Plural religious beliefs: A Comparison between the Dutch and white South Africans

The concept of religious beliefs is distilled from the perspective of one’s belief in God. With regard to this belief in God we propose to distinguish between two dimensions: The personal versus the a-personal characte r of God and his transcendent versus his immanent nature. This leaves us with a plurality of beliefs in God. Does this plurality of beliefs exist in the minds of people in the Netherlands and in South Africa? Together with this we explore the relationship between church involvement and plural religious beliefs in both countries. We have found a sharp contrast between the Dutch and a sample of church-going white South Africans regarding secularization and church involvement. Nevertheless, we have found a highly similar structure of religious beliefs among both people.


INTRODUCTION
A com parative study on multiple religious beliefs between the D utch and white South Africans has to cope with some objections. First, the geographical distance between the two populations is enormous. Second, the Netherlands are considered to belong to the most secular countries of the world, whereas South Africa is very ISSN 0259 9422 = HTS 49/1A 2 (1993) P lu ral religious beliefs religious. But there are very strong historical bonds between both countries as well as a common language. A common core of culture permeates the counsciousness of both peoples. A very strong religious connectedness penetrates the hearts of most church members in both countries. The exchange between D utch and South-A frican theologians is rem arkable. D utch theological publications have always played and are still playing an im portant role on the theological scene in South Africa. Through the ecclesiastical mediation of preaching, worship and catechetics, one may assume that Dutch theological thinking influences the religious beliefs of white South Africans in an eminent way.
In the first part of this article, we analyse the concept of religious beliefs from that of the belief in God. In the second, we will formulate our research questions. In the third, we will determ ine the samples and measurements. In the fourth, we will present the analysis and results and in the fifth our conclusions anti inter pretations.

RELIGIOUS B EU EFS: BELIEF IN GOD
Analysing the concept of religious beliefs, as distinct from a belief in God, wc do not start from a functional, but from a substantive definition of religion, In the former, any goal or value that functions as an ultimate concern in one's life is considered to be religious. In the latter, only relating oneself to a transcendent reality, that is God, is considered to be the core of any true religion (Kaufmaim 1979). Within this substantive belief in God, we propose to distinguish between two dimensions: The personal versus the a-personal character of God and his transcendent versus his immanent nature.

Personal versus a-personal
The mainstream of theology today stresses the persona! character of God. Because the human being as a person exceeds all creatures, no better metaphor can be found to indicate the direction in which G od's being can be adequately approximated. In order not to be misunderstood, God is not regarded as a person next to other per-! sons, as if H e were singular but, instead, in an absolute way, as person, without arti* ■ cle (cf Tillich 1966).
At the same time, some objections have been raised against attrib u tin g the metaphor 'person' to God. From the perspective of political theology some suggest j that this metaphor connotes God as being an absolute monarch. Instead, some pro pose the symbol of the kingdom of G od should be replaced by that of G od's federal H J C Pieierse. P L H Scheepert & J A van der Vox republic in order to stress the image of God as liberating his people and promoting th eir justice and reconciliation (Everett 1989). From the viewpoint of feminist* theology, some suggest that the 'person' metaphor has the connotation of God as an absolute patriarch. Instead, they are in favour of replacing this metaphor by that of God as mother, friend and lover (McFague 1990).
More radical proposals have been form ulated and are still being formulated. The proponents wish to give up any personal or human metaphor and replace it with a-personal symbols, like depth, ground, grounding, source, origin, power, strength, inspiration, fulfilment, et cetera. A serious question here is whether an 'I-You' rela tionship betw een man and G od is any longer possible, or at least thinkable. And also, w hether praying is meaningful. W ithout that a long history of religious, espe cially ritual, practice would be abandoned.
The motives for replacing personal with a-personal metaphors stem from the socalled m odernization process, which is characterised by m an's striving for a rela tionship betw een him and his environment, in terms of his developing rationality (W eber 1920;Habermas 1982). During this process a growing conviction regarding the autonomy of the human subject is emerging. At the same time, a consciousness of the desiibjectivation of the surrounding cosmos is growing (Dux 1982). These two processes have led to a decreasing plausibility of the belief in God as person(al).
Some theologians think that the personal and the a-personal metaphors have to be considered not as real alternatives or poles of a dilemma, but as mutually cor rective and complementary (Schoonenberg 1986;Van der Ven 1991). But the real question is whether this type of scholarly dialectical and synthetical thinking really corresponds to what ordinary people think.

