P160 naturalismclassroomlect

Editor's Notes

• #10: Whether empirical data was to be the basis for theory or the test of theory, [11] we couldn’t be sure that science would end up with a true theory. We saw that we couldn’t test theories without relying on auxiliary hypotheses (some of which might be hard to test themselves)
• #11: that it’s always possible to find more than one theory that fits a given set of observations
• #12: and that our observations of the world thus far give us no secure information about what the world will be like in the future.
• #13: All these problems might make you think the foundational approach is hopeless. Of course, [15] we could give up on justifying the scientific methodology altogether and instead just describe it.
• #14: This is what sociology does: it gives a detailed picture of the customs of the tribe of science.
• #16: There’s no way to give anything but local justification for these customs, though, so the sociologist can offer [18] no reason to think that scientific knowledge is objectively better than the knowledge produced by any other tribe.
• #17: This way of avoiding the difficulties of foundationalism comes at a high cost: we can’t explain what we set out to explain in the first place. While the sociological account of science may tell us something about what makes science distinctive , it doesn’t do anything to explain the success of science. If the success of science is something we would like to be able to explain, maybe we shouldn’t throw up our hands and turn the project of understanding science over to the sociologists. Isn’t there some middle ground we can find between the foundationalist approach and the purely descriptive sociological approach?
• #22: Naturalism is a philosophical alternative to foundationalism. [24] What is the approach the naturalist takes?
• #23: One of the slogans of naturalism is, [25] “Philosophy should be continuous with science.” It’s a nice slogan, but what does it mean? For one thing, rather than standing outside of science and trying to give an independent justification of how science tries to solve problems, the naturalist accepts that the scientific method is what it is.
• #24: Science doesn’t seem to think it needs this to get the OK from philosophy to use the scientific method, so naturalist philosophers will think it’s not a useful philosophical project to try to supply the logical justification for this method.
• #25: Another part of the idea that philosophy should be continuous with science is that philosophers can use results from science as a resource to help answer philosophical questions.
• #26: If you’re trying to answer a question in epistemology (the branch of philosophy dealing with what we can know and how we come to know it), you’re allowed to [29] take as your starting point our best scientific theories of human beings, the workings of their brains, and the workings of their sense organs.
• #27: As this is the best knowledge on the matter right now, the naturalist says, there’s no reason not to use it in our philosophical investigations. There is, as always, a chance these scientific theories will turn out to be wrong. If they do, we’ll have to try to work through the philosophical questions again using whatever scientific theories come next.
• #28: This sounds OK to most people until the naturalist insists that [30] using scientific results as a philosophical resource is a good approach EVEN IN THE PHILOSOPHY OF SCIENCE. Here, some people worry.
• #29: Traditionally, a large part of what the philosophy of science has tried to do is to [31] demonstrate the reliability of scientific knowledge.
• #30: Using scientific results to help answer a question assumes these results are reliable.
• #31: But if you’re assuming certain scientific results are reliable in order to demonstrate that scientific results are reliable
• #32: you’re locked in a circular argument! So you might wonder whether naturalism will be any help at all for philosophers of science.
• #33: Remember, however, that naturalists have abandoned the foundational project. Rather than getting caught in a vicious circularity, [35] naturalism shifts the question.
• #35: Like most “isms,” naturalism comes in many varieties. The more radical variety of naturalism holds that the right thing to do is to replace philosophical questions with scientific ones. For example, radical naturalized epistemologists think the only proper answer to the question of how we know what we know will be a detailed scientific description of how beliefs are formed and how they change. How do you know she loves you? It has to do with your neurons and your brain chemicals, and here are the details. But there are plenty of naturalists who think a wholesale replacement of philosophy with science would be a mistake. After all, some questions are distinctly philosophical. The naturalist thinks that science can help answer these questions, but it shouldn’t replace them.
• #36: Let’s take a closer look at one of these questions to see why science alone couldn’t give a satisfying answer. [38] Which belief-forming mechanisms are good ones?
• #38: Scientists usually get uncomfortable identifying a goal for a mechanism or organ or organism. They’d rather just describe what the mechanism or organ or organism actually does. This is where the philosophers come in. [40] Philosophers can propose a number of possible goals and use these to evaluate policies for forming beliefs to distinguish which are better and which are worse for achieving a particular goal.
• #39: For example, it’s likely that we’ve all formed the belief that the sun will rise tomorrow, and we’ve done so on the basis of lots of observed sunrises. [42] Is the mechanism with which we’ve formed this belief a good one?
