Census definition of household

The household was first defined in 1872 as comprising of those who lived together and ordinarily cooked at the same hearth including their servants and visitors. In 1881 Census it was defined as comprising of all those persons who actually slept in the house or compound on the night of 17th February, 1881. From 1891 till 1941 the term 'family' was used in place of Household. From 1951 Census onward again the concept of household was used in Indian Censuses. In 1971 Census a household was defined as 'a group of persons who commonly live together and would take their meals from a common kitchen unless the exigencies of work prevented any of them from doing so'.

Distinction between Household and Family

We have seen that in the literature of the first phase there is recognition of the situation in which the members of a three or more generational genealogical unit live in two or more households but are bound together by a number of activities.  Not only have empirical investigations of the last two decades confirmed the widespread existence of this situation, a few have also endeavored to analyze it, either wholly or in part (Agarwala, 1955; I.P. Desai, 1964; Gore, 1968; Kapoor, 1965; Madan, 1965; Mayer, 1960; Singer , 1968; Srinivas, 1965).  A few have chosen to delineate only the property-holding group which is frequently composed of two or more households (Bailey, 1957, 1960; Nicholas, 1961).  A few have also given attention to relations between emigrant households and the households residing in the original village or town (Cohn, 1961; I.P. Desai, 1964; Gough, 1956; Madan, 1965; Rao,1968). 

The distinctions between household and family is fundamental to these analyses  Many investigations are concerned only with the households, but even those not specifically concerned only with the households but even those not specially concerned with them have had to accord them a crucial place in observation and analysis.  For example even though I.P. Desai’s investigations (1956, 1964) are avowedly concerned with inter-household relations, he has had to take the household as the basic unit of observation and the starting point of analysis.

The recognition of a distinction between household and family is indeed a major achievement of the studies of the Indian family during the last two decades and writings on households have been preponderant during this period.  A considerable confusion, however still persists because of the tendency of use the words “family” and household” as synonyms. 

Types of Households

Compared to the uniform view of the Indian household prevalent until about 1940, the empirical studies of the last two decades portray diversity of forms- another major achievement. However in their attempts to grapple with the diversity of forms they have produced a plethora of typologies, ranging from the simplest classification of households into nuclear and joint (or extended) to the elaborate typology involving mathematical formulate developed by Mukherjee (1959, 1962).  It is no exaggeration to state that no two scholars’ typologies agree with one another.  Sometimes the same scholar uses different typologies in different studies (Kapadia, 1956, 1959).

Multiplicity of typologies is not necessarily undesirable if the types are sufficiently and accurately defined so that one could be translated into another terms.  Such typologies are rare, however.  In some studies the types re not defined at all (Esptein, 1962). In some typologies the use of such words as uncle, aunt cousin, grandfather and grandson (Agarwala, 1955), “families” (Cohn, 1961), and “other related persons” (Conklin 1969) makes the types inaccurate.  In many the use of the term ‘generation’ is a source of much in accucracy (Cohn; 1961”; I.P. Desai 1955; 1964a; Dube, 1955; Kapadia, 1959; Kapoor, 1965; Karve, 1953; Madan, 1965; Ross, 1961; Sarma, 1951; Sashani, 1961; Srinivas, 1952a).

The dichotomy of elementary and joint (or extended) family forms the foundations of every typology.  In most typologies the joint family is considered, explicitly or implicitly, a group based on patrilineal descent and virilocal (or patrilocal) residence.  Consequently these typologies do not take care of the non-patrilineal and non- virilocal (or non-patrilocal) relatives found in many households.  Some scholars note the existence of households including such relatives, usually a married daughter and her husband and their children (the case of ghar-jamai) but they do not make the definition of the joint family consistent with these households, and they carry on the general discussion of the joint family with its conventional definitions (Cohn, 1961; Madan, 1965; Mayer, 1960; Sarma, 1964).  A few scholars have modified the definition of joint family to accommodate such relatives- they define it as composed of two or more related couples-  by they also carry on the general discussion in terms of partrilineal patrilocal’ unit.  The only scholars who give full recognition to non-patrilineal, non-patrilocal relatives I their typologies in an accurate way are Mukherjee (1962) and this associate Pakrasi (1962).  Unfortunately neither of them has gone beyond the formulation of the types to throw any further light on this important aspect of the Indian family.

After the use of term “joint family” in the sense of the patrilineal, patrilocal group, very little concern is shown for a definition of the maximum genealogical depth- the outer limit- of patrilineal descent in the formation of households.  Most of those who do show concern provide only imprecise definitions (Cohn, 1961; I.P. Desai, 1964a; Dube, 1955; Karve, 1953; Ross, 1961; Srinivas, 1952a).  A major source of this imprecision is the use of the term “generation” in defining the depth.  In some writings the joint family is not clearly distinguished from the lineage (Dube, 1955; Karve, 1953; Srinivas, 1952a).  The only scholars who define the outer limit with considerable accuracy are Madan (1965) and Mayer (1960).

Although in all general discussions the joint family is considered to be a group wider than the elementary family, we rarely find this distinction applied consistently in the actual typologies.  The joint households actually taken into consideration are usually are of three types:

1. parents and one or more married sons
2. tow or more married brothers
3. members mentioned in the type one 1 and 2 plus one or more married sons in the next generation.

