Thursday, December 3, 2009

249. The Baseline Scenarios -- 25: Hunter-Gatherers

The National Geographic article on the Hadza, by Michael Finkel, is accessible in full via the internet at the following address:
http://ngm.nationalgeographic.com/2009/12/hadza/finkel-text. The subtitle is misleading, since the bit about the Hadza "living a hunter-gatherer existence that is little changed from 10,000 years ago" refers simply to the general estimate among Archaeologists that the earliest signs of agriculture date from roughly 10,000 years in the past. The assumptions being that 1. the Hadza have lived as hunter-gatherers during the entire period since the earliest advent of agriculture; 2. hunting and gathering in itself is the principal determiner of culture, and 3. all hunter-gatherers are now living as our ancestors lived -- these are huge assumptions.

Nevertheless, we have good reason to believe that certain aspects of Hadza life haven't changed much over a far longer period -- anywhere from 40,000 to very possibly well over 100,000 years ago. How can we make such a statement? Not on the basis of assumptions, but evidence pointing to strong socio-cultural affinities with HBC, the ancestral culture we've been taking as our (hypothetical) baseline.

We learn toward the beginning of the article that one of the Hadza males is "maybe five feet tall." The mean height of Hadza males has been estimated at 162.24 centimeters, i.e. roughly 64 inches or 5 foot 3 (see table in How universal are human mate choices? Size does not matter when Hadza foragers are choosing a mate, Sear and Marlowe, p. 608). Since Aka Pygmy males are reported as averaging 153 centimeters, or roughly 5 feet, the Hadza don't seem all that much taller. And as we've learned, certain Bushmen groups fall into roughly the same ballpark. Did the Hadza, Pygmies and Bushmen inherit their small height from a common ancestor or is it some sort of adaptation, or even perhaps a coincidence? Only biological research can tell us for sure, so this remains an open question.

The Hadza language is correctly described as an isolate. Despite the use of clicks it has not been possible to group it with Khoisan or any other family. According to Finkel, "Gene­tic testing indicates that they may represent one of the primary roots of the human family tree—perhaps more than 100,000 years old," but I don't know of any research supporting that. On the contrary, the Hadza appear to be a genetic isolate. Consider the following graph, based on an analysis of autosomal microsatellite markers by Tishkoff et al (from Supporting Online Material, figure S12):


(For an enlarged view, right-click and select "Open in new window".) The Hadza are easy to find as they are represented by a bright yellow band roughly one-third of the way in from the left. The fact that they are the ONLY group represented by this color tells us how distinctive their genetic profile is.

The Hadza are a musical isolate as well, as far as I've been able to determine, with a vocal style completely different from that of the Pygmies and Bushmen, and also very different from the call and response pattern so typical of Bantu vocalizing.

[Added 12-6: Recordings of Hadza music were released a few years ago on two CDs, The Hadza Bushmen of Tanzania. Clips from all tracks can be previewed at the Amazon.com web page by clicking on the above link. Volume 1 is limited to solo vocals and instrumentals, but Volume 2 contains several excellent recordings of traditional group vocalizing. In a comment to this post, Maju provides links to a couple of interesting youtube videos of Hadza people singing and dancing: one -- two. The youtube clips sound like typical East/South African call & response Bantu style singing, and the dancing strikes me as rather tentative and half-hearted. Looks to me like a show put on for tourists, based more on Bantu traditions than Hadza ones, which tourists might find puzzling. The recordings on the CD set, especially those on Volume Two, are very different, with much less emphasis on call and response, and with long intricate melodies and drawn out, sustained tones, very unusual for Bantu Africa.]

While there is indeed something very mysterious in this picture, it could be explained, I believe, by a severe population bottleneck that could only have occurred at some very early point in their history as a distinct group. Such a bottleneck, leaving only a very small population to serve as a "founder group," would explain the genetic anomaly as well as the musical one -- the highly group oriented, cooperative style of P/B could have been lost in the wake of a disaster that temporarily isolated a small number of survivors. If such a bottleneck event occurred relatively recently, the genetic picture for the Hadza would not look so different at so many values for K (see graph). This is my own interpretation of the genetic evidence, and I am by no means an expert, so it could be totally wrong, but as far as I can tell a very early bottleneck seems to be the only logical explanation for the Hadza "mystery."

