Hacker News new | past | comments | ask | show | jobs | submit login
Information Theory and the Foundation of Life (quantamagazine.org)
100 points by alexwg on Jan 28, 2017 | hide | past | favorite | 12 comments



> Aging, too, has conventionally been seen as a trait dictated by evolution. Organisms have a lifespan that creates opportunities to reproduce, the story goes, without inhibiting the survival prospects of offspring by the parents sticking around too long and competing for resources.

Can someone with more bio expertise explain this theory to me? I've heard it said before, but it doesn't seem right to me. Some species disperse widely, and the offspring end up far from their parents, not really competing for resources. Wouldn't such species evolve quickly not to age, if possible? Similarly, wouldn't it be far better for non-migrating species not to age but instead evolve to migrate when old?

It seems intuitive to me that keeping an old body from falling apart takes great resources / evolutionary focus, and there are diminishing returns the further out the age. e.g.,., if a species were only 1% likely to make it to old age T in the wild, then it would be hard to select for traits that led to it being in good shape at that age.


The fitness that evolution is driven by and toward is "reproductive fitness".

That is, species are driven by natural selection towards those traits that increase the likelihood that they survive long enough to produce offspring that themselves have a reproductive fitness greater than or equal to their parents.

So natural selection just does not care anywhere close to as much about what happens to an organism after it successfully reproduces, except to the extent that sticking around to raise its offspring improves the reproductive fitness of its offspring and thus its own fitness.

For pretty much every species, there was never any selective pressure to improve physiological maintenance mechanisms beyond what was necessary to achieve the above. So our short-term maintenance mechanisms are fairly robust (e.g., healing injuries, especially while youngish), but longer-term we still have issues with cancer and cognitive decline because the sorts of maintenance needed to prevent them were never selected against except while we were old enough to be parents.

So evolving to "not age" happens, but only to the extent that it improves the reproductive fitness of the population, and usually not much further than that.

As far as the whole migration issue, it's likely just more likely for animals to evolve to reproduce more often than to live longer if there exists a sufficiently large ecological niche for them to occupy. Changing the frequency and degree of reproduction is much easier than fixing aging.

The above was a fairly superficial examination of the issues and there, of course, exists more nuance if you look for it. But I hope this explains the basics.


The idea is that evolution is not about individuals, it is about a population as a whole. Every generation brings new genetic combinations, and the best ones have to survive and spread. From an evolution point of view, you don't want the older individuals to stick around too much once they have passed on their genes.

All other things being equal, there is usually a benefit in being there for a longer time. This is especially true for intelligent and social animals: you know the place, you have a social network. Therefore, older individuals need a handicap for the competition to be level with the new generation. Without aging, a population would simply stall as strong individuals who were there first would systematically squash younger competition.

> It seems intuitive to me that keeping an old body from falling apart takes great resources / evolutionary focus

Actually, it looks increasingly like aging is an additional feature, not a constraint, of life. Some organisms (some shrimps, naked mole rats, galapagos turtles...) seem to have inadvertently switched off aging.


I agree; that line of thinking is only valid if aging is considered to be "beyond the ability to reproduce." If the organisms can still reproduce at the same rate, they are not too old to contribute to natural selection. However, a cumulative distribution over the number of offspring created by age for a particular fitness model will regardless overshadow reproduction in old age with the aggressive reproductive periods during youth, and there are likely diminished evolutionary returns for keeping an organism around longer.

On an ecosystem/gaia scale, there is also the argument that organisms that never die will break the ecosystem by failing to return nutrients to the nutrient cycle. However, this is not necessarily an argument against biological immortality, as there are plenty of other ways to die (starvation, predation, etc.)


To expand on your second paragraph, I would say it's not argument at all because it implies purpose. There is nothing to stop the ecosystem from breaking.

If there was a species that "fixed" aging and continued to reproduce, it would be so successful it would take over and break the ecosystem. This was never observed, not because of "purpose", but most likely because there is entropy limit to extending lifespan (mentioned in the end of article).


If you liked this article, you'd probably also enjoy the book "The Information", by James Gleick.


Thank you for this recommendation. The book is fantastic so far.


I believe I might have been the first person to correctly analyse the thermodynamics of evolution by natural selection: http://thinkinghard.com/evolution/2ndlaw.html (originally http://web.archive.org/web/19961221074218/http://www.xmissio...).


Unfortunately, people often get credit for popularizing ideas rather than being the first to conceive of them.


Wait, what do you think of Schrödinger's use of "negentropy" in "What is Life?" (1944), then?.


I wonder if this can also explain the variable success of different societies and cultures. Perhaps societies in environments of greater disequilibrium will "evolve" cultures that most efficiently model their environments and thus are most efficient at dissipating energy and hence increasing the external global entropy.


I would like to see a new approach to/insights from this idea using the Fluctuation Theorem. Maybe even in the language of David Deutsch's constructor theory..!

(I can already hear the scorn)




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: