formal power series as inverse limits
1 Motivation and Overview
The ring of formal power series can be described as an inverselimit.11It is worth pointing out that, since we aredealing with formal series, the concept of limit used here hasnothing to do with convergence but is purely algebraic. Thefundamental idea behind this approach is that of truncation —given a formal power series and an integer , we can truncate the series toorder to obtain .22Here, the symbol “” is notused in the sense of Landau notation but merely as an indicatorthat the power series has been truncated to order .(Indeed, we must do this in practical computation since it isonly possible to write down a finite number of terms at a time;thus the approach taken here has the advantage of being closeto actual practice.) Furthermore, this procedure of truncationcommutes with ring operations
— given two formal power series,the truncation of their sum is the sum of their truncations andthe truncation of their product
is the product of theirtruncations. Thus, for every integer the set of powerseries truncated to order forms a ring and truncation is amorphism
from the ring of formal power series to this ring.
To obtain our definition, we will proceed in the oppositedirection. We will begin by defining rings of truncatedpower series and exhibiting truncation morphisms betweendifferent truncations. Then we will show that these ringsand morphisms form an inverse system which has a limit whichwe will take as the definition of the ring of formal powerseries. Finally, we will complete
the circle by demonstratingthat the object so constructed is isomorphic
with the usualdefinition for ring of formal power series.
2 Formal Development
In this section, we will carry out the develpment outlinedabove in rigorous detail. We begin by formalizing thisnotion of truncation.
Theorem 1.
Let be a commutative ring and let be a positiveinteger. Then is isomorphicto .
Proof.
We may identify with the subring of consistingof series which have all but a finite number of coeficientsequal to zero. Consider an element of . We may write . Thus, every element of is equivalent to an element of the subring modulo . Hence, if follows rather immediately from thedefinition of quotient ring
that is isomorphic to .∎
Let us call the isomorphism between and which is described above . We now define a few more morphisms.
Definition 1.
Suppose are integers satifying the inequalities. Then define the morphisms as follows:
- •
Define as the map which sends eachequivalence class
modulo to the unique equivalenceclass modulo such that .
- •
Define as the map which sends eachequivalence class modulo to the unique equivalenceclass modulo such that .
- •
For every integer , let be the quotientmap from to .
- •
For every integer , let be the quotient
map from to .
These morphisms commute with each other in ways which aredescribed by the next theorem:
Theorem 2.
Suppose are integers satifying the inequalities. Then we have the following relations:
- 1.
- 2.
- 3.
- 4.
- 5.
[More to come]