October 22, 2017

Religion in the light of a general behavioural model

This paper, presented by Goldsmith at the 20th Annual Meeting of the Society for the Scientific Study of Religion (Somerset Hotel, Boston, Mass. October 23-26, 1969), represents an early attempt to articulate, in highly technical terms, concepts that would become significantly revised and reconsidered in later versions.

Published in Systematics: The Journal Of The Institute For The Comparative Study Of History, Philosophy And The Sciences, Vol. 8. No. 2 September, 1970, Pp. 91-100.

I shall attempt to describe in this paper what I consider to be the role of a model in scientific explanation, explain why such a model should be a general behavioural one, and enumerate the principle features of the variables that it will make use of. I shall then try to determine to what extent the concept-word ‘religion’ satisfies these requirements.

Since time is limited I shall have to make a number of assumptions. The first is that behavioural processes do not occur at random but for a specific purpose—to permit adaption to a given environment. The second is that, in general terms, there is but a single way of achieving this, and it is made use of throughout nature—though the higher the level of organization the more differentiated must be the mechanisms involved.

I refer to the use of a control-system which ensures the monitoring of successive moves or responses so as to correct errors or deviations from the optimum trajectory or from equilibrium taken four-dimensionally, as it must be in a behavioural context. Monitoring is achieved by isolating or detecting data relevant to the system in question, trans-ducting them into the appropriate informational medium and organizing them so as to constitute information.

The role of information in behaviour is best regarded, as Craik showed in (1943),1 as constituting a model or ‘template’ of the system* involved. Monitoring or ‘learning’ is made possible by the constant modification of this model in the light of each new experience in such a way that it becomes an ever better representation of the system, permitting in this way an ever more homeostatic pattern of responses.

* Sub-system plus its environment.

My third assumption is that scientific information is organized for the same purpose and in the same manner. In this way the scientist, instead of searching after ‘truth’, must simply be regarded as trying to improve his model. The behaviour that this model will permit, however, will differ from that of an organism or society functioning as an integral part of a balanced ecosystem in that it will display a degree of homeostasis in excess of phylogenetic requirements—and that by tending to disturb the ecological balance in the immediate interests of a group of individuals, may well be to the detriment of whatever constitutes their four-dimensional unit of phylogeny.

This brings us to the notion of a general behavioural model. There is a tendency during phylogeny for information to become concentrated in a central organization, or brain. The system thereby represented is treated as a single integrated whole and is thus classified in terms of a single classificatory system.

In Hellenic times, scientific information, however meagre and inaccurate, appears to have been organised in this way. Since then, however, there has been a breakdown in centralized control, and science has gradually split up into a host of almost watertight compartments—the jealously guarded preserves of specialists using different terminologies and methods.

However, the present vogue for multi-disciplinary research points to a growing need for the re-unification of these divergent fields of study. This is justified on the grounds that the total ecosystem developed as a single process and thereby constitutes an integrated four-dimensional whole whose different sections and phases can no more be examined in vacuo than can those of a developing embryo. This is further confirmed every day as we observe the widespread and often unforeseen repercussions of radical interferences with the total ecosystem’s basic structure, such as the building of a dam, the mechanization of agriculture, the use of insecticides, the establishment of an all-pervasive state welfare system—such repercussions can only be compared with the side effects on the human organism of antibiotics and other powerful drugs.

I am convinced that to establish the causes of any such interference or to predict its effect, data must be detected from almost every field of study, and must then be transducted into, and organized in terms of, a single terminological system, that which is appropriate to a general behavioural model capable of representing the total ecosystem as a single integrated four-dimensional whole.

Now, how does the use of such a model permit one to interpret and predict systemic changes? The answer is by the use of that method normally referred to as systems-analysis. This means that a systemic change is reflected in the value of one or more of the variables. Our knowledge of the inter-relationship between these and the other variables of the model then enables us to calculate the total change brought about to the model and hence to predict those likely to occur in the system as a whole. For this to be possible, our model and hence the variables it uses must have certain features in common.

The first of these is that it must be functional. Its role is to permit the prediction of changes relevant to the functioning of the system so that the appropriate adaptive responses may be mediated. In the same way, the variables chosen must also be functional: their role is to contribute in a precise way towards enabling the model which they constitute to fulfil its function.

The second feature of our model is that it must display order. As we have seen, behaviour cannot occur without information and if the latter does not display order, neither will the associated responses. Let us look a little more closely into this notion of order. Order is normally equated with limitation of choice. An organism or a society can be regarded as stable if it constitutes a self-regulating unit of adaptive behaviour; and if it is capable of responding adaptively to any predictable environmental challenge.

