To expand upon my yesterday's comment ("Features, Not Tribes."), which i see now was part of a crossbloggal discussion of this, i see the present situation as one in which poets choose (for temperamental or personal-history reasons) from among several presented "dialects" of poetry (& not excluding many of the poetries of the past, or from other cultures); our eclecticism has the cross-dimension of being more pure or more mixed, but all the same it does very little to originate. And of course we choose our tribes, also, & analogously: & one of the reasons we use a certain dialect is that we think we have something in common to discuss with others who use it (see my essay on Blixen).
   But tribal divisions only map onto dialects insofar as poetic centers enforce (or fail to enforce) the dialects that correspond to their perceived canon. And one cannot talk about blixen, to return to this designation for 'dispersed groups', without recognizing that hardly anyone belongs to just one blik, nor does a large percentage exactly correspond to each other--although correlations can certainly be found.
   Further complicating this is the small fact that our animosities mostly map onto mythical entities. (The "avant garde" is only one such.)
   But if we start talking about features, how much less danger is there of falling into name-calling or reifying! For instance, here is one feature: linebreaks/ no linebreaks (="prose poetry"). That doesn't imply anything about the poet's subject matter, sexual political religious orientation, or place of origin... Other features i find interesting: wide parataxis/ limited parataxis; monolingual/polylingual; human-generated/computer-generated; persona/non-persona...
   Just the other day i invented a new form. I take a number of rhime-equivalent pairs (words whose A=1, B=2...Z=26 sums are the same number); & use them as endwords in any order. Additionally, the poems i wrote in this form counted syllables--most of them were in "snowflakes", where i use lines of 3, 5, 7, 9, & 11 in any order (then repeat). --I haven't named it yet.
   Syntax was fairly straightforward, with slight disjunctions between lines and/or stanzas; a few of the referents were esoteric, but most of the poems was intelligible, if a bit vague. I think of the basic dialect as post-Symbolist, with a smattering of Langpo. These poems, as with most of my other output, were not "addressed" as such; they refrain from the tribal. And while rhime is a procedure i have used before, i have only used it in couplets before now.
   Is this "experimental"? Is it "avant garde"? Is it "Quietudinous"? Is it "Confessional"? Is it "NeoFormalist"?
   Is it "Flarf"? Is it "Umbrist"? Is it "Xenon Mercurist"?