Jump to content

Definable real number: Difference between revisions

From Wikipedia, the free encyclopedia
Content deleted Content added
m link ref correctly
rm fallacious statement; add sources; wikiformatting
Line 8: Line 8:
A [[real number]] ''a'' is '''first-order definable in the language of set theory, without parameters''', if there is a formula ''φ'' in the language of [[set theory]], with one [[free variable]], such that ''a'' is the unique real number such that ''φ(a)'' holds in the standard model of set theory (see {{harvnb|Kunen|1980|p=153}}).
A [[real number]] ''a'' is '''first-order definable in the language of set theory, without parameters''', if there is a formula ''φ'' in the language of [[set theory]], with one [[free variable]], such that ''a'' is the unique real number such that ''φ(a)'' holds in the standard model of set theory (see {{harvnb|Kunen|1980|p=153}}).
For the purposes of this article, such reals will be called simply '''''definable numbers'''''. This should not be understood to be standard terminology.
For the purposes of this article, such reals will be called simply '''definable numbers'''. This should not be understood to be standard terminology.


Note that this definition cannot be expressed in the language of set theory itself.
Note that this definition cannot be expressed in the language of set theory itself.
Line 14: Line 14:
==General facts==
==General facts==


Assuming they form a set, the definable numbers form a [[field (mathematics)|field]] containing all the familiar real numbers such as [[Zero|0]], [[One|1]], [[Pi|{{pi}}]], [[E (mathematical constant)|''e'']], et cetera. In particular, this field contains all the numbers named in the [[mathematical constant]]s article, and all [[algebraic number]]s (and therefore all [[rational number]]s). However, most real numbers are not definable: the [[Set (mathematics)|set]] of all definable numbers is [[countably infinite]] (because the set of all logical formulas is) while the set of real numbers is [[uncountable set|uncountably infinite]] (see [[Cantor's diagonal argument]]). As a result, [[almost all|most]] real numbers have no description (in the same sense of "most" as 'most real numbers are not rational').
Assuming they form a set, the definable numbers form a [[field (mathematics)|field]] containing all the familiar real numbers such as [[Zero|0]], [[One|1]], [[Pi|{{pi}}]], [[E (mathematical constant)|''e'']], et cetera. In particular, this field contains all the numbers named in the [[mathematical constant]]s article, and all [[algebraic number]]s (and therefore all [[rational number]]s).


A naive treatment of definability might lead to the fallacious conclusion that most real numbers are not definable.<ref>{{Cite web|url=http://mathoverflow.net/a/44129|title=Is the analysis as taught in universities in fact the analysis of definable numbers?|website=mathoverflow.net|access-date=2016-03-05}}</ref> The usual argument asserts that the set of all logical formulas is only [[countably infinite]], while the set of real numbers is [[uncountable set|uncountably infinite]], therefore there must be real numbers that have no description (see [[Cantor's diagonal argument]]). However, this argument is fallacious, and, in fact, it can be proven, assuming the consistency of [[Zermelo–Fraenkel set theory|ZFC]], that there are models of ZFC in which all real numbers, all sets of real numbers, functions on the reals, etc. are definable.<ref>{{Cite journal|last=Hamkins|first=Joel David|last2=Linetsky|first2=David|last3=Reitz|first3=Jonas|date=2011-05-23|title=Pointwise Definable Models of Set Theory|url=http://arxiv.org/abs/1105.4597|journal=arXiv:1105.4597 [math]}}</ref>
The field of definable numbers is not [[complete space|complete]]; there exist [[Cauchy sequence]]s of definable numbers whose [[limit of a sequence|limit]] is not definable (since every real number is the limit of a sequence of rational numbers). However, if the sequence itself is definable in the sense that we can specify a single formula for all its terms, then its limit will necessarily be a definable number.


While every [[computable number]] is definable, the converse is not true: the numeric representations of the [[Halting problem]], [[Chaitin's constant]], the truth set of first order arithmetic, and [[Zero sharp|0<sup>#</sup>]] are examples of numbers that are definable but not computable. Many other such numbers are known.
While every [[computable number]] is definable, the converse is not true: the numeric representations of the [[Halting problem]], [[Chaitin's constant]], the truth set of first order arithmetic, and [[Zero sharp|0<sup>#</sup>]] are examples of numbers that are definable but not computable. Many other such numbers are known.


