Thursday, May 04, 2006

The earliest life on earth

I didn't attend the media event last night that RSR described because I was at the 3rd part of Science, Education, and the Public. Last night's theme was "The Earliest History of Life: Solution to Darwin's Dilemma," by Dr. William Schopf of the Department of Earth and Space Sciences at UCLA.

Schopf is a major expert on the fossil remains of the earliest life, cells found fossilized as early as 3.45 billion years ago. He set as his goal the solution of one flaw that Darwin cited as potentially fatal to his theory:

if the theory be true, it is indisputable that before the lowest Cambrian stratum was deposited … the world swarmed with living creatures … why we do not find rich fossiliferous deposits belonging to these assumed earliest periods… I can give no satisfactory answer. … The case at present must remain inexplicable; and may be truly urged as a valid argument against the views here entertained.
From the 10th chapter of the 6th edition of the Origin.

The immediate answer to the dilemma is that there are fossil deposits. It appears that hardened outer coverings evolved around the Cambrian, increasing the number and quality of preserved fossils, but Schopf set as his goal the demonstration of evidence for live well before the Cambrian.

In Darwin's day, the search for fossils meant looking for large multicellular remains, typically preserved bones or shells. And that takes us back about 500 million years. The earth is about 4.55 billion years old, so if we envision the entire history of earth as one 24 hour period, Darwin knew about the period between 9 pm and midnight, while we've now gotten back to around 5:30 am. All of human history only stretches back to about 11:58:40.

If indeed life before the Cambrian wouldn't fossilize well, we need to look for other evidence of the earliest life. The first line of evidence Schopf discussed was stromatolites. These are fossilized bacterial mats, accretions of silt and cyanobacteria, built up over millenia. These occur today, but are also common features of the fossil record, stretching back billions of years.

The second line of evidence was based on isotopic ratios. The standard photosynthetic pathway preferentially takes up the lighter isotope of carbon, carbon-12, leaving carbonaceous deposits enriched in carbon-12 compared to limestone (which more closely preserves the balance of carbon-12 and carbon-13 from the atmosphere). Deposits of carbonaceous rock as young as 3.5 billion years ago can be shown to have carbon-12 enrichment, evidence of photosynthesis.

But to show that those deposits come from life, it would be nice to have an actual fossil. And that's what Schopf is confident he found in rocks from 3.5 billion years ago, in the middle of Australia. By taking thin slices of ancient rocks, he and his students could hunt for cellular fossils, and managed to find what looks like a filament of bacteria. The size of the purported cells are self-consistent, and within the range of modern bacteria. The ways that the filaments decay is consistent with how undisputed bacterial fossils decay. Progressively thinner slices show that the cells are disc shaped, carbonaceous, and not consistent with being geological by-products.

His claims were strong, and I don't know what the current thinking is on those specimens. Robert Hazen's description of these fossils in Gen-e-sis: The Scientific Quest for Life's Origins is less confident. While there's broad agreement that these early rocks have carbon of biological origin, it's less clear that these geological structures are necessarily fossils. Schopf has marshaled a lot of evidence for his claim, but there are also plausible doubts. It may be that the chert is a later inclusion in older surrounding rock, and that these filaments could be of inorganic origins.

As Hazen writes: "If your philosophical bias is that life emerges rapidly and is thus ubiquitous in the universe, then any tantalizing sign of life – carbon smudges or scrappy fossils – will provide encouraging supporting evidence. If you think life rare or unique, then the tenuous character of the same data will prevail. Consequently, to prove either case – that an ancient rock does or does not hold evidence for life – is extraordinarily difficult."

My inclination is to accept Schopf's claims. Other researchers in South Africa claim to have found fossilized cells, Schopf's group have a solid track record of ancient fossilized cells, and pushing the earliest history back to 3.5 billion years doesn't really change much. Biogenic carbon can be found in equally ancient rocks, so finding cells in that era doesn't really surprise me. And whether life emerged early and widely is not just a philosophical position, it's an empirical claim, one that these sorts of data provide evidence for, as do lab experiments on producing life.

There's a broad point to be made though, about how we detect things. Schopf didn't follow anything resembling what the IDolators suggest is "the ordinary practice of sorting through things we alternately attribute to law, chance, or design." Dembski likes to imagine that people have some sort of hierarchy, first elimination "law-like" behavior, then randomness, and then you're left only with "design." But let's look at how Schopf actually worked.

We know what life looks like, so Schopf set specific criteria. There had to be particular kinds of carbon, there have to be cells, the cells have to be of roughly equal size. There is what Schopf described as a "cascade of traits" that something must possess before he'd call it life. Each characteristic provides a test, and if the test is falsified, we reject the hypothesis that the item was alive (or we rethink our hypotheses about what life should be like). Failing to falsify some specific hypothesis isn't proof, and that's where the biases Hazen describes kick in. In a Bayesian framework, we'd call those biases our "priors" and they have a major influence on the posterior probability we assign to an object representing "life." In time, however, more evidence will balance out our differing prior probabilities, and we will develop a consensus about when life began, regardless of our personal biases.

IDolators believe it's "the ordinary practice" to successively reject everything but our final hypothesis. But that's not how science works. We test each hypothesis on its own. That's the "ordinary practice." Right now we can't reject that the fossils are life or that they are natural geological phenomena.

Next Tuesday, KU's Joe Heppert will discuss the recently approved science standards and "alternatives" to evolution.