In 1992 the Swedish scholar Axel Hadenius published a book on democracy in Third World countries entitled Democracy and Development (Hadenius 1992). In this book Hadenius accomplished two tasks. On the one hand, for a sample of 132 countries, including almost all sovereign countries in Latin America, the Caribbean, Africa, Asia and Oceania, he established a scale to measure the level of democracy in each and every country. On the other hand, he explained the variations in the level of democracy he was able to detect by a variety of explanatory factors, including socioeconomic demographic and cultural factors, and institutional arrangements. His main finding was that "no single explanatory factor strikes like an iron fist through the material"; instead, "several attributes of different kinds stand out as important" (Hadenius 1992, 146).

It is not the aim of this article to review or criticize the book by Hadenius at large. The volume is rich in findings as well as documentation, and we certainly agree with one reviewer, who states that Hadenius's book "is a worthy successor to Robert Dahl's Polyarchy ... in its efforts to quantitativelyevaluate extent and the determinants of democracy across

Resumé

In a recent book on "Democracy and Development" (1992) Axel Hadenius stumbles over the observation that island states are of special interest: on a scale running from 0 to 10, the average level of democracy is 7.1 for island states while for others it is 3.6. This, according to Hadenius, is due to the fact that island states are far more Protestant dominated than others. Contesting this view, three points are argued in this article. First, the attempt to explain away the relationship between insularity and democracy by introducing Protestantism into the explanatory design is questioned on grounds of methodology, the leading argument being that comparative analyses should not treat differences in factors like Protestantism as matters of degree. A second argument is about concept-stretching: the overall travel capacity of concepts like Protestantism is questioned for the universe of small island states, where custom and the accommodation of traditional authority must be credited with nitrating functions. Secondly, theoretical reasons and empirical illustrations are introduced to suggest that one is well advised to consider small island states as a separate category, imbued with democracy and democratic procedures. Thirdly, due notice is taken of the fact that several small island states, such as Cape Verde, the Comoros, Sao Tome and Principe, the Seychelles and Western Samoa, have undergone transitions to democracy during the time span after Hadenius's research (1988), the level of democracy for small island states therefore being now even higher.

nations" (Weaver 1994, 778). Only one specific finding, which comes about almost incidentally, as a siding, is considered here. When discussing the relation between size and mode of government, Hadenius stumbles over the observation that island states are of special interest. For island states, on a scale running from 0 to 10, the average level of democracy is 7.1, while for others it is 3.6 (Hadenius 1992, 126). This, Hadenius argues, may be a function of size. Islands are as a rule small: findings suggest that the smaller a territory, the more likely it is to be an island (Sutton & Payne 1993, 581). Recognizing, therefore, that the association of islands with democracy may be a consequence of smallness rather than insularity, Hadenius controls for size in terms of area as well as population. However, this makes no difference, and the fact remains that island states are special (Hadenius 1992, 126-127).

Faced with this finding, Hadenius comes up with the idea that the connection between insularity and democracy may be spurious, reflecting effects of other attributes than size. Referring to the fact that his earlier analyses have demonstrated a link between Protestantism and democracy, a link that persists when Protestantism is tested separately together with socio-economic attributes (1992, 120-121), Hadenius now introduces the observation that island states are far more Protestant dominated than others. "This", he believes, "may be the explanation" (1992, 126). And indeed, as Hadenius includes Protestantism in a multiple regression which tests the relevant variable together with a number of socio-economic attributes, the relationship between island states and level of democracy becomes much weaker and is no longer significant. "It may be concluded", Hadenius writes, "that the often very small island states are not, as such, as special as they appear to be on simple inspection of the political geography" (1992, 127). The favourable effect of insularity on the level of democracy is, he maintains, "primarily a consequence of the impact of other variables" (1992, 147).

Readers of Hadenius's book seem to agree. Anders Sannerstedt has suggested that the fact that there are several democracies among the Caribbean island states but only a few in Africa may explain the connections Hadenius has found between democracy on the one hand and insularity, trade with USA and lack of ethnic fragmentation on the other (Sannerstedt 1994, 74). Howard Handelman notes in a book review that Democracy and Development is useful by virtue of its large database and the intelligent manner in which Hadenius tests important hypotheses in the literature (Handelman 1994, 316). However, Handelman also notes that "the reader is occasionally diverted by discussions of obviously spurious (and trivial) relationships such as the positive correlation between democracy and island states" (1994, 316-317). This article, however, does not concur in this view. On the contrary, we argue that the relationship is far from trivial. We argue

that Hadenius makes an important observation which is, in fact, even more important than he seems to realize himself: we argue that size and insularity indeed make a difference for democracy. Specifically, we argue three points.

We argue, first, that Hadenius's attempt to explain away the relationship between insularity and democracy is less than convincing. The introduction of Protestantism into the explanatory design does not really alter the picture, and the fact remains that island units are, by and large, more democratic than other units. We argue, secondly, that there are good theoretical reasons why island units are special in this respect and that one is well advised to consider small island states as a separate category, imbued with democracy and democratic procedures. Thirdly, and outside the context Hadenius's research, we make a point of the fact that in several small island units developments ensuing almost immediately after the time span of Hadenius's research have altered drastically the descriptions given by Hadenius and have increased still more the level of democracy in this state category. In short, we argue that the role of Protestantism may be seriously questioned, that there are good inherent other reasons why small island states are democratic units, and that recent events suggest that Hadenius' classifications of small island states are now in several cases obsolete.

