We characterize the collective and single asset phenomena of a liquid market. We show that neglecting interactions between different assets prevent from grasping the deep structure of the underlying process, which is intrinsically multivariate. The introduction of a mapping from an N asset market to a liquid of N particles, which is studied by means of its (known) kinematic behavior, allows us to separate the background assets dynamics contribution from a truly single asset one. The detrended market liquid behaves effectively as an interacting particle system and, consequently, the pair interaction can be statistically characterized. As an application of our model we present the results obtained by an empirical analysis of the 30 assets belonging to the German stock market index DAX (Deutscher Aktienindex) for the period of time 1987-1999, and the correspondent interaction potential among asset pairs.