Transcendent versus immanent
In traditional theological thinking, the relationship between man and G od is assu med to be asymmetric. God is thought of as sovereignly weighing up m an's good and bad actions and one-sidedly determining man's destiny. In other words» God is regarded as a being of absolute transcendence.
In reaction, modern theologians strive to present an image of God in which the divine attribute of absolute transcendence is replaced by that of immanent transcen dence. For them, G od is not apathetic: he does not find himself outside or above man's life and actions, unaffected by their good and bad days. Instead, he is in their midst. His very nature is to be characterised in terms of immanence. By that, they m ean God is present near man, deeply engaging himself in m an's societal history and his individual biography. But, in this immanence, God is transcendent. By that, P lu ra l religious beliefs these theologians refer to God being sympathetic and compassionate to all human beings in all their concreteness at all places in all times. Me is involved in all their concrete lives, concrete adventures and fate in a way that cannot be achieved by ordinary human beings. In that im m anence. He is transcendent (Schillebeeckx 1989;Hartshorne 1989).
Some theologians even go a step further. They think that G od's transcendence has to be replaced radically by his immanence. For them, God is exhaustively the inspiring force striving for justice and love to be actively present am ong men. They only meet God in man's actions that aim at liberation from alienation and oppres sion. God is there and nowhere else. That is to say, God is neither absolutely nor immanently transcendent. Instead, He is absolutely immanent. For them, praying consists only of reflecting, contemplating, m editating and concentrating on one's own moral and mental processes (cf Van Buuren I963). 1 We intend to conceptualize all these viewpoints on religious beliefs, which one implicitly or explicitly present in the theological debate. We start from both dimen sions introduced previously: personal versus a-personal, and immanently transcen dent versus absolutely transcendent. From these, the conceptualization leads to four analytically distinguishable perspectives on 'belief in God'. (4) pantheism T he term 'panentheism ' used in the diagram above originates from scholars like Hartshorne (1948), Tillich (1966), and Schoonenberg (1986).

RESEARCH QUESTIONS
A fter having distilled four analytically distinguishable perspectives on 'belief in G od', a first set of questions refers to this plurality of beliefs: Does this plurality of beliefs in God exist in the minds o f people in the N etherlands and whites in South A frica? And if these plural religious beliefs do exist for both, to what extent are they comparable? Are there differences and to what extent do they differ?
Next to this general comparison, a second set of questions refers to the differen ces between members of denominations in each of both countries. Are some beliefs in God m ore typical for one denomination, and others more typical for other deno minations? We will explore these questions its well.
A third sot of questions refers to the relationship between a plurality of religious beliefs and the extent to which people are involved in church affairs. Some theolo gians declare that religious beliefs can often be found among people who are not involved in the institutionalized church. In this respect, they sometimes refer to the Constitution on the Church by Vatican II, where church membership was defined in terms of 'having the Spirit' (Lumen Gentium 14). The institutional church does not exhaust this 'having the Spirit'. Therefore, it also exists outside the church (R ahner 1976). O ther theologians weaken the connection between religiosity and church in volvem ent from the opposite side, by pointing at the phenom enon of ecclesial atheism . This refers to the num ber of church m em bers, who do not (seem to) believe anything (Zuiehner 1988). Among social scientists there is also a dispute regarding this relationship. Some authors propose that religious beliefs can develop strongly outside of the institutionalized church because contemporary religiosity is free, persona!, spiritual, on the move and open (T hungtv at 1985). O thers have found that, with a decline in church involvement, religious beliefs seem to disappear (Felling et aI 19N6). Whoever is right is not for us to decide. We will explore the relationship between church involvement and a plurality of religious beliefs in both countries.

