Science and Knowledge
Introduction: Is Science a Reliable Source of Knowledge? Scientia vincere tenebras?
Despite common assumptions, it is not obvious that the typical claims of science are true or that they count as items of knowledge. Suppose something like the following characterization of the various scientific methods: collectively, they are means of obtaining propositional knowledge about the natural world via epistemic activities such as observation, making falsifiable hypotheses to explain the data of such observation, testing those hypotheses for confirmation or falsification, honestly reporting the results of the tests, proper repetition of the process, etc.
An Aporetic Triad
Each of the following statements is defensible, yet they are collectively inconsistent. If any two are true, the third is false. Which would you reject and why?
1. Scientific methodology enables propositional knowledge about the natural world.
2. Propositional knowledge entails epistemic certainty.[1]
3. Scientific methodology does not enable epistemic certainty about the natural world.
Consider some options for handling the problem.
Possible Solutions: Deny (1)
Suppose you deny (1). Yet you accept the common assumption of scientific realism: one of science’s primary purposes is to acquire knowledge, as Quine suggests in the first chapter of Pursuit of Truth by saying that understanding is a central aim of science. In this case, you’re faced with the claim that science is a failed enterprise, at least concerning that vital purpose.[2] Science aims at knowledge but fails to hit the target.
If you don’t like that conclusion, you could deny that science is aimed at obtaining knowledge. Indeed, some antirealists have said that science is not designed for acquiring knowledge. Rather, scientific methodology is a relatively unrefined tool of practical rationality, helpful for collecting evidence, generating rough explanations, pragmatic decision-making, and probabilistic prediction-making. It’s the best tool we have for such efforts (at least with respect to matters empirical), but it’s not wholly reliable, and there are many impurities we haven’t been able to remove from the scientific process.
But if you take this option (i.e., that science is not designed for knowledge or truth, but instead for pragmatic purposes), you’re faced with problems. Here’s one: it’s commonly said that science makes epistemic progress. What sort of epistemic progress, if not progress in obtaining knowledge about the natural world? You might say that science progresses with respect to making predictions and explaining natural phenomena. Yet one wonders what counts as progress if the predictions and the explanations are not based on knowledge. (For more on this questions, see the discussion of instrumentalism below.)
There’s another problem: Why is science epistemically better than, say, myth? Suppose you are an ancient Egyptian farmer living in Lower Egypt. You believe in the Egyptian pantheon. You accept that the fertility of the soil in the Nile Delta is explained by the fact your community worships the goddess Hathor at the local temple and that in return for regular veneration, Hathor blesses the soil. You believe that when the gods copulate, crops grow. You hold that the Nile overflows when the gods drink too much wine at their annual party. These and other myths enable you to explain the geographical patterns of the Nile River (which has a predictable flooding season), anticipate its flooding cycles, and make decisions about agriculture, general health, etc. You take these myths and their practical applications very seriously. Now, these myths are false. Yet they function for the ancient Egyptian as rough-and-ready assumptions which enable him to engage in existentially significant, probabilistic prediction about the natural world.
If science enables you to make practical decisions, predictions, and explanations, and yet science does not provide knowledge, then why is science better than myth-making? Is science a contemporary set of pragmatic naturalistic myths? Useful fictions? As a society, do we overvalue science? Perhaps science is more reliable than ancient myth for making predictions. Even so, if the typical claims of science are not items of knowledge, one might worry that science is merely a contemporary form of myth, though more sophisticated than ancient forms.
Possible Solutions: Deny (2)
The above might seem too costly. It’s crazy, you might think, to say that science is overrated or that it is merely a set of useful fictions. Science is a major aspect of human life. Our civilization relies on it. Its practical value is too high for us to doubt it. You might agree with the scientific realist that science is a trustworthy source of knowledge.
