Header image

john hawks weblog

paleoanthropology, genetics and evolution

Photo Credit: Dental chipping in Homo naledi. Ian Towle and colleagues

Link: Writing about science in an African language

Sibusiso Biyela has a great essay in The Open Notebook recounting the challenges faced in communicating science about a South African dinosaur discovery in one of the major languages of South Africa: “Decolonizing Science Writing in South Africa”.

My job was to write about the discovery for the South African website SciBraai—and to do so in my native language of Zulu.
But there’s no word for “dinosaur” in Zulu. Nor are there words for “Jurassic,” “fossilization,” or “evolution.” Despite the fact that Zulu—or isiZulu, as the language is called in South Africa—is spoken by some 10 million people, it simply doesn’t have the words for communicating science.
So my news piece wasn’t just a news piece. It was an attempt to tell a science story in a language that science overlooked—to help right a societal wrong. It was a small contribution among an increasing number that aim to help decolonize South African science writing. And it was rife with more pitfalls than I could have imagined. The task of describing science clearly, concisely, and accurately—already challenging in English—became exponentially more difficult in my native tongue.

I absolutely love how creating new translations and terminology provides the opportunity to escape the bad thinking of the past. For example, on “dinosaur”:

I encountered trouble, however, with the word dinosaur, which comes from the Greek for “terrifying lizard.” The term is a misnomer: Many dinosaurs bear little resemblance to lizards, and some ancient animals that looked like terrifying lizards, such as the dimetrodon, are actually more closely related to mammals than to dinosaurs. I didn’t want to introduce into Zulu the same misconceptions that already plagued so many English speakers.

It is one of the hardest challenges in writing about science to escape the worn ruts of past writers, which may be easier to follow, but lead readers off in wrong directions. The great opportunity of starting fresh is that a writer can build a new way of representing the science that aligns with today’s concepts and ideas.

Link: Scientists and 'science denial'

A short essay by Kari Fischer from the New York Academy of Sciences, in The Scientist: “Opinion: What You Believe about “Science Denial” May Be All Wrong”.

Scientists should receive more institutional support, training, and career incentives to engage in proactive communication with the public. And when we do speak out, we must remember that we represent not only ourselves, but our institutions, and science as a whole. We should resist the temptation to engage with trolls, or become them ourselves by berating “non-believers.” Ridicule will not foster trust.

Obviously I’m more familiar with my corner of science than with others. Anthropologists in the U.S. live in a bubble. Few listen to non-academic voices, or engage the public in a way that has any direct effect on their research agenda.

These kinds of academic meetings encouraging better methods of speaking to the public have some value. But what is really needed is a wholesale change to the incentive and reward system in science and academia.

Long-distance weather 'teleconnections'

Last week Nature released an intriguing paper that demonstrates super-long-distance correlations in rainfall patterns, showing how the South Asian monsoon is synchronized with East Asia, Africa, and even events in North America. The paper, by Niklas Boers and coworkers, is titled, “Complex networks reveal global pattern of extreme-rainfall teleconnections”.

According to the statistics provided with the paper, it has one of the longest submission-to-acceptance times I’ve ever seen. It was received August 7, 2016, and finally accepted December 4, 2018.

The basic idea is that weather patterns are correlated over short distances, but become less and less correlated as distance increases. Over distances less than 2500 km, the loss of correlation is approximately linear with distance, but over longer distances the loss of correlation is faster than linear—a power law with exponent greater than one. But against this pattern of rapid loss of correlation at long distances, some events stand out as being much more correlated than expected. These types of correlations are identified in the paper as “teleconnections”.

Finding such teleconnections is a big data problem, where the datasets involve weather observations in a dense worldwide matrix. As geneticists would recognize, this gives rise to a problem of multiple comparisons. The paper includes a fun discussion of the multiple comparisons problem as applied to weather network data:

Before proceeding, however, a statistical problem needs to be solved, which arises in all data-driven interdependency analyses, and in particular in networks that are constructed on the basis of statistical similarities. Such approaches are generally biased because of multiple comparisons. In this case, we compare each timeseries with 575,999 other timeseries, which amounts to more than 1011 comparisons in total. Therefore, the network will contain links that—despite corresponding to statistically significant pairwise synchronization values—are present only because of random coincidences, and not because of physical mechanisms.

The method used to address this problem in the paper is to reduce the problem to “bundles” of spatial links, thereby identifying channels of long-distance interaction above the noise of incidental similarities. This method lies behind the identification of long-distance couplings:

After applying our correction technique, a concise teleconnection pattern associated with the northern part of the South Asian monsoon is revealed: in addition to regional links covering most of the Indian subcontinent, we observe pronounced link bundles connecting SCA with eastern Asia, the African tropics, large parts of Europe and the eastern coast of North America, as well as the Southern Hemisphere extratropics. The break between regional and teleconnection scales (Fig. 2) is not affected by this correction (Extended Data Fig. 2).

By examining the time lags separating these correlated patterns in different parts of the world, the authors conclude that Rossby waves are the most important mechanism driving long-distance weather correlations. I didn’t know anything about Rossby waves before reading this paper, and they seem like an interesting phenomenon.

One implication of such long-distance correlations is that what seem like “clusters” of extreme events may actually be a single event with long distance results. That has importance for studying the frequency of extreme weather events of all kinds, which have been suggested as a consequence of a warming climate.

