Take a
Quantificational Modal Logical System that includes Brouwer, and allows free
variables. Call such a system
SQML- (since it is slightly weaker than SQML). It is a theorem of such a system
that <If possibly, only one thing exists, then only one
thing actually exists.>
Informal
Proof:
Suppose that there is a possible world W at which only one thing exists.
So at W it is true that x=y. The necessity of identity
states that ∀x∀y(x
= y → □(x = y)). So it follows that (x=y) →
□(x=y). But by Modus Ponens, it follows that □(x=y).
But by Brouwer, it follows that x=y is true at @.
[1][2]
[1][2]
Even if one thinks that abstract objects necessarily exist along with God, or that God has proper parts, ~TT is still a theorem of SQML-. This is at least because TT entails that (ii) possibly, the only essentially minded entity that exists is God, and (iii) possibly, something is essentially minded and numerically distinct from God.[3]
(Very) Informal Proof: BF, CBF, and □NE are all theorems of SQML- (cf. Menzel SEP proofs). So it is a theorem of SQML- that if something exists at a possible world, then it exists necessarily (i.e., there are no contingent entities). Given this and (ii), it follows that necessarily, some minded entity exists that is numerically distinct from God. This contradicts (i). So ~TT.
[1] Although this
would normally not be licit, since the domain at w is just a singleton set,
this is in fact a licit move.
[2] Kripke proved
this follows from QML and the premise that □∀x□(x
= x). And this premise is actually a theorem of SQML-. Observe: ∀x(x
= x). By the Rule of Necessitation (RN), it follows that □∀x(x
= x). By Converse Barcan Formula (CBF), which is a theorem of SQML-,
it follows that □∀x(x
= x) → ∀x□(x
= x). So it follows by Modus Ponens that ∀x□(x
= x). But by RN it follows that □∀x□(x
= x).
[3] Trinitarians
can read (i)-(ii) as follows: (i) possibly, the only essentially minded
entities that exist are the Father (F), Son (S), and Holy Spirit (HS); (ii) Possibly,
something is essentially minded and numerically distinct from F, S, and HS. The
proof would still, mutatis mutandis, go through.