Malagasy origins

5 minute read

The natives of Madagascar have long been known to have roots in Southeast Asia. Their native language is a derivative of the Austronesian language family, and Despite the fact that the island is geographically close to Africa, some cultural elements show a clear link between Malagasy and the indigenous peoples of Indonesia, especially the language, which is a derivative of the Austronesian language family, which includes the languages of Polynesia. Archaeology shows that the first settlement of Madagascar occurred within the past 2000 years, and much of the native faunal diversity of the island crashed by 1500 years ago. Anthropologists infer that the habitation of the island involved long-distance ocean transits from island Southeast Asia. But the economic and material elements of the island's culture, especially cattle pastoralism, come from nearby Africa. Thus, the island's history is a mixture of different influences.

A recent study of Malagasy Y chromosomes and mtDNA is by now old news, but it is a nice example of admixture analysis using haplotypes. It also illustates the pervasive mixture of peoples, even across oceans, through mating and trade networks. There is a press release from the Wellcome Trust on the study, and an article by Rhett Butler (yes, Rhett Butler) of, that comments on the diversity on the island:

Within Madagascar there is a great deal of variation between ethnic groups from the Indonesian-looking Merina in the highlands to Arabic Antaimoro on the eastern coast to the African Sakalava in the south and west of the country. Despite these differences, the Malagasy language is spoken throughout the country, something which is a bit surprising given the size and ethnic diversity of the island. Sharing a common language is the strongest bond between Malagasy -- people's physical appearance, religious practices, and traditions are highly regional.

The new paper does not question the origins of this ethnic and biological variation; it quantifies the relative paternal and maternal contributions of different source populations to the present Malagasy. From the conclusion:

The above analyses demonstrate that we can consider the Malagasy to be an admixed population derived from two ancestral populations, one African and the other Indonesian; we can now estimate the admixture proportions of these populations. The mutual exclusivity of Y-chromosomal and mtDNA lineages between these two ancestral populations means that we can obtain a point estimate of admixture proportions simply by counting lineages. Of Malagasy mtDNA lineages, 38% (14/37) can be traced to Africa, whereas 51% (18/35) of Y-chromosomal lineages have an African origin. This increases to 55% (18/33) when the two putative recently admixed Y chromosomes are removed (Hurles et al. 2005:899).

The Y chromosome haplotypes with Asian affinities were most similar to others sampled from Borneo, confirming linguistic similarities between Malagasy and some Bornean languages.

The paper also suggests that populations like the Malagasy may facilitiate gene mapping projects:

Admixture between two highly differentiated populations generates long-range allelic associations that decay over time (Chakraborty and Weiss 1988). The amount of linkage disequilibrium (LD) exhibited by an admixed population depends on a number of factors, including proportions of admixture, differentiation between ancestral populations, time since admixture, and demography (Pfaff et al. 2001). It has been proposed that it will be possible to efficiently map genes underlying complex traits by focusing on association studies in admixed populations, and a range of potentially informative populations has been identified (Halder and Shriver 2003). Although the time since admixture in the Malagasy is comparatively long, the high degree of differentiation between the two ancestral populations and the even balance of their contributions suggest that excess LD might still exist. Admixture mapping of genes underlying complex traits is predicated on the observation that the trait itself is differentially manifested among the ancestral populations. Therefore, although most attention has focused on European and African admixture in African Americans (Halder and Shriver 2003), it would be of interest to identify a range of admixed populations that are derived from various different ancestral combinations. The admixture of Indonesian and African lineages present in the Malagasy may be uniquely informative. Further characterization of LD in the Malagasy will be necessary for determining whether the Malagasy can be added to the list of admixed populations suitable for the identification of genes underlying complex traits that are of interest to anthropologists and medical geneticists alike (Hurles et al. 2005:900).

I for one find this passage creepy. Imagine: there is a "list" of admixed populations that are of interest to medical geneticists (I disavow any interest by anthropologists) because of their possible utility in identifying genes underlying complex traits. The purpose of such research is to help people in Western societies by developing pharmaceutical products for sale at high prices. It is the unique history of this population that makes them useful in this way. Thus, geneticists are looking for ways to profit off the history of an entire "list" of indigenous and diadvantaged peoples.

The truth is, from a genetic perspective, there is nothing especially problematic about mapping traits, there is probably no substantial profit to be gained from it, and it is not that much of a timesaver compared to just mapping them in American populations where informed consent rules have some teeth. So why does a geneticist include a passage like this one, with all its apparent anthropological problems? Because geneticists are used to applying for grants, in which they put forward the most practical medical applications of any resulting knowledge, no matter how minor. I think it's time for a reality check: history is history, and medicine is medicine, and mixing the two is an explosive cocktail of misinformation, misgivings, and misapprehension.

The release raises an interesting question with no answer:

But why, if the population is a 50:50 mix, is the language almost exclusively derived from Indonesia?
"It is a very interesting question, for which we have as yet no certain answer, as to how the African contribution to Malagasy culture, evident in biology and in aspects of economic and material culture, was so largely erased in the realm of language," commented Professor Robert Dewar, of The McDonald Institute for Archaeological Research, University of Cambridge. "This research highlights the differing, and complementary, contributions of biology and linguistics to the understanding of prehistory."

This is a critical cultural question that is usually not asked by genetists. Humans may adopt cultures and languages for many reasons, ultimately intrinsically cultural themselves. Language groups cannot be assumed to be equivalent to biological origins, and they clearly are not in this instance. Many genetic studies make the error of assuming that languages and populations are concordant, so studies like this one are a good reminder that the assumption isn't true.


Hurles ME, Sykes BC, Jobling MA, Forster P. 2005. The dual origin of the Malagasy in island Southeast Asia and East Africa: Evidence from maternal and paternal lineages. Am J Hum Genet 76:894-901. Full text online