Suppose you have two ($\ell$-adic) Galois representations $V$ and $W$ and you know that each Frobenius-eigenvalue (at unramified primes) for $V$ is also one for $W$.
Is this enough to obtain a Galois-equivariant map between $V$ and $W$? Or does this only serve as a hope that there could be such a map?