VILLAGE,LINEAGE,AND CLAN VILLAGE,LINEAGE,AND CLAN b two or three families of a second surname are reported to have agnatic grouping in China.Let us take the case of the Tang sete remained,the rest having moved away-a single-surname settle- ments me ed briefly in the last paragraph.It happens that ment has perhaps almost emerged.The generation-depths of the they formed the subject of Potter's study in 1962-3,and we are surnames' found in these multi-surname settlements are as agnatic and follows:one of 16 generations,two of 1,four of 13,thirteen of 2,thirteen of 11,eight of to,five of9,two of8,three of7,two of It is in fact made up of a cluster of cight,not seven,villages 6,two of 5,one of 4,four of 3,and two of 2.It will be seen that there is an average of three and a half surnames per village in the multi-surname settlements,although it would of course be un- recent immigrants among them).All Tang in the eight villages realisticto countas lincages the recently form a single descent group focused upon the central ancestra hall of the complex and nominally under a single lincage head, important to note that the only very old lineage listed (of 0 who is the oldest man in the most senior surviving generation. generations)is in fact an outlier of the Punti (Cantonese-speaking) The relation at a lower level between and agnatic lage of Hang the New Territories.My second example is taken from the Ping Shan area of this plain.Thirty- village and a sub-lineage.As a village it contains (in addition to four villages are listed of which seven are multi -surname settle- the outsiders)8 Tang houscholds;but as a sub-lineage it over- flows into two neighbouring villages of the complex,where a of 20 genera- further 32 of the households belonging to the Hang Mei sub- lineage are found.Hang Mei and its co-ordinate sub-l meapes, tions,one of 28,eight of 27 (all but one of these are Tang settle- ments),one of 26,one of 25,two of 23,two of 22,one of 16, each with its territorial base,compose the Ping Shan lineage, one of I5,one of 14,one of I1,one of 9,one of 4,two of 3,two but is part of a wider corporate bra agnatic of 2,and one of I.In fact,the last six cases ought really to be removed from the list,for they are not deep enough to be lineages of Ping Shan is given in the genealog y as being one of the grand- and must be far too small to be villages. Generation-depths are county 'surname'in each of the multi-surname s founded other lineages, two of 28 generations,one of 27, one at Kam Tin and the other at Tai Po(the latter being a group one of23,two ofII,and one of I.There is an average of just under of Tang which will concern us later on).All the T ang local three surnames per village in the seven multi-surame villa ges.1 lineages tracing their descent to the first ancestor in the county These data cannot of course be fully understood unless we have a share in an estate held in that ancestor's name. Now,we may take the Ping Shan lineage to belong to a large are lineages class of such lineages in the New Territories-and,by extension, betwecn the units which segments of lineages?Clearly we in in southeastern China in general.Some of them,as earlier when we must abandon the usc of the word quotations will have indicated,are very much largern numbe the argument 'lincage'in a loose sense and try to give it a meaning precise than the Ping Shan Tang group;the figure of 1o,o00,although it enough to help in the analysis of the co mplex forms assumed by may be suspect as a good round Chinesenumber,is probably not always very misleading,and it may be recalled tha Krone gave 30,000 as the upper limit for the size of Punti villages in 19250 Hsin-an county.Each group of this class occupies a compact Potter,P'ing Shan. Cf.p.13 above,f
20 VILLAGE,LINEAGE,AND CLAN VILLAGE,LINEAGE,AND CLAN cluster of settlements which have grown up around the original central settlement.One at least of the settlements is likely to be thebat the cof these lineag walled and moated,or was so fortified in the past.(The word descen I agnatically from a common ancestor,the w wai-Mandarin:wei-still a appears in the names of many villages intun being focused on an ancestralhallorother piece ofp in the New Territories,indicating that at some time they have For this larger scale of group of which we have just been walled enclosures and may still be so today.I have calculated from the official A Gazetteer of place Names in Hong 装 and the New Territories that of some I,6o0 traditional Chinese to(tcemto methatuco names in the Colony as a whole,about io include the character and composite lineage'suggest,wrongly, for i.It is that not all forms of unit has been formed by the banding together of local the description of wai,which