1: \begin{abstract}  

2: We obtain optimal Gaussian concentration bounds (GCBs) for stochastic chains of unbounded memory (SCUMs)

3: on countable alphabets. These stochastic processes are also known as ``chains with complete connections''

4: or ``$g$-measures''. We consider two different conditions on the kernel: (1) when the sum of its oscillations is less than one, or (2) 

5: when the sum of its variations is finite, {\em i.e.}, belongs to $\ell^1(\N)$. 

6: We also obtain explicit constants as functions of the parameters of the model.

7: The proof is based on maximal coupling. Our conditions are optimal in the sense that we 

8: exhibit examples of SCUMs that do not have GCB and for which the sum of oscillations is strictly larger than one, or the variation 

9: belongs to $\ell^{1+\epsilon}(\N)$ for any $\epsilon > 0$.  These examples are based on the existence of phase transitions. We also 

10: extend the validity of GCB to a class of functions which can depend on infinitely many coordinates.

11: 

12: We illustrate our results by three applications.

13: First, we derive a Dvoretzky-Kiefer-Wolfowitz type inequality which gives a uniform control on the fluctuations of the empirical 

14: measure.

15: Second, in the finite-alphabet case, we obtain an upper bound on the $\bar{d}$-distance between two 

16: stationary SCUMs and, as a by-product, we obtain new (explicit) bounds on the speed of Markovian approximation in $\bar{d}$. 

17: Third, we obtain exponential rate of convergence for Birkhoff sums of a certain class of observables.

18: 

19: \medskip

20: 

21: \noindent\textbf{Keywords:} concentration inequalities, maximal coupling, chains of infinite order, $g$-measures, categorical time series, empirical 

22: distribution, Dvoretzky-Kiefer-Wolfowitz type inequality, $\bar{d}$-distance, Markovian approximation.

23: 

24: \end{abstract}

25: