It seems to be asked textbook RSA with plaintext and ciphertext in the set $\{1,2,\dots,25,26\}$. That's wholy unreasonable, for in RSA the public key gives the mapping of plaintext to ciphertext, and with such small set it is trivially inverted. Let's see how it can be done nevertheless.
Textbook RSA operates on the set $\{0,1,\dots,N-1\}$ with $N$ the product of distinct primes. In order to use the desired set, I propose to use $N=26=2\times13$, and assimilate $0$ with $26$. We can now use textbook RSA, where encryption goes $C=P^e\bmod N$ ($P$ is the plaintext, $C$ is the ciphertext that the question calls $E$, and $e$ is the public exponent).
$\lambda(N)=\operatorname{lcm}(1,12)=12$, the public key $e$ must be coprime to that, and $e\times d$ must be one above a multiple of $\lambda(N)$, therefore we must have $e$ one of $1$, $5$, $7$ or $11$, with $d=e$ (within some multiple of $12$). $d=e$ means encryption and decryption are the same operation. Beside, ciphertext and plaintext are identical for $e=1$.
For example, with $e=5$, the plaintext $P=2$ is enciphered into $C=P^e\bmod N$, that is $C=2^5\bmod26$, that is $C=6$. And $C=6$ is deciphered into $P=C^d\bmod N$, that is $P=6^5\bmod26$, that is $P=2$.
We end up with a dumb mono-alphabetic substitution cipher with only four possible keys, one of which leaving plaintext unchanged, all with three fixed points $1$, $25$, $26$, and such that encryption and decryption are the same.