In this paper we present a novel probabilistic protocol for path discovery in Mobile Ad Hoc Networks (MANETs). The protocol implements what we call a polarized gossiping algorithm. While in the classical gossip algorithm each node forwards a message with the same probability, our proposal is characterized by a variable gossiping probability, which is high enough only for sustaining the spreading process towards the destination. The gossiping probability of a node is determined by the difference between its proximity to the destination and the proximity to the destination of the node from which the message was received. Differently from other proposals no external location service support, e.g., via GPS, is required. Rather, the proximity is estimated from the ``inside'' of the network using periodic beacons for determining the time elapsed since a node met the destination and the dwell time of a node with the destination. These information are then exploited by nodes to modu...