Recently, Peter Molnar passed away of pancreatic cancer. There are already several solid obituaries and reminiscences of him, and more are on the way. In short, Peter was a leader (if not the leader) in understanding mountain building and its impact on the climate system. He started with seismology, pushed to collect data from some of the most remote and difficult places to work, while also making contributions to plate reconstructions (some of the first with actual uncertainties came from him and Joann Stock). As he recognized the importance of gravitational potential energy in mountain evolution (in part through working with people like Phil England and Greg Houseman) and the continuum nature of deformation within large orogens (in part from work with Paul Tapponnier) he recognized the significance of paleoelevation and began scouting what others were doing in this regard. This led to working with climate scientists and paleontologists to find robust ways of getting around numerous problems in estimating the elevation of ancient mountain belts. This also led to a couple of papers (with Phil England) challenging the inference of surface uplift from erosion. Even as he plunged into unfamiliar regions like paleontology and climate science, he continued to work on instability of mantle lithosphere and its role in continental deformation. As the world of climate science became more familiar, he got interested in some largely climate science type questions such as the cause of northern hemisphere’s Ice Ages (Pleistocene). He was a prolific author with several hundred papers and an h-index over 100. Beyond that, he had an encyclopedic memory for authors and papers, a characteristic that was a huge help in developing personal connections within the field.
I am not here to talk about his contributions (and the many omissions in my summary above should be in indication of that); I list the above to make clear just how successful a scientist Peter was and to provide a hint of the breadth of his interests. I want to talk about how he felt about the scientific enterprise.
So some context. Peter was my PhD advisor at MIT from 1981-1987, and then he has been a colleague on the faculty of the University of Colorado Boulder since he arrived in 2001. At MIT he would be gone for months at a time, off on projects in Asia (mainly). He was far more present at CU.
I still do not fully understand how he got so much done. His door was nearly always open; he regularly would eat lunch with different folks, students or faculty, and discuss whatever was interesting to his meal mates. He enjoyed hosting out of town visitors (a habit enabled by his wife Sara). I bothered him about trivial matters far too often, and he was never dismissive or even impatient. As a graduate advisor at MIT, he said his role was to provide support for us students and we were to go work on what seemed most promising; he only suggested projects when asked, a very different approach than, say, the English system of posting thesis topics that students could apply to complete. So the aid we students provided his research goals was frequently minimal. And yet papers were reviewed, proposals written, literature read and catalogued. He was disciplined to a level few of us would ever reach even for short periods of time. In short, he was a scientist’s scientist. He was fully and emotionally engaged in the study of the Earth. This meant a lot to him.
This led to many conversations about the health of science as an intellectual endeavor. He and those with whom he discussed this were hardly the first to worry about this. M. King Hubbert raised this in a 1963 address. An opinion in Nature in 2016 arguing for less publication spurred discussion in our group. The constant press to publish papers, to get grant money, to admit more students was leading to the diminution of actual science being done by scientists. Peter would remind others of the work of de Solla Price, whose 1962 book claimed that the number of scientists doubled every 10-15 years while the actual growth of achievement doubled only every 30 years or so (Peter would claim the numbers were 15 years to double the number of good scientists and 5 to double the number of scientists; regardless, the factor of three difference was about right). This suggests that good science is drowning in a sea of mediocrity or worse. This greatly worried Peter who, in an email, ended with the thought “I do not envy the younger generation and especially those motivated by quality instead of by finding ways to game a system destined to be nothing (if not a negative quantity) more than beancounting.”
So Peter took on the system, and this is arguably the least visible contribution he made for science, but one that is important and worth sharing. At first it was ridiculing the discussions revolving around tenure or hiring where the only information presented was something like number of papers, number in prominent journals, amount of grant money brought in. “But what have they done?” was the essence of Peter’s response. “Did you read their papers?” became a feared challenge, in no small part because Peter had read the papers and wasn’t shy about sharing his opinion. It wasn’t long before search and promotion committees made a point of reading the papers least the wrath of Molnar descend on them. While Peter despaired to the end that our hiring and promotion procedures were stuck in bean counting, we who were there both before and after Peter’s arrival knew better.
The other aspect of geoscience he lamented was the proliferation of models, specifically ones with lots of free parameters. He would paraphrase a favorite saying he attributed to Greg Houseman: You start with one thing you don’t understand in the Earth, and then you make a model to reproduce it, and then you have two things you don’t understand. While some of his criticism was unfair, the difference between a good model and a bad one is not easily perceived. A good model removes large areas of model space from further consideration, showing that some parameters cannot yield results in accord with observation. But as the number of free parameters proliferates, it becomes that much harder to restrict the limits on any given parameter. At the extreme, these models produce a hypothetical earth that depends on a set of plausible but unconstrained assumptions. Peter was always looking to minimize free parameters through tricks like non-dimensionalization; the goal was to find the fewest parameters that mattered. In this battle, Peter was on the losing end (at least so far) as animations of geodynamical processes that are possible with such models have great appeal to many earth scientists, to say nothing of those outside the field.
Many argue that we need more interdisciplinary work. The problem is that as we get more specialized, there are fewer and fewer people who can see enough of their own field, let alone adjacent ones, to know where crossing disciplines is productive. Peter was one of those scientists who could understand different fields well enough to pursue important research topics across disciplines. So as we try to move forward, it is probably worth keeping in mind the issues that troubled Peter, for those issues might be hindering the development of new scientists capable of following in his footsteps.