We hope that our results will inspire further work on the predictability of EREs arising from these large-scale teleconnection patterns and on their representation in weather and climate models. Many studies have recently raised the concern that the characteristics of extreme events will change under ongoing climate change. A particular challenge in this regard is the discrimination between natural variability and anthropogenic influences6,7. With the increasing temporal lengths of global, high-resolution rainfall datasets, investigating changes of the global rainfall teleconnection structure along the lines of this study should become possible in the near future.

It is possible that the long-distance correlations are actually present within current climate models, and themselves give rise to the increase in extreme weather events in models. Or the Rossby wave-driven correlations might be a mechanism not yet replicated in climate models.

Scientific meetings should happen where research happens

In Nature Evolution and Ecology, Zeresenay Alemseged, Jackson Njau, Brianna Pobiner and Emmanuel Ndiema have a comment that reports on the 2017 conference of the East African Association for Palaeoanthropology and Palaeontology: “Connecting palaeoscientists in eastern Africa and the wider world”. This is a valuable essay and is accessible through the online PDF reader that the Nature journals have started using.

The commentary begins with a strong statement of the injustice of the paleoanthropological enterprise at the end of the last century:

[I]n spite of the undisputed importance of the region to the understanding of human evolution, there was no regional scientific forum that facilitated scientific discourse locally until the birth of the East African Association for Palaeoanthropology and Palaeontology (EAAPP) in 2005. As such, eastern Africa largely served as a data-mining hub in which local research, training and public education saw few of the benefits of the worldwide fame and international funding efforts that the region attracted. A very small number of senior local scholars managed to travel overseas to attend scientific meetings, but most east African scientists, especially students and early-career researchers, were scientifically isolated owing to the prohibitive costs of travel, accommodation and conference registration and cumbersome visa procurement for travel to Western countries. Moreover, there was a complete absence of a platform where international scientists undertaking research in the region could engage with local decision-makers and stakeholders managing research, conservation and curation of palaeoanthropological resources, leading to haphazard and non-strategic collaboration. The rationale behind establishing the EAAPP emanated from these realizations and aimed at addressing these issues.

I added the emphasis to the sentence that mentions “data mining”.

It is sad how little value anthropologists of the past have created for the many nations where they have collected data. The essay points to the funding efforts of the past, but even these have been anemic compared to the scientific value and future potential of fossil discoveries in Africa. The future of discoveries crucially depends on the ability of scientists and other stakeholders to translate heritage into jobs, economic value, and international prestige for nations that face enormous development challenges.

Today’s scientists have to build what the previous generations did not. Previous generations too often acted as if fossil discoveries were like diamond mining. They spoke of “fossil fields being nearly exhausted” and acted like De Beers, creating scarcity of scientific value by limiting access.

That’s a great strategy if you want to keep selling diamonds, but it is a terrible way to build science. Building science means creating value at many levels, not only traditional academic outputs, but building scientific capacity, tourism, and heritage management.

On Twitter over the last few days I have pointed out the analogous situation happening in human genomics and paleogenomics, with a number of scientific meetings this year devoted to the genetics of Africa but being hosted at research institutes in Europe. It is so short-sighted to try to organize a meeting to show an institute is “engaged in Africa” and think that the best way to show this is to pay airfare for Americans to visit Europe.

Link: Falsifiable science and good science

Sabine Hossenfelder has become an outspoken skeptic of the idea that a new, even-bigger-than-the-LHC particle collider will achieve any breakthrough in high energy physics.

Yet many physicists are arguing strongly for a new collider, one that would attain higher energies than the current Large Hadron Collider. They point out the many theoretical models that make predictions about particles within the energy range of a new collider. In response to this argument, Hossenfelder writes: “Just because it’s falsifiable doesn’t mean it’s good science.”.

The other day I got an email from a science writer asking me to clarify a statement he had gotten from another physicist. That other physicist had explained a next larger particle collider, if built, would be able to falsify the predictions of certain dark matter models.
That is correct of course. A next larger collider would be able to falsify a huge amount of predictions. Indeed, if you count precisely, it would falsify infinitely many predictions. That’s more than even particle physicists can write papers about.
You may think that’s a truly remarkable achievement. But the question you should ask is: What reason did the physicist have to think that any of those predictions are good predictions? And when it comes to the discovery of dark matter with particle colliders, the answer currently is: There is no reason.

The current Standard Model in physics explains existing experimental data. Many physicists don’t like the Standard Model because it doesn’t seem natural—it violates their intuition that physical models should be mathematically simple, symmetrical, and “beautiful”. But the Standard Model works. While it might stop working at slightly higher energies, according to Hossenfelder there’s “no reason” to suppose that the particular energies in range of a next-generation collider will lead to new unexplained violations, any more than the LHC has.

Meanwhile, the price tag of a new collider will be extremely high. This price is a reckless gamble on predictions that nearly all physicists believe will be falsified by a collider.

The question is, what could be done instead?

I watch the high energy physics community because they are much better organized to mobilize public funding than evolutionary biologists. They are talking about building a collider that may cost 10 billion dollars. The Human Genome Project cost less than 3 billion.

The Human Genome Project cost less than 3 billion.

The operating budget of the current Large Hadron Collider is approximately 1 billion dollars per year.

The bison bone bed at Gran Dolina

In an earlier post, I looked at work by Mark White, Paul Pettitt, and Danielle Schreve, which considered evidence for Neandertal prey selectivity at five sites. One of those, Mauran, France, was a site where Neandertals repeatedly killed groups of bison, amounting to more than 130 animals over many years.

