First, acquiring enough tokens to singlehandedly determine outcomes of votes is incredibly expensive. The simple act of buying tokens increases demand on the tokens (which have a fixed supply) thus increasing the market price. By the time a bad actor has acquired 51% (or close to it) of the tokens, they will likely have spent the equivalent of hundreds of millions of dollars.

Second, even if they are able to acquire some huge percentage of the tokens, the public nature of the blockchain means we (the public) can actually do some pretty fancy analytics to determine if there's collusion between accounts and identify bad actors (if not by name, at least by account). The community can then decide to "fork" the bad actors out of the system (Essentially the community can deploy a new token that is identical to the old token except the bad actors' account balances are zeroed out. If everyone decides to switch to using the new token, the bad actors' old tokens become worthless and they just spent millions of dollars to slightly inconvenience people).

Specific details of our initial token distribution (how many tokens does 1 ETH buy, etc) are forthcoming, but after that initial distribution, the price of the tokens is determined by the market forces. It's those market forces that (theoretically) make the TCR actually "work" as token holders are incentivized to make the registry as "good" as possible and increase demand for the token (because people want to be on the registry).

-- Nick Reynolds