TheoremBase

The reciprocal and geometric limits follow from the Archimedean property of the real numbers and Bernoulli's inequality; monotone convergence uses the existence of suprema and infima in the real numbers and their approximation; Bolzano-Weierstrass extracts a monotone subsequence by the peak argument; and a Cauchy sequence converges because a bounded tail has a convergent subsequence.

Proof

Each result cited is universally quantified over the data in its own statement.

We work in the setting of The Real Numbers, with the Natural Numbers, Integers and Rationals Identified with Subsets of the Reals, and Completeness. By its clause The Real Numbers, with the Natural Numbers, Integers and Rationals Identified with Subsets of the Reals, and Completeness §reals, R\mathbb{R} is an ordered field, so the notation of Commutative Rings, Fields and Ordered Fields: Standard Notation applies to it and the rules of Rules of Arithmetic and Order in an Ordered Field are in force by Commutative Rings, Fields and Ordered Fields: Standard Notation §ordered-fields. By The Real Numbers, with the Natural Numbers, Integers and Rationals Identified with Subsets of the Reals, and Completeness §identification, N⊆R\mathbb{N}\subseteq\mathbb{R}: a natural number nn standing where a real number is required is identified with the element of R\mathbb{R} it denotes, and by The Real Numbers, with the Natural Numbers, Integers and Rationals Identified with Subsets of the Reals, and Completeness §agreement sums, products and the order agree whether formed in N\mathbb{N} or in R\mathbb{R}. In particular every n∈Nn\in\mathbb{N} satisfies 0<n0<n in R\mathbb{R}, by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §naturals; so n≠0n\neq0, and 1/n=n−11/n=n^{-1} is defined. Elementary rearrangements in the field R\mathbb{R}, such as x−0=xx-0=x or L−x=−(x−L)L-x=-(x-L) (the latter by Rules of Arithmetic in a Commutative Ring: Zero, Signs and Squares, and No Zero Divisors in a Field §signs), are used without comment. In particular, 0+0=00+0=0 and 1⋅1=11\cdot1=1 by Commutative Rings §ring, so −0=0-0=0 by Negatives, Differences, Reciprocals and Quotients §negative, whence x−0=x+0=xx-0=x+0=x for x∈Rx\in\mathbb{R}, and 1−1=11^{-1}=1 by Negatives, Differences, Reciprocals and Quotients §reciprocal.

Clause reciprocals. We show that the sequence (1/n)n∈N(1/n)_{n\in\mathbb{N}} converges to 00 in the sense of Convergent Sequences of Real Numbers §converges.

Let ε∈R\varepsilon\in\mathbb{R} with ε>0\varepsilon>0. By Rules of Arithmetic and Order in an Ordered Field §positive-reciprocal, 0<ε−10<\varepsilon^{-1}. By the Archimedean property The Real Numbers, with the Natural Numbers, Integers and Rationals Identified with Subsets of the Reals, and Completeness §archimedean, applied to x=ε−1x=\varepsilon^{-1}, there is N∈NN\in\mathbb{N} with ε−1<N\varepsilon^{-1}<N. Let n∈Nn\in\mathbb{N} with n≥Nn\ge N. Then N≤nN\le n also as real numbers, by The Real Numbers, with the Natural Numbers, Integers and Rationals Identified with Subsets of the Reals, and Completeness §agreement, so ε−1<n\varepsilon^{-1}<n by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §mixed. Now Rules of Arithmetic and Order in an Ordered Field §positive-reciprocal, applied to 0<ε−1<n0<\varepsilon^{-1}<n, gives n−1<(ε−1)−1n^{-1}<(\varepsilon^{-1})^{-1}, and applied to 0<n0<n it gives 0<n−10<n^{-1}; moreover (ε−1)−1=ε(\varepsilon^{-1})^{-1}=\varepsilon by Rules of Arithmetic and Order in an Ordered Field §reciprocals. Since −ε<0-\varepsilon<0 by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §strict-negative, we have −ε<1/n−0<ε-\varepsilon<1/n-0<\varepsilon, and so ∣1/n−0∣<ε|1/n-0|<\varepsilon by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §absolute-strict. Thus NN, chosen after ε\varepsilon, satisfies the definition, and 1/n→01/n\to0.

