We argue that for a single particle Bell's inequality is a consequence of n
oncontextuality and is incompatible with statistical predictions of quantum
mechanics. Thus noncontextual models can be empirically falsified, indepen
dent of locality condition. For this an appropriate entanglement between di
sjoint Hilbert spaces pertaining to translational and spin degrees of freed
om of a single spin-1/2 particle is invoked. (C) 2001 Elsevier Science B.V.
All rights reserved.