This is precisely why, in mathematics, we generally use "inclusive" definitions, as illustrated in post #2. An equilateral triangle is a special case of isosceles triangle; the latter includes the former. That is, any equilateral triangle can also be called isosceles. See here:
An isosceles triangle is a triangle with (at least) two equal sides. In the figure above, the two equal sides have length b and the remaining side has length a. This property is equivalent to two angles of the triangle being equal. An isosceles triangle therefore has both two equal sides and two...
mathworld.wolfram.com
If we didn't define terms this way, then we would have to state many theorems in complicated ways, such as, "two non-collinear radii of a circle, together with the chord joining their endpoints, form an isosceles or equilateral triangle".
The same is true of quadrilaterals: a square is a special rectangle, which is a special parallelogram, which is a special quadrilateral. So a theorem about parallelograms applies in particular to squares. We can just say, "Opposite angles of a parallelogram are congruent", rather than "Opposite angles of a parallelogram, rhombus, rectangle, or square are congruent".
[You don't need to stop using "radius" to mean the segment. Use your book's or teacher's terms; this is a long-established usage. Check any dictionary.]