My most influential paper was a complete accident

Images: selections from Heard 1992, Patterns in tree balance among cladistic, phenetic, and randomly generated phylogenetic trees. Evolution 46:1818-1826. Orchid photo (Belize) © S. Heard.

 While I think my research is interesting and important, I’m well aware that few of my individual papers are likely to change the world.  Really, this is a normal feature of science: most progress comes from the accumulation and synthesis of many small results, not from a single mind-blowing paper.  But some of my papers have been more influential than others.  It’s interesting (and to be honest, a bit galling) that what I think is my most influential* paper (Heard 1992, Evolution, Patterns in tree balance among cladistic, phenetic, and randomly generated phylogenetic trees) was an accident.

I wrote the paper as a PhD student, as a side project unrelated to my dissertation research.   It turned out to be an accident in two ways.  First, the answer I got from my research had nothing to do with the question I originally asked.  Second, purely by chance I published it just as the field started to get interested in my accidental answer.

balanced and unbalanced phylogenetic treesMy influential paper had to do with a property of phylogenetic trees called “balance”.  A phylogenetic tree represents the evolutionary history of a clade (a group of lineages descended from a common ancestor), with each branching a speciation event.  A particular tree might be balanced (bush-like; top right) or imbalanced (comb-like, bottom right).  Those patterns suggest different things about patterns across the clade in speciation rates.  In a perfectly balanced tree, any two lineages of the same age have exactly the same number of descendents.  In an imbalanced one, in contrast, some lineages have diversified much more than others – for example, at the root of the imbalanced tree at right, a single ancestor gave rise to two lineages – one eventually including seven species, but its sister, only one.

I got interested in tree balance for a weird reason. As a PhD student I went to a lot of talks by systematists and other evolutionary biologists.  At the time (the late 1980s and early 1990s), systematics was still experiencing convulsions as members of two methodological traditions (“cladists” and “pheneticists”) were at each other’s throats**.  I had the impression that phylogenies presented in talks by pheneticists tended to look balanced, while those presented by cladists tended to look imbalanced.  This bothered me, because (with some rare exceptions on both sides), everyone had the same goal: to reconstruct true evolutionary history.  Could it be that the methods were leading them to different answers?

I managed to answer that question: they weren’t – at least, not when the data were informative and the datasets comparable.   I could explain that in more detail; but I don’t want to, because that question, in hindsight, was a boring one.  I’d rather turn to the much more interesting answer I came to by accident.  Here it is, in this graph:

phylogenetic tree balance as a function of tree sizeThe hollow symbols show a measure of tree balance for real phylogenies (from 0, or perfectly balanced, to 1, or perfectly imbalanced).  This clearly depends on tree size (the number of species included), but that mattered only to the boring question.  What’s much more interesting is the difference between the hollow symbols and the solid ones.  Those solid points are the expected tree balance for a model in which all species have equal probabilities of speciation (equal speciation rates) and imbalance arises only through stochastic variation in who speciates when.

orchidThe difference between the data and the equal-rates model is really, really important.  We’ve suspected at least since Victorian times that some lineages of organisms have been more successful at diversifying than others – that speciation rates have been higher (or extinction rates lower) in some lineages than others.  Think, for example, about such hyperdiverse lineages as orchids, warblers, and beetles, or about such depauperate ones as gingkos, penguins, and lungfish.  The difference between the two traces on my graph establishes two things.  First, speciation-rate variation isn’t a special property of a few oddball clades, but rather a general property of life on Earth.  Second, we can quantify that variation with a rather simple metric based on phylogenetic trees (which were at the time just beginning to be common in the literature, but have since become a deluge of biblical proportions).  So my paper did two really interesting and important things – by accident.

That’s exciting enough, but it gets better.  Around the beginning of the 1990s, people were noticing tree shape and thinking about it in a bunch of different ways.  Over the next 25 years, interest grew and grew, as people put various aspects of tree shape to work (just as a few examples) testing for ecological drivers of diversification, calculating phylogenetically weighted measures of conservation value and biodiversity loss, measuring community structure to detect processes like competition and environmental filtering, and even deducing means of spread and evolution of viruses.  (For those interested in all this science, a couple of reviews are here and here.)  Not one of these things had occurred to me when I set out to study tree shape, but my paper was in the literature at the right time to influence a whole lot of subfields that were going a whole lot of places***.

Now, perhaps you don’t care about tree balance, but you’re still reading because you’re convinced that surely I must have some kind of broader point to make.  I do; but I’m afraid it’s not a very original point.  Stories like mine are, of course, why we need to fund “basic” or “curiosity-based” research.  Governments love to pretend that we can do science more efficiently by only funding work that addresses a well-defined economic or societal need – but expecting this to work is monumentally foolish.  Everyone knows, and most scientists have, a story like mine.  It might be a  famous ones, like the accidental invention of Post-It notes; or it might be more obscure, like the accidental discovery of phosphorus by a gold-seeking alchemist****.  It might even be a story that hasn’t finished unfolding yet; after all, it may take years for the accidental importance of your work to become clear.  Science has run on these stories for hundreds of years, with smart people following their curiosity around the twists and turns that nature throws at them.  Actually, in a larger sense accidental discoveries aren’t accidental at all.  They may be individually unpredictable, but in the aggregate, they’re the fruits of an “unleash curiosity” strategy.  In other words, “accidents” aren’t occasional oddities; they’re the backbone of science.

We should tell our accidental stories often, and we should make this strategic and historic context clear.  Society needs to know that there’s no better way to move science forward than to unleash the smart and curious, and wait for “accidents”.

© Stephen Heard ( October 6, 2016

*^It isn’t my most cited paper (and certainly not my most over-cited one).  Two review papers have been cited much more heavily.  But it is, by a small margin, my most-cited primary-results paper; and by a much longer margin it’s steered the direction of its subfield more than anything else I’ve ever published.  By the way: yes, all these papers are paywalled. In the unlikely event that you can’t find a copy: e-mail me.

**^Usually, but not always, metaphorically.  Fortunately, the field rapidly put this debate behind them and never again got into personality-driven conflicts over the merits of rival methodologies. Riiiiiight….

***^It isn’t always directly cited, but I can trace influence in other ways.  To be clear, I’m not claiming I founded any of these lines of work – only that my paper was able to help them along.

****^I promise to tell this story some time.  It involves alchemists and princes, boiling stale urine, skullduggery, and art.

5 thoughts on “My most influential paper was a complete accident

  1. Pingback: Is this blog a “science blog”? If not, what is it? | Scientist Sees Squirrel

  2. Pingback: Open science, phosphorus, and economic impact | Scientist Sees Squirrel

  3. Pingback: Stephen King and me | Scientist Sees Squirrel

  4. Pingback: Where do I find the time to blog? | Scientist Sees Squirrel

  5. Pingback: Well, THAT could have been awkward: job-interview edition | Scientist Sees Squirrel

Comment on this post:

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

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

Twitter picture

You are commenting using your Twitter 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.