It means that these filters will exist in one instance per voice. Polyphony = multiple “voices” (chords etc.) at once. Polyphonic filters means each voice (i.e. every note) will have its own (virtual) instance of such a filter, with independent cutoff and resonance. Closest you can get with “monophonic filters” is using several instances and spread the notes along them. It’s like the difference of pumping a chord through one filter (monophonic) or evey single note of that chord through its own filter (polyphonic) before they are mixed together.