usually implies the existence of a In some cases,indeed,they may have come about in this brick wall and tower;some fortifications were humbler and but it is just as likely that they will have emerged by a pr er;and not all wai have that word as spart of their name.It segmentation,new local lineages being formed as par e therefore.that a yery conidrabeop development by which what was once a local lineage diffcrentiated and dispersed.) ere hudded bchind their wals n pre-itsh tm any rate,the different settlem in one compley Agnation does not stop at the higher-order lineage.As belonged to different agnatic segments of the total group,and body knows who has read a book about China,the mere there may have been a further residential discrimination within of a surname is by itself a fact of agnatic kinship.But on each settle ment such that it was divided up among segments of above the level of any unit we call a lineage we are in th lower orders.To this kind of group I propose to apply the term of clanship,where lineages of like surname may be tied t 'local lineage' But it is fact only the extreme case of the local lineage.The great majority of local lincages are much smaller and less spec- almost devoid of significance.They may,on the other tacularly sited and defended.Some share their village not only be used for genealogical reference and for forming ten with a few ousiders but aso with other local lineages What alliances.But if several local or higher -order lineages in fac defines the whole class of local lineages,great and small,is that bine and establish a common ancestral hall or estate,then they are corporate groups of agnates (minus their married sisters has once more been condensed into lineage bonds.The iving in one settlement or a tight cluster of between a system of physically dispersed segments of a settlements.It will be obvious that'local lineage'does not exhaust corporation and a network of historically-or at any rate g gically es turns upon the m bonds to people not living in their local community.First,as will be fairly a total lineage exists.Otherwise the proliferated segmen local communities;we have in such cases what may convenicntly into autonomous lineages linked only by ties of clanship be called dispersed lincages.Thosoft mmbers,formingaloca lineage,who occupy the main settlement (where the ancestral hall and its associated property are likely to be found)are,so to justified.In general terms,the genealogical informatio
VILLAGE,LINEAGE,AND CLAN VILLAGE,LINEAGE,AND CLAN 23 local or higher-order lineage is relatively complete and relatively a。b。mow置n The Ch'en fixed;that on a clan grouping is comparatively tenuous and subject lineage occupies two. to change.I shall return to this question in a later context;what Membersh needs to be shown now is that the genealogical difference between ip runs to about ten thousand distributed into six fang or sub-lineages.1 And so on.In his summary of his findings clan and lineage is contingent,not necessary;for some of the genealogies, on southcastern China Amyot says that lineage organizations constantly associated with a specific district or hsian of relatively indeed,their genealogical frameworks may be very small dimensions.Considering now the spatial distribution much in debate,as we shall see presently-and yet they function of surname groups in relation to one another,no significant as effectively as other,genealogically better endowed,local line- ages. pattern appears from the point of view of social organization. Groups with the same surname can show up in ten Let us consider some of the ethnography.Yang provides an different of a higher- order lineage placcs.There is no immediate organizational connection between 'of the Lee was in another county, them,certainly not merely from the fact that they have the same some cighty miles away,'where the headquarters of the Lee surname.There is none even when they belong to the same li being only one of the as sub-branches having migrated away from the founding village. Their populations do not have the same kind of interrelationship possess altogether about twenty mow of sacrificial land within I think Amyot uses the word lineage'rather too vaguely, but his analysis is extremely interestir Nanching.'A wide range of relevant material is to be found in and important.The Amyot's study of the Chinese in Manila.He was able to collect ssence of it is that what I have called higher-order incage a good deal of information on Ch'tian-chou,Fukien,from lationships are likely to be confmed to the small areas formed by hsiang (which 'can be crossed on foot in a couple of hours'). nce he on Hui-an county myot says that its several hundred villages and hamlets implying