Van Inwagen's Nod to John Seddon

Much recent work in the epistemology of modality acknowledges the challenge of Peter van Inwagen's mitigated modal skepticism, especially as presented in his celebrated 1998 Phil. Studies paper, "Modal Epistemology". Furthermore, many have attempted to grapple with van Inwagen's arguments in that paper. Despite this, not many note van Inwagen's not-infrequent reference to John Seddon's 1972 Mind paper, "Logical Possibility" as an important source of his skepticism about modal claims remote from ordinary experience.  This is significant, since (i) a common criticism is that van Inwagen's mitigated modal skepticism isn't sufficiently warranted, since he fails to motivate a presumption against modal claims remote from ordinary experience, and (ii) van Inwagen alludes to the arguments in Seddon's paper as a core basis for such a presumption. The aim of this post is thus to exposit the main points of Seddon's paper. The hope is that it will help illuminate the nature and basis of van Inwagen's mitigated modal skepticism.

Scientific theories tell us about the world. As it turns out, however, our well-developed and well-confirmed scientific theories not only reveal the nature of objects and events in the world, but they also reveal some facts about the fundamental structure of reality that are at odds with the sorts of exotic possibilities philosophers are often interested in. For, first, they tell us that the features of things in the world are essentially interdependent and inter-related, in that kinds are often a part of an interlocking, nested hierarchy of entities (in some cases, e.g., in terms of species, genus, class, order, etc.). Given this fact about the structure of physical reality, the alteration of one feature of a thing requires an alteration in others, which in turn requires a change in others, all the way down to the categorical properties at the base: a change in even the “node” features often requires a change in the “root” features, so to speak.

Another fact that our scientific theories reveal about the structure of reality is that natural kinds occupy highly-specific niches, in that a given natural kind is typically “sandwiched” between two neighboring kinds, such that fairly minor alterations to a member of a kind would either make it go out of existence qua that thing, or land it in another natural kind. Given these well-supported facts about things and kinds – interrelatedness of kinds, interdependence of features of a given kind, and highly-specific niches for each kind – it follows that there is little leeway for things to be altered without either making them cease to exist qua that thing or kind of thing, or otherwise requiring widespread, perhaps global, changes in the world’s structure and its history. Call this the Little Leeway Thesis:

The Little Leeway Thesis: The structure of the world leaves little leeway for altering its occupants.

The Little Leeway Thesis poses a problem for exotic thought experiments. For such thought experiments typically represent exotic new sorts of things and states of affairs, or radical alterations to existing things or states of affairs. But if so, then the Little Leeway Thesis gives positive reason to doubt that the metaphysical possibility of such things and states of affairs. And if that’s right, then we have a defeater for any prima facie justification such thought experiments might have enjoyed. Below are two concrete examples in which our scientific knowledge conforms to, and thus supports, the Little Leeway Thesis, and thereby defeats the target possibility claims of high-flying thought experiments:

Case 1: Floating Iron Bar

Changing the properties of a bar of iron so as to make it float requires changing its gravitational properties, which in turn requires changing its specific density, which in turn requires changing either the size of its unit cells or the number and kinds of atoms within them. But to change the former requires a change in species, in which case it’s no longer iron. And to change the latter results in a change in the features determined by the lattice structure of its unit cells (cleavage, crystal habit, optical properties, etc.), in which case, again, we’re no longer talking about iron. Either way, then, changing the properties of iron so as to make it float results in non-iron.  In fact, iron is a transition metal, lying between manganese and cobalt on the Periodic Table. Thus, the range of possible densities for iron is highly restricted by these neighboring transition elements. But such leeway for changes in density isn’t anywhere near what’s required to make iron float. Thus, floating iron bars are impossible.

Case 2: Carnivorous Rabbits 

Changing the properties of a rabbit so as to make it a carnivore would require changing (e.g.) its dental features (it lacks meat-tearing canines) and its digestive features (it’s designed for rapid digestion of vegetation that results in small edible pellets for re-consumption -- digested meat from such a system would be toxic), etc. Further, The Principle of Natural Selection implies that organisms must be broadly adapted to their environmental niches. Now since carnivores are predators while rabbits are prey, a host of the rabbit’s current features are useful only for making it adaptable, and thus viable, as a prey and not as a predator. Thus, it will have to be changed in these respects as well so as to make it adapted to its niche qua predator. So, for example, we will have to alter its scut tail, which functions as a signal to the group in contexts of danger, and thus not an adaptive predator trait, which often requires stealth. It would also require a radical lowering of the rabbit’s birth rate, so as not to quickly deplete surrounding food resources. But at this point in our alteration of the rabbit, (but probably much before this point) it no longer seems to be a member of the class of lagomorphs. Indeed, such alterations in rabbits would require radical changes in the evolutionary history of the lagomorphs. Etc., etc. But if these things are so (and there are many more where those came from), then to change a rabbit from an herbivore into a carnivore would require changing it into something other than a rabbit, indeed it would no longer be anything even remotely resembling a rabbit (once again, a change in the nodes requires a change in the root). Thus, carnivorous rabbits are metaphysically impossible.

In short, our well-confirmed theories imply an interrelation among natural kinds, interdependence of the features of things within a given kind, and highly specific niches for each kind to occupy. And these facts give good evidence that there is little leeway for the possibility of a few features of a thing being altered without a change in thing or kind of thing – indeed, the resultant kind may be impossible, as there is good reason to thing there is no possible niche for it to occupy in many, if not most, cases. But if so, then we have good reason to think that thought experiments that imply otherwise – i.e., high-flying thought experiments -- are unreliable guides to possibility. 

But not only do we have good evidence that particular theories imply the Little Leeway Thesis within those theories; for we also have the basis of a track-record argument from the set of theories to the meta-claim that the Little Leeway Thesis applies to things in general.  But high-flying thought experiments represent things in ways that require the falsity of the Little Leeway Thesis. And if that’s right, then any justification such thought experiments might have conferred on possibility claims is defeated. We can put the problem the Little Leeway poses for high-flying thought experiments in terms of the following argument:

The Argument From Little Leeway:

1. The Little Leeway Thesis is true (supported by well-confirmed theories).

2. If the Little Leeway Thesis is true, then we have good reason to think that the things represented by high-flying thought experiments, though conceivable, are impossible.

3. If we have good reason to think that the things represented by high-flying thought experiments, though conceivable, are impossible, then high-flying thought experiments are unreliable guides to possibility.

4. Therefore, high-flying thought experiments are unreliable guides to possibility.

Now some of our scientific theories are very well-developed, and thus provide good evidence for what’s possible and what isn’t. However, when a theory isn’t well-developed, and thus doesn’t have enough information to say what’s possible and what isn’t, this doesn’t mean that more things are possible. Rather, it just means we’re ignorant about the modal facts here, as the theory is too underdeveloped to read off the modal facts from the theory.  

In short, scientific theories tell us that the structure of the world is tightly integrated, and in a way that implies the Little Leeway Thesis.  But the legitimacy of high-flying thought experiments as evidence of possibility requires the falsity of the Little Leeway Thesis. Therefore, we have reason to think that high-flying thought experiments aren’t good evidence of possibility.

It's beyond the scope of this post to defend Seddon's case for a presumption against modal claims remote from ordinary experience. Here it is enough to note that a case for such a presumption has been made, that van Inwagen appeals to it as grounds for such a presumption, and therefore that critics of van Inwagen have more work to do if they are to dismiss his mitigated form of modal skepticism.