- home - publications - documents to download - architecture - contact


The Haveli - a social history

Sarah Glynn 2001

The final version of this paper was published in Architecture of Rajasthan, Edited by Giles Tillotson, Bombay: Marg, pp 102-111 (2001 Reissued 2004)   

If you look down from the air, or from flying a kite off a high roof, a north Indian town presents a cell-like appearance.  Some cells are large, others small, and they are clustered together sometimes regularly, and sometimes with what seems complete randomness.  Between run the arteries of the town, branching ever smaller to penetrate among the clusters.  These cells are the haveli courtyards, some spacious - perhaps with gardens - and grouped to form palatial mansions, others crammed into their plot so that their central space becomes little more than a light well and ventilation shaft.  The arteries are the roads and galis that give access to the havelis and break down the structure of the town into its different quarters and localities.  As the arteries get smaller, so the spaces they enter become less public, intended only for local users.  The smaller arteries, like the havelis themselves, may even be protected by a gate.

            Courtyard houses, in which verandas mediate between internal and external spaces, are found wherever the weather is warm, from Italy, through the Middle East, to China.  Enclosed fortified houses, protected by guarded gates, are found wherever there is risk of physical attack.  The havelis are the particular variety of fortified courtyard house which developed to suit the social and cultural conditions of rich, or at least well to do, families in what is now North India and Pakistan. The origin of the word is uncertain, its dictionary definitions various, but in the old towns of South Asia it generally suggests the inward looking courtyard houses of which they are so largely composed.

            Few existing havelis can predate the sixteenth century, but the form of those that survive is probably based on that of earlier buildings.  Fundamentally, they have changed little over the centuries.  The first havelis evolved in a feudal type of society, where wealth emanated from the courts, and security of house and city was paramount.  Later builders emulated earlier forms - thus peaceful nineteenth-century merchants built great gateways, not so much for security, but to ape the prestigious symbols of the old aristocracy.  The British Raj introduced both a new peace, and new house types to India.  However, while many decorative details, and even new materials such as cast iron, were adopted from the British, the overall house form remained remarkably constant, until the major changes in life style which occurred this century.  Many British-period houses were still havelis, lived in in a manner which had evolved only slowly since Mughal times, even if the setting for that life now included classical columns.

            This article looks at the material conditions in which these houses were created, and particularly at those topics which must always affect those who design houses: the client, the site, the available materials, the workings of the design and construction industry, and the brief - that is what the house was required to do.  But I do not argue for a strict determinism.  Every designer or group of designers and craftsmen adds their own cultural memories, prejudices and pure inspiration.

            Finally, having examined the havelis’ creation, I want to look at the factors which are destroying them, and at how this precipitate destruction might be stopped, at least in part.

The Client
The haveli household was a complex social organisation, made up of the owner’s extended family, numerous servants and sometimes also slaves.  The household head was generally a nobleman or lesser landowner, or a successful merchant.  His wife presided  over the day to day running of the house, as well as over the activities of all the other women.  She supervised the servants, regulated the store rooms, organised the preparations for festivals and celebrations, and often kept the keys of the safe.  But if she was widowed, except in a few cases when the oldest son was still very young, she had to relinquish all her power.  Polygamy, except when the first wife proved barren, was generally restricted to the very rich.

            Household size varied enormously, but a large haveli might hold as many as two hundred family members, including poor relations.  Uncles and cousins set up their own homes only when conditions became too overcrowded.  It was thought desirable to have very large numbers of servants, both to give prestige, and also to allow each servant to be assigned limited well-defined duties.  While small, poorer households  had to make do with one or two servants, in a rich haveli they could considerably outnumber the family, and included such people as lady companions and story tellers, whose role was to banish boredom and the trials of loneliness for the ladies of the zenana.  Servants usually lived in, except for those considered outcasts, such as sweepers and dhobis.

