5 ms·

Genetics and Eugenics in Frank Herbert's DUNE

120 points by blopeur 8y ago

HelloNurse 8y ago
Very interesting, but fantastic genetic traits like precognition and ancestral memories are (in fiction) a traditional good fit for the Mendelian inheritance pattern of few and high impact mutations.
- gwern 8y ago
  I've idly wondered how you could interpret all the mutations in X-Men, but I don't know enough about the families or inheritance to make any reasonable guesses (I've seen a few cartoons and movies and that's about it). I know mutants generally have mutant kids but don't the powers differ? Which doesn't make much sense and I can't retcon it. I have noticed, however, that in _My Little Pony_ (which I've finally gotten around to watching), it seems like earth/pegasus/unicorn traits are not races or subspecies analogous to human races but instead likely follow a polytomous liability threshold pattern. Why? Because we see the Cakes, two earth ponies, manage to have a pair of pegasus and unicorn fraternal twins, and when queried, mention they have some pegasus/unicorn ancestry. Since the pony type is mutually exclusive (we never see any pony who is both pegasus and unicorn, aside from the alicorns who are special magically-created cases), I don't know of any kind of common Mendelian trait pattern which would give a single type, normally breed true, and allow two earth pony parents with mixed ancestry to have kids who are both unicorns and pegasi (one non-earth-pony offspring would be fine as it would simply indicate a recessive discrete trait, but two different others? I don't see how that works unless you postulate some sort of multi-locus Mendelian trait with dominance of earth trait by both unicorn and pegasus, I guess?). Whereas if it's polytomous liability threshold, then that would be possible albeit unlikely, since they could inherit different fractions of ancestry and get pushed across different type thresholds.
  - komon 8y ago
    As far as the X-Men go, very often children have similar (if not enhanced) powers compared to their parents c.f. Jean Grey and her multiple-universe-spanning offspring with some variation on tele{pathy,kinesis}. However, in the cases where the mutation seems to not be directly inherited, the explanation lies in the "implementation details", as it were. When the X-gene activates (due to puberty, stress, etc.), it releases a protein that goes on to cause mutations in surrounding genes, seemingly at-random but not so random that you have "mutant" teens dropping dead of sudden protein deficiencies.
  - a_t48 8y ago
    Why not two traits? One that determines if (Earth || !Earth) (call it the "Magic" gene) and one that determines (Pegasus || Unicorn) but only if the other gene is active.
    - gwern 8y ago
      I don't really like some of the mechanisms like dominance because it implies something for which we have no canon evidence: biases in discordant offspring towards particular types. And requires some rather specific patterns: in that case, it would require Mr and Mrs Cake to both have exactly 1 Earth/!Earth allele pair (so they become Earth but have a !Earth which can be passed on) and then each has a different Pegasus or Unicorn allele, which then also separately get passed on so they have two kids with !Earth/Pegasus and !Earth/Unicorn pairs. And it seems like if that were how it worked, you would expect way more discordant families to show up. (While polytomous threshold model can accommodate any level of rarity or stochasticness quite easily and naturally.)
  - HelloNurse 8y ago
    Mendelian ponies could have a single gene with 3 alleles, E/U/P, with EE, EU and EP earth, UU and PP respectively unicorn and pegasus and UP either unicorn or pegasus at random (i.e. depending on other genes). If the two parents are EU and EP they are earth but offspring of all three types is possible. An even simpler model: one gene with two alleles, earth and fancy, earth dominant, fancy ponies become unicorns or pegasi depending on rainbow intensity or other non-genetic factors. In the Dune case a very Mendelian concern with preserving a few important mutations in very narrow bloodlines could be combined with a secondary requirement for high IQ, physical fitness etc. that could be (inefficiently) satisfied by focusing on "well bred" aristocracy that benefits from marrying genetically far above average people.
  - drewbuschhorn 8y ago
    I feel stupid saying this, but it was explained to me as being like light through a filter, like for spider man, Parker was always going to get powers (all things being equal) but because the trigger was spider related the powers \ way he manifests them are "spider colored", like light through a keyhole, different keyhole, different image, but same light source. With regard to the X-Men, I think they imply the "present day" is a Cambrian explosion, that then settles down in the far future into the most useful powers becoming common if less godlike levels: immortality, psi, telekinesis, and any other boring wish fulfillment.
badloginagain 8y ago