SAMPLES
In the N etherlands, a two-stage random sample of the Dutch population was con structed within the framework of a research project 'Social and cultural develop ments in the Netherlands 1985'. In the first stage, a number of municipalities were selected in such a wav that the distribution of region (North, East, South and West) and the degree of urbanization (from small vilages to big cities) would be represen ted proportionaiely to the national distribution. In the second stage, people aged from 18 to 69 were randomly chosen from the registers of the municipalities selec ted. About 56 percent (N = 3003) of the respondents approached were willing to be interviewed for one and a half hours, during the winter of 1985-1986. This sample turned out to be representative of the whole Dutch population regarding sex, age and marital status as well as the combination of these characteristics (see Felling et (if 1987). f l o r a l religious beliefs In South Africa there were no means available to construct a strictly compara ble national sample. U nder these conditions and taking account of our main re search questions, we decided to approach a num ber of churches in the Pretoria region requesting them to co-operate in the research. Some churches refused, but o th e r churches responded positively: The Dutch Reform ed C hurch (i)R ('), the M ethodist Church in South Africa (M C), the Presbyterian Church in Southern A frica (PC), and the Roman Catholic Church (RC). These churches had registers containing the addresses of their white members in their communities, hut not for their black members. Out of these registers, addresses were randomly hut systema tically selected: 1 out of 3 addresses for the MC, the PC and the KC. and I out of 2 addresses for the DRC. At the first address, the man was asked to fill in the ques tionnaire, at the second address the woman and at the third address a boy or a girl older than 16 was asked to co-operate. Eventually 262 out of 954 people approa ched were willing to co-operate during the summer of 1989. which am ounted to 27 per cent. Then we removed the blacks from the sample and ended up with 220 valid questionnaires. Unfortunately we do not have adequate census data te» compare our sam ple with. We have to em phasize, though that this sample consists of highly educated respondents who live in the eastern suburbs of Pretoria which is consider ed to be a middle and higher class residential area. Therefore our sample cannot be regarded as representative of the whole Pretoria area.

Measurements
Concerning religious beliefs, a comprehensive conceptualization had been develo ped at an early stage of the research on social and cultural developments (Felling et al 1982;1986). This conceptualization was criticized and then refined by Van der V en (1990), whose conceptual scheme was introduced above. In conformity with this scheme, a number of items were selected, considered to be the most valid indi cators of the particular stances in this scheme, based on previous principal-factoranalysis on the Dutch data (cf Felling et ai 1986:53). These items were translated from Dutch into English so that the English items were considered to be contextual equivalents of the original Dutch ones in the cultural environment of South African church members. Eventually, we inserted into the questionnaire two items indi cating personalistic theism (in appendix 1 referred to as PT-itcms). three items indi cating a-personalistic theism (in appendix 1 referred to as AT-items), and two items indicating pantheism (in appendix 1 referred to as P-items). A combination of items referring to personalistic theism on the one hand and pantheism on the other hand / / J C Pitienc, P L H Sduepen A ƒ A mbs derVen would indicate panentheism. Unfortunately, we did not insert items in the questionnaire that exclusively and explicitly referred to panentheism.
In o rd er to assess church membership, the respondents of both sam ples were first asked w hether they considered themselves to he m em bers of a religious com munity, and then, if appropriate, of which church they considered them selves a member. This two-step quest ion-form at tends to restrict the number o f people who declare themselves to be church members, contrary to a so-called one*step questionformat w here one directly asks of which church the respondent considers himself to be a m ember (Eisinga e ta i 1990). In South Africa we found, according to o u rsam p ling design, m em bers of the churches m entioned above. In the N etherlands we found m em bers of the N etherlands Reform ed Church (N ederlands Hetvojrmde Kerk), the Reformed Churches in the N etherlands (G ereform eerde K erken in Mederland) and the Roman Catholic Church, as well as non-members.
The assessm ent of church involvement was based on a number of relatively straightforward questions, aimed to ascertain the degree to which people w ere/are actively engaged in church matters. In previous research, the two general categories of church involvement and non-involvement had been refined (cf Felling et al 1982Felling et al , 1987.
Within the category of church non-involvement, three types were distinguished: second-generation unchurched, that is people whose parents had not been members nor had they themselves ever been members, first-generation unchurched, whose parents had been members, but they themselves had not been, and form er church members, who had given up their membership recently. These refinements appea red to be useful in the Netherlands but not in South Africa, where a negligible num ber of non-m em bers was found in the registers of the churches. F o r the Nether^ lands, we re-collapsed these three types into one category of non-involvement. And within the category of church involvement, three types w ere also distinguished: marginal members, who considered themselves m em bers but who m erely atten d church activities once or twice a year; modal members, who attended church activi ties regularly, that is at least once a month and core members who attended church activities regularly as well as being active in church community organizations.
Let us first describe the relationship between the denominations and the degree of church involvement in both samples. The results a re contained in Tables la and  lb. ISSN 0259 *>422 -UTS 49 /1 2 (1993) P lu ral religious beliefs  twice a year. Among Roman Catholics we found relatively many such marginal members as we did among members of the Netherlands Reformed Church. la con trast, the percentage of marginal members in our South African sample am ounts to a mere 14,8 per ccnt. Hie contrast with the Dutch sample is even greater when one notices that there are no marginal Roman Catholics in th e South African sample. Next, we find that 41,9 per cent of Dutch church members can be considered modal members, that is they attend church services at least once a month, which percen tage am ounts to 46,9 in our South African sample. Even more striking is the catego ry of core m em bers -that is, the people who attend church services regularly and who are active in church community associations. In th e Netherlands, 22,2 p er cent of the church members can be considered core members, am ong which we found re latively many members of the Reformed Churches in the Netherlands. In the South African sample 38,3 per cent are closely involved in th e church community, among which we found many Methodists and fewer Catholics and Presbyterians, relatively speaking.

