an early stage in the ICPP project--the passage of time obviously can mute and even nullify the effect of extraparliamentary origin on a party's governmental status, witness the British Labour Party, which had an "outside" origin. Therefore, while we continued to code the variable, we came to regard it as a variable to be related to governmental status in a test of the original Duverger hypothesis instead of a "basic variable" within the governmental status cluster.
According to the underlying theory, the most "inside" parties are judged to be those formed by the chief executive of the government for the purposes of legitimizing his leadership or expanding popular participation in his regime. These should be distinguished from parties which are formed by the chief executive for the purpose of introducing responsible but perhaps limited interparty competition. Parties formed by groups of legislators acting as a bloc in parliament are still "inside" the government, of course, but less so than those just cited, and parties formed by individual government officials--for example, dissident legislators, governors, cabinet members-are still further removed.
Parties formed by former governmental officials are the most "inside" of the outside parties, just ahead of parties formed by leaders of prominent, powerful, and legal social organizations that themselves are institutionalized in the society, for example, popular religious groups and large-scale labor unions: These institutions have their own stake in society and are apt to be recognized as representing more legitimate claims than those represented by less powerful social institutions, which may also sponsor the development of parties, but ones which are still further outside the government. Finally, social organizations that have been outlawed or declared illegal, regardless of their size and influence, create parties that are removed from government nearly at the extreme, exceeded only by parties formed by nationals of another country.
One complicating factor in the concept of the locus of party origin enters when we consider parties that are not formed anew but arise from party splits or mergers. When the splits or mergers involve parliamentary members, no problem arises, but one does when the originating parties are "outside" the government. Both may best be handled by regarding the originating parties as social organizations and classifying the status of the merged/split party according to the legality and power of the originating parties, with a premium given to a large size and illegality in obtaining high scores.
Operational Definition. The higher the score on this variable, the more "outside" the condition of origin. The party is coded as to the status of its originators at the time of its creation. Thus, this variable--like those in the institutionalization cluster--receives only a single code for status at time of origin rather than one code for each half of our 1950-1962 time period. When more than one code was applicable, the lowest (i.e., the most "inside") code was assigned.
Coding Results. Table 4.7 for BV207 displays a distribution for "outside origin" that is nearly bimodal, almost half of the parties being formed either by legislators in office (code 4) or by leaders of major social organizations (code 8). If we divide the scale between codes 4 and 5 to conform to Duverger's inside/outside dichotomy, we find that almost two-thirds of the parties in the study were formed outside rather than inside of the government. Substantive analysis of these data should provide a direct test of the Duverger hypotheses as a task for some later time. If one were to make the test, he need not be concerned about the correlation between BV207 and AC207, which was nonsignificant.