It can be a touchy topic whether people who do science or make use of outcomes of science are supposed to believe in science. I can think of at least three ways that we may or may not believe in science.

Knowledge tunnels

Imagine that we try to apply the current state of scientific knowledge to some everyday occurrence. It could be something that happens when cooking, a very particular sensation you notice in your body, anything that you have practical experience-based knowledge about, one of those particular phenomena that anyone notices when in some practical line of work. Every trade has its own specialised terminology and models of explanation that develops based on experience.

When looking closely, we will often find that the current state of scientific knowledge does not have a theory, model or explanation ready for us. This might not be because the problem is so hard to be unsolvable. It might be because the problem has simply been overlooked.

That means that if we want to maintain a world view where science has the answer, we don’t just need to believe in particular statements that science makes. We also need to trust that a scientific explanation can be developed for many cases where one doesn’t exist yet.

I can’t remember who I read who used this image (I think it might have been a mathematician): knowledge is not a solid body of work, but more like tunnels dug out of the mostly unknown. We don’t know most things, but we know some paths dug through this mass.

Whether you believe in it or not

There is another popular saying: the good thing about it is that it works whether you believe in it or not — said by Neil DeGrasse Tyson about science and attributed to Niels Bohr about superstition.

This seems to be true, not just about deep cases such as keeping one set of methodological assumptions for your scientific work and another, say a deeply spiritual religious one, for your personal life. It seems to be true about the way we expect scientists to work and communicate on an everyday level. Scientists do not need to be committed to every statement in their writing or every hypothesis they pursue.

Dang & Bright (2021) ask what demands we should put on individual scientific contributions (e.g. journal articles). They consider assertion norms, the idea that certain types of utterances are appropriate for a certain context, and argue that scientific contributions should not be held to these norms. That is, scientific claims need not be something the researcher knows, has justification for, or believes.

They give an example of a scientist who comes up with a new hypothesis, and performs some research that supports it. Overall, however, the literature does not support it. The scientist publishes a paper advocating for the hypothesis. In a few years, further research demonstrates that it is false. They suggest that in this case, the scientist has done what we expect scientists to do, even if:

• The claims are not accurate; they are false.
• The claims are not justified, in the sense that they do not have the support of the wider literature, and the data do not conclusively support the hypothesis.
• The scientist did not fully believe the claims.

They try to explain these norms of communication with reference to how the scientific community learns. They argue that this way of communicating facilitates an intellectual division of labour: different researchers consider different ideas and theories, that might at times be at odds, and therefore there is value to allowing them to them saying conflicting things.

The rejected norms were picked as the basis of our inquiry since they had been found plausible or defensible as norms for assertion, which is at least a somewhat related activity of putting forward scientific public avowals. So why this discrepancy? In short, we think this is because the social enterprise of inquiry requires that we allow people to be more lax in certain contexts than we normally require of individuals offering testimony, and through a long process of cultural evolution the scientific community developed norms of avowal to accommodate that fact.

Another reason, pragmatically, why scientists might not be fully committed to everything they’ve written is that, at least in natural science, writing often happens by committee. And not just a committee of co-authors, but also by reviewers and editors, who also take part in negotiations about what can be claimed in any piece of peer reviewed writing. Yes, there are people who like to say that every co-author needs to be able to take full responsibility for everything written — but the real lowest common denominator seems that none of the co-authors objects too much to what is written.

Weak citations

Finally, there appears to be two ways to think about claims when reading or writing scientific papers, that may explain that researchers can be at the same time highly critical of some claims, and somewhat credulous about other claims. Think of your typical introduction to an empirical paper, which will often have citations in passing to stylised facts that are taken to be true, but not explored at any depth — those ”neutral citations” that some think are shallow and damaging. Does that mean that scientists are lazy about the truth, just believing the first thing they see and jumping on citation bandwagons? Maybe. Because there is no other way they could write, really.

When we read or write about science, we have to divide claims into those that are currently under critical scrutiny, and those that are currently considered background. If we believe Quine, there isn’t an easy way to separate these two types of ideas, no logical division into hypothesis proper and auxiliary assumption. But we still need to make it, possibly in an arbitrary way, unless we want our research project to be an endless rabbit hole of questions. This does not mean that we have to believe the background claims in any deeper sense. They might be up for scrutiny tomorrow.

This is another reason why it’s a bad idea to try to learn a field by reading the introductions to empirical papers. Introductions are not critical expositions that synthesise evidence and give the best possible view of the state of the field. They just try to get the important background out of the way so we can move on with the work at hand.

Literature

Dang, H., & Bright, L. K. (2021). Scientific conclusions need not be accurate, justified, or believed by their authors. Synthese.