            Until the mid-nineteenth  century, and sometimes after that, both male and female slaves were kept for household duties and personal care, and the grandest households might include concubines and eunuchs to guard the women.  Rama Mehta’s semi-autobiographical novel of 1950s Udaipur describes how the senior maidservant had been given to the haveli as a child, after her desperate father had already watched three of his children die in a famine.[1]  And a Delhi friend recalled to me how his grandparents in Tonk bought 6 or 7 children from distressed families during the famine of 1919.[2]  A famine sale was, perhaps, the most common route into slavery, other than being born a slave’s child.  Generally, social pressures encouraged (though could not guarantee) good care of domestic slaves, even in old age, as they were kept for prestige rather than economic use.

The Site
Havelis are generally town houses, clustered into groups defined by social ties of family, caste and occupation, and the available land is limited.  In a limited area, a courtyard form gives the greatest amount of useable external space (a courtyard is much more useful than a thin strip of land surrounding a villa), as well as providing privacy, security and ventilation.  External courtyard walls can butt up close to each other. 

            Some site restrictions are very specific.  For example, in the planned city of Jaipur most houses are not allowed to have their principal access on the main streets, where the ground floor is almost entirely taken up with shops, and it is only possible to squeeze in a subsidiary access stair.  And plots are often long and narrow so that as many buildings as possible can have a frontage on those main streets.  Narrow gandi galis - dirty alleys - run between the jaipur havelis.  Here the haveli residents threw out their rubbish, and narrow stairways gave the sweepers access to come and clean the haveli latrines.

The materials used depended not only on regional availability, but also on the expected permanence of the construction.  Kaccha, or mud-built, houses were common in Mughal cities, due both to the peripatetic nature of the Mughal court, and to the unreliability of its inheritance laws, which discouraged investment in more permanent structures. But in more feudal Rajasthan, havelis were of stone, or, where the available stone was poor, rubble and render.

            Construction materials affected the design not only in its details but more fundamentally.  For example, where stone was prevalent the designers used grids of columns to keep most spans small enough to be roofed with stone slabs.  Less easily available and perishable wood was required for only a few major spaces.

The Construction Process       
Traditionally, havelis were constructed by master  masons without recourse to drawings.  Gordon Sanderson, who investigated surviving architectural traditions in 1913, was told that the designers of Bikaner’s grand new havelis did not believe in plans.  In Delhi, he found that the master masons made rough plans of room arrangements, and sometimes drew the front elevation, but only drew decorative details for teaching novices.  They were paid a commission on the cost of the materials, and themselves employed the masons who worked under them, visiting the different sites in their charge for a few hours each day.[3] 

            Others in this volume have written on the spiritual element of haveli planning with far more authority than I can, so I will simply add two comments.  The first was made by Sunand Prasad, who argues that the traditional architectural treatises allow for so many possible arrangements ‘that it is apparent that almost anything can be done and that is the point.  There are in the end few prescriptions and practicality is never sacrificed to geometrical purity.’ [4] The second is that many of the patterns created by spiritual diagrams, such as mandalas , are patterns of universal appeal, found in Western architecture as much as in Indian. Perhaps they reflect patterns familiar to us in the natural world.  Except where we can do a serious analysis and have more than circumstantial evidence, we must be wary of giving the mandala primacy.  A mandala is sometimes used to justify and explain universally pleasing forms, as well as to create them; and this post-rationalisation  may be practised not only by the superficial observer, but also, possibly even subconsciously, by the designer himself.

            For aesthetic and formal inspiration designers looked up to the Mughal and Rajput courts, or down to the folk tradition, with its painted and mirror-work decorations.  The British period brought a new palette of forms, to be mixed, hybrid fashion, with the old, or used in a more brazenly Western manner, as individual taste dictated. 

The Brief
There were basic requirements common to every haveli.  Security was achieved through the use of the courtyard, with its guarded gates, and, on a wider scale, by the pattern of town planning.  Grouped havelis could form a secure unit; membership  was controlled, and there were obligations as well as benefits.

