Steps Towards a “Universal Religious Grammar”

Antonianum LXXXIV (2013) xxx-xxx

STEPS TOWARDS A“UNIVERSAL RELIGIOUS GRAMMAR”*

Summary: The new scientific study of religion tries to identify universal patterns at various levels. In-side this program there arises the hypothesis of a possible “universal religious grammar”. To test that hypothesis, eight samples totalling 1700 students, from 8 different religious and cultural populations, have been surveyed using a specifically designed questionnaire of 65 items. The main goal was to as-sess levels and forms of self-transcendence as a measure of religious cognition. The data obtained has been processed applying factor analysis and comparing the ways in which the items become clustered in each case. It is proposed here that some broadly shared factors thus far obtained could represent a prime architectural framework of religious cognition, even if the collected evidence cannot be under-stood as a proof for the existence of an “universal religious grammar”.

Sumario: El nuevo estudio científico de la religión intenta identificar esquemas universales en varios niveles. En este programa surge la hipótesis que indica una posible “gramática religiosa universal”. Para probar tal hipótesis, han sido evaluadas ocho muestras, que suman aproximadamente un total de 1700 estudiantes, de ocho poblaciones de diferente religión y cultura, usando un cuestionario específicamente ideado de 65 ítems. La finalidad principal fue la de apreciar niveles y formas de au-totrascendencia como medida de cognición religiosa. Los datos obtenidos fueron procesados aplicando el análisis de factor y comparando los modos como los ítems llegan a agruparse en cada caso. Se pro-pone que algunos factores ampliamente compartidos hasta aquí obtenidos podrían representar una primera estructura arquitectónica de la cognición religiosa, si bien la evidencia recogida no se pueda entender como prueba en favor de la existencia de una “gramática religiosa universal”.

1. Introduction

The idea of a ‘general religious grammar’, ‘pattern’ or ‘structure’ underlying all religious forms and expressions, is quite modern. However, early versions might be traced back to ancient views on “natural religiosity”, which could re-flect deeply shared human abilities.

* The present paper uses data provided by Liselotte Frisk (Sweden), Erik Sengers (Holland), Boniface N’Guessan Kouassi (Cote d’Ivoire), Fenggang Yang (China), Fatma Sundal (Turkey), Benedictus Simamora (Indonesia), and Anne Kull (Estonia); their generous and voluntary effort has been crucial for this research. We are grateful to Jay Feierman as well, who made very useful suggestions to correct and improve our paper draft.Contact: [email protected]

RIVISTA ANTONIANUM 3-2013.indd 431 28/06/13 10.47

Lluís Oviedo, ofm432

Phenomenology of religion – already in the XX century – offers an inter-esting locus for developing theories about constant religious clues, as some of its classical practitioners not only suggest such a natural tendency, but also try to describe its elementary structure. Key thinkers in this approach have been Rudolf Otto, Mircea Eliade and James Fowler.

The anthropologist Claude Levi-Strauss developed an explicit theory about a ‘universal grammar’ as reflected in religious myths. His work on ‘struc-tural anthropology’ reveals the existence of elementary forms, composed by binary codes, underlying religious and magical narratives, shared by all human cultures. The structural view of the human mind, inspired in contemporary linguistics, has helped to understand many aspects of human cognition and behaviour. Categories distinguishing nature and culture were the foundation of many other distinctions, providing the required mental map to organize life and society; including religion (Levi-Strauss 1963).

More recent theoretical developments can be found in sociology and an-thropology, which have also worked implicitly with the presupposition that religion figures as a sort of “universal trait” in all societies. “Rational-choice theory” and “systems theory” may be able to contribute insight into how that common structure can be understood. Although in these theories the personal dimension becomes secondary, nevertheless concrete assumptions regarding how individuals deal with transcendent goods or how individuals interact as to achieve a social systemic form are at their base. These paradigms are use-ful for describing religious experience in universalized, analytical dimensions (that is, as gain vs. loss), as well as in terms of its own internal communication code (broadly, transcendence vs. immanence). The tested, heuristic power of these methods consistently renders religious-social experience intelligible and describable in terms of its most generalized patterns (Stark & Finke 2000; Luh-mann 2000).

The latest versions presupposing a universal religious baseline resort to biology (Söling 2002; Feierman 2009), neurology (Newberg, d’Aquili, & Rause 1999), linguistics (Rosemont & Smith 2005), cognitive psychol-ogy (Boyer 1999; Atran 2002), or to a multidisciplinary approach (Grassie 2010). In any case, as far as we know, the currently mentioned theories do not provide a ‘grammar’ or ‘basic religious code’, but just a description of our universal tendency to conceive or assume religious beliefs and to behave in consequence. Perhaps the recent work of Justin Barrett comes closer to that goal, as he tries to describe the most common features of a ‘natural religion’ (Barrett 2011, 132 f.)


Manuel Canteras, Steps Towards a “Universal Religious Grammar” 433

The hypothesis of a ‘universal religious grammar’ assumes a very general form. To specify what this “basic structure” could be is not easy. The term “uni-versal grammar” certainly recalls a similar concept used by Noam Chomsky, applied to general language skills (Chomsky 1965, Pinker 1994). Chomsky proposed a model of grammar as a ‘set of constraints’ used to build correct sentences for communication. This process is innate, can be observed in young children, and is embedded in neuronal circuitry; as such it can be described as universal. However, the analogy of religious cognition with a theory on how natural languages work may appear less convincing.