• #40: Psychology can tell us an awful lot about the connection between the sunrises we’ve experienced and what kind of belief we’ll form about tomorrow. The naturalist philosopher takes the question of whether it’s good that we form the belief that we do.
• #41: Psychology can tell us an awful lot about the connection between the sunrises we’ve experienced and what kind of belief we’ll form about tomorrow. The naturalist philosopher takes the question of whether it’s good that we form the belief that we do.
• #43: If your goal is to be able to make plans for tomorrow, believing that the sun will rise tomorrow is a fine belief.
• #44: If, on the other hand, you want to avoid error at all costs, the problem of induction lets us see that this is a risky belief. Notice here that the naturalist is not committed to one of these goals over the other. The point is not to identify and justify the one best set of goals in some absolute way. That would make this a foundationalist project. Rather, the naturalist uses an idea of rationality as an instrument to help you achieve the particular goal you want to achieve. Is it rational to believe the sun will rise tomorrow as a means to make plans for tomorrow? Yes, if holding that belief (versus the belief that the sun won’t rise tomorrow) is likely to make such planning possible.
• #45: We have a flavor of the naturalist approach. Let’s look at how it can be used to counter one of the most troubling claims we saw from Kuhn. [47] It seems intuitive that observation and experiment make science responsive to the real structure of the world. What we observe should be a good test of whether our theories are accurate pictures of the world.
• #47: But Kuhn insisted that [49] observations are theory-laden: [50] what we observe depends on what theory we hold, so [51] our observations CAN’T provide an objective test of our theory, and we may never know what part of our experience is due to the world rather than to our theories about the world. How can we do science together if, possibly, we all see very different things when we look at the same patch of the world?
• #48: But Kuhn insisted that [49] observations are theory-laden: [50] what we observe depends on what theory we hold, so [51] our observations CAN’T provide an objective test of our theory, and we may never know what part of our experience is due to the world rather than to our theories about the world. How can we do science together if, possibly, we all see very different things when we look at the same patch of the world?
• #49: But Kuhn insisted that [49] observations are theory-laden: [50] what we observe depends on what theory we hold, so [51] our observations CAN’T provide an objective test of our theory, and we may never know what part of our experience is due to the world rather than to our theories about the world. How can we do science together if, possibly, we all see very different things when we look at the same patch of the world?
• #50: What can the naturalist say to respond to Kuhn here? It is true that theory tells scientists where to look and what to look for. But this in itself doesn’t stop observation from being a way to TEST a theory; it’s perfectly possible to find something where the theory tells you to look that contradicts the theory. It is also true that theoretical assumptions affect which observations scientists take seriously and which they view as equipment malfunctions. (This is part of holism about testing.) But is Kuhn’s big claim, that beliefs influence experiences themselves, also true? We know of instances where the same pattern of stimulation on the retina can — and does — lead to different observations. Remember some of the examples Kuhn discussed to make this point: [52] the inverting goggles, [53] the doctored playing cards, and [54] the duck-bunny.
• #52: Given a particular input to my eyes, there are multiple possibilities for what I will observe. But since this input leads me to a particular experience, the Kuhnian argument goes, there must be theoretical assumptions that my visual system uses to choose what I actually observe. And if these theoretical assumptions are entangled in what I actually observe, how could I possibly use observations to test those theoretical assumptions?
• #53: In response, the naturalist points to a nice result from psychology, [55] the Muller-Lyer illusion. Confronted with this pair of arrows, observers consistently identify the top arrow as shorter, the bottom arrow as longer. In fact, the segments are exactly the same length. Scientists say that we observe the arrows this way because of background assumptions about the structure of the world that we use unconsciously to process our visual inputs. Isn’t this more ammunition for Kuhn? Doesn’t this show yet again that our beliefs about the world affect what we see?
• #54: The naturalist says, closer scrutiny of the Muller-Lyer illusion helps us see that we don’t have to give up on observation altogether. [56] While certain background assumptions may have a big impact on how we see the world, [57] others will have no effect on perception at all.
• #55: For example, the top arrow still looks shorter than the bottom arrow even if you know the two are the same length.
• #56: Knowing that this is an illusion, or even knowing a detailed theory of optical illusions, doesn’t change the fact that you still see the one as longer and the other as shorter. What this tells us is that our perception is influenced by SOME theories, but not by others.