Comparability of religious beliefs?
Taking account of the description of both samples, we can distinghuish several rela tively homogeneous sub-samples that could be relatively heterogeneous regarding their religious beliefs, based on the criterion of church involvement in each country. But the result would have been difficult to survey. So, in o rd er to assess the com parability of the structure of religious beliefs critically but still be able to survey them, we conducted exploratory principal-factor analysis on all religious belief-items sim ultaneously, separately in three categories: church-going and 'u n ch u rch ed ' Dutch, and church-going South Africans, using conventional extraction criteria (Kim & M ueller 1978(Kim & M ueller , 1984. The results are reported in Appendix l.2 The structures of religious beliefs am ong the three groups m entioned above show a high degree of resemblance. Let us begin with w hite church-going South African people. We find a clear, simple structure among these people with three distinct em pirical dimensions. The first dimension contains beliefs referring to apersonalistic theism. The second dimension contains beliefs referring to pantheism. And the third dimension contains beliefs referring to personalistic theism. T he as sociation between a-personalistic theistic beliefs versus personalistic theistic beliefs is high and positive (Pearson r: ,45, p=,00), whereas these dimensions are insignifi cantly and negatively correlated with the pantheistic beliefs. The com bination of personalistic theism and pantheism does not appear from the analysis. Hence* we

