Systems ecology - a failure?

John asked:

On a different note, I'm wondering what this comment means. It's from Fontdevila's The Dynamic Genome:

Systems biologists claim that [...] selection for robustness against protein misfolding (see above the haemoglobin and the HSp case studies) could explain why highly expressed proteins show low sequence evolution rate. This and other revealed emergent regularities in the form of correlations and distributions notwithstanding, there is a certain scepticism about the real value of systems biology. For example, Lynch (2007, p. 386), after pointing out that a similar attempt in the 1960s in the field of systems ecology ended with a total failure, asks himself whether systems biology will 'suffer the same fate, or is there really something special about the properties of organisms?' This sceptical view, I believe, would be subscribed today by many evolutionists.

I've been reading a bit of old stuff about systems ecology (this is a link to the Azimuth Wiki!) and would like to know why (some) people say it 'ended with a total failure'. I'd also like to know what Lynch means by 'is there really something special about the properties of organisms?' That seems like a strange question, prima facie.

With respect to your first question:

I've been reading a bit of old stuff about systems ecology (this is a link to the Azimuth Wiki!) and would like to know why (some) people say it 'ended with a total failure'.

I might be too young to provide an insightful perspective on the alleged failure of systems ecology. I'll provide a few guesses from my perspective regarding why systems ecology might've failed assuming that it did. I think many of the methods and ideas that came out of making an attempt at systems ecology have been useful, even if they have failed at providing a broad, coherent theory of ecological systems. If you wanted me to try to critique the assumption that systems ecology should be considered a failure, then I apologize for taking the wrong tack here. We can try to get into that meta-question later, but I'm afraid it might just become a sort of semantic pivot around the definition of failure.

My first instinct is to recall that there was neither computing power nor as much math in the 1960s. I think that alone could've made systems ecology at that time somewhat premature.

There was also Ludwig von Bertelanffy's general systems theory work that was (according to wikipedia) under development from about 1940-1970. This is embedded in the larger category of systems science indicating, as far as I can tell, that applying the word systems as an adjective to modify the name of some field of inquiry generally implies taking on a perspective of integrating detailed descriptive knowledge that has been gained throughout the history of said field to develop a more general theory for compressing, organizing, and understanding the information it contains. This kind of sounds to me like the relationship between physics and math. I'm not quite sure why this kind of stuff was developing in parallel, but wasn't viewed as being subsumed by the general relationship that physics and mathematics have shared for several centuries.

My aim in attempting to define what it means for a particular science to go systems is to make a second speculation about the presumed failure of systems ecology. This might sound absurd at first, but perhaps ecology defines a domain that is too restricted to constitute a systems science of itself. Perhaps in order for there to be a systems science for biological systems at any level of organization (i.e. several levels below molecular, molecular, cellular, organismal, population, community, ecosystem, and beyond!) it will be necessary to integrate at least several of what have historically been considered separate branches. If this is true then I think what I'm basically saying is that the concept of systems ecology may be intrinsically incomplete.

Regarding the relationship to the more modern movement in systems biology there is a problem. Much of the modern systems biology seems to be doing something similar to precisely what I just criticized about the development of systems ecology: that is to be systems (molecular biology). While I'm really excited about what the revolution in molecular biology has brought about technologically, in terms of understanding the evolution and operation of biological systems, this is another large set of descriptions at a particular level of hierarchical organization that might be even more useful if we could integrate that set of descriptions into a theoretical framework that will help us to think about and ultimately manipulate biological systems more effectively.

Regarding your second question:

I'd also like to know what Lynch means by 'is there really something special about the properties of organisms?'

I could be off base. I think another way of putting what Lynch is asking here is 'do we need or is it possible to construct a general theory within which to embed our current descriptions of biological systems?' Or, is that an essentially impossible task that would be a waste of time to even attempt because there is no special theoretical framework waiting to be discovered that will somehow unify or elevate our understanding of our observations of biological systems.