"Either-Or" Rather than "More or Less"

Hadenius operationalizes democracy by constructing an index which focuses on the electoral process and the degree of political freedom. Electoral measures include: the level of suffrage, the periodicity, effectiveness honesty of elections; and the openness of elections to candidates parties regardless of their political positions. Political freedom indices include: organizational freedom to form parties and interest groups, media freedom, and the absence of political oppression and violence. The point of departure for Hadenius's discussion of insularity is formed by simple regression analyses, where the size of the population, the area and whether or not the state is an island, are matched against these dimensions of democracy.

The results indicate, as we have noted, that size connects with democracy and that insularity connects even more strongly with democracy. Furthermore,an control reveals that the relationship for island states persists, whereas this is not the case for the size variables. Hadenius is aware of the possibility that this may ensue from different classifications: he takes due note of the fact that in the one case, for island states, he has a dichotomy, while in the others, for population and area, he is dealing

with continuous variables (1992, 126). In order to test the validity of his finding, he therefore reclassifies the population and area variables and distinguishes dichotomously between small and large states. However, this makes no difference. The effect of the size variables is again drastically curtailed when the island variable is included in the regression. Introducing at this point Protestantism as an explanatory factor, Hadenius, however, does not go to the trouble of dichotomizing this factor, which is treated as a continuous variable. In the following we use this assertion as an entrance to our criticism of Hadenius's approach. The criticism is set forth in a number of points:

(a) In some important methodological papers Giovanni Sartori has urged comparative political analysts to avoid the tendency of assuming all differences be a matter of degree (1970, 1044; 1991, 248-249). The maxim that differences in kind are best conceived as differences of degree, and that dichotomous treatments are invariably best replaced by continuous ones, is, according to Sartori, a producer of conceptual messiness (1991, 248). Therefore, before comparisons for "more or less" can be performed it is necessary to define the thing or property to be compared in terms of "either-or". While concurring in this beware of degreeism, Rosemary O'Kane has argued that "without the ability to be certain about what something is not, potentially anything can be pulled into the comparison and the danger grows of making unlike look alike" (1993, 170). It is our impression that Hadenius has not avoided this danger.

(b) Let us proceed by asking why one should assume that there is an association between Protestantism and democracy. On this point, Hadenius is somewhat taciturn. Making reference to Samuel Huntington and other authors, he argues that Protestantism fosters individual responsibility and has since the Reformation promoted a tradition of rebellion against established In short, Protestantism is more tolerant and less dogmatic than other religions (Hadenius 1992, 119). It is obvious, Hadenius states, that religious and other fundamentalism is difficult to reconcile with democracy. which claim to present the absolute right and truth have difficulty in submitting to the democratic mode of decision-making (1992, 119).

(c) Now, if Protestantism paves the way for democracy by promoting a certain mood of thinking about political matters, are we then entitled to believe that any increase in the amount of Protestantism makes a difference? do not think so. For instance, if two countries A and B both have Protestant minorities, the percentages of Protestants being 11 in A and 15 in B and all other relevant and operative factors being equal, does this difference really count? Should we really expect B to be a more democratic nation than A? Resorting to degreeism and correlation analyses

implies an affirmative answer, to which we, however, take exception. Surely some sort of threshold must be operating here, meaning that the proportion of the population that professes to be adherents of Protestantism must be large enough really to leave a mark on the political culture and the political architecture of the nation and to influence the mode of decision-making. To state this differently: for the impact of Protestantism (or any other religion) to count, this religion must have a dominating position of some magnitude. A consequence of this thinking is that countries are to be regarded as either Protestant or not, not as more or less Protestant. Establishing a cutting point is, of course, a matter of convenience and plausibility. We would, for our part, argue that a threshold of 50 percent is a minimum condition, and that a threshold of, say, 70 percent would be a trustworthy device.

(d) Table 1 departs from the dichotomized view that states are either Protestant or not, either small or not, and either islands or not. If and when these dichotomized dimensions are crossed, an eight-fold table emerges, frames eight categories of states. In principle, for each and every category an average democratic score may be calculated on the basis of the empirical results that Hadenius reports, and the various averages thereafter be compared, the comparisons revealing patterns of impact and explanation. instance, if Protestantism is assumed to remain a decisive factor when other factors are controlled for, then the values for cells 1, 3, 5 and 7 must be assumed to exceed the corresponding scores for cells 2,4,6 and 8 respectively. Again, if insularity is assumed to be decisive, the values for cells 1, 2, 3 and 4 must be assumed to exceed the corresponding scores for cells 5, 6, 7 and 8 respectively. Finally, if smallness is assumed to be decisive, then the values for cells 3,4,7 and 8 must exceed the corresponding scores for cells 1, 2, 5 and 6 respectively.

(e) However, it is not possible to carry out in practice an exercise like the one suggested in Table 1. If a threshold value of 70 percent for Protestantism is applied to the 132 states that are investigated by Hadenius, the reason for his negligence to reclassify the Protestantism variable becomes clear.

It is namely the case that all Protestant states are island states!¹ This, then, means that a distinction of vital importance for testing purposes cannot be observed: to determine whether Protestantism rather than insularity holds the key to democracy, Protestant islands should be compared to Protestant mainlands. Furthermore, an exploration of the impact of Protestantism would require that Protestant mainlands are compared to non-Protestant mainlands. Both comparisons, however, remain out of reach. The situation does not improve much if the threshold is lowered to 50 percent, meaning that all countries which reach this lower level are regarded as Protestant countries. Only one mainland country, namely Swaziland, is a relevant test case, and the outcome of this test is not encouraging, as the level of democracy in Swaziland is assigned only 2.6 points by Hadenius (1992, 62) as compared to, for instance, 8.7 points for Muslim Gambia.

