This is a memory corruption error and could result due to the heap corruption. To further investigate this problem set the event in your ini.ora and bounce the instance. Such that when it were to occour next time it would generate a detail trace under udump. Then contact oracle.
event="10235 trace name context forever, level 1"
My personal undestanding of this problem is that the corruption could have occoured at Call heap, PGA heap, session heap or SGA heap. The event trace can only pin point the exact location. On some occassions increasing the java_pool_size and shared_pool_size would help to solve the problem.