A study on sentence processing involving rehearsal, response initiation, and mental transformation required the subject to perform a secondary task concurrently with transformation of a sentence previously given. A secondary task interruption technique was used to allow inferences as to where the planning occurs by measuring increases in task time or pausing time. The primary question of interest was whether cognitive operations in terms of planning were carried out prior to the verbalization of the sentence or applied concurrently as the sentence is produced. The results support the contention that planning in terms of cognitive operations does take place prior to verbalization. The experimental task (transformations) involved lists from a pool of simple, active, affirmative, and declarative sentences together with their negative and passive transformations. The secondary tasks were generated on semantic and syntactic levels. It was found that the location of interruption is the most important variable in the length of secondary task performance time and that secondary task duration times increase with the difficulty of the transformation being made, but not differentially within location. Both location and transformation were found to affect pausing time; the initial pause location and the passive transformation contributed the greatest effect. (SW)