The Taxonomic Tendency The Cross-Roads Hypothesis Biological Evidence Ethnolinguistic Evidence Archaeological Evidence The "Evolutionary Network"

Biological, linguistic and archaeological evidence is considered in the reconstruction of the historical evolution of the "Austric" peoples. The Austric peoples are held to have been the earliest progenitors of Austronesian speaking peoples of the Pacific, as well as the Tai-Kadai and Austro-Asiatic speaking peoples of the Southeast Asian Mainland, who were a culturally related civilizational complex centered in South China and Northeastern Southeast Asia and who fanned out and diverged across the Pacific regions sometime after 10,000 B. C.

This reconstruction will be considered from the point of view of a cross-roads hypothesis that takes to task some basic biases and fallacies that are intrinsic to our tendency to taxonomically classify and to construct evolutionary trees based upon relative differences within our sample data. An alternative "evolutionary network" is offered as a viable, if non-parsimonious solution to some of the basic problems of taxonomic hierarchy.

It is argued that such biases in our evolutionary reconstructions represent oversimplifications of the actual historical complexity of past events and processes, and are largely the product of our own inherent limits of "depth perception."

There are many problems with the construction of evolutionary trees representing human culture historical development. It is clear that our theory or data will precondition the kind of evolutionary tree we construct because of our taxonomic organization of the data and at least tacit labels, and the kind of tree we build will therefore also reflect to some extent our ideas.

Several criticisms of taxonomic reconstruction of evolutionary trees are pertinent:

First, it has no where been proven that there is any necessary or tight linkage between cultural and genetic traits. In fact, more evidence points to the relative independence of language and culture from biological underpinnings. Thus the changes that occur in the patterning of a language or a culture may in no necessary way reflect changes which occur in the genetic composition in a group of people.

Second, unlike speciation in biological evolution, language and cultural change is influenced phenotypically by the diffusion of traits across boundaries. These processes and sources of exogenous change; are not well understood and are extremely variegated and complex, such that few simple one-to-one correspondences or simple cause-effect principles hold.

Furthermore, such exogenous influences have been continuously present in language and cultural change, and cannot be simply accounted for or factored out of our equations for such change.

Third, constant rates of change for language, culture and biology have been presumed but never proven. Rates of change in language and culture are likely to be much more variable, and the patterns of divergence much more complicated, than is usually posited for biological evolution. Because of the lack of constancy of historical change, we cannot calibrate our clocks by a non-relative standard.

Fourth, and finally, lacking a non-relative standard, we have no common ancestral baseline by which we can measure and compare changes in cultures, either through time or across space. The baseline we arbitrarily establish is more likely the reconstruction of our own biases and preconceptions than it is any actually valid representation of the past.

Our scientific preference for clear, exclusive taxonomic classification and for direct causal determination, renders problematic and usually biased our interpretation of the past as: 1) a movement from simple to complex; 2) a continuous rate of endogenous divergence; 3) a derivation from a single common ancestor; 4) a de-emphasis of exogenous factors of change.

A consequence of such bias is to seek and see in the record a homogeneous versus an heterogeneous origin--a single, finite, prototypical baseline from which all subsequent change is to be traced. There is an underestimation of the possible degree of variance and complexity of patterning of the original ancestral populations as we simply do not have a satisfactory measure or knowledge of the full range of variation in the remote past.

Such biases in our evolutionary reconstructions represent oversimplifications of the actual historical complexity of past events and processes, and are largely the product of our own inherent limits of depth perception.

Different techniques of analysis confer different capacities for calculating time depth. In the case of glottochronology and lexicostatistics, the depth of time that can be inferred for change is not completely known, though few scholars would believe that we could see back in linguistic time much deeper than 6,000 BP or at most 10,000 BP The depth perception of the genetic record is held to be much greater, and virtually unlimited. Archaeological records seem to be limited only by the paucity and kind of datable material in the deposits.

But the greater the span of time that must be accounted for, the greater the relative paucity of available data and the greater the gap that must be filled--the more inherently uncertain and vague become the outlines of our history. There is a tremendous difference between the amount of change that occurs in one year, ten years, a hundred years, a thousand years, ten thousand years, a hundred thousand years or a million years. With each decrement, we have a corresponding level of the relative improbability such that, whatever our evidence, our knowledge is that much less representative of the time frame it encompasses.

But there also occurs an effect of fusion upon the horizons of our field of view, such that at great distances differences of time of hundreds of thousands of years become equivalent to what in more proximate points may be a matter of hundreds of years. Past that threshold, something that is 1,000,000 years old might just as well be 2,000,000 years old, though the absolute difference is almost twice.

Thus there is also an inherent theoretical limit to our depth perception such that given whatever information that is available governing a period, we must accept its fragmentary character and the greater improbability of seeing accurately and clearly into the distant past.

The net consequence of this historical myopia is that it confounds our attempts at taxonomic construction. Lacking a stable basis in knowledge, such trees therefore tend to be unstable and this inherent instability is marked by their great proliferation, theoretical preconditioning, and ephemeral quality. One single piece of new material evidence can overturn or upset years of taxonomic reconstruction.

The depth blindness of our reconstructions renders these trees subject to a critical "taxonomic tendency" and deterministic fallacy that precludes either the possibility of alternative complex, multivariate models of change or the possibility of ever capturing the real complexity of the past with any reasonable accuracy.

The taxonomic tendency can be regarded as our proclivity to see things we label in terms of exclusive, non-overlapping categories. The deterministic fallacy follows from this in the tendency to rank order and hierarchically categorize relations. Temporally this translates into a linear progression of time in which things that come before predetermine those things which follow. The net outcome of these pre-dispositions is the spatialized diagram of the "Evolutionary Tree" that stands for change in time.

We get a sense of set-piece stability and stasis that is illusory when we consider the real fluidity of time and change. We get Zeno's paradox, and a stratigraphic conflating of temporal depth as a two-dimensional plane surface--a sounding into the well of the past. We get the inevitable convergence of the curve of variation towards a common ancestor.

It is one thing to reconstruct a prototypical ancestral form from several descended forms. Our depth of view allows this much. It becomes more problematic to reconstruct a super-prototype from presumably related prototypes, and even a fourth-order super-super-prototype. The problem is that we can only push our reconstructions back so far before the evidence upon which we base our reconstructions becomes too flimsy.

We can reconstruct part of the prototypical core of a common stock, but not the ancestral form, because it is mostly likely that the languages from which we base our reconstructions did not have a directly ancestral form, but only an indirect ancestral stock. We can reconstruct only from what we know, and not from what we don't know, and what we don't know always outweighs the little we know, such that each time we learn something new, our representation of the past becomes overturned.