One may also wish to talk about definable [[complex number]]s: complex numbers which are uniquely defined by a logical formula. However, whether this is possible depends on how the field of complex numbers is derived in the first place: it may not be possible to distinguish a complex number from its conjugate (say, 3+i from 3-i), since it is impossible to find a property of one that is not also a property of the other, without falling back on the underlying set-theoretic definition. Assuming we can define at least one nonreal complex number, however, a [[complex number]] is definable if and only if both its real part and its imaginary part are definable. The definable complex numbers also form a field if they form a set.
One may also wish to talk about definable [[complex number]]s: complex numbers which are uniquely defined by a logical formula. However, whether this is possible depends on how the field of complex numbers is derived in the first place: it may not be possible to distinguish a complex number from its conjugate (say, 3+i from 3-i), since it is impossible to find a property of one that is not also a property of the other, without falling back on the underlying set-theoretic definition. Assuming we can define at least one nonreal complex number, however, a [[complex number]] is definable if and only if both its real part and its imaginary part are definable. The definable complex numbers also form a field if they form a set.
Line 25: Line 25:


==Other notions of definability==
==Other notions of definability==
The notion of definability treated in this article has been chosen primarily for definiteness, not on the grounds that it's more useful or interesting than other notions. Here we treat a few others:
The notion of definability treated in this article has been chosen primarily for definiteness, not on the grounds that it's more useful or interesting than other notions. Here we treat a few others:


===Definability in other languages or structures===
===Definability in other languages or structures===


====Language of arithmetic====
====Language of arithmetic====
The [[language of arithmetic]] has symbols for 0, 1, the successor operation, addition, and multiplication, intended to be interpreted in the usual way over the [[natural number]]s. Since no variables of this language range over the [[real number]]s, we cannot simply copy the earlier definition of definability. Rather, we say that a real ''a'' is '''''definable in the language of arithmetic''''' (or '''''[[arithmetical hierarchy|arithmetical]]''''') if its [[Dedekind cut]] can be defined as a [[Predicate (logic)|predicate]] in that language; that is, if there is a first-order formula ''φ'' in the language of arithmetic, with two free variables, such that
The [[language of arithmetic]] has symbols for 0, 1, the successor operation, addition, and multiplication, intended to be interpreted in the usual way over the [[natural number]]s. Since no variables of this language range over the [[real number]]s, we cannot simply copy the earlier definition of definability. Rather, we say that a real ''a'' is ''definable in the language of arithmetic'' (or ''[[arithmetical hierarchy|arithmetical]]'') if its [[Dedekind cut]] can be defined as a [[Predicate (logic)|predicate]] in that language; that is, if there is a first-order formula ''φ'' in the language of arithmetic, with two free variables, such that


:<math>\forall m \, \forall n \, (\varphi(n,m)\iff\frac{n}{m}<a).</math>
:<math>\forall m \, \forall n \, (\varphi(n,m)\iff\frac{n}{m}<a).</math>


====2nd-order language of arithmetic====
====2nd-order language of arithmetic====
The second-order language of arithmetic is the same as the first-order language, except that variables and quantifiers are allowed to range over sets of naturals. A real that is second-order definable in the language of arithmetic is called '''''[[analytical hierarchy|analytical]]'''''.
The second-order language of arithmetic is the same as the first-order language, except that variables and quantifiers are allowed to range over sets of naturals. A real that is second-order definable in the language of arithmetic is called ''[[analytical hierarchy|analytical]]''.


===Definability with ordinal parameters===
===Definability with ordinal parameters===
Sometimes it is of interest to consider definability ''with parameters''; that is, to give a definition relative to another object that remains undefined. For example, a real ''a'' (or for that matter, any set ''a'') is called '''''ordinal definable''''' if there is a first-order formula ''φ'' in the language of set theory, with ''two'' free variables, and an [[Ordinal number|ordinal]] γ, such that ''a'' is the unique object such that ''φ''(''a'',γ) holds (in V).
Sometimes it is of interest to consider definability ''with parameters''; that is, to give a definition relative to another object that remains undefined. For example, a real ''a'' (or for that matter, any set ''a'') is called ''ordinal definable'' if there is a first-order formula ''φ'' in the language of set theory, with ''two'' free variables, and an [[Ordinal number|ordinal]] γ, such that ''a'' is the unique object such that ''φ''(''a'',γ) holds (in V).


The other sorts of definability thus far considered have only countably many defining formulas, and therefore allow only countably many definable reals. This is not true for ordinal definability, because an ordinal definable real is defined not only by the formula ''φ'', but also by the ordinal γ. In fact it is consistent with [[ZFC]] that ''all'' reals are ordinal-definable, and therefore that there are uncountably many ordinal-definable reals. However it is also consistent with ZFC that there are only countably many ordinal-definable reals.
The other sorts of definability thus far considered have only countably many defining formulas, and therefore allow only countably many definable reals. This is not true for ordinal definability, because an ordinal definable real is defined not only by the formula ''φ'', but also by the ordinal γ. In fact it is consistent with [[ZFC]] that ''all'' reals are ordinal-definable, and therefore that there are uncountably many ordinal-definable reals. However it is also consistent with ZFC that there are only countably many ordinal-definable reals.


