In this paper, we analyze the historical process of emergence of proof assistants, and acceptance of proof assistants by the community of mathematicians. Our analysis is done by reflecting on how the notion of proof deepened through the proposal of formalism by Hilbert, formalization of the notion of computability, and especially development of type theory. In this analysis, we view mathematics as human linguistic activity where proofs are produced and communicated by means of both natural languages and formal languages. The complete acceptance of proof assistants by mathematicians is yet to be achieved, but we argue that it should and would happen by properly designing and implmenting a meta proof assistant which can talk and reason about any formal systems.