Main reasons to think about this question:
1. Much of the best science is about UNOBSERVABLE entities (genes, electrons, the center of the earth); what can we know about something we cannot directly perceive?
2. The long history of science is full of theories that led to technological success and new devices, but were shown to be false later on. For example, the geocentric model of the universe described by Ptolemy was used for centuries (and can still be used today) for navigation purposes.
3. Granting that science is technologically successful, we can still ask, WHY is science successful? What is responsible for its success? What explains the success of science?
~ It seems to have negative connotations. E.g., ãIn theory, communism is the best form of government.ä Or ãThatâs just your theory.ä Here Îtheoryâ is being opposed to Îfacts.â
~ But all the major accomplishments of modern science are all called Îtheories.â For example: the general Theory of relativity, the quantum Theory, the kinetic Theory of gases, genetic theory, microeconomic theory.
~ Being-a-theory is irrelevant to being-true: there are true theories and false theories. The word Îtheoryâ by itself does not distinguish between the two.
What is the aim or goal of a scientific theory? Theories do NOT merely DESCRIBE what we see (the Îphenomenaâ); they EXPLAIN why we see what we see; they reveal the hidden order of the universe that causes the phenomena that we can see. [NOTE: Not all philosophers of science agree about this; some claim that the aim of science is empirical adequacy.]
~ Thus, no scientific theory is absolutely certain (ãindubitableä (15)), since basically all go Îbeyond the phenomena,â that is, theories make claims about entities and objects we cannot directly experience.
Signs/ indicators that a theory is true (ãInternal and
External Virtuesä)
~Since we can never tell for certain when a theory is true, we have to judge theories by indirect indicators of truth. A good analogy is a jury listening to a witness in court. The jury cannot go back in time to view the crime happening, thus a jurist cannot directly determine whether a statement a witness makes is true or not. But there are indirect indicators. Kosso classifies the indicators into two types: internal and external.
~ Internal features of a theory are ones that donât depend on the way the world is, external features do. So, whether a particular theory correctly predicts the outcome of a certain experiment or not is clearly an External feature.
1. Logical consistency. This is jargon for Îhas no contradictions.â In the legal analogy, we know that at least part of a witnessâs testimony must be false if he states that the criminal was over 6 feet tall, and later states that the criminal was under 6 feet tall. And we know something the witness said must be false, regardless of any facts about the crime.
2. Simplicity. People generally prefer simpler theories to complicated ones, but are simpler theories more likely to be true? Some people think so.
3. Testability. Some theories can be subjected to tests, whereas others cannot. Consider a Zen example: suppose I claim that when a tree falls in a forest, and no entity capable of perception is around to hear the tree crash, that tree makes no sound at all. How could this theory be tested? It cannot.
~ We asked last lecture, ÎWhat is science?â One answer is: a claim that is subjected to empirical testing. When a claim or theory passes an empirical test, we say that claim has been confirmed (to a degree x÷Îconfirmedâ is much weaker than Îprovenâ).
~ Basic (over-simple?) picture of confirmation: ãHypothetico-deductiveä (H-D) account of confirmation. You have the basic theory. The theory makes a specific experimental prediction. For example, one of the consequences of Einsteinâs special relativity is that clocks moving close to the speed of light will run detectably slower than clocks not moving with such a high velocity. So, if we have two clocks in our lab that are initially synchronized, and then we send one of them on a high-velocity round-trip journey, when it returns to our lab it will show an earlier time than the clock in our lab shows. And if an experiment similar to this (using Îatomic clocksâ) is performed, we find that the clock that took the trip does show an earlier time. So the special theory is confirmed. However, the special theory is not thereby definitely true; that would be a mistaken inference. We know: ÎIf STR is true, then clocks traveling near the speed of light tick slower than ones which do not move so fast.â Finding out that the Îthen·â part of that statement is true does not guarantee that the Îif·â part is true. For consider the following true if-then statement: ÎIf it is raining, then it is cloudy.â Now suppose we look out the window and see that it is cloudy. Can we infer that it is raining? Obviously not. And this argument is exactly the same form as the STR argument.
