Published online by Cambridge University Press: 22 April 2026
It is the purpose of this paper to present some of the results of a statistical comparison of the basic vocabularies of a collection of Malayopolynesian languages. By reason of the very number of its members, the Malayopolynesian family has great importance among the language families of the world. A fair estimate would place the number of Malayopolynesian languages at about 500. If the number of the languages in the world is set at about 4000, then about ⅛ of all the world's languages are Malayopolynesian.
1 This paper, with various modifications, was presented at the 25th International Congress of Orientalists in Moscow (1960) and at the Tenth Pacific Science Congress in Honolulu (1961).
2 Although the main support of this project has been the Tri-Institutional Pacific Program, additional support was received from the National Science Council, the MIT Computing Center, and the Southeast Asia Program of Yale University. The contributors to this project have been so many that the list would be disproportionate to the size of this article. Most important among them is George Grace, who collected a large amount of material in the field, mainly from Melanesia. Others—teachers, missionaries, anthropologists—have sent in materials from Melanesia and elsewhere, all to be processed by my secretary, Mrs. Joanne Cardona, whose devoted effort sped the work immeasurably. I should mention here the contributions of the members of the Summer Institute of Linguistics to the material on the Philippine languages. I hope to list the other contributors in a more complete publication of the results of this project.
3 For an elaboration of this theory, see I. Dyen, Language Distribution and Migration Theory, Lg. 32.611–26 (1956).
4 Lexico-statistic Dating of Prehistoric Ethnic Contacts, PAPS 96.452–63 (1952).
5 An IBM 704 Program for Computing Lexico-statistical Indices (mimeogr., 1960).
5a Except for the possibility of an undiscovered borrowing; see above.
6 It should be noted here that the Chi-square test is not applied directly to the percentages, but to the observed number of positive and negative cognate pairs. The use of the Chi-square test was suggested by A. T. James of Yale University, a mathematical statistician. He nevertheless has reservations as to its applicability to our list comparisons because the test assumes what is not true, namely that the probability of cognation is the same for each pair of words in a pair of lists. Further investigation is now going on into the validity of using the Chi-square test. (The heterogeneity of the lexicostatistical list is being investigated by a combined Chi-square test at the 5% level on thirds of the list with meanings rearranged in a descending order of frequency of cognate pairs; in one of about 30 instances tested thus far the heterogeneity was barely significant, whereas in the remainder it was less than significant.) The linguistic comparatist may nevertheless be satisfied with the crude application of the criterion of a 10% difference; some notion of what such a difference implies can be obtained from the fact that it means an excess of 15 cognates if each of the two pairs of lists have 150 comparable pairs of words and a corresponding difference in other cases.
7 I use here the convention of simply citing the languages A and B side by side to refer to the outcome of comparing the list of language A with that of language B.
8 This is to be regarded as a rule-of-thumb. It does not replace the actual application of the Chi-square test.
9 Atlas van Tropisch Nederland 9 (The Hague, 1938).
10 G. W. Grace, The position of the Polynesian languages within the Austronesian (Malayo-Polynesian) language family (IJAL Memoir 16, 1959).
11 Journal of the Polynesian Society 69.180–4 (1960).
12 This new term has the advantage of avoiding the use of the term Indonesian, which is currently used by some for the official language of the Republic of Indonesia beside the other names Bahasa Indonesia and Bahasa.