It might be useful to point to similar analogies, for example regarding a “universal moral grammar” (Rosemont & Smith 2005; Hauser 2008) and a “social grammar” (Eibl-Eibesfeldt 1989, p. 544). Other human experiences could be examined under that light, like the ‘aesthetic’ or the ‘emotional’ giv-ing place to respective languages. Ritualistic behaviour comes close as a can-didate that can be studied as a universal pattern (Lawson & D’Aquili 1990). However, some doubts are legitimate, since a natural language is not the same as a set of categories helping to make sense of the action’s value, social conventions, or religious feelings. To which extent religious forms might be described as a ‘language’ is an open question. Some arguments can be sug-gested: it configures a “linguistic game” (Wittgenstein); it is a form of com-munication resorting to a distinctive code (Luhmann); it is built on specific semantic terms and codes. Probably the Chomskian approach can furnish some further help, as will be shown. In any case, in all the mentioned research programs, the idea of religion – or moral – as a language has to be under-stood as a heuristic tool, providing some broad ground to come to a deeper understanding of such human features.

A contemporary line of research in the scientific study of human and so-cial traits looks for universal clues, innate cognitive and behavioural patterns, and basic processes underlying human development. This line is in sharp con-trast with the alternative, the one stressing more the cultural differences, and – at its base – the construed character of every human and social form, deeply entrenched with its own environmental and cultural features, and a clue for human mental plasticity (Evans & Levinson 2009; Everett 2012; Prinz 2012). This criticism can be perceived concerning religion as well (Martin 2011, 130 f.). The philosopher of mind David Chalmers has recently gone into the dis-cussion, claiming that the truths about the world are built from very basic ele-ments and forms of reasoning (Chalmers 2012).

Our intention is not to settle a long and deep debate, but just to test the following intuition: that religion can be studied as a broadly shared structure



deeply rooted in the human mind, and being constituted by a set of basic pat-terns or rules serving as the foundation for all further developments and reli-gious behaviours. To this end an ad-hoc questionnaire has been designed and a significant amount of data has been recorded from international samples of students.

The project hypothesis is manifold: - first, that the existence of a basic religious grammar could be perceived

through the persistence of capacities of self-transcendence, irrespective of one’s lack of religious up-bringing, and despite the lack of religious practice;

- second, that such a grammar could be identified through an elementary binary code contrasting forms of transcendence on one hand, to the de-fault view of an immanent or natural reality on the other;

- third, that such a code could identify different versions or expressions in which the mentioned contrast would be articulated;

- and fourth, that these many expressions, building on the elements avail-able, configure religious traditions, and biographical narratives (like proc-esses of conversion or de-conversion).

The collected samples would offer the opportunity to test the related hy-potheses, helping to build a more accurate model.

2. Method, theoretical background, and data sampling

Trying to discern whether a ‘universal religious grammar’ can be traced along different cultures and backgrounds would require extensive and ambi-tious field work able to register as many different religious expressions as pos-sible and to look for commonly shared features that could be articulated into a simple code or set of rules. An alternative to a research program entailing too long a time and too many resources, can be devised – with a more reduced scope and budget – sampling questionnaires from different cultural and reli-gious backgrounds and comparing the results with the application of standard statistical analysis. This simplified approach requires, first, to design an ap-propriate questionnaire able to register general and elementary perceptions of transcendence with its main modulations; and, second, to collect data, with this questionnaire, from as many contrasting cultural and religious milieus as possible, for later comparison.

The main challenge, when trying to describe broad expressions of tran-scendence, consists of defining those rules or codes with whom the religious



mind operates, i.e. which contribute to a kind of knowledge able to capture a transcendent dimension, beyond the habitual, natural or immanent. In this case it becomes very difficult to escape a form of circularity when this project tries to give its first steps: it is necessary to start with some idea of ‘religion’ to be able to establish its basic codes; however it would be rather those codes that should define or characterize religious mind and behavior. This circularity is not only methodological, but reflects some deeper difficulties when dealing with religion: any attempt to objectivize it has to account for subjective aspects (Olivetti 1995). Such a question renders always ‘incomplete’ every explanation of religion. However, trying to overcome this obstacle, an exploratory model is proposed. In this sense, to advance in some way, it is required to state a few hypothesis, which need to be tested through empirical methods, whose results contribute to a redefinition of the departure pattern.

In broad terms, it is convenient to envisage a research program in which the hypothetical forms that assume a basic religious code are suggested, keep-ing in mind the available theories; to proceed successively to elaborate a ques-tionnaire that reflects the main features of that code. The following steps con-sist of collecting data from representative samples of populations with different religious sensibility: religious populations belonging to distinct traditions; and populations with weak religious background or lack of religious formation. The results, after applying factor analysis, can already offer a first panorama that will help to verify and to correct the initial proposal. Following this model and based on the feedback between theoretical proposal and empirical verifica-tion, which in turn contributes to a correction of the initial proposal, ulterior steps can be taken.

There are many theories of religion and its basic forms, and it is not a sim-ple task to build a ‘theory of universal religion’, since that category is too fuzzy and plural. In trying to simplify and reduce the basic characteristics of religious thought, a choice has been made for the most abstract proposals, as those that derive from structural analyses and systems theory. From that point of view, combining the ‘structural anthropology’ of Levi-Strauss and the ‘theory of so-cial systems’ of Niklas Luhmann, a paradigmatic approach can be obtained in which religion becomes a social system that is constituted by communications based on the code that distinguishes what is determined from what is unde-termined; the immanent sphere from the transcendent one (Luhmann 1977). Although the nuances of these theories point to a richer conception of the religious communication and a certain debate arises on the ‘basic distinctions’ conforming the ‘religious code’, this orientation becomes useful when trying to outline the set of rules presiding over religious communication.



