At 368, Reppert argues: If the reference of our terms is indeterminate, then this has the disastrous consequence that we cannot reason to conclusions.
This is surely wrong. Reasoning can be purely formal. (If all flombs are bloops, and all bloops are shimbs, then all flombs are shimbs. The reasoning is impeccable. We can reason even if our terms have no meanings!) Moreover — and perhaps partly in consequence — so long as we restrict ourselves to a single context, and use the same word throughout, we can reason perfectly well even if our meanings are indeterminate. (If this is a rabbit, and that is a distinct rabbit, then there are at least two rabbits. Fine, regardless of Quinean indeterminacy in the meaning of “rabbit”!)
VR: But that's just the problem. Without fixed reference, we don't know whether we have three terms, four terms, five terms, or six terms.
Let's take the following argument.
1. All jackrabbits are gavagais.
2. All gavagais are mammals.
3. Therefore all jackrabbits are mammals.
This argument is valid if the reference of gavagai is invariant between premise 1 and premise 2. But what if in premise 1 gavagai means "rabbit", but in the second it means "undetached rabbit part." Then the argument is invalid.