A lot of the ideas that have been attributed to Aristotle over history were actually later interpretations of his work. One such notion is that Poetry (meaning “literature” or even “art in general”) is imitation, which has been for centuries the foundation of artistic judgments of value based on realism. You’ve heard it before: good art is a faithful representation of its subject, and the less the representation resembles the object, the worst the artistic work is. Perhaps you might even share that opinion. Whether that is right or wrong is not relevant here. The point is that the association of realistic representation with Aristotle’s theories is debatable, and might put too many words in the mouth of the long-deceased philosopher.
“Imitation” is the most common translation of what Aristotle called mimesis. That translation is also debatable. Chapter IV of the Poetics gives the following explanation:
Objects which in themselves we view with pain, we delight to contemplate when reproduced with minute fidelity: such as the forms of the most ignoble animals and of dead bodies.
It’s easy to take that “minute fidelity” for granted. However, the very beginning of the Poetics argue that all art is imitation, including music and dance:
(…) even dancing imitates character, emotion, and action, by rhythmical movement.
(http://www.gutenberg.org/files/1974/1974-h/1974-h.htm#2H_4_0003. Emphasis added.)
Anyone can judge quite easily whether the representation of a likeness or an action is faithful – you can look a portrait and say it doesn’t resemble the person, or watch/read something “based on a true story” and find out it was only loosely based. But how do you judge the fidelity of the representation of an emotion? Take jealousy, for instance. Both Othello and Tape represent jealousy in different ways, and both plays have enjoyed popular and critical success. Which is the most faithful?
There’s more to mimesis than mere fidelity. Perhaps the theory in the Poetics is that representation doesn’t need to be realistic, but believable – which ties in with the idea that “it is not the function of the poet to relate what has happened, but what may happen” (http://www.gutenberg.org/files/1974/1974-h/1974-h.htm#2H_4_0011). Other passages in the Poetics show more concern with some criterion of propriety instead of fidelity. Chapter II, for instance, ends with the classical distinction that “Comedy aims at representing men as worse, Tragedy as better than in actual life” (http://www.gutenberg.org/files/1974/1974-h/1974-h.htm#2H_4_0004). Not “as they are” in actual life.
Now, I can’t read Greek, but I’m convinced that fidelity is not so black-and-white in the Poetics, and that “imitation” is a very imprecise translation for mimesis. I could argue further, but then I would have to write a long, boring academic paper. And perhaps learn Greek. For now, I only want to argue that when someone says “Aristotle says this-and-that”, the truth is more likely to be “Professor So-and-So wrote that Kant said that Boileau believed that Horatio thought Aristotle meant this-and-that”. We're playing a two-thousand-year-old game of Broken Telephone, really. It’s best to make up your mind by actually reading the Poetics (it’s only forty pages long, anyway).