Probably our theory of choice, formulated in this way, is not of great help or utility for our purposes; hence the convenience of a more focused applica-tion. Broadly speaking, what counts is the idea that religion may be heuristical-ly studied as a language, a communication form, and as such it depends on how it establishes a code that, as is usually assumed, operates on basic distinctions. Maybe what is not so simple is to establish the distinctions that give content to this binary code, because the one that separates determinate/indeterminate is too abstract and indefinite. A way to put some more flesh into this elementary and abstract construction consists of proposing a list of distinctions that can configure axes around which the religious communication moves. Reflecting on possible related terms that complement or build up the basic structure, a possible list follows:

· Being - nothing → life - death · Order - chaos · Salvation - damnation [extinction] · Nature - grace · Sin - forgiveness · Absolute - relative · Temporal - eternal · Worldly - extramundane· Natural - supernatural /mysterious · Body - soul · Hope - desperation · Totality - part → universality - limitation · Alterity - isolation.An immediate objection is that in some of the cases, the distinctions can

be operated clearly ‘outside the religious system’, or they only converge in a lateral way with that code. In fact, it can be assumed that some complementa-rity among several of the approaches or proposed distinctions needs to be as-sumed: the contrast between transcendence and immanence seems one of the most basic; but also salvation/damnation (or ‘complete extinction’) appears as pertinent, and even more “specifically religious”, as Peter Beyer suggests (Beyer 2006, 82ff.). On the other hand the distinction absolute/relative could com-plete or give an almost-religious tonality to many other distinctions that can be ‘relatively secularized’. The semantics of the ‘absolute’ can play a decisive role at demarking the field of a proper religious code, or at least, its closest expressions. Looking at this set of contrasting pairs, a grammar makes sense only if it can



build in a recursive way (Chomsky 1965) some sort of larger combinations, i.e. constructing sentences that include variations of more pairs.

As a matter of choice, the designed questionnaire has avoided any reference to the words ‘God’ or ‘divine’ or ‘supernatural agent’. The reason is that this in-strument intended to be as broadly inclusive as possible, since it was addressed to religious and non-religious people, and to members of different religious traditions. A second reason is the kind of ‘minimalist program’ here entailed: it just tries to identify aptitudes of self-transcendence, what is not necessarily related to a personal supernatural being, but could be a previous condition for such a development. Such minimal perception can be perceived in many forms of ‘spirituality’, as something distinctive from institutional religions. It is linked to theological topics pointing to an anthropological structure that is a condi-tion of the possibility to receive and understand an explicit revelation (Rahner 1941). In any case, in the present research, ‘self-transcendence’ reflects again an heuristic tool trying to figure out elementary forms that religion can assume. It simply presupposes an ability in the human mind to move beyond the natural, physical, level of reality, and to point to alternative dimensions.

Trying to put into practice this theoretical model, and being aware of these difficulties, an initial questionnaire of 65 items has been devised. The question-naire covers three basic areas: questions about the initial religious formation, their possible advantages and disadvantages (12 items), questions on religious practice (4 items) and a central body of 49 items in which are proposed state-ments that translate into practical or experiential perceptions of most of the above mentioned distinctions (between 2 and 4 for each distinction). The items were proposed as positive sentences, and the respondents were asked to rank each one on a Likert scale of 5 levels, according to their greater or smaller agreement with the proposed statements. Three demographic items (gender, age and level of studies) completed the questionnaire. This built instrument can be seen at:

http://www.academia.edu/1025674/Questionnaire_Religious_basic_forms

This questionnaire has been administered in a first wave to a sample of 416 students of the University of Murcia, in Spain, in the autumn of 2008. It was a convenience sample, which gathered cases from courses in two Facul-ties. The analysis applied to this broad sample did not reveal big problems with the understanding and the distribution of items in factors. Further steps have been made in the two following years to translate and distribute the questionnaire to convenience samples of students from different countries,



cultures and religious backgrounds. The following table reports the current state of the research:

Country N Religious description language Mean age

Spain 416 Catholic mildly secularized Spanish 21,2

Sweden 103 Protestant very secularized English 25,6

Cote d’Ivoire – Togo 340 Catholics and Muslims French 28,9

Holland 323 Catholic and Protestant very secularized English 20,1

China 108 No religious adscription Chinese 20,9

Turkey 205 Muslim Turkish 22,1

Indonesia 99 Muslim English 21,8

Estonia 99 Christian Lutheran, religion repressed during communism

English 22,2

Total 1693

Samples in Israel and Russia have been attempted, obtaining – to date – a too scarce an answer. In most cases the questionnaires have been distributed to the students in their classrooms, to be completed in situ, or outside the les-son time. The Estonian students from several Faculties were invited to answer an online-survey. The African sample was collected by different means, mostly outside of the academic milieu.

The idea behind this procedure was to verify how these populations were able to identify forms of self-transcendence, gathering or distinguishing be-tween different items; such ways of clustering in the practice could reveal a closely formal organization of the religious mind, at least in the form of a “col-lective mind”. To this end, factor analysis, with Varimax rotation, has been ap-plied to the 65 items of the questionnaire. This method allows for gathering the items into clusters by means of affinity; a coefficient measures the correlation between the selected items and the belonging factor, or their greater or lower affinity. This constitutes an important step in our method, since the way the items are clustered in the different samples would reveal internal dynamics and the structure that assumes the religious dimension in each collective. Usually we selected those factors with alpha > 0,5. The mean of the main components of each factor was calculated.



