This is doctor logic's Typepad Profile.
Join Typepad and start following doctor logic's activity
Join Now!
Already a member? Sign In
doctor logic
Recent Activity
Joseph, I think we're making progress. You dispute my estimation of the numbers (e.g., you think my estimates or classifications are subjective), but you seem to agree with the basic structure of the probability argument. "You're making what is at absolute best a subjective judgment call ("Evolution sure looks unguided!" - Sorry, but to me it looks guided.)," No, it's not subjective. When we consider an unguided process of natural selection combined with random genetic modification, we get common descent, common architecture and common construction. Evolution doesn't subjectively look like this model, it *objectively* looks like this model. "Third, we have definite and clear examples of (limited, human) agents using evolution and evolutionary design principles towards intentional ends." Sure. Suppose I want to design a better radio receiver using a genetic algorithm. Well, for starters, I'll create an environment with a fitness function that favors receivers with higher S/N ratio, sensitivity, etc. Then I'll run my evolution sim and get a design for my receiver. Then, I'll take that design and pair it with a power supply that I designed using traditional methods. Also, I won't populate my sim with independently evolving predators that eat my receivers. As for the relative size of the spaces, consider this. At every point, I have options. I can switch from an evolutionary algorithm to a traditional design algorithm. I can re-implement an evolved algorithm in a different substrate. I can also add manufacturing facilities to my world. I can stock the world with completely new, traditionally designed entities. I can eliminate species at will, and render others invulnerable. That's a lot of different options available to a designer that are not available to unguided evolution at EVERY step. At every step I can switch from evolving to manufacturing, switch from evolving one kind of species to another, and switch from one substrate or technology to another, etc. I can also choose to change my goals from time to time. What does that mean in biological history? The designer could decide that the platypus will be implemented in silicon nanotech instead of organic molecules using non-evolved components. Or the designer might decide that hedgehogs will not breed but be manufactured at thousands of miniature factories across the globe. There is no evidence for such design choices at all. In other words, look at the issue from an atomic point of view. At each step in a naturalistic evolutionary process, only certain things are allowed. With design, virtually anything is allowed *at each step*. This isn't subjective, it's objective. The number of possible worlds consistent with design becomes vastly bigger at every moment in time relative to the number of possible worlds consistent with evolution. In theistic language, a god can intervene (or not) at every instant in time, and veer away from what appears to be unguided evolution. At every instant in time, the number of possible theistic worlds is vastly greater than the number of possible naturalistic worlds. Therefore, the total number of worlds consistent with theism is vastly greater than the number of worlds consistent with naturalism. And the observed world is one of the worlds consistent with naturalism.
Joseph, You say "Calling evolution the "natural alternative" will not fly when the point is being made that the "natural alternative" *is* (and for a Designer, 'could be') a "designed alternative"." This is *precisely* analogous to the sorted versus shuffled analogy. We cannot tell whether a deck was sorted into order or shuffled and happened to fall into order. Right? I mean, all you know is that the deck before you is sorted. Does that mean that we can't be confident that it was sorted versus shuffled? No. We can be almost (but not quite) certain that the deck was sorted. Why? Because "apparently-sorted" accounts for a teeny tiny number of outcomes of shuffling, whereas "apparently-sorted" accounts for 100% of the outcomes from sorting. Likewise, when I look at the species before me, I cannot distinguish the case (i) where the life was evolved unguided and (ii) where an agent designed us using evolution. However, "apparently-unguided evolution" is only one teeny tiny subset of "designed". And, "apparently-unguided evolution" accounts for roughly 100% of unguided evolution. Maybe you're getting hung up on the term science. You'll note that I don't use the term in the argument above. My argument relies on rationality, i.e., deductive and inductive inference, theorizing and probability. Call it philosophy, if you like. It won't change the argument at all.
