The distinction between abstractions and concrete representations is not as emphasized in mathematics as it is in computer science, making it easier to conflate the two. This can happen particularly when studying foundations, where it’s very tempting to view their encodings as the ultimate authority on what mathematical objects really are. We can avoid this trap can by keeping in mind an object’s specification, i.e. it’s essential properties and how it interacts with other objects (Wells 1995). This helps distinguish them from their concrete representations.
Here are some examples of mathematical abstractions and their representations:
Math objects are encoded as sets of sets of … sets in Zermelo-Fraenkel set theory (e.g the Kuratowski construction of an ordered pair is ).
Probabilistic events are modelled by sample spaces.
An abstract vector space $V$ is represented by a concrete system of coordinates $\phi: V \rightarrow {\bf R}^n$ for some basis of $V$.
An abstract group $G$ is represented by isomorphisms on some space $X$, $\phi: G \rightarrow \hbox{Aut}(X)$.
The abstract number systems $N$, $Z$, $Q$, $R$, $C$ are represented by concrete numeral systems, for example the decimal and binary numeral systems.
(Tao 2015)
References
Wells, Charles 1995, Communicating Mathematics: Useful Ideas From Computer Science
Tao, Terence 2015, 275A, Notes 0: Foundations of probability theory