Clause geometric. Let r∈Rr\in\mathbb{R} with ∣r∣<1|r|<1, and let powers be as in Powers with Exponents in the Natural Numbers with Zero §power.

If r=0r=0, then rn=0r^{n}=0 for every n∈Nn\in\mathbb{N} by Powers in a Commutative Ring, a Field and an Ordered Field: Exponent Laws, Factorisation, Geometric Sums, Monotonicity and Bernoulli's Inequality §product, since n≥1n\ge1 by Arithmetic and Order of the Natural Numbers §least. So (rn)(r^{n}) is the constant sequence (0)n∈N(0)_{n\in\mathbb{N}}, which converges to 00 by Limits of Sequences of Real Numbers: Uniqueness, Boundedness, Constants, Tails, Arithmetic, Quotients, Absolute Values, Finite Sums, Order, Squeezing, Domination and Subsequences §constant.

Let now r≠0r\neq0. Then 0<∣r∣0<|r| by Rules of Arithmetic and Order in an Ordered Field §absolute-value. From 0<∣r∣<10<|r|<1, Rules of Arithmetic and Order in an Ordered Field §positive-reciprocal gives 1=1−1<∣r∣−11=1^{-1}<|r|^{-1}, with 1−1=11^{-1}=1 as noted at the start. Put h=∣r∣−1−1h=|r|^{-1}-1. Adding −1-1 to both sides of 1<∣r∣−11<|r|^{-1} gives 0<h0<h by Rules of Arithmetic and Order in an Ordered Field §order-sum, and 1+h=∣r∣−11+h=|r|^{-1}. Since 0<10<1 by Rules of Arithmetic and Order in an Ordered Field §squares, Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §strict-negative gives −1<−0=0-1<-0=0; so h>0>−1h>0>-1, hence −1<h-1<h by Uniqueness of Least and Greatest Elements, Properties of the Strict Order, and Trichotomy for Total Orders §strict-transitive and h≥−1h\ge-1 by Uniqueness of Least and Greatest Elements, Properties of the Strict Order, and Trichotomy for Total Orders §weak-strict.

Let n∈Nn\in\mathbb{N}. Bernoulli's inequality Powers in a Commutative Ring, a Field and an Ordered Field: Exponent Laws, Factorisation, Geometric Sums, Monotonicity and Bernoulli's Inequality §bernoulli, applied with x=hx=h, gives (1+h)n≥1+nh(1+h)^{n}\ge1+nh. Since 0<n0<n by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §naturals and 0<h0<h, we have 0<nh0<nh by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §positive-product. Since 0<10<1 by Rules of Arithmetic and Order in an Ordered Field §squares, 0≤10\le1 by Uniqueness of Least and Greatest Elements, Properties of the Strict Order, and Trichotomy for Total Orders §weak-strict, and Rules of Arithmetic and Order in an Ordered Field §order-sum gives nh≤1+nhnh\le1+nh. Hence 0<nh≤(1+h)n0<nh\le(1+h)^{n}, and by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §reciprocal-order

((1+h)n)−1≤(nh)−1.\big((1+h)^{n}\big)^{-1}\le(nh)^{-1}.

By Powers in a Commutative Ring, a Field and an Ordered Field: Exponent Laws, Factorisation, Geometric Sums, Monotonicity and Bernoulli's Inequality §reciprocal, applied to ∣r∣≠0|r|\neq0, we have ∣r∣n≠0|r|^{n}\neq0 and (1+h)n=(∣r∣−1)n=(∣r∣n)−1(1+h)^{n}=(|r|^{-1})^{n}=(|r|^{n})^{-1}. Hence ((1+h)n)−1=∣r∣n\big((1+h)^{n}\big)^{-1}=|r|^{n} by Rules of Arithmetic and Order in an Ordered Field §reciprocals. Also (nh)−1=h−1 n−1(nh)^{-1}=h^{-1}\,n^{-1} by Rules of Arithmetic and Order in an Ordered Field §reciprocals. Finally ∣rn∣=∣r∣n|r^{n}|=|r|^{n} by Powers in a Commutative Ring, a Field and an Ordered Field: Exponent Laws, Factorisation, Geometric Sums, Monotonicity and Bernoulli's Inequality §sign. Together:

∣rn−0∣=∣r∣n≤h−1⋅1nfor every n∈N.|r^{n}-0|=|r|^{n}\le h^{-1}\cdot\frac{1}{n}\qquad\text{for every }n\in\mathbb{N}.