A 2017 paper by Antonio Rodríguez-Hidalgo and colleagues demonstrated a similar pattern of bison hunting in the TD 10.2 layer of Gran Dolina, Atapuerca, Spain: “Human predatory behavior and the social implications of communal hunting based on evidence from the TD10.2 bison bone bed at Gran Dolina (Atapuerca, Spain)”. There, around 400,000 years ago, hominins left the partial remains of more than 60 bison, forming a bone bed of more than 22,000 specimens.

The results indicate a monospecific assemblage heavily dominated by axial bison elements. The abundance of anthropogenic modifications and the anatomical profile are in concordance with early primary access to carcasses and the development of systematic butchering focused on the exploitation of meat and fat for transportation of high-yield elements to somewhere out of the cave. Together with a catastrophic and seasonal mortality pattern, the results indicate the procurement of bison by communal hunting as early as circa 400 kyr. This suggests that the cognitive, social, and technological capabilities required for successful communal hunting were at least fully developed among the pre-Neanderthal paleodeme of Atapuerca during the Lower Paleolithic. Similarly, the early existence of mass communal hunting as a predation technique informs our understanding of the early emergence of predatory skills similar to those exhibited by modern communal hunters.

It’s not clear whether the hominins at Gran Dolina were early Neandertals like those represented approximately 450,000 years ago in the Sima de los Huesos, or whether they may instead have belonged to some other population. Whoever they were, they must have had a place somewhere nearby where they could kill small groups of bison at two different times of year.

Why nearby? Because they transported basically whole bison bodies into Gran Dolina, where they butchered the animals further and then carried most of the long bones somewhere else. What was left was mostly skulls, vertebrae, and ribs—an unusual proportion of ribs compared to most faunal assemblages. The hominins paused as they butchered to break the ribs and “snack” on red bone marrow.

Why two seasons? Because newborn bison are born at a single time of year, making it possible to work out what season first-year juveniles must have been killed, and the bone bed includes first-year juveniles of at least two different seasons.

Also, there’s this detail:

Human tooth marks on the bison-set have been identified on 192 specimens (Table 5). They are predominantly located on ribs (76.3%) and, to a lesser degree, on unidentified flat bones (7.3%) and hyoids (5.7%), 48.4% of which are associated with other anthropogenic modifications, such as cut marks (Supplementary Online Material [SOM] Table S1). A large range of human tooth marks produced during the consumption of the carcasses have been characterized and recorded, although scored and pits are the most abundant.

The authors emphasize that the bison bone bed is a distinct pattern of evidence about hominin hunting. There are non-bison bones in the layer from many different animals, but these constitute only a small fraction of the total, and none of them have clear evidence of human modifications such as cutmarks. Both the bison and non-bison material has evidence of carnivore modifications. But the representation of large parts of the axial skeleton and the evidence for human butchery show that humans had primary access to the bison carcasses, and carnivores ravaged the bones later, after the humans were done with them. The ribs bear abundant evidence for removal of the viscera, which is rare at other sites mainly because ribs are much less likely to be present.

In total, these show that effective communal hunting of large mammals was part of the behavior of hominins in Europe well before the appearance of Middle Paleolithic toolkits. Rodríguez-Hidalgo and colleagues compare the site to others across western Eurasia:

This is fully consistent with other well-documented and thoroughly taphonomically investigated assemblages from the Middle Pleistocene, such as Bolomor, Cuesta de la Bajada, and Gran Dolina TD10.1 and Gran Dolina TD6 in Spain (Blasco, 2011; Saladié et al., 2011; Domínguez-Rodrigo et al., 2015; Rodríguez-Hidalgo et al., 2015) Schöningen in Germany (Voormolen, 2008; Starkovich and Conard, 2015; Van Kolfschoten et al., 2015) and Gesher Benot Ya'aqov and Qesem in the Near East (Rabinovich et al., 2008; Stiner et al., 2009); in which hunting emerges as the main method to acquire animal carcasses.

As such, these particular hominin-accumulated sites have a very different profile than carnivore-accumulated sites, in that the hominins have been much more selective on a few prey species. That’s compared in this figure from Rodríguez-Hidalgo and coworkers, which looks at the number of individuals of prey species represented (on the x-axis) compared to “evenness”, which is a measure of how equally represented different prey species are. The carnivores here do not focus their attention on single species as often as these hominin accumulations have done:

Prey species evenness from Rodríguez-Hidalgo et al. 2017.

This is notable again for the similarity between the Gran Dolina bison bone bed and the Mauran bison accumulation, from the last interglacial.

The bison bone bed also differs from both earlier and later layers of Gran Dolina. This suggests that the cultural and environmental conditions that gave rise to bison kills near the site were localized in time.

Obviously these kinds of sites where dozens of kills give sufficient information to look at the entire process of hunting and butchery across a limited period of time are relatively rare in the archaeological record.

A site like Gran Dolina tells us that such kill-dense sites cannot be strictly representative, because over hundreds of thousands of years of hominin activity at the site, the bison bone bed represents only a limited time, and very different hunting preferences predominated at other times.

Link: The tension between telling a story and telling readers what is going on

An article by Chip Scanlan at Nieman Storyboard displays an insider’s knowledge of the nut graf, the part of a journalistic article that explains to the reader what the rest of an article is about: “Nut grafs: Overused, misused — or merely misunderstood?”.

The big idea in the article is the tension between writers who want to create a kind of suspense in their stories, and readers who want to know what the heck a story is about before they commit to reading all of it. The so-called “nut graf” is an invention to enable readers to get the gist of an article early on, so that if they choose not to read the rest, they’ll know more or less what they missed. It’s not exactly Cliff’s Notes, but it helps.

