wgdavey@zianet.com

Our primary finding is that languages across the World that are remote geographically, racially, historically, technologically, and linguistically have similar names for parts of the body. Thus they must be derived from a single older language, the "Mother Tongue". Our second important finding is that the languages of the Americas are strongly linked with all the languages of the Old World. In this sense the Americas are not remote lands but one of the most important factors of world languages. These are the major conclusions from a study of the names for one hundred parts of the body or body secretions in a massive database of over 750 languages across the World. The names of bodyparts were chosen because of the simple fact that all peoples have the same bodies wherever and however they live.

A point to emphasize is that the similarity in names is a simple and obvious fact, not a matter of speculation. The writer is not a linguist and the reader does not need to have the specialized knowledge of linguists. In brief, if names sound the same then they probably are the same or "If it looks like a duck, walks like a duck and sounds like a duck then it is probably a duck".

We anticipate that there will be many skeptics regarding our claim for worldwide similar names and so we immediately present two of the simplest and clearest examples of the thousands of lists of links that we have discovered.

The first example shows the remarkable spread of names for Breast of the form "T-T-". For those unfamiliar with Greenberg’s analysis of American languages, Na-Dene, Almosan-Keresiouan, Macro-Carib, Macro-Panoan, and Andean are all American language groups. This Table shows the possible variations in spelling and also shows the widespread name of TITI that may thus be the oldest form of the name.

The second example is equally remarkable but of a completely different form. Here we show where there are pairs of complex names found in widely separated regions. Here we have arranged the data to show some immediately clear conclusions. In particular we have placed the "African" ME or Middle East and the single S.A. examples in the left hand columns since there are clearly implications regarding the "Out of Africa" thesis. We have also placed the American (Am or Amer) examples on the right. These data are clearly worthy of further analysis in conjunction with many others of our studies but the dominance of the Middle East and the Americas is immediately obvious.

Collecting and ordering the names so that they could be conveniently analyzed was an onerous task that occupied several years. It was greatly aided by following an orderly approach and constructing computer-based (Microsoft Word) files. It consisted of several parts, first collecting the initial vocabularies in "Source Lists", then collecting the names for each bodypart in "Primary Lists", and then grouping the information in each bodypart Primary List to group names that began with the same initial letter forming our "Working Lists".

Source Lists: The first task was to collect the initial "Source Data" for each suitable language. Here we listed the available names for 120 different parts of the body in alphabetic order. Only a small number of vocabularies exist in widely available formal dictionaries and most of this information exists in university libraries in specialized journals that are linguistic or anthropological in scope. The vocabularies were very varied in completeness and different names were sometimes found from different sources. Where subtleties of pronunciation were present they were necessarily ignored and the names were recorded as the best "reasonable" English spelling.

From these Source Lists it was clear that some names were infrequent and a list of the most usual one hundred body parts was selected. All the names were then collected in one hundred lists for each part. Here names were grouped according to similar linguistic and geographical associations.

It should be noted that these files are "open ended" since additional vocabularies can be (and have been) added as desired though carrying all new data forward into other files is a formidable task.

Primary Lists: The hundreds of Source Lists in themselves are inconvenient to use directly and it is necessary to collect all the names for a given bodypart so that they may be conveniently compared and analyzed. Thus we have constructed 100 "Primary Lists" where all the names for a bodypart, say "arm", are brought together. In these lists we group all the names for each of our selected regions, say Northern Eurasia, in the order we have listed above. Thus, for example, all the Indo European names are collected together and readily compared. Similar comparisons can be made for other regions and language groups, for example for "Austronesian" languages of South East Asia.

Working Lists: Inspection of the Primary Lists immediately showed that there were many different names for every bodypart and so are not suitable for seeing if there are similar names in different languages. Consequently we re-grouped all the entries in the Primary Lists according to the initial letter in the name. This was a very tedious but essential task and these are the most important lists in our study since they contain the specific information used in all of our studies.

Here we emphasize that it is the initial consonant that is used, not any initial vowel that might be present. This is because our many studies of words have shown that it is overwhelmingly true that similar words are defined by the sequence of consonants, and that the precise vowel that completes the syllable is usually of little importance. On the lighter side, these facts may be some support for the humorous, if sarcastic, comment attributed to Voltaire that etymology is a science where "the consonants count for little and the vowels for nothing". We agree on the vowels but not on the consonants.

Despite this emphasis we must admit that in some cases the similarity in vowels is so striking that we have included vowel similarities in some striking cases.

Under a given initial letter we have ordered the names according to linguistic groupings so that, for example, all Indo-European names beginning with "k" are listed together.

Here we have found that it is not necessary to list according to every letter of the alphabet and that certain initial letters are more frequently found than some others are. This is a convenient point to recognize the probable linkage between some letters, and we use the well-known fact that when names change they tend to do so in established ways. For example "t" preferably changes to "d" (and vice-versa) rather than to another letter. We also find similar changes between "p" and "d", "m" and "n", and some other letters. In brief we collect the names in thirteen groups where we list the names beginning with "t" and then "d", "m" and "n", "k" and "c" and "q" all pronounced as "k", "g", "c" or "ch", "p" and "b", "s" and "z" grouped together, "l", "r", and finally "j" and "y" grouped as a pair.