3. Results

Convergences and divergences become apparent when the hypothesis, built on the theoretical background, is confronted with the outcomes of the empirical test. In our case, one thing is the speculative approaches to the re-ligious experience, and another one is how young people in their own experi-ence conceive transcendence or alternative dimensions to a worldly confined reality. The aim of the attempted survey has been to test whether and how the proposed model works in real settings; the results somewhat confirm and somewhat depart from the initial view.

a) On the persistence of self-transcending capacity

An important and expected result is that not all people have the same levels of religious perception. This can sound already like a disclaimer regard-ing the main thesis that justifies the present research: the existence of a ‘uni-versal religious grammar’. Indeed, it cannot be ‘universal’ a perception that in some cases is shared by only half of the surveyed sample. Taking a glance at the outcomes of more secularized societies, very elementary forms of perception of transcendence reach a meagre mean (about 3 on a scale of 1 to 5); many students feel unable to identify minimal levels of transcendence. This data prompts an unavoidable question regarding the initial program, trying to con-sider religion as a language, which usually is a universal human ability, while religion is not always so. Nevertheless, it can be conceived that such structure lies latent and can be activated at any time in life. It would be universal in the sense that when a person tries to deal with transcendence, he or she would usu-ally resort to – or activate – this language or code; it would be available to eve-rybody, and constrain – in a positive way – the possible outcomes. One is sure: when a distinction is made between an immanent and a transcendent level of reality, we move inside religious communication; and probably very similar rules would be applied on every occasion. In any case, it is important to remind ourselves that a population can be very secular, or even undergo administrative pressure against religious education and organizations, and nevertheless their individuals can keep a ‘not-yet-activated’, ‘half-activated’ or ‘fully activated’ ca-pacity for self-transcendence; its activation will depend on many factors in the own environment and personal history.

The former question requires further analysis and deepening. A result to take into account in the present research is that in most of our samples the fac-tor of ‘religious practice’ was distinct from other factors pointing to diverse ex-



pressions of self-transcendence. This outcome can be explained pointing to the fact that, in many cases, students ranking very low in the indicators of religious practice and religious up-bringing, or even critical regarding religious educa-tion, were able nevertheless to identify levels of transcendence. In other words: items corresponding to a general ability to identify forms of self-transcendence were often matching in separate factors as those in which items of religious practice and up-bringing were more paramount. This point can be explained, first as a proof concerning the presence and resistance of forms of ‘spirituality’, or fuzzy religious expressions not linked to more institutional religions. How-ever, a different reading could point to the persistence of ‘elementary religious structures’ that can work autonomously from other circumstances and even social pressures. It is interesting in this respect that samples like the Chinese, the Dutch and the Swedish, which show low levels of religious practice and upbringing, show nevertheless higher means in the factors pointing to self-transcendence.

b) On the modulations of religious mind

The second important insight, after checking the results of our surveys and applying a factor analysis, is that in all the samples this statistical procedure manages to distinguish dimensions of transcendence that are more or less re-lated, but at the same time can be separated when a sufficient number of cases have been gathered. The demarcation lines do not always correspond to those that were in mind when the instrument was designed, but nevertheless demon-strates the existence of articulations or modulations of the religious perception of a collective. Expressing the same idea in negative terms: the results do not provide a unique cluster of religiousness; they do not gather all the items of self-transcendence into a single factor; neither do they appear as scattered in a formless dispersion of data; they come articulated into clusters that in most cases reveal an internal logic or coherence, that was more apparent when the samples reached some dimensions (more than 200 cases). What emerges from this data is that individuals can combine their religious perceptions or experi-ences in a free way: more of this dimension, less of the other one; however when taking the individual approaches together, some internal logic or pattern arises.

The questionnaire intended to measure aspects or dimensions of religi-osity through a spectrum of items that are hypothetic indicators of self-tran-scendence orientations or perceptions. In broad terms, the proposed questions tried to check for several aspects: confessional religion, feeling of transcend-



ence, sense of absolute, of eternity, wish of salvation, sense of limits, need of forgiveness and of openness to alterity. These attributes are measured in a very approximate way, and perhaps it will be not possible to go much further. In-deed the distribution of items in factors is not always neat, but sufficient to consider separate clusters.

A caveat is required at this point. The presence of an item in different factors does not mean that the factors were determined in a confusing way, but rather that that item is somewhat ‘confusing’ and is less able to distinguish between one factor and the other one. This situation reflects several causes: the always fuzzy character of spiritual perception, often of vague contours; se-mantic difficulties at understanding the question; or perhaps it suggest that we need a correction and better formulation of that item.

It is important to remember that factors occupying the first places are those explaining greater variance, giving more contrast among the respondents, while the last factors show what is more common, the shared characteristics or the less variable elements.

Starting with the first and largest collected sample, corresponding to Spanish students with Catholic background, several tendencies were observed leading to more robust observations that could serve as a basis for further theo-retical development.

The following table shows the results of the factor analysis applied to the Spanish sample, and the factors that reached alpha > 0,5. Communalities were rather low, as was the mean of variability (see Appendix table). Nevertheless, broadly speaking, these results reflect to some extent the schema of distinc-tions that was suggested as expressing nuances or versions of the fundamental distinction between transcendence and immanence.

Factor order Factor designation Number of Items R> 0,4

alpha % var.

1 Religious practice 11 ,926 23,8

2 Transcendence 14 ,921 6,3

3 Refusal of religious up-bringing 6 ,839 4

4 Hope, absoluteness 5 ,753 3,4

5 Eternity 3 ,529 3

6 Alterity, community 4 ,586 2,7

7 Sense of limits 4 ,710 2,5

8 Immensity cosmos 3 ,591 2,2



9 Sense of guilt 3 ,664 2,1

10 Moral absolute 2 ,562 2,1

11 Religious upbringing out-of-family 2 ,656 2

12 Chance and necessity in nature 2 ,750 1,9

13 Immanent hope 2 ,602 1,8

Main factors from the Spanish sample; Varimax rotation; alpha > 0,5.