(f) Since almost all Protestant states in Hadenius's sample are islands, it follows that Protestantism cannot explain why islands are democratic. The two competing independent variables coincide, and the one can therefore not take precedence over the other. One may in fact suggest that a synchronous appearance of Protestantism and democracy is an outcome of a third background factor. Hadenius makes reference (1992, 130, 196) to a study by Myron Weiner, who suggests that the high level of democracy in island states derives from the fact that the majority of these were under British rule. Hadenius does not concur, maintaining that the duration of the colonial period is of greater interest whereas it makes little difference who was in control (1992, 128-133). Be this as it may, we wish here to point at a chain of reasoning which is hinted at by Hadenius, as he notes that Great Britain was a sea power, as he also notes that the Union Flag therefore flew in many islands, as he remembers that island states are in large measure Protestant, and as he concludes that one can undoubtedly assume a historical connection here (1992, 130-131). Rather than islands being democratic because they are Protestant, from this would follow that islands are Protestant as well as democratic because they have been colonies. question, then, becomes one that concerns the specific impressionability island units.

In Conclusion

The fact that a unit is an island does not in itself make this unit democratic. To explain why islands are democratic, one needs (1) to identify factors or qualities that are generally linked to democracy, and (2) to show that these factors or qualities are present in an island context to an unusually high extent. Hadenius has done precisely that, pointing at Protestantism as a key factor. Whereas we do not in the least question the general assumption

that there is a link between Protestantism and democracy, we argue, however, (1) that islands are not Protestant to the extent Hadenius makes us believe; and (2) that other factors and features that are connected with insularity go a long way towards explaining why insular units are receptive to Protestantism as well as democracy. Both arguments are explored further in the next chapter, which focuses on a subset of island states, namely small island states.

To these rather specific remarks concerning method and design, we
would like to add a final and more general methodological interjection:

About religious affiliations in Vanuatu, Handbook of the Nations (1989, 312) gives the following characteristic: 'most at least nominally Christian'. We believe that the reservation that is contained in the expression "at least nominally" is well founded and relevant in the present context as well. We namely believe that Sartori's notion of concept-stretching is applicable here: the very concept of Protestantism may be troublesome in the context of islands. Discussing the question "What is comparable", Sartori makes a point of stressing that the question always is "comparable with respect to which properties or characteristics" (1991,245-246); in the context at hand, the question is whether or not Protestantism always entails the qualities of tolerance and lack of dogmatism which are of importance for the promotion of a democratic disposition. We believe for our part that the concept of Protestantism does not travel well in all cases.

The issue cannot be discussed at length here, and a few comments must suffice. We may refer to the case of Tonga as an illustration. This country is predominantly Protestant (Derbyshire & Derbyshire 1989, 749). However, the social order of Tonga has not been imbued with sentiments open to the essence of the Protestant message. We quote a brief passage from a chapter on social life in the country: "For hundreds and hundreds of years the people of Tonga have lived under a system of social taboos. Every man must obey his eldest brother, who in turn must obey his father, who in turn must obey his eldest brother, who in turn must obey his father or whoever is the eldest among his father's brothers still living . . . Total and unquestioning obedience or fakaapaapa is the essence of this social order" (Niu 1988, 307-308). In this case, then, custom and the accommodation traditional authority obviously must be credited with a kind of filtration function vis-å-vis incentives and initiatives.

This function is probably to be found in other similar cases as well. It has been said that the incorporation of customary values and practices was one of the most difficult and complex intellectual and technical problems in the process of constitution-making in the Pacific island states, custom achieving prominence because of a general wish that the constitution should be related to the social and economic conditions of the country and based on its values (Ghai 1988, 39-42). It is also a well-known fact that in many

island cases the missionaries based the local church organization upon village and elan structures, which proved an effective tool for church purposes (e.g. Samana 1988, 318-319). An obvious consequence of this would be that the church purposes, those relating to organization as well as those relating to the messages and the commandments of the church, were influenced and coloured by such strains of custom and tradition that are in less than perfect agreement with orthodox interpretations and do not accept any rebellion against established authorities. We do not state this as a faet, but certainly as a highly plausible hypothesis.

In a brilliant concluding chapter, Hadenius discusses the merits as well as the pitfalls of his statistical research design (1992, 143-157). He readily admits that reports that build on mass data from many countries inevitably become sweeping, presenting "a bird's-eye view of the empirical landscape" (1992, 155). A qualitative analysis of a few cases, he states, provides a far more penetrating and variegated picture of the situation under study (1992, 155). This is of course true, as is Hadenius's remark that the two research trends can at best complement each other. Our hesitation about the travel capacity of the concept of Protestantism does therefore not in the least constitute a critique of Hadenius's way of conducting his study. If one undertakes a study of almost all countries of the world, assumptions about travel capacities are certainly justified. We do feel, however, that the conclusions that follow in this case from the assumptions are not well founded and do not provide an increased insight.

