0. 1 Purpose 

The scope of this essay is ambitious ; its results are modest [1]. It represents an attempt to delineate a framework for synchronic and diachronic analyses of French kinship. The conclusion should be taken as a working hypothesis to be tested further and revised or discarded. Ten centuries of semantic and social history are here reviewed concisely in order to stress some exploratory guidelines. My endeavor resembles that of March Bloch in Les Caractères originaux de l'histoire rurale française. Like French rural history, French kinship deserves a tentative synthetic treatment despite the unavoidable shortcomings inherent in such an enterprise. 

The following excerpts from Bloch's preface define a strategy relevant to the present purpose. The reader should peruse them carefully and bear them in mind as methodological limitations fundamental to this essay. 

Un historien averti des difficultés de son métier ne se décide pas sans hésitations à retracer en quelques centaines de pages une évolution extrêmement longue, en elle-même obscure et, par surcroît, insuffisamment connue... 

Dans le développement d'une discipline, il est des moments où une synthèse, fût-elle en apparence prématurée, rend plus de services que beaucoup de travaux d'analyse, où, en d'autres termes, il importe surtout de bien énoncer les questions, plutôt, pour l'instant, que de chercher à les résoudre... Le jour où des études plus approfondies auront rendu mon essai tout à fait caduc, si je puis croire qu'en opposant à la vérité historique des conjectures fausses je l'ai aidée à prendre conscience d'elle-même, je m'estimerai pleinement payé de mes peines. 

Seuls les travaux qui se bornent, prudemment, à un cadre topographique restreint peuvent fournir aux solutions définitives les données de fait nécessaires. Mais ils ne sont guère capables de poser les grands problèmes. Il faut, pour cela, des perspectives plus vastes, où les reliefs fondamentaux ne risquent point de se perdre dans la masse confuse des menus accidents...

L'histoire est avant tout, la science d'un changement. Dans l'examen des divers problèmes, j'ai fait de mon mieux pour ne jamais perdre de vue cette vérité. Cependant il m'est arrivé, notamment à propos des régimes d'exploitation, de devoir éclairer un passé très lointain à la lueur de temps beaucoup plus proches de nous. « Pour connaître le présent », disait naguère Durkheim, en tête d'un cours sur la famille, « il faut d'abord s'en détourner ». D'accord. 

Mais il est des cas aussi où, pour interpréter le passé, c'est vers le présent, ou, du moins, vers un passé tout voisin du présent qu'il sied, d'abord, de regarder (Bloch [1960] vii-ix). 

The exploratory framework is substantiated as firmly as possible. However, gaps remain ; they will be bridged with suitable generalizations. Rough as it is, this preliminary mapping will prove fruitful if it can generate better analyses. 

0.2 General Remarks on the Study of Kinship Terminologies 

A kinship terminology is a sociological and semantic field, i.e., an institution. It is indeed a verbalized system of relationships recognized as such, sanctioned, and persisting from generation to generation as a “social fact”. Aside from the obvious point that the components of such relationship systems are people linked to each other through variations on biological themes, not much more can be said of kinship that would hold true across societies. This is enough to distinguish kinship from other social systems, e.g., from relationships of production, political systems, etc., but further specifications on this general plane are well-nigh impossible (Maybury-Lewis [1965a], 254-255, 259 ; cf. Leach [1961] Ch. 4). 

Morgan's basic dichotomy (1871), however, may serve to establish two broad classes of kinship systems : that between (mainly) “descriptive” or genealogical structures versus (mainly) “classificatory” or categorical ones. Further, it might be argued that genealogically-structured kinship systems are “complex structures”, whereas categorically-structured ones are “elementary structures” as contradistinguished from the former by Levi-Strauss (1949). In effect, because of their relatively closed character — they constitute a taxonomy of kin types — genealogically-based systems are incompatible with prescriptive alliance ; in contrast, categorically-based systems —which constitute a taxonomy of positions (cf. Needham [1966]) — are relatively open and thus perfectly compatible with prescriptive alliance : in fact, they need an extrinsic sociological closure of one type or another (Hocart [1937] ; Leach [1958] ; Maybury-Lewis [1965b]). 