The first factor in this sample corresponds to religious practice and up-bringing. It is the factor that discriminates the most among the Spanish stu-dents. The second one is a broad container that includes 14 items describing forms of self-transcendence: capacity to project oneself into a different dimen-sion, and sense of mysterious connection with the whole of reality; furthermore other items point to the ‘need of changing our mind to understand reality’, or the need of extramundane help. The most revealing feature of this outcome is that such a dimension of transcendence, despite its fuzzy contours, does not gather a series of items that could be expected into this general cluster. Some interferences can be observed with the factor describing a sense of hope and absolute; but it is still surprising that factors such as those reporting a sense of limitation, sense of eternity, the immensity of cosmos, or yearning for a moral absolute, are not integrated into a general factor of transcendence. Some other traits, like the sense of guilt and the need to deal with it, or the sense of personal limitation, appear as autonomous and can or cannot be linked as accessory or supplementary elements to the central perception of transcendence.

All the described results point to a complex structure of the sense of tran-scendence, which can be perceived in different modulations. Some thesis can be formulated thereafter in the following terms: · The religious code consists of a core of perceptions and elementary intui-

tions about the possibility of a distinct realm, beyond and distinguishable from the immediate reality, which can be felt on several occasions and related to certain kind of experiences – like awe, personal limits, and great expectations.

· Around this central core, some other aspects can be switched on or off, depending of the extension of the code: the sense of eternal life or life after death, the sense of absoluteness, projection to others, infinity, and need of salvation, or getting rid of evil. Religious practice is related to a more engaged activation of the religious code.



· Some aspects of this ‘grammar’ are more substantial or basic, i.e. belong to everybody, and some others are linked to religious upbringing, socializa-tion, and obviously one’s own religious tradition; they appear in different degrees of intensity and awareness.

· The broad experience of transcendence can be fragmented into distinct dimensions and areas, extending or shrinking that experience, reaching more or less other dimensions of human knowledge and life. Some form of Chomskyan recursivity can be hypothized to explain the extension and combinations of basic elements.

c) On variations of the main theme.

Following our program, we have selected heterogeneous populations that allow us to test and further deepen the main and the derived hypothesis. It is worthwhile to give a look at the other seven collected samples to see whether the initial impression is more or less confirmed. The samples register great vari-ations in the number of cases and reliability; however it is convenient to take a look at them all together and to compare them.

In most samples (Turkey, Sweden, Africa, China, Indonesia, Estonia), the first factor is the perception of transcendence, or the broad container gathering a huge number of items corresponding to this perception. The second is usually the ‘religious practice’ factor. In some societies the ability for self-transcend-ence discriminates more than the religious practice, whose level – high or low – is almost taken for granted in the respective cultural milieus.

The factor analysis reveals striking contrasts between more Western coun-tries and the rest (see the summary table in Appendix). The samples of Africa, China, Indonesia and Turkey, give low levels of communality (the lowest being Africa and China); the items in these cases do not gather with sufficient inten-sity, but remain rather loose. The factors become less defined and differentiated among them. The Spanish, Swedish, Dutch and Estonian samples, in contrast give neater clusters, in which items manage to collapse into factors with high levels of communality. This difference could express somewhat a cultural con-trast, two orientations of the religious mind: in one – the Western – individu-als manage to better differentiate aspects of religious perception or experience, while the other cultural milieus do not develop such a sense of differentiation and keep a more mixed or entangled view of transcendence expressions. How-ever, this is one interpretation, the other one could be linked to difficulties of



understanding of the items, or to the reduced number of cases in some of the samples. This point needs to be settled with more data collection.

Deepening this last line of argument, the factor designed ‘Broad sense of transcendence’, gathers more or less items in each sample. This is already an inter-esting result: such a big container of “transcendence” reaches a higher number of items in the Turkish and the Indonesian samples. This can be understood as an indication of the way to organize religious ideas or feelings in different cultural or religious backgrounds. Indeed, both samples do not reveal too many distinc-tions in the way the students conceive forms of transcendence. In other words, their conception is more integrated and so less differentiated. Probably it’s not by chance that both samples correspond to a similar religious tradition. Compar-ing the Spanish and the Dutch sample, for example, it is striking that in both European areas the students were more prone to distinguish between different expressions of transcendence. However it is difficult to deduce a logical pattern: in the Swedish sample, the ‘transcendence factor’ gathers 15 items; in the Span-ish 14, and in the African, just 9. The Turkish case reveals, in this first factor, the presence of traits like “connectedness”, “sense of limits”, “animism”, which would appear separately in other surveyed samples. The formerly suggested interpreta-tion can become a hypothesis: certain kinds of religious populations or traditions will keep a more integrated view of religion and spirituality; while other reli-gious traditions are expected to distinguish and differentiate more between such aspects. At the moment it is not clear which feature determines a tendency to greater or smaller differentiation. Since one of the main hypothesis we are work-ing on is that the universal religious grammar is composed of a core of an elemen-tary sense of transcendence, and that other “modules” or secondary expressions of transcendence can be switched on or off, a second derived idea is that some ways of religious development or expression could mean more or less levels of differentiation in this field.

Going on with our analysis, the big transcendence factor reveals a quite familiar pattern. A striking fact is that in almost all the samples, from very different religious and cultural background, this first factor gathers several ex-pressions of self-transcendence that are always clustered in the same way. After analysing the eight available samples, “transcendence” is the label we give to a factor appearing always in the first or the second place of the list. In seven of the eight samples (the Indonesian one is shorter and less reliable) a shared number of items emerge:

· There are many things happening that we cannot explain with the usual logic· In this world there is more than just what can be seen and felt



· When I look at some landscapes I feel there is something more than what I can perceive

· Living beings are connected in a mysterious way· The soul is immortal· It’s hard to think that the world we know is just the result of spontaneous

processes and by chance

Can we say that this constitutes the “hard core” of the religious experi-ence and its central element, the basic motive of its grammar? Well, the work still proceeds on a hypothetic level. However – just at this level – what is re-vealed in these results is that the heart of the perception of transcendence – as opposed to immanence, or the lack of this awareness – is constituted by an elementary perception of “otherness” or a distinct dimension to the real, be-yond the empirical and immediate world, the one perceived through the sens-es; furthermore, this dimension links, in a mysterious way, living beings. As a consequence, the world can hardly be explained just by spontaneous processes (mysterious causes or agencies seem to be suggested in this context). And, as an intimately connected element, that awareness involves the perception of a spiritual dimension in humans.

