Hani and Akha are considered to be of the Tibeto-Burman language family, closely related to the Yi (Lolo), also to the Lisu and Lahu in western Yunnan, Burma, and Thailand (Matisoff, 1978, 1983; Bradley, 1979). Yunnan has a larger number of Tibeto-Burman language related minority groups who in earlier centuries were characterised by the Han as 'Man' (barbarians) or Ch'iang (Sainson, 1904; Wiens, 1967). A distinction was made between 'Wu-man' (black bone barbarians - non-sinicized or wild) and 'Pai-man' (white bone barbarians - sinicized). Important amongst these are the Naxi and Bayi, who live near the city of Dali and Dali lake and were probably the dominant groups in the Nanchao Kingdom 750-1250 AD (Wiens, 1967; Backus, 1981). Hani speak dialects of a language mutually intelligible with Akha. Chinese linguists have developed a Romanized script, which is in moderate use locally, though Chinese is the language used in school.
Akha - excepting those who have become Christians - use clan names as lineage and family names. They are less often used to identify sub-groups. These are numerous and mostly are not localised, even less so since the 'drifts' of the last 50 years. Clan names are personal, genealogical, names, and not totemic or animal names as with the Lisu. Some of the older, more prestigious clan names, as we shall see later, go back 30-45 generations to a founding ancestor, others 15-25 generations or fewer as a result of fission in the clan. In some cases Akha may use what we might call 'super clan' names, referring to common ancestors of major, older clans of between 45-55 generations, in contradistinction to clans which branched off later but still in very early times. All Akha genealogical names go back to an apical ancestor and common founder of all Akha and Hani tribal groups, Sm-io. For the older clans this is 55-60 generations ago.
There are three main typologies, or criteria for typologizing,
used to distinguish major Akha sub-groups which have substantial differences
in customary law and dialect. These 
are
used rather promiscuously in the literature and in practice by various parties.
Thirdly,
and most especially by Akha themselves, genealogical ('super clan' and clan)
names may be used to indicate differences in language, customary law, language,
and (at least until recently) geographical location. Examples are the Adzjaw
Akha and the Akheu (or Akhui), the latter not even wanting to call themselves
Akha. These are distinct from descendants of Dzeugh'oe, the Mangpo Akha, and
those who claim descent from the Mazeu (or Majeu).
These
criteria for classifying sub-groups sometimes overlap in use as the consequence
of intermarriage and social and spatial drift. In Thailand until the 1950s
most Akha were Ulo Akha from he Dzjeugh'oe and Majeu sub-groups. Subsequently,
immigration of refugees from Burma lead to a considerable increase in Loimisa
Akha, mostly of the Dzjeugh'oe sub-group. Akha immigrants from China in the
same period included a considerable number of Phami Akha of the Mangpo sub-group
and a smaller number of Akheu, also known as Udzang Akha.
A survey on the Akha types of Laos is being done by the DORESAKA project.
We now focus on the hypothesis that it is the (Hani/)Akha genealogical and kinship system (guidzuieu) and related ancestor service (apoelaweu), underlying so to speak the lines of transfer of knowledge we have just reviewed, which ultimately explain the striking unity of Zangr and Hgangr between Akha groups, and which are at the core of their structural unity and survival.
In the following I examine the characteristics of what I call 'the Akha ethnic alliance system', in terms of (a) its structural characteristics, (b) its ideology as expressed in ancestor service, and (c) its status as an alternative political and economic system for groups marginalized in a 'region of refuge' at the edges of expanding state systems. It is viewed here not so much from an abstract theoretical or academic perspective, but rather in terms of its historic and potential future usefulness for the survival of the Akha and related Hani peoples.
All who have studied and compared this patrilineal descent system have been struck by its consistency. Over many years I have collected, from men of a variety of clans and lineages, their 'counting of generations', and in general they have corresponded well. They are also confirmed by genealogies collected by others (Lewis, 1969-70, Vol. 1; Kammerer, 1985; Boucherie, 1995; Davies, 1909; Tooker, 1988; Dauffe, 1906; Henry, 1903; Duy Thieu, 1996). More striking still has been the (re-)discovery by Akha, and more recently Hani, themselves that their genealogical systems fit together so well (Li Xi Xian, 1995).
