PeroK said:
I would say that the definition of the area of a rectangle is length times width. If you want to try to prove that, then you would need a more fundamental definition of area.
I would guess that the underlying theory of areas that
@logicgate is assuming is along the lines of an abstract
measure.
You start with an axiom that the area of a 1 by 1 square is 1. [Or, more generally ##k## with ##k## depending on the choice of units for length and area]
You add an axiom that says that if you join two geometric figures so that the two overlap, if at all, only at their boundaries then the area of the combined figure is the sum of the areas of the component figures.
This puts you in a position to evaluate the area of any rectangle with sides of rational length. Either by joining finitely many identical squares to make a rectangle. Or dividing a rectangle into finitely many squares. Or both.
To make the jump to sides of irrational length, we might choose to exploit a new axiom:
"A rectangle whose perpendicular sides both have a positive length has a positive area"
This puts you in a position to sandwich the area of a rectangle with irrational sides in between the areas of sets of rectangles with rational sides. Then you use the magic of convergent sequences, Dedekind cuts or least upper bounds to argue that if the rectangle has an area at all (as it must, given the axiom above), that area must be the real number given by ##\text{width} \times \text{length}##