I always assumed that the BG were directing genetic averages across whole populations, and/or had such a deep understanding of genetics that they could predict the exact result of two peoples genes mixing. If you could determine with 99% accuracy the exact genome of a child by their parents genomes alone, would you be able to optimize humans across given a relatively small set?
- philipov 8y ago
  Except that gene mixing is a highly quantum mechanical process, and so we are precluded from predicting the result of a single event. We can only predict the probability of a particular mix of genes. They could direct averages, but not individual events.
  - dbasedweeb 8y ago
    Of course with the spice and mentats, they could probably use some limited form of prescience to explore possible avenues the future was likely to take. It is a universe in which magic, by another name, works.
    - philipov 8y ago
      Once you accept that premise, throw rationalizations out the window. There's no point theorycrafting, just make up anything you like. Whether or not something is possible is a function of what is needed for the story. The answer to any such question is, "Do you want it to be?"
      - dbasedweeb 8y ago
        I enjoy Dune a lot, but what you describe is absolutely the basis for any science in the story. It’s less science fiction and more high fantasy.
      - krapp 8y ago
        There is "hard" and "soft" magic in fantasy, just as there's "hard" and "soft" science in SF. A story with magic in the premise still needs that magic to have rules and limitations, even if they're not communicated to the reader - even the soft magic of myth and fairy tales needs to have some sense of internal logic.
downer56 8y ago
This is the best write-up of Dune that I’ve ever read. Some of gwern’s stuff is kind of overcooked to a fault, and with intense focus, misses the broader prevailing drift of the topic at hand, but gwern’s got me on this one. It’s a good read.
- gcb0 8y ago
  thought genetics and the BG religion is more a deus exmachine than the main plot. the main plot is an alegory of all the wars USA entered so oil prices remained under control. the oil must flow.
madengr 8y ago
Had to stop reading due to spoilers. Currently on Children of Dune. Alternating between that series and Asimov.
- pmoriarty 8y ago
  Children of Dune and Dune Messiah are, imo, the weakest of Frank Herbert's Dune series. Persevere through those and you'll be rewarded in his later books.
  - ry_ry 8y ago
    It probably doesn't bear repeating at this point, but Brian Herbert's dreary offerings are pretty dismal and should be avoided by all but the most ardent Dune aficionados jonesing for a spice-fix and/or sci-fi masochists hoping to negate any enjoyment they derived from the Herbert novels. Full disclosure: I didn't enjoy the Brian Herbert stuff.
lnanek2 8y ago
The author claims 20 generations would be enough to breed the powers shown and thousands is unrealistic: " even the most pessimistic estimates of how many generations it might take to drastically increase average human intelligence or almost entirely eliminate a nasty recessive by R.A. Fisher might be 20 generations . " But some of the powers, like remembering ancestors memories barely even exist in humans today. The field of epigenetics has proven that a starving parent can alter how the genetics of even grandchildren express themselves. So some basic functions like wrapping DNA tighter or loser and adding a methyl chemical group exist to pass information across generations, but this is like an extra bit or two per gene. To extend the functionality to the level of recording memories would take almost as long as it took for DNA to become as fancy as it did from the original RNA molecules that could produce more of themselves without any enzymes or DNA having developed yet. So almost as long again as life itself has taken, not 20 years.
- gwern 8y ago
  > But some of the powers, like remembering ancestors memories barely even exist in humans today. That would be a valid point... except why are they still breeding for things like intelligence? That was my point. Thousands of generations later, they're still breeding for basic stuff like enhanced intelligence, which makes little sense under pretty much everything we know about intelligence's genetics. > The field of epigenetics has proven that a starving parent can alter how the genetics of even grandchildren express themselves. The extent to which transgenerational heritable epigenetics matters in humans has been wildly exaggerated by the media and the field is rife with overinterpretations, small samples, and non-replication. Not to mention, even that hype didn't exist when Herbert was writing it, so he couldn't've been thinking of it. I am fairly sure he was gesturing towards a more Akashi-Records-style psi mechanism with access granted by ancestry, not actual encoding of memories into spare DNA (which among other things would mean no ancestor would remember anything after having a kid). > To extend the functionality to the level of recording memories would take almost as long as it took for DNA to become as fancy as it did from the original RNA molecules that could produce more of themselves without any enzymes or DNA having developed yet. Neurons already re-edit and splice their chromosomal DNA, so the mechanisms are already there for at least part of it.