~ Now we see that one cannot say ãThe difference between myth and science is that science proves its claims, whereas myths do not,ä at least if Îx is provedâ means Îx is certain.â Scientific theories are more-or-less confirmed, but not proven to be true.
~ But what happens when a theory makes a prediction, and we do the experiment, and we get an actual result different from the predicted result? Suppose we did the clock experiment, and the clocks came back reading exactly the same times. Is the theory definitely discredited? Suppose I assert ãIf something is a groundhog, then it is brown,ä and then someone shows me a black groundhog. My assertion seems definitely disproved. (Going back to the big question ãWhat is Science?ä, Karl Popper thought the essence of science was producing theories that were very falsifiable, that is, made very risky predictions.)
~ However, consider the following example. Part of basic chemistry asserts that boiling water, at standard atmospheric pressure, cannot rise above 100 degrees Celsius (before turning into steam). Now suppose Iâm in chemistry lab, and Iâve got some boiling water in a beaker. I want to test whether the theory in my textbook is true or not, so I take a thermometer out of the drawer and stick it into the beaker. The thermometer reads 120 degrees Celsius, and I am excited that I have made a new scientific discovery, and the scientists who wrote my textbook are wrong; water boils at 120 degrees, not 100. But is that the right inference to draw? Probably not: this thermometer could be malfunctioning. Or even further, the laws of thermal expansion relied upon in making thermometers could be wrong, i.e., mercury doesnât expand exactly the way we thought. Or there could be some unknown substance in the air in my laboratory that is raising the boiling point for water. There are many other background assumptions that must be made in order to effectively test the claim that water boils at 100 degrees, any of which might be false. This is known in philosophy as the Quine-Duhem problem. The moral of the story is that we often cannot test a specific scientific claim Îin isolationâ or by itself; we have to test whether it PLUS a lot of our background assumptions about the stuff are true. So we likely will not say we have disproved the scientific claim that water boils at 100 degrees; instead, we will probably say we have disproved our assumption that this thermometer is accurate and functional.
~ Real science makes interesting maneuvers when confronted with discrepancies between prediction and result. Uranusâs position was different than Newtonâs theory initially predicted. But these discrepancies led to the discovery of Neptune. But scientists have other reactions to discrepancies too. For example, Mercuryâs observed orbit was very slightly different than Newtonâs theory predicted. And scientists observed this about 50 years before Einsteinâs general theory of relativity. So there was a stretch of about 50 years where Newtonâs theory contradicted observed phenomena, but astronomers still did not reject Newtonâs theory. So sometimes there can be falsifying data, but scientists themselves still hold on to the falsified theory.
~ What does the following mean: ÎThe truth of a theory is underdetermined by the dataâ? Answer: ãmany competing theories can account equally well for the evidenceä (87). The example of the heated metal bar (Fig. 5.1) shows an example of underdetermination in action.
~ 2 versions of underdetermination thesis:
A. Any scientific theory is underdetermined by all observed phenomena that we have seen thus far (e.g., metal bar example).
B. Any scientific theory is underdetermined by all observable phenomena that a human being could ever possibly observe.
~ 3 responses to the underdetermination problem (= 3 views on what theories claim).
1. Instrumentalism: Theories are useful instruments for predicting future observations, and whether theyâre literally true or false is irrelevant. On this view, the idea of electrons shouldnât be taken seriously; itâs just a convenient tool for building working nuclear reactors or particle accelerators.
2. Empiricism: Theories are literally true or false, but we can never know which, so we should be agnostic about whether electrons exist.
3. Realism: Theories aim to be approximately true. Electrons really do exist. Such people must have some way to deal with the underdetermination problem, since this problem threatens realism (but not 1 or 2).