¹ See his paper above, pp. 435–6. This conclusion is puzzling to say the least: either the appropriate tests are positive or they are not, and, accordingly, the theory is good or not. One can assess how crucial a factor will be not from some implicit average of the number of negative and positive tests, but from the strength of the relationship observed in valid tests.

² Lemieux, Vincent, “Les dimensions sociologiques du vote créditiste au Québec,” Recherches sociographiques, VI (1965), 181–95.CrossRefGoogle Scholar For his view about the tests of the theory at the provincial level, see his “Réponse,” Recherches sociographiques, VII (1966), 363–5.

³ Despite the presence of a strong two-party system (the Liberals in opposition had previously kept 47 per cent of the votes), two new parties together polled in that election 34 per cent of the votes (the Parti Québécois, 23 per cent, and the Ralliement Créditiste, 11 per cent).

⁴ The general treatment of these conditions is of course to be found in Smelser, Neil J., Theory of Collective Behavior, (New York, 1963CrossRefGoogle Scholar). They have been dealt with in the case of Social Credit in the Rise of a Third Party, (Englewood Cliffs, 1971), chaps. 6ff.

⁵ The same is true of urban areas, except that the boundaries of perception are more distant and less well demarcated.

⁶ Rise of a Third Party, 64, and fn. 6.

⁷ Ibid., 65.

⁸ It should be obvious that I am here only concerned with political movements acting through political parties.

⁹ On the notions of real and derived phases in collective behaviour, see Smelser, Theory of Collective Behavior, passim, esp. 356–8. More will be said about this below.

¹⁰ We do not necessarily mean protest orientations due to one-party dominance, but also protest votes which would normally have been channelled into a traditional opposition, but which, given the dynamism of a new radical movement, now take this route. Of course, if one-party dominance also prevails then the derived phase could be of much greater magnitude.

¹¹ See chap. 3, Rise of a Third Party.

¹² See ibid., for details.

¹³ Ibid., 65.

¹⁴ Based on Peter Regenstreif's poll, reproduced in Lemieux, Vincent, Gilbert, Marcel, and Blais, André, Une élection de réalignement: l'election générate du 29 avril 1970 au Québec (Montréal, 1970), 88.Google Scholar The other data presented there are also relevant, 88ff. See also Blais, André, Gilbert, Marcel, and Lemieux, Vincent, “The Emergence of New Forces in Quebec Electoral Politics,” in Canada: A Sociological Profile, ed. Mann, W.E. (2nd ed.Toronto, 1971), 537–4.Google Scholar

¹⁵ The new factor may have been more important however for the moderate success of the Bloc Populaire in Quebec in 1944 (14 per cent of the votes).

¹⁶ I should add that the comments made previously regarding the similarities and differences between the “special model” which has been more fully developed here and Macpherson's model still hold; see Rise of a Third Party, 69–70, in particular fn. 28.

¹⁷ Ibid., 64–5.

¹⁸ Ibid., 39.

¹⁹ Above, p. 420.

²⁰ Rise of a Third Party, chap. 5. Notice how this hypothesis appears contradictory to the one-party dominance hypothesis (ibid., 73). Indeed the “discovery” of the one-party-dominance model was a serendipitous finding emerging from our work on the former hypothesis.

²¹ Similar remarks could probably be made about the Ralliement National (also a Social Credit group) in the 1966 Quebec provincial election (ibid., 53).

²² There are indications that the federal Social Credit party currently keeps pulling up the provincial wing of the party. Moreover, the 1970 provincial election probably marked, with the decline of the Union Nationale, the opening of an era of one-party dominance, which is likely to favour the Ralliement Créditiste; see Pinard, Maurice, “The Ongoing Political Realignments in Quebec” in Quebec Society and Politics, ed. Thomson, Dale C. (Toronto, 1973), 119–38.Google Scholar

²³ On the other hand, what prevented a serious decline of the Quebec provincial Liberals after the rise of the Union Nationale was the existence of a strong federal Liberal wing.

²⁴ Above, p. 432.

²⁵ Rise of a Third Party, 26 and chap. 3, passim. Blais’ criticisms can be found in his paper above, p. 430. (Blais grossly misrepresents my position, as any reader can verify, when he writes that I have recourse to this model “only when a third party arises without there being one-party dominance.”)

²⁶ The other factor would of course have been a previously weak opposition party. The reader should note the similarities between some of the cases classified under the so-called two-step model: a long tenure of office by a strong traditional party followed by one term of office by the weak one, this term of office terminating during a period of strain, and finally the emergence of a third party. See in particular Rise of a Third Party, 26, 41, and 48.

²⁷ On this last paragraph, see Lipset, S.M., The First New Nation (New York, 1963), 294, 300ff.Google Scholar The plurality-electoral system, on the other hand, is more likely to affect the size of the representation obtained by, and the long-term viability of, a third party than to affect its emergence in terms of popular support.

²⁸ Above, pp. 428–9, 432, 436–7. The hypothesis that limited partisan attachments are related to the rise of “flash” parties has also been suggested in a comparison of France and the United States; see Converse, Philip E. and Dupeux, Georges, “Politicization of the Electorate in France and the United States,” Public Opinion Quarterly, 26 (Spring 1962), 1–23.CrossRefGoogle Scholar We also found of course that nonpartisans were more likely to have turned to Social Credit in 1962; see Rise of a Third Party, 167–9.

²⁹ This is not contradictory to the statement that one-party dominance is likely to lead to long tenures of office. When a weak opposition party has about 33 per cent of the votes, it is relatively easy to gain (or lose) more than 5 per cent of the votes, but this has little chance of producing a change of government.

³⁰ I found that stable voters were in general less likely to vote Social Credit than were party changers; see Rise of a Third Party, 75–8.

