interest groups.’’ More generally, they said they had ‘‘in mind phenomena
occurring within such organisations as legislatures, political executives, armies,
bureaucracies, churchesand the like’’ (Almond and Coleman 1960 , 33 ; emphasis
added). Very early in the book, they had simply said that ‘‘instead of ‘institutions,’
which again directs us towards formal norms, [we prefer] structures’’ ( 1960 , 4 ).
A few years later, in 1966 , in theirComparative Politics, Almond and Powell referred
in a similarly casual manner to ‘‘formal and institutionalchannels of access which
exist in a modern political system. The mass media constitute one such access
channel’’ ( 1966 , 84 – 5 ; emphasis in original). Parties, legislatures, bureaucracies, and
cabinets are then mentioned as being also formal and institutional channels.
Despite what has been frequently said about the attitude of ‘‘behaviorists’’
vis-a`-vis institutions, there is therefore not so much a negative approach to these
elements of political life as a ‘‘taken-for-granted’’ standpoint. Easton, in the index
to hisPolitical System( 1953 ), lists a number of points at which institutions are
mentioned (he does not do so at all in his subsequentA Systems Analysis of Political
Life( 1965 )), but in the body of the text, as when he discusses the works of Bagehot
or Bryce, the word institution is not even mentioned: he appears to assume that the
kinds of bodies which these authors referred to are ‘‘institutions.’’
Thus behaviorists did not deny that institutions had a role, but, by introducing
the ‘‘broader’’ notion of structure, Almond (and indeed, by introducing the notion
of system, Easton and Almond) brought about a distinction which was bound one
day to lead to questions about possible diVerences between these concepts. The
notion of institution was rendered controversial by the sheer fact that a second
notion was introduced without abandoning theWrst, but, in the 1960 s and 1970 s,
the point had not been reached when one could say, as Rothstein did in his
contribution to the New Handbook of Political Science of R. E. Goodin and
H.-D. Klingemann ( 1996 ), ‘‘Political Institutions: an Overview:’’ ‘‘whichever story
political scientists want to tell, it will be a story about institutions. A central puzzle
in political science is that what we see in the real world is an enormous variation,
over time and place, in the speciWcs of these institutions’’ ( 1996 , 134 – 5 ). Possibly the
Wrst text which truly raised the issue was that of Lawson,The Human Polity,
published in 1985 , where it is stated under the subtitle of institutions: ‘‘Aninstitu-
tionis a structure with established, important functions to perform; with well-
speciWed rules for carrying out these functions; and with a clear set of rules
governing the relationships between the people who occupy those roles’’ ( 1985 , 29 ).
The nature of the debate was of course transformed by March and Olsen’s work,
Rediscovering Institutions,publishedin 1989 , after the article entitled ‘‘The New Insti-
tutionalism’’ was published in theAmerican Political Science Reviewin 1984 .The
inXuence of the volume has been extraordinarily large, but the text is also extraordin-
arily laconic, not to say more than laconic, about what institutions are. TheWrst
sentence echoes the phraseology of Almond and Coleman, a generation earlier, the
words ‘‘such as’’ being used as substitutes for a deWnition: ‘‘In most contemporary
about institutions, mainly, but not exclusively, political 719