Star formation is often described as the outcome of a balance between gravity and turbulence. In this talk, I will argue that this picture is fundamentally misleading. I will show that molecular cloud turbulence is not the agent regulating star formation, but largely a manifestation of gravitationally driven flows. Molecular clouds, clumps, and cores participate in a global hierarchical collapse, forming a multiscale accretion flow. Within this picture, several long-standing puzzles dissolve: the origin of the Schmidt–Kennicutt relation, the apparent constancy of the star formation efficiency per free-fall time, and the scaling laws of molecular clouds. In addition to the molecular-line traditional indicators of molecular cloud dynamics (mass, linewidths), I will show how the stellar kinematics of young stars can give us clues of the actual molecular cloud dynamics. Finally, if time allows it, I will discuss recent insights on whether MOdified Newtonian Dynamics (MOND) plays a role in the dynamics of molecular clouds and young stellar clusters.
