PROVENANCE:  information origins, sources & destinations Canada and the Web

ISSN 1203 - 8954

Integrating Western and Traditional information processing:
Observations from Papua New Guinea

Stuart Hawthorne and John Evans

Western information systems implemented for the use of indigenous PNG users experience a significant decline in efficiency when transferred to local control. This observation will come as no surprise to people who have been familiar with the day-to-day workings of government and private organisations in Papua New Guinea over the last 30 years. In fact, though somewhat provocative, we would risk venturing the speculation that today there is not one imported Western information system in the whole country that is maintained and run by indigenous users at its designed operational efficiency level.

This is not a situation that PNG should welcome, even if this claim was not completely true. Information management arrangements like records systems in private companies and government departments, classified catalogues and access systems in libraries, legal and transaction records, organised files of financial and commercial papers, and so on, are essential to the effective maintenance of good governance, social order, education and commercial enterprise. As well, efficiently kept records are central to the maintenance of an acceptable commercial environment involving overseas trading partners. We feel that if PNG is ever to improve upon the poor economic and social conditions under which it currently labours, an improvement in information management is essential.

We do not see this problem as necessarily being the ’fault’ of anyone; rather it is a manifestation (though, in PNG, a badly managed manifestation) of a problem that has arisen elsewhere. The literature covering the implemention of Western designed information systems in developing countries frequently attests to the difficulty of matching the world view of the local community with the way knowledge is represented in the system. This difficulty arises because the local perspective is, or is historically derived from, a community-centered orally-based method of information sharing and use. It is holistic and essentially deductive. This contrasts with the inductive, reductionist nature of Western information systems. This fundamental difference between conceptual perspectives is why we see in practice a persistent divide between the imported service model and the kind of service that would really meet the needs of the people.

One of the difficulties in fixing this problem is that to date all system design has been based on the Western model, principally by Western practitioners who have not yet found their way across the divide. This is meant as a statement of fact and not meant critically. As many Westerners have found, we ourselves included, successful integration of the two views is a very difficult task. The main reason for this one-sided input is that in traditional society, the information ’system’ was (and is) not a separate conceptual entity like it is in the West but was distributed as part of the social fabric of the community. Traditionally, information management and use was based more on an ethos than a methodological construction. Because there was no ’system’ as such, no tradition of analytical dissection of system elements and of system design ever developed. In this regard too, as alluded to by Kinasa (1997), it is reasonable to expect that, since information sharing and use in traditional society was so closely related to survival of the group, once something was found to work, there would have been a strong reluctance to change it. To do so could have proved hazardous to group security. In our view, this legacy from historical times persists today in the apparent general disinterest of indigenous PNG people in systems efficiency and in the dearth of indigenous system designers.

If the burden of work is to currently fall upon Western practitioners, then another major difficulty is that Western system designers simply do not seem to know enough about the issues involved. While there is voluminous comment in the literature, much of this consists of descriptive reports or covers case-specific implementations. Invariably, these are tightly bound to the cultural aspects of the particular community in question and it is difficult to extract or identify useful parameters that can be applied in other cases where a different culture exists. We have a suspicion that there are general culture-independent principles that apply to all community-based (historically oral) information systems employing distributed storage. Until these are identified, we do not believe that Western designers have a sound enough theoretical base from which to proceed. What is needed is a rigorous and consistent foundation for developing effective integrated systems under any culture employing distributed information storage and use. Our collaborative team is seeking to identify the universal principles that make up this consistent foundation. Our research perspective is that, to large extent, it is not so much ’cultural’ differences that impact upon integration of perspectives as it is the architectural and structural differences between systems, which in turn impact upon the functional behaviour of information users

The purpose of this short paper is to record a number of general observations about Western and Melanesian information processing that reflect this functional, as opposed to cultural, research perspective. It is our hope that feedback and comment on these observations will lead to new insights and be able to be used to more tightly focus our research efforts.

The remainder of the paper consists of two parts, each part containing four observations. In the first part, ’Information Issues’, we compare different aspects of the holistic perspective of traditional society with the Western reductionist view and note the differences that arise in the manipulation and perception of ’information’. In the second part, ’System Issues’, we note some of the design issues we think mitigate against successful transfer of Western systems to the traditional domain.