The system is patrilineal in the sense that only male successors who continue to have sons are enumerated, together with their genealogical names. As is characteristic of Tibeto-Burman genealogical systems - as opposed to Chinese, Tai, and others - the last syllable of the father's name forms the first syllable of the male children's names (or in the distant past the last two syllables forming the first two).
Characteristically, Hani and Akha genealogies start from M, Heaven or 'heaven principle'. It is not counted as an ancestor, but as a kind of original principle. This is followed by M-Ma (generation 1) the 'wide, great Heaven', and M-gh'ang 'middle of heaven' (generation 2, of whom Akha stories relate that this was an Emperor). There follow eleven generations leading to Sm-io at generation twelve. Most Akha and Hani agree that the first twelve generations are 'mythical' ancestors (neither men nor spirits), and are also ancestors of peoples other than Hani and Akha. Sm-io is the 'apical ancestor' of both Akha and Hani, through whom they (and maybe related ethnic groups) distinguish themselves from others. Sm-io is followed by twelve generations of ancestors with three syllable names, like those known to us from other Tibeto-Burman groups from the Nanchao kingdom: thus Oe-toe-loe (in generation thirteen) is followed by Toe-loe-dzm (in generation fourteen) etc. (Backus, 1981: 58). They end with Dzoe-tang-pang at generation twenty-five. It seems to be some kind of ruler - whether man or woman is not clear. It also refers to a 'fortified high place' on a primordial migration route (Dzoebaw, 1982). It is not clear who these first twelve ancestors after Sm-io are. They are certainly not neh (spirits). One hypothesis is that they are Yi/Lolo ancestors, as genealogical lines of some Hani sub-groups originate there (Boucherie, 1995, Vol. 1: 33a).
Akha and Hani believe they can speak of real recognisable people and original named clans (agu/adzjeu) only after Tang Pang. Tang Pang is also the place in the genealogy from which to count in order to establish how long a particular clan has been in the system. Today, as in the past, a man marrying an Akha woman can become an Akha, if they have a son. He can then attach himself at Tang Pang as the first generation of a new clan. Further requirements are rigorous. The man has to give up his previous ethnicity, be fluent in the Akha language, accept and have knowledge of Akha customary law. He must start ancestor service and accept related obligations (Mao Youquan, 1996). It rather resembles a change of nationality status in the contemporary world.
After the Tang Pang generation the genealogical system branches into several major directions with long lines of up to forty-five generations (in which case totalling up to seventy since M and M-ma). Major branches of up to forty-five generations after Tang Pang (fifty-eight generations from Sm-io) are Tang Pang Mang, Tang Pang Ma, Tang Pang Sjang, Tang Pang Ji etc. Tang Pang Mang is seen by the older Akha clans as their link with Sm-io and Tang Pang. Other branches such as those starting from Tang Pang Siang and Tang Pang Sjang clearly lead to Hani groups. There are however also 'mixed' groups (leading to both Hani and Akha groups) derived for instance from the Tang Pang Ji branch (Boucherie, 1995, Vol. 1: 34a). As for people calling themselves Akha, those branches which lead to older clans seem to split into two major branches at Boesoeleh, eight generations after Tang Pang, and twenty-one generations after Sm-io. After this, between the twenty-sixth and twenty-ninth generations after Sm-io, most three syllable names change into two syllable names and the principal older and more prestigious names emerge. Amongst them is Mazeu, in one branch, Mangpo in another, while most others are derived from a branch starting with Dzeugh'oe which includes Dzjawbang. This period of three generations is for most members of the older Akha clans about twenty-eight to thirty generations ago, or about 700-750 years. As we discussed earlier, this was the period of Dzjawbang, the exodus of the Akha to the South-West and their increased marginalization, the unification of customary law, and the formation of alliances between clans, in which Abaw Mazeu or Majeu played a leading role (Lewis, 1969/70, Vol. 1:63).
The case of the ancestor Dzeugh'oe in this period is particularly interesting. This ancestor had so many sons, from whom so many contemporary named clans are derived, that one is tempted to make the hypothesis of an alliance of a number of clans under that name, rather similar to the case of Tang Pang (see figure XXX).
