An Interview with Jannie Hofmeyr in Three Parts
What is life? This is a question biologists have almost forgotten to ask. Apart from a few lonely souls who occupy themselves with astrobiology or the origin of life, we no longer seem to wonder about life itself. Maybe we have lost the forest for the trees? Maybe we consider the problem solved? Genetic programs all the way down. Dialectical Systems is a forum that unites philosophers, cognitive scientists, biologists, and others who still care about the nature of the living, the central question of biology. Our small community owes much to Robert Rosen (Wikipedia), one of our most outstanding intellectual ancestors.
It is difficult to categorize Rosen. His work stands truly on its own, a unique, foundational, and extremely original contribution to the science of the organism. His 1991 book “Life Itself” is often discussed, has accumulated hundreds of citations, yet, it is rarely really understood, and its true meaning and impact remain heavily disputed. Here, I join biochemist Jannie Hofmeyr (Stellenbosch University, Google Scholar, Wikipedia) for a three-part conversation about Rosen, his continuing importance, and Jannie’s recent work that not only renders Rosen more accessible, but also connects his abstract theories to our empirical knowledge of biochemistry and structure of the cell.
From the 1950s onward, Rosen set out to apply the principles of relational biology, as originally elaborated by his supervisor and mentor Nicolas Rashevsky (Wikipedia), to the organization of whole living systems. He adopted category theory as his tool of choice. That is certainly one of the hurdles on the path toward a better understanding of his work. What he attempted was a high-level functional analysis of the relations between the components that make up an organism, that make it alive.
Category theory describes such relations by mathematical mappings into which all kinds of abstract structures can be packed. Rosen arrived at a peculiar set of such mappings, which he called a metabolism-repair, or (M,R)-system. His simple diagram of mappings became iconic among those who understood the importance of his insight. It captures what Rosen called “closure to efficient causation” (and which others have called organizational closure), an organizational principle that remains at the heart of the best theories of the organism today.
Rosen’s conjecture was that all living systems (and only living systems) are closed to efficient causation. Closure, therefore, is the defining feature of life. He even derived his own idiosyncratic notion of “complexity” from this conjecture, considering only living systems (and systems that contain living beings) to be truly complex (while non-living systems are merely complicated).
The key point of Rosennean complexity is that the behavioral repertoire and evolutionary potential of a living system cannot be fully captured by algorithmic computer models. Or, in his own words, organisms have no largest model. Only partial aspects of their dynamics can be formalized. Organisms can and will always surprise us. Interventions on complex systems always have unintended side effects. Our formal descriptions of these systems remain forever incomplete. Every once in a while, a living system will do something that is unexpected, not yet captured by our formal model. This is Rosen’s central insight, and also his most controversial legacy. Rosen’s work has been interpreted in different ways. Its technical and abstract nature sometimes make for challenging reading.
In the first part of our interview, Jannie and I introduce Rosen’s idiosyncratic diagrammatic formalism and talk about his famous conjecture, providing our own interpretation of his writings, which we believe to be both close to its author’s original intent, and the most useful rendering of his arguments for current work in empirical biology and biochemistry.
In the second part of our conversation, we move on to Jannie’s own extension of Rosen’s work (see here, here, and here), which shows that the original relational diagram does not map to actual cellular processes. Jannie offers an alternative topology of the model (also mentioned, although only in passing, in Rosen’s “Life Itself”), and shows how each of its components can be identified with particular aspects of the living cell. Jannie also explains why this refutes a simple reductionist approach to the understanding of cellular architecture, and circumscribes those aspects of the cell that can be understood by purely mechanistic means (and those that can’t).
In the third and last part of the interview, we switch gears to speculate about the implications of Rosen’s and Jannie’s work on some rather fundamental questions, such as the (non-)computability of living systems and a post-mechanistic science that we will need to properly understand life and its open-ended evolution.
Our conversation is best consumed in its full length, since each part builds on the preceding ones. However, in case you are interested in specific parts of the conversation, we have provided labeled chapters on YouTube that allow you zoom in to the relevant part of each recording.
We remain in the hope that our conversation may inspire more people to engage with Rosen’s highly fascinating opus. An extended list of relevant references is provided in the description of each YouTube video.
My sincerest gratitude goes to Jannie Hofmeyr, not just for sharing his deep insights into the works of Rosen and the organization of living cells, but also for the many hours of inspiring discussions (and many a glass of wine) that we enjoyed during our stay at the Wissenschaftskolleg zu Berlin in 2014/15. This is where my interest in life itself got properly rebooted. In the background of all this is my late supervisor and best of all mentors, Brian Goodwin (Wikipedia), who would have loved this conversation. Last but not least, I would like to thank Andrea, Matteo, Charbel, and Leo for featuring our conversation on Dialectical Systems. I very much hope to contribute more content to this fantastic and much-needed initiative in the future.
Comments on “Mapping Theories of Life into Cell Biochemistry”
Ugo Corda says:
Rosen’s work always fascinated me, and it looks like Jannie’s model is a great enhancement of Rosen’s model.
Having said that, it is not sufficient for a scientific model of reality to sound great, it also needs to make verifiable predictions. The best way to achieve that in this particular case would be, as discussed in part 3, to create a living organism from scratch using the knowledge derived from Jannie’s model. I understand the ethical implications of going down that road, but it is something already underway in any case (see all the current efforts in synthetic life). It would actually be interesting to know, in that respect, which models are being used by synthetic life researchers, if any, and if they know of Jannie’s model.
Leaving synthetic life aside, Jennie’s model would still need to prove itself by making some novel and/or better predictions compared to rival scientific models. Just as an example, let’s remember that reductionism, with all its defects and limitations, is what most of today’s medical practice is based on (with all its defects and limitations), so a rival model would have to do better than that.
It is fine to talk about anti-reductionism and post-mechanistic approach, but we need to show concrete results that are superior to what we get from competing models of reality. Otherwise it will be just a fancy narrative, no matter how much it can change our perspectives on reality (many fancy narratives do that anyway, and it is no guarantee that those narratives have any scientific value).