This implies that choice is limited, i.e. if in given conditions one knows what the correct response should be one can predict its occurrence with at least statistical precision. For an organism to be capable of behaving in this way, a given environmental signal has to be detected, trans-ducted and organized in its brain in one way and one way only—that which will give rise to the correct response. If this occurs its model of the system can be said to display order. In exactly the same sense of the term, an organization of scientific information must also display order. This implies that any modification of the value of a variable must affect that of the others in but a single way.

It is evident that for this to be possible the variables must be precise. If they are vague and are used in a different way by different people at different times, these conditions would not be satisfied.

Our model must have yet another feature: its variables must be closely inter-related. In order to establish the effect of a modification in the value of a variable on that of the others, the relationship between them must be clearly established. This inter-relationship can be examined under different headings. I am particularly concerned in this paper with that obtaining between general and particular behavioural processes. Behavioural processes whether they be phylogenetic, ontogenetic or mediated by the brain, appear to involve the gradual differentiation of functions previously fulfilled in a more general way. They can thus be regarded as proceeding from the general to the particular, from which it must follow that their various stages must be organized in hierarchical fashion. If our model is to be capable of representing such a process it must also be so organized. Thus particular processes will be related to general ones in that they are derived from them by the process of differentiation.

This brings us to yet another necessary feature of our model: its relevance to the sub-system. I think this is best formulated by stating that its sensitivity to systemic changes must be directly proportionate to their importance. Another way of stating this is that the model must have the highest possible information value vis-a-vis the sub-system. Yet another way of stating this is that they must be as far as possible ‘intelic’, i.e. relevant to the integral.

What is true of the model vis-a-vis the sub-system is also true of the variables vis-a-vis the model. Indeed, each variable must be so designed that a modification in its value must have the maximum effect on that of the others. If it does not, it can be considered as irrelevant to the model as a whole. Since the model is organized hierarchically, the more general the traits represented the greater must be the number of more particular traits affected by it—and hence the more relevant must be the model to the system. It must follow that the more general the variables used the greater the extent of the system in space and in time on which one can predict the effects of a given systemic change.

On the other hand, the greater the number of particularities represented, the greater the improbability of the systemic changes that our model will be capable of predicting.

Now there is very little point in trying to predict improbable situations if one cannot first of all predict probable ones correctly, and since in accordance with the law of economy a model must make use of the minimum number of variables for it to accomplish a particular function, it must follow that these must be as general as possible.

To sum up: for systems analysis to be applicable to the total ecosystem, a general behavioural model is required whose variables must be functional, precise, closely inter-related and to the extent that they represent behavioural processes of traits, hierarchically organized and as general as possible.2

The use of the term “Religion”

Having now seen what must be the principle features of the variables of a general behavioural model, let us see to what extent the concept-word ‘religion’ qualifies.

I think it is true that most of the concept-words used to refer to cultural phenomena such as mind, learning, knowledge, democracy, fascism, etc., were not designed specifically as the variables of a scientific model, but were rather acquired by us as part of our cultural heritage and as such can better be regarded as more appropriate to a much simpler, vaguer and more subjective model: that corresponding to the world view of our particular culture.

It is the thesis of this paper that religion also falls within this category.

Religion, however we wish to define it, is clearly some aspect of a cultural pattern. Let us therefore first see what we mean by this latter concept.

A society can be regarded as a unit for certain types of behaviour mediated at this level just as can a family, an individual, a cell, and a molecule for other types of behaviour. What is more, as I have already stated, their respective behaviour patterns can all be represented by the same model using the same set of variables. That this is not obvious at first sight is probably due to the fact that we are used to thinking in terms of complex and disordered societies such as our own. However, in certain conditions, a simple society can display a very high degree of order—probably as much as can a non-human animal society. An Australian aboriginal tribe, for instance, is quite capable of adapting to predictable environmental changes without the aid of any coercive institutions of any kind. It constitutes a self-regulating unit of behaviour in the same sense that a biological organism does. It is quite possible to examine those cultural traits that permit this ordered behaviour and then to follow what happens to them as the society becomes more complex, begins to disintegrate, and as coercive institutions are introduced in an attempt to compensate somewhat ham-fistedly for the lost social order.

It is suggested that the term ‘culture’ be used to refer to a social behaviour pattern. This almost coincides with its current use in anthropological literature, though this concept-word is not usually formulated in such functional terms.

The role of a culture is therefore to ensure the maintenance of the society’s ordered structure and by the same token its adaptation to its changing environment. At the same time, like all behaviour patterns, it must be divisible functionally into an organization of information, or set of beliefs, that constitutes a model of the sub-system and a corresponding pattern of responses. These must all be hierarchically organized just as they are in the case of all other behavioural patterns.

