Previous computerized models of child language acquisition have sought to determine how children acquire grammatical word categories (GWCs). The current study seeks to determine if statistical structure can be corroborated as a factor in GWC acquisition. Previous studies examining statistical structure have dealt with word order rather than GWC order and only examined an overall success rate. The present study examines how well a computer model of child acquisition of GWCs was able to reorganize scrambled sentences back into the correct GWC order using transitional probabilities extracted from adult language input. Overall, a 50% success rate was obtained, but when broken down by utterance length, utterances up to eight words in length had a success rate much higher than chance. Thus, it is likely that statistical structure informs children's acquisition of GWCs.