The Category of Small Island States

As pointed out by Mattei Dogan and Dominique Pelassy, all comparativists conceptually homogenize their field (1984, 133). In this chapter we introduce argue in favour of a slightly different choice of series of nations than the ones preferred by Hadenius. Whereas Hadenius in his analyses keeps small and large states apart, and distinguishes between island states and other states, he does not explicitly deal with small island states as a separate category. We argue, for our part, that a combination of smallness and remoteness holds quite specific democratic promises. In defending this view in empirical terms, a first task, then, is to identify a population of small island states.

When distinguishing dichotomously between small and large states, Hadeniusmakes of the limits one million for population and 1000 square kilometres for area (1992, 126, 195). This is probably as good a solution as any. Here, however, another yardstick will be used which is described at length in a separate research report (Anckar 1991), the details of which are not to be repeated here. Suffice it to say that size is determined in terms

of average figures for a total population of island states, those island states being considered small, which are below the averages for area and population and which satisfy some further quantitative as well as qualitative criteria. When this device is applied to the sample of states investigated by Hadenius, 29 out of a total of 132 states emerge as small island states.²

The average level of democracy in these 29 small island states, as measured by Hadenius's index, is high indeed: 7.3 points as against 5.9 for other island states. The relevance of the remoteness dimension is supported by the fact that the average democracy level is 8.0 for remote small islands as against 6.3 for other small islands.³ As is evident from Table 2, the small island states do stand out as a distinct category. On the one hand, whereas other states position themselves only to a small extent within the highest levels of democracy, the small island states occupy in almost full numbers these same levels. A total of seven countries score a full 10 points. This democratic nobility consists of small island states only, and this, in turn, means that 24 percent of the small island states and no other states are ranked in this highest category. The same pattern repeats itself when moving down one step on the ladder of democracy. A total of 28 nations score 9 points or more, and 17 of these nations are small island states. The implication of this is that 59 percent of the small island states are in this category, whereas the corresponding figure for other states is 11 percent. Finally, a total of 39 nations score 8 points or more; 18 of these nations are small island states. This means that 62 percent of the small island states belong to this category, whereas the corresponding figure for other states is 21 percent. Whereas about two-thirds of the small island states are in the three upper categories, only one-fifth of the other states place themselves in their company. It follows, then, that the distribution is the reverse when the lowest levels of democracy are in focus. Only one small island state scores less than 1 point as against 19 other states, the percentages being 3 as against 19. Four island states score less than 2 points as against 41 other states, the corresponding percentages being 12 as against 38.

Before turning to a discussion of the factors promoting democracy in small island states, we return briefly to the issue of Protestantism. Given that islands are different in a number of ways (e.g. Anckar 1995b), the differences decreasing the extent to which individual factors can be observed in isolation, one is well advised to search for regions that are homogeneous enough in terms of geography, culture and history to allow the assumption that at least some vital components are controlled for. We have therefore compiled average levels of democracy for the 9 small island states in the Caribbean and the 10 small island states in the Pacific, discriminating further between strong Protestant states (percentage of Protestants being 70 or more), Protestant states (percentage of Protestants being 50 or more but less than 70), and other states (all other religious configurations). The results are presented in Table 3. As the number of units per category is quite small, definite conclusions are out of order, and the figures have a heuristic value only. It is, however, worth noting that the results do not indicate any impact whatsoever of the level of Protestantism the level of democracy. Within both regions, non-Protestant states are about equally as democratic as strong Protestant states. It is also worth noting in this context that Protestantism is not a dominating feature in the population of small island states. Out of 29 units, only 8 are strong Protestant units, whereas 4 are Protestant and no less than 17 represent other religious configurations.

Let us, then, turn to the question of conceivable links between smallness, insularity and democracy. We may recall that Hadenius conceptualizes democracy in terms of honest and open elections and unrestricted political freedoms. Such devices are promoted by the existence of a tolerant, considerate, understanding and moderate political climate, which has a capacity for reciprocal anticipation and is foreign to despotism, dictates and zero-sum solutions. In small island states, we argue, such a climate is likely to emerge and prevail. Again, our argument is stated in a chain of points:

(a) More than other units, remote units are likely to promote feelings of
fellowship and a sense of community. When people live at distance from

the outside world, they are likely to develop a spirit of community and solidarity: they share a feeling that they are, so to speak, alone in the world and thrown upon their own resources. Remote units are therefore often cohesive units, marked, perhaps, by a certain reservedness. Discussing the English-speaking islands of the southeastern Caribbean, Tony Thorndike describes the political culture of the islands: "It is a political culture that is essentially inward-looking and insular, where the surrounding sea is a barrier rather than a highway" (1991, 110).

One mechanism at work here is about problems and problem-solving. It
is, for instance, a very reasonable assumption that remote units have a
special concern for the maintaining of networks for transport and communication
the outside world (e.g. Hamilton-Jones 1992). Generally
speaking, remote units must come to terms with problems that are outcomes
of remoteness and therefore are, to some extent at least, special in nature.
Special problems call for special solutions; these problems and solutions,
then, become a part of the daily life of the inhabitants of remote units.
They are, as members of communities with special characteristics, subjected
to the consequences of these characteristics as well as accustomed to dealing
with the consequences. Remoteness thus becomes an unifying factor, a
frame of reference that is shared by every member of the community,
(b) A spirit of fellowship and community is also promoted by the citizens
being able to orient themselves towards political life and the political
apparatus. Such an ability is no matter of course. The costs in time,
inconvenience, energy, money, embarrassment, anxiety, fear, and so on,
of gaining information and knowledge about politics and social affairs may
prove too high for many members of a society (Dahl & Tufte 1973, 41).
However, members of small units possess rather unique possibilities to
form structured sets of preferences: the units are, because they are small,
less complex and more lucid than bigger units. One manifestation of this
is that big units are likely to develop a greater number of organizations and
organizational subunits, which maintain networks of interactions, power,
influence and specialization which are difficult to detect, identify and
understand (Dahl & Tufte 1973, 36-39). In this respect and in others, small
units appear more simple, elementary and easy of access. The implication
of this is not only that members of small units are well informed about their
society. As Giovanni Sartori has pointed out, information is not knowledge:
at a minimum, knowledge implies a grasp of, and a mental control over,
the information that is not contained in the information itself (1987, 117).
The implication is that members of small units may develop such a mental
control. They are more likely than members of big units to understand the
way means relate to ends and they are more likely to understand the
consequences that different lines of action may carry.