³¹ This point is supported by data of the Social Credit study: there were more party changers between the two old provincial parties in weak federal districts than in strong ones; see the marginal N's in Table 5.3, Rise of a Third Party, 77.

³² I grant however that partisan attachments could also be eroded by severe strains, as for instance greater variations in income and more rapid urbanization (Blais, above pp. 436–7). But these strains cannot be the major source of variations in partisanship. Moreover they should manifest themselves in greater political change, which would then have its own effect (Rise of a Third Party, 109–11). Note also that Converse and Dupeux saw the limited party identifications of France as resulting from the lack of family socialization and thus as an intervening factor between the latter and the rise of “flash” parties. In my view, they rejected too easily the idea that it could be linked to a fluid and complex multiparty system, on the basis of a simple comparison with Norway. I would suggest that the lack of political socialization simply hides the transmission from generations to generations of limited identifications, which are themselves related to the party system. At least in Canada, as we have argued, a multiparty system is likely to be a one-party-dominance system producing weak party identifications.

³³ Instead, with a surprising lack of rigour for a strict methodologist, the author simply asserted that while the first difference between West and East “would give support to Pinard's theory,” the second difference was a “more striking and basic” one (how was this established?), and he went on to conclude a few paragraphs later that the East-West difference in third-party fortunes could “thus hardly be accounted for by one-party dominance”; Blais, above, p. 433.

³⁴ They were classified as cases of “suddenly weak” opposition parties and were among those which then led me to suggest the special model I have now more fully developed.

³⁵ The basis on which Blais can therefore refer to White's evidence to reject the one-party-dominance hypothesis is still a mystery to me.

³⁶ Unless of course for some reason the emergence of class or communal sentiments were to be related to the former strength of the old parties.

³⁷ Blais considers not only the PQ vote but the PQ and RC vote together. But the conducive factors for the latter do not lead one to expect any relationship there either. See Blais, above, pp. 434–5.

³⁸ Above, pp. 403–6, especially p. 406.

³⁹ The strong correlation-coefficient nevertheless obtained in this case is subject to suspicion given the non-linear distribution to be observed in Table IV. Moreover, one could expect some diffusion impact of the partial abdication of the Liberals even in districts in which they did not abdicate. Finally, I wonder if urban-rural and ethnic controls would not also be appropriate here.

⁴⁰ On this point, Blais cites Blalock, but he misses Blalock's main point which is that in linking a theory to specific indicators one must develop an auxiliary theory which has its own assumptions, See Blalock, Hubert M. jr, Theory Construction (Englewood Cliffs, 1969), 151–4.Google Scholar What I will be arguing below is that while my assumptions appear valid, those of Blais are often very questionable.

⁴¹ Rise of a Third Party, 25.

⁴² Blais contends I had used seats as a test of the model. This is not correct. I simply started collecting the data on seats as I went along collecting data on votes and kept doing it even after that data soon appeared to have no face validity as a measure of the strength of a party. Again compare what Blais says, above, p. 426, with what I actually wrote where he refers to Rise of a Third Party, 63, fn. 2.

⁴³ Note incidentally how this last redefinition of one-party dominance drastically alters the distribution of cases as weak or strong opposition parties (in Table II there are 9 weak and 37 strong cases; in Table III, these figures become 21 and 23 respectively). See pp. 425, 427, above.

⁴⁴ See above, p. 427, fn. 14.

⁴⁵ I have on the other hand reported that my two indicators in the tests on Social Credit in Quebec were strongly related to one another; see “La faiblesse des Conservateurs et la montée du Crédit Social en 1962,” Recherches sociographiques, 7 (Sept.-Déc. 1966), 361.

⁴⁶ It should be stressed that in the name of rigour Blais, contrary to White, has systematically disregarded any qualitative evidence of the weakness of opposition parties, such as failures to contest seats, factionalism and splits in a party, severe leadership crises, moribund organizations, etc. I agree with him when he finally writes that his procedure involves “the danger [of losing] touch with reality” (above, p. 431).

⁴⁷ Above, p. 426, fn. 10.

⁴⁸ Ibid. Blais writes here (fn. 11) that I acknowledged this possibility. This is wrong: in the reference he cites, I was discussing the independent variable, not the dependent one; besides even for that variable I mentioned the serious problems involved.

⁴⁹ Ibid.

⁵⁰ We might add that the strong opposition party column includes the following “strong” parties: Newfoundland, 1959, where the Conservatives were at an average of 33.6 per cent of the votes; Saskatchewan 1956, where the Liberal opposition was at 33.8 per cent; Quebec, 1935 and Quebec, 1936, where the strength of the Conservatives was around 33 per cent (in this case, see Rise of a Third Party, 52). This is therefore a conservative test.

⁵¹ I also considered a few elections at the federal level, but they were not chosen randomly. Blais is therefore right in excluding them from a more rigorous test.

⁵² I may not have been very rigorous in the latter case, but on the basis of that statement I fail to see how according to Blais I did “not always stick to [my] own criterion.” The examples he gives to make his point all witnessed the appearance or improvement of a third party; above, p. 425.

⁵³ Ibid., p. 433.

⁵⁴ Blais seems to finally recognize that logic in the following paragraphs (p. 434ff.).

⁵⁵ The results of that test are not presented in his paper, but he has kindly given me the authorization to refer to them.

⁵⁶ His final comments concern his negative evaluation of the tests at the constituency level. I have already argued that White's evidence is in line with what I had previously observed and is now better accounted for by the revised model. As for my own evidence, I do not think that Blais can cite Lemieux's comments on it as rejecting the theory; quite the contrary, as noted above. Finally Blais's evidence about the 1970 Quebec election concerns an exception now accounted for in the revisions presented above.