There is a tendency each time we do this to augment our uncertainty and decrease our empirical foundation. Each time we achieve a more normal shaped curve of distribution, but each time the degree of variation in the curve becomes narrower and narrower. The artifact of our effort is that we come up with a strong sense of a shared central tendency, but one that revolves around a narrow axis of only a small subset of invariable forms.

Building evolutionary trees from this perspective will always have the same result of yielding a common, narrowly defined ancestor, losing the sense of original variability and the range of variation. Evolutionarily we know that divergence has been a continuous process. It is unlikely that two forms branched off from one another at the same time, though branching occurred through time. Many more offshoots were formed than are available now to our senses. The clean tree with the same point of divergence, the same ancestor rather than a related stock, is the deception of our own treelike representations--historically or evolutionarily a highly improbable event.

An evolutionary tree based exclusively upon the principle of irreversible divergence is bound to look symmetrically balanced and uncomplicated in its representation. In human cultural evolution, development has never been only irreversibly divergent, but possible processes of reconvergence related to borrowing has meant that we can speak of an evolutionary network of entangled shoots leading away and back to one another.

Finally, in our culture historical reconstructions it is probably unnecessary to try to extend our phylogeny past the super-family to some earlier proto-language. We may say that if Indo-European, Austric and Sino-Tibetan may be remotely related to some ancient ancestor--the original form or root stock is nonexistent except perhaps as the barest shadow on the surface of the sand. By any working definition, we can still maintain that these three language phyla are essentially, for all intents and purposes, unrelated--or rather that their hypothetical relationship is irrelevant because its reconstruction cannot be demonstrated. Another way of putting this is that their relationship, or degree of correspondence, is so indirect as to be meaningless.

It is not necessary to belabor the convolutions of the physical relativity of space-time to recognize in such tree construction the gross oversimplification of complexity and simplistic reification of order that is the outcome of reductionism. And yet, we would be very hard pressed indeed to come up with viable and satisfactory alternatives in the modeling of the complexity of our past. There is perhaps only one principle that we can be sure of--whatever period we are talking about, the past had probably never been as orderly as our present representations of it.

The attempt to construct an evolutionary model of human cultural biological evolution tends therefore to exclude the problems of historical process. Unlike species, languages, cultures and peoples may mix in productive ways. The borrowing between peoples, languages and cultures has been a constant source of exogenous change upon the historical developments of people.

In an especially complex historical context such as Southeast Asia, that has long served as an interregional cross-roads of trade and migration, historical processes may be inherently problematic, especially if we take into account the very plausible hypothesis of repeated reconvergence and subsequent back-borrowing from distantly related languages, cultures and peoples coming back into contact after a considerable period of separation and divergence.

The hypothesis entertained in this paper is that Southeast Asia has been the origin point of a single super family of remotely related languages, cultures and peoples, and the host to the intrusion of a second super family. The crossing-over and back-crossing between languages, cultures and peoples within this region has resulted in a complex, variegated and virtually unsolvable jigsaw puzzle of what and who are related, and what has been borrowed, lost and returned.

The "Cross-roads" hypothesis that Southeast Asian civilization, though possibly sui generis to somewhere in the Northern reaches of the Southeast Asian Mainland, may have actually had multiple points of origin and many different beginnings, is entertained in terms of a comparison of linguistic and biological and archaeological reconstructions--alternative "time-lines" and "trees" may be considered on the basis of different theoretical interpretations.

The cross-roads hypothesis is based upon the notion of the continuous existence of a widespread and relatively stable "root stock" or common civilizational "trunk" in Southeast Asia, from which at various periods separate branches diverged and frequently intertwined and reconverged, and toward which other alien influences intruded.

In considering this model, we must separate processes of linguistic and cultural divergence that represent the limbs and branching structures of the tree, from the more basic processes of aging that are affecting the development of the structure of the main trunk (as well as the entire tree) from which all related languages are divergent.

There is little reason to presume that the social landscape which existed before 10,000 BC in Southeast Asia was any less inherently complex in terms of its branching and variegation, and, from an analytical standpoint, any less problematic than what subsequently emerged during the proto-historical and historical phases.

As a cross-roads region, Southeast Asia has always been a great mixing pot of different peoples, languages and cultures. The capacity for Southeast Asian civilization to absorb the new and the alien and to maintain a sense of cultural continuity and identity with the old and familiar is nothing less than remarkable.

We can lay the foundation of our cross-roads hypothesis in positing a basic core of civilization, of cultural, biological and linguistic traits which are inherently conservative and highly resistant to substantial change over time, and in the excoriation of multiple outer layers which are more peripheral and therefore ephemeral products of historical influences.

We must distinguish between exogenous and endogenous sources of change, such that, by definition, we may claim that exogenous change, among other things, tends toward assimilation, or synchronous substitution and replacement, whereas endogenous change tends toward drift--or diachronic modification and alteration. We may further say that whereas exogenous change tends to be disintegrative and random, endogenous change tends to be more nonrandom and integrative.

Replacement of the core will proceed at a much slower and more constant rate and will remain less complete than replacement of the peripheral layers. The likelihood of survivals of basically unchanged elements from the very earliest times is highest in the core. Though both the core and the periphery will be subject to drift, the likelihood remains greatest that remotely related groups will share a few core characteristics which are not the result of diffusion, and that the more closely related in time, the greater the likelihood of shared core characteristics. In other words, people of a common heritage will most likely possess a common core.

Core characteristics may be most visible, and most conservative, among those groups that are the most outside or isolated the cross-roads of historical change and happenstance.

Drift can be referred to as natural historical divergence that is the result of endogenous factors of change and alteration. Though drift results in diachronic replacement, such replacement is always one of homologous derivation through continuous modification rather than of analogous conversion. Rates of drift will be variable and partially dependent upon outside influences, such that those groups that are from the exogenous point of view peripheral to the cross-roads, possess cores which are also most peripheral and therefore subject to the greatest amount of drift.

This is the consequence that drift and assimilation are not completely independent processes. The region of their interdependence defines an intermediate zone in which change is definable as amalgamation--the fusion process between the new and the old, the familiar and the foreign.

The very factors which contribute to the great external conservation of forms upon the periphery and the rapid assimilation of forms at the cross-roads, tends also to contribute to processes of greatest internal drift at the periphery and greatest internal conservation at the cross-roads. A consequence of this may be something of an evolutionary paradox--peripheral groups may always have core characteristics that are more reconstructable, but cross-road cores may be more likely to retain survivals of characters that do not need to be reconstructed--closer to the original prototypical form.