Different cultural patterns will then be classified in accordance with the different ways they achieve this end; and different cultural traits in accordance with the specific function they fulfil within the cultural pattern of which they are part.

The question is—can this be said of the term ‘religion’?

Does it constitute a specific type of cultural pattern that differs in any functional way from other types?

Alternatively does it constitute a particular cultural trait that fulfils a specific function in a given cultural pattern?

I think that it does neither. We refer to something as a religion simply because it appears to us like a religion—i.e. because it has certain features, sometimes mainly of a terminological nature, that we associate consciously and unconsciously with religious things. It is thus a subjective and not a functional classification. A consequence of this is that we find the term sometimes applied to cultural patterns that have functionally very little in common, for instance, as we shall see, that of a simple ordered society and that of a complex disordered one. At the same time it is often not applied to cultural patterns fulfilling the same functions.

Take the case of early Christianity and Marxism. They undoubtedly have much in common, as is pointed out in an article in the Journal for the Scientific Study of Religion by Mary Barbara Zeldin;3 so much so that a classification which includes one and not the other of these functionally similar movements is of little use in a general behavioural model.

Now, if the term ‘religion’ does not appear to be used functionally when applied to the culture of complex and disordered societies, this is even more the case when it is used to refer to that of a simple ordered one. To illustrate this, let us see what are the functions vis-a-vis the individuals and the societies themselves of various cultural traits which in a simple ordered society are normally classified as ‘religious’.

Let us consider, for instance, those associated with ancestor worship, initiation rites and dynamism or mana. Ancestor worship undoubtedly has a stabilizing effect on the members of a society by providing them with the consoling thought that their relatives, rather than having been annihilated at death, have simply graduated to a new and more prestigious age-grade, thereby remaining in some way members of their family, clan and tribe. It also provides them with a psychologically necessary goal-structure, for by giving the ancestral spirits their due in the form of sacrifices and by carefully observing the intricacies of tribal traditions, they will obtain their cooperation in warding off diseases and other disasters. At the same time and by the same token, this institution does much to ensure the continuity of the society, effectively complementing the prestige of the council of elders and the power of public opinion in maintaining the set of cultural values on which the very survival of the society as an ordered unit of behaviour depends.

Initiation rites appear to provide a traumatic experience to mark the series of abrupt changes in psychological and social status occurring at birth, puberty and the onset of middle age. Circumcision, for instance, occurs shortly after birth among the Jews, at puberty among many East African tribes, and at forty among the Gallas of Ethiopia.

Such a traumatic experience undoubtedly favours the abandonment of the behaviour pattern associated with the initiate’s status and the development of a new one appropriate to that which he now acquires. Seen from the point of view of the society taken four-dimensionally, it contributes towards the temporal differentiation of its parts.

Dynamism fulfils an important role in the establishment of a society’s goal structure. Many of the rituals and ceremonies performed by its members appear to be interpretable as a means of increasing its stock of mana, while the role of many of its taboos is to prevent behaviour that may lead to a reduction in its stock of this essential substance. In addition, the associated ceremonies undoubtedly have powerful integrative functions and both they and the taboos confer on the society an identity of its own by distinguishing it from its various neighbours and hence contributing to the maintenance of its cohesion and hence of its ordered-structure.

What is to be gained by stating that these traits fall under the heading of religion? What would be the information value of such a statement? Undoubtedly very low, for it does not provide any information on how these traits contribute towards the achievement of the social goals. It is my thesis that it merely states that they look ‘religious’. They affect us in such a way that we are led mainly unconsciously to classify them in this manner.

The subjective nature of this classification is further emphasized by the fact that it appears to be peculiar to the members of complex and disordered societies.

Indeed, as far as I can gather, and I have checked this with members of quite a number of simple societies, the latter’s vocabulary does not normally contain a word corresponding to our concept of religion. This would lead one to suppose that the concept itself is alien to such societies, probably because they have no use for it, i.e. because there is no advantage to be gained by classifying any of their cultural activities in this manner.

If religion is a subjective rather than a functional variable, by the same token it is not a precise one. If different types of societies tend to have different notions of religion, the same is true of individuals within a given society. Indeed ‘religion’ is invariably defined by writers on the subject in different and mutually incompatible ways, and one must conclude that if people cannot agree as to how the term is to be used it cannot possibly constitute a precise variable of a general behavioural model.

This brings us to the question of hierarchy. If the term religion is to be classified hierarchically, it must be possible to relate its use in the context of a complex society to its use in that of a simple one, in such a way that the former be a differentiation of the latter. However, we have seen that the term has very low information value in the latter case. Let us look a little more closely into the reason for this.

The cultural pattern of a simple ordered society appears to form an integral whole, as does that of any ordered unit of behaviour. All writers appear to agree on this point; they only tend to differ as to whether or not this culture is to be labelled ‘religious’.