Evidently some writers hate the idea because the nut graf is usually written in an omniscient voice that pulls readers out of a narrative.

We toss out the word “story” every day. “Great story!” “Why wasn’t my story on the cover?” And the highest praise, “I wish I could write that story.” But as the legendary writing coach Jack Hart noted, most journalistic pieces are not stories, but articles, well reported and organized, accurate and fair. But no child ever looked up from their pillow at night, eyes wide with excitement, and beseeched, “Daddy, tell me an article!”
Stories have characters, not sources; settings, not addresses; dialogue, not quotes. Instead of nut grafs, they use transitions—a term from the musical world— subtle, elegant turns that mark the passage from one scene, subject, or place to another.

This is a tension that often occurs in science writing. When I teach, I like to assign narrative pieces by writers that tell human stories. But I also like to assign textbook chapters and scientific articles.

Stories are valuable, and they can do things that more neutrally written articles cannot. But a story can be the ultimate cop-out for a writer who is addressing scientific subjects. And the idea that an article must be “well reported and organized, accurate and fair” belongs to a mythological world that probably never existed.

A written piece is a machine for thinking. There will be no single right way to write, because readers are rightly suspicious of any writer’s motives and abilities. Narrative can be an effective way to convey some information, but readers should be wary of the biases hidden in stories about individuals, especially when the stories are a sugar-coated way of introducing a scientific subject.

Neandertal slaughters

Back in 2016, Mark White, Paul Pettitt and Danielle Schreve published an interesting analysis in which they compared how Neandertals hunted and butchered animals at five “kill sites” in France, Germany, and Poland: “Shoot first, ask questions later: Interpretative narratives of Neanderthal hunting”.

The five sites represent different species of prey animals: bison, horse, rhinoceros, reindeer, and aurochsen. The sites vary in geological age from the last interglacial some 120,000 years ago up to the height of the last glacial, around 50,000 years ago.

Individually, each site shows Neandertals making effective use of geographic features of the landscape, such as changes in topography, narrow side branches to valleys, and marshes next to steep hillsides, which enabled them to channel fleeing animals into situations where they were cornered, and then to kill indiscriminately.

Our conclusions indicate that Neanderthals did not necessarily pre-select individuals from a herd, who they then isolated, pursued and killed, but rather ambushed whole groups, which they slaughtered indiscriminately. There is strong evidence, however, that Neanderthals were highly selective in the carcasses they then chose to process. Our conclusions suggest that Neanderthals were excellent tacticians, casual executioners and discerning diners.

As a group, these sites show Neandertals maximizing the chance of successful kills by using topography, while minimizing need for chase and tracking injured animals.

Kill sites are very different from cave sites or occupation sites. The kill sites have evidence for butchery tools and the remains of carcasses that were not transported away by Neandertals to other places. They may represent many instances in which Neandertals encountered and killed animals in the same geographic place, and indeed several of these sites are inferred to represent a large number of visits by Neandertal groups, over many years.

As such, these sites present a very particular kind of evidence about Neandertal behavior. These are among the places where Neandertals were most lethal, using their knowledge of the landscape and of animal behavior to give them an advantage.

For example, the site of Mauran, France, represents the accumulated remains of more than 130 bison, killed over a millennium or more by Neandertals who were using the local landscape to trap and ambush groups of animals:

The original excavators have already used the landscape and character of the bison assemblage to provide a reconstruction of Neanderthal hunting at Mauran (Farizy et al., 1994). In this account, the topography at the site — a rocky limestone barrier fronted by open vegetation and marshy ground — provided a natural trap into which Neanderthals could drive and corral bison (Farizy et al., 1994 see Fig. 2). The stratigraphy and differential bone preservation were taken to indicate that the site represented hundreds of separate events over several centuries with individuals and small groups taken each time, rather than a few massive North American jump-style slaughters.

The Taubach, Germany, site preserves evidence of multiple single-animal kills of rhinoceros over time. This site shows that for these large herbivores, at least, Neandertals were targeting young individuals with a strategy that separated them from other adults.

But sites like these do not document a single-kill strategy for any of the medium-sized herbivores that were the Neandertals’ main prey animals. The authors boldly put forth a challenge:

However, we eagerly await a convincing Middle Palaeolithic example of a targeted, isolated killing of a medium-large gregarious herbivore.

What they are emphasizing is that Neandertals made use of their knowledge of the social and flight behavior of these animals to kill them. Separating one animal from a group, using persistence hunting methods as wolves do, would not have been a good strategy for the physical abilities of Neandertals. Hunting communally, killing animals at a topographical and seasonal advantage, and making use of the most valuable parts of carcasses was.

Nonetheless, these communal kills probably were not all of Neandertal hunting behavior. To begin with, they don’t summarize the entire diversity of the behavior of the herbivores. It is equally part of the social behavior of many of the medium-sized herbivores to have single bachelor males and bachelor groups. With such groups, opportunities to obtain single prime-age animals or groups of prime-age animals would not have been uncommon. Sites that represent repeated kills over many years might include such instances within the overall pattern but they would be obscured within the statistical distribution of all the others.

Rock shelters and caves, which are not kill sites, present a record that is a different compound of events over time. The faunal remains at these sites were transported by Neandertals from primary kill sites. Those transport choices bias the record to some extent. They also represent multiple kill sites, which in some instances were places where different species of animals were the preferred prey.