            Courtyard planning also helped to keep the interior private, but the need to separate men and women required more complicated arrangements.  Where possible, women were kept to their own inner part of the house, the zenana.  This might be a rear courtyard, or an upper floor.  High walls, with pierced screens (jalis), for air and limited views, helped maintain privacy, along with the careful use of bamboo blinds (chiks) and curtains (pardas).

            Havelis were designed so as to temper the worst excesses of the hot climate, with a range of spaces, from external courts (chowks) and terraces to verandas to inner rooms, for use at different times of day and in different seasons.  Through breezes were encouraged, and often the only doors were on store rooms.  Chiks could be used to provide shade whilst maintaining air flows, and pardas could be hung over openings if more warmth was required.  Some grander Mughal houses had cool underground rooms, but less extravagant cooling methods included screens of khas grass soaked in water, hand held fans, and awnings.  Jharokhas, bay windows bracketed out from the walls and often filled with jalis, helped those sitting behind to catch any possible breeze, as well as to glimpse the world beyond.  Bracketed access balconies and overhanging upper stories not only maximised  space, but provided welcome shading.

            Besides these practical considerations, the havelis played an important role as status symbols, so size and richness of decoration were highly valued.

            Most of the rooms in the haveli were required to be flexible in their use, with soft furnishings and charpois being moved around as season and mood dictated.  So in summer, for example, beds would often be taken out at night to the courtyards and terraces.  There were, however, certain specific requirements to be accommodated.

            Every well appointed haveli had a grand reception room for receiving formal visitors and for festival gatherings or family celebrations.  Some had a further reception room in the zenana, and some had screened viewing windows to allow the women to watch certain of the men’s festivities.  Smaller reception rooms were required for business offices, and raised platforms (chabutras) served as the inevitable waiting room.

            In a large haveli a separate chowk or ground floor might be set aside for servants’ quarters and stabling, but often servants were expected to find themselves a space to lie down as best they could, after the master’s family had retired to rest; a personal servant might sleep on the floor beside her mistress’s bed.  It was considered important to move around in style, and the stables had to accommodate horses, bullocks and elephants, while European landaus and phaetons, and later cars, joined the more traditional palanquins, bullock carts and tangas.  At the other end of the scale, small havelis simply hired transport when it was needed.

            Small enclosed rooms were built as kitchens, but they had no chimneys - only grilles in the wall - and were very hot and smoky.  However, much of the preparation and the washing up could be done in the veranda or outside in the chowk.  The joint family all ate together, though the women did not usually start until the men had taken their fill.  The men sometimes ate in the zenana, and sometimes separately, for example if they had guests.  Servants often cooked and ate in their own quarters.

            Hindu havelis included a puja room for the family gods; and a tulsi bush in the courtyard, holy to Vishnu, formed the object of many of the women’s devotions.

            For festivals and family celebrations the havelis had to be able to accommodate large numbers of guests for several days and nights.  Stone rings, built into the walls, provided anchorage for the awnings which were stretched across courts and terraces, and wooden platforms covered in matting were used to extend the floor areas.

Fifty Years of Decay
The times of such festivities seem long past, though they are still well within living memory.  The decline of the havelis has mirrored the decline of the society which created them, and the seeds of social change were planted well before Independence, though the last fifty years has seen them come to maturity.

            The post-Independence constitution brought an end to court culture, and the new political climate encouraged a measure of land reform.  Although from the point of view of the peasantry this was very far from complete, it cut deeply into the incomes of the land-owning families.  Members of the old aristocratic families, who had formerly relied on the income from their land-holdings and the perks of their position in the court, had to search elsewhere for jobs.  Others were attracted away by the new possibilities opening up for well-connected Indian men.  These new work patterns helped break up the joint family system and destroy the dominance of the older generation, and thus the havelis’ in-built resistance to change.

            The spread of modern education and Western views was aided by the radio, which brought new ideas into the depth of the house.  This helped break down the purdah system, which has vanished from most of India with remarkably little comment. For some women, the defining moment came with their decision to take an active part in the freedom struggle.