• #57: In particular, scientists have identified “low-level sets of assumptions about the physical layout of the world” as very important to perception. Neuroscientists tend to locate these sets of assumption in what they call the perceptual module of the brain. Other sorts of theories and assumptions would be less likely to affect perception, although they could make a big difference in how we interpret and respond to what we perceive.
• #58: The Muller-Lyer illusion doesn’t show that a scientific theory could NEVER affect what we experience, but it suggests that such high-level theories are less likely to have such an effect.
• #59: Rather, [60] the perceptual module and the assumptions it builds in will [61] shape what we experience,
• #60: and [62] our scientific theories will most likely shape [63] our interpretation of that experience.
• #61: and [62] our scientific theories will most likely shape [63] our interpretation of that experience.
• #62: And, [64] if the perceptual module output is fairly uniform from person to person – regardless of what scientific theories or other higher-level assumptions they make – [65] this output could provide an intersubjective basis for theory choice. Thus, scientific decisions would be much more objective than Kuhn’s account would lead us to believe.
• #63: Of course, just because we agree about observations doesn’t necessarily mean they give us reliable knowledge about the world. So, the community of scientists could have perfectly objective ways to test theories against observations without those theories necessarily being true. This is a problem we’ll return to. We’ve just looked at how naturalists have deployed a bit of science, the Muller-Lyer illusion, to show that it’s not unreasonable to think we could use observations to test our theories. The naturalist approach moves us out of the Kuhnian pit of subjectivity and makes it plausible that we could offer some explanation for the success of science. Naturalists have also followed the sociologists, looking at the social structure of science in order to understand science’s success. Let’s explore a couple naturalist pictures of the tribe of science.
• #65: The naturalist philosopher of science David Hull draws on [67] sociology and organizational psychology
• #66: to show that [68] the special features of science arise from a combination of competition and cooperation.
• #67: You’ll recall that the sociologist [69] Robert Merton described the tribe of science as governed by four norms (universalism, communism, disinterestedness, and organized skepticism)
• #68: and [70] a reward system (establishing your priority for a result or idea).
• #69: Hull’s picture of science starts by suggesting an [71] alternative reward system:
• #70: the goal scientists strive for is [72] to have their ideas USED (and cited) by other scientists. Developing a convenient form of an idea someone else put forward first is still an important contribution to science. And, getting to an idea first when that idea is a “dead end” that doesn’t lead to new findings isn’t so great. Following Merton’s norms would keep scientists in pretty close contact with each other and reality (especially through the norm of organized skepticism). The reward system is supposed to be what motivates people to join the tribe of science in the first place.
• #72: Striving for the reward seemed to encourage subverting the norms: you’d be less likely to waste time checking someone else’s work, less likely to share your preliminary findings lest a competitor scoop you, and at least a little tempted to cut corners by making up your results.
• #73: How does Hull’s modification of the reward system make more sense of science? [75] Getting your idea used requires cooperation.
• #74: Your own ideas depend on the ideas of others
• #75: both as a starting point upon which to build
• #76: and [78] as support for new ideas
• #77: You need to build your ideas before others can use them, and you can’t build ideas others will use without relying on the ideas of others. So the self-interests of the members of science are entangled — if the reward we each seek is the use of our ideas, we can’t achieve this reward without the cooperation of the others who actually use our ideas.
• #78: Why is this entanglement of self-interests important? Hull argues that this is what generates the special features of science. Under Merton’s picture, it was difficult to see why scientists would spend much time replicating and checking results, since this investment of time and effort wouldn’t result in new publications or priority claims. However, in a system where [80] each scientist strives to produce work that is used
• #79: it makes sense to [81] check the work of others before you use it as a basis to set up new work of your own. You wouldn’t want to base your new power plant design on a shaky theory like cold fusion; you’d want to make sure cold fusion actually worked first. Moreover, scientists have a better chance of seeing their results used if they can assure others that they will be a reliable basis for their work.
• #80: Moreover, scientists have a better chance of seeing their results used if they can assure others that they will be a reliable basis for their work. Thus, [82] you have an incentive to check your own work carefully before you put it out.
• #81: So, according to Hull, the norm of organized skepticism flows from the peculiar reward for use. Hull’s account makes it clear why fraud is such a serious matter to the tribe of science. If someone makes up a finding and I try to use it as the basis of my work, this will undermine the usefulness of my own work. Once the fraud is discovered, all the honestly done work that attempts to build on it will be undermined as well, and no one else will use this work. This means a bunch of other scientists will be especially mad at the cheater!