The list of factors more or less follows the pattern already observed in the Spanish control case. However different patterns can be observed: the Dutch group gives four main factors highlighting a sense of limitation: moral, in nature, cognitive, and of humanity; the sense of guilt is present in most samples, as is the sense of wholeness and eternity, or wonder. However, beyond this point, the way in which the items become clustered varies in great measure from one sample to the next, beyond some basic and expected distinctions. The most fundamental distinction, already underlined, divides between religious practice and upbring-ing, on the one hand; and sense of transcendence, or the other. However, moving on from the Spanish sample, it is harder to identify an organization of items in factors that could reflect quite smoothly the schema that served as the inspira-tion for this questionnaire. This perception allows different interpretations: that we need bigger samples of at least 400 cases, or more likely; that the translations were not too accurate and could not transmit all the foreseen nuances; or that the students from different cultures could not always grasp in equal form the intended meaning of each question, especially when formulated in a foreign lan-guage. However, beyond methodological issues, a deeper question arises: that the alluded circularity between religious expression and religious schemas is at work in this case as well; the cultural mediation becomes too strong to be ignored.



Some of the factors are arguably suspicious of being on the side of im-manence. This can be the case of one factor that has been labelled “Sense of alterity”; it gathers items expressing a longing to overcome loneliness and shar-ing with other people. Our aim at including these items was to assess in which measure such tendency could be considered a form of self-transcendence that could be more or less linked to other expressions of spirituality, as suggested by a modern philosophical tradition stressing the value of alterity. In the present survey this is not the case. We can discuss how much such a desire to live to-gether or to overcome loneliness can be taken as a sign or indicator of self-transcendence; however in most samples we lack any item revealing such an af-finity, at least for the surveyed population. In this sense the openness to alterity can hardly be considered as a form of real transcendence.

Further results can be provided that confirm other parallel research in course, to which they appear as complementary. It is worthy to comment on the issues of gender and of ‘religious pro-sociality’. The variable of gender does not usually correlate in a significant way with variables of religious practice or transcendence. It is possible to identify in some of the samples (Spain and Hol-land) a greater correlation with factors of alterity or sociality, in the sense that females are a little bit more sensitive to alterity and relationship. In some way this result confirms the idea that religious perception is more universal, or less dependent on demographic variables such as gender.

The issue of prosocial leaning and religion has been discussed often in the last decade. From the collected data, broadly speaking, the factor that expresses social attitudes or a longing for alterity is usually separated from the factors clustering items of religious practice or sense of transcendence. It is important to highlight the meaning of this data. As in the former result, the ability for self-transcendence appears as something relatively autonomous or differenti-ated from many other variables or factors, and this reveals a kind of ‘independ-ent faculty’ of human mind, following its own rules and codes, and less a ‘by-product’ or an ‘epiphenomenon’.

Summarizing our results, the applied method does not provide enough evidence for the ‘universality’ of the proposed model of ‘religious grammar’, but only in a limited way. If all the samples would cluster their factors in a very similar way, then the hypothesis under examination would find more support. This is not the case, and therefore it is necessary to assume the strong imprint culture imposes upon religious structure and expression. Only at a deep level or in very formal terms could the view of a ‘universal religious grammar’ be legitimate. However this last point presupposes the distinction between form



or code, and content, in religious expressions. This goes far beyond the scope of the present paper, and probably needs a different methodological treatment.

4. Discussion: Religious universal grammar and its many implications

The present research is a first attempt – to our knowledge – to build a theoretical model of ‘religious grammar’ and to test empirically its applicabil-ity and extent. The proposed model resorts to a binary code of basic distinc-tions and to a system of recursivity able to link and integrate more features. It has produced several relevant results at an exploratory level, pointing to the heuristic value of the proposed model and to levels of religiosity that could be codified in a distinctive way.

Nevertheless, this research has clear limits, and it is important to rec-ognize them. The limits of the method are apparent. Some are linked to the non-existent budget with which this research project has been carried out. The good will and free collaboration of many colleagues has contributed to trans-late, distribute and collect the questionnaires. Most of the data processing has been the work of the two main authors.

The limits come to light when the reliability of the samples is tested. Re-sorting to students for this kind of survey has been already criticized, since these ‘people’ appear as little representative of the whole of their societies (Henrich, Heine & Norenzayan 2010). Some issues can arise from difficulties of transla-tion; or perhaps because of the tricky understanding of an English questionnaire by students of different languages. The sampling is another issue of concern, when the mentioned limits are considered. In any case, we consider attested the exploratory weight of this data; furthermore a great amount of evidence has been provided regarding the main hypothesis formulated at the beginning.

One of the aims of the present research has been to help to discern in which measure a commonly shared cognitive structure in the religious mind can be pre-supposed as underlying every form of religious behaviour. Religious studies have been divided on this issue. Some scholars point to the priority of ritual practices as the universal patterns that give rise to forms of the religious mind, that would be rather derived or even a ‘by-product’ of that learned and transmitted reitera-tive behaviours. Probably, at the current point of our research we will be unable to settle the question. In any case, it is convenient to go beyond this point. Some religious behaviour, like rituals, follow quite rigid patterns of repetitive perform-ance; in these cases a ‘basic ritual grammar’ can be induced from the observation of many sacred rites. However, religious behaviour is expressed in a much broader spectrum of attitudes and actions, like preaching, praying, doing penance, read-



ing, or engaging in charities. When that behaviour becomes so complex, then more cognitive elements come into play, more distinctions are operated, and de-cisions are made, even if this form of reflection may move inside quite predictable patterns or structures (Feierman 2009).

