We present a new dynamical picture that identifies the formation of the exoticcc-containing states XYZ with the confinement-induced hadronization of a rapidly separating pair of a compact diquark and antidiquark. This picture combines the advantages of diquark-based models, which can accommodate much of the known XYZ spectrum, with the experimental fact that such states are both relatively narrow and are produced promptly. It also naturally explains the preference of some of the exotic states to decay to ψ(2S), rather than J/ψ, in terms of a simple wave-function overlap effect. [3]. These states do not fit into the standard nonrelativistic quark model of a singlecc pair with separation r interacting via a potential V (r), which had been successful in accommodating all of the previously known charmonium states [4-6]; instead they are believed to be tetraquark (ccqq ′ ) states currently named X, Y , or Z 1 . Notable evidence supporting this identification includes the facts that X(3872) is an extremely narrow (Γ < 1.2 MeV) J P C = 1 ++ state but is tens of MeV lighter than the nearest quark-model candidate χ c1 (2P ) [6], and the recent confirmation at LHCb [7] of the charged J P = 1 + state Z(4430) as a resonance decaying into π − ψ(2S). This first verification of the existence of exotic hadrons, which possess neither meson (qq) nor baryon (qqq) valence structure, is an exciting advance for QCD; a key challenge is to uncover the dynamical structure of these states.One can imagine the binding of a (q 1 q 2q3 q 4 ) state to occur in a variety of ways. First, the four valence quarks can all interact democratically, which one may * Electronic address: sjbth@slac.stanford.edu † Electronic address: dshwang@sejong.ac.kr ‡ Electronic address: richard.lebed@asu.edu 1 In the current nomenclature, the neutral(charged) states observed in B decays are labeled X(Z), whereas the Y are the neutral, J P C = 1 −− states observed in initial-state radiation e + e − processes.call a "true" tetraquark. However, simple SU(3) color group theory shows that the combination of two quarks (each a color 3) and two antiquarks (each a color3) can form an overall color singlet in only two independent ways-matching the color structure of factorized two-meson states (q 1 q 2 )(q 3 q 4 ) and (q 1 q 4 )(q 3 q 2 ). In large N c QCD, this fact has long been used to argue that narrow tetraquark states do not occur, since the four-quark source operators needed to create them are saturated by two-meson states. Weinberg has recently showed [8] that this argument contains a loophole; however, his scenario requires modifications that go beyond the usual large N c counting rules and structures [9][10][11].The color factorization property of the tetraquark naively suggests a two-meson molecule structure for the observed states. Indeed, many of the XYZ states lie close to such thresholds (e.g., m X(3872) ≈ m D + m D * ≈ m J/ψ + m ω ), suggesting a molecule with a small binding energy E b via a van der Waals-type attraction [12]. However, the typical scatteri...