The present paper will present an overview of the state of the art in language attrition research, attempting to provide an integrated picture of the attrition process in the light of Dynamic Systems Theory. Recent findings suggest that a fully developed L1 system is a powerful attractor state, as defined by Dynamic Systems Theory. Once this state has been attained, L2 influence or activation problems may cause small temporary shifts to the overall system, but it will generally settle back to its original state. Such a view could explain the dramatic differences observed between the L1 knowledge of pre- and post-puberty migrants. While speakers who have attained school age, but not puberty, will generally be able to use most grammatical features of their L1 in a target-like way, this knowledge may not yet have ‘settled’ into the mature attractor state, and therefore remain vulnerable to attrition.