By the clause reciprocals proved above, 1/n→01/n\to0. By Limits of Sequences of Real Numbers: Uniqueness, Boundedness, Constants, Tails, Arithmetic, Quotients, Absolute Values, Finite Sums, Order, Squeezing, Domination and Subsequences §arithmetic, with λ=h−1\lambda=h^{-1}, the sequence (h−1⋅1/n)n∈N(h^{-1}\cdot1/n)_{n\in\mathbb{N}} converges to h−1⋅0h^{-1}\cdot0, which is 00 by Rules of Arithmetic in a Commutative Ring: Zero, Signs and Squares, and No Zero Divisors in a Field §zero; so it is a null sequence. By Limits of Sequences of Real Numbers: Uniqueness, Boundedness, Constants, Tails, Arithmetic, Quotients, Absolute Values, Finite Sums, Order, Squeezing, Domination and Subsequences §domination, applied with cn=rnc_{n}=r^{n}, L=0L=0, bn=h−1⋅1/nb_{n}=h^{-1}\cdot1/n and K=1K=1, we conclude rn→0r^{n}\to0.

Clause monotone. Let (an)(a_{n}) be a sequence in R\mathbb{R} and S={an:n∈N}S=\{a_{n}:n\in\mathbb{N}\} the set of its terms. SS is nonempty, since a1∈Sa_{1}\in S.

First let (an)(a_{n}) be nondecreasing and bounded above. By Bounded Sequences of Real Numbers §bounded, SS is bounded above in the sense of Bounds, Least and Greatest Elements, Suprema and Infima for a Partial Order §bounded. By The Real Numbers, with the Natural Numbers, Integers and Rationals Identified with Subsets of the Reals, and Completeness §supremum, the nonempty set SS, being bounded above, has a supremum s=sup⁡Ss=\sup S, which is an upper bound of SS by Bounds, Least and Greatest Elements, Suprema and Infima for a Partial Order §supremum.

Let ε∈R\varepsilon\in\mathbb{R} with ε>0\varepsilon>0. By Arbitrary Positive Slack, and Approximation of Suprema and Infima, in the Real Numbers §epsilon-above, applied to the nonempty set SS, which is bounded above, there is an element of SS greater than s−εs-\varepsilon. That element is a term aNa_{N} with N∈NN\in\mathbb{N}, so s−ε<aNs-\varepsilon<a_{N}. Let n∈Nn\in\mathbb{N} with n≥Nn\ge N. Then aN≤ana_{N}\le a_{n} by Monotone Sequences and Subsequences: Comparison of All Terms, Growth of the Indices, and Subsequences of Subsequences §monotone, and an≤sa_{n}\le s since ss is an upper bound of SS. By Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §mixed, s−ε<ans-\varepsilon<a_{n}, so −ε<an−s-\varepsilon<a_{n}-s by Rules of Arithmetic and Order in an Ordered Field §order-sum. Also, adding −s-s to an≤sa_{n}\le s gives an−s≤s−s=0a_{n}-s\le s-s=0 by Rules of Arithmetic and Order in an Ordered Field §order-sum, and 0<ε0<\varepsilon, so an−s<εa_{n}-s<\varepsilon by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §mixed. Hence ∣an−s∣<ε|a_{n}-s|<\varepsilon by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §absolute-strict. Since NN was chosen after ε\varepsilon, this shows an→sup⁡Sa_{n}\to\sup S in the sense of Convergent Sequences of Real Numbers §converges.

Now let (an)(a_{n}) be nonincreasing and bounded below. By Bounded Sequences of Real Numbers §bounded, SS is bounded below. By The Real Numbers, with the Natural Numbers, Integers and Rationals Identified with Subsets of the Reals, and Completeness §infimum, the nonempty set SS, being bounded below, has an infimum t=inf⁡St=\inf S, which is a lower bound of SS by Bounds, Least and Greatest Elements, Suprema and Infima for a Partial Order §supremum.

