This control scheme, introduced by O. Khatib already in 1987, enables to control the endeffector in task-space while simultanously tracking a secondary (e.g. joint-space) task, which has null effect on the primary task.
When controlling highly redundant robots such as humanoids, this control framework can be extended to a multi-level hierarchy, as proposed by L. Sentis.

Literature: Khatib et al. and Sentis et al.