The Austronesian languages are sub-divided into a number of smaller families, just as Indo-European is divided into Indo-Iranic, Hellenic, Armenian, Germanic, Celtic, Italic, and so on. The principal division is between the Formosan languages (the indigenous languages of Taiwan) and the Malayo-Polynesian languages, which are the languages found outside of Taiwan. Malayo-Polynesian is by far the widest-spread family of Austronesian, and was the most widely distributed language family in the world in pre-modern times, found from Madagascar to Rapa Nui, and from Hawai'i to New Zealand. Indonesia and the Philippines are the countries with the largest numbers of Malayo-Polynesian-speaking people.
'Formosan' is an umbrella term for several first-order sub-families of Austronesian; there is disagreement about their number (Robert Blust, if I remember correctly, says that there are nine Formosan families), but Paiwanic, Atayalic, Rukai (the most divergent extant Austronesian language), and Tsouic are some of the most realistic candidates. These are all sub-families of the same order as Malayo-Polynesian, which includes all of the Austronesian languages outside of Taiwan, from Indonesian and Tagalog to Hawaiian and Malagasy. The diversity on Taiwan is a key part of the evidence for an Austronesian origin there. The consensus is that proto-Austronesian was spoken in Taiwan or southeastern China around 3500 BCE, and that it spread south, in the guise of proto-Malayo-Polynesian, about five hundred years later. By 2000 BCE, speakers of Malayo-Polynesian languages were almost certainly present throughout the islands of Indonesia, probably moving to the area in different waves associated with the Bornean languages and the Nuclear Malayo-Polynesian languages.
Malayo-Polynesian breaks down into several second order sub-families, most of which have been disputed at some point. It was formerly divided into 'Western', 'Central', and 'Eastern', reflecting the distribution of the languages (i.e, 'Western' in the Philippines and western Indonesia, 'Central' in eastern Indonesia, and 'Eastern' in the Pacific), but these terms have now been discarded by most linguists and replaced by, among others, Philippine languages, Sama-Bajaw languages, and Nuclear Malayo-Polynesian. Nuclear MP contains Central-Eastern Malayo-Polynesian (CEMP) as a further sub-family. CEMP contains all the Polynesian languages, in addition to most of the languages of eastern Indonesia.
I am particularly interested in CEMP languages, especially those languages formerly labelled 'Central' MP (CMP). That is to say, those languages spoken exclusively within the Indonesian provinces of Nusa Tenggara Timur and Maluku, both in southeastern Indonesia. These are some of the poorest provinces in modern Indonesia, which is partly due to their dry climate, low population density, and the fact that they have relatively few salable natural resources.
|A map of the Nuclear Malayo-Polynesian languages - slightly out-dated, as it divides into Western MP, Central MP, and so on, which are now not accepted as valid clades (but a good map nonetheless!). The gold ring indicates the area in which CMP languages were once considered to be spoken; that is the area I am talking about here. Note the strange distribution on Alor and far eastern Timor - that is where non-Austronesian languages continue to be spoken, although I think CEMP languages are still in the majority there. (h/t: Wiki user: Kwamikagami)|
This is a contiguous area stretching from Flores to Aru (fairly near coastal New Guinea), and is noted by ethnographers for the overlapping cultural attributes found there in early- and pre-colonial times. Among these are:
- Headhunting, as in, raids conducted at the end of the dry season for the purpose of taking heads, animals, slaves, and valuables, with an overt religious purpose associated with the female property of fertility and the male property of esteem/status
- Recursive-complementary dualism applied to gender, religion, aesthetics, headhunting, textiles, the sexual division of labour, relations between hinterlanders and coast-dwellers, and much else (more about this later)
- Diarchic (as opposed to monarchic) socio-political leadership
- Sacrificial poles, typically in the centre of the village, and typically shaped like a fork
- Ancestor-focused religion (in addition to ancestor-focused descent groups, both matri- and patrilineal, to use Robin Fox's system of classification)
- Boat symbolism for the house and village
- Poetic parallelism (the use of paired speech to enhance a point in oral poetry and oratory)
- Some form of cross-cousin marriage, usually matrilateral cross-cousin marriage, whereby men would marry one of their classificatory mother's brother's daughters (MBDs - more about this later, too!); this is closely related to the principle of recursive-complementary dualism
- Bridewealth thought of as payment for the offspring of a marriage
- Defensive village architecture, likely related to the prevalence of headhunting, and including both stone and organic walls, and ladder-access to hill- and ravine-top village sites
Those in the know might compare this list to those previously offered by others, including F. A. E van Wouden (whose Sociale Structuurtypen in de Groote Oost (1935) is the original statement of eastern Indonesian socio-cultural unity), J. P. B de Josselin de Jong, and James Fox. Mine is, of course, far from complete, and is going on what I can drag up from memory. Either way, this part of Indonesia is curiously uniform in certain respects (there is naturally plenty of variation, but most of it occurs within common parameters). That's part of what makes it interesting.