A great example is Abric Romaní, a rock shelter near Barcelona. Juan Marín and coworkers recently examined the age profiles of horses and deer excavated from the site: “Neanderthal hunting strategies inferred from mortality profiles within the Abric Romaní sequence”.

The equids display prime dominated profiles in all of the analyzed levels, whereas the cervids display variable profiles. These results suggest that the Neanderthals of Abric Romaní employed both selective and non-selective hunting strategies. The selective strategy focused on the hunting of prime adults and generated prime dominated profiles. On the other hand, non-selective strategies, involved the consumption of animals of variable ages, resulting in catastrophic profiles. It is likely that in the selective hunting events were conducted using selective ambushes in which it was possible to select specific prey animals. On the other hand, encounter hunting or non-selective ambush hunting may have also been used at times, based on the abundances of prey animals and encounter rates.

OK, that was two “on the other hands” in the abstract.

The core result is that the horses are strongly biased toward prime age adults, while the deer are a mix of prime age adults, juveniles, and older adults, a catastrophic profile. The deer look like they could have been hunted in ways reflected by the kill sites discussed by White, Pettitt and Schreve. The horses, on the other hand, look like very selective exploitation. That might mean that the Neandertals were hunting bachelor groups, or they had situations that enabled them to hunt lone adult horses more effectively.

Marín and coworkers further suggested that the reason for exploitation of juvenile deer may have been economic rather than dietary:

Binford [24] observed hunting events in which the Nunamiut (Tulekana and Kakinya) exclusively hunt young reindeer in order to obtain soft leather for clothing. Lithic use-wear analyses at Abric Romaní show that worked skins existed within the sequence, with work on fresh leather being more common [134]. In addition, lithic functionality studies in level Ja relate denticulate and notch features to the hardening of hides [81]. In the Abric Romaní sequence, although young individuals have been identified in almost all of the studied levels, they do not reach 71% of the total, as in level I. Therefore, in this level, the hunting of cervids seems to have been specifically intended to obtain this prey of low economic return, possibly to obtain their hides.

A similar suggestion was made by White and colleagues, who recognized that Neandertals may have killed animals in larger groups occasionally for valuable parts such as hides.

In contrast to the paper by White and coworkers, John Speth (2018, as well as earlier papers) has emphasized that Neandertals could not have been indiscriminate in their choice of age and condition of prey animals, for nutritional reasons. Unlike carnivores, hominins cannot subsist on diets with high protein proportions and must seek out prey animals with fat available.

I wanted to point to that argument here, and I will return to it at greater length. A brief consideration suggests there’s no contradiction between the ambush hunting patterns documented by White and coworkers, and the need to consume high proportions of dietary fat. The notion of overkill followed by selective consumption would have enabled Neandertals to choose the most fat-rich parts of carcasses for immediate consumption.

Endarkenment now

An article in Salon by Phil Torres attempts to source and check quotes in Steven Pinker’s best-selling book released last year, Enlightenment Now: “Steven Pinker’s fake enlightenment: His book is full of misleading claims and false assertions”.

In brief, Pinker borrowed a quote from Bailey, who didn’t cite the original source and who lifted the quote from its original context to mean the opposite of what Zencey had intended. This led Zencey to confess to me, “how this guy [i.e., Pinker] managed to become a public intellectual in fields so far removed from his expertise is something to wonder at.”
If this were a single misdeed, one could perhaps forgive it. But it’s not the only error of this sort within just one page in [Enlightenment Now].

The piece goes on to examine a number of claims within a chapter of the book that covers “existential risk”. The Salon essay links to a longer, in-depth examination</em> of part of that chapter.

The present document does precisely this by dissecting individual sentences and paragraphs, and then placing them under a critical microscope for analysis. Why choose this unusual approach? Because, so far as I can tell, almost every paragraph of the chapter contains at least one misleading claim, problematic quote, false assertion, or selective presentation of the evidence.

I have mixed feelings about writers like Pinker. From results like this, it would seem to be extremely challenging to be a responsible scholar while writing books with a broad interdisciplinary scope. To be candid, it is easy for a critic who has a personal animus against an author to go through any book and find “errors” that are actually disagreements of opinion or emphasis. The more prominent the author, the more likely such critics will exist, like trolls on the internet.

From experience, I can say it is not possible to write on the internet very long without attracting critics. A scholar who makes writing public begins a conversation. Any honest scholar should have the humility to acknowledge that no research plan will turn up every relevant fact. Exposing written work to the public will bring out observations, facts, and references that a writer may have missed.

Pinker’s critics have varied axes to grind, and it’s important to examine those motivations when assessing their criticisms of his work.

But not all critics are trolls, and not all disagreements are matters of opinion. I don’t think the degree of flubs that are coming out in Pinker’s work can be explained away as inevitable results of public exposure, and I don’t think he is uniquely targeted by critics with politics that disagree with his. Any writer who aspires to have his work read by hundreds of thousands of people, whose words may influence political and business leaders, should be held to the highest standard of accuracy.

For me, the bottom line is that the kind of money harvested by Pinker’s books should support a few fact-checkers and research assistants to check the footnotes and provide additional sources.

A related thought today on Seth Godin’s blog: “The honor code”:

An honor code: The simple expectation that we trust you, that you call your own fouls, that you act honorably even if you think no one is watching…
As we think about implementing this, we need to decide between, “people are so dishonorable, it makes no sense to trust them” and, “the only way to help people become more honorable is to trust them.”

“Calling your own fouls” is an important concept to good scholarship. It requires self-examination. Likewise, a rigorous adherence the first law of holes.

