A Reflection on Russell's Ramified Types and Kripke's Hierarchy of Truths

Both in Kripke's Theory of Truth ktt 8] and Russell's Ramiied Type Theory rtt 16, 9] we are confronted with some hierarchy. In rtt, we have a double hierarchy of orders and types. That is, the class of propositions is divided into diierent orders where a propositional function can only depend on objects of lower orders and types. Kripke on the other hand, has a ladder of languages where the truth of a proposition in language Ln can only be made in Lm where min. Kripke nds a xed point for his hierarchy (something Russell does not attempt to do). We investigate in this paper the similarities of both hierarchies: At level n of ktt the truth or falsehood of all order-n-propositions of rtt can be established. Moreover, there are order-n-propositions that get a truth value at an earlier stage in ktt. Furthermore, we show that rtt is more restrictive than ktt, as some type restrictions are not needed in ktt and more formulas can be expressed in the latter. Looking back at the double hierarchy of Russell, Ramsey 11], and Hilbert and Ackermann 7] considered the orders to cause the restrictiveness, and therefore removed them. This removal resulted in Church's Simple Type Theory stt 1]. We show however that orders in rtt correspond to levels of truth in ktt. Hence, ktt can be regarded as the dual of stt where types have been removed and orders are maintained. As rtt is more restrictive than ktt, we can conclude that it is the combination of types and orders that was the restrictive factor in rtt.