Brodie, "The reliability of science to judge whether a deck of cards has been shuffled or sorted presupposes the intentions of an agent. Without previous knowledge of how decks of cards are typically ordered by their creators, we would be in absolutely no position to discern whether or not any particular deck showed signs of intentionality." I think you're missing my analogy. Intentionality isn't part of the card deck analogy. The card deck simply illustrates reasoning with probability. The way I am applying it is as follows. In model 1, there is an agent with some unspecified intent. In model 2, there is no agent , but there is the prediction of common descent, etc. Clearly, the space of a zillion different intents in model 1 includes results consistent with model 2. However, there are a zillion different intents in model 1, and we have no reason a priori to suppose that "unspecified intent" should be weighted towards producing common descent versus the other zillion possibilities. Again, there's no way to tell the difference between unguided evolution and an agent who intended to create an apparently unguided evolutionary pathway. But this is irrelevant because it has not been established that we should fine tune our intents to creating an apparently unguided evolutionary history. This is precisely analogous to the shuffling case. Yes, shuffling could produce an ordered deck, and we would be none the wiser. However, it is incredibly unlikely this would happen unless we have reason to say why shuffling (in this particular case) would lead to a sorted deck.
Joseph, "I may as well say 'I can imagine a googleplex number of alternate ways an undesigned world would look, and there's no a priori reason it should look like this, ergo the world is designed.' " The fact that you are saying this means you have not understood any aspect of my argument. There are at least two important points you've missed. 1) How many natural alternatives are there versus designed alternatives? As I explained, for every unguided evolutionary pathway, there are many many more design alternatives. The evolutionary possibilities are a tiny (*really* tiny, 1:N) subset of the possibilities open to a designer. That means, weighting members of each class equally, unguided evolution is far more probable once we know that evolution occurred. This means we have some 1 in N likelihood that the world was designed. 2) Yes, there are countless ways that unguided evolution might have unfolded. Why believe any particular one? Wouldn't that be fine-tuning? Yes. But, as I explained, fine-tuning is okay as long as you get a payoff for the fine-tuning in the form of predictions. Evolutionary biology does just that. 3) In principle, a design theory could overcome this hurdle by making a prediction that has M:1 leverage, M>N. As we know, "design theorists" don't want to talk about the designer, so they have no hope of ever making any predictions. "A Designer on the level we're talking about? It's simply out of the question." Argumentum ad simply out of the question? :)
Joseph, I find it a very queer notion that science cannot be applied to agents. A person's character consists of regularity about intentions, powers, foresight, etc. We don't normally apply science to personal character, but it is applicable because we can construct statistical models of an individual. That's why we can say that a person is acting out of character, or counter to past goals. That's why stories make sense when the hero predicts the next move of the villain based on the villain's goals and personality. Personality is predictive. "Not when science is in no position to know what the intentions of the designer are, what degree/power of intervention/foresight/etc said designer may have had available and how much was exercised." In the case where there are multiple possibilities, we are rational to assign those possibilities equal weight. That means considering all possible intentions, powers, foresights, etc. to have equal weight. For some designers, it is likely that they would evolve life on Earth, but this misses the point. I shouldn't assume the designer is someone who would do this, when there are so many other options that are, a priori, equally likely. Why shouldn't I equally well assume that sophisticated designers would never ever evolve stuff? I have no reason to prefer one to the other, so I weight them equally in my consideration. If we had a successful, predictive theory that explains why we ought to give more weight to a designer whose designs are indistinguishable from unguided evolution, then we could say that finding something that looks unguided is not decisive. Of course, we lack even a theoretical model like that, let alone a predictive or successfully predictive model. "The truth of common descent and the truth of evolutionary mechanisms... does nada, nothing, zero, zip to demonstrate a lack of teleology, intention, mind, or design at work in natural history." You're just stating your conclusion. And my argument explains why. If teleology were involved, as with the cards, it is, a priori, far more likely the world would look different. Do you have a counterargument or not? I don't need to make this problem artificially sterile to reach my conclusion. I just have to treat like as like. There's no rational reason to prefer each evolver god to the relative googol of manufacturer gods and hybrid manufacturer-evolver gods. So, I am rationally justified in concluding that life was not designed. This is argument is subject to revision, should reason for a different statistical weighting come along. However, rationality does not wait for different statistical weights. Rationality must proceed based on the simplest weighting consistent with the data. We don't assume it is equally likely that the Sun will rise as not rise tomorrow. Why? How do we know the distribution isn't flat until next Thursday, where the distribution becomes zero and the Sun doesn't rise? It could be a peculiar distribution. However, it is not rational to assume a peculiar distribution when a non-peculiar one will do. "even while noting that the power of 'natural selection' is under dispute even by orthodox evolutionary biologists " No, it isn't. Not in any relevant way. But, hey, if you are making a case that evolutionary biologists think the power of NS is a threat to the theory of evolution, please feel free to include references.
