Substance and reality

Bill has an interesting post here on the notions of reification and hypostatization in the course of which he says a couple of things which attracted my attention.

Firstly, he says

Or consider the internal relation being the same colour as. If two balls are (the same shade of) red, then they stand in this relation to each other. But this relation is an "ontological free lunch" not "an addition to being" to borrow some phraseology from David Armstrong. Internal relations have no ontological status. They reduce to their monadic foundations. The putatively relational fact Rab reduces to the conjunction of two monadic facts: Fa & Fb. To bring it about that two balls are the same colour as each other it suffices that I paint them both red (or blue, etc.) I needn't do anything else. If this is right, then to treat internal relations as real is to commit the fallacy of reification. Presumably someone who reifies internal relations will not be tempted to hypostatize them.

There seems something odd about this.  Let's suppose Armstrong is right.  Then the internal relation being the same colour as is not real and it's a fallacy to think it is.  Does this mean that the balls are not the same colour?  In what sense is the relation not real?  Is it imaginary?  Do relations fall into the exclusive categories of the real and the unreal?  I have to confess that I'm struggling to understand this.  Here's an argument in favour of the reality of being the same colour as:  it makes an existential commitment.   To say that a is the same colour as b is to say that there exists a colour C such that Ca & Cb.

Secondly, he says

To hypostatize is is to treat as a substance what is not a substance. So the relation I just mentioned would be hypostatized were one to consider it as an entity capable of existing even if it didn't relate anything. Liberals who blame society for crime are often guilty of the fallacy of hypostatization. Society, though real, is not a substance, let alone an agent to which blame can be imputed.

It struck me that the relation of crime to society is closely analogous to that of accident to substance.  One might say that (a) crime is an accident on society.   But if society is not a substance what 'ontological status' are we to grant to society and crime?