Thursday, December 14, 2006
Popper and Experimenter’s Regress
(I'm posting this here in case my hard drive goes belly-up again. Everyone check for brainfarts!)
Karl Popper’s landmark take on science is both striking and inspiring, and many scientists consider its development an act of heroism. The influence of his theories is such that they have been invoked in scientific debate in order to evaluate one position or another, a rare feat for a philosopher in any field (Godfrey-Smith 57). Given his impact and his unique and arresting take on the matter, Popper’s views make for a pertinent and illuminating point of reference when dealing with other works on the subject, particularly those which are more sociological in their approach.
Popper’s philosophy of science is more or less two-fold. His first concern was with what he called the “problem of demarcation”, or the problem of distinguishing science from non-science. His solution to this problem lay in falsification: in order for a theory to be scientific, the hypothesis must be potentially refutable given certain possible observations. In other words, theories must take “risks” in order to be scientific. These risks usually amount to generalizations that prohibit the occurrence of certain kinds of events, so that if said events are observed, the theory is refuted (Popper 41). Falsificationism also dictates the objective of scientific testing: “Every genuine test of a theory is an attempt to falsify it, or to refute it (Popper 42)”.
In keeping with his solution to the demarcation problem, Popper described scientific change as the repetition of two simple stages, conjecture followed by attempted refutation. In the first stage, a scientist would endeavor to explain some aspect of the world by means of a falsifiable hypothesis. In the second stage, the hypothesis would be subjected to critical testing—or testing that aimed to show that the hypothesis was false—and would then be either refuted or kept on the books for more testing in future.
Popper’s other topic of interest was the inconclusive nature of science’s inductive methodology. He stressed repeatedly that science could not confirm a theory, regardless of how many times that theory agreed with observation (Popper 44). The scientific method could only refute a hypothesis; it could never show it to be true. Theories that withstood repeated attempts at falsification are “corroborated”, but corroboration only means resistance to falsification on multiple occasions. It does not imply confirmation.
Popper’s theories have often come under fire both directly and indirectly from those who investigate science primarily from a historical or sociological perspective. Harry Collins and Trevor Pinch offer such an account of science in the fifth chapter of The Golem: What You Should Know About Science. In this chapter, they reflect on the controversy that occurred in the early1970s over the detection of gravitational radiation, and they use it to illuminate a phenomenon that they call “experimenter’s regress”, or the social process of theory acceptance on the ‘research frontier’, a process which generally produces a verdict that is used to measure how well experiments have been run.
Beginning in 1969, Professor Joseph Weber of the University of Maryland claimed to have found evidence for gravitational radiation, the likes of which had been predicted by Einstein’s general theory of relativity, though not to the extent that Weber observed. Distinguishing between oscillations produced by gravitational radiation and those produced by the random vibrations within the device used to measure said radiation (vibrations which are bound to occur at any temperature above absolute zero), however, proved a trying and ambiguous task. Many scientists repeated the experiment, but the controversy came to an end more because of the rhetorical skill of one scientist—Richard Garwin—than because of any observational evidence to contradict Weber’s claims. Collins and Pinch quote one scientist as saying:
Another states that, “Garwin talked louder than anyone and he did a very nice job of analyzing his data… they sort of convinced everybody (Collins and Pinch 105).” It is apparent that the acceptance of a theory is by and large a social process, then. Note that this is in direct contrast with Popper, who asserts that, “…in science, only observation and experiment may decide upon the acceptance or rejection of scientific statements, including laws and theories (Popper 45).”
However, while the resolution of experimenter’s regress in the face of ‘novel phenomena’ differs from Popper’s account of science, a more marked difference lies in the way ‘science’ is conducted following the resolution. In the aftermath of experimenter’s regress, the social consensus becomes “the correct outcome”, and its production or lack thereof is used to determine how well experiments have been done:
In Collins’ and Pinch’s characterization of science, observation and experiment do not always determine whether or not a theory is correct. Instead, the scientific community reaches a consensus as to the correctness of a theory, and the production of the outcome that has been agreed upon becomes a way to measure how well an experiment has been done. Experiments, after this point, are no longer used to test theories; theories are instead used to test experiments. This is, admittedly, a rather dangerous practice, especially if current methods of experimentation are known to be more exact than those that were used when consensus was forged. Moreover, this approach contrasts drastically with Popper’s view of science:
What Popper’s view of science fails to account for, then, is the fallibility of the experiments themselves. Most anyone who has taken a laboratory course knows that there are many things that can go wrong during an experiment, whether it be due to equipment malfunction or human carelessness.
