The next state in the process depends solely on the current state and the transition probabilities.
In a Hidden Markov Model, each state has a set of possible outputs that can be generated.
As with the transitions, each output also has a probability associated with it.