quote:

Originally posted by Fil2:

quote:

But that's not the full quote. It's both edited and taken out of context.

Edited yes, but only to make my point clear but the quote is not out of context.

If you read my post more carefully, you will find that, like you, I am arguing that "A theory has to be supported by evidence, it doesn't become fact simply because it cannot be disproved". Oppenheimer doesn't supply evidence. He uses Capelli's data, reassigns the STR values to his own haplogroups definitions without explanation, presents them as scientific fact and explains what the inferences are. Capelli found something like 70 R1b1c haplotypes which Oppenheimer reduced to something like 16 haplogroups. Why should for example, the STR sequences, 13 24 15 10 12 13 appear in the same haplogroup as 13 24 15 11 12 13, R1b-13 in Oppenheimer's definition, whereas he assigns 13 24 14 11 12 14 to haplogroup R1b-14a. What is the rationale behind it? Why should one be a descendant of Rory whilst the other is a descendant of Ruy? In response to the statement, "Other geneticists have yet, to my knowledge, to criticise his work so it is well to bear it in mind.", I simply point out that without an explanation of this method, it cannot be verified or rejected. There's nothng conspiritorial about that statement.

quote:

Originally posted by Fil2:

quote:

"How many of the first immigrants were cremated"

Unless cremation was a common practice in their homelands, and I don't believe it was, very few.

The number of cremation urns found on the Continent is large. From the bronze age to the early pre roman iron age the great mass of recorded burials, in the hundreds of thousands, are cremations. In the early centuries AD, the number of inhumations increases but cremation remains a popular practice. Urnengräber, ie. in cremation urns, Brandgrüber, the burnt material from the pile and Brandschüttungsgräber, the same but including the burnt grave goods, are common in many parts. It is precisely the analysis of the different styles, especially the comparison with urns found in Frisia with those found in Sancton and Heworth in Yorkshire and urns found in Anglen and Fyn with those found in Sancton as well as the fused anglo-saxon styles of the Weser Elbe region with those found in several sites in Yorkshire which provided the evidence which archaeologists such as Myres, Swanton and Albrectsen attempted to use to answer who came, where they came from and at what period in time they came.

best

Harry A

quote:

The converse of publishing a book which is not subject to peer review is also that it is not seen by the scientific community as a contribution to the science. Even if Oppenheimer was 100% correct, the first person to write the same thing in a peer reviewed publication will get the credit, not Oppenheimer...Oppenheimer, far from being cranky is very clever.

So are we dealing here with simply a failure of Oppenheimer to go through the 'proper channels'? I read with interest the paper that Harry referred me to:

quote:

"Geographic Patterns of R1b in the British Isles – Deconstructing Oppenheimer".
http://www.jogg.info/32/campbell.pdf

Harry, you say 'typical of the criticism'. From this I assume you can cite other papers by other scientists, precisely the thing I am looking for?

My problem is that if we consign Oppenheimer to the wastebin of pseudohistory amd pseudoscience then we may be ignoring his contribution to this most important of debates. Oppenheimer infers the following from his work:

1. The genetic pattern of the British Isles was determined well befor the arrival of the Anglo-Saxons.
2. There is a roughly east-west split that dates back as far as the late-paleolithic/early neolithic.
3. The Angles contributed only a small fraction of our genetic heritage.

I fully accept that his views may be wrong but at the back of my mind I have a niggling doubt: why on earth is a University as prestigious as Oxford happy for him to continue to be associated with it?

quote:

why on earth is a University as prestigious as Oxford happy for him to continue to be associated with it?

Oppenheimer is a paediatrician by training, and it is in that role that he is associated with Oxford University. His popular writings on genetics and human prehistory are totally unrelated to his association with Oxford University.

quote:

Harry, you say 'typical of the criticism'. From this I assume you can cite other papers by other scientists, precisely the thing I am looking for?

Hi Duncan,

I don't want to start going through them all but here is another:
http://scienceblogs.com/gnxp/2007/03/bad_historical_population_gene.php

Other than that go to http://dna-forums.org and just search on Oppenheimer.

The important point to realise here is that, as with the JOGG, these are all personal views and Oppenheimer is just as entitled to hold a view as anyone else. Peer science is about publishing the hypothesis along with any assumptions made in creating that hypothesis, the method or experiment, the results and a discussion or interpretation of the results or possible conclusions that can be drawn from the results. Other scientists are given enough information to test what has been published.

So, because Oppenheimer has written his ideas in a book, peer scientists won't respond. All one can see are the views of other people who may be qualified or not.