The greatest variation is not likely to be found in the oldestcore, but in the intermediate region in which those core characters merge with the peripheral. These are the non-primary core characters.

Another factor is the saltational nature of the overall process. Short-term stability may belie long term fluctuation, while long-term stability may be based upon short-term transitional periodicity. Another way of seeing this is that a group may not always remain at the cross-roads or the out-lands, and that to some extent, one group's cross-roads may be another group's periphery. Structures of the long run are not always predictable by the patterns of the present, and the current unfolding or stability is not always governed or predetermined by the long term patterning of change.

The problem remains to locate the core of a group. In this case of detection and discovery, several minor hypothesis may be presumed.

First, the most basic core (primary core) is common to all groups and all peoples who are related by common definition, no matter how remote the actual relationship. A corollary of this is that if cores are discovered which seem basically unrelated, this is evidence of a lack of remote relationship and evidence for a multi-lineal evolution of unrelated groups.

Second, interregional and regional distributions of cores may be identified which are unique for one branch of people or another. A corollary of this is that such secondary cores may be the diagnostic features of that branch or sub-grouping, and may locate that group to a common prototypical form.

A third principle is that two or more cores cannot occupy the same space at the same time, such that if there is evidence of one core intruding or taking over another, it must be construed as an event of assimilation, possibly involving the ethnocide or even genocide of the previous group, or of amalgamation. A corollary of this is that when two cores do come within proximate contact, conflict and chaos is likely to ensue--whether destructive or constructive in its consequences.

A fourth principle is that multiple non-primary cores may be added over time, and the historical pathway of a distinct groups development will be marked by the temporal accretion of such non-primary cores. A corollary of this is that the history of a distinct local group is embedded in a complex stratigraphy of layering of non-primary cores about some often hidden heart.

We may speak of the superceding of old cores by the emergence of newer cores, as the covering over of the older heartwood by the newer sapwood. Though the developmental history of this process may be quite variable for different groups, the capacity ("growth") and course ("age") of embedding, remains equivalent and the same for all groups.

A corollary of this is that all complexes are at core equally old, though some cores may be more embedded than others, and, regardless of the absolute age, all complexes still have the same capacity for growth. Another way of putting this is that more primitive forms though relatively older, may have more relative potential whereas more derivative forms, though relatively young, may have less relative potential.

We are still left with the problem of locating the core of all complexes, and the cross-roads hypothesis anchors this common core to those forms and functions that are the most basic, prototypical or necessary, and the most shared by correspondence between all groups. Linguistically, we might locate the basic core as those traits that are most basic of the human body, the elements of the environment, those relative markers or dichotomies, and those relationships that are most socially basic. Basic core characters are likely not to be conceptually abstract or based on anything that is able to be borrowed. They are unmarked, unelaborated and nonspecific concrete referents well-rooted in basic forms of expression, perception and conception.

There is a neoteny about these basic core characters. A child's detailed recognition of the forms of animals, faces, natural designs, insects, and basic man-made objects lends credence to the notion of natural categories, a level of intellectual sophistication preceding the formal acquisition of color terms, of numbers and mathematical manipulations, etc. Similarly, the child's natural language and its structure also reflects this basicness--a bee as being the basic form for both a distant bird, a fly, a bee and a spider.

Core characters share a feature that might be referred to as a high rate of correspondence of prototypical form and function. Correspondence derives its value from its associational context with the character complex in which it takes part and in which it is embedded. Correspondences can be said to be something more than mere analogues, but not quite completely homologous structures. Correspondence might be taken as the measure of proto-typicality or "basicness" of a trait or trait complex--it is that set of characters for which there is the greatest likelihood of common correspondence with other sets of traits, and therefore the greatest likelihood (or ratio) of genetic, non-convergent relationship.

Correspondence can be taken as an indirect measure of the degree of relatedness between two complexes. More directly, it is a measure related to complexes, or to its constituency, more than to its individual components. It is a measure of the basicness of its constituency structure, a measure of it degree of internal coherence or strength of integration--the linkage between its components is relatively stable and unbroken--and a measure of its equivalence of value.

Indirectly, correspondence measures the degree to which similar traits in similar complexes perform similar kinds of functions. Correspondence as an indirect measure of relatedness is rooted in the "structural covariance" of its basic factors, even across different forms of data.

Bones of early modern Homo Saipiens have been found in association with Bac Sonian contexts in Northern Vietnam and Malaya (Burling, 1967:p. 19) But without doubt earlier modern humans ancestral to the Australian Aborigines must have passed through, and perhaps occupied, the regions of Southeast Asia between 50,000 and 30,000 BP, as evidenced by the early remains at Niah Cave in Sarawak. Burling speculates that later mongoloid groups coming from the north pushed back, displaced and partially absorbed the earlier negritoid and "Australoid" races that are held to have unique physical features that can be found occassionally among some contemporaneous Southeast Asians. The negritoids survive today as the Semang of Malaysia and the Aeta of the Philippines, and the indigenous peoples of the .Andaman Islands off the tip of Burma. "When these characteristics are shuffled among a large population, an occasional individual is likely to turn up who looks rather Negritoid...." (Burling, 1967: pg.19)

Next in this conjectural sequence are the "Veddoids" resembling the people of South India and Sri Lanka, who are held to have either descended from or mixed with Australoid elements in Southeast Asia. They are held to be represented by the Senoi of Malaya "wavy-haired, narrow headed, dark skinned agriculturalists practicing shifting cultivation. And, as in the case of the Negrito type, this type has varying degrees of dilution over much of Southeast Asia..."(Buchanan, 1967: pg. 27)

The Melanesoids, ancestral to the Melanesians, are held by some to have come next to Southeast Asia, but if so, then they were only a transitional type. They were succeeded by the Austronesian or "Indonesian" type who spread throughout the Western Pacific as well as venturing into peninsular and insular Southeast Asia. Arriving perhaps earlier than 3,000 BC, perhaps concurrently with the arrival of the more "Mongoloid" peoples from the north.