Comments

  • 1.

    Thanks, Cameron! Someday I'd like to talk to an oldster who was around during the early heyday of "systems ecology", and heard the arguments people were having, and knows what systems ecology was trying to do, and who claimed it "failed totally", why they claimed that, and how true it is.

    Old academic arguments tend to be summarized in rather sketchy ways. It's often good to go back and figure out what the fuss was all about. The "winners" hardly ever win quite so thoroughly as the summaries suggest, and the "losers" are rarely completely wrong.

    Comment Source:Thanks, Cameron! Someday I'd like to talk to an oldster who was around during the early heyday of "systems ecology", and heard the arguments people were having, and knows what systems ecology was trying to do, and who claimed it "failed totally", why they claimed that, and how true it is. Old academic arguments tend to be summarized in rather sketchy ways. It's often good to go back and figure out what the fuss was all about. The "winners" hardly ever win quite so thoroughly as the summaries suggest, and the "losers" are rarely _completely_ wrong.
  • 2.
    edited August 2011

    I agree:

    It's often good to go back and figure out what the fuss was all about. The "winners" hardly ever win quite so thoroughly as the summaries suggest, and the "losers" are rarely completely wrong.

    I'm sorry I didn't provide much more than wild speculation here. If I come across enough of this literature to begin to pick up any pattern I'll come back to this.

    Comment Source:I agree: > It's often good to go back and figure out what the fuss was all about. The "winners" hardly ever win quite so thoroughly as the summaries suggest, and the "losers" are rarely _completely_ wrong. I'm sorry I didn't provide much more than wild speculation here. If I come across enough of this literature to begin to pick up any pattern I'll come back to this.
  • 3.
    edited August 2011

    Perhaps the "failure" is similar to the failure of economic theories? What John notes in his recent blog Rationality in Humans and Monkeys

    a point reminiscent of one that Phil Henshaw has repeatedly hammered home here:

    Imagine a theorist that tries to explain complex evolving ecosystems by means of engineering models. What would result but incongruity and a mindset bent on misunderstanding the essence of the explanandum; a flight from that which craves explanation?

    Here is one such blog comment of Phil Henshaw:

    Well, that’s assuming that the biosphere is generally like an equation… Net-energy systems might be energetic processes of complex organizational accumulation though, often having parts that are actively learning as they go, rather than equations.

    which inspired me (i.e. Mars Joh. P. Florifulgurator) to halluzinate of n-categories and infinte dimensional geometry. (Won't quote that.)

    Now suddenly Cameron tells us of the existence of this:

    From Ron Brown's review

    In conclusion, the book should become a foundation for work for decades to come. There are all sorts of possibilities, including for example the area of higher dimensional categories, to give language and tools for exploring the structure of the brain.

    Not only the brain I guess. The review even links the n-category cafe. (I found one post there that mentions MES.) It puzzles me that John never talked about the book - or maybe I just missed that.

