It’s completely fine to specify priors for transformations of parameters & data, are you receiving some kind of error when you do so?
This specification should compile without issue:
(events[i] / pop[i]) ~ lognormal(mu[i], sigma)
The only situations where you need to be careful are when a non-linear transformation is applied, in which case you would need to add a jacobian adjustment.
For more information on this, see this forum post and this section of the Stan manual
EDIT:
If you want to specify a prior on (events[i] / pop[i])
and also use that quantity elsewhere in the model, you just need to construct it in the transformed parameters
block:
transformed parameters {
vector[N] rate = events ./ pop;
}
model {
rate ~ lognormal(mu, sigma);
}