There is even held to be an infusion of a "Nesiod" group, or "Brown Race" who were "olive-skinned" with wavy hair and narrow noses and originated in Western Europe, through the Mediterranean and the Middle East to northern India. "The group appears to have diffused into Southeast Asia from the northwest and may have been the bearer of the megalithic culture of Assam and Chota Nagpur. Today its extent is from Southern China (many of the minority peoples are non-Mongoloid and belong to this group) to the Indonesian Islands...." (Buchanan, 1967:pg. 28)

But the latest group held to have arrived in Southeast Asia were the "Alpine-Mongoloid" peoples--broad headed, black straight hair, epicanthic folds, prominent cheekbones. They are held to have moved slowly and steadily southward from the second Millenium BC, spreading down the river valleys of the Irrawaddy (600 BC), the Chao Praya, the Mekong and the Red River, and down the coastlines, "being deflected from Yunan by the solid group of Nesiot peoples established there." "This southward drift of Mongoloid peoples has continued up to the present time and the migration of Chinese into the area in the last century in response to the demands for plantation and mine labor and the new commercial opporutinties, can be regarded as merely the continuation of this thousand-year old overflow towards the tropics." (Buchanan, 1967: pg. 29)

Constructions of such racial sequences are dubious at best, and unenlightening at worst--they remain essentially unproveable. The aspect of the genetic and cultural diversity and heterogeneity of Southeast Asia is more informative--we would expect with a deep history of migration and mixing, a broad and pervasive "Southeast Asian base" to have formed characterized by such diversity and heterogeneity, among which exist perhaps isolated pockets of "purer" forms. "The general result of these movements is reflected in the great diversity, racial and cultural, of Southeast Asia's population. Mixing of peoples of the most diverse physical types has been taking place over many many millennia; so, too, has the inter-weaving of many diverse cultural elements." (ibid.: 27) It would be expected that the range of variance within and between local and regional populations is much greater than the differences separating regions.

One must legitimately ask, when given the lack of complete objectivity of such racial data and its analytical interpretation, whether the racial sequences inferred about the past in Southeast Asia were not as much the product of our own prejudices about racial progression and superiority rooted in a Great Chain of Being, with the most primitive black Australoids and Negritoids coming first, followed by the brown Veddoids and Nesoids, and then by the yellow Alpine Mongoloids,  and the concomitant advancement of civilization.

One criticism of this approach is that we do not have at any point in time a good idea of the genuine total variance of human populations. We cannot simply assume that at some arbitrary point, 10,000 or even 30,000 years ago, there was less overall genetic variance than there was at some later point, 3,000 BP or even today. Not being able to assume increasing racial variance with time, or the original lack of variance, means that we cannot say with any legitimate certainty which "race" came first, based alone upon extant evidence of "types."

Luigi Cavalli-Sforza has been the most sophisticated proponent of the consociation of genes and culture in history, and he has given us a rough map of the Southeast Asian populations based on relative differences in blood alleles. His work is more interesting and more empirically valid that the previous taxonomies shown, and therefore warrants consideration before building his taxonomic scheme. He has dealt quantitatively and mathematically both with the problem of genotypic distribution and change through time, as well as with the issue of cultural transmission, the Neolithic transition, and the question of the relationship of language, archaeological, cultural and genetic characteristics.

First, Cavalli-Sforza has posited a definite correlation between genetic heritage and cultural heritage, based upon the way that both have depended upon the same factors of isolation and migration, themselves the function of history, ecology and geography. (1986: pg. 14) The main factor of this correlation, beyond isolation and migration, are the adaptive radiations of various groups at various times, and the alleged cultural conservatism "so that people settling in new areas will retain much of their culture even for long periods of time." (ibid.: pg. 14). In this association he emphasizes the role of "demic diffusion" of people who displace others rather than of the cultural diffusion of cultural characters, evidenced by the relatively slow and steady rate of the spread of the Neolithic revolution in Europe from its origin area in the Middle East.

Relying upon "principal component" analysis of genetic composition of various populations, his genetic map of clinal variation almost exactly corresponds to the map of the spread of the Neolithic. The picture that emerges agrees with the prior existence in the area of expansion genetic variation, the capacity of farmers to increase to high densities, a slow migration rate and the assimilation of local hunter-gathers--either by marriage or acculturation.

Cavalli-Sforza also deals with the geography of the distribution of cultural complexes, and the "fine clustering" that is related to the greater local versus broader areal or local affinities that are probably nonrandom in distribution if only slightly disturbed by random movements of mobility. Again, assume the inherent conservatism of cultural complexes, such distributions could be achieved by "demic diffusion." "...In the complex history of the settlement of Africa, some groups have expanded much more than others, and inevitably split a number of times in tribes that have occupied different areas, split again, moved again and so on. A highly conserved cultural trait will then spread with the people. Naturally there cannot have occurred only fissions, but also fusions, which will generate greater complexity; but with groups rapidly expanding in relatively unpopulated areas, fissions are likely to have been more numerous than fusions..." (L. Cavalli-Sforza "Diffusion of Cultures and Genes" in On Evolutionary Anthropology: Essays in Honor of Harry Hoijer, 1983. Edited by B. J. Williams 1986: pg. 24)

Cavalli-Sforza correlates demic descent with a certain cultural complex of characters, especially family, kinship and house-building style, which were traits least correlated with environmental adaptations and the complex associated with such adaptation--subsistence strategies, technology and house-building materials. His correlation in this regard is indirect though, because he first correlates language differentiation, as highly conservative, with genetic differentiation, and then language distribution with the previously mentioned trait complexes. The factors of whether these correlations are by demic descent or "reciprocal interaction between the traits" (pg. 27) is not clear, but such correlations provide evidence of slow rates of evolutionary change, either during movement or fission.

In conclusion, he supports the high correlation between some cultural traits, language, kinship and family, and genetic transmission, because of co-transmission--"that is, people carry both with them. The vagaries of demographic expansions associated with new cultural developments, permitting rapid demographic increase in some groups, can only magnify the geographic variation of genes and culture while retaining their essential associations, i.e. their covariation." (pg. 33)

In another work, (L. L. Cavalli-Sforza and M. W. Feldman Cultural Transmission and Evolution: A Quanti-tative Approach . 1981), Cavalli-Sforza makes the case for the close association between genetic and linguistic transmission, and the relative stability of language change over time. He correlates genetic data derived from the study of 120 alleles (Luigi Luca Cavalli-Sforza, Alberto Piazza, Paolo Menozzi, and Joanna Mountain "Reconstruction of Human Evolutin: Bringing together Genetic, Archaeological and Linguistic Data" in Proceedings of the National Academy of Sciences, USA. Vol. 85: pgs. 6006-4, August, 1988) linguistic data based upon evidence for "superfamilies" and Archaeological evidence. He creates an evolutionary tree in which 42 groups are related, showing close association with the linguistic and genetic data for time of splitting and sub-grouping.

His reconstruction of Southeast Asian groups contains an original super-cluster that splits into mainland and insular southeast Asian groups, "a fairly tight cluster of six populations...on bootstrapping, Filipinos are lost 29% of the time, Malaysians 23%, and Indonesians 7%" (1988, pg. 8004) Pacific Islanders clustered into a loose group of three, and New Guineans and Australians which "remain together more than 50% of the time."

