By means of decomposition we investigated using an experimental design (n = 160) the impact of combinations of processing and retention requirements on performance. Retention-during-processing spans (i.e., dual task requirements with equal importance of both evaluated components) were compared with the "reading" (final word) span of read sentences. In a correlational design, the relationships of these intentional spans and a unintentional indicator of working memory (the pronoun inference distance) were compared with fluid intellectual performances. The effects of processing on the span may not be restricted to articulatory suppression of rehearsal. While the substitution of retention of an additional processing-independent item improved the span, this did not hold for items resulting from processing. Thus, the span is determined by coordination of retention with processing, but neither by an activity without retention nor by retention per se. Coordination requires resources in addition to those required by the activities to be coordinated. Strategies of coordination differing in their efficiency result in different spans. Both unintentional processes and intentional ones contribute to the dual task performance of the retention-during-processing type. The best predictors for fluid verbal and nonverbal intellectual performance are the unintentional (pronoun) inferences and the complex intentional retention-during-processing span. The results are in favor of a working memory model in terms of a set of functions.