# Combination Theorem for Cauchy Sequences/Inverse Rule

## Theorem

Let $\struct {R, \norm {\, \cdot \,} }$ be a normed division ring with zero: $0$.

Let $\sequence {x_n}$ be a Cauchy sequences in $R$.

Suppose $\sequence {x_n}$ does not converge to $0$.

Then:

$\exists K \in \N : \forall n > K : x_n \ne 0$

and the sequence:

$\sequence {\paren {x_{K + n} }^{-1} }_{n \mathop \in \N}$ is well-defined and a Cauchy sequence.

## Proof

Since $\sequence {x_n}$ does not converge to $0$, by Cauchy Sequence Is Eventually Bounded Away From Non-Limit then:

$\exists K \in \N$ and $C \in \R_{> 0}: \forall n > K: C < \norm {x_n}$

or equivalently:

$\exists K \in \N$ and $C \in \R_{> 0}: \forall n > K: 1 < \dfrac {\norm {x_n} } C$
$\forall n > K : x_n \ne 0$

Let $\sequence {y_n}$ be the subsequence of $\sequence {x_n}$ where $y_n = x_{K + n}$.

By Subsequence of a Cauchy Sequence is a Cauchy Sequence then $\sequence {y_n}$ is a Cauchy sequence.

So $\sequence { {y_n}^{-1} }$ is well-defined and $\sequence { {y_n}^{-1} } = \sequence { \paren {x_{K + n} }^{-1} }_{n \mathop \in \N}$.

Let $\epsilon > 0$ be given.

Let $\epsilon' = \epsilon C^2$, then $\epsilon' > 0$.

Similarly, $\sequence {y_n}$ is a Cauchy sequence, we can find $N$ such that:

$\forall n, m > N_2: \norm {y_n - y_m} < \epsilon'$

Thus $\forall n, m > N$:

$(1): \quad 1 < \dfrac {\norm {y_n} } C, \dfrac {\norm {y_m} } C$
$(2): \quad \norm {y_n - y_m} < \epsilon'$

Hence:

 $\displaystyle \norm { {y_n}^{-1} - {y_m}^{-1} }$ $<$ $\displaystyle \dfrac {\norm {y_n} } C \norm { {y_n }^{-1} - {y_m}^{-1} } \dfrac {\norm {y_m} } C$ $(1)$ above $\displaystyle$ $=$ $\displaystyle \dfrac 1 {C^2} \paren {\norm {y_n} \norm { {y_n }^{-1} - {y_m}^{-1 } } \norm {y_m} }$ $\displaystyle$ $=$ $\displaystyle \dfrac 1 {C^2} \norm {y_n \paren { {y_n }^{-1 } - {y_m }^{-1} } y_m}$ Norm Axiom $(\text N 2)$: Multiplicativity $\displaystyle$ $=$ $\displaystyle \dfrac 1 {C^2} \norm {\paren {y_n {y_n }^{-1 } - y_n {y_m }^{-1} } y_m }$ Ring Axiom $(\text D)$: Distributivity $\displaystyle$ $=$ $\displaystyle \dfrac 1 {C^2} \norm {y_n {y_n}^{-1} y_m - y_n {y_m}^{-1} y_m}$ Ring Axiom $(\text D)$: Distributivity $\displaystyle$ $=$ $\displaystyle \dfrac 1 {C^2} \norm {y_m - y_n}$ Inverse Property of a Division Ring $\displaystyle$ $<$ $\displaystyle \dfrac 1 {C^2} \epsilon'$ $(2)$ above $\displaystyle$ $=$ $\displaystyle \dfrac 1 {C^2} \paren {\epsilon C^2}$ Definition of $\epsilon'$ $\displaystyle$ $=$ $\displaystyle \epsilon$

So:

$\sequence { { {y_n}^{-1} } }$ is a Cauchy sequence in $R$.

$\blacksquare$