Joseph, I don't think I'm making the assumption you're accusing me of. You're saying that it's not impossible for a designer to use evolution to accomplish a design goal. I totally agree. Dealing 5 cards in order off the top of the deck doesn't prove with absolute certainty the deck wasn't shuffled. Maybe the preparer did indeed shuffled the cards. Science cannot tell the difference between a shuffled deck and a sorted one when the card sequence is identical. However, science can say whether or not it is *likely* to have been shuffled. Likewise, we can come up with a scenario in which a designer uses evolution as the mechanism of design. However, we have to create a fine-tuned model to do this. Here's what I mean by fine-tuning. By analogy, when the shuffling proponent says that the deck was shuffled, the proponent is arbitrarily fine-tuning his shuffling model to refer only to those cases of his shuffling model in which the first 5 cards on the shuffled deck turn out to be 2, 3, 4, 5, 6 of clubs. The proponent is right that the deck is possibly shuffled. However, this fine-tuning of the solution space doesn't escape the probability argument at all. Just because the deck is possibly shuffled, and we can't tell a shuffled deck from a non-shuffled one (when the card orders are identical), doesn't mean we don't know (with high probability) the deck wasn't shuffled. We do know the deck is very probably sorted. Fine-tuning can beat the odds when the fine-tuned model successfully makes predictions that are unlikely to be true a priori. Evolution has done that because descent (and even more so, common descent) is not likely to be true a priori. In contrast, design advocates are happy to fine-tune their model of a designer, but they can't make any predictions to justify the fine-tuning.
Bob, I disagree. Although my argument would not apply to say, the ultimate laws of physics, it does apply to anything and everything that leads to us specifically. In other words, we are accidents. And I think the argument goes way beyond saying that design is superfluous. That's like saying that, in the card case, the lucky shuffler is superfluous. We would not say that the Sorter theory merely makes the Shuffler theory superfluous. We would say that the Shuffler theory is all but ruled out.
Joseph, "If science doesn't discover a lack of teleology, but rather some adherents presuppose it, it changes the discussion dramatically." There is no teleology, and we can be virtually certain of it. Here's why. Suppose I have a deck of cards, and I tell you that the deck is either shuffled or sorted by suit and rank. I start dealing cards off the top of the deck. 2, 3 ,4, 5, 6 of clubs. Is the deck sorted or shuffled? It would be irrational to conclude that the deck is shuffled. It is possible (one part in 310 million, I think), but incredibly unlikely. More generally, when we're considering two possible scenarios, and there are many outcomes consistent with one scenario, and only the observed outcome consistent with the other, we're rationally going to side with the latter. And there are many (trillions at least) ways of designing life instead of evolving it. First of all, a designer doesn't need descent, let alone common descent. We don't breed cars, we manufacture them. Second, a designer can use completely different architectures and raw materials from one design to another. Rabbits could be plastic, and deer could be nuclear powered, but they're not. They're blood relatives of ours, sharing most of our DNA. Finally, the utility of life is, as far as we can tell, survival, which is the one utility consistent with evolution. This means that Darwin's theory of evolution for all practical purposes rules out teleology. And it's not just a wishy-washy philosophical argument. It's an argument that has an extremely high probability of being true. Theistic evolution is implausible.
