EDB — 02M

view in whole PDF view in whole HTML

View

English

E87

[02M]Difficulty:*. 1 Consider the following quotient of the family of all integer valued sequences

\[ {\mathbb X}=\{ a:ℕ→ℕ \} / ∼ \]

where we define \(a∼ b\) iff \(a_ k= b_ k\) eventually in \(k\).

We define the ordering

\[ a⪯ b \iff ∃ n \mbox{ s.t. } ∀ k≥ n, a_ k≤ b_ k \]

that is, \(a⪯ b\) when \(a_ k≤ b_ k\) eventually. This is a preorder and

\[ a\sim b \iff (a⪯ b\land b⪯ a) \]

so it passes to the quotient were it becomes an ordering, see Prop. [1Z7].

Let \(a^ k\) be an increasing sequence of sequences, that is, \(a^ k⪯ a^{k+1}\); we readily see that it has an upper bound \(b\), by defining

\[ b_ n = \sup _{h,k≤ n} a^ k_ h ~ . \]

We can then apply the Zorn Lemma to assert that in the ordered set \(({\mathbb X},⪯)\) there exist maximal elements.

Given \(a,b\) we define

\[ a∨ b = (a_ n∨ b_ n )_ n \]

then it is easily verified that \(a⪯ a∨ b\). So this a direct ordering, see [06N].

We conclude that the ordered set \(({\mathbb X},⪯)\) has an unique maximum, by [06S].

This is though false, since for any sequence \(a\) the sequence \((a_ n+1)_ n\) is larger than that.

What is the mistake in the above reasoning? What do you conclude about \(({\mathbb X},⪯)\)?

Download PDF
Bibliography
Book index
  • eventually
  • preorder
Managing blob in: Multiple languages
This content is available in: Italian English