In the previous post, I explained why, assuming that there is necessity in the world, relations of necessity are perfectly knowable on the basis of experience, without any recourse to abductive reasonning. That certainly works in favour of an endorsement of natural modalities for the empiricist: at least, epistemic arguments against it are not conclusive. But still, an empiricist could resist this kind of commitment by assuming that there is no necessity in the world from the start. She could interpret modal discourse in a pragmatic way, rather than assume that modal statements have truth values. In this post, I want to explain why she shouldn't.
As I said in the previous post, I don't think that the existence of necessity in the world itself (as opposed to relations of necessity if there is necessity in the world) can be confirmed or disconfirmed by experience. However, no position really comes without some general metaphysical framework, and this includes all empiricist positions: the idea that all knowledge comes from experience is not itself confirmed or disconfirmed by experience. The idea that our theories will continue to be empirically adequate in the future is not confirmed or disconfirmed by experience. The empiricist is only willing to assume the minimum necessary, and to refrain from speculations. One could frame this as some kind of transcendental argument: we need some basic assumptions to make sense of the world, but we shouldn't assume more.
This is exemplified by van Fraassen's defense of constructive empiricism. He does not say that his position is itself confirmed by experience, but he rephrases empiricism as a position about the aim of science, as a collective endeavour: its aim, according to van Fraassen, is to produce empirically adequate theories (not true theories). A scientist can well be a realist, but needs not be to be a good scientist. However, a scientist must at least assume that our theories are empirically adequate, and will continue to be in the future. Now if one thinks that scientific practice is a rational activity, one should also believe that our best theories are empirically adequate. This is a minimum.
I think this is the correct way of thinking about epistemological positions. But I think that it supports modal empiricism rather than constructive empiricism.
The notion of observable
A first reason to think so is that constructive empiricism is quite unstable. Van Fraassen says that it is the position according to which our theories are empirically adequate for all observable phenomena in the universe (past, present, future, observed or not). He also says that modal statements have no truth values: there is no fact of the matter about what could have happened, it's just a way of talking about the models of our theories (at least when the claims are supported by scientific theories). But then, what does "observable" means? Doesn't it mean: something that we could observe?
This criticism was raised by several commentators (for example Ladyman). Here is van Fraassen's response: that a phenomena is observable implies that we could observe it, but it doesn't mean that. It's only that we can derive pragmatic assumptions from the fact that a phenomena is observable. However, "observable" is not itself a modal notion. It is an objective feature of phenomena. Van Fraassen doesn't think that it is an a priori notion, or a linguistic notion (as logical empiricists did, and were criticised for), and endorses semantic realism: we really talk about unobservable entities as we do about observable ones, it's just that we cannot know about them. He claims that our scientific theories can well help us know which kinds of phenomena are observable or not.
I am not convinced by this. Van Fraassen doesn't want the linguistic notion of the logical empiricists, but I'm not sure we can abandon it, and maintain that there is still an observable/unobservable distinction, without either being a realist about scientific theories, or endorsing modalities.
Indeed, if "observable" is not a priori, and if scientific theories themselves (from cognitive sciences, or biology) tell us about what is observable or not, how can we make sense of this notion while maintaining that these theories are not true? Imagine a theory T that implies "phenomena of type X are observable". If we are constructive empiricists, we only believe that these theories are true in what they say about observable phenomena. Then what we actually believe about T is a tautology: that "phenomena of type X, if they are observable, are observable".
That's not the end of the story, since "observable" also implies modal statements, such as "it could be observed in the right conditions". That's an escape for the modal empiricist: she needs not be a realist. If T is modally adequate, it has informative modal consequences about non-observed phenomena. But for the constructive empiricist, things are more problematic: if these modal statements have no truth values, then the only concrete implications are that "phenomena of type X, when in the right conditions, are observed" and so all we believe about empirically adequate theories is that they are true about what is observed in the right conditions. Empirical adequacy would boil down to our theories being correct about what is actually observed only. That is much more restrictive.
An empiricist could well restrict her empirical adequacy to what is actually observed (not all observable phenomena), but according to van Fraassen, this is too restrictive to make sense of scientific practice. That's where the aim of science enters the picture.
