"Of the many tendencies imputed to symbolism this is the most characteristic -- out of an acuter perception of what all poets have always known, that words are insufficient if their power is bounded by their meaning, emerged an audacious doctrine which branded their representative function as inferior, and sought to shift the poetical interest from what they signify to what they may suggest. In the Parnassian system description was paramount, and feeling sprang from it immediately: the emotion which symbolism pursues bears no constant relation to the objects represented or the ideas expressed; rather it aims at the recovery of vanished moods by curious incantations, by the magical use of verbal atmosphere. To fashion a true likeness of the material world it holds a vain and illusory undertaking: It values sights, sounds, scents, and savours for their secret affinities with states of the soul .... "Three years of on-and-off study, and I still can't quite figure out what I think of the symbolists. Plato, Aristotle, Descartes; Eliot, Woolf, Faulkner, and the French classicists: the philosophers and the "idea" artists are a piece of cake ("idea" artists is a careless term that I don't care to correct beyond saying that I'm talking about "artists who have rational ideas, though their art is not limited to the rational; I'm not talking about platonists). It's these irrationalists that confuse me, these strange artists who believe in some sort of incantatory art--the symbolists--or the anti-rationalists--the Dadaists, surrealists, even (in my opinion) the more extreme existentialists. There's nothing more frustrating to an academic than not "understanding" something. Artists like the symbolists make good art and write terrible theory.
And is this terrible theory? Honestly, I'd be inclined to think so. The Parnassians tended to not be such good writers. They were more like skilled minaturists whose gorgeous description rarely attained its actual goal: that of making "feeling spring from description". But while the symbolists claimed to reject description, the idea that one can use words without "describing" anything is pure nonsense. Words are descriptive in a fundamental way: what you're doing when you're speaking is describing concepts which without language would be unformulated, and without commonly understood words would be incommunicable. These concepts in turn do not derive from some "idea cloud" floating somewhere overhead (sorry Socrates), nor do they have their roots purely in the individual psyche, at least not in any practical sense (even if one is to admit the rational possibility that our interaction with everything around us is "in our heads"--sorry Gilbert Ryle--we still act as though it were real). Concepts have their roots in those sights, sounds, scents, and savours that the symbolists value, but you can't separate these things from their physicality. If you want to communicate their "secret affinities with states of the soul," you have to deal with the thing which has the affinity as it is. To evoke these affinities you don't, it's true, want to have the physical itself as the final object of description. But you will need to use the physical as a means to communicate that final object, if only because concepts are incommunicable if undescribed--that is, if unarticulated.
In short the symbolists a.) rejected Parnassian theory, but b.) went right ahead and put Parnassian theory into practice in a much more vigorous manner than the Parnassians did. The claim to use words "magically" (and yes, there was plenty of occult background chez some of those fellows; more the painters than the poets) to move beyond the material world is a frustratingly illogical one. ("Magic? Really? We've come to that now, have we?") But in point of fact, the symbolists accomplished something a bit more rational than their theory would indicate (which is why someone like the eminently rational T.S. Eliot claim to have found his artistic voice through reading symbolist poetry). They rejected Parnassian theory because the Parnassians applied their theory in a limited manner. For them, description was something like what Flaubert understood description to be, which is great for a novel, but I think rather stupid in poetry. The symbolists realized that the power of words to evoke is not restricted to situations in which those words are being used to describe a specific thing. Rather, you can use them as elements, notes, say in a musical composition. You do need phrases (we're not talking about atonal-ism yet), but you don't need to limit and order those phrases to form a description (think classical music) or to a narrative (think Romanticism in both music and poetry). You simply need to design a progression of moods, which cannot be achieved without recourse to description, but which need not get bogged down in one particular description or another. The perfect musical metaphor for this? Debussy. It's no wonder that they were contemporaries. Nor, to be sure, that japonisme was all the rage in Europe between 1870-1914.