Maine's, Morgan's and Rivers' influences have led quite a few anthropologists into the ethnocentric pitfall of treating all kinship systems as if they were genealogically-based. Lévi-Strauss, Dumont, and Needham, among others, have shown that there is more than genealogies to kinship. This is not the place to discuss Lounsbury's attempt to justify the general validity of the new phase of the genealogical approach labelled componential analysis (Lounsbury [1965]; cf. Leach [1958] ; Schneider [1965] ; Levy [1965]). But the problem is related to a broader issue which must be mentioned in this context of semantic analysis. 

Like kinship analysts, semologists are in conflict on the definition of units of signification. Their respective concepts of “dictionary” epitomize the debate (Quine [1953] especially Ch. 3; Katz and Fodor [1963]; Greimas [1964, 1966]; Jakobson [1966]). Thus, some contend that “words” cannot be used operationally as semantic units and that a broader segment of discourse (proposition) is needed, and their argument runs somewhat like that of category-analysts in kinship studies. In conformity with the traditional vein of European philology, others hold that words are indeed proper semantic units and must be handled as such (cf. Darmesteter, [1898]); in anthropological linguistics, lexicostatistics rests on similar grounds. But then, more refined treatments of “words” have been developed which, in point of fact, already resorted to the analytic device which was used afterwards by componential analysts of kinship, namely the concept of “distinctive features” (Jakobson ; Levi-Strauss [1945] ; Hjemslev [1953] ; Greimas [1964] ; see below, 3.2.2). Both types of analysis in semantics and in logic will doubtless continue to coexist for a long time: they work on different levels and their respective objectives are slightly unkeyed (cf. Maranda [1966a, 1966b]). In the analysis of kinship, this distinction may not be at issue: genealogical and categorical systems do not seem to be mutually reducible, as they do not overlap functionally. The consequence is that the content of the former will not be amenable to analyses other than genealogical, and that the content of the latter will be distorted if approached componentially. 

It would be highly conjectural to suggest that genealogical systems evolve as a rule from previous categorical stages, although such seems to have been the case for Indo-European (Hocart [1928 ; 1953] 178-182), and although the passage from subsistence to cash economies might be influential in such transformations. Actually, genealogical increase through time means that kinship loses ground as framework for social interaction, this function being gradually assumed by other specialized institutions. The concomitant semantic process would be that terms are then extended from a nucleus instead of being restricted by specifiers, which is criterial in the distinction of genealogical from categorical systems (Hocart 1953: 178-182). 

Be it as it may, French presents some residual traces of categorical dimensions (residual because their gradual disappearance is historically attested: below, Chapters 2 and 3). The system is definitely genealogical on the whole, though, and must be approached accordingly. — It is significant, in this respect, that the modern anthropological meaning of généalogie was already attested in the thirteenth century.

So far, most semantic analyses of kinship systems have been received skeptically on sociological grounds. A common reply is that semantic analysis is justified in its own right. But then, it is counterargued, semantics has no bearing on the understanding of the totality of social facts and remains a rather sterile exercise. Lounsbury (1965) tried to show the sociological relevance of his approach : he does not provide, however, an explanatory model as powerful, as Leach's categorical one (Leach [1958]) whose challenge he set out to accept. 

This study is both semantic and sociological. The hypotheses formulated on the basis of lexical investigations are correlated to social history. In conclusion, it is pointed out that the structure of French society and the structure of French kinship terminology covaried. Therefore, if pure semantic analysis yields some information, its validity remains very limited. 

0.3 Abstract 

The five chapters form three parts: Part One (Chapters One and Two) is essentially historical and sociological; Part Two (Chapter Three) introduces a formal model for the analysis of the kinship terminology and puts it to use; finally, Part Three (Chapters Four and Five) returns to historical and sociological domains to test the conclusions of the formal analysis. The Conclusion (Chapter Six) summarizes and evaluates the attempt. 

