It will, I think, be evident that functioning within the systems of kin terms and naming there is another system: a system of sex differentiation. In kinship terminology, there is a consistent distinction between male and female kin. The only exceptions to this are the terms for younger siblings, spouse’s younger siblings and parents-in-law. Nevertheless, even these terms may be distinguished accordingly by the use of the sex categories phau’khwa and phau’my as modifiers. Similarly, in the naming system, the distinction is established through the use of sex markers prefixed to autonyms. In the case of teknonyms, the sex distinction is made in the use of the kin terms pa, mo, phy and phi.

The entire corpus of linguistic categories which make distinctions between sexes, in Palokhi, is represented in these two systems. However, some categories which are used in these systems in the class “humans” are also used to make sex distinctions in the class “animals”. These categories cannot, therefore, be examined independently because some of them are common to both classes.

If the corpus of terms which establish sex differences in their various usages is examined in its own right as a system, some intriguing features become apparent.

To begin with, there are only three basic contrastive pairs which express sex differences in the class “humans” and the class “animals”. They are:

The elements in these pairs may well be regarded as fundamental semantic units since they cannot be reduced further to any other constituent units of meaning. Furthermore, in their respective classes, they can only be defined semantically by their relationship to each other. As we have seen, phau’khwa and phau’my are general sex categories, while pa and mo are kin terms. The other kin terms phy and phi, which are also sex differentiated, cannot be regarded as fundamental terms because they can be represented in Karen by other forms such as “father’s father”, “father’s mother” and so on. The term mo, on the other hand, is the same for “female animal” and “human mother”. Similarly, the term pha, “male animal” and the male sex marker for autonyms is the same. They are, very clearly, polysemous terms.

Another important feature is the asymmetry between the categories “unmarried male” and “unmarried female” (phau’khwa/phau’my), and the sex markers (pha/nau) for the autonyms of males and females who are unmarried, and who are married but have no children. The term phau’khwa describes both the general sex category “male human” as well as “unmarried male human”. The general sex category, “female human”, on the other hand is phau’my while the category “unmarried female human” is different, that is, mykoe’nau. The term mykoe’nau, therefore, is a marked term.

It may be noted that the term mykoe’nau has the additional meaning of “pubertal” applied specifically to female humans (see also Jones [1961:123]) and that there is no equivalent term for males. Furthermore, although the term is partly derived from the general sex category (or, more accurately, its minimum free form —my—), nau itself presents every semblance of being a female marker of its own, serving the same functions as my. It occurs, for example, in the kin term waenau (elder female sibling). As we have seen, the complementary term for waenau is waecau’; unlike nau, however, cau’ does not appear elsewhere outside of kinship terminology.^[7] Nau, therefore, is clearly also a marked term. Accordingly, the term mykoe’nau thus possesses a high degree of redundancy, as it were, in terms of markedness.

The asymmetry in the sex markers for autonyms (pha/nau) lies in the fact that pha also describes a sex category in a different class (namely, “animals”) while nau is a marked term which is to be found only in the class “humans”.

Another noteworthy feature in the system of sex categories is the distinction that is made between being unmarried and being married in the class “humans”. This is to be seen in the terms for “unmarried human male”, “unmarried human female” (phau’khwa/phau’my) and “married human male”, “married human female” (phau’khwapgha/phau’mypgha). If we also consider the case of autonyms and teknonyms, it will be seen that this distinction is taken further. Autonyms and their sex markers are used only for males and females who are either unmarried or who are married but have no children. Teknonyms, on the other hand, indicate not only the sex of the person concerned but also the fact that they are married and have children.

Other than the major contrast animal/human which forms the two classes within which these sex terms are found, all these terms are organised according to a single natural dimension of contrast, namely, male/female. The symmetries and asymmetries between the various terms which correspond to this contrast suggest, however, that there is some other ordering characteristic or dimension of semantic contrast at work in the class “humans”. This can be no other than the distinction between “unmarried” and “married” which appears consistently in the class. More precisely, the distinction is that between (umarried) + (married-without-children)/(married-with-children) + (married-with-grandchildren) as the set of sex markers for autonyms and teknomyms indicate.

While this distinction is appropriate to the class “humans” and, indeed, essential to indicate the kinds of discriminations made in this class, it is clearly inappropriate to the class “animals”. Yet, as I have noted above, there are some terms (mo and pha) which may be found in both classes. The unmarried/married distinction therefore introduces a somewhat false dichotomy because it restricts further comparisons and contrasts between the categories in both classes which is analytically necessary, given the occurrence of mo and pha in both classes. A more general contrast applicable to both classes which allows the distribution of these two terms to be represented, and which can also subsume the unmarried/married distinction, is thus required. The only logical contrast which permits this is the contrast non-procreative/procreative or non-reproductive/reproductive. Such a distinction, I suggest, is the other dimension of semantic contrast operating alongside the animal/human and male/female distinctions which together account for the particular configuration of sex categories and their uses in Palokhi. For present purposes, I shall use the terms “non-procreative” and “procreative” to express this distinction.

Table 3.2 shows the various sex terms organised according to the three dimensions of semantic contrast which underlie the various uses of these terms in Palokhi. The symbol (+) indicates marked categories, while pha and mo as applied to animals make no distinction between those which have offspring and those which do not.