Akha history of increased contact with Mon-Khmer groups from this period onwards, finds confirmation in the genealogical system. Of some Akha clans, such as the Amaw it is still related that an older clan 'met them in the forest' but that they are Akha now. Another story tells of a first Akha man who was looking for a wife and met a forest woman, who proved to be a pehseu or 'vampire' and ate her husband; but a second man was more successful, and this lead to the institution of the custom of separation between men's and women's sleeping spaces. This seems to refer to marriages between 'more civilised Akha men' and women from 'still cannibalistic' Mon-Khmer groups (Aboe, Kh'aboe). It must be emphasised that pehseu are not spirits but people in the Akha language (cf. Kammerer, 1985: 32).
There is thus evidence that during their marginalization and encounters with Mon-Khmer-speaking groups in higher areas of southern Yunnan, older Akha clans intermarried with them. In addition some may have been incorporated as jakh'a (slave-children, servants, dependants). Two further arguments for this are: first, that there are two current forms of adoption (padaweu, to incorporate a lineage), namely (a) incorporation into, or genealogically close to, one's own lineage, and (b) linking with Tang Pang which is a more honorific form; and secondly, the existence, until quite recently, of some remnants of ranking, and a related prohibition of marriage with clans deemed to be 'too low'.
Shorter branches, of fifteen to twenty generations or fewer after Tang Pang (thirty to thirty-five from Sm-io) are thus 'late-comers'. Examples are Tang Pang Bui, Tang Pang Dzju, and Tang Pang Ka, all of about fifteen generations in depth. These are groups which might have entered the Akha system about 375 years ago, as generally one generation (from birth of father to birth of son) is reckoned as twenty-five years. These three cases claim to know that their ancestors entered the system late, and even today are aware of their previous ethnicity. The first claim that they were poor, marginalized Yunnanese Chinese, but are now Akeu-Akha. The second group claims they were Tai-related; and the third group that they were Wa people adopted as Akha.
Akha themselves are thus well aware, thanks to their genealogical system and mnemesis, that the Akha ethnic system is built up of on the one hand much older 'original' Akha groups, and on the other hand clans and lineages which have been adopted, but are fully Akha now.
We can thus hypothesise safely that several stages of the Akha genealogical system seem to coincide with Akha 'internal' historical mnemesis as expressed in oral texts. These are:
It is clear that the Akha ethnic alliance system is not a purely biological one. Nonetheless Akha do not agree with contemporary anthropological wisdom which would suggest that peoples in this region tend to change ethnic identity more easily than elsewhere (Banks, 1976).
One important conclusion to be drawn is that Hani/Akha ancestor service is the ideology and the backbone of the ethnic alliance system. Hani and Akha honour their ancestors nine to twelve times each year in ceremonies (apoelaweu) held to mark important occasions. These are performed by each family in the village in honour of its 'line of life'. In addition the Dzoema, Phima, and Badzji honour and remember their ancestors, those lines of people who handed down their knowledge and teachings through 'lines of knowledge'.
These ancestors are not gods or spirits, or devils. They are people who have died but yet can still be present in some way, and on some important occasions. It is clear from the stories about ancestors that they made mistakes, and that learning from these mistakes led, over the centuries, to the building up of Akha Zang.
This ancestor service is not religion, nor superstition. It is the means for the Akha and Hani to recognise their roots, their history, and the manner of their survival. It is also a symbol of their unity, their 'cultural citizenship' in a situation of diaspora (Li Xi Xian, 1995). It is for this reason that a rule of avoidance and non-cohabitation exists in Akha villages between Apoelaw-Akha, who maintain ancestor service and therewith also Zang, and Kali-Akha who have thrown away their ancestor-baskets, and therewith also customary law, morality, and traditional knowledge, as requested by some Christian sects. The latter are considered by those who maintain ancestor service to have destroyed their identity (Kammerer, 1990).
The question of how to maintain and adapt Hani/Akha ancestor service to new situations, such as urban life and modernised agriculture, was discussed extensively at the Second International Hani/Akha Culture Studies Conference in 1996 (Kukeawkasem, 1996).
© Leo Alting von Geusau, based on Geusau (2000)