john hawks weblog

I have a review of Marlene Zuk's new book, Paleofantasy, in this week's Nature: "Evolutionary biology: Twisting the tale of human evolution" [1].

I can't replicate my review here, but for people who have access to Nature I thought I'd bring attention to it. And if you don't have access, I wanted to share a couple of my reactions.

It was a fun book for me to read. Zuk brings a light-hearted skepticism to a broad array of topics in human evolution. She took as her focus a collection of "paleo-advice" ideas: barefoot running, paleo diet, back-to-nature parenting advice. She then added some uncritically-accepted scientific notions about our evolution, such as the idea that agriculture was "the worst invention ever devised". To each of these topics, she brings an array of recent science questioning or disproving the assumptions. The result is not to debunk ideas, but to give a fuller (and more nuanced) perspective on how much we know (and don't know) about our evolution.

The serious issue underlying all these topics, which Zuk recognizes, is the difficulty of reconstructing Pleistocene environments. Some hypotheses assume a fairly detailed model of ancient environments -- the so-called "environment of evolutionary adaptedness". But ancient humans lived in an array of environments, more different than each other in many ways than different parts of today's globalized world. We are unquestionably living in environments no ancient humans knew, in population size, density, disease, lifespan, and many other ways. But in other ways, our difference from some ancient people is trivial compared to their diversity. Are we well-adapted to live in cities? Perhaps not in some ways, but maybe in others.

Probably the best part of my review to share is the end:

As an anthropologist, I observe that Zuk's use of the term 'fantasy' is just an emphatic way of describing the hypothesis-forming that is essential to evolutionary science. We play with hypotheses, explore their predictions and try very hard to falsify them. So it is, in a way, unremarkable that so many hypotheses proposed by anthropologists about ancient environments now seem to be wrong — and, in a few cases, even ridiculous.

It means that science is working. Genomics, high-resolution climate records, and microscopic and isotopic evidence have changed our understanding of what the past has to offer. With that in mind, let the next round of palaeofantasies begin.

Zuk's "very brief" overview of human evolution is a lot shorter than in other recent books on the topic. I found this to be a merciful change -- how many times do I really need to read about the Australopithecus-to-humans timeline? Readers who don't already know the basic timeline are unlikely to pick up the book, I would guess. Still, if you're looking for a "latest news" about early humans, this book is not directed that way. Where it excels is its coverage of recent evolutionary changes and the shifts in Holocene environments and genetics.

The book is not without its weak points. Without quite enough of the "paleo-advice" topics to carry the whole story, there were some real differences in tone across the chapters, with some a bit drier than others.

People coming to this book for "the right answer" about ancient environments are not going to find it. There is no right answer, at least not a scientific one, for many of the topics covered here. Zuk has done well to talk to a range of scientists, covering these different aspects of our evolutionary history, and discuss the reasons for their disagreement.

I wish scientists would do that for themselves more often!

References

A new paper by Anna Druzhkova and colleagues examines the ancient mtDNA sequence of a putative 33,000-year-old dog from Razboinichya Cave in the Altai region: "Ancient DNA Analysis Affirms the Canid from Altai as a Primitive Dog" [1]. The paper's analysis is a simple application of phylogeography, showing that the mtDNA of the Altai dog fits in a clade with a number of pre-Columbian New World dogs:

The domestication of dogs from the grey wolf is well accepted [1]. However, the timing, location and number of domestication events is still actively debated [2]–[5]. The archaeological record provides unequivocal dog remains beginning about 14,000 calendar years (cy) ago [6]–[7] requiring a domestication that predates agriculture. Putative dog remains ranging in age from 31,000 to 36,000 cy [2] [8]–[9] have been questioned as potentially representing aborted attempts at domestication, or morphologically unique wolves [4]. A full mitochondrial genome analysis of modern dogs suggests an origin in southern China around 16,000 years ago [10], whereas an extensive nuclear genome-wide SNP analysis supports a Middle East and European origin [11], which is more in accordance with archaeological data. Here we isolated, sequenced and analysed 413 nucleotides of the mitochondrial DNA control region from a putative dog specimen dated as approx. 33,000 cy from the Altai Mountains in central Asia. Only a single specimen - namely the Goyet dog (36,000 cy [2]) predates the Altai dog and hence it is thus far the second oldest known specimen assigned morphologically to the domestic dog [8].

The evidence of dog domestication has developed piecewise over the last several years. A number of Upper Paleolithic skeletal specimens have morphological dimensions inconsistent with wolves, but comparisons of the genetics of recent dogs has tended to argue against such early domestication.

In the current paper, the mtDNA similarity of the Razboinichya canid and pre-Columbian American dogs is pretty persuasive evidence that this specimen came from an early population ancestral to the dogs of northeast Asia, which would later enter the New World. This paleontological specimen shows that the mtDNA phylogeny of modern-day dogs does go way back into the Late Pleistocene, which argues against a single recent domestication. Still, the mtDNA is not the strongest possible source of evidence, since present-day dogs can be found across many of the clades that include mtDNA from wild wolf populations.

Curiously, Druzhkova and colleagues did not include the Goyet canids in their mtDNA comparisons. An analysis of 57-bp of the mtDNA of these dogs was carried out by Germonpré and colleagues [2], showing that the Belgian Upper Paleolithic dogs have a diverse range of mtDNA haplotypes, across several clades of the wolf genealogy. The current paper bases its mtDNA cladogram on 400-bp sequences, so they aren't strictly comparable, but it is nevertheless interesting that the other putative early dogs are not part of this clade including pre-Columbian dogs and the Altai specimen.

