American Political Science Review Vol.81 likely,therefore,that the highly educated politics or public policy can nevertheless have many more unfounded preferences develop preferences by getting them from than do those who have far fewer prefer- Berelson,Lazarsfeld,and McPhee's(1954) ences about subject matter of which they well-known ("But,"as James Stimson know little. [1986]reminds us,"we keep forgetting it") Rational people,I have argued,support two-step flow of communication from their way of life.By answering two ques- activists to less attentive citizens.Stimson tions,they are able to discover their cul- (1986)shows that"mass perceptions track tural identity:Who am I?(free to negoti- activist positions."His thesis is that this ate or bound by a group?)and What social connection "accounts for the riddle should I do?(follow detailed prescriptions of inattentive electorates who seem to that vary with role or decide for myself?). know much of what they need to know to Knowing who they are and are not-the make policy informed choices."Wholly in cultures to which they do and do not the spirit of cultural analysis,Stimson belong-helps them to begin sorting their concludes "that many of the things that preferences.Cultural identity enables matter in political life...have very little individuals to answer for themselves the to do with individual psychological proc- crucial quantitative and qualitative ques- esses.They are macro behaviors,such as tions about preferences:How many are mediated cognitions,that require for they expected to have?What kind should understanding a focus on between'rather these be?Fatalists know that they do not than 'within'individual effects"(pp.4,19, need to know anything (it won't matter) 20).All I would add is that "between... except what others tell them to do.They individual effects"become "within indi- are prescribed to,not prescribing.Mem- vidual effects." bers of hierarchies can rule in whatever "System"or "person"blame are dead goes with their station and rule out what- giveaways.The slightest clue as to ever does not.By relying on others whose whether the authorities and the institu- duty it is to take care of whatever they tions vis-a-vis individuals are at fault neglect and by positive reinforcement of helps people know whether they want to this nonparticipation-it is normative not go along with egalitarian or hierarchical to act above (or,for that matter,below) or market policies.Anyone who thinks your station-both groups come to learn that attribution of blame to "the system" how much of what kind of preferences or to individuals is not diagnostic should they are expected to have and how much consult Table 2 from Verba and Orren,in they can leave to the authorities.Indi- which they address that very question to a vidualists are expected to figure out for variety of elites.The differences could themselves whether and to what extent hardly be greater(Wildavsky 1985b). participation is worthwhile.There is no If it were necessary to go back to the onus on nonparticipation.s cultural source each time a new preference Overall,it cannot be too difficult to is involved,building back up to the actual arrive at preferences on most matters, preference through some sort of chain of because everyone does it.Just as we con- inference,many people could not manage sider our connections with those who the complexity;hence there would be far advocate petitions as a quick way of fewer preferences.Consequently,con- determining whether we would feel com- certed political action would be a rarity. fortable in signing,so do people in general Near universal preference formation re- learn how to know what they ought to quires that preferences be inferred from prefer without knowing much about it. all possible directions.Culture is the India People who do not pay much attention to rubber man of politics,for it permits pref- 16