Fixed-point lemma for normal functions: Difference between revisions

Content deleted Content added
m Dated {{Need consensus}}. (Build p613)
Reverted to revision 444557017 by JRSpriggs: We have consensus; see the article limit ordinal. (TW)
Line 1:
The '''fixed-point lemma for normal functions''' is a basic result in [[axiomatic set theory]] stating that any [[normal function]] has arbitrarily large [[fixed point (mathematics)|fixed point]]s (Levy 1979: p. 117). It was first proved by [[Oswald Veblen]] in 1908.
{{Need consensus|date=August 2011}}
 
The '''fixed-point lemma for normal functions''' is a basic result in [[axiomatic set theory]] stating that any [[normal function]] has arbitrarily large [[fixed point (mathematics)|fixed point]]s (Levy 1979: p. 117). It was first proved by [[Oswald Veblen]] in 1908.
 
== Background and formal statement ==
A [[normal function]] is a [[proper class|class]] function ''f'' from the class Ord of [[ordinal numbers]] to itself such that:
* ''f'' is '''strictly increasing''': ''f''(&alpha;) < f(&beta;) whenever &alpha; < &beta;.
* ''f'' is '''continuous''': for every nonzero limit ordinal &lambda; (i.e. &lambda; is neither zero nor a successor), ''f''(&lambda;) = sup { f(&alpha;) : &alpha; < &lambda; }.
It can be shown that if ''f'' is normal then ''f'' commutes with [[supremum|suprema]]; for any nonempty set ''A'' of ordinals,
:''f''(sup ''A'') = sup {''f''(&alpha;) : &alpha; ∈ ''A'' }.
Indeed, if sup ''A'' is a successor ordinal then sup ''A'' is an element of ''A'' and the equality follows from the increasing property of ''f''. If sup ''A'' is a limit ordinal then the equality follows from the continuous property of ''f''.
 
A '''fixed point''' of a normal function is an ordinal &beta; such that ''f''(&beta;) = &beta;.
 
The fixed point lemma states that the class of fixed points of any normal function is nonempty and in fact is unbounded: given any ordinal α, there exists an ordinal β such that β ≥ α and ''f''(β) = β.
 
The continuity of the normal function implies the class of fixed points is closed (the supremum of any subset of the class of fixed points is again a fixed point). Thus the fixed point lemma is equivalent to the statement that the fixed points of a normal function form a [[club set|closed and unbounded]] class.
 
== Proof ==
The first step of the proof is to verify that ''f''(γ) ≥ γ for all ordinals γ and that ''f'' commutes with suprema. Given these results, inductively define an increasing sequence &lt;α<sub>''n''</sub>&gt; (''n'' &lt; ω) by setting α<sub>0</sub> = α, and α<sub>''n''+1</sub> = ''f''(α<sub>''n''</sub>) for ''n'' ∈ ω. Let β = sup {α<sub>''n''</sub> : ''n'' ∈ &omega;}, so β ≥ α. Moreover, because ''f'' commutes with suprema,
:''f''(β) = ''f''(sup {α<sub>''n''</sub> : ''n'' &lt; ω})
:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; = sup {''f''(α<sub>''n''</sub>) : ''n'' &lt; ω}
:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; = sup {α<sub>''n''+1</sub> : ''n'' &lt; ω}
:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; = β.
The last equality follows from the fact that the sequence &lt;α<sub>''n''</sub>&gt; increases.
Line 40 ⟶ 38:
| authorlink = Oswald Veblen
| title = Continuous increasing functions of finite and transfinite ordinals
| journal = Trans. Amer. Math Soc.
| volume = 9
| year = 1908
| pages = 280&ndash;292
| id = Available via [http://links.jstor.org/sici?sici=0002-9947%28190807%299%3A3%3C280%3ACIFOFA%3E2.0.CO%3B2-1 JSTOR].
| doi= 10.2307/1988605
| issue = 3