Information issues

Information significance  Under the reductionist viewpoint, cognisance of information takes place in terms of its techno-economic utility. Under the communal approach, cognisance of information takes place in terms of its position in a universal spatio-temporal domain of interest to that community. The difference this makes to what is considered significant can be appreciated using an analogy drawn from art. Imagine an artist’s model standing with one arm akimbo. Artists assign discrete recognition to the space that lies between the bent arm and the body. Even though this space is ’empty’, this separate identification affirms that the space is part of the total picture, and has an equal status with other ’solid’ parts of the view. In this, artists are in accordance with the holistic perspective. By contrast, under the reductionist view, empty space (that is, ’nothing’) has no techno-economic value in its own right. If significant, it is significant only in terms of the adjacent ’solid’ elements that it comes between. In short, in the West, detail is acknowledged immediately only if it has potential use in a bigger picture later. Under the holistic perspective, every detail in the big picture is acknowledged now because individual details may be useful in later particular contexts. Clearly, in our view, this difference impacts upon what is selected for inclusion in the datastore, whether this datastore be an electronic one or a mental one.

Information transferance

 Historically, under the orally-based communal approach, a centralised controlling system with a single interface is an alien concept. Rather, information is distributed amongst the members of the community so the process of having access to this information (that is, retrieval) is also distributed, with the product of retrieval being a distillation of the concatenation of individual contributions. Since what the individual can ever come to know privately in total can only be an a posteriori copy of what the group has come to know collectively in public, the essence of information transfer is principally deductive. In direct contrast, the Western approach is mainly that the group learns from the individual. For example, the overriding criterion for evaluating the worth of any particular piece of research in the West is that it must ’add to the body of knowledge’. Thus, the essence is inductive.

Information storage

 Under the distributed communal approach, and allowing for social rank, it is not too far wrong to say that all members of the community group know a little about everything. While perhaps some particular specialisations might exist (for example, knowledge about battle tactics might be retained mainly or only by men), generally there is no strict categorisation of information storage. If there is any specialisation, then usually it is a smaller group who specialises, not an individual. The traditional approach to storage is multi-disciplinary and the material that is stored is qualitative, being stored in the minds of all of the individuals that make up the group. Because of the mental basis of this storage and because the human intellect operates on the relationships between concepts, the interrelations between elements of information are as significant as the information itself in understanding what is known. It is for this reason, and because of its distributed storage, that information is more a social construction than a statement of independent fact. In a way, traditional people do not deal with their mental contents as ’information’ in the same way as Westerners do. Rather, it remains much closer to being ’knowledge’ since the relevant set of concepts is removed from the knowledge store with all of its relationships intact. Thus, it tends to remain as subjective knowledge rather than become objective information. This leads to a fluidity of meaning and an intensional aboutness that does not have the same assumed permanency that it does in the West. This contrasts with the Western mono-disciplinary method which requires ’objective’, quantitative information to be stored in mutually exclusive categories. In the West, the emphasis is on stating information as bald facts divorced from as many of their (deemed irrelevant) external connections as possible. The establishment (and indeed, the re-establishment) of the relations between these facts and other facts is seen more as part of the process of using that information, rather than of the information itself.

A consequence of this tendency to perceive and use information in the same form that it is stored is that spatio-temporal relations that are used to identify information at a given time in a sense always continue to be part of that information. This is because they have become part of the public holistic background (or, if you like, the reality that is the group universe or cosmos) against which information is stored and retrieved by individuals. When brought up against the reductionism of the West, this way of looking at things causes difficulties in such ubiquitous areas as, say, period contracts, where different obligations of the parties may apply at different times. These difficulties spill over into the efficient use of records systems where, for example, the necessity to employ a time-based filter on a retrieved set of contract records may not be realised. In the West, manufactured realities such as period contracts are exploited so often that they are not consciously seen as artifices. For a person holding the traditional perspective however, these spatio-temporal manipulations have the effect of distorting the known reality. This is because under the traditional view, the spatio-temporal domain characterises, or is, the external world and the historically derived expectation is that this changes from without, not from within.

Information relevance

 One of the necessary conditions for successful articulation of knowledge through group cooperation has to be that each individual contribution to the central pool of information is given equal consideration by the group [1] If this condition is not satisfied, the willingness of individuals to cooperate at optimum level may be affected, to the detriment of the group. The ’equal consideration’ condition implies an assignment of responsibilities. The individual is responsible for putting forward all information that has any possible connection with the topic and which may be potentially useful. The group is responsible for determining the worth, and therefore relevance, of individual contributions in regard to the group consensus towards which it is working. The important point is that the individual clearly understands that it is not his or her responsibility to determine relevance. To do so would be to usurp the group’s eminence. This is in direct contrast to the Western ’stick to the point’ approach in which an individual is expected and encouraged to provide only that information which (he or she thinks) is relevant. In the West, the individual is expected to presume relevance; in the traditional community, the individual is expected not to assume relevance.

