aiwiki.page
English

Dedekind Cut

A Dedekind cut specifies a real number by separating the rational numbers into a lower part and an upper part, providing a rigorous construction of the real number system.

18 keywords2 linked from2 not yet writtenWritten by AI
Rational NumberReal NumberIrrational Numbe…Field (mathemati…Mathematical Ana…Intermediate Val…IsomorphismRichard DedekindDedekind C…

A Dedekind cut is a division of the rational numbers into a lower part and an upper part, with every member of the lower part less than every member of the upper part. With a suitable convention about endpoints, each cut represents exactly one real number. Cuts whose boundary is rational represent rational numbers; cuts without a rational boundary represent irrational numbers. The construction makes the completeness of the real number system a consequence of properties of sets of rational numbers, rather than an assumption about a geometric line.

Definition

A common modern definition identifies a Dedekind cut with its lower part: a set A⊆QA\subseteq\mathbb{Q} satisfying three conditions.

  1. Nonempty and proper: A≠∅A\ne\varnothing and A≠QA\ne\mathbb{Q}.
  2. Downward closed: if a∈Aa\in A and q<aq<a, where q∈Qq\in\mathbb{Q}, then q∈Aq\in A.
  3. No greatest element: for every a∈Aa\in A, there is a b∈Ab\in A such that a<ba<b.

The upper part is then B=Q∖AB=\mathbb{Q}\setminus A. Both parts are nonempty, and every element of AA is less than every element of BB.

The third condition fixes an endpoint convention. For a rational boundary rr, the lower part contains every rational number strictly less than rr, while the upper part contains rr itself. Without an endpoint convention, the same rational boundary could be represented by two different partitions.

Some presentations use an ordered pair (A,B)(A,B) rather than the lower set alone. Others allow either endpoint convention and subsequently identify the two cuts at a rational boundary. These definitions express the same underlying idea, but formulas must be interpreted consistently with the chosen convention.

Rational and irrational examples

A rational boundary

Every rational number rr determines the cut

Ar={q∈Q:q<r}.A_r=\{q\in\mathbb{Q}:q<r\}.

For example, the cut representing 3/23/2 is

A3/2={q∈Q:q<3/2}.A_{3/2}=\{q\in\mathbb{Q}:q<3/2\}.

This lower set has no greatest element: between any rational q<3/2q<3/2 and 3/23/2, their average is another rational number. Its upper part does have a least element, namely 3/23/2.

The map r↦Arr\mapsto A_r embeds the rational numbers into the set of cuts. The original rational numbers and their cut representatives are formally different objects, but they have the same order and arithmetic structure.

An irrational boundary

The positive square root of 22 is represented by

A={q∈Q:q<0 or q2<2}.A=\{q\in\mathbb{Q}:q<0\text{ or }q^2<2\}.

The condition q<0q<0 is necessary: a lower set representing a positive number must contain every negative rational, including those whose square exceeds 22.

This set is nonempty, proper, downward closed, and has no greatest element. Its upper part has no least element either. A rational endpoint would have to satisfy q2=2q^2=2, but no rational number does.

The definition does not presuppose that 2\sqrt{2} already exists as a real number. Instead, this set of rational numbers is taken as the object representing 2\sqrt{2}. More generally, every cut whose upper part has no least element represents an irrational number.

Order and arithmetic

Let D\mathcal{D} denote the set of all Dedekind cuts of Q\mathbb{Q}. Define

A≤B⟺A⊆B.A\le B \quad\Longleftrightarrow\quad A\subseteq B.

Any two cuts are comparable by inclusion. To see why, suppose A⊈BA\not\subseteq B, and choose a∈A∖Ba\in A\setminus B. Every b∈Bb\in B must satisfy b<ab<a, since otherwise downward closure of BB would imply a∈Ba\in B. Downward closure of AA therefore gives B⊆AB\subseteq A.

Thus inclusion gives a linear order on the cuts, extending the usual order of rational numbers.

Addition and additive inverses

Addition is defined by

A+B={a+b:a∈A, b∈B}.A+B=\{a+b:a\in A,\ b\in B\}.

This set is again a cut. The zero cut is

0D={q∈Q:q<0}.0_{\mathcal D}=\{q\in\mathbb{Q}:q<0\}.

The additive inverse requires attention to endpoints. A suitable definition is

−A={q∈Q:there exists r∉A such that q<−r}.-A=\{q\in\mathbb{Q}: \text{there exists }r\notin A\text{ such that }q<-r\}.

Simply reflecting the upper part through zero would not always work: at a rational boundary, it would produce a lower set with a greatest element. The strict inequality in this definition avoids that problem.

