Michela Massimi has a long article at Aeon defending scientific realism.
The time for a defence of truth in science has come. It begins with a commitment to get things right, which is at the heart of the realist programme, despite mounting Kuhnian challenges from the history of science, considerations about modelling, and values in contemporary scientific practice. In the simple-minded sense, getting things right means that things are as the relevant scientific theory says that they are.
…We should expect science to tell us the truth because, by realist lights, this is what science ought to do. Truth – understood as getting things right – is not the aim of science, because it is not what science (or, better, scientists) should aspire to (assuming one has realist leanings). Instead, it is what science ought to do by realist lights. Thus, to judge a scientific theory or model as true is to judge it as one that ‘commands our assent’. Truth, ultimately, is not an aspiration; a desirable (but maybe unachievable) goal; a figment in the mind of the working scientist; or, worse, an insupportable and dispensable burden in scientific research. Truth is a normative commitment inherent in scientific knowledge.
Scientific realism is the belief that scientific theories describe the real world. For a realist, when general relativity talks about space being warped, there is actually a thing out there being warped. When particle physics talk about an electron, it is referring to a real thing out there that definitely exists.
The main alternative to scientific realism is instrumentalism, which holds that scientific theories are frameworks, tools, for predicting observations, with no particular guarantee that they describe reality. Specifically, the reality of any statements the theory makes beyond predictions that can be tested, are suspect. Often the testable predictions arise from the theory’s mathematics whereas the non-testable ones arise from the accompanying language narrative, but there can be non-testable mathematical predictions and testable narrative ones.
I’ve written before that I actually find this distinction invalid, because our understanding of reality is itself just another mental model. Our brains build models of the world, the accuracy of which we can only measure by how well they predict future experiences. I can see no other measure of truth. (If you can think of another one, please let me know in the comments.)
Emotionally I am a scientific realist. I try to maintain a commitment to truth and do see science as the pursuit of truth. Indeed, I think it’s the best tool we have for learning about the truth. But part of being faithful to truth means acknowledging the limitations of what we can know about it. And my questioning of the distinction between realism and instrumentalism makes me, to a hard realist, a non-realist.
So, intellectually, I’m an instrumentalist. All we ever get are conscious experiences. A successful scientific theory predicts future experiences. All statements about objective reality are theoretical. But the nature of that reality beyond what can be tested about our theories, are unknown.
Which brings me to this statement in Massimi’s article:
Constructive empiricists, instrumentalists, Jamesian pragmatists, relativists and constructivists do not share the same commitment. They do not share with the realist a suitable notion of ‘rightness’.
As an instrumentalist, I say baloney! Massimi mischaracterizes all alternatives to realism as accepting arbitrary notions. While there are details about James’ pragmatic theory of truth I’m unsure about, I find the overall idea sound. What is true is what works, what enhances our ability to make accurate predictions.
A commitment to this standard is just as much a commitment to “getting it right,” but with the added benefit that adherence to it can be tested. Ultimately instrumentalism, and similar philosophies, are simply epistemic humility, which I think is crucial for actually getting it right.
Unless of course I’m missing something?