A different reason for concern emerges from the observed pattern show-ing a less differentiated ‘religious mind’ among non-western students in our samples. This point can be linked to the frequent claim in the anthropological study of religion that many populations do not follow the Western pattern of ‘religious distinction’: the separation between a secular and a sacred realm as a condition of the religious mind and the social structure (Asad, 2003; Wood-head, 2010, 29). This claim, if verified, would challenge any universalistic view, in favour of more culturally driven forms. The issue goes far beyond the object of the present paper. Several voices have recently risen against the scientific agenda and its reductionist view regarding religion, unable to account for the vast variety or religious experience (Martin 2011; Bellah 2011).

A final issue that transcends the frame of our research, and nevertheless becomes deeply relevant concerns the models of human cognition. We have chosen a more computational model based in dual processing; however, no-body is completely sure about which is the most reliable model of human cog-nition. Alternative paths can be proposed, as is the case of probabilistic and fuzzy reasoning. In some cases the evidence points to a more Boolean approach in the way we humans process relevant information (Day 2007). Again, we need to restrict ourselves to the hypothetic field: it is like waiting for a simple ideal type, which probably just reflects some aspects of human real cognition, but does not exhaust its many forms and mechanisms.

The important point is whether and how the life of an individual or a community changes when the self-transcending code is activated, and hence, a distinction is operated between worldly things and affairs, and an alterna-tive realm, opening to different views, hope, and able even to re-frame many worldly affairs. In our opinion, the existence of that grammar opens up many possibilities, as can be traced in the religious history of humanity; it provides the basic conditions giving rise to religious ideas, on which a proper engage-ment or behaviour can be conceived. In any case, such grammar needs an ap-propriate elaboration as a condition for their effective behavioural and even social exercise, something that requires much more than an elementary code; this last idea provides only a simple syntax; ulterior developments are a differ-ent thing and require a deeper historical reconstruction.

The last observation leads to some adjustment in the current discussion regarding the religious mind and its workings. The cognitive approach to reli-



gion has generally stated the capacity of the human mind to process informa-tion following some patterns that lead to the inclusion of ‘supernatural agents’ into the usual view of reality. Most of these processes become clearly uncon-scious or move in the field of bias and default mental positions. The research introduced here delivers a slightly different view: the human mind processes information following dual patterns or distinctions. In this case, the ‘religious mind’ is not different from other ways in which relevant information is proc-essed in order to deal with our environment or to cope with difficulties and threats. It just introduces a code operating by its own distinction, which can be formulated in various ways, but for reasons of opportunity ‘transcendence vs. immanence’ seems more pertinent. This distinction sets the conditions for further spiritual or religious developments. That is not yet ‘religion’ as we know it in its historical expressions, but just an elementary structure on which these organized religions are built, surely resorting to more distinctions and a much higher elaborated code.

What the former development suggests is that the cognitive basis for reli-gion is to be looked for somewhere else, and that this program becomes useful when fuzzy forms of spirituality would be discriminated from more elaborated religious expressions. By the same token, this system allows for a better un-derstanding of the difference between unconscious and conscious processes in religious cognition and behaviour. The revealed unconscious structure does not prevent that conscious processes, always bounded inside this code, may be developed and give rise to an infinite set of cultural elaborations.

Some additional lessons can be learned in the process. The assumed meth-od implies that religious categories are formed in a mind inside a culture, or rather a sub-culture of young students. What the factor analysis shows is how a ‘social mind’ works, or how a mind connected with other people’s minds op-erates. This is quite evident when considering that the factors are the result of comparing a broad sample of questionnaires: what emerges is not the sin-gle mind construction, but how this particular mind, on average, can operate distinctions and choices that are reflected in a broadly shared mentality. The consequence is that this method – and we suspect everyone – cannot avoid the social dimension of religious experience and to isolate ‘the religious mind’.

5. Directions for further research

The difficulties found and the mixed quality of some results probably point to one characteristic of religion, already highlighted and that clearly affects every attempt at studying it: its circularity and its unavoidable social and cultural dimen-



sions. When part of the outcomes show intricacies in the translation, the expres-sion and the understanding of elementary experiences of self-transcendence, what is meant, among other things, is how much such experiences are deeply entrenched with and configured by the cultural and religious frame in which they are lived or felt. Probably this difficulty is insurmountable and poses a limit to every research program looking for a universal cognitive religious code. In any case this should not deter further attempts at looking for such cognitive structures and exploring the religious mind and its relationship with religious behaviour, as one of the most intriguing issues that anthropological research can try to fathom.

More questions arise at the end of this research. Are there independent lines of evidence for a religious grammar other than what we measured and factor analyzed? Surely other ‘innate grammars’ need to be considered, if this program is to be fully advanced. The case for human social behaviour is a good example; sometimes it is missing or limited in persons with autism. Then, more research is needed to settle whether the hypothetic self-transcending grammar (or pre-religious) is a phenotype that has to get activated by something in the environment.

Furthermore, it might be possible to formalize, at least in an ideal form, the structure of the religious mental code. Possible candidates are the Laws of Form of George Spencer-Brown (1969), built on the logics derived from ‘drawing distinctions’; and the ‘Neural networks’ logic, that could help to understand how codes interact and form networks of higher complexity and complementary meaning.

One last consideration would be made regarding the applied method. The authors express their conviction about the utility of an approach that combines in a sequenced way theoretical development and empirical testing to further reshape the interpretative frame. In many cases, recent experience indicates that often the scientific study of religion has been unable to work on both sides: theory building and empirical or field research. When it does just one it some-times gets meagre and often confusing results. This can be better solved when both approaches can be combined in a singular research program. It happens sometimes that in trying to test a hypothesis some sidelines emerge bringing unexpected results, not contemplated in the working theories.