When we leave aside the genetic, linguistic and musical mysteries, we see a great many striking similarities with Pygmy and Bushmen culture, strongly suggesting survivals from HBC. For one thing, we learn from the National Geographic article that a significant number of Hadza, one-quarter, "remain true hunter-gatherers," with strong evidence that all Hadza lived solely by hunting and gathering in the not too distant past.

Anthro­pologists are wary of viewing contemporary hunter-gatherers as "living fossils," says Frank Marlowe, a Florida State
University professor of anthropology who has spent the past 15 years studying the Hadza. Time has not stood still for them. But they have maintained their foraging lifestyle in spite of long exposure to surrounding agriculturalist groups, and, says Marlowe, it's possible that their lives have changed very little over the ages (p. 104).

"The Hadza do not engage in warfare," "have plenty of leisure time," "live almost entirely free of possessions," "collect honey," and hunt with poison arrows, cultural characteristics shared among all three of our baseline "feeder" groups, EP, WP and Bu. Additionally, they "recognize no official leaders," favor "individual autonomy," and rarely stand on ceremony for events such as weddings, funerals and other rituals. In this latter respect they do differ somewhat from HBC, where we would expect to find rituals such as the Elima girl's initiation, or the Molimo festival, as practiced by the Mbuti, or the Xhoma male initiation ritual of the Ju/'hoansi Bushmen. Moreover, "there are no Hadza priests or shamans or medicine men." Perhaps such traditions were lost in the same bottleneck that affected their genetic profile and their music.

Unlike most Pygmies and Bushmen, who tend toward monogamy, the
Hadza rarely marry for life: "Most of the Hadza I met, men and women alike, were serial monogamists, changing spouses every few years." Again, the relative looseness of marital relations could possibly be due to an extended period early in their history when social norms may have broken down.

"There was a bevy of children in the camp, with the resident
grandmother, a tiny, cheerful lady named Nsalu, running a sort of day care while the adults were in the bush. Except for breast-feeding infants, it was
hard to determine which kids belonged to which parents."

This approach to raising children, with involvement by individuals other than the parents is not uncommon for most Pygmy and Bushmen groups and may well have been a feature of HBC child-rearing.

"Gender roles are distinct, but for women there is none of the forced subservience knit into many other cultures. . . Among the Hadza, women are frequently the ones who initiate a breakup—woe to the man who proves himself an incompetent hunter or treats his wife poorly."

The status of women among the Hadza appears extremely close to their status among EP, WP and Bu -- and thus HBC -- i.e., a state of relative equality, with very clear overtones of subservience in certain areas, as reflected in sometimes very different gender roles and limitations.

We have already discussed Hadza "beehive" huts: "During the rainy season, they construct little domed shelters made of interwoven twigs and long grasses: basically, upside-down bird's nests."

Core HBC values such as avoidance of conflict and strong emphasis on sharing are also present among the Hadza: "Most conflicts are resolved by the feuding parties simply separating into different camps. If a hunter brings home a kill, it is shared by everyone in his camp."

I was especially intrigued by the scarification on the forehead and nose of this Hadza woman (see photo on p. 119):

For one thing, it tells us that the Hadza too, employ scarification, another link with HBC. Moreover, the scarification pattern reminds me of a very similar type of linear design found very commonly in cave art from many parts of the world. For example:


Gargas Cave, French Pyrenees


Llonin Cave, Spain



El Pindal, Spain


Koonalda Cave, Australia

Experts have long puzzled over the meaning of these lines, and I'm wondering what meaning this particular scarification pattern might have for the Hadza. I'm not claiming to have uncovered a connection, but who knows -- if such patterns were commonly used by HBP, and had some sort of meaning for them, then it's possible they could have had more or less the same meaning for their rock-painting descendents. Just a thought -- but worth looking into, I should think.

14 comments:

Maju said...

Just checking Wikipedia fast for the haploid genetic data of the Hadza: they seem most related to Pygmies (mtDNA L2a1, Y-DNA B2b) and Sandawe (mtDNA L4g), with a quite strong male input from Bantus (E1b1a, c. 30%).