• #82: The norm of communism flows from the reward system, too: once my ideas are published, I’ll want to share them widely, since failing to share them means that these ideas won’t be available for others to use or cite. And, the norm of universalism fits nicely with this reward system. Not only do newer scientists cite established scientists (rewarding the established scientists by using their ideas), but their new work making use of old ideas puts them in a good position to be rewarded themselves (perhaps the established scientists will take notice of these extensions of their work and cite the new scientists).
• #83: Hull’s picture of science gives more satisfying answers to certain philosophical questions than does Merton’s. How can we explain why the beliefs generated within the tribe of science fit the world as well as they do? And how can we resolve the tension between basic human self-interest and the desire for credit and the sort of disinterested scrutiny the tribe of science needs to exercise to produce an accurate picture of the world? If the reward driving the tribe of science is use, then the ideas that survive within this tribe will be the ideas that are used.
• #84: To be usable, an idea must [86] survive testing and attempts to replicate the predicted outcomes.
• #85: It must fit reasonable well with the other ideas in use.
• #86: And it must lead us to new ideas.
• #87: These features are the mark of an idea that also fits well with the real world.
• #88: Notice that in Hull’s system scientists don’t need to be saints.
• #89: It’s entirely possible that an individual scientist will be less skeptical about his own ideas.
• #90: But, before his ideas are used, [92] other members of the community will subject them to close scrutiny. So, Hull’s picture of science show us how the psychology of individual scientists can lead to the community-level properties that lead to successful scientific knowledge.
• #91: Hull’s naturalistic picture of science seems to be an improvement over Merton’s. Merton’s description made the success of science (plus the peaceful coexistence of the norms and the reward system) more mysterious. [93] By identifying the peculiar reward scientists seek, Hull can [94] explain why science has the norms it does, as well as [95] how those norms propel science in the right direction.
• #92: Hull’s naturalistic picture of science seems to be an improvement over Merton’s. Merton’s description made the success of science (plus the peaceful coexistence of the norms and the reward system) more mysterious. [93] By identifying the peculiar reward scientists seek, Hull can [94] explain why science has the norms it does, as well as [95] how those norms propel science in the right direction.
• #93: Hull’s naturalistic picture of science seems to be an improvement over Merton’s. Merton’s description made the success of science (plus the peaceful coexistence of the norms and the reward system) more mysterious. [93] By identifying the peculiar reward scientists seek, Hull can [94] explain why science has the norms it does, as well as [95] how those norms propel science in the right direction.
• #94: For our last example, let’s look at how [96] Philip Kitcher uses naturalism to answer a normative question about the social structure of science.
• #95: The scientific resource Kitcher uses is modeling techniques from game theory (techniques widely used in scientific fields like population genetics and economics).
• #96: The philosophical task is to determine, in a scientific field with competing research programs, the best distribution of researchers among rival programs.
• #97: There are two related questions. [99] First, how should the individual scientist decide which research program to join? Remember that Laudan says he should pursue the program with best current rate of puzzle-solving success. Human nature tells the scientist to join the research program most likely to bring me the biggest reward.
• #98: Next, what distribution of workers is best for the scientific field? Of course, the one that’s most likely to lead to a successful solution to the problem.
• #99: Kitcher takes up a normative question here: [101] How should science be set up to ensure that individual scientists will make the choices that are best for science as a whole? To answer this question, he looks at what kind of reward system will get rational scientists looking out for their own interests to act in a way that harmonizes with the interests of the group.
• #100: The community’s goal is to solve a scientific problem. Let’s say there are two research programs aimed at solving this problem. We don’t know which (if either) research program will actually solve it, but one may be more promising than the other. [102] Let’s say the first has an 80% chance of success and the second has a 15% chance of success. What’s the best distribution of labor between the two research programs?
• #103: Indeed, we can get a more precise optimal distribution of scientists between the projects if we take account of [105] “decreasing marginal returns”. For any research program, you need a certain amount of scientific labor to get it going and keep it running.
• #104: But after a certain point, each additional scientist won’t make much difference to that research program’s chance of succeeding.
• #105: Once the most promising research program hits that point, [107] the rest of the scientists should work on the other research program. We’ve got a pretty clear picture of the distribution of scientific labor that would be best for the community.