When we humans design, we go through a process that has a lot in common with evolution. First, we imagine the desired goal. Then, we brainstorm. We mix up ideas (as if at random), giving some priority to ideas that have worked in the past. Then we simulate the function of each combination of ideas, weeding out the ideas that plainly can't meet the goal. Finally, we implement and test the ideas we think have a good chance of working to see if they meet our original goal. Evolution is very similar to this process except that there's no simulation, and no imagination of the final product. Evolution has to implement every combination, even the ones that won't work, because it cannot tell ahead of time what is going to work. Furthermore, because there's no "goal comparator" to see if the final product meets the goal, there's only one goal that's compatible with mindless evolution: survival. I agree that evolution is not teleological, and does not "design". For a process to be labeled "design", the process must have some ability to simulate for fitness relative to a goal. Evolution doesn't have that. You say... "The products of evolutionary processes, therefore, exhibit no design." That's an interesting phrase. IMO, no product "exhibits design" per se. Rather, a product may exhibit utility relative to a goal. (For example, fruit bats exhibit superb utility relative to their survival niche. A pocket watch exhibits great utility for accurately measuring time and displaying the time visually.) However, idiomatically, we often refer to "products-with-utility" as "designs". I think Dawkins and others do this with evolved products because evolution is a mechanism that creates products with survival utility. When scientists refer to evolved products as designs, I read it as an idiom. When scientists say that rattlesnakes are a good design, they mean that rattlesnakes are products with good survival utility.
Bill, "A free act is not an uncaused act but one that is caused by the agent as ultimate source of the action." Perhaps you could contrast this with a quantum decay. When a neutron decays, there are certain constraints on the final state. There's conservation of momentum, energy, charge, etc. However, some aspects of the decay (like the directions of the radiating decay products) are not determined by the initial state or by timeless factors. Let's suppose that these undetermined aspects of the final state are fundamentally (as opposed to epistemically) random. In that case, the neutron's initial state can be said to have caused the final state only in the sense that it *enabled* the final state. That is, there would be no final state of decay products without the original neutron. However, the neutron did not *determine* the final state. Hence, the end result is not *determined* by anything whatsoever, even if it is *enabled* by something quite obvious. Similarly, it would seem that an uncaused act is one in which the choice is not *determined* by the agent, but is only *enabled* by the agent. And this leads us back into the jaws of the logical complementarity of determination and randomness. Every aspect of an agent choice that is merely enabled by the agent (i.e., not predetermined by the agent) is fundamentally random because there's nothing at all for it to depend upon. (We've exhausted the set of all things by considering determining factors both in time and outside of time.)
Bill, "While I was sitting there talking to him I had the thought that I could throw the coffee in his face. Of course I didn't do it. I had excellent reasons not to do it and no reason to do it. Yet I was aware of my freedom in the strong sense." (Sorry, italics aren't working anymore.) Is this really the strong sense you spoke of in an earlier post? Weren't you aware that, had your preference sided with the aesthetics of making a symbolic act of free will and throwing coffee in his face, then you could have done so? It really seems that if, on every ground, you prefer to not throw coffee in his face, then to throw the coffee would be to violate your free will, not affirm it. As for my big picture view on FW... First, I think the phenomenology doesn't conflict with determinism. Second, and more importantly, I think the concept of free will being a third category after determinism and randomness is logically incoherent. I think you can generally define an outcome as "determined" by saying that it is fixed by factors in time and in the past, or by constants that are out of time. The only factors that remain are those in the future, and that doesn't help. Having exhausted all factors (all those in time + all those outside of time), the outcome can only be random in the most fundamental sense. I'm fine with fundamental randomness, but I don't think that randomness = freedom for most people.
Peter, If, when Philoponus says choosing is illusory, he means that a belief in the strong sense of "Could have done otherwise" is illusory, then I agree. If he says that all choosing is illusory, then I don't buy it. I think you're begging the question against deterministic choice in the above. If determinism is true, we still weigh our actions from an epistemic point of view. We call that weighing "deciding". You can't roll "strong 'could have done otherwise'" into the definition of deciding without begging the question. Here's how I think it works. At the end of the decision-making process, the decider knows what he will do (or try to do). At the start of the decision-making process, the decider is not aware of which choice will be made. In this picture, the perception of a choice is the perception of an epistemic question. "Which action of mine will I find to be best?" To become aware of which action will be preferred, the decider has to simulate the outcome of each action. After the simulation, the decider is aware of what action will result in the most preferable outcome. At this point, the decider knows which action is preferred, and, therefore, which action he will perform. Note that the decider does not know at the start of the process, but does know later. This doesn't mean that the decider isn't choosing. If my model is write, then "choosing" is *defined* to mean the mental activities I described above. We are not free to say that a deterministic version of the above is not choosing without begging the question. You say: "Therefore, Philoponus is *committed* (an illusion, I suppose) on pain of inconsistency to hold the view that the researcher does not *decide* whether to accept the hypothesis based upon the evidence. His brain already accepts the hypothesis before any *decision* is made by the researcher based upon the evidence." I don't think that's right, and (correct me if I'm wrong Philoponus). I think Philoponus is suggesting that he was bound to DECIDE on the eventual conclusion. Suppose you're trapped on a desert island where there are two wells. One well seems normal, and the other is marked with "Danger Poison" signs. Which well will you drink from? Your decision is determined as of now when I write this. You may weigh your options, but you're going to decide to drink from the well that isn't marked "Danger Poison". This is the case because you want to live (I assume) and drinking from the first well is preferable. (I cannot imagine that you would prefer to drink from the first well, but deliberately drink from the second anyway.) So you still make a decision (a weighing), but you're predictably going to weigh in favor of the safe well.