A further Y-DNA lineage E1b1b1g is mostly restricted to East Africa but being part of E1b1b1, the same large haplogroup that is found through the Mediterranean and being at lower apportions among the Hadza-Sandawe composite than among their agriculturalist neighbors, guess that can be considered as some sort of "Nilotic" or otherwise East African specific marker but with a "recent" arrival date (LSA or Epipaleolithic, I guess). Not sure if it's present among the Hadza or it's just a Sandawe lineage anyhow.

If such a bottleneck event occurred relatively recently, the genetic picture for the Hadza would not look so different at so many values for K (see graph). This is my own interpretation of the genetic evidence, and I am by no means an expert, so it could be totally wrong, but as far as I can tell a very early bottleneck seems to be the only logical explanation for the Hadza "mystery".

I think that you are reading too much into K-means clusters' depth:

"As it is a heuristic algorithm, there is no guarantee that it will converge to the global optimum, and the result may depend on the initial clusters".

What you get is an approximation of how a population (in this kind a global population) tends to cluster with each other. The greater the depth the more the clusters and the finer the detail. At low K levels the information is always minimal and depends on may factors. Usually realistic genetic clusters only show up at relatively large depths (and you can see how the initial clusters literally vanish from many populations, which become a new clearly distinct one, as you go into further depths).

What showing up early on suggests to me, is that the cluster has been isolated for long (or that makes a good deal of the whole sample - not in this case), not that it's recent. A bottleneck could distort this evaluation, I guess, but I don't trust the very concept of bottleneck, specially when applied to humans and specially in a continent with such an old population as Africa.

IMO, it's much more likely that they have been a distinct, reproductively isolated (up to a point) group for very long. However, I reckon that there is some mystery to this long isolation because, at least today, the region is rather densely populated.

Maju said...

Unlike most Pygmies and Bushmen, who tend toward monogamy, the Hadza rarely marry for life: "Most of the Hadza I met, men and women alike, were serial monogamists"...

I don't really get the difference between monogamy and serial monogamy. Monogamous couples divorce (sometimes at least) and that means serial monogamy, unless divorced people don't marry anymore, what is unlikely.

Are you claiming that Pygmies and Bushmen mostly live happy marriages forever like some birds? That would be most strange, IMO, specially with the reported high levels of domestic violence among Pygmies.

I don't think you have dwelt muh into this matter when dealing with the other populations studied so far, so I would not mind an explanation or further discussion of this matter.

DocG said...

Maju, I'm not a pop. genetics expert and must admit that I don't fully understand how STRUCTURE graphs are produced. Thanks for the clarification.

As I understand it, however, the Hadza are regarded as something of a mystery since their autosomal genetic profile doesn't seem to fit with any other group or groups. I've discussed this with Sarah Tishkoff and she admits that she's puzzled.

"What showing up early on suggests to me, is that the cluster has been isolated for long (or that makes a good deal of the whole sample - not in this case), not that it's recent."

I didn't say it was recent. I think there was probably a very early bottleneck followed by a long period of isolation, yes.

I'm not sure why you are so skeptical regarding bottlenecks. If the Hadza were simply "a distinct, reproductively isolated (up to a point) group for very long" then why wouldn't their autosomal picture resemble that for either the Pygmies or the Bushmen, who also show signs of long isolation? Unless you are suggesting their ancestry is unique?

Whenever there is some sort of disaster that doesn't kill off an entire population, a population bottleneck may easily occur. Why is that so difficult to accept?

DocG said...

"I don't really get the difference between monogamy and serial monogamy. Monogamous couples divorce (sometimes at least) and that means serial monogamy, unless divorced people don't marry anymore, what is unlikely."

Here is how Turnbull describes Mbuti marriages: "Mbuti marriages tend to be somewhat unstable until the first child is born, and indeed are not considered as marriages by the Mbuti until then, separation and the formation of new unions being the rule rather than the exception" (Wayward Servants, p. 73). So you are right, it may be difficult in such cases to distinguish between monogamy and serial monogamy.

Maju said...

Re. the Hadza, they are pretty much unique (though matri- and patri-lineally most related to Pygmy and Sandawe) and that demands an explanation. However I do not have it - at least not one that looks uncontroversial.