Darwin Day in St. Louis

Next Saturday, February 9, I will be in St. Louis speaking at the Darwin Day event at Washington University: “Institute for School Partnership: 2019 Darwin Day Celebration”.

DATE: Saturday, February 9, 2019
TIME: Registration & continental breakfast 8 am. Program 8:25 am to 1:00 pm
LOCATION: Room 202, Life Sciences Building, Washington University Danforth Campus

If you’re in the St. Louis area and interested in evolution, it should be a great event. I’ll be talking about “New discoveries and insights into our African origins”.

Family Tree DNA database now available to FBI investigators

Salvador Hernandez reports for Buzzfeed that Family Tree DNA has now opened its database of genetic information from more than a million users to the FBI: “One Of The Biggest At-Home DNA Testing Companies Is Working With The FBI”.

In December 2018, the company changed its terms of service to allow law enforcement to use the database to identify suspects of “a violent crime” such as homicide or sexual assault, and to identify the remains of a victim.
In a statement, Bennett Greenspan, the president and founder of Gene-by-Gene, Family Tree's parent company, said the firm would not be violating its terms of privacy to its customers, despite the FBI's access.
"We came to the conclusion that if law enforcement created accounts, with the same level of access to the database as the standard FamilyTreeDNA user, they would not be violating user privacy and confidentiality," Greenspan said.

That’s a devious interpretation of the terms of service. Since anyone can upload data to the service, it is probably already true that people have uploaded genome data that doesn’t belong to them. It might therefore have been trivial for the FBI to work within the Family Tree DNA service even without any formal permission from the company.

Then again, here’s a shot from the company’s current website:

Family Tree DNA privacy promise


How worried about this should anyone be?

When it comes to it, the science is clear that once a critical number of people have voluntarily shared their genome data, essentially every individual will have a third cousin or closer relative in the database. Crimes today are being solved not because criminals themselves have uploaded their data, but because their distant relatives have done so.

With more than a million individuals, the Family Tree DNA database kickstarts that process. It’s a larger dataset than the public ones that law enforcement agencies used last year to catch the Golden State Killer and others.

Family Tree DNA is not, however, unique or essential to the process. I don’t think there’s much question that a million people in the U.S. would voluntarily provide their genomes to a law enforcement database if it were marketed to them. “Help us catch the killers!” It really wouldn’t take much more than that, and the government could have its own version. Or the present freely available upload sites would just have to grow larger, which they already are on track to do.

Such databases have different blind spots, since genealogy buffs contribute their DNA for different reasons than the genome neighborhood watch. But we are inevitably within a couple of years of law enforcement being able to track down a third cousin of any genetic sample they collect.

In light of this inevitability, it would be wise for the FBI and government to think carefully about how they want citizens to participate in the process. The Family Tree DNA process today requires issuing a warrant to obtain information about distant relatives of a suspect DNA sample.

I think it is unwise to create a situation where courts are issuing such warrants to distant relatives solely because of partial DNA matches. It’s entirely avoidable by relying upon people who actually volunteer to help authorities search for criminal matches.

None of these tactical issues adjust the underlying reality: We are very near the point when every individual will be identifiable through DNA matches, even if that individual has not contributed his or her own DNA samples to any database.

NSF grants track language used in abstracts

Hmm… this is interesting from David Markowitz: “Text analysis of thousands of grant abstracts shows that writing style matters”.

Two other results were telling about the NSF data. First, using fewer common words was associated with receiving more award funding, which is inconsistent with the NSF’s call and commitment to plain writing.
Second, the amount of award funding was related to the writing style of the grant. Prior evidence suggests that we can infer social and psychological traits about people, such as intelligence, from small “junk” words called function words. High rates of articles and prepositions, for example, indicate complex thinking, while high rates of storytelling words such as pronouns indicate simpler thinking.
NSF grant abstracts with a simpler style – that is, grant abstracts that were written as a story with many pronouns – tend to receive more money. A personal touch may simplify the science and can make it relatable.

Correlation is not causation, standard exceptions apply, your mileage may vary.

Genomes and the complicated history of baboons

Today, Science Advances has released a paper by Jeffrey Rogers and coworkers on the genome diversity of six species of baboons: “The comparative genomics and complex population history of Papio baboons”.

This paper represents a significant advance in scientific knowledge of the history and evolution of baboons across Africa. The genus Papio arose around the same time as our own genus, Homo, and the diversity of baboons across Africa today reflects a history of divergence and mixture during the last million years.

Our team pointed to an earlier stage of knowledge of baboon population history in our 2017 paper, Homo naledi and Pleistocene hominin evolution in subequatorial Africa”. Our interest has been the geographic distribution of Homo species in Africa, in light of the occurrence of H. naledi in the later Middle Pleistocene.

Anthropologists have recognized baboons as a very relevant comparison to hominins for more than a hundred years, and in particular the work of Clifford Jolly has focused upon baboon population differences as useful models for the adaptive differences that may have separated ancient hominins. I wrote about Jolly’s perspectives on hybrid zones and introgression in baboons back in 2005: “Look to the baboons; there will you your insights find!”

Baboons of course are not alone as relevant comparisons to hominins. Carnivores and ungulates both show some similar biogeographic patterns to baboons, at least based upon mitochondrial DNA variation. But the whole-genome analysis of these species within Africa is only beginning. So getting a clear whole-genome picture of baboon population history is one of the first views of species that lived in the same habitats as ancient humans.