An example of peer criticism to a paper published in a scientific journal is quite different. For example, investigating the genetic impact of indo european speakers in India, in 2006 Sengupta et al published a paper in the American Journal of Human Genetics entitled, Polarity and Temporality of High-Resolution Y-Chromosome Distributions in India Identify Both Indigenous and Exogenous Expansions and Reveal Minor Genetic Influence of Central Asian Pastoralists

http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1380230

Denise R Carvalho-Silva, Tatiana Zerjal and Chris Tyler-Smith of the Wellcome Trust Sanger Institute responded in the Journal of Biosciences.

http://www.ias.ac.in/jbiosci/mar2006/1.pdf

Their criticism is based on Sengupta's conclusions drawn from inferences of the origin and dating of certain lineages. Where and when a lineage was created and when it entered another population is one of the biggest unknowns in genetics.

The weakest assumption in Oppenheimer's book too is the dating and the hypothesis that the Basques are the sole representatives on the european paleolithic.

Jim Wilson (Genetic evidence for different male and female roles during cultural transitions in the British Isles 2000) explained that this was based on 3 assumptions, the Vasconic language, the assumed ancient mtDNA lineages of the Basques and the highest frequency of blood group O and rhesus cde, which "is thought to represent the contribution of Paleolithic Europeans", ie. another assumption. He warned "We know of no other study, however, that provides direct evidence of a close relationship in the paternal heritage of the Basque- and the Celtic speaking populations of Britain."

Santos Alonso addressed this hypothesis in 2005, (The place of the Basques in the European
Y-chromosome diversity landscape) whilst Oppenheimer's book was being finalised. "However, the idea of a Basque genetic pool that shows little influence from both the Neolithic and later population flows, has spread through the literature as a circular argument that has led to use the Basque population as the representative gene pool of the first modern human settlers of Europe. Thus, with the general aim to investigate the evolutionary history of the Basques we studied in this work a large
Basque sample ...".

His conclusions were: "Contrary to previous suggestions, we do not observe any particular link between Basques and Celtic populations beyond that provided by the Paleolithic ancestry common to European populations, nor we find evidence supporting Basques as the focus of major population expansions."[i]

To the best of my knowledge, there has been no challenge to study and other studies are starting to suggest that the Basques may have been just as influenced during the neolithic as many other european populations.

quote:

I fully accept that his views may be wrong but at the back of my mind I have a niggling doubt: why on earth is a University as prestigious as Oxford happy for him to continue to be associated with it?

He hasn't committed any scientific treason. It's not wacky pie in the sky stuff. All he is saying is that [i]if[/ the Basque hypothesis is true and if Försters new and experimental use of the methods of molecular biology can be applied to linguistics and if we take one interpretation of Caesar's de Bello Gallico, then the scenario he describes could be a possibility. He's not even saying it is the case.

In any event, it's no more outrageous than Bryan Sykes' claim to have met a bloke who used to drink in a pub in North Wales with the last two remaining Neaderthals alive.

best

Harry A

quote:

Originally posted by Harry Amp:
He hasn't committed any scientific treason. It's not wacky pie in the sky stuff. All he is saying is that if[/ the Basque hypothesis is true and if Försters new and experimental use of the methods of molecular biology can be applied to linguistics and if we take one interpretation of Caesar's de Bello Gallico, then the scenario he describes could be a possibility. He's not even saying it is the case.

All well and good but with regard to the theme of this discussion Oppenheimer's major claim is that the genetic divisions identified by Capelli et al. are not a result of the Anglo-Saxon invasion of Britain but of much earlier population shifts. This is his main hypothesis and his methodology, no matter how questionable (Campbell's critique in The Journal of Genetic Genealogy being a good example) is designed to support this analysis. What I'm saying here Harry is that until the scientific community tackle Oppenheimer, which for very good reasons they may not, then we have to do that job. Until Oppenheimer does follow conventional scientific protocol he is in limbo - free from scientific criticism because he has not written a scientific paper. He is therefore in the realm of pseudoscience and pseudohistory. I don't think for one minute we are dealing with another Graham Hancock but let's not throw the baby out wi' yon bathwater. Curiously Oppenheimer, may find himself in the same place academically as that other amateur historian of recent times, Michael J. Harper...

quote:

Originally posted by Harry Amp:
In any event, it's no more outrageous than Bryan Sykes' claim to have met a bloke who used to drink in a pub in North Wales with the last two remaining Neaderthals alive.

