"Abstraction" is a term that bears multiple meanings. Since the later half of the 20th century, however, the main interest of philosophers in abstraction has been related to the use of abstraction principles, for the role they play in the foundation of mathematics: given a suitable abstraction principle and some appropriate definitions, it is possible to derive second-order Peano's Arithmetic. This result set the field for a rebirth of the view that arithmetic knowledge is analytic, a view which is tempting for its epistemological implications. A couple of problems this view faces are that of justifying the accepting of an abstraction principle as being analytic, and that of explaining, given the analiticity of an abstraction principle, just how it can be that an analytic truth implies the existence of objects (indeed of infinitely many). In trying to solve the latter, Linnebo has introduced the concept of thin objects, which are lightweight objects, in the sense that nothing (or at least not much) is required for them to exist. Fine also provides a defense of abstraction, but of a different sort, closer to the original conception of it, by employing his theory of arbitrary objects (which are objects possessing all the properties common to a certain range of objects). Following Linnebo's suggestion and Fine's defense of Cantorian abstraction, I intend to argue that a proper conception of arbitrary objects may shed a light upon all the different sorts of abstraction - and, most importantly, upon abstraction principles.