The earlier description of the Razboinichya canid by Ovodov and colleagues [3] suggested that the specimen was part of an early domestication event that was "arrested" by the Last Glacial Maximum.

We suggest that the pre-LGM Goyet and Razboinichya canids are unlikely to be the ancestors of post-LGM dogs. These canids most probably are both “proto” or incipient dogs that did not persist long enough to found enduring lineages, since no putative dog remains have been found at adjacent sites in western and central Europe and in Siberia occupied during the LGM. The ecological changes caused by progressive cooling almost certainly caused social and settlement pattern changes severe enough to have disrupted the domestication process and prevented the evolution of fully domesticated dogs.

Such a scenario would reconcile the early skeletal evidence for dogs with the conclusion that recent dogs come from a small mtDNA population.

But I think it's too soon to conclude that today's dogs don't have deeper Pleistocene roots. As zooarchaeologists have been finding more and more possible evidence of dogs, they may be filling in the record (for example, with apparent dogs from the Gravettian Předmostí site [4] and from the later Upper Paleolithic of Kesslerloch, Switzerland [5]). I wonder whether a good actualistic study of dog deaths and remains in small-scale human societies would give rise to clearer expectations about how many dog skeletal specimens we should expect from Upper Paleolithic contexts.

References

The Raw Story reports on a new study of 40,000 community college students in Washington state, which concludes that online courses are not as effective as classroom-based courses for this student population: "Research shows everyone does worse with online learning". In particular, the story emphasizes that minorities and men do worse in the online setting.

“We found that what the students really wanted more of was a connection with their instructor. They wanted more guidance from their instructor. They wanted their instructor to be able to help motivate them with their passion and their caring for their students and how the students did,” Jaggars continued. “The thing about college students, you know, they come in with a lot of anxiety and insecurity about whether college is the right place for them, whether they can do this kind of work, they need an instructor that’s really supportive and enthusiastic about the material and communicates that enthusiasm to the students.

“If you think about a MOOC, you know, 200,000 students, and one instructor, I’m not really sure how that, you know, how that connection can be made,” Jaggars said.

Apples to oranges, I would say. A massively open online course (MOOC) is not going to replace small group instruction for students whose motivation is low. Nor will they replace small group instruction for motivated students in very specialized areas. There is a process of "learning how to learn" in college.

What I find interesting is how professor-focused this model of learning actually is. Appearing today in Inside Higher Ed is an essay by Andrew Ng, one of the founders of MOOC purveyor Coursera: "Learning from MOOCs". He emphasizes that professors who adapt their material to the MOOC format discover just how much they have been doing in the classroom that really isn't about learning the material:

Adelman discovered that in putting his course online, he became more focused on what students are experiencing, even though he wasn’t in direct contact with them. “When I lectured, I had to ask myself at all times ‘What is it that I want my students to learn?’ In the old-fashioned lecture hall I was an entertainer, more self-focused rather than teaching-focused, but I was not conscious of this dynamic until I put a course online for the first time,” he says. “For me, the lectures alone were a source of continuous learning and adaptation.”

If we take seriously the idea that students learn in different ways, that they come to us with different skills and competencies, then we must recognize that some will be cheated by any mode of delivery. Being "an entertainer" in a lecture hall is a great way to reach some students, to communicate enthusiasm with the subject. Others will be turned off by this style, will wonder why they are paying to listen to this professor who loves the sound of his own voice.

MOOCs and other online courses today are reinventing lots of wheels. For example, creating compelling lecture-centered content is a storytelling and film editing problem, not specifically a lecture problem. Finding good solutions that will work in courses is a matter of bringing in expertise from these specialties.

But one area where today's MOOCs are experimenting with genuinely new innovations is in the way that students interact with each other. When we consider the differences between an online course and its classroom equivalent, the professor really is not much of a difference. The biggest difference is the number and pattern of contact with other students. Some students do great in isolation, but others work better in a group. A course that shapes the pattern of online interactions among students thereby shapes the way they can learn from each other's progress. But how?

But through today's technological advancements, online courses are very much alive. They are part of an ecosystem that, if nurtured through community discussion forums, meetups, e-mails, and social media (like Google+ hangouts), can flourish and grow. This allows each class’s community to take on a life of its own, with a distinct culture that’s defined at least as much by the students as the instructor, and which even skillful instructors can only guide, but not control. Nearly every instructor that I’ve spoken to has been surprised by the deep desire of students to connect with each other as well as with the teaching staff and professor.

Lots of different modes of interaction, different courses experimenting with different patterns. Some online courses have a huge range of student interactions, from total self-study to the creation of real-life meet-ups to discuss the course. Others have had notable failures, from total crashes of online commenting systems to abusive students writing anonymously on course message boards.

I am very excited about the potential of technology to create new modes of teaching. I think that online presentation can reach new communities of learners who are not served by universities or community colleges, but who are no less deserving of great learning opportunities. But creating these new learning environments will inevitably siphon off some people who previously could only obtain college-level educational opportunities at great financial expense.

At the same time, I am skeptical about effectiveness of online content. The range of experimentation now is very wide. That shows that there are lots of attempts at innovation, but very little selection favoring the best approaches. We need rigorous attempts to outline the conditions in which particular online communities facilitate learning.