His phylogenetic reconstruction of genetic evolution in Southeast Asia records and early split of Southeast Asians with New Guinea and Australian populations at a genetic distance of about .017 at a time estimate greater than or equal to 40,000 BP. A subsequent split separated the New Guineans and Australians 25-30,000 and later Mainland Southeast Asians from Pacific peoples between 15,000 and 30,000. Subsequent divisions of Polynesians and then Micronesians/Melanesians, and later South Chinese, Austro-Thai, Indonesian, Malaysian and Filipino groups, and then finally the Austro-Thai into the Mon Khmer and Thai groups.

Though the linguistic evidence does not exactly fit this model, it is apparent that the Austric superfamily, with a possibly earlier relationship with IndoPacific and Australian families, supports the view of the grouping of Austronesian languages with Austorasiatic and Daic languages. The only misfit are the South Chinese languages (Miao-Yao) that are grouped with the SinoTibetan but which at least one linguist, Benedict, has associated with Austro-Thai.

Unfortunately, there is very little nonrelative, clear-cut stratigraphy in language, and no method as absolute as radiocarbon dating. Linguistics does have a small arsenal of techniques of comparative linguistics, the age-area hypothesis, lexicostatistics and glottochronology, which are of some precise value in determining common genetic relationship and in reconstruction prototypical ancestral languages. Unfortunately, it is the thesis of this paper that such techniques remain relative, cross-linguistically problematic, and of a relatively shallow parallax or depth perception. In general, they are based on some set of presumptions about a constant rate of replacement of basic cognates, the inherent conservatism of language, and the continuous rate of divergent, endogeneous morpho-phonological conditioning of words and structural change in syntax.

The ethnolinguistic map of Southeast Asia is as obscure and confusing as the physical evidence, though on firmer ground than the former. I have come across at least five different phylogenetic trees of the historical relationship and classification of the different languages found in Southeast Asia. What this represents is the lack of thorough linguistic description and comparison of the languages. Each tree perhaps rises from and gives rise to a different picture of the historical development and sequences of Southeast Asian peoples. The pattern is confused, because it is not known which pattern of movement, in which direction, obtained for each language group.

There are of course, with any classificatory problem based upon the lack of complete evidence, the lumpers and the splitters. Splitters might see as many as eight or more language families. Robbins Burling divides them into eight families--Miao-Yao, Tibeto-Burmese, Thai, Vietnamese-Muong, Mon-Khmer, Malayo-Polynesian, Kadai, and Andamanese. LeBar, Hickey and Musgrave have as many as eleven different language families just for the mainland region, including Mon-Khmer, Viet-Muong, Semang-Senoi, Cham, Malay, Thai, Kadai, Chinese, Tibeto-Burman, Karen, Miao-Yao. Shorto lumps the languages of the mainland into five families: Indonesian, Mon-Khmer, Tibeto-Burman, Thai, and Miao-Yao. Buchanan's picture is somewhat different, recognizing three major families--the Austro-Asiatic, including the Mon-Khmer, the Malayo-Polynesian, including Indonesian and Cham, and the Sino-Tibetan, including Vietnamese, Thai and Burmese. Grouping at the next higher level of the superfamily is also variable. Buchanan's Austro-Asiatic and Malayo-Polynesian speakers were the original inhabitants of the region, followed by the Sino-Tibetan Languages which came to predominate and push the others of the Mainland into the highlands of the Annamite Cordillera except for the Lower Mekong region which remained Mon-Khmer.

According to Shorto, Thai is the most recent, widespread and least differentiated language, whose introduction is a matter of historical record, driving like a wedge down the center of the peninsula, shattering the formerly continuous communication network of localized cognate speech groups, spreading from a point of origin somewhere along the Vietnamese-Chinese border--the area also representative of the greatest linguistic diversity. Burling notes that the Kadai language is found in the same general region, giving credence to the notion that they may be both historically related.

LeBar, Musgrave and Hickey identify four superfamilies--Thai-Kadai, Sino-Tibetan, incorporating Chinese, Tibeto-Burman, Karen and Miao-Yao, Austro-Asiatic, including Mon-Khmer, Viet-Muoung and Semang-Senoi, and Malayo-Polynesian including Cham and Malay. Frank M. Lebar, Gerald C. Hickey, and John K. Musgrave Ethnic Groups of Southeast Asia, Vols I & II Human Relations Area Files, 1964)

Tibeto-Burman languages are concentrated in the northwestern region of Southeast Asia and are attributed to a late migration into the region. Because they are much more diverse, they are considered to have had a longer history than the Thai language. Malayo-Polynesian, though highly dialectically differentiated in the insular regions of Indonesia, are represented by only a few languages in South Vietnam, and shows greatest diversity in Melanesia.

Mon-Khmer is generally grouped as an Austro-Asiatic language and is held to be the oldest in the region, with age-area correlation supporting a central origin of the Mon-Khmer somewhere in the northern highlands of Thailand and Laos. These languages are related to the Munda languages of India.

Viet-Muong, is widely regarded as a close relative, or cousin of Mon-Khmer, though it has been in an especially ambiguous category, assigned by some to the Thai or even the Tibeto-Burman languages.

At an even higher level, some linguists have related the Austro-Asiatic with Austronesian languages. Robbins Burling suggested the possibility of the remotest relationship of all the languages in one great super-family. "Such an hypothesis should probably not be dismissed completely, though it seems more likely that the similarities can be explained by mutual borrowing through out the long course of history of these languages.... As a result, languages can come to resemble each other in many ways, even though they go back to different antecedents." (Robbins Burling Hill Farms and Paddi Fields: Life in Mainland Southeast Asia. 1963: pg. 26. )

More recently, linguists have come to believe that Austro-Asiatic and Austronesian were the respective continental and maritime descendants of a common Austroasian ancestor. Joseph Greenberg has traced a single ancestor "Austric" for Austro-Asiatic, Austronesian, Daic and Miao-Yao, all of which diverged fairly early. Thai diverged more recently from Austronesian.

The Austric Hypothesis was first proposed by Schmidt, and has recently been revitalized by the work of Benedict on Austro-Thai and Headley, and has been more recently picked up by Greenberg and Meritt Ruhlen.( Merrit Ruhlen A Guide to the Languages of the World 1976) Ruhlen does not trace the structure back to the earlier Austric, but has to main families of language--Austro-Asiatic and Austronesian. Austro-Asiatic comprises four brances--Munda, Mon-Khmer, Malaccan and Nicobarese. Mon-Khmer, the most widespread and diverse branch, consists of sub-branches--Palaung-Wa, Monic, Khmuic, Viet-Muong, Katuic, Bahnaric, and Pearic. Proto-Mon-Khmer, from which all the sub-branches descended, is held to have constituted a unified language during the second Millennium BC.