Let ε∈R\varepsilon\in\mathbb{R} with ε>0\varepsilon>0. By Arbitrary Positive Slack, and Approximation of Suprema and Infima, in the Real Numbers §epsilon-below, applied to the nonempty set SS, which is bounded below, there is an element of SS less than t+εt+\varepsilon. That element is a term aNa_{N} with N∈NN\in\mathbb{N}, so aN<t+εa_{N}<t+\varepsilon. Let n∈Nn\in\mathbb{N} with n≥Nn\ge N. Then t≤ant\le a_{n}, and an≤aNa_{n}\le a_{N} by Monotone Sequences and Subsequences: Comparison of All Terms, Growth of the Indices, and Subsequences of Subsequences §monotone, so an<t+εa_{n}<t+\varepsilon by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §mixed. By Rules of Arithmetic and Order in an Ordered Field §order-sum, t≤ant\le a_{n} gives 0≤an−t0\le a_{n}-t, and adding −t-t to an<t+εa_{n}<t+\varepsilon gives an−t<εa_{n}-t<\varepsilon. Since 0<ε0<\varepsilon, Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §strict-negative gives −ε<−0=0-\varepsilon<-0=0. Thus −ε<0≤an−t<ε-\varepsilon<0\le a_{n}-t<\varepsilon, so −ε<an−t-\varepsilon<a_{n}-t by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §mixed, and ∣an−t∣<ε|a_{n}-t|<\varepsilon by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §absolute-strict. Hence an→inf⁡Sa_{n}\to\inf S.

Clause bolzano-weierstrass. Let (an)(a_{n}) be bounded, and fix M∈RM\in\mathbb{R} with ∣an∣≤M|a_{n}|\le M for every n∈Nn\in\mathbb{N}. Then every sequence of the form a∘σa\circ\sigma, with σ:N→N\sigma:\mathbb{N}\to\mathbb{N}, is bounded with the same MM, since its terms are terms of (an)(a_{n}). By Limits of Sequences of Real Numbers: Uniqueness, Boundedness, Constants, Tails, Arithmetic, Quotients, Absolute Values, Finite Sums, Order, Squeezing, Domination and Subsequences §two-sided, every such sequence is therefore both bounded above and bounded below. We let N\mathbb{N} carry its order, a total order with strict relation << by Arithmetic and Order of the Natural Numbers §partial-order and Arithmetic and Order of the Natural Numbers §trichotomy, as in the preamble of the definition of subsequences.

Call m∈Nm\in\mathbb{N} a peak if an≤ama_{n}\le a_{m} for every n∈Nn\in\mathbb{N} with m≤nm\le n, and let PP be the set of peaks, a subset of N\mathbb{N}. Either for every N∈NN\in\mathbb{N} there is a peak mm with N≤mN\le m, or there is N0∈NN_{0}\in\mathbb{N} such that no m∈Nm\in\mathbb{N} with N0≤mN_{0}\le m is a peak. We treat the two cases separately. In both, a least element of a subset of N\mathbb{N} is unique and is written min⁡\min, as in Bounds, Least and Greatest Elements, Suprema and Infima for a Partial Order §least.

Case 1: for every N∈NN\in\mathbb{N} there is a peak m≥Nm\ge N. For u∈Nu\in\mathbb{N}, the set Pu={m∈P:u≤m}P_{u}=\{m\in P:u\le m\} is nonempty by the case hypothesis. So it has a least element by Arithmetic and Order of the Natural Numbers §well-order, and min⁡Pu∈N\min P_{u}\in\mathbb{N} is defined, min⁡\min being used properly at the nonempty subset PuP_{u}; hence p:N→Np:\mathbb{N}\to\mathbb{N}, u↦min⁡Puu\mapsto\min P_{u}, is a map by Sets and Maps: Ordinary Notation §maps. Apply Recursion on the Natural Numbers Starting at One §recursion with the set N\mathbb{N}, the element c=p(1)c=p(1), and the map g:N×N→Ng:\mathbb{N}\times\mathbb{N}\to\mathbb{N}, (k,u)↦p(u+1)(k,u)\mapsto p(u+1), given by Sets and Maps: Ordinary Notation §maps. It yields a map σ:N→N\sigma:\mathbb{N}\to\mathbb{N} with

σ(1)=p(1),σ(k+1)=p(σ(k)+1)for every k∈N.\sigma(1)=p(1),\qquad\sigma(k+1)=p(\sigma(k)+1)\quad\text{for every }k\in\mathbb{N}.

