N
NowAgnostic
Guest
I, for one, am getting tired of all the endless evolution vs. ID threads. The same hackneyed points get raised over and over again, no matter how many times the other side thinks they’ve been refuted. To be clear by “supernaturalism” I mean the action of the “God of the philosophers” - an infinite being.
The question is, are naturalism and supernaturalism hypotheses which can be confirmed or disproven by scientific methods, or are both positions a priori philosophical positions unfalsifiable by evidence. Because, if the latter, everyone is just spitting in the wind attempting to argue based on evidence - the only arguments possible are philosophical and logical ones. And the answer is that both positions are unfalsifiable and you simply can’t argue based on science but can only argue it philosophically.
Now there’s of course the simpler argument and the more complicated one. The simpler argument revolves around Popper’s criterion of falsifiability. It’s nice when you get evidence consistent with a hypothesis, but it doesn’t allow you to conclude that particular hypothesis is true, since there could be another one which explains it even better. So what counts is when you get evidence inconsistent with a hypothesis which allows you to strike it off the list so to speak, and a hypothesis withstands the “test of time” by surviving repeated attempts at falsification. So, if a certain hypothesis is unfalsifiable in principle and can never be crossed off the list, then there is never any real test for it; it is untestable and therefore unverifiable.
The fancier argument involves using Bayes’ Theorem to attempt to mathematically quantify the evidence for various models. An unfalsifiable model expands the space of what it can model to be able to handle every single possible observation, and thus it has no explanatory nor predictive power. An example: you have 20 data points - 20 values of the independent variable combined with 20 values of the dependent variable. If you can model it with only 2 parameters (a line, say) with sufficient accuracy then the model explains the data well. But if you model it with 20 parameters (say, a linear combination of 20 basis functions) your model explains nothing. Your model would perfectly fit any 20 points. For the model to have “value”, its value must consist in the fact that it modeled well your actual data, but it might not have.
Now, supernaturalism is intrinsically unfalsifiable. There is not a single observation which could possibly disprove the existence of God. Let’s imagine the worst nightmare possible for some of the theists on this forum. Let’s imagine there were a completely verified theory of abiogenesis, there were absolutely no problems with evolution such that even higher life forms were seen to evolve in the lab, and a completely satisfactory scientific explanation for the origin of the universe. None of this would disprove God in the slightest way. Both atheists who claim that the current scientific knowledge somehow “does away with the need for God” or theists who fight tooth-and-nail the latest scientific discoveries because they believe the atheists when they claim that are wrong.
Since supernaturalism is intrinsically unfalsifiable, naturalism, being its opposite, is also intrinsically unfalsifiable. There is not a single observation which could possibly disprove naturalism. (No, not even alleged miracles.) True, there is not a scientific explanation for everything currently. But it is impossible to prove that one can never be found in principle. Theists claim naturalists operate “on faith”. So what? That’s not a disproof. Moreover, there is a fundamental misunderstanding of what scientific models are. Scientific models, or “laws” with enough evidence, are merely methods for finding a description of reality using the “minimum description length” or least amount of information. So, for instance (in Newtonian physics) the trajectory (position at each point in time) can be reduced to the initial position, initial velocity, mass, and forces acting on the body. All science is simply an information theory. The theist’s claim of evidence against naturalism basically boils down to the claim that some event (such as abiogenesis, evolution, or a miracle) violated a scientific law. They would have a point if laws were 1) prescriptive, rather than descriptive; 2) known completely; and 3) known a priori. But they aren’t; scientific knowledge is incomplete and more importantly, it is empirically inferred from the data. The inferences can change with new data, and so every claimed “violation” of a scientific law can simply be an occasion to refine the model to better fit the data; and in fact, that has happened many times in physics. The other claim that is repeated over and over again is the more moderate one that something, while not absolutely impossible, is highly unlikely to have occurred by mere “chance” under naturalism and therefore supernaturalism is the more likely hypothesis. But how likely is the event to have occurred under supernaturalism? This cannot be estimated because supernaturalism is unfalsifiable. Therefore, no inference can be made about the probability of naturalism. Back to the modeling analogy; let’s say my 19-parameter fit doesn’t model the 20 points very well. I can’t therefore conclude the 20-parameter model is a better model even though it fit the data perfectly. The 20-parameter model simply can’t be compared with any other model.
