The phenomena of null subjects in child grammars of English are examined in the context of Nina Moss Hyams' proposals about these structures within the framework of generative grammar. Some problems with these analyses are examined and an alternative analysis is proposed. It is noted that Hyams predicts that children learning a language requiring lexical subjects will produce sentences with nonlexical subjects before they reset that parameter; two flaws in this analysis are described. An alternative account proposes that the absence of lexical subjects in early English production is not a consequence of the pro-drop parameter but the result of the inability of child grammar to case mark the subject. It is argued that the nominative case is assigned by the addition of tense but that child grammars are unspecified for tense. It is further argued that lexical subjects emerge when the child has acquired the feature of tense, and the trigger is the presence of modals in the input. (MSE)