In one of the Philosophy Colloquium lectures I attended in the past few weeks, "A Grue-some Riddle of Induction," the lecturer presented some of the difficulties associated with the instantial model of inductive reasoning and suggested in conclusion that reasoning from causality could be a more productive and reliable procedure. The example of the fictional attribute "grue," which is defined as an object's greenness when observed at any time before a given future date and its blueness when observed at any time after that date proves the uncertainty of instantial reasoning. Almost no one would believe that an attribute such as grue exists, but a purely instantial manner of reasoning is utterly unequipped to disprove it: is all of your predicates are drawn solely from what is observable and that "future date" in grue's definition has not yet arrived, you cannot prove that every object we see as green now is not in fact grue. Examining the riddle from the perspective of Aristotlean causality, however, seems to offer some tentative resolution, though it cannot, it seems to me, resolve the question completely.
Technically Aristotle speaks of four "causes" of being: material, form (what it is for a thing to be), goal (why it exists), and the efficient cause (the thing that causes a change [addition of an attribute to a preexisting being, creation of a being] to start). I don't find all of these equally helpful for analyzing grue, however, (though this is admittedly not something I've thought about exhaustively by any means) so I'll focus on the material cause and the efficient cause.
Considered from the point of view of the "material" of colour, grue's postion becomes more tenuous. Of course, one must first agree that colours are caused by varying wavelengths of light hitting the retina of the eye in order to come to some sort of agreement about this subject. Then, if one agrees that green and blue are caused by different wavelengths of light, one must also agree that grue would necessarily have to include in its definition a change in material. It cannot therefore be a simple substance of which green and blue are complex predicates, but must itself be a complex predicate in which exists a change from one simple attribute to another. Moreover, if there is a change, there must be some efficient cause to initiate that change. There must be some reason that the green changes to blue in order for something to be grue.
The biggest potential problem with this argument is the opportunity it allows for an opponent to claim that arguing from our knowledge of the way colours work scientifically is in fact arguing in some sense from instantial reasoning. This brings me to another point. It seems, inconviently enough, that you can't really escape from performing some degree of induction when reasoning about material things. You have to reason from what you see and experience in order to have scientific knowledge about a subject. The only substantial difference between science and mere casual observation is in the degree of rigour involved in the exploration of a thing. It is in this way, then, that I can take an example, assume some basic characteristics of this example, and then use Aristotle's method of analyzing that thing by its causes. Notice that in the previous paragraph I specifically stated that some basic agreement on the way colours work is necessary in order for a person to agree with and follow an Aristotlean argument against the concept of grue. Reasoning from instances is a necessary part of our rational life, misleading as it may occasionally be. The best we can do, I believe, at least in reasoning about material objects, is to be as careful as possible when making instantial claims, and to try to limit the number made. On that basis we can build up ideas about what we experience using other methods of reasoning, such as causality, but without agreement on the instantially-based assumptions which underly and form the foundation of our arguments, even the most well-reasoned discussion can be easily undermined and refuted.