Peter, You say: If Philoponus' position is true, then he is not free to let the evidence decide. It seems to me that this statement is problematic on two grounds. First, if the evidence is doing the deciding, then Phil's decision is still deterministic. Presumably, the evidence is deterministic, e.g., the Pythagorean theorem isn't changing with time, but is determined. Second, your argument seems to be a non sequitur because you have not established that the material processes in Philoponus's brain aren't accounting for the evidence. A reductionist will say that, when Phil is being rational, Phil's brain is deterministically accounting for the evidence. If I have some simple components that follow simple rules of configuration, I can create a system from those components that implement very sophisticated rules.
Bill, But when we choose the course with the preferred outcome, we have the sense that we could do otherwise. In your post "Weak and Strong Readings of 'Could Have Done Otherwise'", you sided with a strong interpretation. However, I don't think that jives with the phenomenology. I think the strong reading might be a linguistic error. The terms "choice" and "predetermination" conflict, but only in a different context. For example, suppose I go to a board meeting to choose whether we'll do A or B, but I do not know that the board chairman has already determined he will ignore the outcome of the meeting and do A, no matter what. In that case, the "choice" made by the board is nullified by the "predetermined" decision of the board chairman. Clearly, the will of the board is defeated. They think they're making a decision when they're not. But that language game in which predetermination and choice conflict is not relevant to the free will of an individual. Setting aside decisions where I don't prefer one choice over another, I certainly don't have the intuition that I could have chosen against my preferences. Rather, I always mean the weak reading, i.e., that had I preferred an alternative, I could have picked the alternative. So, I can't agree on the phenomenology of the decision. Yes, I am making something happen (e.g., reaching for a beer) according to my preferences, but I don't also have the sense that my preferences are coming out of thin air, or that I would have decided against my overall preferences.
I don't see a problem in saying that a decision-making process can be deterministic. That process looks something like this. 1) We perceive that we have several potential courses of action. 2) We simulate the outcome of each course. 3) We choose the course with the preferred outcome. (Preferences being determined by other past factors.) 4) We execute the chosen course. and in the case of free will... 5) Our chosen course often results in the effects we foresaw, i.e., our simulations were mostly accurate, at least in the short term. Now the only thing here that seems remotely problematic is (1). How can there be potential courses of action when physics has determined which course we will eventually decide upon? Well, it seems to me that the "potential" nature of each course can be of an epistemic nature. Suppose I'm going to choose to cross the road when the nearest light turns red. The light turns red, I look around me, and observe that a car is turning on red and heading my way. I perceive a potential choice. I can either start crossing the road and hope the car stops for me, or I can wait til it passes. If determinism is true then it is already determined which course I'll take. However, until I do the simulating and the preferring, I'm not going to *know* which course I will take. Naturally, I decide to wait for the car to pass. In hindsight, knowing me well, you would easily have predicted that I would be cautious and wait for the car to pass. And the reason you know what I would have done is that you would be familiar with my decision-making processes, i.e., that I generally choose to be cautious. Nothing in this process is illusory. Choice is not illusory. The only thing that is illusory (for some) is the idea that had I gone back in time to the moment of my decision, kept every parameter identical, and yet chosen otherwise. I think our real intuition about free will is that, if we went back in time, and our preferences were different, then we could have chosen differently.