Two broad sociological dimensions are particularly emphasized: (1) social organization as geared to the relationships of production, in which kinship was embedded in traditional French society, and (2) verbalized relationship systems, both in their descriptive aspect, i.e., terminology, and their normative aspects, the exploratory one as in oral literature, and the jural one, as in law. The point is, therefore, to investigate the native definition of kinship-in-context in French society (Part 1); to describe it as accurately as possible with the help of a model which allows for the measurement of states (Part II); to investigate the dynamism underlying the institutionalization of the many faces of kinship (as related to marriage, succession rights, authority, exploitation of land, etc.) as this process is socially verbalized and sanctioned (Part III); and, finally, to sketch a structural definition of French kinship in a perspective taking into account the diachronic and synchronic interplay of the factors involved in its evolution (Conclusion). 

Sociological data and verbalized systems thus form the two main dimensions juxtaposed here. How French society formulates and evaluates kinship relationships is compared to the latter's social and historical positions.

The political system of medieval France, like that of most other European countries at the time, was strongly hierarchic. Pyramidal social organization was the rule, enacted in vassalage. In contrast, on the level of the household (or atomistic unit of exploitation), lateral solidarity prevailed (the freresche). Thus, a hierarchy of lords lived out of their vassal's work, whose kinship structure maximized manpower and minimized the flank offered to the fisc (below, 1.3.3). 

With the advent and rise of bourgeoisie, the foundations of a more egalitarian society were laid. Concomitantly, laterality decreased while lineality was emphasized, first in town houses and then more generally. Lineages of plutocrats were gradually constituted where generation and descent began to override siblingship. Then, alliance relationships shrank too. 

On the whole, lateral solidarity prevailed in a strongly hierarchic social order until the progressive democratization of society, due chiefly to urban life, was accompanied by an increase of the lineal axis and by a simultaneous decrease of collaterality. The process is not without parallelism with what can be observed today in cases of passage from a subsistence to a cash economy. The latter is in effect usually accompanied by a growing desire to restrict one's obligations in order to consolidate one's economic position, and solidarity focuses on the nuclear family and the direct line. The hypothesis proposed can be represented roughly as in Fig. 0.1.

Fig. 0.1 French Society Before (A) and After (B) the Consolidation of Bourgeoisie. Vertical Lines stand for political organization, “X” for family structure: see text. 

The kinship terminology evolved along lines congruent with the transformation of the social order. In effect, during the two hundred years which saw the consolidation of the French bourgeoisie (eleventh-thirteenth centuries), closer collaterals were distinguished from relatives in the direct line, singled out, and segregated to marginal positions as new and exclusive names appeared for lineals. Conversely, in the centuries that followed, step relatives lost their designations and were relegated with affines. Thus, a double process worked out successively: in both cases the terminology drove away from the set of primary relationships those which began to lose privileged treatment. This was done by the coinage of new terms to count out some collaterals on the one hand, and by the withdrawal of other terms whereby step-relatives — therefore counted out of the nuclear family but distinguished from affines — were altogether lumped and counted out with the latter. Thus, as the lineal terminology became more and more refined, the affinal terminology became poorer, and if the former seems to have reached a state of stability since the end of the last century, the confusion of the latter still persists. 

Among semantic fields, kinship terminologies as have long enjoyed privileged treatment in social anthropology; similarly, the obvious relevance of law is recognized. But there are other expressions of “collective representations” which do not always receive the attention they deserve, such as proverbs and folktales. 

Folklore in general is a “social fact” for it exists in its own right as an institutionalized device to process and retrieve information encoded and stored by the collectivity. Cosmologies vary from society to society, as do political systems, kinship, etc. ; likewise, conceptions of social order and the cultural axioms on which these rest are also stated in relatively stable corpora of lore owned and managed, so to speak, by the group as such (below, 4. 1). These function as systems of quasi-norms, usually stated in terms bolder and more excessive than the true norms codified in law. They carry socially approved, socially encoded, and socially transmitted outlooks which often betray the principles at the roots of formal jural prescriptions. It is well-known that proverbs are frequently quoted in native African justice courts, and quite a few French proverbs are also legal aphorisms (below, 4.2). The “if ... then” structure implied in a great many folkloric items — “if one does this, then that will follow” — bears witness to their exploratory character as well as to their regulatory function. 

