Time for a revision? Maureen O’Malley and Russell Powell on Major Transitions, part 3

Maureen O’Malley and Russell Powell say that the major transitions framework is in need of repair. They have a point, or rather several good points. I have looked at their criticisms of three different versions (the original framework as laid out in the book by John Maynard Smith and Eörs Szathmáry, Rick Michod’s ‘evolutionary transitions in individuality‘ framework, and Szathmáry’s revised ‘Major Transitions 2.0‘). But what is their proposed fix, and will it have the intended effect?

Figure 4 from O'Malley and Powell 2016. Two major aeons of evolution (modified from Falkowski 2006). ‘Gya’ stands for ‘billion years ago’; the date for the origin of photosynthesis may need to be pushed back (see Crowe et al. 2013).

Figure 4 from O’Malley and Powell 2016. Two major aeons of evolution (modified from Falkowski 2006). ‘Gya’ stands for ‘billion years ago’; the date for the origin of photosynthesis may need to be pushed back (see Crowe et al. 2013).

First, a quick recap. O’Malley and Powell’s main criticisms are:

  1. Disunity: the major transitions are not a coherent category (or natural kind, in philosophese). I agree that this is true for Maynard Smith and Szathmáry’s original formulation, not so much for Michod’s revised framework.
  2. Progressivism: the major transitions framework advocates a ‘ladder of life’ view of evolution. I agree that some readers may mistakenly infer a progressivist agenda, but disagree that it is inherent in any version of the framework.
  3. Missing events: some transitions that meet the criteria for major transitions have nevertheless been excluded from most discussions. This is, surprisingly, largely true: chloroplast acquisition, for example, is rarely discussed in this context.

These problems, in O’Malley and Powell’s view, would be largely mitigated if

…metabolism as an interactive force [were] afforded a more central place in philosophical and scientific theorizing about major transitions.

How does this fit with any of the criteria for major transitions? Does it result in “…entities that were capable of independent replication before the transition [and] can replicate only as part of a larger whole after it,” in new means of transmitting information, or in a transition in individuality? It doesn’t, and O’Malley and Powell reject the idea that it should:

We conceive of the MTE project very differently. In our view, the purpose of any account of major evolutionary transitions is to sketch an answer to the following Big Question: ‘‘How did the biological world get here from there?’’

…These difficulties stem not only from theoretical disunity, but also—and more fundamentally—from giving up on the grander explanatory ambitions of major transitions theory. MTE in its revised and original forms carves out only one small slice of the Big Question (namely, hierarchy), and then carves out only one part of an answer to that one small slice of that one Big Question. [emphasis added]

and later:

…if the aim of major transitions theory is to give an account of game-changing events that have shaped the large-scale organization of life on earth—that is, to sketch an answer to the Big Question—then we think the theoretical limitations of a strict ETI-based approach—or even a slightly more relaxed one such as MTE 2.0—are too severe. [emphasis added]

‘How did the biological world get here from there’ is indeed a Big Question; it’s just not the big question that the major transitions framework was designed to answer. The framework for giving ‘an account of game-changing events that have shaped the large-scale organization of life on earth’ might be described as ‘evolutionary biology’; no smaller category will suffice. The major transitions framework is a part of evolutionary theory in the same way that behavioral ecology, neutral theory, evo-devo, and inclusive fitness are parts of evolutionary theory. Each ‘carves out only one part of an answer’; none claims to explain, by itself, the entire sweep of biological history.

By setting the bar so high, O’Malley and Powell have inflicted an unfair burden on the major transitions framework. The ‘grander explanatory ambitions’ are ones they have imposed, and the criticism of not explaining everything would apply equally to all of the subfields I’ve listed above (and many others). The Big Question is too big, and the ‘grander explanatory ambitions’ too grand, to expect major transitions theory to answer.

But what of their proposed fix of focusing on metabolism? Metabolism, and major metabolic innovations, are hugely important, no disagreement there. The origins of electron transfer mechanisms, anabolic and catabolic pathways, and photosynthesis, in O’Malley and Powell’s words, “…transformed the biogeochemical world and the space of evolutionary possibility.”

In part 1 I said of adding metabolic transitions to the major transitions framework that

I think the proposed cure is not only worse than the disease, but less effective than some less invasive treatments.

Now I’ll explain why I think that. The problem that including metabolism is intended to solve is that the major transitions framework “…overlooks other equally important types of events that it should aim to explain.” There is no question that it excludes important types of events. But are these types of events it should aim to explain? Why should it? The major transitions framework, especially Michod’s ‘transitions in individuality’ version, is formulated to explain the nested hierarchy observed in some branches of the tree of life, not the entirety of biological and biogeochemical evolution.