formal divisions laid down by governments,? and since srouped into about fifty hsiaug,a term which he translates as ship’. the eterm sing may desi gnate eeither a complex the areas within which higher-order lineage systems are likely to of villages and hamlets forming some kind of unity,or again,the mayhave itttodo with bureauricalydica I suggest that we refer to our crucial territories as largest village of this complex from which the latter derives its name.It is usually a market An av age hsing can be I am able to offer another example,on a smaller scale,from Taiwan. crossed on foot in a couple of hours.'Most of the population of Wolf,setting the background to his study of a villag near Taipei,tells us that the villa Hui-an is divided among ten major surnames ('lineages'),cach orient themselves towards of which 'occupies predominantly at least one of the Hsien the outside world in terms of the village's status as one of the five lin(a Hokkien word)which make up the'district'of Chi-chou village complexes.There are twelve such lineage hsiang.The Thisdistricthas never been bythe bureaucracy but its boundarics are those whichmark the line between the villagers'immediate private world and the larger impersonal 1bid,p.43. of the Ibid.,pp.52 f. ,p.40. che word a Amyot,p.ct,Pp,37任 wrong connotations for Americans
24 VILLAGE,LINEAGE,AND CLAN VILLAGE,LINEAGE,AND CLAN world of society at large'.Chi-chou appears to have been settled 25 in the late cighteenth century by men of the sumname Ong(pre may arrogate preponderant control in the marketing com- munity' sumably,Mandarin:Wang)and in early times the inhabitants of the'district'made up a single lineage.In time it segmented and It may well turn out,after the matter has bcen fully explored spread,and people ofother surnames began entering the district that in fact vicinage and standard marketing area are usually tosteAsaresult,today Chi-chou'consists of five independent congruent and that they provide us with the key to understanding Ong lineages each of which occupies a community in which a how local lineages are normally grouped together.There is third of their neighbors are families of another surname'The contrary evidence (how explain,for example,that units of the unity of Chi-chou rests on the common origin of the Ong people higher-order lineage of which the Nanching Lec,in Yang's while the separate identity of each of the five ln is a product of account,were members,were separated by eighty miles?),but the fact that it is built around a core of Ong familics who make it may well be that most of this evidence relates,not to lineages a local lineage.I take it that the five Ong local lineages together which are predominantly peasant in their composition,but to those whose affairs are in form (or formed)a higher-order lineage,which,as the dominant of subst the hands of gentlemen and merchants agnatic group in the district',provides it with its organizationa stance.It is part of Skinner's argument that men with wider framework. political and economic horizons will operate within areas defined Arguing with cogency and great scholarly clegance for the in relation to higher-order market centres;and it may follow ng areas in the non-official structure of the case Chinese rural socicty,Skinner has seized upon Amyot's remark that the largest village in a hsiang (itself also called by that term) county (such that local lineages very widely separated we is usually a market.And he has put forward the view very groupd together)were dependent on the fact that then persuasively that lineage organization above the level of the leaders of these local lineages hadan active interest in the superior village depends for its conformation on the catchment area of the market town of Sham Chun,which now lies just to the north of lowest level of market town.'My suggestion here is that,sincc the frontier etween China and the Colony of Hong Kong.Some peasant families have much social intercourse within their standard (perhaps all)of the local lineages belonging to the surnames marketing community but little outside it,interlineage ties con- Tang and Man (Mandarin:Wen)were members of geographi- cally tained within a single marketing systemare likely to be perpet- wide-ranging higher-order lineages. uated whereas bonds between localized lineages sited in different The Chinese term is applied toboth lineages and clans (as standard marketing areas tend to erode with time.'And he thesc terms are used by anthropologists);so too are tsung and tsung-tsu;the Chinese nouns by themselves are not unambiguous pointers to the precise nature of the about a Fukien'standard marketing community',by suggesting thatthe dominant localized lincage within a composite lineage used. e groups and quasi-groups for Yet Creel has recently argued that in anent Chna the terms and s(which in modemn Chincse are both translatable as 'surname')indicated precisely the dis- et品 ft to mv eptember 1964 I am profoundly grace tinction between clan and lineage which needs to be made in the copyor analysis of modern Chinese agnatic organization.The hsing andi业many ways a p 1 Skinn on Taiwan of SL: cong and
