Robert Sokal, 1926-2012

Biostatistician Robert Sokal died on April 9. Like most ecologists, I knew Sokal primarily through his canonical statistical textbook with Jim Rohlf, Biometry. But he was of course much more than the author of a classic textbook. In his research, he was a pioneer of clustering methods, originally in the context of pheneticist phylogenetics. His personal life was dramatic; he fled Nazi Germany as a youth and was raised in China, eventually becoming part of one of the world’s finest evolutionary biology groups at Stony Brook University. The trend towards increasing quantitative rigor in ecology is a long term trend; Robert Sokal was hugely important in driving that trend.

Remembrances from Joe Felsenstein and Michael Bell, Chris Jensen, and Greg Mayer.

10 thoughts on “Robert Sokal, 1926-2012

  1. Thanks for this, some good stuff in those links. Nobody mentioned the Sokal Hoax however.

    I had completely forgotten about Numerical Taxonomy until I read this. I wonder where phenetics sort of stands amongst the evolutionary biologists these days.

    • C’mon now Jim, you know that’s a different Sokal.

      There aren’t really any pheneticists any more. People estimate phylogenies in other ways.

      • Oh, ok, sorry. I thought you were making a bad joke. The Sokal hoax was by a physicist, Alan Sokal if memory serves.

      • Never under-estimate the depths of my ignorance.
        But yeah, Alan Sokal sounds correct.

  2. Anyway, the book is titled Numerical Taxonomy, not Numerical Cladistics. Whatever I read of it, if anything, was long ago, but given the title, is it not possible that they were developing a system for taxonomy that was not-necessarily geared toward phylogeny construction, i.e. more general than that? After all, you don’t have to be phylogenetically oriented to construct a taxonomic system based on phenological similarities (phenetics).

    • Well, I think both pheneticists and cladists back in the 1960s had some philosophical views on what they were doing that look a little odd today, when most everybody agrees that phylogeny reconstruction is the goal, and that fitting models of DNA sequence evolution is the way to achieve that goal. Philosopher and historian David Hull wrote a book called Science as a Selection Process, about how science advances by a process analogous to evolution by natural selection. He used the phenetics vs. cladistics wars as a case study. A somewhat unusual choice of case study in some ways, given that a lot of what was being fought over wasn’t really resolvable by data. But I found it a quite interesting book, with a lot of trenchant and historically-grounded remarks about the sociology of science and scientists.

      • Thanks Jeremy. Somewhat relatedly, I wonder what the state of the science is in terms of classifications that are purely functional/ecological in orientation, i.e. no real attempt to track phylogenetic history at all. This is where I can see things like Sneath and Sokal being exceedingly useful, at least potentially. I guess I should look at some issues of Functional Ecology.

      • Well, “trait-based” approaches are trendy right now in community ecology. Don’t worry about taxonomic identities, worry about what “functional traits” your organisms have. But the choice of traits, and the classification of species into “functional groups”, tends to rely mostly on intuition and data availability (e.g., what traits are easy to measure). There could be a role for Sneath-Sokal type approaches here somewhere. But what ecologists mean by “functional traits” is rather broadly and loosely defined, so this could be one of those cases where trying to be more quantitative amounts to trying to precisely measure something that we can’t actually precisely define.

      • Hmmm, interesting, and ties into the discussion about appropriate choice of an index. I’ve tried reading the book “Plant Functional Types”* more than once and sort of gotten stopped each time by difficulties related to what you describe there (i.e. why did these folks pick this set of traits and how would you measure them?). Part of it was just my own impatience however.

        * Smith et al., 1997. Plant Functional Types.

Leave a Comment

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.