Nor was it intended to be a list of the most important changes; if it were, it should include not only the origins of metabolic pathways and photosynthesis, but also of flight, predation, terrestrial life…add your own favorite. And this is the crux of why I say that the proposed cure is worse than the disease. By adding metabolic changes, which do not meet any of the criteria of major transitions, we would effectively be changing the criteria for inclusion from ‘results in a new kind of individual’ (in Michod’s version) to ‘is really important.’ We would change a category that is something like a natural kind to something that is nothing like a natural kind. In this, the proposed solution would actually worsen one of the problems O’Malley and Powell identify, that of disunity.

What about ‘less effective than some less invasive treatments’? O’Malley and Powell have identified some real and important problems with the major transitions framework. In its older version, it includes events that don’t meet its stated criteria. In Michod’s revision, some of the events that meet the criteria are rarely discussed. This seems to me to suggest an obvious solution: why not just use Michod’s criteria of transitions in individuality and include the (mostly) missing transitions? O’Malley and Powell are not satisfied that revised versions focusing on transitions in individuality solve the problems they’ve identified:

The revisionist approach, as we have noted, is to purge all inconsistent events from the list and retain only events that constitute ‘genuine’ ETIs. But this would mean including all known transitions in individuality, and it is not clear how ‘major’ some of these transitions are (e.g., tertiary plastids). Furthermore, if transitions in individuality are what drive up levels of hierarchical complexity, as the revisionist model supposes, then this might imply that tertiary plastid endosymbioses should be regarded as the ‘culmination of evolution’—the utmost hierarchy of previously independent units of selection becoming integrated into a single unit. [my emphasis]

Yes, this would mean including all known transitions in individuality. Good. No, it is not clear how ‘major’ (in the sense of ‘really important’) some of these transitions are, but as I said in part 2, ‘really important’ is not one of the criteria.* We shouldn’t be worrying about what is the ‘culmination of evolution’ (there is no ‘culmination of evolution’), but if tertiary plastid symbioses turn out to be the highest hierarchical level (the one that includes the largest number of increasingly inclusive levels), what of it? This is only a problem if we want the framework to be progressivist, and we all agree that we don’t.

O’Malley and Powell don’t think that simply applying Michod’s criterion consistently is a good solution, but I find their reasons unconvincing:

To make matters worse, some MTE reformers believe that every interspecific mutualism (‘ecological fusions’, in which evolutionary fates are entwined) would need to be included in a revised model consisting of ETIs alone (e.g., Kerr and Nahum 2011: 128–129), and (still worse) that ETIs must also include the myriad of events in which non-kin mobile genetic elements are integrated into genomes (e.g., Ågren 2014). Including these smaller events is clearly a mistake if ‘major’ transitions are being sought, even if the criteria for being ‘major’ remain rather fuzzy.

Because these things have been suggested doesn’t mean we are bound to abide by them. If they result in new kinds of individuals, we should include them; if not, not. There is nothing fuzzy about this definition, except for the inevitable difficulty in definingindividual‘. The criterion of being ‘major’ (in the sense of ‘really important’) is the problematic element here, and that, not true transitions in individuality without biosphere-altering implications, is what should be discarded.

These problems indicate to us that whatever major transitions are, shifts in individuality cannot be the sole criterion for their designation. In response, one might hold that an evolutionary event is a major transition only if it implicates an ETI and has major evolutionary consequences.

Or better still, forget about consequences and apply the ‘transitions in individuality’ criterion consistently. Some ETIs (evolutionary transitions in individuality) have had major consequences, others not, or not yet. I would rather throw out the bathwater of consequences than the baby of shifts in individuality.

In summary, O’Malley and Powell bring up some good criticisms of the major transitions framework. Disunity is a real problem for some versions, although I think Michod’s revisions have largely resolved that. Most versions are missing events, and the absence of plastid acquisition from most discussions is especially troublesome. I don’t, however, think that either the oxygenation of the Earth or metabolic innovations should be included within this category.

There is utility, I think, in grouping transitions that result in a new kind of individual: they are something like a natural kind, and we may learn something by seeking generalities that apply to all of them. Rick Michod would argue, and I would agree, that we have learned some things by doing so. Including events that clearly had important implications, but did not result in a new kind of individual, creates more problems than it solves. It worsens the disunity problem that O’Malley and Powell have identified in the original framework, and it expands the category of major transitions to the point that it is no longer useful.