(c) The emergence of feelings of tolerance and understanding is facilitated

if and when open channels of communication exist between those who govern and those who are governed. In this respect, it is important to note that the lucidity of small units not only promotes the insight of unit members. It also promotes the ability of unit leaders to survey what is going on in their societies. The leaders of small units are thus more likely than leaders of big units to acquire a knowledge of general dispositions; factors that are related to smallness are contributive in this respect. For instance, the relative simplicity of articulation structures and the limited extent of specialization make the incoming information about preferences more genuine and comprehensible, as it is filtrated through a relatively small amount of intervening structures and agents. The remoteness of island units and the unifying sense of community that comes out of remotenessmerits in this respect also: the distance between those who govern and those who are governed is lessened as the two segments share a frame of reference which has emerged from shared problems and problem conceptions. Also: in small-scale societies ordinary citizens can deal more directly with top leaders (Dahl & Tufte 1973, 87), and this offers preconditionsfor effective citizen control.

(d) Small units may of course be heterogeneous in terms of, for instance, population subsets, cultural diversity or regional diversities.⁴ However, as a rule, small units are more likely to be homogeneous than big units (Dahl & Tufte 1973, 91-109). One implication of this is that small units, more than big units, are likely to develop open and flexible political processes which offer the participants opportunities to know and understand each other. The homogeneity of small units namely promotes knowledge, consideration anticipation. In a by now classical paper Willmoore Kendall and George Carey explain why this is the case (1968, 17). First, in homogeneous which share interests and values, an individual may acquire a knowledge or an estimate of the reactions of others simply by observing his or her own reactions. Secondly, if a homogeneous society has in its bosom groupings that are differently affected by policy decisions, society members, because of the overriding homogeneity of the society, still have greater opportunities to comprehend the composition and the structure of the society and thereby to foresee the potentially conflicting interests. Finally, in homogeneous societies one may expect from the society members a high degree of sympathetic identification with each other, implying a willingness to understand beforehand the probable effects of action and a greater effort to feel others out.

Recent Cases of Democratization

Hadenius investigates the varying levels of democracy at a certain point in

time, which is the year 1988 (1992, 2). Since, however, changes in terms of the level of democracy occurred in some countries during that year, he is forced to choose a still more precise point in time, namely the end of 1988 (1992, 38). From a methodological point of view, this is of course a sound and defendable solution. Whereas there are ample means for dealing in sophisticated empirical manners with variances along the temporal dimension in the variables which are considered (e.g. Bartolini 1993), there are no corresponding methods for dealing empirically with the future. The best one can do is to acquire data which are as recent as possible. However, the best is not enough. The world that the social scientist tries to observe and understand is in a state of constant flux, which entails challenges to the scientist and often undermines the validity of his or her results. This is true also in Hadenius's case. His praiseworthy ambition to bring his investigation as close as possible to the present day has in fact by now lost some of its relevance and vitality.

True, very many years have not passed since 1988. These years have, however, witnessed in quite a few cases a transformation of less than democratic small island states to democratic entities. These transformations, convey a still brighter picture of the democratic capacities small island states than the one presented in Hadenius's research. Of the 29 small island states, 12 score less than 9 points in Hadenius's classifications, and quite a few of these 12 cases receive gradings that fall short even of the threshold of 5 points. This is true of the Maldives (4.0) and Fiji (3.1) and six more islands, which score 2 points or less. Whereas some of the 12 cases, e.g. Bahrain and Brunei (Mani 1993), still today appear more or less immune to ideas and processes of democratization, some other cases have experienced quite remarkable democratic progress. In the following, in no specific order, these cases are briefly discussed:

The Case of Western Samoa

In Hadenius's classification Western Samoa is given 6.0 points (1992, 62). A closer examination of Hadenius's argumentation reveals that this somewhat low score is based solely on the fact that about 80 percent of the population of the country was still in 1988 excluded from the franchise, as the right to vote was reserved for matais or titled family heads, who elected 45 out of the 47 Members of the Fono or National Parliament, the remaining two Members being elected by universal suffrage of some 2000 voters on the so-called individual voters' roll (Derbyshire & Derbyshire 1989, 760; de Backer 1991, 31; Hills 1993). Attempts to change this system were introduced on several occasions since the independence of the country in 1962, but they remained unsuccessful for a variety of reasons, not least a fear in the minds of many that universal suffrage would challenge and erode

the faa-Samoa, the Samoan way of life. In 1990, however, as a change in the electoral system became the issue of a plebiscite, the Electoral Law was finally and drastically changed. All Western Samoas of twenty-one years or more would now be able to cast their vote, every three years, for one of the members of the Fono. The first elections based on this new system of universal suffrage were effected in April of 1991 (de Backer 1991). Obviously, in classifications employing more recent data, Western Samoa would join the group of small island states which score a full 10 points, or would, at least, come very close to this group.