H l C Pietene, P L HScheepers & J A van der Ven
Let us turn now to the results of this analysis with the denominations and church involvement as independents, and the beliefs mentioned above as dependents, sepa rately for the Dutch and the white church-going South African people. L et us start with the theistic views,  It appears that there are significant differences regarding theistic views betw een non-m em bers and members of D utch denom inations. Differences in th e N ether lands can be simply described as following the religious-orthodoxy continuum: The unchurched subscribe the least to theistic beliefs, followed by the Roman Catholics, hâve no indications of the existence of panent he ism in the minds of white Somli African church-going people.
Among Dutch church-going people this structure is even more simple. We find only two dimensions. One contains both personalistic and a-personalistic theistic beliefs. The other dimension contains the pantheistic beliefs. The corrélation be tw een both dim ensions is m oderate and negative (Pearson r = -,25, p = ,00). As among the South Africans, we have no indications that paneniheism exists among the Dutch church-going people.
The belief-structure of Dutch unchurched people is highly similar: There is also one dimension containing both personalistic and a-personalistic theistic beliefs, and another distinct dim ension containing pantheistic beliefs. But am ong these un churched people, the correlation between both dimensions is moderately positive (Pearson r =,17, p = ,00).
Reviewing these results, we found more sim ilarities than differences. The belief-structure of both the church-going and unchurched Dutch is simple and clear. O ne dimension may be summ arized as containing theistic beliefs and another as containing pantheistic beliefs. Hence, both the church-going and unchurched Dutch do not distinguish between personalistic and a-personalistic theism. This is the one and only difference with the w hite church-going South-A fricans. Among these people we did find this distinction. So, in terms of plurality, we could state that the plurality of beliefs is higher in South Africa because of this distinction. But we con sider this difference minor, because the correlation within this subsample between these dimensions is so high. Hence, we consider it, for the sake of comparison, justi fiable to distinguish two belief-dim ensions for each of our sub-populations: one theistic view and one pantheistic view.

Different levels of religious belief?
Next, we computed summed Likert scores for both dimensions and for all respon dents with valid scores on all items. The scale for the theistic view ranges from 5 to 25, and the scale for the pantheistic view ranges from 2 to 10. By means of one-way analysis of variance, we discovered that white church-going South African people had significantly higher scores on the theistic beliefs (23,3 versus 16,2, F -ratio-209, p=,00), whereas the Dutch had significantly higher scores on pantheistic beliefs (5,8 versus 5,1> F-ratio=15, p = ,00). This result conformed with our expectations. But what we thought o f as a more interesting issue was the matter of how these beliefs varied across both denominations and across levels of church involvement in both samples. That is why we executed two-way analysis of variance for both samples.

H J CPietmc, P L H Scheepen A J A von derVcn
In the N etherlands there appear to be major differences between both denom ina tions and church involvement. First, it appears that the D utch Rom an Catholics hold fairly strong pantheistic beliefs, stronger than those of non-members. Second* it appears that these pantheistic views are much less popular among members o f the N etherlands R eform ed Church and the R eform ed Churches in the Netherlands. These people tend to cling more to theistic beliefs, as we have shown in T able 2a. T hird, we ascertain a fairly simple relationship between pantheistic beliefs and church involvement: The less one is involved in church affairs, the m ore one sub scribes to pantheistic views. Hence, it appears that th e people who are loosely attached to church as an institution in the Netherlands, that is marginal members, hold fairly strong pantheistic beliefs.
These findings certainly do not hold for the South African sample, where in fact we found no significant differences at all. Hence, our sample holds fairly homo geneous pantheistic views; that is, they reject these views. Yet there are notable trends. The first is that pantheistic beliefs are least popular among Roman Catho lics and most popular among Methodists. The second tendency is that pantheistic views are subscribed to most strongly by modal members. This holds for atl main line denominations, except for the Roman Catholics.
A more distant view on the levels of pantheistic beliefs in both samples shows us a particular paradox. Whereas the Roman Catholics in the N etherlands are fairly loosely attached to the church and in favour of these views, Roman Catholics in South Africa appear to be closely involved in church matters and opposed to these beliefs.
In terms of plurality, again we have to conclude that the Dutch hold less homo geneous views than the South Africans. In the last part of this paper we will try to interpret these findings.