Austro-Thai is the hypothetical super-family which embraces all of Kam-Tai and Austronesian languages. Next to Indo-European, it is the most widely dispersed language family in the world. Ruhlen divides Austro-Thai into "Kam-Tai" and "Austronesian" some time before 5,000 BC, and traces its origing to South central China. Kam-Tai is further subdivided into "Tai-Kam-Sui" and Kadai. From Tai-Kam-Sui the Tai language families split off from the Kam-Sui, the former giving rise to Thai in the Southwest, Nung in the Central and Yay in the north, and the latter leading down to Lakkia. Kadai has lead down to Laqua.

The Austronesian family is divided into two major branches, the Western or Indonesian, and the Eastern or Oceanic. Austronesian dispersal across the Pacific began sometime after 5,000 BC, with the occupation of Formosa, by which time the Kam-Thai languages had already diverged. Proto-Austronesian speakers found Micronesian and Polynesian Islands unoccupied, occupying the Philippines, and the presence of Melanesian and Indo-Pacific speakers who originally occupied a wide area of the Indonesian Archipelago. Austronesian speakers obliterated the Melanesian peoples. The greatest area of diversity today of Austronesian languages is bi-focally situated in Western New Guinea into Eastern Indonesia and Eastern New Guinea until neighboring New Britain. By 3,000 BC the original Austronesians had diverged into several distinct groups. One group constituting the "proto-Oceanic" made its way to the area of east New Guinea-New Britain. The other group went southward to somewhere near the Celebes and became the eastern branch of Austronesian.

It is hypothesized that the age-area hypothesis belies the original homeland of Austronesian peoples because of a backward expansion during which the original north to south expansion was followed by a south to north movement, obliterating traces of the original languages. The present diversity of these languages is held to represent only a portion of the total diversity that originally prevailed.

Evidence for this "double dispersal" hypothesis lies in the presence of Austronesian of great antiquity lying upon the periphery of the central focal area of Western Austronesian. From here, subsequent to 3,000 BC, these languages spread westward through southwest Indonesia and Borneo and northward back to the Philippines, obliterating the previous Austronesian languages. Formosa today is split into two branches of Austronesian, the ancient Atayal descended from the original proto-Austronesian form, and the rest belonging to the second northward expansion. Indonesian subsequently split into Western Indonesian and Hesperonesian that subsequently split into Eastern Indonesian and Northern Indonesia, comprising the Philippine and Formosan languages.

The Eastern Branch of Austronesian began to disintegrate in about 2,500 BC into separate language groups that came to disperse across the entire Pacific region. By 1,500 BC Fiji was occupied by Eastern Oceanic speakers, who several centuries later became the first "Proto-Polynesians" first occupying Tonga. The settlement pattern of Polynesia is a textbook case of historical linguistic methods. Tongans settled Samoa, then Samoas inhabited the Marquesa Islands by 300 AD, as well as the Tokelau Islands, the Ellice Islands and other Islands in Micronesia and Melanesia. By 500 AD Easter Island and Tahiti were settled by the Marquesans. New Zealand was settled from the Cook Islands by 800 AD, and Hawai from the Marquesas by 1,000 AD.

Robert Blust, in his paper on Austronesian Culture History, ( Robert Blust "Austronesian Culture History: Some Linguistic Inferences and Their Relations to the Archaeological Record." in World Archaeology, Vol. 8, No. 1, pp. 19-43.) uses the comparative method in the corroboration of archaeological evidence in the reconstruction of the culture history of the proto Austronesian peoples. Despite a major disconformity between Archaeological and Linguistic evidence relating to iron and a number of other material items, including rice, which has been attributed to the inherent complexity of the regions involved, there are a number of areas in which both lines of evidence point to similar conclusions.

An early paper by Hendrik Kern (1889) presented lexical evidence of an original Austronesian speech community inhabiting an environment of sugarcane, coconut, bamboo, rattan, cucumber, stinging nettle, deriss root taro, banana, pandanus, yam and such animals of the dark-haired langur were found." He argued that this homeland must have been insular or littoral, otherwise there would not be so many far-flung cognates relating to the sea, marine life and maritime technology. He attributed to these speakers the domestication of the dog, pig and fowl, and knowledge of rice and iron. "In this way he was able to limit the class of possible homelands to the coastal zone of the South-East Asian mainland.

An alternative view was developed by Murdock (George P. Murdock "Genetic Classification of the Austronesian Languages: A Key to Oceanic Culture History" in Ethnology 3.2, pp. 117-26.) who, based on the great concentration of these languages in Eastern New Guinea, believes their presence predates the borrowing of these cultural traits from "peoples of a different language and superior culture" on the South-East Asian mainland. This hypothesis shows remarkable disregard for the wide sharing of basic cognates related to many of these cultural items which would require assignment to Proto-Austronesian of these items.

Blust outlines Proto-Austronesian culture history as follows: they occupied settled villages containing houses and public buildings, "an unwalled or low-walled building where people met, public business was transacted and strangers spent the night," a "boat house," a "house of retirement for women during menstruation and after childbirth," a yam shed, a small shrine house on poles with only one side of roof only. Houses were made of stilts and entered by a ladder. A gabled roof contained a ridgepole, covered by an inverted log or bamboo rain shield, thatched with sago leaf. A hearth was built on the floor to one corner, with one or more storage shelves for pots, firewood, etc. above it. Inhabitants slept with a wooden head rest or pillow. They had the pig, dog, fowl and hunted, made pottery, plaited mats and baskets, wove true fabrics on a simple back loom, mended torn materials with needle and thread, tattooed themselves, chewed betel, and drank a kind of intoxicating beverage. Some form of native script may have been invented and used on perishable materials. Evidence suggests they possessed a well developed maritime technology, cultivated a variety of root and tree crops, rice and millet which they pounded with a wooden mortar. Bow and sharpened bamboo spikes were used in hunting and warfare, and "headhunting with is associated complex of religious ideas must certainly have existed as early as 2000 B. C., they may have had slavery and some degree of social stratification. Various tropical skin diseases were prevalent. The resulting mixed picture suggests a somewhat "polymorphous economic base incompatible with the somewhat rigid notion of 'progress' from one exclusive level to the next." (pg. 37)

The pig, as evidenced by skeletal and dental remains and by similarity of cognate form, is an example of striking corroboration between linguistic and archaeological evidence.