120 points by blopeur 8y ago

Very interesting, but fantastic genetic traits like precognition and ancestral memories are (in fiction) a traditional good fit for the Mendelian inheritance pattern of few and high impact mutations.

gwern 8y ago
I've idly wondered how you could interpret all the mutations in X-Men, but I don't know enough about the families or inheritance to make any reasonable guesses (I've seen a few cartoons and movies and that's about it). I know mutants generally have mutant kids but don't the powers differ? Which doesn't make much sense and I can't retcon it. I have noticed, however, that in _My Little Pony_ (which I've finally gotten around to watching), it seems like earth/pegasus/unicorn traits are not races or subspecies analogous to human races but instead likely follow a polytomous liability threshold pattern. Why? Because we see the Cakes, two earth ponies, manage to have a pair of pegasus and unicorn fraternal twins, and when queried, mention they have some pegasus/unicorn ancestry. Since the pony type is mutually exclusive (we never see any pony who is both pegasus and unicorn, aside from the alicorns who are special magically-created cases), I don't know of any kind of common Mendelian trait pattern which would give a single type, normally breed true, and allow two earth pony parents with mixed ancestry to have kids who are both unicorns and pegasi (one non-earth-pony offspring would be fine as it would simply indicate a recessive discrete trait, but two different others? I don't see how that works unless you postulate some sort of multi-locus Mendelian trait with dominance of earth trait by both unicorn and pegasus, I guess?). Whereas if it's polytomous liability threshold, then that would be possible albeit unlikely, since they could inherit different fractions of ancestry and get pushed across different type thresholds.
- komon 8y ago
  As far as the X-Men go, very often children have similar (if not enhanced) powers compared to their parents c.f. Jean Grey and her multiple-universe-spanning offspring with some variation on tele{pathy,kinesis}. However, in the cases where the mutation seems to not be directly inherited, the explanation lies in the "implementation details", as it were. When the X-gene activates (due to puberty, stress, etc.), it releases a protein that goes on to cause mutations in surrounding genes, seemingly at-random but not so random that you have "mutant" teens dropping dead of sudden protein deficiencies.
- a_t48 8y ago
  Why not two traits? One that determines if (Earth || !Earth) (call it the "Magic" gene) and one that determines (Pegasus || Unicorn) but only if the other gene is active.
  - gwern 8y ago
    I don't really like some of the mechanisms like dominance because it implies something for which we have no canon evidence: biases in discordant offspring towards particular types. And requires some rather specific patterns: in that case, it would require Mr and Mrs Cake to both have exactly 1 Earth/!Earth allele pair (so they become Earth but have a !Earth which can be passed on) and then each has a different Pegasus or Unicorn allele, which then also separately get passed on so they have two kids with !Earth/Pegasus and !Earth/Unicorn pairs. And it seems like if that were how it worked, you would expect way more discordant families to show up. (While polytomous threshold model can accommodate any level of rarity or stochasticness quite easily and naturally.)
- HelloNurse 8y ago
  Mendelian ponies could have a single gene with 3 alleles, E/U/P, with EE, EU and EP earth, UU and PP respectively unicorn and pegasus and UP either unicorn or pegasus at random (i.e. depending on other genes). If the two parents are EU and EP they are earth but offspring of all three types is possible. An even simpler model: one gene with two alleles, earth and fancy, earth dominant, fancy ponies become unicorns or pegasi depending on rainbow intensity or other non-genetic factors. In the Dune case a very Mendelian concern with preserving a few important mutations in very narrow bloodlines could be combined with a secondary requirement for high IQ, physical fitness etc. that could be (inefficiently) satisfied by focusing on "well bred" aristocracy that benefits from marrying genetically far above average people.
- drewbuschhorn 8y ago
  I feel stupid saying this, but it was explained to me as being like light through a filter, like for spider man, Parker was always going to get powers (all things being equal) but because the trigger was spider related the powers \ way he manifests them are "spider colored", like light through a keyhole, different keyhole, different image, but same light source. With regard to the X-Men, I think they imply the "present day" is a Cambrian explosion, that then settles down in the far future into the most useful powers becoming common if less godlike levels: immortality, psi, telekinesis, and any other boring wish fulfillment.