CONCLUSIONS AND DISCUSSION
We have ascertained a sharp contrast between the Dutch and the sample of white church-going South Africans. It appears that half of the Dutch population lives in a secularized world, far from the religious institutions. In the Netherlands, among the people that are more or less involved in the church, 35,9 per cent belong to the m ar ginal members, 41,9 per cent to the modal members and 22,2 per cent to the core members. Among the sample of the white South Africans, the figures are different: Only 14,8 belong to the marginal members (less than half those in the Netherlands), 46,9 per cent to the modal members (about as many as in the Netherlands) and 38,3 P lu ra l religious b cb cís then members of the Netherlands Reformed Church and the strongest theistic b e liefs are found among members of the Reformed Churches in the Netherlands. In our South African sample it appears that only m em bers of the M ethodist church subscribe less to theistic views than members of the other denominations who hoid fairly homogeneous views. And, both in the Netherlands and in South Africa, it a p pears, simply, that the more one is involved in church matter, the stronger o n e's theistic beliefs are.3 In term s of plurality, we would conclude th at there is less agreem ent among Dutch than among the people in our South African sample who hold fairly homogeneous views. Now, let us proceed with the pantheistic beliefs. per cent to the core members (almost twice those in the Netherlands). Set aside the position of the modal members, it seems hard to imagine greater differences.
Yet, we have ascertained a highly similar structure of religious beliefs among both people. We have found a theistic domain of belief, in which the personal and a-personal aspects go together. A part from that, we have found a pantheistic do m ain. Lastly, it is also worthwhile to mention what we did not find, namely in the domain of panentheism. Let us reflect on these results.
With regard to theism, among the Dutch we found no distinction between the belief in the personal and the a-personal aspects of God. This distinction was pre sent among the white South Africans. However, both aspects appeared to be highly associated among the latter. These findings sharply contrast with the Christian tra dition. In this tradition, the belief in God is uniformly interpreted in terms of per sonal images of G od, like Father, Creator, Redeem er, Liberator. Legislator, King, Judge, Friend. H e is supposed to think, feel, want and act in a way that transcends human conditions and boundaries, but in a m anner that nevertheless is marked by characteristics that stem from 'la condition humaine', although in an analogical way. T he relevance of o ur research finding is that this persona! image of God is not dis tinguished from an a-personal image by the Dutch, and appears to be highly associa ted with an a-personal image by the white South Africans.
This may lead to re-interpreting G od's image in terms of the tension between its personal and a-personal aspects. One of the theo-'logicat' questions implied in that refers to the nature of the antonym 'personal/a-personal\ What is the meaning of 'a'personal? Does this 'a-' refer to an opposition that is contradictory (black/notblack), contrary (black/white) or complementary, respectively implying (black/notw hite)? This question is not specifically theological, it is logical in nature, it per m eates a lot of semantic fields in all kinds of disciplines. Some theologians think that 'personal' and *a*personaP refer to a complementary opposition. They point to the complex nature of the concept 'immanent transcendence', which holds imma nence and transcendence together. The aspect of transcendence may refer to God as 'personal', whereas the aspect of immanence may relate to G od as 'a-personal'. The antonym 'personal/a-personal' may be considered complementary as the anto nym 'immanent transcendence' are. A whole new perspective of rethinking old, tra ditional theological concepts em erges here (Schoonenberg 1986;Van d er Ven 1991).
A nother interesting point that arises from our research refers to the fact that pantheism exists isolated from personal and a-personal theism. It is a phenom enon apart in both the Dutch and the South-African populations. But we arrived at ano ther interesting insight. Theism is connected with active church involvement, pan H J CPietme, P L H Scheepen A J A van der Ven theism more or less with ecclesial distanciation. The more one is involved in church activities, the more one is committed to theism and the less to pantheism and vice versa. This leads to the question of how this relationship can be understood. Two hypotheses may be form ulated. Logically speaking, the less one is involved in church activities, the less one is confronted with the main religious stream that is present in the Christian tradition, namely theism. From this, pantheism may be un-> derstood in term s of religious erosion, which is caused by decreasing church involve? mem. This interpretation has its point of departure in a more or less conventional approach to the relationship between religiosity and church, in which church in volvement functions as an indicator of religiosity. But ah alternative hypothesis, in order to understand what is happening, might be plausible as well: a more non-conventional one. Perhaps pantheism belongs to the underlying structure of people's religiosity, which is covered and oppressed by ecclesiastical mechanisms. Perhaps this deeper layer of religiosity, which refers to G