13.07.2015 Views

Mancosu - Philosophy of Mathematical Practice (Oxford, 2008).pdf

Mancosu - Philosophy of Mathematical Practice (Oxford, 2008).pdf

Mancosu - Philosophy of Mathematical Practice (Oxford, 2008).pdf

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

what structuralism achieves 361This says nothing to specify what polynomials are beyond the paradigmaticallyuninformative assertion that one <strong>of</strong> them is called X. ItdoessayR[X]isaring,so there are other elements, like X 2 = X · X,and3X 2 + X, described by theiralgebraic relation to X. Intuitively u a evaluates these at X = a. That is, sinceu a is a ring morphism:u a (X 2 ) = a 2 and u a (3X 2 + X) = 3a 2 + a and so on.But nothing in Fact 1 specifies the elements <strong>of</strong> R[X] as sets. It specifies whereR[X] fits in the pattern <strong>of</strong> rings and morphisms.Lang states Fact 1 in italics without calling it a definition or theorem oranything else (Lang, 1993, p. 99). If pressed, he would probably call it atheorem. Indeed either one <strong>of</strong> his <strong>of</strong>ficial definitions <strong>of</strong> polynomials can bemade rigorous and used to prove the Fact. But Lang does not like using themin pro<strong>of</strong>s. Fact 1 itself gives a rigorous definition <strong>of</strong> R[X] up to isomorphismwhich he does like to use.Philosophers usually define an isomorphism <strong>of</strong> structured sets as a one-tooneonto function preserving structure in each direction, as in (Bourbaki,1939). Mathematicians today use a simpler and more general definition, whichhowever agrees with Bourbaki’s in this case (Lang, 1993, p.54). For the sake<strong>of</strong> generality it speaks <strong>of</strong> morphisms instead <strong>of</strong> structure preserving functions: Everystructure S 1 has an identity morphism 1 S1 : S 1 → S 1 .Amorphismf : S 1 → S 2is an isomorphism if it has an inverse, thatisamorphismg : S 2 → S 1 whichcomposes with f to give identities:2fg1 g f1 12 21 11 2The need for generality will appear in Section 13.3.1 and the accompanyingcase study. For now, note how directly this concept <strong>of</strong> isomorphism suits Fact 1.Theorem 3. Suppose a ring R[X] and function c satisfy Fact 1, and so do anotherR[X ′ ] and c ′ . Then there is a unique ring isomorphism u : R[X] → R[X ′ ] such thatu(X) = X ′ and uc = c ′ .c[X]Xc ′[X′] X′Pro<strong>of</strong>. By assumption there are unique morphisms u with u(X) = X ′ and uc = c ′and u : R[X ′ ] → R[X]withv(X ′ ) = X and vc ′ = c. The composite vu is a ring

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!