“实数集中的Borel集全体的势为连续统势”这一论断是否蕴含“实数集存在一个不可测子集”?
神奇,这个问题让我发现了选择公理研究中两本经典参考读物的错误。首先回答题主的问题:“实数集中的Borel集全体的势为连续统势”这一论断的确蕴涵“实数集存在一个不可测子集”,证明放在最后。
但我想先指出的一点是,按照现代选择公理蕴涵关系的标准参考读物的说法,这个蕴涵关系是不成立的。下面说原因。
在Paul Howard的Consequences of the Axiom of Choice中,给出的蕴涵关系是Form 43(也就是Axiom of Dependent Choice, DC)蕴涵Form 363(“Borel集的集合与实数集自身等势”)
386 PART V: REFERENCES FOR RELATIONS BETWEEN FORMS
361 358 (3) note 18 (N2T)
363 358 (3) note 18 (N2T)
8 360 (1) clear
8 361 (1) G.Moore [1982] p 325
0 362 (7) This project
361 362 (1) G. Moore [1982] p 325
43 363 (1) G. Moore [1982] p 325
它引用的文献是Moore的Zermelo's Axiom of Choice,其中第325页不加引用地断言了这个蕴涵关系
Table 4. Deductive Relations Concerning the Denumerable Axiom of Choice and the Principle of Dependent Choices
Axiom of Choice
⇟
For every x,Cℵˢ
↙
Cℵˢ⇟
↛Principle of Dependent Choices (1.1.2)
↓
In ℝ, there are exactly 2ℵ₀ Borel sets.
↓
(2.3.4) In ℝ. there is a measurable set
that is Borel set.
Cℵ¹ m ≤ ℵ₁ or ℵ₁ ≤ m
↓ ↓
↛↛Principle of Dependent Choices (1.1.2)
↓
Lōwenheim-Skolem Theorem
↛ℵ₁ ≤ 2 ℵ₀
→Kōnig’s Infinity Lemma (4.5.6)
→Urysohn’s Lemma (4.6.6)
Principle of Dependent Choices (1.1.2)
↓
(1.6.6) If (M,<) has no
subset of type *ω, then
it is a well-ordering.
↕
数学联邦政治世界观提示您:看后求收藏(同人小说网http://tongren.me),接着再看更方便。