od's presence in the midst of the world and hum an life, can develop itself in as much as it is liberated from eccle siastical pressure, that is ecclesiastical dogmatism, moralism and discipline. From the history of the Jewish and the Christian religion, we know o f a great num ber of examples in which all kinds of pantheistic understreams have been removed by the official tem ple and church authorities. The most striking illustration may be found in the so-called Baal religion in the O ld T estam ent times in which 'm other earth1 was associated with religious images, affections and rites. This kind o f 'natural theo logy' has always been confronted with resistance and condemnation by the official church in Christian times, especially in Calvinist circles. In this interpretation, the am bivalent relationship between 'natural' religiosity and church plays a major role. Religiosity and church do not always go hand in hand. The latter sometimes domes ticates, restricts and oppresses the former. The question is whether pantheism may be seen as an indicator of increasing secularisation or of increasing liberated religio sity. O ur study ascertained that pantheism is strongly present among both non members and marginal members in the Netherlands. But our finding that panthe ism is subscribed to more strongly by marginal members than by non-members may be interpreted as an indicator of pantheism as a mode of liberated religiosity. H ere, a new horizon of a-conventional thinking emerges.
An interesting result of our research -a negative one -refers to what we have not found. We found no indication of what we have called panentheism. That does not mean that it is not a relevant concept. Some highly regarded theological scho lars, who were very sensitive to what was religiously going on among people, such as H artshorne, Tillich and Schoonenberg, proposed that panentheism as a concept exactly expresses the metaphors and feelings that are implied in modern man's b e lief in G od. It refers to the tension between the internal and the external dimen sion of religiosity. The internal dimension relates to the spark or even flame of di vine inspiration that exists in the inner core of man and which perm eates his whole life. G od is more interior to man than man is intimate to himself, as Augustine said (interior intimo m eo). T he external dimension refers to G od's all-overshadowing presence that covers the whole world and holds all human beings together. G od is all-encompassing. At the same time, God lives within man and surrounds him (cf D evenish & G oodwin 1989). This concept is highly im portant not only from an Christian perspective. The dialogue between Christian and Buddhist thinkers points in exactly the same direction. It has been called metatheism. The word 'm eta' re fers to transcending the personal aspects of theism into the direction of God being an open mystery that has both internal and external dimensions (Kruger 1989). We expected to find this concept by factor analysis of the items on religious beliefs, especially by exploring some combinations of items that belong to personal theism and pantheism, because panentheism is supposed to be a dialectical synthesis of both. But we did not succeed. The reason may lie in the fact that we did not submit items that refer directly and exclusively to panentheism to our respondents. Here, an interesting perspective for further new steps in the field of religious research is dawning.
The results of our research, and the above reflection on the question of our con cept of God, put a challenge to our reformed thinking on God. The challenge is to engage in a thorough theological study of our reformed tradition, our interpretation of biblical revelation, and a hermeneutical-theological interpretation of our under standing of God in the the light of contemporary pluralistic beliefs among the peo ple to whom we communicate the Gospel. Such a study would enrich our theologi cal account of what we as theologians and the members of the church believe about God.
Endnotes H J C Pieterze, P L H Scheepen A I A h o t der Ven of the convergence criterion, for the South African sample. That is why we per* formed principal-component analysis for this sample.
3. There are two exceptions, though, in the South African sample, as indicated by the significant interactive term. One is that, in the Presbyterian Church, modal members subscribe more strongly to theistic views than do core members. The other is that in the Dutch Reformed Church, modal members subscribe more strongly to these views than both marginal and modal members. As there are so few exceptions that disturb the clear analytical picture, we consider them neg ligible.

Appendix 1 Factor analysis on religious beliefs
White South African churched (N=117) AT There is something like a supreme being controlling life AT There is something beyond this world AT I believe in the existence of a supreme being P To me, God is nothing but the valuable in human being P God is not up there but only in the hearts of the people PT There is a God who occupies Himself with every human being personally PT There is a God who wants to be God for us Dutch churched (N=1382) AT There is something like a supreme being controlling life AT There is something beyond this world AT I believe in the existence of a supreme being P To me, God is nothing but the valuable in human being P God is not up there but only in the hearts of the people PT There is a God who occupies Himself with every human being personally PT There is a God who wants to be God for us