Suppose an SxT matrix Beta and SxS matrix G. Is there a problem with the

following?

for (s in 1:S) { Beta ~ dnorm(0, 0.0001) } # initialize first column

for (t in 1:T) { Beta <- G %*% Beta } # calculate remaining columns

getting an error earlier that the compiler can't resolve Beta, "undefined

ancestor node or directed graph"

thanks