The Case of Sao Tome and Principe

Sao Tome and Principe is a former Portuguese colony off the Gulf of Guinea. The country is given a harsh treatment by Hadenius, indeed: it is one of eleven countries which score 0.0 points (1992, 62). According to Hadenius, the country did not have open elections, organizational freedoms and the freedom of opinion. This is certainly a true description of the situation at the end of 1988. Sao Tome and Principe was a one-party state, with small opposition groups operating from outside the country (Hodges & Newitt 1988, 102-104). In short, the country was devoid of almost all democratic qualities: to the democratic image of the small island states of the world, Sao Tome and Principe was a disgrace.

Later, however, developments have occurred that alter the classifications.When country became independent in 1975, it adopted a Marxist-Leninist regime, which, however, due to an economic liberalizationin wake of an economic crisis, had to give way to a more open political climate. In 1987, plans were approved for a constitutional reform, which, however, stopped short of introducing a multiparty system (Hodges & Newitt 1988, 112). It soon became evident that the plans had to be altered in this respect. In 1989 a multiparty system was adopted, and this change was ratified by a referendum in August 1990 (Cahen 1991; Traore 1992, 15). Some months later, in January 1991, the former single party was defeated at the legislative elections by the Party of Democratic Convergence which gained 54 percent of the votes (Traore 1992, 15; Blaustein & Flanz 1993, vi), and in December 1992 the first multiparty municipal elections of the country were held (Blaustein & Flanz 1993, vii). Sao Tome and Principe is heavily dependent on external aid and is still a democracy in its infancy with fragile democratic institutions. However, the present political situation is by now regarded as pretty stable, and any danger of a coup is said to be lessened by a programme of reforms in the army (NTW,⁵ 166). Sao Tome and Principe probably does not yet deserve to be ranked among the top countries in Hadenius's scheme, but it is, on

the other hand, quite obvious that the country has by now moved quite far
from a democratic zero point.

The Case of Cape Verde

Like Sao Tome and Principe, Cape Verde is a former Portuguese colony, which secured its independence in 1975. Hadenius does not think much of democratic life in Cape Verde, and rightly so. The score assigned to Cape Verde is very low, only 1.7 points. A variety of reasons are operative here. In 1988 elections were not open, organizational freedoms were nonexistent, of violence and oppression were reported, the freedom of opinion was circumscribed. As Hadenius points out, although a limited criticism could exist in the country in certain areas against those in power at a lower level, "there can be no talk of calling into question the government the policy it pursues in the relevant spheres" (1992, 57).

Today, however, Cape Verde is very different. Without much pressure from the street, the African Independence Party decided in 1990 to abandon its constitutional single party status, and to open the elections to the National Assembly to members of all parties (Blaustein & Flanz 1994, vi). Cape Verde thus emerges as an example of a smooth transfer to democracy, where a ruling party in a one-party state freely cooperates in holding elections risking its own defeat. The defeat materialized as early as in January 1991, when the Movement for Democracy Party achieved Africa's first political switch in an election, taking twice as many seats as the single party in power (Cahen 1991). One month later the Cape Verdean voters confirmed their votes, electing in the presidentials a political newcomer over the incumbent president (Traore 1991; Blaustein & Flanz 1994, vi), and later in the year the first multiparty local elections were held (Blaustein & Flanz 1994, vi). The peaceful transfer of power was possible not least because the socio-economic situation of the country was better than most others in Africa. An observer enumerates some favourable conditions: "The economy was relatively sound with income per capita increasing. Quality of life had improved, life expectancy had increased and infant mortality had dropped. The country had even embarked on a reafforestation to rehabilitate the ecology" (Medhanie 1993, 63). However, in August 1993 reports emerged of an armed coup against the government, and although the reports were quickly denied, they did serve to strengthen rumours that the Movement for Democracy is troubled because of internal divisions (NTW, 46).

The Case of the Comoros

The Indian Ocean islands of the Comoros are given only 2.7 points by

Hadenius. Again, this is for good reasons. At the end of 1988, the country, which in 1978 banned all political parties and turned into a one-party state (Newitt 1984, 66), represented an authoritarian state type. According to Hadenius's classifications the elections were anything but open, organizationalfreedoms circumscribed, political violence and oppression occurred. In the 1987 elections candidates in opposition to the government party were allowed to stand, but although they obtained 35 percent of the vote, the government party won all the seats (Derbyshire & Derbyshire 1989, 381).

The situation underwent a sudden change when in late 1989 President Abderemane was assassinated to be succeeded, in accordance with the Constitution, by the President of the National Assembly. This change of government unleashed a demand for multiparty democracy and suddenly no less than 16 political parties came into existence, eight candidates running in the presidential elections one year later (Oyowe 1991, 30). A period of political instability followed, marked by difficulties in installing cohesive governments, tensions between various parties and fractions, and even an attempted coup d'état by the Supreme Court in August 1991 (Oyowe 1991, 30-31). One observer has even found it fitting to state that "democracy is proving too big a price to pay for an archipelago as poor as the Comoros" (Oyowe 1991, 31). Be this as it may, and although the Comoros are facing dire economic difficulties, there is little doubt that the country has undergone and is undergoing a process of democratization. Measuring the level of democracy in the Comoros in 1994 would yield quite different results than those reported by Hadenius.

