, ELSE pick 5".
Note that there's no RNG here, not even a pseudo-RNG. All this program does is pick a number different from what was predicted it would pick; it's a "spiter". And it's perfectly predictable what number this robot would pick in any situation. OTOH, it is impossible for your computer to predict what it would pick; there are two and only two scenarios. In scenario 1, your computer predicts that it picks 5, but it picks 6. In scenario 2, your computer doesn't predict it picks 5 (e.g., predicts it picks 6), but it picks 5. In neither scenario does your computer predict what it picks.
So it's in this sense that I propose you may be confusing determinism with fatalism; determinism specifies specific future actions as a result of prior states; fatalism specifies specific futures regardless of anything, and it's critical to point out that this includes regardless of prior states. This scenario and a spiter (like my robot) highlights the distinction. Both determinism and fatalism propose specific futures, but determinism's future is a result of something.
3. Principle of Alternate Possibilities
It is just logic, how can you defend free will?
There is a thing that we do, and it works a certain way. We make choices for some definition of "make choices", and we act with intent for some definition of "act with intent". The problem here is really about whether the thing we do when we make choices counts as "free will" choices, and whether the thing we do when we act with intent counts as exercising free will.
The argument you're presenting seems to assume the "principle of alternate possibilities" (PAP), the notion that in order to count as free will, alternate choices must be "possible". I'll be honest with you, I have big problems with PAP myself; the suggestion that a thing which will not in fact happen (alternate possibility) nevertheless has being in some sense escapes me... it seems to me that this is suggesting that a thing somehow both has and will never have being.
The problem here, though, is that not everyone subscribes to PAP. Harry Frankfurt, for example, argues against PAP with his Frankfurt cases. Essentially, he's arguing that free will does not actually require there to be alternate possibilities. This allows for the fact that the "thing that we do" when we make choices in a deterministic universe can be said to count as free will choices; which is the compatibilist position.
So the argument against free will in your post is not actually as solid as you make it out to be. Specifically, here is a highlight of its flaws:
- The universe may not in fact be deterministic
- Even if it were, determinism is different than fatalism; determinism specifies specific futures resulting from prior causes, not in spite of anything. One could argue (and many do) that free will is a question of what level of control we have over what happens... that's result of territory, which is in line with compatibilism
- Not everyone agrees that the specificity of the future is relevant to whether or not the kinds of choices we make are "free will" choices (i.e., agrees with PAP).
Here is an example of an argument against determinism and in favour of free will.
In a deterministic universe the past, the present, and the future are all fully determined. This means that the total information content of the universe remains constant over time. For example, Laplace's demon would know everything about the future state of the universe, so that the future universe could contain no new information.
This is contrary to our scientific theories, specifically thermodynamical law. The second law of thermodynamics states that information, in form form of entropy, increases over time.
Like all arguments on the subject, I'm sure that this argument is not without its problems.
Your position is one of the three sectarian views that lead to inaction that the Buddha declared 2500 years ago, in Anguttara Nikaya 3.61. That is the logical moral argument anyways..
The issue with the position that 'only the past affects the present and the future', is that it denies active choice in the present outright. But active choice (life) is shown at the quantum mechanical level, and even philosophically, you would have to deny the perceptions of individual frames of reference to say 'their choice does not affect their outcome'.
The past and the present affect the present and the future. Then individual action remains connected to the result of that action.
If you define free will as the generation of decisions from nothing then you have begged the question as to free will's nonexistence, if you are also presuming that every event is generated from some prior event.
When other respondents invoke quantum mechanics they are essentially attacking the latter point, i.e. determinism - that every event is generated from some determinable prior event. The whole argument of (at least the Copenhagen-style interpretation of) QM is that such prior events aren't determinable, but rather only probabilistic causes of the present state. According to Copenhagen QM there is no deterministic chain of cause and effect knowable ahead of time but rather probabilities (with any appearance of iron determinism only materialising on hindsight).
Personally while I think the QM counter is a red herring because we live in the macro Newtonian world rather than the micro QM one, the broader point about begging the question is valid. Why should we define free will as decisions generated from nothing? What justifies that characterisation? There is a kind of cognitive chauvinism going on - the suggestion that true decision-making, thought and therefore free will must have come from nothing. But nothing in the Newtonian world is like this; even if everything in the QM world potentially is.
The current top answer introduced me to compatiblism and maybe that's what I'm arguing for. My specific argument for compatiblism however rests on calling out as question-begging the defining of free will as decisions generated from nothing, while conceiving causality in terms that forbid anything from being generated from nothing.
I think the other answers cover the more philosophical and logical aspects of the argument quite well. I would like to consider in more depth a particular claim that you made:
If you would upload a brain to a computer and run it it would make the same decisions as the human version.
The problem with this claim is that it is utter nonsense. In fact, if you picked a program and a fixed set of inputs, and ran the program on those inputs over and over and over again, you would find that many times, the program would produce different outputs. As a professional software engineer for several decades, I can assure you this really happens in the "real world", and is not just a theoretical possibility.
You see, there are a lot of things going on here. The most obvious is that no device has perfect reliability. Sometimes, an electronic circuit will fail in a subtle way, producing the anomalous result. Other times, an unexpected event will occur, like a cosmic ray will flip one or more bits inside a computer. Barring stray radiation, we should consider the "normal" failure case.
Obviously, we cannot produce the exact same electronic circuit millions of times without fail. The instructions to etch a circuit don't change over the course of a production run, and yet, the output is not reliable. In any given wafer, dozens of chips have defects, some of which render it completely useless. So to start with, we don't even have deterministic manufacturing. Thus, the idea of building a device that perfectly reproduces the behavior of some other device is probabilistic. "Oh, but we just wait until we have a good device." Yes, that is the typical strategy. But as I pointed out above, a single device does not always produce the same designed behavior!