When the terms are ordered in this way, it becomes immediately apparent that they do not simply constitute a system of sex and gender differentiation. They also form a system of social classification for humans based on a logic of sexual difference, generational difference and the difference between states of “non-procreativity” and “procreativity”.^[8] This classification is significant for what it establishes simultaneously. First, it isolates “non-procreative” female humans from all other categories, particularly the categories which share the feature “procreative” across classes. Second, it establishes, as it were, that the property “procreative” is common to animals, male humans (unmarried or married) and female humans with children, through the distribution of the polysemous terms pha and mo.

In the class “humans”, the nature of the relationship between the set of marked terms (that is, non-procreative female) and all other terms is the most crucial aspect of the classificatory scheme. The categories phy and phi, as I have mentioned before, are not basic contrastive pairs according to the semantic differentiations which underlie this scheme; they may be regarded as extensions of the terms “father” and “mother” according to the logic of generational difference and they are, therefore, not a crucial feature of this scheme. Any significance they possess in the sub-class “procreative” derives from the term pa and mo.

The relationship between the set of unmarked terms and the terms for non-procreative female humans are represented in Figure 3.1 which summarises the essential features of Table 3.2.

The bi-directional arrowed lines stand for “equivalent” relationships. The uni-directional arrowed lines stand for “non-equivalent” relationships and a substitution of terms in the direction of the lines according to sociolinguistic usage following the transition of individuals from “non-procreative” to “procreative” statuses.

These relationships lie at the core of the system of classification. In terms of the pragmatics of the uses of these terms, what is important is the regular substitution of terms as individuals go through their life cycles which recognises not only their sex at all times, but their change in “states of being”, that is, from “non-procreative” to “procreative”.

There is, however, a further significance to this cluster of terms. Much of the difficulty in apprehending it lies in the translation labels which may be used to represent the Karen terms. It will be more readily apparent if the Karen terms are rendered as follows:

Although somewhat clumsy, these labels serve to indicate the essential meanings and distinctions which are made by the Karen terms in their various uses. Other than the distinction between “non-procreativity” and “procreativity” and the singling out of non-procreative human females, it will be apparent that embedded in this group of terms is a distinction between the nature of reproduction in the two classes “animals” and “humans”. Where animals are concerned, there is a general recognition of the procreative role of females (or “motherhood”) but not that of males. In the case of humans, however, the roles of both males and females are indeed recognised. As I show in my discussion of the domestic ritual ‘au’ ma xae (see Appendix A), the Palokhi Karen do indeed say that it is impossible to establish the “paternity” of domesticated animals since any number of males could copulate with a single female to produce offspring. This is not to say, therefore, that the Palokhi Karen are unaware that males and females (whether animal or human) are both equally necessary in sexual reproduction. They are quite aware of this biological fact. The point I wish to stress is that this group of terms, as cultural categories, carries with it cultural definitions of the nature of animal and human reproductive processes. Central to these definitions is “paternity” and “maternity” in human reproduction, and “non-paternity” and “maternity” in animal reproduction.

To return to the system of sex differentiation: it will be noticed that while the system is based on the logic of the differences discussed earlier, there is nevertheless an apparently arbitrary distribution of terms as well as markedness and non-markedness across classes.

This is a good illustration of the kinds of problems not uncommonly encountered in componential analysis and the referential approach to the study of semantics which Fox (1975:118–9) has noted. The reason, as Fox also points out, is that “… it is not taxonomic generality but polysemy — the property of a symbol to relate to a multiple range of other symbols — that becomes the criterion for hierarchical inclusion”. This is precisely the case with the Palokhi classificatory system. Where Palokhi sex categories are concerned, polysemy or “interlinkage” to use Frake’s term (1969, cited in Fox [1975:119]) includes relationships of homonymy, antinomy and shared semantic fields. It is this which gives the system the appearance of having an arbitrary distribution of terms and markedness.

The apparently arbitrary distribution of terms across and within classes, therefore, is in fact simply the “surface” manifestation of an embedded system of classification that rests on the semantic associations produced by a remarkable parsimony in what can only be called a selective distribution of lexical items. The system is, in other words, “motivated”: it is not just a linguistic artefact; it is sociologically meaningful.

As empirical systems, kinship terminology, naming systems and systems of sex (and gender) differentiation may be — and indeed often are — treated as linguistic systems in their own right as I have done here to some extent. The principal value of investigating these systems as such rests in the contributions which may be made toward an understanding of the human universals which underlie their formulation. My purpose in examining these systems in Palokhi, however, has been rather different, namely, to arrive at what is sociologically meaningful about these systems in the particular context of Palokhi Karen society. In this restricted context, it is not the features of these systems taken individually which enable us to determine what is sociologically meaningful but, rather, those which emerge from their interdigitation. The classification of males and females in terms of “non-procreative” and “procreative” states of being is one such feature. The distinctiveness of “non-procreative” female humans, through redundant marking, is the other.

When, however, such emergent motivated features exhibit a demonstrable relation with features of the same order in a different, non-linguistic system, it can only be concluded that these features are not only sociologically meaningful but culturally significant as well. In Palokhi, such a relation of homology does indeed exist between the classification of males and females and a system of dress and colour symbolism.