Lluis Oviedo ofmManuel Canteras



References

Asad, T. (2003). Formations of the Secular: Christianity, Islam and Modernity. Stan-ford: Stanford University Press.

Atran, S. (2002). In Gods We Trust: The Evolutionaly Landscape of Religion. Oxford, New York: Oxford University Press.

Barrett, J. L. (2011). Cognitive Science, Religion and Theology: From Human Minds to Divine Minds. West Conshohocken, PA: Templeton Press.

Bellah, R. N. (2011). Religion in Human Evolution: From the Paleolithic to the Axial Age. Cambridge,MA, London U.K.: Harvard University Press.

Beyer, P. (2006). Religions in Global Society. Abingdon, New York: Routledge.Boyer, P. (1999). Cognitive Aspects of Religious Ontologies: How Brain Processes

Constrain Religious Concepts. In Tore Ahlback. Åbo (ed.), Approaching Reli-gion. Scripta Instituti Donneriani Aboensis XVII: I (pp. 53-72).

Chalmers, D. (2012). Constructing the World, Oxford, New York: Oxford University Press.Chomsky, N. (1965). Aspects of the Theory of Syntax, Cambridge, Ma: MIT Press.Day, M. (2007). Let’s Be Realistic: Evolutionary Complexity, Epistemic Probabilism, and

the Cognitive Science of Religion. Harvard Theological Review, 100(1), 47-64.Dupoux, E. & Jacob P. (2007). Universal Moral Grammar: A Critical Appraisal.

Trends in Cognitive Science, 11(9), 373-378.Eibl-Eibesfeldt, I. (1989). Human Ethology. New York: Aldine De Gruyter.Evans, N. and S.C. Levinson (2009). The myth of language universals: Language di-

versity and its importance for cognitive science, Behavioral and Brain Sciences, 32-5, 429-448

Everett, D. L. (2012). Language: The Cultural Tool. New York: Pantheon.Feierman, J. (Ed.) (2009). The Biology of Religious Behavior. Westport, CT: Praeger.Grassie, W. (2010). The New Sciences of Religion: Exploring Spirituality from the Out-

side In and Bottom Up. New York: Palgrave MacMillan.Hauser, M. D. (2008). Moral Minds: How Nature Designed our Universal Sense of Right

and Wrong. London: Abacus.Henrich, J., Heine, S. J. & Norenzayan A. (2010). The Weirdest People in the World?

Behavioral and Brain Sciences, 33(2-3), 61-135.Lawson, E.Th., & R.N. McCauley (1990). Rethinking Religion: Connecting Cognition

and Culture. Cambridge: Cambridge University Press.Lévi-Strauss, C. [1963] (1976). Structural Anthropology I, II. New York: Basic Books.Luhmann, N. (1977). Funktion der Religion. Frankfurt a.M.: Suhrkamp.Luhmann, N. (2000). Die Religion der Gesellschaft, Frankfurt a.M.: Suhrkamp.



Martin, D. (2011). The Future of Christianity: Reflections on Violence and Democracy, Religion and Secularization. Farnham U.K., Burlington VT: Ashgate.

Newberg, A., d’Aquili, E. G. & Rause, V. (2001). Why God Won’t Go Away: Brain Sci-ence and the Biology of Belief. New York: Ballantine Books.

Olivetti, M.M. (1995). “Filosofia della religione”, in P. Rossi (ed.), La filosofia 1, Tori-no: UTET, 137-220.

Pinker, S. (1994). The Language Instinct. New York: William Morrow.Prinz, J.J. (2012), Beyond Human Nature: How Culture and Experience Shape our

Lives, London, New York: Allen Lane.Rahner K. [1941] (1997), Hörer des Wortes, Sämtliche Werke 4, Freiburg i.B.: Herder.Rosemont, H. Jr., & Smith H. (2005). Is There a Universal Grammar of Religion?

Grand Rapids, MI: Eerdmans.Söling, C. (2002). Der Gottesinstinkt – Bausteine für eine evolutionäre Religionsthe-

orie. PhD Thesis: http://bibd.uni-giessen.de/ghtm/2002/uni/d020116.htm, accessed: 10.11.2009Stark, R. and R. Finke (2000). Acts of Faith: Explaining the Human Side of Religion,

Berkeley, Los Angeles, London: University of California Press.Spencer-Brown, G. (1969). Laws of Form. New York: Dutton.Woodhead. L. (2010). Implicit Understandings of Religion in Sociological Study and

in the Work of Hugh McLeod. In Brown, C. G., & Snape, M. (Eds.). Seculariza-tion in the Christian World. Essays in Honour of Hugh McLeod (pp. 27-40). Farn-ham, UK, Burlington, VT: Ashgate.



Appendix

Number of significant factors (alpha > 0,5) for each sample, and data re-lated to ‘transcendence factor’

Country N. Factorsα ≥ 0,5

Alpha factor trans.

% exp. fact. trans.

% exp. all factors

Order fact. trans.

Spain 13 ,916 10,03% 63,51% 2Sweden 11 ,866 10,84 76,13 1Cote d’Ivoire 7 ,634 4,59 33,34 3Holland 9 ,849 5,67 65,9 3China 10 ,763 6,58 41,54 2Turkey 5 ,819 13,56 66,72 1Indonesia 5 ,41 8,72 54,74 1Estonia 9 ,818 10,25 61,37 1

Questionnaire and more statistical analysis of respective factors at:http://www.academia.edu/1025674/Questionnaire_Religious_basic_forms


Steps Towards a “Universal Religious Grammar”

Documents