# How exactly the hidden state works in an RNN ? How to decide on how many past instances to consider?

I am unable to grasp the working of RNN because in different tutorials, it is explained differently.

Please correct me as I have considered that:

In a Many to one Model: if a sentence is like My name is Shady becomes Mon nom Shady then how are the inputs were fed to model? IMO it was that My produced Mon by multiplying with w0 and produced some hidden state as h1 by multiplying with a0. Then (h1 * a1)+ (name * w1) produced the nom. I want to ask that how is the state h changing? is h_i at the ith time a single float that the i+1_th input will use? If so,then why do they say that it uses previous N states when it is using a cumulative function of all the previous states that have already occured???

And if the sentence is paragraph of 10000 words, will the last input use cumulative function state which is a state defined by all the previous 9999 inputs??