That's not so outrageous. I saw a bus load travelling down to the playoffs a week last Sunday...

quote:

What I'm saying here Harry is that until the scientific community tackle Oppenheimer, which for very good reasons they may not, then we have to do that job. Until Oppenheimer does follow conventional scientific protocol he is in limbo - free from scientific criticism because he has not written a scientific paper.

Many geneticists would agree with some of what he writes. The celtic and former celtic speaking populations of north west europe are genetically very similar to the germanic speaking populations at many levels. It's all a question of degrees, how similar is 'similar' and of interpretation, why are they similar?

You can see the an example of this on the following plot of aDNA:

http://bp1.blogger.com/_Ish7688voT0/R6Nt4XlfrwI/AAAAAAA...I/s1600-h/pc300k.jpg

The Irish and Germanic markers are separated from each other but taken together, they are both well separated from the Italians, Greeks or Spaniards. One could argue that they are similar or quite different depending on what story one wanted to tell. The UK population is neatly spread over both camps.

Oppenheimer choses to state that they are similar and argues his case. He argues that the paleolithic population of the UK was little influenced by immigration in the neolithic.


This has been argued by geneticists like Richards for some time, eg. "... the immigrant Neolithic component is likely to comprise less than one-quarter of the mtDNA pool of modern Europeans." (from Richards et al. 2000, Tracing European founder lineages in the Near Eastern mtDNA pool.)


Other geneticists, such as Chikhi argue the converse, ie. the neolithic expansion dramatically affected the paleolithic population, eg. "We also argue that local hunter-gatherers contributed less than 30% in the original settlements." (from Chikhi et al, 2002, Y genetic data support the Neolithic demic diffusion model)

Because this is all peer reviewed science, it can all be challenged. Usually challenges are to the assumptions made in the model, very rarely, the method and sometimes the interpretation. But they continue to argue. Barbujani summarised the situation in 2006 with: Population genetics: DNAs from the European Neolithic.

Oppenheimer is free from such challenges for the reasons discussed. He is quite free to take Richard's view and reject Chikhi's view and write it into a book accordingly. Someone else may do the opposite, hence you have historians such as Starkey claiming that 85% of the population was replaced by invading Saxons.

best

Harry A

quote:

hence you have historians such as Starkey claiming that 85% of the population was replaced by invading Saxons.

Which would be how many in round numbers?

quote:

Which would be how many in round numbers?

Depends on the population of Roman Britain which in turn varies according to whom one is listening.

Some place it as high as 3.9m, based on the number of villas found, the number of people each villa could support, the number of coins in circulation etc. If that figure is true, Starkey would need something of the order of 3.3m germanics. However, the 3.9m romano british figure is often proposed by those who argue for acculturation, so a high figure is required to reduce the likelihood of a massive migration, on the basis that such high numbers are not credible.

Archaeologists such as Härke argue for a population crash in the roman british population in the 4th cent so that the AD410 figure is something between one and two million. Faulkener's study of the buildings falling into disuse lends support to this. This would still require a migration of 850,000 to 1.7m germanics however which is still much greater than any other migration of the period and therefore seem unlikely.

I think Starkey must have taken his 85% figure from Weale's study on Anglo Saxon mass migration where Weale found that, in one part of England, the Anglo Saxons contribute between 50% and 100% to the modern local genepool, although he stresses that the figures do not clarify whether this means extermination, ethinic cleansing or simply added to the genepol existing at the time. Capelli's study found that the figure of 50-100% was not uniform however and that some areas, such as the south east experienced lower immigration, around the 30% mark.

Härke estimates that immigration was at roughly at around 10%, between 150,000 and 250,000. If this is the case, the question then becomes why does such a small number at the time grow into the larger numbers that we see today?

However, chain migration, ie. where people follow in the footsteps of their ancestors can too be a factor and some rough and ready figures from West Heslerton support the idea that an early low number of settlers may have been added to constantly over a period of a couple of hundred years. Additionally, the romano british population may have been depleted by emmigration to Brittany and Galicia and by the effects of the Justinian plague.

But, all population models are inadequate and we have no real idea at what age germanics or romano britons had children or how many children they had. The two groups may well have had different fertility rates, birth rates and life expectancies. So, all the models include a large amount of guesswork.

best

Harry A

quote:

Capelli's study found that the figure of 50-100% was not uniform however and that some areas, such as the south east experienced lower immigration, around the 30% mark.

