Good questioning. I don't know for sure what eventually lead us to adopt the paradigm of RLC circuits, but do know that the physics of electrical circuits deals with the way energy flows. In a circuit energy can be stored in the form of magnetic or electrical fields, and so the inductor and capacitor manage that capability. Energy can also be lost, and thus the resistor provides that means of modeling.
In a generalized way, the capacitor and inductor have a conceptual symmetry about how energy is stored. Why not the resistor regarding loss? Actually there is a conceptual element that fits the bill: the memristor. Although memristors have been realized, I don't believe they have been the subject of practical electrical circuits.
If we could conceive of other ways in which energy can flow in an electrical circuit, perhaps we could invent new components to consider.
I should also mention that the modeling in terms of RLC is idealized in another manner - as 'lumped' parameters. We deal with R, L, and C in our circuit models as pure elements of one or another. In real components and circuits things are not so pure. Capacitors have a little inductance & vice versa. Resistors have capacitance at high frequencies. Depending on what we seek to predict with our models determines whether we can get by with the lumped parameter models or if we need to model more precisely.