Instead, you can deny (2). If so, you should be prepared to show why the various arguments for epistemic infallibilism fail. This is no easy task. (See here for my summary of some arguments to support infallibilism.)[3]
Possible Solutions: Deny (3)
Would you deny (3)? Well, you’d have lots of explaining to do. For example, you’d need to show why it makes sense to believe that science generates propositions supported by evidence that is sufficient to meet the very high demands of epistemic certainty. Clearly, the many claims of science which have turned out to be false do not meet this high standard. Which specific claims of science are matters of epistemic certainty? What is the evidence for these claims? Is the evidence strong enough to guarantee the truth of the claims? I wouldn’t want the task of defending the denial of (3).
Possible Solutions: Back to (1)
The epistemic infallibilist has a way out of this aporetic stalemate. He could deny (1) and accept some version of scientific antirealism, such as instrumentalism. This wouldn’t be a rejection of science, nor of its rationality, but only of scientific realism.
Here is a good place for some definitions. Roughly, instrumentalism is the view that science does not provide us with knowledge about the objective natural world, nor do scientific terms refer to real objects in the objective world. Rather, science is useful for epistemic activities such as explaining natural phenomena, conducting laboratory research according to various operational stipulations, making predictions, and providing workable solutions to certain carefully constructed problems. On this view, science is useful; its instrumental value is high with regard to certain ends toward which we can progress, but knowledge is not one of those ends. Instrumentalism is a version of scientific antirealism, which is opposed to scientific realism, the view that science provides propositional knowledge and truth about the objective natural world, and that scientific terms refer to objects in that world.
In addition, the infallibilist might hold that the debate between scientific realism and antirealism is an example of a philosophical debate that we humans can’t resolve, and thus that the most reasonable position is agnosticism about the claim that science provides knowledge of the objective world. (There are other kinds of antirealism, and various forms of instrumentalism. I won’t go into these here.)
We have plausible arguments for scientific realism and for scientific antirealism. For example, the realist can argue that scientific realism is the best explanation for why scientific theories often work with respect to making predictions, explaining phenomena, etc. But the antirealist might object that scientific theories don’t need to be true or known to accomplish these ends. Moreover, one can provide arguments in favor of antirealism, such as an inductive argument sometimes called the “pessimistic induction.” Assume that the natural world is uniform and that our cognitive faculties are like those of past scientists. The history of natural science is a history of error; many positions taken as true by scientists turned out to be false, even though these positions were reasonable to accept at the time. Many claims of science thought to be items of knowledge turned out to be false and thus not cases of knowledge. Thus, probably, many present claims of science are false. Ditto about the future. Science, therefore, is not a reliable means to obtain knowledge.
Note: This is not necessarily a reason to deny the value of science. One could hold that science is generally useful for predictions and practical decisions concerning the natural world, even if science doesn’t provide knowledge, strictly speaking.
Note also: I am not attempting to defend antirealism here. I’m just saying that it is a defensible position – at least in some form. The advocate of epistemic infallibilism has some room to deny scientific realism by appealing to a defensible form of antirealism.
However, one doesn’t need to be an epistemic infallibilist to deny (1). Say Smith is an invariantist and a fallibilist about the nature of propositional knowledge. Yet he holds that the standards for knowledge are uniformly quite high, though they don’t demand infallibilism. Smith thus could hold that science generally fails to meet these lower (but still relatively high) standards for knowledge, and hence that science is not a reliable knowledge-producing enterprise.
Conclusion
To make a long entry short, it is not evident that the typical claims of science are true or that they count as instances of propositional knowledge. This is an important point to keep in mind for those who want to be intelligent and critical thinkers with respect to science, given that our culture tends to fall into the various intellectual vices of scientism.
[1] See here for what I mean by “epistemic certainty.”
[2] I’m assuming that Quine used “understanding” as a synonym of knowledge. Some philosophers argue that there’s a difference. I won’t pursue that point here.
[3] You might also read A Cumulative Case Argument for Infallibilism, by Nevin Climenhaga, in Skeptical Invariantism Reconsidered, Eds. Christos Kyriacou and Kevin Wallbridge, 2021.