What is science for? A common answer is that science aims to discover the truth about the world. From this perspective, known as scientific realism, a good theory is one that accurately describes reality, including the parts we can't see. Scientists talk about electrons, gravitational waves, and black holes. A scientific realist believes these things actually exist, and that our best theories are our best descriptions of them.

This seems straightforward. But in 1980, philosopher Bas van Fraassen proposed a different view in his book The Scientific Image. He argued that science has a more modest, practical goal. His philosophy, called constructive empiricism, challenges the idea that we need to believe in the literal truth of our theories.

Constructive empiricism suggests that the goal of science isn't to find the capital-T Truth, but to develop theories that are "empirically adequate." This is the core concept of van Fraassen's philosophy.

In other words, a theory just needs to get the observable stuff right. It must align with all observations and experiments we have made in the past, present, and future. This is what van Fraassen calls "saving the phenomena." The theory has to account for everything we can see and measure.

This leads to a crucial distinction between belief and acceptance. A scientific realist believes a successful theory is true. A constructive empiricist simply accepts it. To accept a theory means to commit to using it to explain and predict observable phenomena. You use it as your guide to the empirical world.

What about unobservable entities like quarks or neutrinos? The constructive empiricist remains agnostic. The theory of quantum mechanics is incredibly successful at predicting what we observe in particle accelerators. For the empiricist, that is enough to accept it. Whether quarks are "real" is a metaphysical question that science doesn't need to answer.

Van Fraassen's critique of scientific realism is that it asks for too much. It requires a leap of faith from a theory's empirical success to its literal truth. Why, he asks, should we believe in the unobservable parts of a theory when all we can ever test are its observable consequences? The assertion that a theory is empirically adequate is a much weaker, and therefore more defensible, claim than the assertion that it is true.

Constructive empiricism offers a pragmatic way to think about science. It focuses on what makes scientific theories useful and successful—their ability to correctly describe the world we experience—without getting entangled in metaphysical debates about what might lie beyond our observational reach.