Sandawe and Hadza were part of the ancestral East African population, it seems. This population may well have been very diversified since the MSA, as is apparent from the marked differences between these two survivor groups. So, tentatively, I can consider that the main agriculturalist impact in the area did not happen until the Bantu colonization (previous Nilotic pastoralists, like the Maasai, were scattered probably). The Bantus, who were obviously more intensive, had clear interaction with the Sandawe (historically the Bagogo - reported to have very low densities in the 19th century), but the Hadza maybe remained apart as some sort of naturally occurring "buffer state" between the two post-neolithic groups.

It's very tentative; I'd need to be much more knowledgeable on East African history and prehistory to estimate how correct or wrong this may be.

Why am I so wary of bottlenecks? First, because a bottleneck, to be such thing, must be a most dramatic catastrophe, almost wiping out the population (>90% sudden mortality). This kind of event is most unlikely (though may have happened in some cases). And, second, because all the claims so far of genetic "bottlenecks" I've read about are in fact dealing with founder effects (and the related processes of drift and fixation), for example the almost topical "bottleneck" at the out-of-Africa episode, which is much better explained by founder effect plus drift.

Whenever there is some sort of disaster that doesn't kill off an entire population, a population bottleneck may easily occur. Why is that so difficult to accept? -

It must be a disaster that almost wipes out the population. The usual catastrophe that kills 0.1%, 1%, 5% or even 30% of the population (like the Black Death in some places) is not any bottleneck, just a "minor incident" without much demographic influence. In all written history I know of no such event anywhere, except in the case of massive genocide (generally a too modern occurrence).

However in the last 200,000 years there have been some catastrophes of magnitude we can barely imagine or understand: climatic change, sometimes sudden, supervolcano explosions... So I don't dare to discard some sort of bottlenecks at some moments in prehistory but they should affect very wide areas, if not all the planet, so it's hard to imagine a bottleneck that affected only the Hadza.

With enough long-lasting isolation we'd get the same result. The problem is to explain why the Hadza have been so isolated all the time (I speculate about it above).

Maju said...

Re. marriage, I was thinking that the institution should be difficult to consolidate while childcare (and generally the whole economy) is groupal and not based on the nuclear family, right?

Anyhow, it's curious that what you report for Pygmies appears similar to the practices of Basques some 200 years ago (ref. Wilhelm von Humboldt): couples only married when there was child coming. Normally the more stable boyfriend (not necesarily the father) just assumed his responsability.

German Dziebel said...

"The Hadza are a musical isolate as well, as far as I've been able to determine, with a vocal style completely different from that of the Pygmies and Bushmen, and also very different from the call and response pattern so typical of Bantu vocalizing."

Does it show similarities to any other musical tradition at all? When you say "completely different" from P/B, it sounds like their style is monophonic. Do you have a sample we could listen to?

German Dziebel said...

Hadza kinship (see, e.g., http://www.sscnet.ucla.edu/history/ehret/kinship/african_kinship_data.htm) is Bifurcate Merging (with a shift to Hawaiian in the +1 generation female set and in the 0 generation patrilateral set). I do see strong similarities with some Khoisan kinship systems (grandparents equal cross-cousins equal grandchildren) but no similarities with Pygmy kinship.

Hadza kinship terminologies look contracted and skewed as compared to expanded symmetric prototypes found outside of Africa. This may indeed signify a bottleneck. Their sibling set (niye/niyeko) belongs to typical African type A, which is again a highly derived one, according to my typology.

Maju said...

German: have you browse YouTube? I could find this one and this one.

DocG said...

Maju: "Sandawe and Hadza were part of the ancestral East African population, it seems."

Yes, it certainly seems as though both East and South Africa were originally populated by groups directly descended from HBP and more or less similar to them culturally (though morphologically and genetically and also musically distinctive in certain ways). Groups like the San, Hadza and Sandawe appear to be survivals from this population, which must have been very large and widespread prior to the Bantu expansion.

There are other very interesting groups, such as the Masai, Datoga, and many other herders, in addition to the Khoi-Khoi herders once widespread in Southern Africa, all of which probably represent outgrowths from the earlier hunter-gatherer groups of the East and South, just as the Bantu are apparently outgrowths from the Pygmy (or Proto-Pygmy) groups of Central Africa.

DocG said...

