Since no suggestions have been made, consider this.
Special relativity deals with spacetime with no gravity. The trajectories of bodies moving uniformly when plotted on a spacetime diagram are straight in the euclidean sense.
Einstein realized that objects falling freely in a gravitational field also experienced force-free motion even though accelerating, so their trajectories were also in a sense straight lines although in the euclidean sense they would appear curved. This requires that the SR line element ds2= dt2 - dx2 - dy2 - dz2 must become [itex]ds^2 = \sum g_{ab}dx^adx^b[/itex] where the metric coefficients gab can change from place to place, or even with time. The signature of Minkowski space is preserved.
If two masses are interacting through gravity their worldlines ( geodesics) will approach. The rate of approach turns out to given completely by the Riemann tensor.Thus the effect of gravity can be encapsulated in this tensor.
So, in short the motivation is to include gravity in SR while still maintaining SR in some sense.