I like the idea of staying within the dimensions of discourse, but I think you make my point with the classic example. When you draw a triangle (basically a 2-dimensional object) upon a sphere (a 3-dimensional object) you distort the triangle in the 3rd dimension. One may then measure the distance between any two corners of the triangle in one of two ways: The distance as measured along the surface of the sphere (the geodesic), and the straight line distance (on a Euclidean plane) through the sphere.
By extension then, if I "draw" a 3-dimensional pyramid on a 4-dimensional surface, would I then have three ways to measure distances between the points? Further, if we use the method of determining distances as the determinant of being flat or curved, then what about the measurements about the pyramid? Would it be flat, curved, or something else?
So, how do I deduce that the triangle is flat or curved from your information? And how do I extend that logic to a 4-dimensional object?