How does constructive empiricism differ from instrumentalism?

Question

Constructive empiricism is the belief that science aims to give theories which are empirically adequate (i.e. agree with observables), and accepting a theory means believing that it is empirically adequate.

Instrumentalism is the belief that the value of scientific theories is measured by the extent to which they help us make empirical predictions.

Is there any difference between the two? They both would reject the idea that science can say anything meaningful about unobservable entities. They also both seem to agree that there is some sort of reality that we have access to with regard to observables, and that science is valuable insofar as it helps us understand these observables. I don't see any difference.

Answer 1

There are different versions of instrumentalism, but a typical version states that the only factual content that scientific theories have is observational (or empirical). In other words, all the non-observational statements entailed by a theory are to be understood non-literally; for instance, all non-observational statements have to be translated into the observation language first before they can be understood literally.

Constructive empiricism, in contradistinction, states that scientific theories are to be understood literally and thus have factual content beyond observational statements. However, there is no way of finding out whether the non-observational content of scientific theories is true.

§1.2 in the Stanford Encyclopedia entry on constructive empiricism (https://plato.stanford.edu/entries/constructive-empiricism/) gives a quick summary of this distinction.

Answer 2

That's a nice question. Not an expert, so make sure you check the answer.

I believe the difference is mainly one in semantical difference about what scientific statements/theories should be taken to be about.

They both would reject the idea that science can say anything meaningful about unobservable entities.

(Firstly, I'd say that "instrumentalism" isn't a completely clear label. So this might not go for all stances that get call that.)

Generally though this is not true or not precise enough. Instrumentalism would say that statements or theories about unobservable entities aren't really truth-apt or are metaphorical. Whereas constructive empiricism would see no semantic issues but have some epistemological reservations.

This paragraph of a SEP article is relevant for your question. More there.