If the aim of science was to produce theories that are empirically adequate for observed phenomena only, what would be rational? Should we test various aspects of our theories, various situations? Or should we refrain from observing anything new? The latter would certainly be rational: the best way to ensure that our theories account for what is observed is not to observe anything. But that's not what scientists do: they keep on producing new situations to know if the predictions of the theory is correct. The reason, according to van Fraassen, is that the aim of science is to produce empirically adequate theories for all phenomena, not only those actually observed. Then, of course, making new tests makes sense: we just want to be sure that our theories correctly accounts for remote phenomena, so we attempt to observe them, or to reproduce them in our laboratories.
Extending empirical adequacy to all observable phenomena is required if one wants to make sense of scientific practice. But as we saw, it's hard to make this move and to maintain that modal statements have no truth value.
Furthermore, I think that van Fraassen's argument can be extended to defend modal empiricism. Here it is: if the aim of science was to produce adequate theories for actual phenomena (not possible ones), what would be rational? Should we implement situations that would not occur naturally, to test our theories? Or should we only test our theories for phenomena that, we know, do occur naturally, somewhere in the universe? The latter would be perfectly rational, but scientists do implement artifical situations to test their theories. For example, when Alain Aspect makes an EPR experiment, he does not attempt to know what happens somewhere in the universe: he is only trying to test some possible consequences of the theory. The fact that an EPR experiment could occur naturally or not seems irrelevant. That wouldn't make sense if the aim of science was to account for actual phenomena only, but if the aim of science is to produce modally adequate theories, this makes perfect sense: we will test all possibilities regardless whether they would occur naturally or not.
I think the right way of understanding the role of intervention is to view it as the implementation of possibilities. When scientists control parameters in order to reveal causal relations, they implement various possibilities in order to acquire modal knowledge. Compare Kepler's laws, which derive from the observations of planets, and Newtonian physics, which concerns manipulable objects as well as planets. The former is more easily interpreted as "phenomenal laws" describing a pre-existing harmony in the solar system, and the latter as "fundamental laws" describing causal relations between objects (associated with forces). Causal relations are modally loaded. Arguably, the possibility of manipulation plays an important role, even though the theory can later be extended to non manipulable objects. There are area of science where intervention is not possible, for example cosmology, but these are the area where modal knowledge is problematic: it's not easy to differentiate facts (about the initial conditions of the universe) and laws in cosmology.
So the best way to make sense of experimental interventions is to assume that scientists are attempting to gain modal knowledge, by testing possibilities. For this reason, I think that modal empiricism is the best way to make sense of scientific practice as a rational activity. Scientists need not be realists to be good scientists, but at least, they must assume that our best theories are modally adequate, and that the aim of science is to produce theories that would work in all possible situations. Technology is certainly the most obvious manifestation of the success of science, and we all believe that our technologies work well, and will continue to do so. Now a good technology is one that would work whatever the contingency: we don't know what will happen in the universe, do ideally, it should work in all physically possible situations.
Conclusion: why be sceptical?
I defended that endorsing natural modalities is the best way to make sense of scientific practice. Another reason to endorse them is that modal discourse is ubiquitous, and that there is no a priori reasons to be sceptical about it. Attempts to reduce meaning to relations between actual observations fell out of fashion because of important criticism, including, for example, indispensability arguments regarding theoretical terms, or the impossibility to reduce dispositional terms, and contemporary empiricists are generally semantic realists: they think we should "interpret theories literally", even if we don't believe they're true. Why, then, not apply this attitude to modal discourse? Isn't modal discourse indispensable as well (in explanations, for example, or for understanding dispositional terms)? The in principle impossibility of modal knowledge could be a reason to be sceptical, but the same would apply to unobservable objects. Furthermore, I showed in the last post that it is possible to acquire modal knowledge by induction. Then apart from dogmatism, there is no reason for an empiricist not to be a modal empiricist.
If we take stock, we'll see that lots of arguments work in favour of modal empiricism as the best epistemological position available: it makes sense of scientific rationality, better than standard empiricism does, while realism is not required, and it can respond to the no-miracle argument, while retaining the advantages of empiricism when it comes to theory change (structural realism does this as well, but it collapses to empiricism). It can also resist arguments of underdetermination.
In future posts, I will just assume that modal empiricism is true, and say more about how we can interpret scientific theories (and modalities) if we are a modal empiricist.