Building is one of the most fundamental of human activities. Houses are the clearest everyday embodiment of our needs, presuppositions and ideals, a conversion of raw materials and social cooperation into a design for living. In the West, human civilisation is invariably assumed to be marked by the creation of permanent buildings and be inseparable from it. The United Nations ethnocentrically defines them as part of our universal human heritage. Case studies abound to show that in other cultures this is not necessarily so.

Since at least the 1960s, there has been a gradual resurgence of academic interest in other people's houses, broadening more recently into wider environmental concerns and the study of the constructed landscape. Archaeologists and anthropologists are both keen to read the structure of society easily and directly from its houses. Development workers want to make their dwellings more culturally bearable. Architects are picking through the ruins looking for new and marketable ideas in a world that hates architects. Optimists might see all this as a healthy dismantling of academic barriers, pessimists as unprovoked border raids by envious neighbouring disciplines that hastily cart off their booty under a hail of hostile arrows. Others again would see the whole phenomenon as another proof of that unhealthy and parochial British obsession with the house as the focus of identity. (Napoleon was wrong about those shopkeepers. The British are, above all, a nation of aspiring householders. Africans who spend their lives struggling to pay for a wife are always incredulous at the British preparedness to go through the same enduring torment for a mere house.) Whatever one's view, the spate of publications flows unabated, and there has long been a need to bring them all together in some sort of systematic overview. But what sort of overview? An encyclopaedia typically presents both a mass of factual information and classifies it in a way that not only makes it accessible but constitutes something like a current definition of that which can be known.

The Encyclopedia of the Vernacular Architecture of the World comes in three large volumes. The first is dedicated to "Theories and Principles". Naturally, it begins with a sensible discussion of the whole concept of "vernacular" architecture. Few social science books dare to offer definitions these days, realising that they are too much like Aunt Sallys attracting the academic, iconoclastic, hooligan element, but here we have one: "Vernacular architecture comprises the dwellings and all other buildings of the people. Related to their environmental contexts and available resources, they are customarily owner- or community-built, utilising traditional technologies. All forms of vernacular architecture are built to meet specific needs, accommodating the values, economies and ways of living of the cultures that produce them."

This definition is immediately striking for the way in which it reruns many of the old, unhelpful distinctions that have dogged the definition of that vexed category "tribal art" over the past 30 years. We have self/other, art/craft, individual/tribal, universal/culturally specific, aesthetic/functional and so on. In all of these, editor Paul Oliver leans towards the second of each pair, a traditional -not to say old-fashioned -"anthropological" view. Unsurprisingly, many of the entries of the later sections are by anthropologists or flensed from explicitly anthropological works.

ÁñÁ«ÊÓƵ

The first few chapters are something of a mess. Nothing is gained by the constant criss-crossing of entries under headings such as "Phenomenological", "Spatial", "Structuralist", "Meaning", "Religion and Beliefs", "Rite and Ceremonial". It is true that they have the considerable virtue of brevity, but they are very shallow and repetitive while still managing to be mutually contradictory. From the point of view of an innocent seeker after insight, they are cumulatively confusing. Contributors seem to have been handed the same reading list (notably M Eliade and M Griaule) and a very few ideas to go with them and seem to be working at such a level of abstract generality as to appear positively evasive. The main theoretical input seems to be Levi-Strauss's least interesting idea, that of "house societies". One almost expects to see archetypal interpretations of motifs and -Oh Lord, yes -there they are at the back, mercifully brief.

Normally, old hands have a firmer grasp. Oliver on "Language" and odd-jobbing on many other entries, Labelle Prussin on "Nomadism" and "Gender" suggest they are distilling a considerable reading or at least have an eye for focused coherence. Many of the other contributors in these early sections do not.

ÁñÁ«ÊÓƵ

However, matters rapidly improve with the treatments of environment and technology as these are both more extended and more discursive. Line drawings, plates, practical examples do good, factual, encyclopaedic work. We learn about the uses of dung and oyster shells, felt and blood glue, oil drums and beer bottles (but not the African use of bottle-tops for tiling). We learn of the ways in which "disaster housing" following a natural calamity often affects vernacular architecture, such as changing round houses to square to fit the corrugated iron. There is enough information on window types, squinches and pendentives, braced and balloon framing and sisal plaiting to revolutionise if not your fieldwork, at least your DIY. It is one of the strengths of the work that it declares explicitly its intention to consider houses not just as static things but as acts of construction, sites of exchange, part of developing domestic cycles, markers of ownership and encoders of cultural knowledge -that is, total social facts.