The disconformities of this picture relate to the apparent cultural loss of iron and rice, supplanted by sago, and the loss and subsequent partial reacquisition of the back loom, as these peoples fanned across the Pacific. The likelihood of distant groups referring to iron and to a "knife/sword" by the same cognate by chance or borrowing remains small, unless the borrowing was simultaneous from the same apparently widespread source. The early origins of Austronesian, dating the separation of the Atayalic from others between 4,500 to before 6,000 BP, makes it unlikely that the original referent of these words were to "iron" or even the earlier "bronze". Later separations of Batak and Iban, 3,500, and the Northern Sarawak languages, 3,000. Other cognates, notably "blacksmithing," "anvil" and "rust" support this picture of a major disconformity.

Similar kinds of contradictions appear with rice ("Rice plant, unhusked rice, husked rice, pound rice and cooked rice"), millet/barley, weaving ("Loom, weave, cloth, weavers "sword", 4000 BP), writing systems ("Write, paper, book, letter, write."), headhunting (4,000 BP) and the blowpipe. All show evidence of an early common origin in the language.

Rice, in the form of Oryza sativa, the Asian cultivated species of rice, shows greatest diversity of varieties in a belt "extending from the Assam-Meghalaya region of India to the mountain ranges of Southeast Asia and southwestern China." Early maturing, drought resistant varieties probably emerged between 15,000 and 10,000 BP along the Northern and Southern Slopes of the Himalayas. Annual ancestral forms of O. Sativa emerged along the periphery of wild annual progenitors, in the southern reaches of the Himalyas and "to a lesser extent in southern and southwestern China. Alternating periods of drought and pronounced temperature variation accelerated the development of the annual forms of O. Sativa in northeastern and eastern India, northern Southeast Asia and southern China..." (M. S. Swaminathan "Rice" in Scientific American Jan. 1984..)

While it seems highly unlikely that so many different groups would share by accident or even borrowing the same complexes, it is also highly implausible that a common proto-culture 6,000 BP would share an advanced iron working technology and a writing technology. Some other kind of explanation must be sought, either an intrinsic bias in lexicostatistic methods, or a complicated history of simultaneous widespread diffusion and subsequent migration, involving possibly "back migration." This is a clear instance in which the parsimony and logic of our methods is contradicted by the complexity of archaeological and historical evidence.

It is in this regard that we can consider the kinds of trees we can build with more substantial archaeological evidence. The evidence from Niah Cave can be taken as a type-site for the prehistory of the region. Tom Harrison's 1967 progress report (Tom Harrison "Niah Caves: Progress Report to 1967" in The Sarawak Museum Journal. XV, 1967 pp. 95-6.) puts the presence of early stoneage and the earliest modern Homo saipiens saipiens at 40,000 BC, and evidence of occupation until about 30,000 BC. A second "Mesolithic" period evidenced by an extensive burial complex dates "prior to 4,000 BP and not earlier than 20,000 BC. This cultural complex is characterized by presence of advanced flakes, edge-ground pebble tools at the lower levels. In strata prior to 2,025 BP are found Neolithic burials in wooden or bamboo coffins, with pottery, quadrangular adze, mats, nets. By about 1,200 BC there is evidence of bronze, elaborate pottery vessels and urn burials. This is soon replaced by iron technology, import ceramics, glass beads and death ships before 700 AD. Site looting and Malay texts are found at 1400 AD and glass bottles, indicative of whites, by 1860 AD.

By this evidence, and associated evidence from other sites, we may speculate that an early occupation of Southeast Asia occurred around 40-30,000 BC of peoples remotely related to the Australoid peoples. A second clear phase of occupation occurred between 20 and 10,000 BC. We might suggest people remotely related to the Negritoes.

A third phase of occupation commences within the fifth millennium BC, and continues until the middle of the second millennium. This Neolithic culture is characterized by round axe, pottery, mats and nets, and undoubtedly reached Borneo by sea. From then on continuous contact by sea lead to successive phases of bronze (1200 BC) and iron technology (500 BC-500 AD?)

If we compare this with Higham's chronological table for sites of Mainland Southeast Asia, (Charles Higham The Archaeology of Mainland Southeast Asia: From 10,000 BC to the Fall of Angkor. 1988: pgs. xv-xvi ) we see a very similar agreement. Early Hunter-Gatherers, for example at Spirit Cave, before 10,000 BC with a limited range of stone tools, wooden implements for hunting and gathering, fishing and trapping. Shellfish and gathering wild plants were important to these small band sized groups. Between 5,000 and 1,500 BC there is evidence for extensive coastal settlement at Khok Phanom Di that culminated in a complex maritime cultural orientation, rice, exchange, ranking and elaborate mortuary ritual.

There is overlap with settlement sites from General Periods A (3,000 BC) and B (2000-500) along the interior regions of the major streams and valleys (Ban Chiang, Phung Nguyen, Samrong Seng) with small settlements and weak ranking, stone implements and shell indicating exchange with coastal communities, and giving way to bronze working sites with mines in the hills, casting in lowlands, ranking in small communities, signified by jewelry and bronze implements and cultivated rice.

There is then General period C beginning from 500 BC and characterized by iron-working, interregional contacts with China and Indian Civilizations, social ranking, agriculture. Bronze drums, body plaques, bowls and drinking vessels, and chiefly burials in boat coffins. From AD 200 until AD 1500 General Period D corresponds to the rise of Mandalas in Northeast Thailand, the Chao Phraya valley, coastal Vietnam, the lower Mekong, with court centralization, Indian inspired religions, statecraft and Sanskrit.

Given these alternative trees, what conclusions can be drawn from the data that would allow a viable reconstruction of the long-term history of Southeast Asia? From the standpoint of chronology and sequence, the archaeological data is the most dependable. The problem with it is the lack of knowledge of the total range of variation of the original population. For instance, are the samples we possess representative of the full range of variability present at the time of their deposition, or of what existed in the time periods between our radiocarbon dates?

On the other hand, there is a certain modicum of complementariness of the linguistic and genetic record with the archaeological record. Comparative reconstruction of proto-language families allow us to backtrack and gain some knowledge of the range of linguistic variation during previous periods. The Genetic evidence provides us with some sense of the most conservative rates of change and the degree of "absolute" differences separating current groups.

Things to be considered in each record:

Genetically, there is always some degree of biological intermixing of gene pools such that the dominant population will tend to absorb the minority population

Population flow will go in the direction of the downward gradient of the asymmetrical power differential. Great radiation of human complexes occurred because of the adaptive advantage that they conferred to those who possess them.