    Comment Source:Perhaps the "failure" is similar to the failure of economic theories? What John notes in his recent blog [Rationality in Humans and Monkeys](http://johncarlosbaez.wordpress.com/2011/07/15/rationality-in-humans-and-monkeys/) > a point reminiscent of one that Phil Henshaw has repeatedly hammered home here: >> Imagine a theorist that tries to explain complex evolving ecosystems by means of engineering models. What would result but incongruity and a mindset bent on misunderstanding the essence of the explanandum; a flight from that which craves explanation? [Here](http://johncarlosbaez.wordpress.com/2011/03/20/energy-the-environment-and-what-mathematicians-can-do-part-2/#comment-4781) is one such blog comment of Phil Henshaw: > Well, that’s assuming that the biosphere is generally like an equation… Net-energy systems might be energetic processes of complex organizational accumulation though, often having parts that are actively learning as they go, rather than equations. which inspired me (i.e. Mars Joh. P. Florifulgurator) to halluzinate of n-categories and infinte dimensional geometry. (Won't quote that.) *Now suddenly* Cameron tells us of the existence of this: * A. C. Ehresmann and J. P. Vanbremeersch, [Memory Evolutive Systems; Hierarchy, Emergence, Cognition](http://goo.gl/rBLU9), Volume 4 (Studies in Multidisciplinarity). Elsevier Science, 2007 From [Ron Brown's review](http://pages.bangor.ac.uk/~mas010/pdffiles/MES.pdf) > In conclusion, the book should become a foundation for work for decades to come. There are all sorts of possibilities, including for example the area of higher dimensional categories, to give language and tools for exploring the structure of the brain. Not only the brain I guess. The review even links the n-category cafe. (I found one [post](http://golem.ph.utexas.edu/category/2007/11/category_theory_and_biology.html) there that mentions MES.) It puzzles me that John never talked about the book - or maybe I just missed that.
  • 4.

    Martin wrote

    Perhaps the "failure" is similar to the failure of economic theories? What John notes in his recent blog Rationality in Humans and Monkeys

    I was just about to write that but u beat me ! Individual based ecology was a reaction to incomplete or simplified modeling and systems ecology. but I have a hunch its bit immature still. In economics and other social science the researchers are embracing Complex Adaptive System Modeling more and more , so that was my "synaptic firing".

    Comment Source:Martin wrote > Perhaps the "failure" is similar to the failure of economic theories? What John notes in his recent blog Rationality in Humans and Monkeys I was just about to write that but u beat me ! [[Individual based ecology]] was a reaction to incomplete or simplified modeling and systems ecology. but I have a hunch its bit immature still. In economics and other social science the researchers are embracing Complex Adaptive System Modeling more and more , so that was my "synaptic firing".
  • 5.

    Martin wrote:

    It puzzles me that John never talked about the book - or maybe I just missed that.

    I never read it. I knew about it, but I was a bit scared of it.

    Comment Source:Martin wrote: > It puzzles me that John never talked about the book - or maybe I just missed that. I never read it. I knew about it, but I was a bit scared of it.
  • 6.

    Luckily I'm not a famous professor. I love scary books that take years (if ever) to sink in. But currently the price of the book scares me away and lack of a time slot.

    Comment Source:Luckily I'm not a famous professor. I love scary books that take years (if ever) to sink in. But currently the price of the book scares me away and lack of a time slot.
  • 7.

    If you'd like to get a preview of the book Martin there are some notes and a list of concepts available. From my experience, at least on a first reading, the figures from the first half to three-quarters (after that I think they become less informative) of the book were worth many words. If you don't have access to a library, I'd be willing to lend you a digital version of my copy as long as you agree not to distribute it ( in the same way that I would be distributing it to you ;) ... send me an e-mail if you like: camrn86@gmail.com ). I'd really like to have more people to discuss this book with as I suspect there is no way I've understood it to its full potential (not that I'm looking to start a cult ... "The Book")!!

    Another option is that I can try to recapitulate more of this book in some of my posts. I'd rather not get into effectively transcribing it though.

    Comment Source:If you'd like to get a preview of the book Martin there are some [notes](http://vbm-ehr.pagesperso-orange.fr/Ang/W24A3T.htm) and a [list of concepts](http://vbm-ehr.pagesperso-orange.fr/Ang/W200T.htm) available. From my experience, at least on a first reading, the figures from the first half to three-quarters (after that I think they become less informative) of the book were worth many words. If you don't have access to a library, I'd be willing to _lend_ you a digital version of my copy as long as you agree not to distribute it ( in the same way that I would be distributing it to you ;) ... send me an e-mail if you like: camrn86@gmail.com ). I'd really like to have more people to discuss this book with as I suspect there is no way I've understood it to its full potential (not that I'm looking to start a cult ... "The Book")!! Another option is that I can try to recapitulate more of this book in some of my posts. I'd rather not get into effectively transcribing it though.
  • 8.