These separate approaches are no more than is to be expected from the deductive/inductive differential but they frequently produce consequences which adherents to each viewpoint find baffling. For example, onlookers might have been requested to prepare an eyewitness report of a car accident. For a typical Westerner, the fact that a passenger in the car wore a green shirt would not seem to be relevant to the request and therefore may not be mentioned. But a person with a holistic perception might indeed mention this fact, as well as the fact that he has a shirt just like it, that his sister bought it for him the last time she was in town, and that his friend, who comes from that town, had a similar car accident three years ago. The traditional group expects all these details to be passed to it even though they may well be discarded later in reaching the group’s understanding of what happened. Initially, however, all details have to be put forward by the individual. A Westerner receiving such an account initially thinks that everything that is reported is relevant information. Realisation that some information is not relevant becomes frustrating because of the Western view that it is the responsibility of the person giving the account to filter out irrelevant details. The Westerner thinks the traditional person has failed to stick to the point. On the other side, there is frustration with the Westerner (in this case being seen as occupying the central group role) because of the apparent failure, after having faithfully been given a surfeit of facts, to discard details that are not required for a final understanding of what occurred[2]. Each side thinks the other side is shirking its responsibilities. For any communication between the respective sides, either the Westerner thinks relevance determination has already taken place but it has not, or the traditional person thinks relevance determination has not occurred but it has[3].

The processing and structural issues surrounding relevance determination constitute one of the major problem areas in the reductionist/holistic dichotomy and are near fatal to operating Western information systems efficiently in developing countries. The Western information-keeping rationale is based on classification coordinate points that imply thematic and paratactic closure. The assumption of the exclusiveness of this closure is central to how Western classification works. Western classification schemes are simply not geared to operate efficiently on information that admits non-exclusiveness. In theory, it is possible to generate classification indicators that encompass disparate non-exclusive information. In practice however, it is extremely difficult to do this without inadvertently letting in other unintended references and without compromising control of the tactic scope of the classification coordinate. In such cases, the coordinate ceases to be thematic and becomes perspective, resulting in non-exclusive coordinates (that is, their referring scope overlaps or is overlapped by the scope of other coordinates). The gross decline in system efficiency that this leads to is unacceptable under contemporary Western practice.

System issues

In our view, many Western information systems implement philosophical and methodological approaches, either in the design of the system or the classification scheme, which introduce inefficiencies. Our intention here is not to discuss whether these approaches may or may not be justified in terms of obtaining preferential advantage elsewhere in the system. Rather, we wish simply to note how some approaches, introduced ostensibly to increase efficiency, will complicate the differences between holistic and reductionist information systems.

Information systems are intellectual systems

 The provision of an information system is provision of a conduit for the transfer of intellectual perceptions. The domain of intellectual perceptions is characterised by almost infinite variability and subtlety of meaning. Given this, the probability of effecting an accurate transfer between one intellect and another would at first glance seem to be very close to zero. The difficulties can be seen even in its simplest form, as follows: I see something happen in the world. I use my intellect to interpret that event. I write down this interpretation to create a record. I give the record to you. You use your intellect to interpret the information in the record in order to reconstruct the world event. The end result is that what you understand occurred is based on your interpretation of my interpretation of the event. In other words, no matter how highly blessed we may be in our respective intellectual capacities, the best understanding you will ever have, and can have, is based on a surrogate view of a surrogate view. Contemporary approaches to system design add to this inherent imprecision by interpolating another level of complexity, in the form of the classification system (that is, the set of intellectual perceptions of an independent indexer), between the original author and the end user. The provision of an information system always complicates an already complex area in this way. If on top of all this is loaded the requirement that the end user employ an extra-cultural rationality, it is not difficult to see why the total intellectual burden placed on an end user quickly becomes unmanageable.

Problems with quantitative classification

 Almost without exception, Western information systems fall short of explicating in applied form the ideal philosophical knowledge/information model. Or to say this another way, we know what ideally is to be done but cannot seem to find the appropriate ways to interface the intangible qualitative requirements of the subjective intellect with the tangible requirements of the quantitative system. In practice the volume of material to be accommodated and the pressures of the commercial workplace have meant that systems which are easily implemented and which can produce a result quickly are favoured. For these reasons, we see the persistence of many quantitative approaches developed in the 1950s and 1960s, this being readily evidenced by the on-going introduction of ever-’improved’ metadata systems. However, as the documentalists came to realise in the 1950s, there is a fundamental problem with most quantitative methods including the metadata approach. The problem is that such systems record details about a record, not of it, (that is, perspective description rather than thematic information) and therefore add to the problem of imprecision, as is readily demonstrated by any search across the Internet. Again, this is something designers of integrated systems need to remain aware of since there is already a major task in maintaining precision in the transition from ’part’ description’ to ’whole’ description. Many classification approaches in Western systems will simply exacerbate this problem.