This part of Capelli et al.'s study has always struck me as fascinating. The areas that we would assume to be the heartland of Saxon immigration actually have low percentages of north-west european YDNA whereas the area roughly denoted by the later Danelaw is much higher. These are also the areas supposedly settled by Angles. Capelli et al are aware of this and are prepared to entertain the possibility that the markers could be due to Danish viking rather than Anglo-Saxon immigration. Until it is possible to distinguish the time of arrival it is impossible to distinguish the Danish from the Anglo-Saxon DNA.

quote:

You can see the an example of this on the following plot of aDNA:

http://bp1.blogger.com/_Ish7688voT0/R6Nt4XlfrwI/AAAAAAA...I/s1600-h/pc300k.jpg

The Irish and Germanic markers are separated from each other but taken together, they are both well separated from the Italians, Greeks or Spaniards. One could argue that they are similar or quite different depending on what story one wanted to tell. The UK population is neatly spread over both camps.

An interesting little picture. Where does this come from? Do you know what the 'AJA' refers to? The differences between the Spanish and the Irish populations tell a different story than the one found in Oppenheimer which shows close similarities between the 'celts' of Britain and the Spanish.

quote:

Where does this come from? Do you know what the 'AJA' refers to?

It's from Analysis and Application of European Genetic Substructure Using 300 K SNP Information by Chao Tian et al.

http://www.pubmedcentral.nih.gov/articlerender.fcgi?too...ed&pubmedid=18208329

AJA = Ashkenazi Jewish Ancestry

best

Harry A

quote:

This part of Capelli et al.'s study has always struck me as fascinating. The areas that we would assume to be the heartland of Saxon immigration actually have low percentages of north-west european YDNA whereas the area roughly denoted by the later Danelaw is much higher.

Yes, the Capelli study makes a note of that too. However, the analysis referred to, the PC Plot is only one analysis. The MCMC method, whilst still showing a similar heterogentiy in the British isles, gives the following admixture results for the south, when measured against the North germany/Denmark sample:

Chippenham 70.8%
Faversham 49.5%
Midhurst 24.4%
Dorchester 36%
Cornwall 57.7%

These are of course higher. In fact, the mean of the medians for the english study centres shows an admixture of 54% but, the main finding is that the distribution is far from evenly spread.

Chippenham is in the land of the Gewissae, the group who were to become the West Saxons so its 70.8% admixture figure is in keeping with that theory. Dorchester and Midhurst of course remain low, but these were areas the West Saxons expanded into, in a political way, not a settlement way.

A possible explanation for Faversham's 49.5% admixture figure, and also Midhurst and Dorchester's is that 'Saxons' from Merovingian Gaul settled in those areas during the 3rd quarter of the 6th cent. These 'Saxons' would probably already include acculturated Belgic peoples, Morini, Menapi etc. who may not be genetically different from Britons.

However, Capelli's results rely, as did Weale's study the previous year, on the hypothesis that R1b results are bascially indigenous. Newer, higher resolution analysis is starting to split the R1b group into new haplogroups, each potentially with its own origin.

As Jim Wilson wrote a few years ago, 'we are on the cusp of learning something new and very interesting about R1b'. Unfortunately, the whole new story is emerging in dribs and drabs and taking longer than one hoped for.

best
Harry A

What about Robin Mckie’s Face of Britain study? This seems to broadly agree with the Goldstein/Capelli findings.

quote:

What about Robin Mckie’s Face of Britain study? This seems to broadly agree with the Goldstein/Capelli findings.

Well it's not Robin McKie's study, he just wrote a quick book based on early results of the Wellcome Trust's Peoples of the British Isles project. This study is on going and so far they have collected around 3250 of the projected 3500 samples. Full results are expected to be published in 2009. You can see in their Newsletter 1 some of the early findings:

http://www.peopleofthebritishisles.org/press

This is a major study but it is different from the yDNA studies such as Weale's or Capelli's. It goes further and includes a number of autosomal markers. When McKie wrote his book, only a handful of areas had been surveyed.

best

Harry A

Harry Amp -- just a note of appreciation for your huge ongoing contributions to this thread. I'm learning a huge amount.

cheers,
Jonathan

quote:

Originally posted by Harry Amp:
.. the analysis referred to, the PC Plot is only one analysis. The MCMC method, whilst still showing a similar heterogentiy in the British isles, gives the following admixture results for the south, when measured against the North germany/Denmark sample:

Chippenham 70.8%
Faversham 49.5%
Midhurst 24.4%
Dorchester 36%
Cornwall 57.7%

These are of course higher.

