I'd suggest you don't. Gibbs' attempt at deciphering the Voynich manuscript wasn't bad simply on the account of it's premises, but because it was something lazily slapped together without much further thought.
But a "logographic" approach to deciphering the manuscript might not be off the table.