The question is, are naturalism and supernaturalism hypotheses which can be confirmed or disproven by scientific methods, or are both positions a priori philosophical positions unfalsifiable by evidence. Because, if the latter, everyone is just spitting in the wind attempting to argue based on evidence - the only arguments possible are philosophical and logical ones. And the answer is that both positions are unfalsifiable and you simply can’t argue based on science but can only argue it philosophically.
Now there’s of course the simpler argument and the more complicated one. The simpler argument revolves around Popper’s criterion of falsifiability. It’s nice when you get evidence consistent with a hypothesis, but it doesn’t allow you to conclude that particular hypothesis is true, since there could be another one which explains it even better. So what counts is when you get evidence inconsistent with a hypothesis which allows you to strike it off the list so to speak, and a hypothesis withstands the “test of time” by surviving repeated attempts at falsification. So, if a certain hypothesis is unfalsifiable in principle and can never be crossed off the list, then there is never any real test for it; it is untestable and therefore unverifiable.
The fancier argument involves using Bayes’ Theorem to attempt to mathematically quantify the evidence for various models. An unfalsifiable model expands the space of what it can model to be able to handle every single possible observation, and thus it has no explanatory nor predictive power. An example: you have 20 data points - 20 values of the independent variable combined with 20 values of the dependent variable. If you can model it with only 2 parameters (a line, say) with sufficient accuracy then the model explains the data well. But if you model it with 20 parameters (say, a linear combination of 20 basis functions) your model explains nothing. Your model would perfectly fit any 20 points. For the model to have “value”, its value must consist in the fact that it modeled well your actual data, but it might not have.
Now, supernaturalism is intrinsically unfalsifiable. There is not a single observation which could possibly disprove the existence of God. Let’s imagine the worst nightmare possible for some of the theists on this forum. Let’s imagine there were a completely verified theory of abiogenesis, there were absolutely no problems with evolution such that even higher life forms were seen to evolve in the lab, and a completely satisfactory scientific explanation for the origin of the universe. None of this would disprove God in the slightest way. Both atheists who claim that the current scientific knowledge somehow “does away with the need for God” or theists who fight tooth-and-nail the latest scientific discoveries because they believe the atheists when they claim that are wrong.
Since supernaturalism is intrinsically unfalsifiable, naturalism, being its opposite, is also intrinsically unfalsifiable. There is not a single observation which could possibly disprove naturalism. (No, not even alleged miracles.) True, there is not a scientific explanation for everything currently. But it is impossible to prove that one can never be found in principle. Theists claim naturalists operate “on faith”. So what? That’s not a disproof. Moreover, there is a fundamental misunderstanding of what scientific models are. Scientific models, or “laws” with enough evidence, are merely methods for finding a description of reality using the “minimum description length” or least amount of information. So, for instance (in Newtonian physics) the trajectory (position at each point in time) can be reduced to the initial position, initial velocity, mass, and forces acting on the body. All science is simply an information theory. The theist’s claim of evidence against naturalism basically boils down to the claim that some event (such as abiogenesis, evolution, or a miracle) violated a scientific law. They would have a point if laws were 1) prescriptive, rather than descriptive; 2) known completely; and 3) known a priori. But they aren’t; scientific knowledge is incomplete and more importantly, it is empirically inferred from the data. The inferences can change with new data, and so every claimed “violation” of a scientific law can simply be an occasion to refine the model to better fit the data; and in fact, that has happened many times in physics. The other claim that is repeated over and over again is the more moderate one that something, while not absolutely impossible, is highly unlikely to have occurred by mere “chance” under naturalism and therefore supernaturalism is the more likely hypothesis. But how likely is the event to have occurred under supernaturalism? This cannot be estimated because supernaturalism is unfalsifiable. Therefore, no inference can be made about the probability of naturalism. Back to the modeling analogy; let’s say my 19-parameter fit doesn’t model the 20 points very well. I can’t therefore conclude the 20-parameter model is a better model even though it fit the data perfectly. The 20-parameter model simply can’t be compared with any other model.
? Who’s side are you on? Why do you argue that God can be disproven?