Maju: "all the claims so far of genetic "bottlenecks" I've read about are in fact dealing with founder effects"

I think you are too hung up on the terminology, which is admittedly somewhat loose. By "bottleneck" I refer to any severe reduction in population that appears to have led to a founder effect. It could be the result of a disaster, but also of the banishment of a small group, or the isolation of a small group for any one of a variety of possible reasons, such as a few people losing their way or being cut off by a hostile group, or children being abandoned, etc.

I'm not saying such events are necessarily common, and in many cases the isolated or surviving lineage may have ultimately died out. But in a few cases it could have survived to produce a new and possibly very different sort of population. It seems clear that the proto-Bantu must have broken away from HBP due to a "bottleneck" or, if you prefer, "founder effect" of more or less this kind. It's hard to imagine any other source of significant differentiation, morphological, genetic or cultural, since simple drift cannot in itself produce sharp differentiations. While one might assume that migration alone could produce such differences, look at the vast and complex Bantu migration, which only spread Bantu morphology, genes and culture throughout SSAfrica.

DocG said...

Thanks, German, for reminding e that I needed to find some examples of Hadza music. And thanks, Maju, for the youtube links. But imo they are misleading, because both the musical style and the dancing seem very mainstream Bantu to me. And the music is very different from that on the CDs of Hadza traditional music I've heard, accessible via Amazon.com. I've added links to all the above, along with some more brief commentary, to post 249.

German Dziebel said...

"By "bottleneck" I refer to any severe reduction in population that appears to have led to a founder effect. It could be the result of a disaster, but also of the banishment of a small group, or the isolation of a small group for any one of a variety of possible reasons, such as a few people losing their way or being cut off by a hostile group, or children being abandoned, etc."

If a population is subdivided, and all human populations are subdivided into "kinship-groups" (in addition to other natural subdivisions such as geography), then the "founder effect" is just a continuation of the "linear effect," which may affect population diversity levels long before a demographic bottleneck occurs. For small hunter-gathering populations, the linear effect is a significant factor.

Maju said...

Victor: the Hadza are not that different when it comes to haploid DNA (see for instance Tishkoff'07) and the sharp differences you mention only show up in autosomal clustering. I don't think that it is an artifact (that could be) but I do think it does just mean relative reproductive isolation in the last many milennia.

[Btw, notice how this paper suggests a Sandawe root for E1b1b and a Hadza root for B2b, with Pygmy clades being derived].

A similar case happens with Basques and Iberians: they show up as autosomally distinct in Europe (K=5 and K=3 respectively in Bauchet'07) but they do not significatively differ from their neighbours in haploid DNA. This kind of autosomal uniqueness is, in my opinion, a product of "recent" isolation, which homogenizes locally the genome. Though I would not be able to decide how recent (several milennia in any case, quite more for the Hadza because they show up in the global structure, what Basques or Iberians do not).

In other words, I think that haploid genetics give normally a better idea of the remoteness of affiliation than autosomal data, which may just indicate "recent" intra-population homogenization by means of normal drift in the context of relative isolation.

Regarding the use of the term and concept of bottleneck, I think it is important to make a difference. Drift may produce rather sharp differences alone, specially in such small populations as the Hadza, provided enough reproductive isolation. The key is, I think, that the population should have been relatively isolated in the last many milennia (i.e. in Europe the three Western clusters can perfectly correspond to the three Paleolithic provinces).

They Hadza would show up later in the K-means analysis if they had larger apportions of non-Hadza genetics. That happens in fact with the Khoi-San, which we know are older in the sequence of divergence from the common ancestral root but who do seem to have incorporated more non-Khoisan genetics in their autosomal pool.

In any case it's just my opinion.

While one might assume that migration alone could produce such differences, look at the vast and complex Bantu migration, which only spread Bantu morphology, genes and culture throughout SSAfrica.

Actually, we'd need a good study on Bantu genetics to tell for sure. For what I can see, Bantus, specially in the SE picked up other groups' genetics. An extreme case are the Xhosa, who often look more Khoisanid than typically Negroid but in the Pygmy study Mozambicans also showed up as a different autosomal cluster and in general Southern African Bantus have some significative non-Niger-Congo Y-DNA, B2a specially, (and probably even more mtDNA), which they must have picked along their migrations.

I'd really like to read a comprehensive paper of Bantu genetics because I do suspect that they incorporated significative pre-Bantu genetics, specially in the East and SE (where there were probably pastoralists already living when they arrived, with too high densities to be just displaced).