"All triangles are isosceles" is a famous geometric fallacy (see below). Unlike many other fallacies its flaw is subtle and hard to spot, so it is often used as a cautionary example against the "danger in diagrams", e.g. in Greenberg's text, but always, it seems, without attribution. It feels so much in the spirit of Euclid that one would think it may go back to antiquity, and Euclid did write a book of fallacies called Pseudaria, now lost. But no, only four geometric fallacies survive from antiquity, and "all triangles are isosceles" is not one of them.
So where did it come from, who came up with it? Or if that can not be traced, what is the earliest known appearence? I am also interested in tracking the origins of other geometric fallacies.
"Proof":
If the angle bisector at A and the perpendicular bisector of BC are parallel, then ABC is isosceles. On the other hand, if they are not parallel, they intersect at a point, which we call P, and we can draw the perpendiculars from P to AB at E, and to AC at F. Now, the two triangles labeled "alpha" in this figure have equal angles and share a common side, so they are equal by angle-side-angle. Therefore, PE = PF. Also, since D is the midpoint of BC, the triangles labeled "gamma" are equal right triangles by side-angle-side, and so PB = PC. From this it follows that the triangles labeled "beta" are right triangles with equal leg and hypotenuse, so equal to each other. Thus, we have BE+EA = CF+FA, meaning the triangle ABC is isosceles. (From mathpages)