Every j∈Nj\in\mathbb{N} is 11 or of the form k+1k+1 with k∈Nk\in\mathbb{N}, by Arithmetic and Order of the Natural Numbers §predecessor. So every value σ(j)\sigma(j) is a value of pp, and hence a peak.

Let k∈Nk\in\mathbb{N}. Since σ(k+1)∈Pσ(k)+1\sigma(k+1)\in P_{\sigma(k)+1}, we have σ(k)+1≤σ(k+1)\sigma(k)+1\le\sigma(k+1). Also σ(k)<σ(k)+1\sigma(k)<\sigma(k)+1 by Arithmetic and Order of the Natural Numbers §successor. So σ(k)<σ(k+1)\sigma(k)<\sigma(k+1) by Arithmetic and Order of the Natural Numbers §partial-order. Thus σ\sigma is strictly increasing, and a∘σa\circ\sigma is a subsequence of (an)(a_{n}). Moreover σ(k)\sigma(k) is a peak, and σ(k)≤σ(k+1)\sigma(k)\le\sigma(k+1) because σ(k)<σ(k+1)\sigma(k)<\sigma(k+1), by Arithmetic and Order of the Natural Numbers §partial-order; so aσ(k+1)≤aσ(k)a_{\sigma(k+1)}\le a_{\sigma(k)}. Hence a∘σa\circ\sigma is nonincreasing. It is bounded below, as noted above, so by the clause monotone proved above it converges, to inf⁡{aσ(k):k∈N}\inf\{a_{\sigma(k)}:k\in\mathbb{N}\}, which exists by The Real Numbers, with the Natural Numbers, Integers and Rationals Identified with Subsets of the Reals, and Completeness §infimum, this set of terms being nonempty and, by Bounded Sequences of Real Numbers §bounded, bounded below.

Case 2: there is N0∈NN_{0}\in\mathbb{N} such that no m≥N0m\ge N_{0} is a peak. We first show the following. If u∈Nu\in\mathbb{N} and N0≤uN_{0}\le u, then there is n∈Nn\in\mathbb{N} with u<nu<n and au<ana_{u}<a_{n}. Indeed, uu is not a peak, so there is n∈Nn\in\mathbb{N} with u≤nu\le n and not an≤aua_{n}\le a_{u}. The order of R\mathbb{R} is total by Ordered Fields §ordered-field and Partial and Total Orders on a Set and the Associated Strict Relation §total, so au≤ana_{u}\le a_{n}. If an=aua_{n}=a_{u}, then an≤aua_{n}\le a_{u} by reflexivity of the order of R\mathbb{R} (Partial and Total Orders on a Set and the Associated Strict Relation §partial), contrary to the choice of nn; so an≠aua_{n}\neq a_{u}, and thus n≠un\neq u. Hence au<ana_{u}<a_{n} by Partial and Total Orders on a Set and the Associated Strict Relation §strict, and u<nu<n by Arithmetic and Order of the Natural Numbers §partial-order.

For u∈Nu\in\mathbb{N} let

Tu={n∈N:u<n and (u<N0 or au<an)}.T_{u}=\{n\in\mathbb{N}:u<n\ \text{and}\ (u<N_{0}\ \text{or}\ a_{u}<a_{n})\}.

If u<N0u<N_{0}, then u+1∈Tuu+1\in T_{u} by Arithmetic and Order of the Natural Numbers §successor. Otherwise N0≤uN_{0}\le u by Arithmetic and Order of the Natural Numbers §trichotomy, and TuT_{u} is nonempty by the preceding paragraph. In either case TuT_{u} has a least element by Arithmetic and Order of the Natural Numbers §well-order, so min⁡Tu∈N\min T_{u}\in\mathbb{N} is defined, min⁡\min being used properly at the nonempty subset TuT_{u}, and q:N→Nq:\mathbb{N}\to\mathbb{N}, u↦min⁡Tuu\mapsto\min T_{u}, is a map by Sets and Maps: Ordinary Notation §maps, with u<q(u)u<q(u) for every u∈Nu\in\mathbb{N}. Moreover au<aq(u)a_{u}<a_{q(u)} whenever N0≤uN_{0}\le u, since then u<N0u<N_{0} fails by Arithmetic and Order of the Natural Numbers §trichotomy.

