On the dynamics of Pérez Laraudogoitia's supertask (Q1977822)

In two papers \textit{J. Pérez Laraudogoitia} [Theoria, Segunda Epoca 11, 85-89 (1996; Zbl 0917.03002); Br. J. Philos. Sci. 48, 49-54 (1997; Zbl 0979.70004)] described the remarkable time evolution of a system consisting of a countably infinite number of point particles situated in a finite interval along the \(x\)-axis. One of the boundary particles moves towards the others, and an infinite number of perfectly elastic collisions occur in a finite time. The completion of the time evolution of such a system constitutes a ``supertask'': the performance of an infinite number of individuated actions in a finite time. Pérez Laraudogoitia argued that, at the completion of the supertask, all the particles are stationary, thus violating the conservation of energy. He further concluded that, at some indeterminate time, a system consisting of an infinite number of initially stationary particles can spontaneously ``self-excite'', i.e. can exhibit motion at some subsequent time. Here, in order to explain the strange behaviour of the system under consideration, the authors use an embedding in a Hilbert space where the equations of motions become singular, hence non-Newtonian, at the time the supertask is scheduled for completion. This fact leads the authors to examine wether any self-excitation or indeterminism is possible within the context of autonomous differential equations. As a result, the global formalism based on Hilbert space embedding recognizes a singularity caused by the infinite density of particle collisions in the neighbourhood of \(t=1\). The resulting motion remains singular and non-Newtonian. A Newtonian system, necessarily non-singular, will not move from its initial stationary state, so the supertask is not time-reversal invariant and self-excitation does not occcur. According to the authors, their global analysis makes it quite clear that, unlike the thermodynamic limit in statistical mechanics, the infinite set of the supertask at \(t=1\) is not the limiting case of a finite system. In addition, its state at \(t=1\) ``cannot be deduced from the motion at earlier times. Newtonian dynamics has behaved as it should. For all times \(t < 1\) it tells us that energy is conserved. But for time \(t=1\), about which it cannot speak, it remains silent''. From the mechanical point of view, this results is obvious, independent from how the system is evaluated (Hilbert space or not, collisions via Dirac-distributions or ``smoothened''). A control volume consisting of \(n\) particles, where an addidional particle goes in but none comes out, cannot be energy conserving. Moreover, since there is no damping nor any other energy consuming device considered, but only a sequence of perfect elastic impacts, the system cannot be Newtonian at all, no matter how large \(n\) is (and neglecting the fact that for sufficient high \(n\) the sequence of individual and separated impacts does not make much sense). If the whole motion is stopped at \(t=1\), this can only be achieved by a ``deus ex machina'' which is then, indeed, non-Newtonian and possibly able to create any unpredictable motion.