It's not a question of inconsistency, it's a question of validity.
Physics is a mathematical model based on certain axioms. The aim of this model is to provide predictions about the real word. "What will happen if I throw a stone? Where will it land?" - these are questions about the future. Physics will then use a certain framework of definitions (mass, force, acceleration, rigid object...) and axioms (such as Newton's laws) to give you some numbers and curves; and then if you throw the stone corresponding to this model, you will know what happens before it actually happens.
Every model has a certain range of validity. If you give it parameters outside of this range, it will fail to give you predictions. For example, the singularity in the center of a black hole is a point outside of the mathematical validity of the Einstein equation; therefore, general relativity cannot predict what happens in that point, since it cannot be applied to that point.
When you are talking about an axiomatical immovable object and an axiomatical unstoppable force, you include two limit cases in the theory:
- Irresistible Force: F = lim(F --> infinity) - an infinite force.
- Immovable Object: m = lim(m --> infinity) - a rigid object with infinite (inertial) mass.
Now, what happens when the two meet each other? Your theory will certainly involve an axiom regarding this. In fact, since you're already talking about forces, you implicitly decide to use the Newtonian axioms (since they define forces); so we should stick to Newton's laws. Within this model, a certain parameter named acceleration (a) is the aim of prediction (and derived parameters such as speed etc., all of secondary interest for us now); and the axiom called Newton II gives the relationship between this and our two previously defined parameters, the mass m and the force F:
But what does this model do when both F and m are infinite? Well, we get
- a = lim(F -> infinity) / lim(m -> infinity)
But calculus - the mathematical framework this model uses - tells us that this quotient can be anything! As such, "prediction" we get ist a can be literally anything. Our model is unable to make predictions about the movement of the object!
So what does this mean? Is our model broken?
No - it works as it should. It has a rigorous mathematical background, and its predictions work for almost all cases.
But this specific case is an edge case where this theory is simply no longer valid. Such cases, where we end up with unstoppable forces acting on immovable objects, are where the theory breaks down - we simply can't get useful answers by using it. And to overcome this issue, one must find a new theory that includes the case in its validity range.