A better solution, I think, is to include all of the events, and only the events, that result in a shift in individuality. This would result in the inclusion of two of O’Malley and Powell’s missing (types of) events, the acquisitions of chloroplasts and mitochondria. It would also, as they point out, result in the inclusion of events lacking major evolutionary consequences. This is only a disadvantage if we think ‘really important’ should be one of the criteria, and I don’t. Having big consequences is surely contingent, and we can’t say which recent transitions in individuality might have big consequences in the future. Applying the transition in individuality criterion consistently would not only be philosophically coherent; it would also increase the number of events from which we can draw inferences. By accepting that major transitions theory only addresses one part of the Big Question, we arrive at a repair that is not just easier, but better.


*In this sense the name of Maynard Smith and Szathmáry’s book, The Major Transitions in Evolution is unfortunate (and I don’t think I’m the first to point this out). There have been lots of major (in the sense of important) transitions in evolution, and by singling out a short list as ‘The’ major transitions, they invited this sort of misunderstanding. In this sense, Michod’s terminology, ‘evolutionary transitions in individuality’ is, IMO, a big improvement.


Stable links:

Ågren JA (2014) Evolutionary transitions in individuality: insights from transposable elements. Trends Ecol Evol 29:90–96.

Crowe SA, Døssing LN, Beukes NJ, Bau M, Kruger SJ, Frei R, Canfield, DE (2013) Atmospheric oxygenation three billion years ago. Nature 501:535–538.

Falkowski PG (2006) Tracing oxygen’s imprint on earth’s metabolic evolution. Science 311:1724–1725.

Kerr B, Nahum J (2011) The evolution of restraint in structured populations: setting the stage for an egalitarian major transition. In: Calcott B, Sterelny K (eds) The major transitions in evolution revisited. MIT Press, Cambridge, pp 127–140.

Michod RE (2000) Darwinian Dynamics: Evolutionary Transitions in Fitness and Individuality. Princeton University Press, Princeton, NJ.

O ’Malley MA, Powell R (2016) Major problems in evolutionary transitions: how a metabolic perspective can enrich our understanding of macroevolution. Biology & Philosophy, 31, 159–189.

Szathmáry E (2015) Toward major evolutionary transitions theory 2.0. Proceedings of the National Academy of Sciences, 10104–10111.


  1. says

    So what we really have here are two different major approaches to the transitions partially mixed in with a third. The first is Maynard Smith and Szathmáry’s notion of increasing complexity in terms of combination and specialization. Individual molecules combining into cells and then the molecules specializing, then cells combining into larger cells (prokaryote to eukaryote with eukaryotes absorbing mitochondria and chloroplasts), then cells combining into collectives and the cells within the collectives specializing to create multi-cellular organisms. Then the organisms interact and form societies – like ant colonies and human civilizations.

    The second is that of metabolism. Life is fundamentally chemical, and in any chemical equation, overall entropy must increase. In order for there to be decreased entropy in biological systems and thus increased complexity, there must be a method of increasing the entropy of the environment. Thus new energy sources and ways of extracting energy are necessary. The main example of this is the oxygenation event, where oxygen in the atmosphere dramatically increased at least partially due to photosynthesis, with that oxygen then providing energy for a dramatic increase in more complicated organisms.This is the position taken by O’Malley and Powell, with specific chemical pathways laid out in detail.

    Lets’ call these two approaches MTC (Major Transitions in Complexity) and MTM (Major Transitions in Metabolism) with the understanding that metabolism is also dealing with complexity, only more indirectly. What both of these approaches ignore is major transitions in the Process of Evolution. This is where plasmid transfer and jumping genes and choromsomal shuffling fit in. Maynard Smith and Szathmary touch on these transitions, but they dont’ fit neatly within the grouping/specializing theme they outlined. They are evolved methods of altering the genome that operate above the base pair level.

    This is what makes the development of language different. If viewed within the framework of Major Transitions in Complexity, the rise of human civilization is a variant on the coming together of organisms and specialization that we find in ants, bees, and termites. I’m not sure how a Major Transitions in Metabolism would view human civilization – was the acquisition of fire a major transition then? What about atomic energy? Viewed as a major transition in the process of evolution, human civilization is a much different phenomenon altogether. It is a transition from one form of replicator to another. As such, it is a meta event. The new replicator is potentially capable of transition of all three sorts described above.

Leave a Reply

Your email address will not be published. Required fields are marked *