Fustel de Coulanges4 in his famous study of the Ancient City recognises the unity of its cultural pattern. He wrote:

The state and religion were blended together so completely that it was impossible, not only to imagine a conflict between them, but even to distinguish one from the other.”

(“Cet État et cette religion etaient si complètement confondus ensemble qu’il etait impossible, non seulement d’avoir l’idée d’un conflit entre eux, mais même de les distinguer l’un de l’autre.”)

It was the religious element that he considered dominant.

Edwin Smith,5 writing on the simple ordered tribal societies of Africa, equally accepts the notion of their integrated cultural pattern; but, on the other hand, prefers not to apply the term ‘religion’ to them. He considers the African’s attitude towards the world of spirits, for instance, as being “a purely secular one”, and the application of words such as ‘worship’, ‘sacrifice’, ‘prayer’, which have a highly specialized significance in English, to the African’s ancestral system as being—to quote Driberg—“both a linguistic and a cultural offence”.

Whether we wish to apply the term religion or not to the cultural pattern of a simple ordered society appears to be mainly of terminological interest. What is important is that the cultural pattern does not appear to be split up into two parts, the religious and the non-religious, as it is in our type of society, but rather constitutes a single integrated whole.

Now, if the term religion has higher information value when applied to the culture of a complex and disordered society, I think it is because in such a case there is such a cultural division, though its exact nature is not in fact indicated by the use of this term.

A possible explanation for this division is that as a society disintegrates, its individual members seek to create substitutes for cultural traits that have now disappeared and which previously provided them with specific psychological satisfactions.

In this way, a new set of cultural traits is built up which is not an integral part of the culture as a whole, in that by providing psychological satisfactions for the individual members of the society it is not by the same token contributing towards its ordered structure, as does ancestor-worship for instance. I use the term ‘heterotelic’ to refer to these substitute cultural traits in opposition to the term ‘homeotelic’.6 In this way a cultural pattern breaks up into what can be regarded functionally as two separate sets of cultural traits, and it is possible that the term ‘religion’ in a complex and disordered society could be used to refer to the ‘heterotelic’ set. This would provide us with a useful variable for our model. It must be noted that we would not be using the term exactly in accordance with its current usage—which is usually inevitable if we made use of functional variables as opposed to ‘subjective’ ones.

This brings us to the notion of importance or generality. Our model, as we have seen, must make use of variables that have maximum importance and hence generality. We have already seen that the generalities of a behaviour pattern develop before its particularities. It is therefore by examining a simple society that fulfils its various adaptive functions in a general way that we can hope to understand the more differentiated functions that appear at the level of the complex society, and it is the variables referring to these general functions that must constitute the basic variables of our model.

However, when we apply this term to the cultural pattern of a simple ordered society, we appear to be doing exactly the opposite. We are making use of a term that may be useful in describing the particularities (whether heterotelic or not) of that behavioural process that is the development of human cultural forms, for the purpose of classifying its homeotelic generalities. What we in fact require is the very opposite—a variable in terms of which we can classify the generalities of this behavioural process and which we can then gradually differentiate so that it can also apply to its developing particularities.

Back to top


To conclude, the concept-word religion is not a functional classification of cultures or cultural traits. It is rather a subjective concept-word telling us how we are, mainly unconsciously, affected by it. It is not a precise concept-word, but is rather used in a different way by different societies and different people within these societies. It cannot be related to the other classifications in terms of which we would describe cultural behaviour; in particular it cannot be organized hierarchically. In addition, if one were to regard the development of human cultural behaviour as a single long-term process, it would be more applicable to its particularities and possibly its more heterotelic ones than to its homeotelic generalities.

For these various reasons, the term ‘religion’, at least as it is used today, would not be a satisfactory variable for a general behavioural model.

Back to top


1. Kennith Craik, The Nature of Explanation, Cambridge Univ. Press, 1952.

2. Edward Goldsmith, The Theory of a Unified Science, Ch. 10, “The Concept of Information”.

3. Mary Barbara Zeldin, “The Religious Nature of Russian Marxism”. Journal for the Scientific Study of Religion, 1969, VIII (1), p. 100.

4. Fustel de Coulanges, La Cité Antique, Libraire Hachette, Paris, 1927, p. 194.

5. Edwin W. Smith, African Ideas of God, Edinburgh House Press, London, 1950, pp. 25, 26.

6. Edward Goldsmith, The Theory of a Unified Science, Ch. 9, “The Concept of Order”.

  • Twitter
  • Facebook
  • Digg
  • Reddit
  • StumbleUpon
  • Diaspora
  • Identi.ca
  • email
  • Add to favorites
Back to top