A study was conducted to test the validity of the Linguistic Features Hypothesis, which predicts that unmarked personal pronouns will be learned first by children, yielding the following order of acquisition: first person before second, followed by third; singular before plural; and subjective case before objective or genitive. A review of the literature generally supports this view. This study, motivated by the lack of a comprehensive approach to pronoun acquisition research, studied comprehension data for "I,""you,""he,""she,""it,""we," and "they." The subjects, 30 middle-class male and female English-speaking children from 2 1/2 to 4 1/2 years of age, were tested for comprehension of the genitive and objective cases through commands eliciting behavioral responses, and for comprehension of the subjective case through commands eliciting verbal responses. No correct responses were found to "it," and this pronoun was dropped from further analysis. The oldest subject group (4 1/2 years) performed better than the two other groups (3 1/2 and 2 1/2 years). No support was found for the prediction that first person precedes second. However, singular forms were comprehended significantly better than plural ones. Finally, no effects were found for case. Thus only partial support was found for the Linguistic Features Theory. (AM)