The Case of the Seychelles

Responding to international and local pressures, the one-party regime, plagued since its inception in 1977 by real and imaginary plots of counter coups (Moine 1990; Houbert 1992, 100), agreed in 1992 to end the nation's one-party state system, and a Constitutional Commission was established to prepare a new constitution. However, conflicts soon ensued in the commission between the leading party and the opposition, and the oppositionthereafter from the commission, which then finished its drafting of the constitution in little more than one week. The parties constituting the opposition objected to the draft, inviting two foreign experts, who in a memorandum as well as in public appearances pointed out defects in the draft constitution, which thereafter failed in a referendum on 15 November 1992 (Blaustein & Flanz 1992, v-vi). According to one of the experts, the draft failed to achieve many of the fundamental purposes for which constitutions are written in modern democratic systems. Fundamentalrights freedoms were not effectively guaranteed, several

international norms were violated in this respect, provisions for political plurality were faulty and inadequate, the draft described a political scheme designed to prevent true democratic debate (Blaustein & Flanz 1992, 119— 121). In January 1993, however, the Constitutional Commission reconvened,and first multiparty elections for 16 years were held in July 1993 (NTW, 169). Hadenius gives the Seychelles 2.0 points only; the new political order, to the extent that it will prevail and prove robust, obviously deserves a much better assessment.

Of course, the presentations in this chapter do not constitute a criticism of Hadenius's work. He has no doubt in his book provided adequate descriptions of the states he has investigated. By the irony of fate, however, events following almost immediately after the closing point in time of Hadenius's research have caused some small island units that appear in his data as unlucky creatures of democracy to abandon undemocratic components of their regimes and turn to democratic processes and structures. short, the relation between insularity and democracy that Hadenius able to track down, has been strengthened during recent years.

In fact, to the cases discussed above a couple of marginal notes can be added. On the one hand, in Vanuatu the 1991 national elections signalled a turning-point in the post-independent politics, the election featuring an intense multiparty competition and the country entering the politics of coalition-building to form a government (Premdas & Steeves 1992). Vanuatu well in Hadenius's classifications, being given 8.8 points. Democratic weaknesses are, according to Hadenius (1992, 68), to be found primarily in the sphere of the freedom of opinion; the recent developments may well serve to remove these weaknesses. Concerning, on the other hand, the Kingdom of Tonga, which scores an indifferent 5.4 points in Hadenius's list, social changes, including the rise of an educated and articulate middle class have produced politicians who are impatient with their limited role and are subjecting the king's government and policies to closer scrutiny than ever before (Campbell 1992; Hills 1993). It may well be the case that Tonga faces in the future a need to advance further on the path to a full-fledged democracy.

---

NOTES

1. For information about the religious profiles of the various states, the handbooks Hela vårlden i fakta '86, Handbook of the Nations (9th edition) and The Times Guide to the Nations of the World (1994) have been consulted and checked against each other.

---

2. Namely: Antigua and Barbuda, Bahamas, Bahrain, Barbados, Brunei, Cape Verde, Comoros, Cyprus, Dominica, Federated States of Micronesia, Fiji, Grenada, Kiribati, the Maldives, Marshall Islands, Mauritius, Nauru, St Kitts and Nevis, St Lucia, St Vincent and Grenadines, Sao Tome and Principe, the Seychelles, Singapore, the Solomon Islands, Tonga, Trinidad and Tobago, Tuvalu, Vanuatu and Western Samoa.

3. Remoteness is operationalized here in terms of isolation rather than geographical distance: the measure used is the number of foreign destinations from the capital airport or the main airport in all small island states. The average number of destinations is 27; those states having 20 destinations or less are classified here as remote. Data are from Anckar (1995a).

4. The small island state Mauritius is a case in point. To quote an authority on Mauritius: "In a society of class, religious, color, caste, and linguistic differences, there is no end to the demands on political leaders from the different Mauritanian communities" (Bowman 1991. 68).

5. The abbreviation NTW refers to The Times Guide to the Nations of the World (1994).

ACKNOWLEDGEMENTS

This article is based on a paper delivered at the XXVIIth Annual Meeting of The Finnish
Political Science Association, Åbo, 12-13 January 1995. The authors acknowledge with thanks
useful comments from other panellists at the meeting.

REFERENCES

Anckar, D. 1991. "Världens små östater. Populationen jämte jämförelsepopulationer", Åbo:
Meddelanden från Ekonomisk-Statsvetenskapliga fakulteten vid Åbo Akademi, Ser. A:350.

Anckar, D. 1995a. Managing Intensity? Constitutional Moralities of Small Island States.
Manuscript, forthcoming.

Anckar. D. 1995b. "Om öar", Finsk Tidskrift Vol. 118 Nr. 5 (forthcoming).

de Backer, R. 1991. "Western Samoa. Universal Suffrage at Last", The Courier, No. 128,
July-August, 29-36.

Bartolini, S. 1993. "On Time and Comparative Research", Journal of Theoretical Politics 5,
131-167.

Blaustein, A. & Flanz, G., eds. 1992, 1993, 1994. Constitutions of the Countries of the World.
Release 92-8: Seychelles Supplement; Release 93-3: Såo Tome and Principe; Release 94-3:
The Republic of Cape Verde. Dobbs Ferry, New York: Oceana Publications.