5 VILLAGE,LINEAGE,AND CLAN VILLAGE,LINEAGE,AND CLAN 27 wasa large and rather ooscommon descent group,showing an attitude of solidarity which in a specific situation might or of branches are supported by historical evidence of another kind. On the other hand there are good reasons for supposing that some might not produce united action.But the much smaller shil was, of the links 'established'in the genealogies are e fabricated.The in the fullest sense of Max Weber's terminology,a orporate group Not being a sinologue and knowing very little about go。a ancient China,I cannot say how far Creel's distinction is justified accuracy and creative imagination. by the evidence,but it may y be that if a sinologue turned a critica (and anthropologically focused)eye on the modern Chinese Consider the genealogy which Eberhard treats at length:the General Genealogy of the Wu Clan of Ling-nan in its latestedition terminology for descent groups,he might be able to assert,as Creel hason for the usage,that there is in facta ceare (c.1933).The Wu clan of south China is spread widely in the linguistic distinction in Chinese between what we call clans and lineages than we have thought.(It has been suggested to me by one Chinese scholar that in origin back into the first millennium B.c.,but nfined himself ne parts of China clans ar .hsing in his compilation to the eight'great houses'of Wu in the two and lineages fsut;but I do not think that the southeastern Chinese provinces.He knew that there were n this region aino the euusite sinogency ore Wu than he for genealogically to find all the villages but he tried gical knowledge,I regretfully leave s where Wu lived and to establish the names this problem of terminology hanging in the air. of their first settlers.Through Eberhard's account we can see how Descent and gencalogics justify both lineages and clans,but the compilers had to struggle to fit together various kinds of while it is easy ough to in c the genealo y of a lineage both genealogical data to makecoherent system. structurally and historically,the genealogies linking lineages The genealogy reaches down to the ighty-fifth generation dogans may be a pule What do they dohat but the 'great houses'themselves trace back their individual do they exist?Written genealo embracing origins to twenty-eight and generations,in the case of the deepest, cnormous populations and spanning a thousand or more years are by no means uncommon in China.In a recent major work fourtcen,in the case of the shallowest.All thegreat hou are arranged to show how branches ramify within Eberhard has analysed tw such compilations of direct nterest to while the compilers them,but 'constructed a frame which would accept all students of southeastern Chinese socicty. In tying to understand what these giant genealogiesae about, Wus now existing,disregarding their ethnic origin,there ar other brancheswhich the genealogists could not integrate into we may ask which of the following assumptions is correct this framework,and where they admit that the degree of rela- First,the genealogies are statements of historical truth,documents tionship remains unclear'.a Each of the 'great houses'has by for studying the geographica further ramification produced a number settlements,some ramifying of agnatic bra nches. concentrated,som cattered,and at this level of the gencalogy pective constructions of the relations between lineages,making it is possible that the'facts'are historical:Wehave no reason to historical'sense of the present distibution of lineages on the ground.In reality both assumptions s em partly to beborne out. doubt the assertion of the genealogy that each of these settlements was created by a membero the main house,or a member of a On the one hand,some of the data on the geographical movement known daughter settlement. 1H.G.Creel,'The Beginnings of Bureaucracy in China:The Origin of the A genealogy such as this is concerned with what some anthro- Mis Hul-chem pologists prefer tocall fission.That is tosay, ang Lit,The T N28置2a7Caa3 Depite on this terminolog to'reconstruct',the manner in which branches send out further branches,each resulting new unit being independent of its parent its general title,the book rests very heavily on the two genealogies. 1Op.t,pp.50-8. bd,p.57. 3 Ibid.,p.67. .,p.118