'PC' is Principal Componenet analysis but what is the 'MCMC' method and why should its results differ from the former? A Y Chromosome Census of the British Isles

quote:

However, Capelli's results rely, as did Weale's study the previous year, on the hypothesis that R1b results are bascially indigenous. Newer, higher resolution analysis is starting to split the R1b group into new haplogroups, each potentially with its own origin.

This higher resolution analysis is, of course, what Oppenheimer did in his book. What other studies do you know of that are attempting to do this?

I think we all wait with baited breath to find out the results of Bodmer's survey. As the Face of Britain book points out, there may not be too much time left to chart the patterns with increased social mobility complicating populations.

quote:

'PC' is Principal Componenet analysis but what is the 'MCMC' method and why should its results differ from the former? A Y Chromosome Census of the British Isles

A PC plot is a method of compressing multidimensional data into two dimensions which makes it easier to visualise what's going on. However, there are a number of distortions with this technique, some deliberate, some unintended. Consequently it is not possible to read a PC plot as one would a typical XY graph, which is what people try, wrongly, to do.

MCMC stands for Markov Chain Monte Carlo and is a computationally intensive method used, in this case, to calculate admixture. If you look in Capelli'ss study, you will see a link to the supplementary data referred to.

quote:

This higher resolution analysis is, of course, what Oppenheimer did in his book. What other studies do you know of that are attempting to do this?

Oppenheimer doesn't do this at all. In fact, by far his main source of data is that provided Capelli. Oppenheimer creates his own groupings as I mentioned above creating much fewer than Capelli in fact. The benefit of higher resolutions is that the rather large number of R1b samples can be split into more geographically specific regions. Oppenheimer attempts to do this with six STRs but doesn't explain why he groups them in the way he does. As sated earlier, much progress has been made but it still remains a slow progress.

quote:

As the Face of Britain book points out, there may not be too much time left to chart the patterns with increased social mobility complicating populations.

But there are improvements in sampling methods too. In particular, one technique used in north west england where modern samples were taken from people who had an ancestry in the area which could be verified by surnames taken from medieval court rolls and so on. This technique proved successful as an attempt to create a proxy medieval population in a modern mixed population.

Excavating Past Population Structures by Surname-Based Sampling: The Genetic Legacy of the Vikings in Northwest England

http://mbe.oxfordjournals.org/cgi/content/abstract/msm2...qRJ9j6mF&keytype=ref

best

Harry A

quote:

Originally posted by Harry Amp:
Oppenheimer doesn't do this at all. In fact, by far his main source of data is that provided Capelli. Oppenheimer creates his own groupings as I mentioned above creating much fewer than Capelli in fact. The benefit of higher resolutions is that the rather large number of R1b samples can be split into more geographically specific regions.

Sorry Harry, are we talking about the same book? Oppenheimer's stated aim (I don't have his book with me at work) is to take the Capelli et al. data and to subject it to greater scrutiny. He distinguishes sub-groups within each major grouping and dates them to different periods. Therefore he comes up with more, not fewer groupings. I'll be able to refer you to specific pages once I get the text to hand.

quote:

Sorry Harry, are we talking about the same book? Oppenheimer's stated aim (I don't have his book with me at work) is to take the Capelli et al. data and to subject it to greater scrutiny.

My response was to your statement about resolution:

"This higher resolution analysis is, of course, what Oppenheimer did in his book."

This is not the case and cannot be the case, since 85% of Oppenheimer data for the British Isles uses Capelli's data. Capelli's study works at a resolution of 6 STRs: DYS 388, 393, 392, 19, 390 391. Oppenheimer has to work at that resolution, he hasn't got 10, 15 or 24 STRs.

I haven't read the book but if he is working at a higher resolution, please let me know what the STRs he uses are.


Regarding the groupings of the various STR values, Capelli's dataset is now widely published on the web. You can see the dataset here:

http://freepages.genealogy.rootsweb.ancestry.com/~gallgaedhil/Capelli.htm

R1xR1a1 means all of R1 with the exception of R1a1, essentially what used to be called R1b and then R1b1c and so on.

As you can see, Capelli has many different STR values for R1xR1a1, for example first on the list above is: 10,13,11,15,24,10. 2nd on the list is 10,13,11,15,25,10 ie. a change at DYS 390. There are several dozen of these.

Oppenheimer on the other hand reduces this wide range of STRs into a dozen or so groupings for R1xR1a1, his R1b-9, Rox and R1b-10 Ruy and so on. But, what is the rationale behind these groupings and what is the scientific basis? Why not divide Capelli's R1xR1a1 data into 24 or 36 groups and tell a different story?

best
Harry A