First, a quick introduction to baboon species. There are six of them:

  1. Papio papio in extreme west Africa,
  2. Papio anubis across most of the Sahel to Sudan and Ethiopia, and from there south into the Lake Tanganyika area,
  3. Papio hamadryas in Eritrea, the Afar region, and across the Red Sea in the Arabian Peninsula,
  4. Papio cynocephalus on the eastern coast of Africa from Somalia to Mozambique,
  5. Papio kindae from Angola to Zambia, and
  6. Papio ursinus in southern Africa.
Map showing geographic distribution of six species of baboons in Africa from Rogers et al. 2019

Rogers and coworkers looked at the genomes of a total of 17 baboons, which represented “2 to 4 individuals” from each of the 6 species. This was not a sample chosen to probe geographic diversity within each species. Only two individuals (both from the Aberdare region of Kenya) were chosen to examine historically recent hybridization and introgression. So it is a very restricted picture of variation, and that is important to keep in mind when trying to make sense of the phylogenetic inferences in the paper.

The study provides a composite tree giving a topology for the relationships of the six species as well as approximate times when they diversified:

Baboon composite phylogeny from Rogers et al. 2019

The main features of this picture include:

  1. A primary split between northern (P. hamadryas, P. anubis and P. papio) and southern (P. cynocephalus and P. ursinus) lineages around 1.4 million years ago.
  2. A subsequent hybridization of a northern and southern branch to form P. kindae. Neither of these branches is closely aligned with any of the other five extant lineages.
  3. A ghost lineage from the base of the genus contributing around 10% of the ancestry of P. papio.
  4. The speciation of today’s species date to between around 400,000 and 800,000 years ago, with the exception of the hybrid origin of P. kindae which was within the last 100,000 years.

These events were sketched out using f-statistics and the CoalHMM software, both of which have been used for hominins as well as chimpanzees and gorillas, so this tree is very comparable to the tree presented for chimpanzees and bonobos by de Manuel and coworkers (2016), for example.

Additional analyses in the paper look at phylogeny using different methods, including Bayesian and parsimony phylogenetic analyses. These give rise to various results that are mostly unreconcilable, and it’s not obvious to me that they add anything to the paper, since none of them are capable of handling the degree of mixture that the f-statistics infer for P. kindae.

The big limitation of the study is the lack of geographic coverage of variation within each species. The small sample size should give rise to more tree-like phylogenetic results than a broader sample. The same is true of ancient hominin genomes: We have only a handful of ancient genomes, and the results are very treelike. But the introgression from Neanderthals and Denisovans in living people samples a broader number of populations from these groups and is not so simply treelike as the high-coverage ancient genomes themselves.

So I think we still have a way to go to really understand the importance of hybridization in the ancestry of today’s baboons.

A hint of upcoming results comes from the two Kenya individuals of P. anubis examined in the study. Both individuals reflect historically recent introgression from P. cynocephalus. The passage describing this is very interesting:

Our results also shed new light on the historical dynamics of hybridization between P. anubis (a northern clade species) and P. cynocephalus (a southern clade species), which has previously been reported in southern Kenya near Amboseli National Park (17). Behavioral observations and microsatellite-based analyses support recent introgression from P. anubis into P. cynocephalus since the 1980s (25, 26). Our analysis of genome-wide haplotype block sharing indicates that a P. anubis individual from the Aberdare region of Kenya, more than 200 km north of Amboseli, is also admixed with P. cynocephalus, carrying ~546 Mb of nuclear DNA derived from P. cynocephalus (fig. S7). If we assume that this resulted from a single admixture event, then it is estimated to have occurred about 21 generations (~220 years) ago. However, other more complex explanations are also possible. The second individual from the P. anubis Aberdare population also carries P. cynocephalus haplotypes, but these shared genomic segments are fewer and shorter and likely result from more ancient introgression. Consistent with other studies (27), our findings suggest that there have been multiple episodes of gene flow involving these two species over a considerable time span and that the effects of past hybridization extend far beyond the current hybrid zone. This complexity may well be representative of the complexity of other known baboon hybrid zones (10, 12, 15, 18, 19, 28).

The heterogeneity between two individuals in the same population with respect to recent introgression is really striking. These two genomes emphasize that the movement of individuals and spread of genes between two hybridizing species are more of a turbulent interface than a smooth cline. We have a hint of this turbulent interface in the Oase genome results from Romania, an individual with a high and recent degree of Neanderthal ancestry.

The taxonomy of baboons seems to be relatively stable now. Twenty years ago, there was substantial debate about how many species should be recognized across Africa. At that time much less was known about the reproductive fitness of hybrids. Jolly (2001) emphasized the “indiscriminate” hybridization of different populations of baboons in captivity :

In captivity, all allotaxa appear to hybridize indiscriminately (Jolly, unpublished data), and there is no evidence for hybrid breakdown, behavioral incompatibility, or intrinsic sterility. Similarly, there is no evidence that Papio, baboon allotaxa ever avoid interbreeding when they meet in the wild, though many boundary areas have yet to be investigated. The fact that documented baboon hybrid zones are narrow, in spite of the lack of obvious, intrinsic barriers to gene-flow, strongly suggests that they are the result of secondary contact following range oscillations (Barton and Hewitt, 1985; Hewitt, 2001; Harrison, 1993).

The current story is very different from Jolly’s (2001) account. As discussed by Rogers and coworkers, scientists have observed a number of indications that hybridization among species of baboons is not fitness neutral.