The general tendency to consider only these units as joint households is so general that most scholars have had difficulty in classifying the households falling between the elementary family and the “real” joint families (Cohn, 1961; I.P. Desai, 1964a; Dube, 1955; Epstein, 1962; Karve, 1953; Sarma, 1951; Sashani, 1961; Srinivas, 1952b).  And sometimes some or all of them are placed in a new pigeon-hole (Cocklin, 1968; Gough, 1956; Kolendra, 1968; Morison, 1959; Ross, 1961; Sahani, 1961).  Whatever their fate in classification, the general discussion of the Indian household is carried on in terms of two ideal types, the elementary and the “real” joint family.

Only a few works give attention to the distinction between households composed of complete elementary family and those composed of the various types of incomplete elementary family.  Cohn distinguishes three sub-types of nuclear family: (1) single aged adult, (2) wife and unmarried offspring and (3) husband and unmarried offspring (1961, p. 1052).  Conklin has a type called “sub-nuclear family” for households of one member and of one parent and children, but husband wife households are included in “nuclear family” (1968).  Kolenda’s “sub-nuclear” type includes only households of one parent and children and adds a separate type for single-member households, but husband-wife households are included in “nuclear family” (1968, p. 346).  Kulkarni separates uni-member households from all other households (1960).  Nicholas distinguishes two types called “irregular family” and “nuclear family”; the former includes “only the widow, or the widow and her children, or occasionally also a man who resides with her”.  Nicholas does not specify the relationship of the ma with the widow nor does he tell us why he considers such families “irregular” (1961 P.1057).  Mayer includes the complete elementary family and the family of a widower and his children among what he considers to be important types of households and refers casually to households composed of a widow (1960, pp. 179-180).  Madan mentions “non-familial households” of a man and his adopted son or a widow and her son, or a childless couple (1962a, p. 89).  I.P. Desai’s typology in his Mahuva book includes the types of “uni-member” and “nuclear family” as including “one or both parents with children’ (1961).

Each of these writings refers at best to only a few of the possible types of the incomplete elementary family, and none tries to assess their significance.  Mukherjee (1962) and Parkasi (1962) are the only scholars who recognize almost all of the various types of the incomplete elementary family, and among them only Mukherjee tries to assess their significance.  All the scholars however, tend to neglect the households composed of the I incomplete elementary family in the general discussion the Indian household.  The households of some or all of the various types of the incomplete elementary family are regarded as “irregular” (Nicholas 1961) “unimportant” (Mayer, 1960), “non-familial” (Madan, 1962a; Mukherjee, 1962; Parkasi, 1962a; Kulkarni, 1960), or “families only through courtesy” (I.P. Desai, 1964a).  This is an unrealistic attitude.

Recently, Kolenda has introduced another complication.  She defines the nuclear family as “coupe with or without unmarried children”.  That is to say a husband-wife unit is also a nuclear family.  And this is done deliberately.  She states, “We should use the same definition for the nuclear family in India as in the United States, if we are ever to compare Indian family structure with American family structure” (1968, p. 346).  This means she abandons the standard anthropological definition in favour of a definition of just one culture.  This is not likely to facilitate scientific discussion.

While the elementary family, of whatever definition is considered a characteristic of Western culture, no attention whatever is given to a crucial fact of Indian culture that an elementary family may be genealogically complete but socially incomplete.

Thus while a great deal of effort has gone into the classification of household compositions, the typologies used in the classifications suffer from a number of deficiencies.

The Development Cycle of Households

The essence of the idea, which a household of one type develops in course of time into one or more households of another type, is present in many works and many have also pleaded for studying the Indian Family in the framework of the developmental cycle and employing the idea to gain a fuller understanding of the Indian Family. Cohn (1961) rightly considers the type of family existing at a certain time not a fixed entity but rather a structural contingent resulting from the operation of a combination of several factors, such as life expectancy, economy, interpersonal relations in the family and migration.  His description of the developmental process however is sketchy.  Onklin (1968, 1969) cogently emphasizes the importance of the role of the cycle of births, marriages, and deaths In the formation of households, but does not describe the formation process at any length and completely neglects the role of partitions in the process.  Ross states that “the typical middle and upper middle class urban Hindu moves through a series of family forms, but does not use the idea for any other purpose.  Epstein (1962) views the cycle as having only two phases, the elementary family and the joint family.  Although I.P. Desai (1964a), Gore and Sarma formulate a number of types of households, they freeze them into two types, elementary and joint while describing the developmental cycle.  Gould’s description of the developmental cycle of the north Indian Family (1968) includes only one cycle comprising three types of households, namely the nuclear family, which develops into the family of the father and his two or more married sons, which after the father’s death spits into as many nuclear families as there are brothers.  In effect this cycle is composed of the dichotomy of the nuclear and the “real” joint family.

Madan is the only work (1965) that describes the developmental cycle in greater detail and also uses it to gain a deeper understanding of other aspects of the households and family.  He describes three courses of development: one arising out of the absence of a son in a nuclear family, the second arising out of the presence of only one son, and the third arising out of the presence of two or more sons.  All other courses of development are considered “unusual” and are therefore neglected.

Recently Kolenda has rightly questioned the idea of a single family cycle (1968, p. 347).  She has also indicated certain factors making for a multiplicity of cycles, most of which are strikingly the same as those which I have discussed in my analysis of the developmental process of households in Radhvanaj; the time of a man’s death in relationship to the ages and marriages of his children, the number of years between the latter in their ages, variation in obligation of children to parents and to each other depending upon birth-order and sex, and a number of norms-for example those concerning the proper time for married sons to break away from  parents, or married brothers to break away from each other.  Unfortunately Kolenda has not yet applied these ideas to an analysis of the developmental cycle in a field situation.