Were one to constantly subject both outcomes and experiments to the kind of rigorous scrutiny that Popper would have applauded, scientific progress would likely grind to a halt. Not only would outcomes provide no definite proof for hypotheses, but it would also be impossible to conclusively falsify a theory. This too, though, would contrast with Popper’s account of science:
One could argue that Popper concerns himself primarily with “the problem of demarcation”, and that the evaluation of experimental procedure in another issue in and of itself, one that Popper simply left unaddressed. However, Popper does articulate clearly that experimentation is not scientific unless it genuinely attempts to falsify a hypothesis; and this is plainly not the case in the most practical classes in schools worldwide.
Popper emphasizes that science’s inductive method cannot confirm or establish theories; that at best, theories can be corroborated, or resist falsification. And while most scientists would agree wholeheartedly with this, they for practical purposes tend to treat corroborated theories as if they have been proven, using them as benchmarks to test other theories, and, perhaps more perversely, to test experiments themselves. This practice is not only unscientific according to Popper; it is fundamental inversion of what Popper advocates.
Science clearly requires a set of criteria to evaluate the quality of both theories and experiments. However, if one is to believe Collins and Pinch, the way that scientists currently tend to go about conducting these evaluations is decidedly unscientific in nature. Experimenter’s regress is resolved by a combination of rhetorical mudslinging and other social factors, and the outcome is then used as criteria for experimental quality. If this process is science, it is not Popper’s science.
Works Cited
Collins, Harry and Pinch, Trevor. The Golem: What You Should Know About Science. London: Cambridge University Press, 1998.
Godfrey-Smith, Peter. Theory and Reality. Chicago: The University of Chicago Press, 2003.
Popper, Karl. “Science: Conjectures and Refutations”. Introductory Readings in the Philosophy of Science. Ed. E.D. Klemke, Robert Hollinger, David Wyss Rudge. New York: Prometheus Books, 1998.
Karl Popper’s landmark take on science is both striking and inspiring, and many scientists consider its development an act of heroism. The influence of his theories is such that they have been invoked in scientific debate in order to evaluate one position or another, a rare feat for a philosopher in any field (Godfrey-Smith 57). Given his impact and his unique and arresting take on the matter, Popper’s views make for a pertinent and illuminating point of reference when dealing with other works on the subject, particularly those which are more sociological in their approach.
Popper’s philosophy of science is more or less two-fold. His first concern was with what he called the “problem of demarcation”, or the problem of distinguishing science from non-science. His solution to this problem lay in falsification: in order for a theory to be scientific, the hypothesis must be potentially refutable given certain possible observations. In other words, theories must take “risks” in order to be scientific. These risks usually amount to generalizations that prohibit the occurrence of certain kinds of events, so that if said events are observed, the theory is refuted (Popper 41). Falsificationism also dictates the objective of scientific testing: “Every genuine test of a theory is an attempt to falsify it, or to refute it (Popper 42)”.
In keeping with his solution to the demarcation problem, Popper described scientific change as the repetition of two simple stages, conjecture followed by attempted refutation. In the first stage, a scientist would endeavor to explain some aspect of the world by means of a falsifiable hypothesis. In the second stage, the hypothesis would be subjected to critical testing—or testing that aimed to show that the hypothesis was false—and would then be either refuted or kept on the books for more testing in future.
Popper’s other topic of interest was the inconclusive nature of science’s inductive methodology. He stressed repeatedly that science could not confirm a theory, regardless of how many times that theory agreed with observation (Popper 44). The scientific method could only refute a hypothesis; it could never show it to be true. Theories that withstood repeated attempts at falsification are “corroborated”, but corroboration only means resistance to falsification on multiple occasions. It does not imply confirmation.
Popper’s theories have often come under fire both directly and indirectly from those who investigate science primarily from a historical or sociological perspective. Harry Collins and Trevor Pinch offer such an account of science in the fifth chapter of The Golem: What You Should Know About Science. In this chapter, they reflect on the controversy that occurred in the early1970s over the detection of gravitational radiation, and they use it to illuminate a phenomenon that they call “experimenter’s regress”, or the social process of theory acceptance on the ‘research frontier’, a process which generally produces a verdict that is used to measure how well experiments have been run.
