Why do scientific theories often include entities that cannot be directly observed?

In the philosophy of science, the inclusion of unobservable entities in scientific theories is a central point of debate between scientific realists and antirealists (or instrumentalists).12

The Realist Perspective

Scientific realists argue that theories should be viewed as literal descriptions of reality. They contend that unobservable entities—such as electrons, nuclei, or quarks—are real components of the world. For realists, the explanatory power of these entities is essential; we can only understand phenomena like chemical bonding or the success of scientific predictions if we accept that these theoretical constructs actually exist and function as described. They often argue that the immense success of scientific theories would be "miraculous" if the underlying entities they postulate were not, at least approximately, true.21

The Antirealist Perspective

Conversely, instrumentalists argue that scientific theories are merely tools or "instruments" for generating predictions about observable phenomena. From this viewpoint, theoretical terms are not necessarily intended to refer to real, unobservable objects but serve as formal devices within a mathematical calculus. Some antirealists, such as those following constructive empiricism, suggest that science aims only for "empirical adequacy"—the ability to correctly predict observable outcomes—rather than uncovering the ultimate, unobservable nature of reality.12