T HIS paper presents some observations on public opinion and on public opinion polling as currently performed. It is hoped that these observations will provoke the discussion for which, I understand, this meeting has been arranged. The observations are not along the line of what seems to be the chief preoccupation of students of public opinion polling, to wit, the internal improvement of their technique. Instead, the -observations are designed to invite attention to whether public opinion polling actually deals with public opinion. The first observations which I wish to make are in the nature of a prelude. They come from a mere logical scrutiny of public opinion polling as an alleged form of scientific investigation. What I note is the inability of public opinion polling to isolate "public opinion" as an abstract or generic concept which could thereby become the focal point for the formation of a system of propositions. It would seem needless to point out that in an avowed scientific enterprise seeking to study a class of empirical items and to develop a series of generalizations about that class it is necessary to identify the class. Such identification enables discrimination between the instances which fall within the class and those which do not. In this manner, the generic character of the object of study becomes delineated. When the generic object of study is distinguishable, it becomes possible to focus study on that object and thus to learn progressively more about that object. In this way the ground is prepared for cumulative generalizations or propositions relative to the generic object of investigation. As far as I can judge, the current study of public opinion by polling ignores the simple logical point which has just been made. This