CMP is no longer considered a valid classification - the problems of CMP have been obvious for decades, with plenty of linguists taking the seemingly valid stance that CMP languages were unified more by contact with non-Austronesian languages than by anything. Far from sharing a proto-language, they actually shared nothing more than a similar substrate effect (making CMP a still-potentially useful clade from an ethnologist's point of view).
In fact, there are still some extant non-Austronesian languages spoken in this part of the world, including Fataluku in East Timor and Abui and Klon on Alor, and while these show no demonstrable relationship to any other languages in the region (there have been some attempts to relate the Timorese languages to the Trans-New Guinea phylum, but these are uncertain), it seems reasonable to conclude that they were present in eastern Indonesia before the arrival of Austronesian speakers, especially in the context of the population genetic evidence (which shows that haplogroups associated with the spread of Austronesian languages are in the minority in eastern Indonesia).
This CMP=substrate stance has been rejected in turn, and replaced by the placing of the CMP sub-families inside a larger grouping with the 'Eastern' languages, Chamorro, and others. This is largely on the basis of findings from 'big data' analyses of Austronesian languages, primarily the Austronesian Vocabulary Database.
So now we have CEMP, which includes all of the languages of Oceanic - i.e, all of the Micronesian, Melanesian, and Polynesian languages - as a sub-category. I lifted this diagram from Wikipedia for ease of understanding (user: Kwamikagami seems to be the one to thank for this neat image):
| Central–Eastern |
As you can see, there is a Greater CEMP family, including both Chamorro and Palauan, spoken in the Pacific. These were formerly included in the 'Western' branch, but are no longer. Note also Eastern MP, which includes two sub-branches: the geographically extremely circumscribed Halmahera-Cenderawasih languages, spoken on Halmahera (North Maluku province) and around Cendrawasih bay, in Indonesian New Guinea (incidentally, cendrawasih means 'bird of paradise'); and the Oceanic languages, spread over an absurdly wide area from New Guinea to Rapa Nui. This placement of the Oceanic languages accords fairly well with the population genetic evidence, I think.
Now look at the rest of the tree. The languages formerly covered by 'CMP' are actually sub-divided into several sub-families ('Timor-Babar', for instance), which is troubling for ethnologists. Actually, most of these sub-families already existed and were uncontroversial, but it's troubling that they relate to one another in the way that they do. Look at the Sumba-Flores languages stuck way out there like that! Sumba and Flores show so many cultural and socio-political similarities (i.e, the above-mentioned ones) to Timor, Sawu, and Maluku Tenggara that it's hard to believe their languages are less closely related than are Timorese languages with Hawaiian and Fijian. That's a bit of a startling find. But it seems to be the case.
This is troubling because all of that cultural unity seen in eastern Indonesia has to be the result of more recent sharing, rather than being due to derivation from a common core associated with speakers of a proto-CMP language. There's no proto-CMP to reconstruct and no common set of traits to be associated with it. This seems like it's a problem for understanding eastern Indonesian prehistory, and indeed it has been. But if you consider the rest of the tree, we clearly have some license to use Polynesian, Micronesian, and Melanesian evidence to better understand Indonesia, because eastern Indonesians and Oceanians do share a common ancestor. The common ancestral language can be reconstructed (proto-CEMP), and we can use Oceanic ethnographic and ethnohistoric data to make better sense of eastern Indonesia data (and vice versa). That's important, because Oceanic speakers had separate cultural trajectories to eastern Indonesians. We can also look at evidence from western Indonesia (say, from Toba Batak communities), and perhaps make a stab at reconstructing Nuclear MP life.
So it's not so bad. And, in any case, the fact that eastern Indonesians share such cultural systems because of more recent contacts is an interesting discovery in itself, and points to something already known by archaeologists and ethnohistorians: there was once a great deal of trade and other intra-regional contact in eastern Indonesia.
A good example of the mutual utility of eastern Indonesian and Polynesian ethnographic data is Patrick Kirch's use of data on temples from Flores to better understand Polynesian religious change over the centuries (specifically, the way in which the earliest Polynesian temples seem to have been situated on sites formerly occupied by domestic residences; see this quite-good book for his article on this). We can look at evidence from throughout the CEMP world, stretching across the earth from Flores to Hawai'i, and make all sorts of fascinating inferences about what these people share, why they share it, and what ancient CEMP life was like. I think that's great.
The Oceanic evidence does have to be used with significant provisos. It is clear that there were massive changes that occurred in Austronesian-speaking societies as they entered the wider Pacific, and these changed attributes were inherited by later Polynesian groups. The abandonment of pottery by late-Lapita/ancestral Polynesian groups is one of these changes. But, used with caution, these data can tell us quite a bit about what ancient eastern Indonesians once said and did, so it would be wrong to ignore them.
In the next post in this series, I'm going to look briefly at the non-Austronesian languages of eastern Indonesia and examine the possible impact of their autochthonous speakers on the developments of eastern Indonesian culture. This is also a good opportunity to involve the genetic data, of which there is a small but tantalizing quantity.