Beginning in 1969, Professor Joseph Weber of the University of Maryland claimed to have found evidence for gravitational radiation, the likes of which had been predicted by Einstein’s general theory of relativity, though not to the extent that Weber observed. Distinguishing between oscillations produced by gravitational radiation and those produced by the random vibrations within the device used to measure said radiation (vibrations which are bound to occur at any temperature above absolute zero), however, proved a trying and ambiguous task. Many scientists repeated the experiment, but the controversy came to an end more because of the rhetorical skill of one scientist—Richard Garwin—than because of any observational evidence to contradict Weber’s claims. Collins and Pinch quote one scientist as saying:
…As far as the scientific community in general is concerned, it’s probably Garwin’s publication that generally clinched the attitude. But in fact the experiment they did was trivial—it was a tiny thing… But the thing was, the way the wrote it up… Everybody else was awfully tentative about it, it was all a bit hesitant. And then Garwin comes along with this toy. But it’s the way he writes it up you see. (Collins and Pinch 104)
Another states that, “Garwin talked louder than anyone and he did a very nice job of analyzing his data… they sort of convinced everybody (Collins and Pinch 105).” It is apparent that the acceptance of a theory is by and large a social process, then. Note that this is in direct contrast with Popper, who asserts that, “…in science, only observation and experiment may decide upon the acceptance or rejection of scientific statements, including laws and theories (Popper 45).”
However, while the resolution of experimenter’s regress in the face of ‘novel phenomena’ differs from Popper’s account of science, a more marked difference lies in the way ‘science’ is conducted following the resolution. In the aftermath of experimenter’s regress, the social consensus becomes “the correct outcome”, and its production or lack thereof is used to determine how well experiments have been done:
The student can have a good idea whether or not her or she has done an experiment competently by referring to the outcome. If the outcome is in the right range, then the experiment has been done about right, but if the outcome is in the wrong range, then something has gone wrong. In real time, the question for difficult science, such as the gravity wave case and the others described in this book, is, ‘What is the correct outcome?’. Clearly, knowledge of the correct outcome cannot provide the answer. Is the correct outcome the non-detection of gravity waves? Since the existence of gravity waves in the very point at issue, it is impossible to know this at the outset (Collins and Pinch 98).
In Collins’ and Pinch’s characterization of science, observation and experiment do not always determine whether or not a theory is correct. Instead, the scientific community reaches a consensus as to the correctness of a theory, and the production of the outcome that has been agreed upon becomes a way to measure how well an experiment has been done. Experiments, after this point, are no longer used to test theories; theories are instead used to test experiments. This is, admittedly, a rather dangerous practice, especially if current methods of experimentation are known to be more exact than those that were used when consensus was forged. Moreover, this approach contrasts drastically with Popper’s view of science:
Confirming evidence should not count except when it is the genuine test of the theory; and this means that it can be presented as a serious but unsuccessful attempt to falsify the theory (Popper 42).
What Popper’s view of science fails to account for, then, is the fallibility of the experiments themselves. Most anyone who has taken a laboratory course knows that there are many things that can go wrong during an experiment, whether it be due to equipment malfunction or human carelessness.
Were one to constantly subject both outcomes and experiments to the kind of rigorous scrutiny that Popper would have applauded, scientific progress would likely grind to a halt. Not only would outcomes provide no definite proof for hypotheses, but it would also be impossible to conclusively falsify a theory. This too, though, would contrast with Popper’s account of science:
Only the falsity of a theory can be inferred from empirical evidence, and this inference is a purely deductive one (Popper 46).
One could argue that Popper concerns himself primarily with “the problem of demarcation”, and that the evaluation of experimental procedure in another issue in and of itself, one that Popper simply left unaddressed. However, Popper does articulate clearly that experimentation is not scientific unless it genuinely attempts to falsify a hypothesis; and this is plainly not the case in the most practical classes in schools worldwide.
Popper emphasizes that science’s inductive method cannot confirm or establish theories; that at best, theories can be corroborated, or resist falsification. And while most scientists would agree wholeheartedly with this, they for practical purposes tend to treat corroborated theories as if they have been proven, using them as benchmarks to test other theories, and, perhaps more perversely, to test experiments themselves. This practice is not only unscientific according to Popper; it is fundamental inversion of what Popper advocates.
Science clearly requires a set of criteria to evaluate the quality of both theories and experiments. However, if one is to believe Collins and Pinch, the way that scientists currently tend to go about conducting these evaluations is decidedly unscientific in nature. Experimenter’s regress is resolved by a combination of rhetorical mudslinging and other social factors, and the outcome is then used as criteria for experimental quality. If this process is science, it is not Popper’s science.
Works Cited
Collins, Harry and Pinch, Trevor. The Golem: What You Should Know About Science. London: Cambridge University Press, 1998.
Godfrey-Smith, Peter. Theory and Reality. Chicago: The University of Chicago Press, 2003.
Popper, Karl. “Science: Conjectures and Refutations”. Introductory Readings in the Philosophy of Science. Ed. E.D. Klemke, Robert Hollinger, David Wyss Rudge. New York: Prometheus Books, 1998.
