# Logical expression that non-existent things don't do anything?

How could we express the idea that "something that doesn't exist cannot possibly do anything" using a logical argument? And if so, it is at all possible to prove that kind of proposition?

• Comments are not for extended discussion; this conversation has been moved to chat. – Philip Klöcking Jan 20 '19 at 13:57
• It depends on what you mean by 'exist' and 'do'. – Gary Reist Jan 23 '19 at 19:31
• @GaryReist Nothing out of the ordinary. Dictionary definitions. Are questions supposed to come with definitions of the terms used? And then, definitions of all the words used in those definitions? – Speakpigeon Jan 23 '19 at 19:36

I am not quite sure if I understand what "expressing an idea using an argument" means. I interpreted your question as " please formalise and discuss provability".

Using FOL (First-order logic)

You probably have noticed that in such a case formalising existence with the existential quantifier is no good. This is of course due to the fact that all terms denote something. For example, for the same reason it is inappropriate to formalise "Anton searches Pegasus" naively into FOL: a valid conclusion would be "there is something that Anton searches", but this is of course not true and thus unwanted.

A common solution is to use a unary predicate E meaning "real" existence, while the existential quantifier only signifies the domain of discourse.

``````E(x) ... x exists.
D(x) ... x cannot possibly do something.

∀x(¬E(x) → D(x))
``````

You will notice that I ignored the modalities. I think the intention of "something that doesn't exist cannot possibly do anything" is not so far off from "something that doesn't exist cannot possibly do anything". If you insist that it should be there,

``````D(x) ... x does something

∀x(¬E(x) → ¬◇D(x))
``````

You will buy all the trouble with quantified modal logic doing that.

Is anything of this provable? Of course not, you need to specify first how the D and E predicate behave.

An important remark

Note what that means: you are ending up imposing conditions on the existence and doing predicate, just to be able to prove what you want to prove. This does not achieve what you want (judging from this and your other questions in this forum). You are putting exactly in what you are getting out.

I think one cannot stress this point enough. There is another quite famous example: you might have heard of Gödel's ontological proof of god. (There is an excellent discussion of this by M. Fitting in his "Types, Tableaus and Gödel's God") The point is basically the same. Gödel introduces a "being positive" predicate in a complicated logic and then proves that there is something that has all positive properties and calls it god. The appropriate response to this is that one hasn't proved the existence of god but shifted the question to "why should I accept this logic?" While such an exercise might be interesting, it doesn't answer the question that one set out to settle.

"something that doesn't exist cannot possibly do anything" is a statement about the world. Logic gives an exact description of what follows from what, which is interesting, but it cannot prove anything about the world. That is simply not what logic was introduced for and one shouldn't use it for that.

• Yes, thanks, that's exactly what I wanted to verify. And, yes, garbage in , garbage out. But we don't have to insist on feeding garbage to it. Science seems like feeding empirical evidence into logical implications. Seems to work well enough. – Speakpigeon Jan 19 '19 at 20:20

By way of an example: let's assume that 'being identical to itself' is a property - in fact, the essential property - all things/objects enjoy. Now define a (unary) predicate, say, P, to mean an object is not equal to itself: obviously, no object enjoys this property P. Then you can easily (intuitionistically!) prove

For all x, P(x) implies Q(x)

for any phrase Q you fancy.

So, in fact, it seems more appropriate to maintain that something that doesn't exist can be/do absolutely anything!

• a.c.bruno "You can easily prove"? Could you make that proof explicit? – Speakpigeon Apr 19 '19 at 13:57
• (1) Assume x is a proper class; (2) use powerset or pair to prove there do not exist proper classes; (3) by ex falso, infer P(x); (4) close/discharge the hypothesis (1), noticing x was arbitrary, to conclude the universal statement – user35066 Apr 19 '19 at 14:25
• @Speakpigeon edited to make it clearer and considerably more general – user35066 Apr 19 '19 at 15:59