• #106: The question now is whether the reward system encourages individual scientists (who are trying to secure the biggest reward) to choose research programs in a way that produces the best distribution for the group.
• #107: Here’s a reward system that wouldn’t work to produce the best distribution: Each worker on the research program that succeeds gets a fixed “prestige” payout, while each worker on the research program that fails walks away with nothing.
• #108: This kind of payout won’t give the best overall labor distribution because [110] each scientist, not wanting to walk away empty-handed, will try to bet on a winner.
• #109: Individual scientists will all clamor to join the more promising research program, since this is the one with greater chances of producing a reward.
• #110: No one will take the risk of joining the less promising research program and getting no payout at all.
• #111: Kitcher says individual scientists will be more likely to make choices leading to the optimal distribution if the reward is structured differently. [113] Instead of getting a fixed amount of prestige for working on the research project that succeeds, each worker on the successful research project gets an equal SHARE of a fixed prestige payout.
• #112: In other words, the more workers on the project, the more people get shares of the prestige, and the smaller each of those shares will be. How does this encourage individual scientists to making the choices that benefit the group? [115] The individual scientists have little incentive to join a crowded group, even if the research project that group is pursuing is very promising. Instead, going with a riskier but less crowded group increases the individual’s expected payoff.
• #113: How does this encourage individual scientists to making the choices that benefit the group? [115] The individual scientists have little incentive to join a crowded group, even if the research project that group is pursuing is very promising. Instead, going with a riskier but less crowded group increases the individual’s expected payoff.
• #115: How does this encourage individual scientists to making the choices that benefit the group? [115] The individual scientists have little incentive to join a crowded group, even if the research project that group is pursuing is very promising. Instead, going with a riskier but less crowded group increases the individual’s expected payoff.
• #116: Michael Strevens offers a refinement to Kitcher’s game-theory model of the choice. [116] Strevens points out that a research program’s chance of solving a problem is NOT a fixed probability. [117] Rather, it depends in part on the number of workers pursuing the program.
• #117: Michael Strevens offers a refinement to Kitcher’s game-theory model of the choice. [116] Strevens points out that a research program’s chance of solving a problem is NOT a fixed probability. [117] Rather, it depends in part on the number of workers pursuing the program.
• #118: If you join a group that has just started pursuing a research program, there’s a lot more hard work for you, but you’re also more likely to make a real contribution to the group. On the other hand, [119] if you join the group late, most of the hard work may already be done. Your presence may do very little to raise the chances of success. But under Kitcher’s reward system, if the project succeeds, EVERYONE in the group gets an equal share of the payout. Even if you didn’t make much difference to the success of the project, you get the same payout as the people in the group who did make a difference. You are a free-rider: you’ve figured out how to see to your own interests without contributing to what’s best for the group.
• #119: If you join a group that has just started pursuing a research program, there’s a lot more hard work for you, but you’re also more likely to make a real contribution to the group. On the other hand, [119] if you join the group late, most of the hard work may already be done. Your presence may do very little to raise the chances of success. But under Kitcher’s reward system, if the project succeeds, EVERYONE in the group gets an equal share of the payout. Even if you didn’t make much difference to the success of the project, you get the same payout as the people in the group who did make a difference. You are a free-rider: you’ve figured out how to see to your own interests without contributing to what’s best for the group.
• #120: ] So Strevens says the reward system should be adjusted slightly. If you’re in the group that solves the problem, you get a share of the fixed prestige that’s proportional to your contribution. What this means is that you get a larger share of prestige if you joined early and made a big difference, a smaller share of prestige if you joined late and made very little difference.
• #121: Kitcher’s model of scientists’ decisions, and Strevens’ refinement of this model, look a lot like models you might actually find in economics or population genetics. These naturalist philosophers deploy tools from science (in this case, game theoretic models) to answer a philosophical question (how can science coordinate what’s good for the individual scientist and what’s good for the community?). Not only do these philosophical discussions not get bogged down like foundationalist ones, but the conclusions of these philosophical discussions might actually be useful to scientists! we’ve talked about naturalism, an approach that sees scientific results as a resource to help answer philosophical questions. We’ve seen some of the advantages of this approach over foundationalism and over strictly sociological approaches. And, we’ve examined how naturalism helps answer deep philosophical worries about relation between theory and observation. Naturalism is a movement in progress; time will tell what features of science naturalism will be able to illuminate and what problems it will run into. But, this is the approach taken in lots of the current research in the philosophy of science, so it’s a movement worth watching.