Because the base-emitter voltage of a BJT in its operating region will be affected by base-emitter current, and vice versa, changes to the base-emitter voltage of a given transistor will affect the collector-emitter current. On the other hand, the amount of base-emitter voltage change required to affect a given collector-emitter current change is often huge and unpredictable; it will vary enormously with temperature, aging, the phase of the moon, etc. By contrast, within a transistor's "linear" operating region, doubling the base-emitter current will roughly double the collector-emitter current. Not absolutely-precisely double it, but pretty close. Such behavior is far more predictable than the relationship between base-emitter voltage and base-collector current.
A FET or MOSFET by contrast doesn't have any gate current except for currents resulting from leakage or stray capacitance. Those currents aren't exactly zero, but manufacturers generally try to minimize them. As such, it's not really possible to characterize the transistor's response to different levels of gate current. The relationship between the gate-source voltage and the drain-source current isn't nearly as predictable as the relationship between base-emitter current and collector-emitter current in a BJT, but it's a still apt to be the most predictable way to characterize the device's operation (it's a lot more predictable and consistent than the comparable relationship on a BJT).