This paper examines Wittgenstein's conception of a language game. It is contended that there is, as yet, no linguistic science and that a great deal more deep philosophical imagining and "playing" is needed before we have enough knowledge to develop a real science of language. Similarly, superficial and over-literary British re-interpretations of Wittgenstein's original thinking are disputed. To make the conception of a language-game clearer, Wittgenstein's first five interconnected language-games, given in the "Brown Book," which illustrate St. Augustine's child-language consisting only of names, are analyzed in detail. In progressing from Game 1 to Game 5, Wittgenstein variegates the notion of the child learning ostensively--and of subsequent relearning by yet more demonstration--and progressively widens the notion of a name. Speech is finally compared with a continually changing sequence of language-games, each game being complete in itself, and the flow of speech itself giving the cues for change from game to game. This comparison is asserted to be Wittgenstein's first insight. (Author/FWB)