Type of record held

  The fact that the bulk of the records of libraries, government departments, the legal community and commercial enterprises are expository documents is usually not taken into account at system design time. (The documents we are mainly concerned with are qualitative texts that contain their meaning within continuous prose, not tabular or field-specific quantitative material such as financial records). An important characteristic of expository texts is that they contain or should contain explicitly all the propositions needed to pass the message of the text. For example, an expository text dealing with the maintenance of ejection seat assemblies in jet fighters has to overtly say everything that needs to be said. All interpretation has to done for the reader by the author. There should be no possible inferred proposition that can add to the meaning of the message of the text, otherwise (in the case of ejection seat maintenance, for example) the results could be disastrous. By contrast, a narrative story nearly always contains implicit propositions that are intended to be inferred and which become part of the meaning of the total text as the reader understands it. For example, a reader is subtly encouraged to infer a tacit proposition that sets out what the hero and the heroine did when they lived happily ever after at the end of the story. In an important sense therefore, an expository text is ’complete’ in that it includes or should include only and all pertinent aspects of the world that it requires. Given this, it can be seen that not to use the contents of the text as a basis for classification of expository texts but instead use external quantitative indicators such as metadata classifiers is to nominally add to the meaning of that text. For holistic records, this approach is going to further complicate their classification since they already have a tendency to provide too much non-exclusive information.

Maintenance of context

 A detailed appreciation of the role context plays in giving meaning to information is often not evident in information systems. While it is true that some systems seek to account for immediate context, the importance of maintaining global context is usually subordinated to the quantitative demands of the system. In many contemporary systems based on a computerised windows user interface, it is relatively easy to come across instances where a record or part of a record is taken out of evident context, even if only temporarily, so that it can be attended to in some way. This is not to suggest that this is necessarily inappropriate or that it is necessarily fatal; nevertheless, in larger or smaller measure, context is not maintained. The ’lost in hyperspace’ problem of hypertext information systems is probably the best-known outcome of not maintaining context. The maintenance of context is important in any information system but it becomes of central importance to the present discussion. This is because one of the main structural differences between the reductionist focus of the Western techno-economic perspective and the holistic focus of the traditional perspective is one of contextual scope. For the latter, maintenance of the wider context is fundamental to how the events of the world captured in a representational medium are perceived and understood. System designers will do no favours for traditional users if they transplant systems in which the maintenance of contextual integrity is hard enough to deal with under Western rationality, let alone the traditional one.


Kanasa, B. (1997). Collecting Local History: Taxonomy of the Zia Knowledge System, (accessed 1 March 2002)

Richards, J. C. & Schmidt, R. W. (1983). Conversational analysis. In J. C. Richards & R. W. Schmidt (Eds.), Language and communication (pp. 117–154). London: Longman.

Shannon, C.E. & Weaver, W. (1949). The mathematical theory of communication. Urbana: University of Illinois Press.


Stuart Hawthorne is a records management consultant in Brisbane, Australia and has a background in information science and philosophy. He lived in Papua New Guinea for 28 years.

John Evans
Email:  Dr. John Evans


© Copyright N. S. Hawthorne and John Evans 2002

[1] This conclusion is arrived at quantitatively through analysis of system requirements and takes no account of any behavioural understanding that may be known in this area. In regard to this however, it is useful to point out that one can readily see this condition articulated in practice in the courtesies of the consensus-oriented ’Pacific way’ meeting strategy.

[2] If we analyse situations like this in terms of the rules of conversational implicature (see Richards & Schmidt, 1983, pp. 119-124) and Grice’s maxims of cooperative behaviour (p. 120), a good case can be made that there is no real transgression of the rules by either side. In regard to the maxim of quantity, though there is a different understanding of what the ’sufficient’ amount of information to be passed and received actually entails, each side nevertheless, is faithfully satisfying the set of rules they believe is governing the exchange. Each side is genuinely seeking to cooperate with the other. The problem is that each side assumes the same set of rules is being used by everyone.

[3] It is instructive to note that we in the West succeed in baffling ourselves, let alone others, in our use of induction. For example, no one has yet satisfactorily explained how we can determine a priori the relevance or otherwise of experimental variables without at least partial presumption of the very theory we are seeking to establish. Indeed, this is what occurs when we exclude mention of the green shirt from our accident report.

© authors
Originally published online at
Re-published by Provenance the web magazine 2002

building information structures & collections

· Provenance the web publication home page
· Back issues
· Development projects
· Reference sources
· Media resources
· Credits

Contact us by email

· Front page Provenance
· Back issues
· Development projects
· Reference sources
· Media resources
· Credits