Apply Recursion on the Natural Numbers Starting at One §recursion with the set N\mathbb{N}, the element c=N0c=N_{0}, and the map g:N×N→Ng:\mathbb{N}\times\mathbb{N}\to\mathbb{N}, (k,u)↦q(u)(k,u)\mapsto q(u), given by Sets and Maps: Ordinary Notation §maps. It yields a map σ:N→N\sigma:\mathbb{N}\to\mathbb{N} with σ(1)=N0\sigma(1)=N_{0} and σ(k+1)=q(σ(k))\sigma(k+1)=q(\sigma(k)) for every k∈Nk\in\mathbb{N}. Then σ(k)<q(σ(k))=σ(k+1)\sigma(k)<q(\sigma(k))=\sigma(k+1) for every kk, so σ\sigma is strictly increasing, and a∘σa\circ\sigma is a subsequence of (an)(a_{n}).

For k∈Nk\in\mathbb{N} we have 1≤k1\le k by Arithmetic and Order of the Natural Numbers §least. So either k=1k=1, or 1<k1<k and then σ(1)<σ(k)\sigma(1)<\sigma(k) by Monotone Sequences and Subsequences: Comparison of All Terms, Growth of the Indices, and Subsequences of Subsequences §monotone. In both cases N0≤σ(k)N_{0}\le\sigma(k), and therefore aσ(k)<aq(σ(k))=aσ(k+1)a_{\sigma(k)}<a_{q(\sigma(k))}=a_{\sigma(k+1)}. Hence a∘σa\circ\sigma is nondecreasing, by Partial and Total Orders on a Set and the Associated Strict Relation §strict. It is bounded above, as noted above, so by the clause monotone it converges, to sup⁡{aσ(k):k∈N}\sup\{a_{\sigma(k)}:k\in\mathbb{N}\}, which exists by The Real Numbers, with the Natural Numbers, Integers and Rationals Identified with Subsets of the Reals, and Completeness §supremum, this set of terms being nonempty and, by Bounded Sequences of Real Numbers §bounded, bounded above.

In both cases (an)(a_{n}) has a convergent subsequence.

Clause cauchy. Let (an)(a_{n}) be a sequence in R\mathbb{R}, with Cauchy sequences as in Cauchy Sequences of Real Numbers §cauchy.

Convergent implies Cauchy. Let an→La_{n}\to L with L∈RL\in\mathbb{R}, and let ε∈R\varepsilon\in\mathbb{R} with ε>0\varepsilon>0. Then ε/2\varepsilon/2 is positive by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §halving. Choose N∈NN\in\mathbb{N}, after ε\varepsilon, with ∣an−L∣<ε/2|a_{n}-L|<\varepsilon/2 for every n≥Nn\ge N. Let m,n∈Nm,n\in\mathbb{N} with m,n≥Nm,n\ge N. By Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §triangle-three-points,

∣am−an∣≤∣am−L∣+∣L−an∣,|a_{m}-a_{n}|\le|a_{m}-L|+|L-a_{n}|,

and ∣L−an∣=∣−(an−L)∣=∣an−L∣|L-a_{n}|=|-(a_{n}-L)|=|a_{n}-L| by Rules of Arithmetic and Order in an Ordered Field §absolute-value. Then Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §strict-sum gives ∣am−L∣+∣an−L∣<ε/2+ε/2|a_{m}-L|+|a_{n}-L|<\varepsilon/2+\varepsilon/2, which is ε\varepsilon by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §halving. So ∣am−an∣<ε|a_{m}-a_{n}|<\varepsilon by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §mixed, and (an)(a_{n}) is a Cauchy sequence.

Cauchy implies convergent. Let (an)(a_{n}) be a Cauchy sequence.

A bounded tail. Applying the Cauchy property with ε=1\varepsilon=1, which is positive by Rules of Arithmetic and Order in an Ordered Field §squares, gives N1∈NN_{1}\in\mathbb{N} with ∣am−an∣<1|a_{m}-a_{n}|<1 for all m,n≥N1m,n\ge N_{1}. Let ρ:N→N\rho:\mathbb{N}\to\mathbb{N} be the map n↦n+N1n\mapsto n+N_{1}, given by Sets and Maps: Ordinary Notation §maps as n+N1∈Nn+N_{1}\in\mathbb{N}, and let b=a∘ρb=a\circ\rho, that is, bn=an+N1b_{n}=a_{n+N_{1}}.