Your physical(ist) example of a pairing problem doesn't seem to be analogous. The light bulb problem is analogous to the original pairing problem in the case that our instrumentation cannot distinguish b1 from b2. Imagine that the fMRI machines in which Tim and Tom are laying are wired together so that they show just a single numerical sum of the activities of their brains instead of showing two pictures as they normally would. In that scenario, no one would say that any theory of mind was at fault for making the brain effects indistinguishable. I tried to think of another physical scenario that would do what you want. The double-slit experiment came to mind, but that physical duality says that multiple causes effectively become a single cause under certain microphysical conditions. IOW, it won't really suit your purpose because we're then considering a situation analogous to a mind meld. Thank you for the clear description of intrinsic causality. I assume that in your picture, the will causes a person to choose A instead of ~A without there being a law specifying that A would be chosen over ~A. That is, the will is the intrinsic cause of the choice. The problem seems to me that, to the extent the selection was not lawful, it was fundamentally random. In other words, if A causing B without the event being an instance of a general law, is just like having a one-time law that says A causes B in this unique situation at this time. That instance of causation becomes a brute fact of the universe, and brute facts are precisely random facts. That would imply that choice-making is fundamentally random, like a quantum decay. Indeed, quantum decays are isomorphic to this situation. You have some global constraints from conservation of energy/momentum/charge/etc, but beyond those constraints, the outcome is described by a one-time, unique law for each decay. That's the recipe for fundamental randomness. Is there an escape?
Bill, yes, I know that this is what incompatibilists believe, but I don't think it's inspired by our intuitive experience of "could have done otherwise". I think there are a couple of intuitions in the game, both of which I find reasonable: I1) When I made decision X, I could instead have made decision Y, assuming I had desired Y instead of X. I2) My decisions have an effect. I don't think there's an intuition I1B: I1B) When I made decision X, I could instead have made decision Y, even had I desired X instead of Y. (Nonexistent) (Note that I1B does not contradict I1.) I think incompatibilism is inspired by I2, not I1. I think that, starting from I2, incompatibilists infer that I1B somehow ought to be our intuition (though I don't think it is). (That is, if A causes agent B, then agent B causes C, C was destined to happen, even before agent B considered A.) In contrast, a compatibilist says that I1 and I2 are both correct intuitions, and will argue that decisions do have an effect in a deterministic universe. (If A causes agent B, then agent B causes C, B's deliberation still has an effect, namely, C.) I'm just saying that I don't have the intuition that, if everything had been identical at the time of my decision, I could have chosen something different. Maybe, I'm just weird. You say, "Deliberating, I have the sense that it is up to me what happens." I, too, have that intuition. However, if what I am is a function of the past (or of timeless constants), I don't see a contradiction. (If A causes agent B, and agent B then causes C, C would not have happened if A had not caused agent B, and B's deliberation, as an intermediate state.) Indeed, I have no intuition that my present state is not a function of the past (or timeless constants). I don't have complete knowledge of my present state or how it came to be, so I lack any intuition for broad claims about what affects that state. Rationally, however, I can see that, if I exclude past causes and timeless factors, I've completely exhausted all the factors that could possibly affect my decision (barring the future affecting the present). If a decision depends on an empty set, it would be fundamentally random. In other words, fundamental randomness seems to me to be the logical complement of determinism. There's no third, "free" category.
The starting point for this post was the "strong" notion of "could have done otherwise." The problem with that reading is that it is incoherent. Suppose I believed that choice A satisfied my desires better than choice B, so I made choice A. Now we ask whether I could have done otherwise on the strong reading. If I could have believed that A satisfied my desires better than B, but somehow chose B, that would precisely be a failure of my will. To argue the strong reading is to deny free will. I sincerely doubt that anyone's intuition of free will would include the possibility that they could have had the same beliefs and desires at time T, but would have made a different choice than they had. When I say I could have chosen differently at the ice cream shop last night, I don't mean that I could have recognized that a chocolate ice cream would best satisfy my desires, but buy a coconut ice cream instead. I really mean that had I recognized that coconut ice cream would have satisfied my needs better, then (and only then) would I have chosen to buy the coconut ice cream.