First I must define precisely what I mean when speaking about ethnographical investigation of. As any other ethnographer, I understand that such an investigation has to cover the whole cultural adaptation of the population of [1]. How deep and broad the investigation must be carried out depends 1. on the degree of differentiation of the given population, i. e,. an objective condition; 2. on how far the investigator must go with his description of the elements constituting the culture; this description cannot become clear until the relative value and relative weight of elements are defined after the elements have been grouped into complexes; 3. on the degree of preparedness of the investigators, their observation and analysis.
The problem of unit of investigation is closely connected with the first condition of tkLe population. Theoretically speaking among a culturally uniform population investigation of a single family is quite sufficient. However, in we are facing a strong social differentiation and an ethnical differentiation, to use the improper term «regional» differentiation [2]. If we adopt the dynamic point of view for social differentiation, the latter presents no difficulty for the ethnographer who differentiated groups. With the ethnical differentiation the situation is far more difficult.
How should an ethnographer approach this problem? First of all there must be done a preliminary work of distinction of groups from some point of view. Since the 1anguage is perhaps the easiest and the most simple character, it may be put at the basis of the map which at least will show the approximate territorial limits of groups which are distinct at least in respect to the spoken language [3]. Should other cultural elements and physical characters of the population be different too, we have to accept distinct units and here also «regional units». If our investigation reveals that, besides the linguistical and other cultural characters, the distinction involves consciousness of the kind in its various forms, we have to separate the units found from those which are the object of the ethnographers final inquiry, i.e. the ethnical units in their crystallized form and units of incomplete formation. We may expect from such an investigation that there will be revealed groups speaking not only Chinese dialects with sometimes undefinable boundaries, but also non-Chinese languages, and an ethnical map, the first important result of inquiry, will be ready. If we have such a map we may not rely upon trying village after village, but proceed with our investigation of ethnical units where it will be possible to distinguish easily common elements from the accidental, and often misleading, occurrences, and cautiously avoid publication of too many partial investigations which in future may burden our knowledge with undigested material and so-to-say refuse left be the experimenting apprentices in ethnography [4]. When the preliminary work is over, the ethnographer may further proceed to the grouping of units according to various methods, e. g. according to some cultural complexes related; for instance, dialects can be grouped into larger classificatory units, connected with the changes of leading ethnoses in whole China; should there be found sharply distinct social groups extending beyond the boundaries of ethnical units, the ethnographer may analyse socially differentiated groups and classify them accordingly; should there be important and determining geographical conditions, the ethnographer may use regionalism as a mean of building up larger classificatory units, e. g. the valley of the Yangtse River, the coastal region etc.; should there be revealed determining cultural factors, e.g. the cattle breeding complex, the millet and rice areas, even the form of government etc., the latter may be used as objective bases of classification of population according to the cultural cycles.
By this I do not want to assert that the village as a unit should not be used for other purposes, especially such ones as those pertaining to this mode of settlement, e. g. the common law, administration, means of communication, economics, census etc. where inferences were made having as their object the vi11age unit and where inferences cannot be generalized unless one is certain that a single example may warrant conclusions. Moreover, in the beginning of his work the ethnographer has to go through the stage of preliminary work. Village, when it exists, is a territorial and administration unit, it is an economic unit in a lesser degree, like the work on the family as a unit of larger units is necessary before proceeding to the clan, village etc. The life of a single family may ethnographically constitute the subject of a special study, but such a luxury is not practically needed. However, the life of a village taken out of the ethnical unit cannot form an independent subject for this life, cannot be fully understood within itself and should be studied within the reality in which it only forms a constituent unit.
1. Under the term «ethnographic» I understand a semantic adjective of «cutltural complex» as it is observed in ethnically and socially differentiated populations.
2. My reluctance of using the term «regional» is for the reason that the same region may shelter different ethnical units, or at least populations affected in various degrees by the process of ethnos which have to be distinguished. The investigation carried out in a «region» with a differentiated population living in it, actually will deal with the units based upon ethnical differentiation.
3. The main difficulty — both theoretical and practical — here consists in finding out the dialectical boundaries which in a number of cases will appear as intermediary zones of mixed dialects — the fact which commonly occurs. However, if one assumes a dynamic point of view as to the nature of ethnical unit, dialect etc. it will not hinder the classification needed by the ethnographer.
4. Not everything needs to be published at once. Not before the ethnographer becomes familiar with the subject of his study, he begins to distinguish what is of importance (relative value and relative weight of the elements in the complexes) and what is not of importance for taking upon himself the responsibility of publication and of leaving the raw material unpublished. It is here presumed that the investigator is prepared for this kind of work.