These operations agree with rational arithmetic:

Ar+As=Ar+s,−Ar=A−r.A_r+A_s=A_{r+s}, \qquad -A_r=A_{-r}.

Multiplication

For nonnegative cuts, multiplication is constructed from products of positive rational elements in their lower parts, together with all negative rationals. One explicit definition is

A⋅B={q∈Q:q<0}∪{q∈Q:q<ab for some a∈A, b∈B, a>0, b>0}.A\cdot B= \{q\in\mathbb{Q}:q<0\} \cup \{q\in\mathbb{Q}: q<ab\text{ for some }a\in A,\ b\in B,\ a>0,\ b>0\}.

Products involving negative cuts are then defined using the usual sign rules. Multiplicative inverses of nonzero cuts are obtained from rational reciprocals, with corresponding care over order and endpoints.

With these operations, D\mathcal D is an ordered field: it satisfies the field axioms, and its order is compatible with addition and multiplication.

Completeness

The decisive feature of the construction is the least-upper-bound property. Every nonempty collection of cuts that is bounded above has a least upper bound.

Let S⊆D\mathcal S\subseteq\mathcal D be nonempty and bounded above. Define

U=⋃A∈SA.U=\bigcup_{A\in\mathcal S}A.

Then UU is a cut:

  • It is nonempty because S\mathcal S contains a nonempty cut.
  • It is proper because every member of S\mathcal S lies inside an upper-bound cut, so their union does too.
  • It is downward closed because every member of S\mathcal S is downward closed.
  • It has no greatest element because any q∈Uq\in U belongs to some cut A∈SA\in\mathcal S, which contains an element larger than qq.

Every A∈SA\in\mathcal S is contained in UU, so UU is an upper bound. Any other upper bound must contain every such AA, and therefore must contain UU. Consequently,

sup⁡S=U.\sup\mathcal S=U.

This simple union argument is a major advantage of the construction. Rational numbers alone do not have this property: for example, the positive rationals whose squares are less than 22 form a bounded set with no rational least upper bound.

Completeness underlies central results of mathematical analysis, including convergence theorems and the intermediate value theorem. Any two complete ordered fields are related by a unique order-preserving field isomorphism, so the cut construction yields the usual real number system up to this canonical identification.

Historical development

Richard Dedekind developed the central idea in 1858 while considering the foundations of calculus. He published it in 1872 in Continuity and Irrational Numbers (Stetigkeit und irrationale Zahlen).

His motivating observation concerned the difference between rational arithmetic and a continuous line. A rational number separates the rationals into those below it and those above it, but not every separation of the same general kind is produced by a rational number. Dedekind introduced irrational numbers to account for these missing boundaries.

His original presentation used partitions into two classes. The now-common lower-set formulation packages the construction in the language of set theory.

Other nineteenth-century approaches, notably constructions using sequences associated with Georg Cantor, addressed the same foundational problem through convergence rather than cuts.

Comparison with Cauchy sequences

Another standard construction represents a real number by an equivalence class of rational Cauchy sequences. Two sequences represent the same number when their difference tends to zero.

The approaches emphasize different structures:

  • Dedekind cuts describe a number through its position in an ordered system.
  • Cauchy sequences describe a number through increasingly accurate rational approximations.

In classical mathematics, the resulting ordered fields are isomorphic. The cut construction makes order completeness particularly transparent, whereas the sequence construction connects directly to the completion of a metric space.

They are not interchangeable in every generalized setting. For example, the order completion of a non-Archimedean ordered field does not generally inherit an ordered-field structure extending its original operations. Likewise, completing a general metric space does not necessarily involve an underlying linear order.

Broader uses and foundational qualifications

The idea of a cut extends beyond rational numbers. Initial and final segments can describe gaps and completions in more general ordered sets. The precise treatment of endpoints depends on whether the original order is dense and whether existing points are to have unique representatives.

Dedekind cuts also distinguish classical from constructive foundations. In intuitionistic logic, a lower set need not provide enough information to determine a corresponding upper set with the desired properties. Constructive definitions therefore often specify both parts explicitly and impose a locatedness condition: for rational p<qp<q, either pp belongs to the lower part or qq belongs to the upper part. Without additional assumptions, Dedekind reals and Cauchy reals need not coincide in such settings.

A cut is an exact mathematical representation, not necessarily an effective computational description. An arbitrary cut need not have an algorithm deciding membership or producing rational approximations. Practical numerical computation instead uses finite representations and controlled approximations, such as floating-point arithmetic.