22.01.2014 Views

Metatheory - University of Cambridge

Metatheory - University of Cambridge

Metatheory - University of Cambridge

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

6. Completeness 55<br />

It will make our lives easier if we introduce a bit more terminology. Where<br />

Ω are some sentences, say that Ω are jointly polarising iff both: 6<br />

(p1) every sentence among Ω is either an atomic sentence or a negation <strong>of</strong> an<br />

atomic sentence; and<br />

(p2) if a sentence is among Ω, then its negation is not among Ω<br />

Where Ω are jointly polarising, say that A is in the field <strong>of</strong> Ω iff A is an<br />

atomic sentence such that either A or ¬A is among Ω.<br />

It is worth pausing for a moment to think about how this terminology,<br />

and this result, connects with SimpleSearch. When ∆ ⊬ ⊥, SimpleSearch<br />

generates a pro<strong>of</strong>-skeleton with at least one open line. From each open line,<br />

we can read <strong>of</strong>f some jointly polarising sentences, Ω. Moreover, when D is any<br />

sentence among ∆, every atomic subsentence <strong>of</strong> D is in the field <strong>of</strong> Ω. At this<br />

point, we claimed (in the pro<strong>of</strong> <strong>of</strong> Lemma 6.6) that there is bound to be some<br />

valuation that makes all <strong>of</strong> Ω true. That was right, and the more general result<br />

is simply this very easy observation.<br />

Lemma 6.8. Suppose Ω are jointly polarising. Then there is exactly one<br />

valuation <strong>of</strong> the sentences in the field <strong>of</strong> Ω that makes every sentence among<br />

Ω true.<br />

Pro<strong>of</strong>. Define a valuation, v, as follows: for each sentence A in the field <strong>of</strong> Ω,<br />

• v makes A true if A is among Ω<br />

• v makes A false if ¬A is among Ω<br />

Given (p2), v is well-defined. Now, if B is a sentence among Ω, then B is either<br />

A or ¬A, for some sentence A in the field <strong>of</strong> Ω. Either way, v makes B true.<br />

Moreover, any valuation that assigned a different value to A would make B<br />

false. Hence v is unique.<br />

■<br />

Continuing, though, to think through how SimpleSearch relates to this terminology:<br />

the next stage in the pro<strong>of</strong> <strong>of</strong> Lemma 6.6 was for us to show that the<br />

(unique) valuation that makes all <strong>of</strong> Ω true also makes all <strong>of</strong> ∆ true too. This<br />

required a bit <strong>of</strong> work (some <strong>of</strong> which I left as an exercise). But let me now<br />

part company with the pro<strong>of</strong> <strong>of</strong> Lemma 6.6, and instead make a much more<br />

simple observation.<br />

We are considering a situation where all <strong>of</strong> the atomic subsentences <strong>of</strong> some<br />

sentence (in this case, D) are in the field <strong>of</strong> some jointly polarising sentences<br />

(in this case, Ω). We can obtain a very easy and general result about such<br />

situations.<br />

Lemma 6.9. Suppose Ω are jointly polarising and that every atomic subsentence<br />

<strong>of</strong> A is in the field <strong>of</strong> Ω. Then either Ω ⊨ A or Ω ⊨ ¬A.<br />

Pro<strong>of</strong>. Using Lemma 6.8, let v be the unique valuation <strong>of</strong> the sentences in the<br />

field <strong>of</strong> Ω that makes all <strong>of</strong> Ω true . Either v makes A true, or v makes A false.<br />

If v makes A true, then since v is the only valuation that makes all <strong>of</strong> Ω true,<br />

Ω ⊨ A. If v makes A false, then it makes ¬A true, and similar reasoning shows<br />

that Ω ⊨ ¬A.<br />

■<br />

6 NB: this terminology is not standard among logicians.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!