Bowman, L. W. 1991. Mauritius. Democracy and Development in the Indian Ocean.
Darthmouth: Westview Press.

Cahen, M. 1991. "Vent des iles. La victoire de l'opposition aux iles du Cap-Vert et å Sao
Tome e Principe", Politique africaine 43, 63-78.

Campbell, I. C. 1992. "The Emergence of Parliamentary Politics in Tonga", Pacific Studies
15, 77-97.

Dahl, R. A. & Tufte, E. 1973. Size and Democracy. Stanford: Stanford University Press.
Derbyshire, J. D. & Derbyshire, I. 1989. Political Systems of the World. Edinburgh: Chambers.

Dogan, M. & Pelassy, D. 1984. How to Compare Nations. Chatham, New Jersey: Chatham
House Publishers.

Ghai, Y. H. 1988. "Constitution-Making and Decolonisation", in Ghai, Y. H. ed., Law,
Politics and Government in the Pacific Island States. Suva: Institute of Pacific Studies,
University of the South Pacific.

Hadenius, A. 1992. Democracy and Development. Cambridge: Cambridge University Press.

Hamilton-Jones, D. 1992. "Problems of Inter-Island Shipping in Archipelagic Small Island
Countries: Fiji and the Cook Islands", in Hintjens, H. M. & Newitt, M. D. D., eds., The
Political Economy of Small Tropical Islands. Exeter: University of Exeter Press.

Handbook of the Nations, 9th Edition, 1989. New York: Gale Research Inc.

Handelman, H. 1994. "Book Review: Democracy and Development" (by Axel Hadenius),
The Journal of Politics 56, 314-317.

Hela världen i fakta '86, 1985. Stockholm: Bonnier Fakta Bokförlag.

Hills, R. C. 1993. "Predicaments in Polynesia: Culture and Constitutions in Western Samoa
and Tonga", Pacific Studies 16, 115-129.

Hodges, T. & Newitt, M. 1988. Sao Tome and Principe: From Plantation Colony to Microstate.
Boulder, Colorado: Westview Press.

Houbert, J. 1992. "The Mascareignes, the Seychelles and the Chagos, Islands With a French Connection: Security in a Decolonised Indian Ocean", in Hintjens, H. M & Newitt, M. D. D., eds., The Political Economy of Small Tropical Islands. Exeter: University of Exeter Press.

Kendall, W. & Carey, G. W. 1968. "The 'Intensity' Problem and Democratic Theory", The
American Political Science Review Vol. LXII, 5-24.

Mani, A. 1993. "Negara Brunei Darussalam in 1992", Southeast Asian Affairs, 95-109.

Medhanie, T. 1993. "The Patterns of Transition to Democracy", The Courier, No. 138,
March-April, 63-66.

Moine, J. 1990. "Seychelles: elements de la situation", L'Afrique et l'Asie modernes. Vol.
166, Automne, 18-33.

Newitt, M. 1984. The Comoro Islands. Struggle Against Dependency in the Indian Ocean.
Boulder, Colorado: Westview Press.

Niu, L. M. 1988. "The Constitution and Traditional Political System in Tonga", in Ghai, Y.
H., ed., Law, Politics and Government in the Pacific Island States. Suva: Institute of Pacific
Studies, University of the South Pacific.

O'Kane, R. H. T. 1993. "The Ladder of Abstraction: The Purpose of Comparison and the
Practice of Comparing African Coups d'État", Journal of Theoretical Politics 5, 169-193.

Oyowe, A. 1991. "The Comoros. In Dire Economic Straits", The Courier, No. 130, November-December,

Premdas, R. & Steeves, J. S. 1992. "Vanuatu: The 1991 Elections and the Emergence of a
New Order", The Round Table No. 323, July, 339-357.

Samana, U. 1988. "Decentralisation, Decolonisation and Development in Papua New Guinea:
The Morobe Experience", in Ghai, Y. H., ed., Law, Politics and Government in the Pacific
Island States. Suva: Institute of Pacific Studies, University of the South Pacific.

Sannerstedt, A. 1994. "Teorier om demokratisering", in Sannerstedt, A. & Jerneck, M., eds.,
Den moderna demokratins problem. Lund: Studentlitteratur.

Sartori, G. 1970. "Concept Misformation in Comparative Politics", The American Political
Science Review 64, 1033-1053.

Sartori, G. 1987. The Theory of Democracy Revisited. Part One: The Contemporary Debate.
Chatham, New Jersey: Chatham House Publishers.

Sartori, G. 1991. "Comparing and Miscomparing", Journal of Theoretical Politics 3, 243-257.

Sutton, P. & Payne, A. 1993. "Lilliput under Threat: the Security Problems of Small Island
and Enclave Developing States", Political Studies Vol. XLI, 579-593.

The Times Guide to the Nations of the World, 1994. London: Times Books

Thorndike, T. 1991. "Politics and Society in the South-Eastern Caribbean", in Clarke, C,
ed., Society and Politics in the Caribbean. London: Macmillan.

Traore, A. 1991. "Cape Verde. A Mudanca-Change", The Courier, No. 127, May-June, 10-11.

Traore, A. 1992. "Sao Tome & Principe. An Alternative to Cocoa?", The Courier, No. 136,
November-December, 11-16.

Weaver, R. K. 1994. "Review: Democracy and Development" (by Axel Hadenius), American
Political Science Review 88, 778-779.