Can you generate confidence intervals for time series ETS forecast components?

Question

Suppose you fit a time series with the ets function from the forecast package in R:

library(forecast)
fit <- ets(USAccDeaths, model="MAM")
summary(fit)

# ETS(M,A,M) 
# 
# Call:
#   ets(y = USAccDeaths, model = "MAM") 
# 
# Smoothing parameters:
#   alpha = 0.4761 
# beta  = 0.0326 
# gamma = 1e-04 
# ...more output...

Is it possible to generate confidence intervals for the parameters (e.g. alpha, beta, gamma)? If methods exist, are they restricted to certain taxonomies? That is, can this method be extended to ARIMA models?