Some metaphysical truth claims seem undetermined by all the naturalistic considerations we can adduce in their favour. Does this imply that the ultimate metaphysical facts are fixed pragmatically by our choice of conceptual scheme or ideology? Or are cases of underdetermination symptomatic of a deeper problem with the idioms and methodologies used in naturalistic metaphysics.
Here’s an example of underdetermination taken from the end of my article Sonic Art and the Nature of Sonic Events. SANS argues for a located event theory of sound (LET) but suggests, in closing, that our choice between two versions of LET is a) empirically significant (since it entails commitment to different claims about where and when sounds occur) but b) metaphysically unmotivated. According to Casey O’Callaghan ‘relational event view’ sounds are periodic disturbance of a medium. Thus sound events are caused by vibration events. However, according to Roberto Casati and Jerome Dokic’s account, sounds are vibration events.
While Casati and Dokic’s view implies that there is a sound located in a vibrating tuning fork contained in an evacuated jar; O’Callaghan’s implies that there is none. Either can cite support from folk intuitions. Thus most folk would probably judge that there is no sound in the evacuated jar. However, were the air in a jar containing a vibrating tuning fork to be regularly evacuated and replenished they might well perceive this as an alteration in the conditions of audition of a continuous sound, rather than the alternating presence and absence of successive sounds (perhaps this is one for experimental philosophy fans).
Thus the two theories trade truth values between particular judgements, while doing seemingly equal justice to the conceptual framework within which sounds are identified, located and sorted.
Similarly, while there are decent phenomenological, acoustic and psychological grounds for supporting event theories – as opposed to proximal theories (which identify sounds with sensory affects) or medial theories (which identify them with the waves which generate these affects for humans within the 12 and 20,000 hz range), it would be self-deceiving to claim that this bundle of folk intuitions and empirical assumptions amount to anything conclusive. While we might insist that one of these accounts must be the true one regardless of underdetermination, it is hard to see why we should embrace bi-valence here unless there is also conclusive evidence that weighs in favour of one or the other account.
Does this imply that there is simply no fact to the matter here and does this slack extend to other similar cases underdetermination (e.g in debates around personal identity or vagueness)?
However, accepting alethic as opposed to epistemic underdetermination just inserts a supplementary metaphysical claim alongside the other candidates: namely, that it is a fact that there is no fact to the matter about when and where sounds are. So embracing relativism or ‘undecidability’ offers no way out here. It seems, then, that if there are different ways of parsing the ontological pie here, it is because there is a sonic reality that transcends the above mentioned idioms. They are not wholly wrong, clearly, but they all somehow fail to capture the ‘being’ of sound.
Which raises the question? Well, if sound can be validly described as an located event, a sensory affect, or wave motion, but is really none of the above, what is it? And how can we arrive at an adequate metaphysics of sound?
The short answer, I think, is that an adequate metaphysics needs to do more than reconcile these claims but show how they adequately express an underlying auditory reality. Thus, as Casati and Dokic show, it is already possible to accommodate some of the intuitions associated with the medial theory within the LET by allowing that sound waves are the means by which our senses acquire information about sound events. This isn’t a metaphysical concession as it stands since sounds are still parsed as located events rather than as the transmission of acoustic energy within a medium, and the ontological relativity remains.
A better metaphysical account would need to go some way to reconciling LET with medial accounts insofar as they each capture something of the underlying metaphysics of sound which each either obscure or over-simplify. Both medial theories and LET identify sounds with events or processes but locate these differently in the total spatial and temporal ordering of events.The ontological difference, here, remains factual rather than formal. Otherwise put, this may be a problem specific to the event-being of sound. It is, perhaps, because sounds are events that they smear across locative boundaries. I have argued in SANS that this problem arises also when we come to demarcate containing systems in which located events are located. Should we locate count the computer within a digital-audio system as part of the resonator ( it is causally necessary for the alternating current that produces sound in the speaker) or count only the minimal supervenience base for the sound under any possible set of physical laws (e.g. universes where the same effect could occur as a result of ritual magic)? Although a spatially smeared-event of this kind is still an event, it is one that already defies adequate representation. This looks as if I am positing a kind of noumenal event behind the conceptually articulated auditory phenomenon, and perhaps that is the most honest position to adopt here. Sound is clearly not ‘unrepresentable’ or unknowable in the way that the Kantian noumenon is usually construed; but , equally, it seems to elude the conceptual partitions employed by analytic metaphysicians of sound.