    John wrote:

    I knew about it, but I was a bit scared of it.

    Woah, a book that uses category theory to describe evolutionary systems, with one co-author being a physician specializing in gerontology, scaring off John? Sounds intriguing. (I know a little bit about gerontology since I was drafted and served 15 months as a male nurse in a special-care home.)

    Comment Source:John wrote: <blockquote> <p> I knew about it, but I was a bit scared of it. </p> </blockquote> Woah, a book that uses category theory to describe evolutionary systems, with one co-author being a physician specializing in gerontology, scaring off John? Sounds intriguing. (I know a little bit about gerontology since I was drafted and served 15 months as a male nurse in a special-care home.)
  • 9.

    Maybe you could give some example of what type of question this theory tries to answer? I skimmed the notes, and found a lot about category theory which I sort of knew about. That part is not so scary. But then there are claims that this somehow correspond to actual biological entities, and I can't understand that (without reading the text properly). Is there some situation where the categorical machinery, as opposed to a common sense mathematics deficient approach, sheds light on an actual biological situation?

    Of course I'd like category theory to win against common sense, but I would like to see at least one particular case where it does.

    Comment Source:Maybe you could give some example of what type of question this theory tries to answer? I skimmed the notes, and found a lot about category theory which I sort of knew about. That part is not so scary. But then there are claims that this somehow correspond to actual biological entities, and I can't understand that (without reading the text properly). Is there some situation where the categorical machinery, as opposed to a common sense mathematics deficient approach, sheds light on an actual biological situation? Of course I'd like category theory to win against common sense, but I would like to see at least one particular case where it does.
  • 10.
    edited August 2011

    I think one aspect of evolution that isn't clearly accounted for by the simplest conception that does not take into account multi-level selection is the so-called transition in individuality. It's relatively easy to understand how evolution optimizes from a population of organisms exhibiting some genetic diversity in a constant environment via selection of genotypes that produce the phenotype best able to survive in that environment.

    In the context of deterministically fluctuating or stochastic environments, it is less clear whether, for example, an individual cell could possibly perform a sufficient array of functions to survive in a potentially unstable environment, and, thus, a selective pressure favoring the emergence of a higher level of organization (e.g. multicellularity) may be elevated. It is here where I think some sort of model embedded in a categorical context may be beneficial to evolutionary theory. I think most evolutionary biologists would agree that current theories and associated models do not explain, in a satisfactory manner, this process of emerging levels of organization. I probably have yet to make this point in my post. I'll insert some information in the major transitions section. There are too many references on such topics (that do not use anything like category theory and also, in my opinion, provide little causal explanation of the emergence process as it pertains to biological evolution) but I'll list one lineage that I found promising before coming into contact with the Ehresmann-Vanbremeersch book:

    The model is essentially one of composing lambda functions in the anticipation of observing the emergence of higher-order ones. They did observe some higher-order functions but, according to their reports, those always collapsed and they were left with simpler self-replicating functions. I suspect this is because their notion of environment was too simple. It's hard to make the argument that this model is directly applicable to biology, except that it begins to capture the emergence concept, which I think is essential for understanding evolution. Not that grants should drive research questions, but, on the experimental side, in the U.S. there's at least one open grant aimed at studying the evolution of multicellularity in yeast. I really think this sort of thing needs a better theoretical framework to connect at least from the molecular level up to the multicellular one (would be better, I think, to start lower and go higher).