In the Archaeological record, periods of temporal overlap in cultural horizons are those transitional eras of spatial-historical contact which signal the inauguration of a new set of developments in the integration of the region.

Linguistically, the odds of widely dispersed language families sharing a common set of cognate complexes by accident or simple processes of borrowing are highly unlikely, and thus constitute the analytical basis for the reconstruction of their shared root, but it is also apparent that entire complexes such as iron working, weaving or writing, may be borrowed from a single source at one point in time and subsequently rapidly and widely diffused across the spatial network as part of a common trade language that is far-flung. These complexes will become subsequently embedded in the distribution of the families to be indistinguishable in the reconstructions from earlier, original complexes. Such introduced complexes may come to occupy the space of the complexes which they have displaced.

It is also apparent that we must take into account certain important discrepancies and occurrences in each of the different kinds of records.

First, we must consider the possibility of culture loss considered earlier--the dropping out of a rice complex or a writing complex among the Polynesians such that their basic cognates would be retained though no archaeological evidence may be found. Such cognates would represent anachronistic survivals.

Second, is the notion of the backtracking of a culture such that there occurs subsequent reconvergence of previously divergent cultural complexes. This is akin to an evolutionary tree in which the branching may be seen as spatially reversible, if not diachronically reversible.

The consequence is that the long term spatial distributions and patterns may not accurately reflect the complexity of the temporal patterns of diffusion and distribution. Reconvergence is a possibility always to be considered in human cultural evolution, while by definition it is impossible in biological evolution. In this regard, the question of the differential likelihood for reconvergence of different trait complexes must be entertained--such that homologous trait complexes which are similar in form and function may enjoy a higher rate and more complete reconvergence, and it is precisely in these trait complexes that we can expect the highest rate and likelihood of reconvergence to occur.

When reconvergence occurs, we are left with either a telescoping of the history of divergence and the consequent contradictions of the side-by-side presence of very old and newer forms.

We end up with a patterning of cultural evolution which resembles more of a directional transition network of interconnecting branches than a branching tree. We might speculate on what kind of surface patterning such a network would yield:

1. A shattered and variegated distribution in which very new and old elements are both peripheral and central, widespread and isolated.

2. A pattern in which the greatest distribution of related languages may not be indicative of the greatest age.

3. A pattern in which the nodes with the greatest number of interconnections may be the most derivative, confused and least historically indicative.

4. Boundaries between groups will tend to be fuzzy such that the dominant group will be marked by the absorption of the characteristics of the subordinate group, and the subordinate or peripheral groups will become characterized by borrowing from the central groups.

5. On borrowing or diffusion, versus divergence, character complexes will undergo adaptive modification that is similar in form to endogenous processes of divergence.

The primary characteristic of an evolutionary network that distinguishes it from the more conventional tree is that any single offspring may be the actual descendant of more than one ancestral parent. It can be shown that in the cases of multiple genesis, the dominant and proximate parent will be the most apparent and probably associated. The second corollary is that though many languages may show direct divergence from a single common ancestor, and only few will be genuine Creoles, all languages will be, probabilistically, at least indirect offspring of such mixed marriages. One theoretical consequence of such networks is that, depending upon our primary inferences, alternative trees may be reconstructed for any given set.

There are two sets of related theoretical issues that we must deal with in regard to such evolutionary networks. The first is the issue of analogy and homology, which will be taken up in the following section, and the issue of historical dependence and independence, which will be taken up in the next chapter.

The problem of analogy and homology; has received some debate in anthropological theory, and centrally concerns the problem of comparison and control. (See, for instance, Richard A. Gould and Patty Jo Watson "A dialogue on the meaning and use of analogy in ethnoarchaeological reasoning" in Journal of Anthropological Archaeology I: 355-381, and Alison Wylie, "The Reaction against Analogy" in Advances in Archaeology: Method and Theory, vol. 8, 1985: pp. 63-111.) To what extent can analogy inform us about our past, and what is the role of analogy in reconstruction, and then how much is our science predicated upon the inferences of homology?

There are few textbook cases of an evolutionary tree of unmuddled divergence that would make a classic case for homology. One of the few is the history of Polynesia Proper and the strong congruence between archaeological and lexicostatistical data showing clearly the evolutionary pathways of the human settlement of this area. On the other hand, there are few clear and uncomplicated cases in which the reverse case of analogy in the effect of borrowing upon a single language. One of the few of these that stands out is the history of English and its several successive phases of borrowing from outside languages.

The historical reality of most languages, cultures and peoples, if we go back far enough, is one of a mixed analogical/homological patterning. The further back one traces one's ancestry, the stronger the probability will become of analogical borrowing which created the basis of the root stock of the complex in question. The point is that we can rarely clearly and cleanly separate the problem of convergence by borrowing and divergence by endogenous change, and in most cases of complexes of characters, this is usually the case. Even in cases of individual characters, we may have the dual effect of both borrowing and refashioning that accompanies inheritance and reproduction.

We are left with several basic principles that we must contend with in our model building and reconstructions. The first is that we are always dealing with polythetic sets, asymmetrically stratified samples, fuzzy boundaries and prototypical effects that must always be somehow accounted for in our formulas of relative distances. Second, space and time may not necessarily be directly translated into one another, such that distances in space may not be equivalent to actual temporal distances, and vice versa, spatial proximity may not be the equivalent of temporal shallowness. The consequence of this is that we are dealing with a ratio variable of space/time. For every quantum of space we are dealing with, there is a corresponding nonequivalent quantum of time that must be inferred. Third, we must deal with the conditionality of analogous/homologous relationships such that for any given character, set or complex of characteristics, we have a probabilistic ratio of the relative analogous versus homologous features to contend with. Fourth, neither analogy nor homology are defined clearly in terms of either identity or difference, but also by a ratio of the degree of similarity/difference. Likewise, we can imagine forms of relationship that are inherently intermediate. Borrowing is a kind of homology that has analogous implications, and convergent evolution is a kind of analogy that has homologous implications.

Though all this seems to run against the grain of simplifying solutions of our science, there may be a measure of sanity in our approach to the inherent chaos and complexity of our history. In modeling our evolutionary networks, we can simulate the consequences and test the effects of various combinations and sets of presumptions and inferences that dictate alternative pathways through the system. Furthermore, we can minimally anchor such a system to a finite data-base derived from absolute archaeological data by which we can perform our tests and increase the precision and plausibility and probability of our inferences.

Southeast Asian Sources: Critical Essays in Culture Historical Reconstruction