Let us consider a single particle in an interferometer. If one of the two p
ossible paths is blocked and the particle is detected, we know that the par
ticle has followed the path which is not blocked. This would be an interfer
ence-free "which-path" information experiment. However, we no longer have a
n interferometer, since one path is blocked. An alternative is to interact
with the particle, but this would change its momentum and as a consequence
the interference fringes would disappear, as discussed by Feynman. We can a
lso consider two particles entangled in direction. Knowing the path followe
d by one of the particles, it is possible to know the path followed by the
other. On the other hand, when this information is erased, interference can
be observed. However, this is a two particle interference: no single parti
cle interference can be observed. Retrodiction experiments are also possibl
e, but these are not conclusive. Here we propose a much less intuitive expe
riment in which, without blocking one path or directly interacting with the
particle, it is possible to know the path which is being followed by the p
article in the interferometer. According to quantum mechanics, this is suff
icient to lose the single particle interference. The same idea can be used
to test the local pilot wave interpretation, to test quantum nonlocality un
der new conditions, and to devise an interferometer for a two-photon wave p
acket. This last result strongly suggests that there must be some connectio
n between the deBroglie wavelength of an N-particle wave packet and entangl
ement.