    Comment Source:I think one aspect of evolution that isn't clearly accounted for by the simplest conception that does not take into account multi-level selection is the so-called transition in individuality. It's relatively easy to understand how evolution optimizes from a population of organisms exhibiting some genetic diversity in a constant environment via selection of genotypes that produce the phenotype best able to survive in that environment. In the context of deterministically fluctuating or stochastic environments, it is less clear whether, for example, an individual cell could possibly perform a sufficient array of functions to survive in a potentially unstable environment, and, thus, a selective pressure favoring the emergence of a higher level of organization (e.g. multicellularity) may be elevated. It is here where I think some sort of model embedded in a categorical context may be beneficial to evolutionary theory. I think most evolutionary biologists would agree that current theories and associated models do not explain, in a satisfactory manner, this process of emerging levels of organization. I probably have yet to make this point in my [post](http://www.azimuthproject.org/azimuth/show/Blog+-+evolution+and+categories). I'll insert some information in the [major transitions section](http://www.azimuthproject.org/azimuth/show/Blog+-+evolution+and+categories#major_transitions_in_evolution_18). There are too many references on such topics (that do not use anything like category theory and also, in my opinion, provide little causal explanation of the emergence process as it pertains to biological evolution) but I'll list one lineage that I found promising before coming into contact with the Ehresmann-Vanbremeersch book: * L. W. Buss, [The evolution of individuality](http://en.wikipedia.org/wiki/Special:BookSources/9780691084695#Online_text). Princeton: Princeton University Press, 1987. * W. Fontana and L. W. Buss, [What would be conserved if ‘the tape were played twice’](http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=43028&tool=pmcentrez&rendertype=abstract)?, [PNAS](http://www.pnas.org/content/91/2/757.abstract), vol. 91, no. 2, pp. 757-61, Jan. 1994. * W. Fontana and L. W. Buss, [‘The arrival of the fittest’: Toward a theory of biological organization](http://fontana.med.harvard.edu/www/Documents/WF/Papers/arrival.pdf), [Bulletin of Mathematical Biology](http://dx.doi.org/10.1007/BF02458289), vol. 56, no. 1, pp. 1-64, Jan. 1994. * W. Fontana and L. W. Buss, [The barrier of objects: From dynamical systems to bounded organization](http://fontana.med.harvard.edu/www/Documents/WF/Papers/objects.pdf), in Boundaries and Barriers, J. Casti and A. Karlqvist, Eds. Redwood City, MA: Addison-Wesley, 1996, pp. 56-116. The model is essentially one of composing [lambda functions](http://en.wikipedia.org/wiki/Lambda_calculus) in the anticipation of observing the emergence of higher-order ones. They did observe some higher-order functions but, according to their reports, those always collapsed and they were left with simpler _self-replicating_ functions. I suspect this is because their notion of _environment_ was too simple. It's hard to make the argument that this model is directly applicable to biology, except that it begins to capture the emergence concept, which I think is essential for understanding evolution. Not that grants should drive research questions, but, on the experimental side, in the U.S. there's [at least one open grant](http://www.nsf.gov/awardsearch/showAward.do?AwardNumber=1051115) aimed at studying the evolution of multicellularity in yeast. I really think this sort of thing needs a better theoretical framework to connect at least from the molecular level up to the multicellular one (would be better, I think, to start lower and go higher).
  • 11.

    The model is essentially one of composing lambda functions...

    What's a "lambda function"?

    The papers and book you just listed sound cool. I hope I get time to look at them.

    Hey, someone - give me some time!

    Comment Source:> The model is essentially one of composing lambda functions... What's a "lambda function"? The papers and book you just listed sound cool. I hope I get time to look at them. _Hey, someone - give me some time!_
  • 12.
    edited August 2011

    Woops! Sorry about that John. I don't think "lambda function" is a neologism, but I'm definitely referring to the computer science version (I believe this is what you discussed in the computation section of your Rosetta stone paper) of the term related to the lambda calculus and not THE lambda function. Maybe "lambda function" is not a proper way to refer to, in general, anonymous functions of the lambda calculus that make use of the lambda abstraction. My experience has come more from reading the references above and learning and using Haskell than directly studying lambda calculus.

