Interpretations and pertinent data on the clustering phenomena obtained in previous experiments are reviewed in order to determine why subjects reorganize material in recall. When variables which augment or diminish the effects of pre-established association and category relations on intellectual processes were contrasted, prominent relationships reduced subjective organization. In word pair experiments, category relations and strong association tendencies excluded idiosyncratic bases of organization. When associational and categorical relations between members of a word pair provided a basis for clustering in free recall alternative to the bases, subject effected subjective organization or idiosyncratic pairing. Although relatively highly associated pairs carried most of the clustering, synonyms did not represent effective categorical groups. Specific, conflicting, and inappropriate modifiers disrupted associational and categorical relationships, thus decreasing clustering. Since an induced set and the homogeneity of lists may influence the use of certain inter-word relations, the entire range of the intentionality variable, from incidental learning to specific, induced sets, should be examined in order to fix precisely the conditions under which pre-established associations and categories will and will not work.