The map ρ\rho is strictly increasing. Indeed, n<n+1n<n+1 by Arithmetic and Order of the Natural Numbers §successor, hence n+N1<(n+1)+N1n+N_{1}<(n+1)+N_{1} by Arithmetic and Order of the Natural Numbers §order.

Moreover, for every n∈Nn\in\mathbb{N} we have N1<N1+n=n+N1N_{1}<N_{1}+n=n+N_{1} by Arithmetic and Order of the Natural Numbers §difference and Arithmetic and Order of the Natural Numbers §commutative. So ∣bn−aN1∣<1|b_{n}-a_{N_{1}}|<1. By Rules of Arithmetic and Order in an Ordered Field §triangle and Rules of Arithmetic and Order in an Ordered Field §order-sum,

∣bn∣=∣(bn−aN1)+aN1∣≤∣bn−aN1∣+∣aN1∣≤1+∣aN1∣.|b_{n}|=|(b_{n}-a_{N_{1}})+a_{N_{1}}|\le|b_{n}-a_{N_{1}}|+|a_{N_{1}}|\le1+|a_{N_{1}}|.

Thus (bn)(b_{n}) is bounded.

A convergent subsequence. By the clause bolzano-weierstrass proved above, (bn)(b_{n}) has a convergent subsequence b∘τb\circ\tau, with τ:N→N\tau:\mathbb{N}\to\mathbb{N} strictly increasing and b∘τb\circ\tau converging to some L∈RL\in\mathbb{R}. Put σ=ρ∘τ\sigma=\rho\circ\tau. By Monotone Sequences and Subsequences: Comparison of All Terms, Growth of the Indices, and Subsequences of Subsequences §composition, σ\sigma is strictly increasing and b∘τ=(a∘ρ)∘τ=a∘σb\circ\tau=(a\circ\rho)\circ\tau=a\circ\sigma. Hence aσ(k)→La_{\sigma(k)}\to L.

The whole sequence converges to LL. Let ε∈R\varepsilon\in\mathbb{R} with ε>0\varepsilon>0; then ε/2\varepsilon/2 is positive by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §halving. We make three choices, in this order, each after ε\varepsilon.

First, by the Cauchy property, choose N2∈NN_{2}\in\mathbb{N} with ∣am−an∣<ε/2|a_{m}-a_{n}|<\varepsilon/2 for all m,n≥N2m,n\ge N_{2}.

Second, by aσ(k)→La_{\sigma(k)}\to L, choose K∈NK\in\mathbb{N} with ∣aσ(k)−L∣<ε/2|a_{\sigma(k)}-L|<\varepsilon/2 for every k≥Kk\ge K.

Third, by Arithmetic and Order of the Natural Numbers §trichotomy, let kk be the larger of KK and N2N_{2}, so that K≤kK\le k and N2≤kN_{2}\le k. Then k≤σ(k)k\le\sigma(k) by Monotone Sequences and Subsequences: Comparison of All Terms, Growth of the Indices, and Subsequences of Subsequences §index, hence N2≤σ(k)N_{2}\le\sigma(k) by Arithmetic and Order of the Natural Numbers §partial-order.

Now let n∈Nn\in\mathbb{N} with n≥N2n\ge N_{2}. Then ∣an−aσ(k)∣<ε/2|a_{n}-a_{\sigma(k)}|<\varepsilon/2 by the first choice, as n≥N2n\ge N_{2} and σ(k)≥N2\sigma(k)\ge N_{2}, and ∣aσ(k)−L∣<ε/2|a_{\sigma(k)}-L|<\varepsilon/2 by the second choice, as k≥Kk\ge K. By Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §triangle-three-points, Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §strict-sum and Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §halving,

∣an−L∣≤∣an−aσ(k)∣+∣aσ(k)−L∣<ε2+ε2=ε,|a_{n}-L|\le|a_{n}-a_{\sigma(k)}|+|a_{\sigma(k)}-L|<\frac{\varepsilon}{2}+\frac{\varepsilon}{2}=\varepsilon,

and so ∣an−L∣<ε|a_{n}-L|<\varepsilon by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares §mixed. Hence N2N_{2} satisfies Convergent Sequences of Real Numbers §converges for ε\varepsilon, and an→La_{n}\to L. So (an)(a_{n}) is convergent.

Citations

Loading…

Dependencies

Uses0

Loading…

Comments

Log in to comment.

Loading…