== See also ==
== See also ==
Line 48: Line 48:
* [[Constructible universe]]
* [[Constructible universe]]
* [[Entscheidungsproblem]]
* [[Entscheidungsproblem]]
* [[Tarski's undefinability theorem]]


==References==
==References==
{{Reflist}}

* {{Citation | last1=Kunen | first1=Kenneth | author1-link=Kenneth Kunen | title=[[Set Theory: An Introduction to Independence Proofs]] | publisher=North-Holland | location=Amsterdam | isbn=978-0-444-85401-8 | year=1980}}
* {{Citation | last1=Kunen | first1=Kenneth | author1-link=Kenneth Kunen | title=[[Set Theory: An Introduction to Independence Proofs]] | publisher=North-Holland | location=Amsterdam | isbn=978-0-444-85401-8 | year=1980}}
* [http://rothkamm.com/download.cfm/turing/Turing.pdf Alan Turing, "On Computable Numbers, With An Application to the Entscheidungsproblem", ''Proceedings of the London Mathematical Society'', 1936] ([[Alan Turing|Turing's]] original paper distinguishing computable and definable numbers)
* [http://rothkamm.com/download.cfm/turing/Turing.pdf Alan Turing, "On Computable Numbers, With An Application to the Entscheidungsproblem", ''Proceedings of the London Mathematical Society'', 1936] ([[Alan Turing|Turing's]] original paper distinguishing computable and definable numbers)

Revision as of 13:40, 5 March 2016

A real number a is first-order definable in the language of set theory, without parameters, if there is a formula φ in the language of set theory, with one free variable, such that a is the unique real number such that φ(a) holds in the standard model of set theory (see Kunen 1980, p. 153).

For the purposes of this article, such reals will be called simply definable numbers. This should not be understood to be standard terminology.

Note that this definition cannot be expressed in the language of set theory itself.

General facts

Assuming they form a set, the definable numbers form a field containing all the familiar real numbers such as 0, 1, π, e, et cetera. In particular, this field contains all the numbers named in the mathematical constants article, and all algebraic numbers (and therefore all rational numbers).

A naive treatment of definability might lead to the fallacious conclusion that most real numbers are not definable.[1] The usual argument asserts that the set of all logical formulas is only countably infinite, while the set of real numbers is uncountably infinite, therefore there must be real numbers that have no description (see Cantor's diagonal argument). However, this argument is fallacious, and, in fact, it can be proven, assuming the consistency of ZFC, that there are models of ZFC in which all real numbers, all sets of real numbers, functions on the reals, etc. are definable.[2]

While every computable number is definable, the converse is not true: the numeric representations of the Halting problem, Chaitin's constant, the truth set of first order arithmetic, and 0# are examples of numbers that are definable but not computable. Many other such numbers are known.

One may also wish to talk about definable complex numbers: complex numbers which are uniquely defined by a logical formula. However, whether this is possible depends on how the field of complex numbers is derived in the first place: it may not be possible to distinguish a complex number from its conjugate (say, 3+i from 3-i), since it is impossible to find a property of one that is not also a property of the other, without falling back on the underlying set-theoretic definition. Assuming we can define at least one nonreal complex number, however, a complex number is definable if and only if both its real part and its imaginary part are definable. The definable complex numbers also form a field if they form a set.

The related concept of "standard" numbers, which can only be defined within a finite time and space, is used to motivate axiomatic internal set theory, and provide a workable formulation for illimited and infinitesimal number. Definitions of the hyper-real line within non-standard analysis (the subject area dealing with such numbers) overwhelmingly include the usual, uncountable set of real numbers as a subset.

Other notions of definability

The notion of definability treated in this article has been chosen primarily for definiteness, not on the grounds that it's more useful or interesting than other notions. Here we treat a few others:

Definability in other languages or structures

Language of arithmetic

The language of arithmetic has symbols for 0, 1, the successor operation, addition, and multiplication, intended to be interpreted in the usual way over the natural numbers. Since no variables of this language range over the real numbers, we cannot simply copy the earlier definition of definability. Rather, we say that a real a is definable in the language of arithmetic (or arithmetical) if its Dedekind cut can be defined as a predicate in that language; that is, if there is a first-order formula φ in the language of arithmetic, with two free variables, such that

2nd-order language of arithmetic

The second-order language of arithmetic is the same as the first-order language, except that variables and quantifiers are allowed to range over sets of naturals. A real that is second-order definable in the language of arithmetic is called analytical.

Definability with ordinal parameters

Sometimes it is of interest to consider definability with parameters; that is, to give a definition relative to another object that remains undefined. For example, a real a (or for that matter, any set a) is called ordinal definable if there is a first-order formula φ in the language of set theory, with two free variables, and an ordinal γ, such that a is the unique object such that φ(a,γ) holds (in V).

The other sorts of definability thus far considered have only countably many defining formulas, and therefore allow only countably many definable reals. This is not true for ordinal definability, because an ordinal definable real is defined not only by the formula φ, but also by the ordinal γ. In fact it is consistent with ZFC that all reals are ordinal-definable, and therefore that there are uncountably many ordinal-definable reals. However it is also consistent with ZFC that there are only countably many ordinal-definable reals.

See also

References

  1. ^ "Is the analysis as taught in universities in fact the analysis of definable numbers?". mathoverflow.net. Retrieved 2016-03-05.
  2. ^ Hamkins, Joel David; Linetsky, David; Reitz, Jonas (2011-05-23). "Pointwise Definable Models of Set Theory". arXiv:1105.4597 [math].