Another topic of broad interest is the origin of reproductive isolation among incipient species (1). One expectation for the genus Papio is that, given the timing of the radiation and the degree of morphological and behavioral differentiation among species, incipient barriers to gene flow may be evident between some pairs of species. Studies of the present-day hybrid zone between northern clade P. anubis and southern clade P. cynocephalus find no readily apparent barriers to reproduction between these species (17, 26). However, studies of captive P. anubis × P. cynocephalus hybrids document significantly elevated frequencies of craniodental anomalies in hybrids, especially hybrid males, indicating some degree of genetic incompatibility (39). Field studies of the hybrid zone between P. ursinus and P. kindae describe a deficit of hybrid individuals carrying Y chromosomes from P. ursinus and mtDNA from P. kindae compared to the converse (18). This suggests that when hybridization began between these two forms, some type of barrier (premating or postmating) reduced the frequency or fertility of matings by male P. ursinus with female P. kindae, while the converse mating type was more successful (18). Last, P. anubis and P. hamadryas differ substantially in their social organization and social structure (11, 28, 40). Among anubis baboons, both males and females are polygamous. Hamadryas societies are multi-level, with “harem”-like, one-male breeding units (OMUs) as basal social entities. In these OMUs, the single adult male defends exclusive access to one or more adult females. Other differences in sex-specific dispersal and social relationships are also observed (11). Despite the dramatic differences in social systems, these species hybridize in the wild (28). Hybrid males can achieve substantial reproductive success, at least in groups consisting mainly of hybrids (19). There is no clear evidence for a barrier to gene flow between the species, although the geographic distribution of phenotypically recognizable hybrids is narrow.

The various regional populations of chimpanzees (Pan troglodytes) are classified as subspecies, and they originated across roughly the same time span as these species of baboons. Chimpanzee regional populations do not exhibit the same variation in mating system, coat coloration, and morphology as baboons. So there is a good phylogenetic species concept (PSC) argument for baboons being different species that is not there for chimpanzees. Meanwhile, the observations noted by Rogers and colleagues are evidence that the baboon species should be recognized under the biological species concept (BSC). Chimpanzees and bonobos are different species under both BSC and PSC criteria, but neither concept suggests that the various subspecies of chimpanzees should be recognized as species instead.

The divergence among Neanderthals, Denisovans, and African ancestors of modern humans also took place across approximately the same time period, around 600,000–700,000 years. With both chimpanzees and Neanderthals, there is evidence for occasional introgression of genes among the ancient populations.

Aside from recent hybridization and the possible “ghost lineage” contributing to P. papio, the baboon picture of ancient gene flow is not clear from the results presented by Rogers and colleagues. The mismatches among Bayesian and parsimony phylogenetic results, and the different results obtained by looking at Alu insertion data, all suggest that incomplete lineage sorting or ancient reticulations may have been very important to the baboon pattern of variation.

To summarize, this paper provides important new data about the divergence of baboon species. Much more data would be valuable, especially samples that would tell us more about the geographic variation within baboon species and the long-term record of hybridization and introgression at the boundaries of these species.

Link: Wired on the legacy of Jim Watson

A fascinating read in Wired by C. Brandon Ogbunu: “James Watson and the insidiousness of scientific racism”.

That one person separates me, an African-American computational biologist, from James Watson—Nobel Laureate and mouthpiece of racist opinions—presents a quandary. For years, I have reveled in the powers of DNA, yet one of the people most associated with its discovery has made abhorrent comments about my race. The dilemma raises several questions: How does it feel to be a black scientist who owes much to James Watson in general, and in my case, is linked to his specific pedigree? Is it much ado about nothing, or might the black scientist occupy a special place in modern conversations about scientific racism?

I like the final thought experiment.

Link: How accurate are age-at-death estimates for older adults?

IEEE Spectrum covers a new research study looking at the accuracy of a method for determining the age-at-death of skeletal remains: “Errors Found in Forensic Software Meant to Assess Age-of-Death of Skeletal Remains”.

Research conducted by biological forensic scientists at North Carolina State University and the University of South Florida has uncovered “serious problems” in a recently released forensic software application available online called DXAGE that is supposed to predict the age-at-death of skeletal remains based on bone mineral density.
The study, published in the Journal of Forensic Sciences, reported that the software’s predicted ages could be off by 14.25 years on average when DXAGE-generated results were compared against known samples. The system’s accuracy was particularly poor for the remains of elderly individuals.

Since the software does nothing more than apply a set of algorithms to observations taken on bone, its results can be no better or worse than the variability of age-related changes in bone. That variability is really big! Biological anthropologists recognize that we cannot accurately estimate the age of older individuals, errors of 15-20 years on age estimates are very common.

So my reaction to this story is, “How did a method for age-at-death estimation that made claims of greater accuracy get started in the first place?”

It is just misleading to think that software gives better accuracy than an expert armed with the same statistics. It’s the white coat phenomenon. Unfortunately, the misperception of “computer precision” has huge influence on juries.

Bone mineral density changes are an interesting consequence of aging, and they have been the topic of some debate as applied to the ages of Neanderthals and other prehistoric people. When we look at ancient populations, there are the possibility of nutritional differences, lifestyle differences, and genetic differences that may have influenced both the peak bone density and the pattern of age-related bone loss.

All these are reasons why we cannot be very definitive about the age-at-death of most older adults in the fossil record. Once they have progressed to a point far enough beyond third molar eruption that tooth wear is not an accurate indicator of age, there are few biological indicators that would not also vary across populations for many reasons.