Not long ago, I spent days vainly trying to hunt down some simple data to deal with an enquiry concerning lavatory arrangements around the world. Sanitary arrangements are a prime diagnostic of many basic, cultural ideas. The group that excretes together keeps together. Under "Sanitation and Hygiene" a great deal has been piled up in one place (not, alas, the well-documented Minang fishpond privvies). Yet, curiously, when such topics are dealt with, the announced multifaceted, interdisciplinary approach immediately dissolves to be replaced by that of a simple public health inspector. Disease vectors displace ritual purity and circulation drives out exchange. The ethnographic and the architectural components tend to automatically sort themselves out again into separate sections.

The introductory volume ends with "Typologies", an attempt to classify structures by "Elevations", "Forms", "Plans", "Spatial Relationships", "Structural Types" and "Uses and Functions". Within the subdivisions different approaches again predominate. "Spatial Relationships" is prey to a very doubtful neo-evolutionism. "Uses and Functions" inevitably assumes a functional world. Having defined the vernacular in terms of the ordinary, a certain amount of work now has to be done to incorporate the display of status and rank back into the scheme of things and allow even a cursory treatment of palaces and temples. Oliver obviously worries that such a scattergun technique may not be carving nature at the joints. Because all classifications leak, the test is rather whether or not they make material easily accessible, and on the whole these do. The arbitrariness of the classification in these sections is made into a virtue by picking and mixing examples as widely as possible. So Guatemala jostles Madagascar, the United States, Nigeria, Chile and Borneo to their mutual benefit.

The remaining two volumes follow the reverse route, classification by geographical area guiding us down to concrete examples that can be explored from several different perspectives. Geography has been reprocessed, however, into something like "human geography", allowing the world to be divided into discrete cultural areas in a manner that disdains national boundaries. Fair enough: in Africa and Asia, they never made more than a limited historical and political sense. A map on the back cover handily displays and locates the various zones and subzones. The West is included, of course, suitably subdivided but typically displaced in time as a mix of historical traditions.

ÁñÁ«ÊÓƵ

This is the bulk of the book, both filling up the remaining two volumes and crying out for more space. In the interests of balance, each subzone (e.g. Guinea Coast, Madagascar and Islands) seems to have been allocated about 20 entries, which seems a little harsh on some and generous towards others. Each entry is about 100 lines and may be either an original composition or a deflated version of the standard literature on the buildings of a particular culture. (The same contributors' names constantly appear at the bottom of sections, suggesting the latter is more common than the former.) Typically, each entry consists of the location of the people, a brief sketch of their way of life and the geographical milieu. Then comes a straightforward material description of the house and the internal division of space, a dash of anthropomorphic symbolism, some status and rank, location in the wider village and remarks on change or history. Each entry comes with photographs or excellent plans/ drawings/diagrams, or both. In other words, the initial definition of vernacular architecture closely controls what is deemed relevant in the individual entries. What you seek is what you get.

Or is it? Given that a book might easily be written about every kind of house round the world, this encyclopaedia takes on an impossible task. It is hard to know for whom the present level of information would be the right one, and one of the jobs of an encyclopaedia is to get the reader beyond such a basic level to the relevant specialised literature. So each entry comes with references. Unfortunately, these are not individually very numerous. If you want to find out about the Torajan house, for example, one of the most frequent objects of reference, you might start in volume 2 under Indonesia West. That would give you three bibliographical references. A search through the "Index of Cultures, Habitats and Locations" would refer you to half a dozen other parts of the text that deal with roofs, symbolism, fertility and wood, but yield possibly only two more titles. You would still not get directly to the basic text of Hetty Nooy Palm. And even there you would not find that, for Torajan builders, the important difference between a house and a ricebarn is not just that one faces north and the other south. It is that a house may be built with banter and joking while a barn must be built with a straight face. For mice are playful animals and they would be attracted to the rice.

Given the enormous price of this work and the importance of colour, which is mentioned frequently in the entries, it is a shame to have it in black and white only. It comes with bibliographies, glossary and index. It should be seen as a major pioneering work.

Nigel Barley is assistant keeper, Museum of Mankind.

ÁñÁ«ÊÓƵ