            The havelis became associated with backwardness and the stifling restrictions of tradition.  They failed to conform to western ideas of privacy, and the possibilities of electric cooling made their open planning redundant.  Cars got stuck in the narrow alleyways.  And as the wealthy chose to move to modern suburbs the town centres entered a spiral of decline.

            Some towns were deserted at partition, or abandoned as the court culture vanished and sources of employment dried up, but in many places, as the haveli owners moved out others moved in, and the havelis succumbed to the twin insatiable pressures of a growing population and commercialization.

            In a typical story of haveli decline, the fuse is lit when the joint family splits up and divides the haveli between them. Parts of the family set up commercial businesses, bringing strangers into the once-private core.  Sections may be rented out or sold to people with little interest in the haveli.  With no overall maintenance, communal areas quickly decay, and there is little incentive for the maintenance of individual parts; and with no overall control, there is nothing to stop piecemeal and inappropriate alterations or encroachments on communal spaces.  Eventually the haveli becomes just another part of the bazar.  Others become godowns or slum housing, or are simply knocked down and redeveloped.

            The unconsidered appreciation of some of the havelis’ decorative details has added a further threat, as houses are robbed of their doors and shutters, leaving them more vulnerable to the weather and to further, less systematic, vandalism.

The Future
There are isolated examples of preservation - a cultural centre in Udaipur and a few hotels - but there is a limit to the number of cultural centres and hotels a town might benefit from, and conversion to satisfy modern hotel standards can be very destructive.

            In looking for new uses it is vital to maintain the essential characteristics of the building.  These may be summarised as its architectural form and decoration;  the open space of the chowk , and its peace which contrasts with the bazar outside; and, finally, the various types of space (terraces, verandas, inner rooms) in which to enjoy cooler periods and endure hotter ones with only minimal electrical help.
            These characteristics can only be maintained if each haveli is kept under a single control mechanism which can look after the shared areas, carry out general repairs, and regulate structural and decorative alterations.  But this need not mean wholesale, top down restoration and gentrification,which,  even if architecturally successful, brings its own problems.  Most of these buildings are already homes to numerous families, who often cannot afford to move out of the city centres.  In many cases, a more suitable model might be that of the co-operative housing society.  If those who live in the buildings are personally involved in their preservation, and decisions which effect the building are taken jointly, then, not only will existing communities not be broken up, but  the improvements may have a greater chance of lasting.  In effect the haveli would become a smaller scale version of the controlled residential neighbourhood.

            When I express an interest in the conservation of Indian architecture, I am often asked if there are not more urgent needs in the country.  I answer with another question: are Indians not good enough to enjoy fine architecture, or is this to be another privilege reserved only for rich Western nations?  But when we fight to preserve these buildings, we must always do this in a way that both allows them to be enjoyed by as many ordinary local people as possible, and also preserves the fragile economy of the numerous poorer families who eke out an existence in the inner cities. 



Cooper, Ilay Rajasthan: The Painted Towns of Shekhawati  Ahmedabad: 1994

Jain, Kulbushan and Minakshi Indian City in the Arid West Ahmedebad: 1994

Mehta, Rama Inside the Haveli  Delhi: 1977

Pramar, V.S. Haveli: Wooden Houses and Mansions of Gujarat Ahmedabad: 1989

Prasad, Sunand The Havelis of North Indian Cities unpublished study for the Royal College of Art, London: 1987

Sanderson, Gordon Types of Modern Indian Buildings Allahabad: 1913

Tillotson, Sarah Indian Mansions: A Social History of the Haveli 1st edition 1994, Delhi: 1998

Tod, James Annals and Antiquities of Rajast’han 1st Edition 1829, London: 1914

Varma, Pavan K. and Sondeep Shankar Mansions at Dusk: The Havelis of Old Delhi Delhi: 1992

Wacziarg, Francis and Aman Nath, Rajasthan: The Painted Walls of Shekavati Delhi: 1982


1. Mehta p6-7

2. Professor Rahman interviewed by the author, December 1987

3. Sanderson p13 and p7

4. Prasad p10