Further Grouping of Words; Doubles, Triples and Quads:

Having all names for a given bodypart in a single list is convenient but this is still not sufficient to show the close similarity of many names. For example some lists are cumbersome and unclear and in one extreme case we see that there are 170 names for "nose" that begin with the letter "N". Consequently the last task is to collect together names where the second, third and even fourth consonant (or syllable) is the same.

Thus at the end of each "working list" of identical first letters we collect the examples where the first two consonants are the same, then, as appropriate, we separately collect those with three or even four similar consonants. In short hand terms we describe these as "doubles", "triples", and "quads". As expected the lists become progressively shorter as the names become more complex

Taking the instance of 170 examples of "initial N" for Nose, in about 40 of these the initial "N" is followed by "S" or "Z", while there are three groups of about 20 doubles where the initial "N" is followed by "K", "N", or "R". In other words, roughly half of the names are of a "two-syllable" form, either "N-S", "N-K", "N-N", or "N-R". And within each two-syllable group, there are between three and seven examples of "three-syllable" names of the form "N-S-K", "N-K-R", and "N-R-S". In this case these are no "four-syllable" examples.

As we have noted, the basic lists used in our analyses are the "doubles". The numbers of such doubles can vary considerably for different bodyparts. In part this is because certain vocabularies which have a sexual connotation are not known. This may be because some of the earlier researchers may have avoided these topics and also perhaps because women might have a reluctance to discuss these with the male investigators. However, even the shorter bodyparts lists contain more than 20 items and many others have 70 or more. The total number of such links is about 5400 or an average of about 50 per language. Inside these lists are about 1600 triples. In other words, there are about 5400 lists of links that show how the languages of the world are related.

As can be seen the total amount of information contained in these lists is very formidable; in computer terms it is over 100 Megabytes. The reader must appreciate the strength of using a computer to store the information so that one may shuffle the names essentially as one wishes and group and re-group the data as appears necessary to address a given question. However it cannot be emphasized too strongly that the analysis of these data is not simply a matter of "feeding it into the computer". In the writer’s experience there can be major mistakes in uninformed and uncritical acceptance of computer-based data "analysis" simply because the data can be presented in an sophisticated manner. There is no known method of substituting such approaches for good judgement.

Basic Principle: All of our analyses begin by collecting all "doubles" names (that automatically include any Triples or Quads) for all bodyparts and a given language or a given region in a single document. Thus we produce documents entitled as "All Basque Links" or "All Papua Links" that contain all such connections worldwide. Such a procedure would be impracticable without computer-based storage of the data, and even so it takes several days to compile a single "All Language" list and far longer for regional lists.

These documents are analyzed in many different ways depending upon the issue or question that is being addressed and interesting and/or significant points raised by the content of the lists themselves. In general each "All Language" list is used in many different studies.

We have found it useful to summarize the broad linkage of a single language or region in these documents in a semi-standardized listing of the number of links with the major regions (discussed later), that is with the Americas, SEAS, Australia, Papua, the Mid East, and Eurasia. The Americas links are usually complied from those with each of the thirteen language groups defined by Greenberg ("Language in the Americas"). In these lists the total number of links is given but we also make an allowance for the differences between the regions by dividing the total number of links by the number of languages in each region. This is not an unambiguous procedure and is discussed in detail in several of our studies.

Comment on Data: The use of computer based files and the massive presentation of linked names is not simply a matter of "overkill" in showing worldwide linkage between languages since the large number of examples allows analyses that are not possible with a small sample. In particular the magnitude of the links is vitally significant in such issues as the strong linkage of the Americas to all of the languages of the Old World. We believe that the statement that there is a 60% or 70% linkage of a given language with the Americas has far more significance than "there are strong links".

Very importantly with such a large set of data we can compare and analyze linkages between entire regions of the world. For example we have listed all links between Africa and the entire rest of the World, and we have also listed all links of the Americas with the rest of the world. In this manner it is possible to show how languages are linked across the World and address the questions of large-scale migration in a comprehensive manner.

This statement of the important aspects of worth of such a large set of data should not be taken as a statement that we have resolved or understand the overall picture presented by languages. We have cast some light on certain issues, particularly the roles of Africa and the Americas, but in general we have been able to better present the questions rather than supply the answers.

Finally, we cannot emphasize too strongly that we believe that publication of the source data and the hundreds of derived documents is beyond presentation in simple paper form. Apart from the sheer bulk (literally many hundreds or even thousands of pages) of printed pages, even if the printed data was available it could not be usefully viewed, reviewed, and analyzed with printed sources. Useful publication of anything other than summaries is only likely to be practical by including computer readable files together with the printed text.

RETURN TO CONTENTS