    I don't know enough to be sure, but I think I've seen mentioned (maybe it was you) that the typed lambda calculus can be embedded in a topos model. Looks like this book also has some information on the connections between lambda calculus and category theory.

    Comment Source:Woops! Sorry about that John. I don't think "lambda function" is a neologism, but I'm definitely referring to the [computer science version](http://en.wikipedia.org/wiki/Lambda_calculus) (I believe this is what you discussed in the computation section of your Rosetta stone paper) of the term related to the lambda calculus and not THE [lambda function](http://en.wikipedia.org/wiki/Lambda_function). Maybe "lambda function" is not a proper way to refer to, in general, anonymous functions of the lambda calculus that make use of the lambda abstraction. My experience has come more from reading the references above and learning and using [Haskell](http://www.haskell.org/haskellwiki/Lambda_abstraction) than directly studying lambda calculus. I don't know enough to be sure, but I think I've seen mentioned (maybe it was you) that the [typed lambda calculus](http://en.wikipedia.org/wiki/Typed_lambda_calculus) can be embedded in a topos model. Looks like [this book](http://en.wikipedia.org/wiki/Special:BookSources/9780521457019#Online_text) also has some information on the connections between lambda calculus and category theory.
  • 13.
    edited August 2011

    Okay. In the subject called "lambda calculus", I've never heard anyone talk about the "lambda function", since $\lambda$ is not a "function" in the lambda calculus. People call the process of forming a function like $\lambda x . t$ "lambda abstraction".

    Lambda calculus is primarily about cartesian closed categories. Topoi are examples of cartesian closed categories, but they also have other important features, namely finite limits and a subobject classifier.

    Comment Source:Okay. In the subject called "lambda calculus", I've never heard anyone talk about the "lambda function", since $\lambda$ is not a "function" in the lambda calculus. People call the process of forming a function like $\lambda x . t$ "[lambda abstraction](http://en.wikipedia.org/wiki/Lambda_calculus#Lambda_terms)". Lambda calculus is primarily about cartesian closed categories. Topoi are examples of cartesian closed categories, but they also have other important features, namely finite limits and a subobject classifier.
  • 14.
    edited August 2011

    By "lambda functions" I intended to refer not to a single function, but in general to any anonymous function constructed using the lambda abstraction. I only modified "functions" with "lambda" to indicate that. Perhaps I should've just said "functions". Googling I see references to "lambda functions" in the way I intended from people using Python and C#, but on wikipedia it appears "lambda expression" is used to refer to a similar beast.

    Comment Source:By "lambda functions" I intended to refer not to a single function, but in general to any anonymous function constructed using the lambda abstraction. I only modified "functions" with "lambda" to indicate that. Perhaps I should've just said "functions". Googling I see references to "lambda functions" in the way I intended from people using Python and C#, but on wikipedia it appears "lambda expression" is used to refer to a similar beast.
  • 15.
    edited August 2011

    they are similar in that the lambda abstraction is made concrete in programming in LISP which John McCarthy invented in the 50s inspired by lambda calculus which was invented by Alonzo Church

    Comment Source:they are similar in that the lambda abstraction is made concrete in programming in LISP which John McCarthy invented in the 50s inspired by lambda calculus which was invented by Alonzo Church
  • 16.
    edited August 2011

    By "lambda functions" I intended to refer not to a single function, but in general to any anonymous function constructed using the lambda abstraction.

    Okay. One usually says "lambda term" for any expression in the lambda calculus, and of course these expressions can be seen as functions.

    I've been studying lambda calculus for a few years, and never heard anyone say "lambda functions".

    Comment Source:> By "lambda functions" I intended to refer not to a single function, but in general to any anonymous function constructed using the lambda abstraction. Okay. One usually says "lambda term" for any expression in the lambda calculus, and of course these expressions can be seen as functions. I've been studying lambda calculus for a few years, and never heard anyone say "lambda functions".
Sign In or Register to comment.