Twenty-four language-disordered children were trained, through modeling with a problem-solving set, to produce a question form involving a wh- word-who, what, or where-in a structure requiring either auxiliary is or auxiliary does. Results indicated the subsequent use of the trained auxiliary across wh- words and, to a lesser extent, the subsequent use of untrained as well as trained wh- words. These results suggest that the training of one multioperation structure may result in the acquisition of two partially independent linguistic operations. Such patterns of acquisition could result in an increase in the efficiency of language training.