We appeal to the "logic" of computers, because what could be cleaner than representing everything in binary? Well, there is no binary logic inside a computer. Rather, there are wildly varying voltage levels and current flows, and we pretend that when the voltage is above a certain level, we have a "1", and when it's below, we have a "0". And in the cases where it's ambiguous, or we find a 1 when there should be a 0, we rely on error correction to get us back into the desired state. Unfortunately, error detection and correction is not foolproof, either. Most schemes only detect 1 or 2 bits of error per data word, and can only correct 1-bit errors. And yet, multi-bit errors are occurring all over the planet, even as we speak.
So, if we cannot count on the reliability of a precision-machined instrument like a modern CPU to behave in the intended fashion, how can we pretend that we can duplicate the behavior of something as messy as a brain? A microchip is sensitive to temperature, vibrations, and power conditioning. A brain is sensitive to temperature, vibration, pH, osmolality, and hundreds of various hormones and other biochemicals floating around the bloodstream. The brain is not a philosophical logic machine. It's a very messy control device for bodies that sometimes spits out something that we call "logic". In fact, much of the time, the brain is occupied just paying attention to what the body is doing. This means that the "uploaded brain" would also need to receive all of the same proprioceptive inputs as the brain-in-a-body did. And that means to give it identical inputs, you would need to physically model the entire body, throughout the entire history up until it made the decision you wish to predict. That isn't just infeasible, it is physically impossible.
You see, there is determinism, and then there is determinism. The kind of determinism you are suggesting is what we might call "reproducible determinism". It implies a kind of information superiority which allows you to take an arbitrary past state, and produce an arbitrary consequent state, without having to "experience" all of the intermediate states. A more accurate term would be reducible determinism, because this ability to predict is predicated on the existence of a kind of "closed form" representation of the system to which you can reduce it. An equation, if you will, that gives you S(t) given any S(k < t). And this is where everything falls apart.
It turns out that "reducible determinism" only describes a very tiny portion of our universe. It only describes things with simple, linear, separable effects. While, the vast majority of events and processes are chaotic. All that means is, given some state S(t), you are hard-pressed to say what S(t+k) looks like without manually iterating the dynamical rules which describe the system. There is no "shortcut to the future". Even worse, the sensitivity of the rules means that if you are off in your description of S(t) by even the tiniest amount, your final prediction can be wildly, laughably wrong.
You see, this "other" determinism, the one which actually permeates the universe, is what we could call "chaotic determinism". And it's a real b!+c#, because even though you may know all the rules of the game, you still have no idea what will happen until it happens.
Computers are subject to chaotic effects, too. Good luck trying to predict the precise voltage level in any part of a CPU during its operation. Now, we use feedback to limit the chaos to carefully controlled states. But what about the brain? All indications are that the brain actually relies on chaotic firing patterns to achieve its results (and calling into serious question a computer's ability to accurately emulate one). That means we need arbitrarily precise measurements in order to properly model its behavior, let alone demonstrate any kind of "determinism". And yet, any kind of measurement we make also disturbs the state of the brain, altering its behavior.
So even if "reducible determinism" does, in fact, describe something as messy and complicated as a brain, such determinism is functionally inaccessible to us, which then leads us to question the utility of such a concept. The only way to demonstrate, convincingly, that a brain actually is deterministic would be to make a prediction about its future state. Not a vague, hand-wavy prediction like: "There will be at least 10 million neural firings in the next minute", but more like: "The test subject will say the word 'cheese' in 3, 2, 1...now!" Of course, given what we know, it is effectively impossible to make such a demonstration, thus rendering the notion of deterministic brains an empirically undecidable proposition at best.
If you would upload a brain to a computer and run it it would make the same decisions as the human version.
I don't think this logic is true. There would be unseen causes that affect our decision making.
(Just for example) If there are three kinds of miseries (adhyātmika, adhibhautika and adhidaivika) they would be affecting our decision making also.
Often our decisions will change if we get enough/more time to think over something. Do you think a computer would also do this?
Why do people still believe in free will?
Even identical twins' decisions may not be the same always. Though there may be commonness, people are different and so there will be freewill. The power within each individual and individual differences are the reasons for this.
If we assume that macro-scale matter like we have in brains only acts in ways that are caused
This assumption is questionable. As Thomas Breuer has shown, from the point of view of any observer, the behavior of a system in which he is properly included, is not predictable, neither deterministically, nor probabilistically.
This means that such system produces events which are not caused by any distinguishable physical states in the past.
Because it is a religious Judeo-Christian concept. ref here Over 40% of the American people still believe that Adam and Eve were the first two people on earth and reject Darwinian evolution. Article on creationism here.
Genesis 2:16-17 New International Version (NIV) And the Lord God commanded the man, “You are free to eat from any tree in the garden; but you must not eat from the tree of the knowledge of good and evil*, for when you eat from it you will certainly die.”
(*) Good and evil is most likely a mistranslation for "all knowledge" ref here.
It is accepted that if you could go back in time there is only a certain level of probability (not 100%) that people would have the same chemical reactions in their brains therefore the same thoughts/emotions/actions. Keep in mind that there are no convincing results on self-aware computers for now, that is to say computers that know that they are computers. Also keep in mind that machines have been able to predict choices even before those choices hit a person's consciousness. Reference to neuroscience of free will here.
The very concept of free will has clashed not only with physics, neuroscience or computer sciences but with sociology, psychology and psychiatry. eg. "Homosexuality, in some countries is prosecuted because it is still believed that the individual has free-will to choose their sexual orientation". ref here gay marriage in the usa ref here
To defend the concept of free-will as told in the old testament nowadays would require an act of faith.