A kinship terminology, a corpus of laws, and an oral literature are intimately connected with the way in which a society sees and attempts to define itself. They are, therefore, related to social organization as “conscious models” (Boas [1911]; Levi-Strauss [1958] 308-310). The analyst who would overlook them would seriously affect the validity of his approach (Cf. Gluckman [1965]).

The model introduced in Chapter Three has its limitations. In the first place, it will not deal with categorically-structured systems, since it is based on a count of genealogical links. Second, it can measure only closed terms (as contradistinguished from open-ended ones), i.e., it is incapable of handling the categorical terms which may accrue in an otherwise genealogically-structured system. This restriction does not apply to French, however, as far as the borders of the system are defined by civil law, but it has some bearing with respect to common usage (below, 3.4). 

My combinatorial model measures meaning by reference to contrastive features as defined by named primary relationships (alliance on the one hand, consanguinity on the other, which is subdivided into descent, generation, and siblingship). A semantic universe of kinship is generated by the combination of primary relationships into permissible strings. This is done to the seventh order, in fact, two degrees higher than would be necessary for French. The result is a matrix of possible cases to which the French encoded (actually expressed) relationships are compared. The comparison yields a specific profile of the French kinship terminology ; then, since a time dimension is available, two historical states of the terminology are contrasted and their differences are measured. 

In order to carry the measurement further, a ratio Intension/Extension is proposed. Like the combinatorial model, the ratio can apply only to genealogically-structured systems for it depends on the same prerequisites. It is most appropriate to investigate such systems, however, since it enables the quantification of the degree of ambiguity/unambiguity accruing in each term, and hence reveals built-in factors of merging and transference of meaning. On the whole, the advantages of the formalization attempted in Chapter Three are to make possible the measurement of a semantic field both synchronically and diachronically. Not only are the results obtained entirely replicable : the yardstick can also be used for quantitative comparative analyses of genealogical systems. 

If categorical systems, on the other side of the kinship dichotomy (0.2), can be related to “elementary structures”, as suggested above, this implies that they must be handled as open classes whose closure is to be looked for on a higher sociological level. Finite mathematics might be used in their formalization in a way that remains to be fully developed (cf. Weil in Levi-Strauss [1949]; Kemeny, Snell, and Thompson [1957]; White [1962], Flament [1965]; Courrège [1966]). 

The conclusion proposes an integrative view of French kinship which makes more explicit the rudimentary hypothesis set forth in Fig. 0.1. The bringing together of the diachronic and synchronic aspects with the help of a “topological formulation” is inspired by Leach's programmatic article (1961: Ch. 1). 

The fundamental variable in topology is the degree of connectedness. Any closed curve is ‘the same as’ any other, regardless of its shape ; the arc of a circle is ‘the same as’ a straight line because each is open-ended. Contrariwise, a closed curve has a greater degree of connectedness than an arc. If we apply these ideas to sociology, we cease to be interested in particular relationships and concern ourselves instead with the regularities of pattern among neighbouring relationships. In the simplest possible case if there be a relationship p which is immediately associated with another relationship q then in a topological study we shall not concern ourselves with the particular characteristics of p and q but with their mutual characteristics, i.e., with the algebraic ratio p/q. But it must be understood that the relationships and sets of relationships which are symbolized in this way cannot properly be given specific numerical values (Leach (1961] 7-8). [2] 

Throughout this essay, the terms “social structure” and “social organization” are used according to Levi-Strauss (1958) and Firth (1951). Other definitions will be found in the text (medieval institutions, “coding costs”, “stable messages”, etc.). 

For the sake of economy and clarity, geographic and dialectal areas as well as sociological substrata are not singled out. The focus of this study is the structure and development of standard French kinship as defined in Chapter Two.