badloginagain 8y ago

I always assumed that the BG were directing genetic averages across whole populations, and/or had such a deep understanding of genetics that they could predict the exact result of two peoples genes mixing. If you could determine with 99% accuracy the exact genome of a child by their parents genomes alone, would you be able to optimize humans across given a relatively small set?

philipov 8y ago
Except that gene mixing is a highly quantum mechanical process, and so we are precluded from predicting the result of a single event. We can only predict the probability of a particular mix of genes. They could direct averages, but not individual events.
- dbasedweeb 8y ago
  Of course with the spice and mentats, they could probably use some limited form of prescience to explore possible avenues the future was likely to take. It is a universe in which magic, by another name, works.
  - philipov 8y ago
    Once you accept that premise, throw rationalizations out the window. There's no point theorycrafting, just make up anything you like. Whether or not something is possible is a function of what is needed for the story. The answer to any such question is, "Do you want it to be?"
    - dbasedweeb 8y ago
      I enjoy Dune a lot, but what you describe is absolutely the basis for any science in the story. It’s less science fiction and more high fantasy.
    - krapp 8y ago
      There is "hard" and "soft" magic in fantasy, just as there's "hard" and "soft" science in SF. A story with magic in the premise still needs that magic to have rules and limitations, even if they're not communicated to the reader - even the soft magic of myth and fairy tales needs to have some sense of internal logic.

downer56 8y ago

This is the best write-up of Dune that I’ve ever read. Some of gwern’s stuff is kind of overcooked to a fault, and with intense focus, misses the broader prevailing drift of the topic at hand, but gwern’s got me on this one. It’s a good read.

gcb0 8y ago
thought genetics and the BG religion is more a deus exmachine than the main plot. the main plot is an alegory of all the wars USA entered so oil prices remained under control. the oil must flow.

madengr 8y ago

Had to stop reading due to spoilers. Currently on Children of Dune. Alternating between that series and Asimov.

pmoriarty 8y ago
Children of Dune and Dune Messiah are, imo, the weakest of Frank Herbert's Dune series. Persevere through those and you'll be rewarded in his later books.
- ry_ry 8y ago
  It probably doesn't bear repeating at this point, but Brian Herbert's dreary offerings are pretty dismal and should be avoided by all but the most ardent Dune aficionados jonesing for a spice-fix and/or sci-fi masochists hoping to negate any enjoyment they derived from the Herbert novels. Full disclosure: I didn't enjoy the Brian Herbert stuff.

lnanek2 8y ago

The author claims 20 generations would be enough to breed the powers shown and thousands is unrealistic: " even the most pessimistic estimates of how many generations it might take to drastically increase average human intelligence or almost entirely eliminate a nasty recessive by R.A. Fisher might be 20 generations . " But some of the powers, like remembering ancestors memories barely even exist in humans today. The field of epigenetics has proven that a starving parent can alter how the genetics of even grandchildren express themselves. So some basic functions like wrapping DNA tighter or loser and adding a methyl chemical group exist to pass information across generations, but this is like an extra bit or two per gene. To extend the functionality to the level of recording memories would take almost as long as it took for DNA to become as fancy as it did from the original RNA molecules that could produce more of themselves without any enzymes or DNA having developed yet. So almost as long again as life itself has taken, not 20 years.

gwern 8y ago
> But some of the powers, like remembering ancestors memories barely even exist in humans today. That would be a valid point... except why are they still breeding for things like intelligence? That was my point. Thousands of generations later, they're still breeding for basic stuff like enhanced intelligence, which makes little sense under pretty much everything we know about intelligence's genetics. > The field of epigenetics has proven that a starving parent can alter how the genetics of even grandchildren express themselves. The extent to which transgenerational heritable epigenetics matters in humans has been wildly exaggerated by the media and the field is rife with overinterpretations, small samples, and non-replication. Not to mention, even that hype didn't exist when Herbert was writing it, so he couldn't've been thinking of it. I am fairly sure he was gesturing towards a more Akashi-Records-style psi mechanism with access granted by ancestry, not actual encoding of memories into spare DNA (which among other things would mean no ancestor would remember anything after having a kid). > To extend the functionality to the level of recording memories would take almost as long as it took for DNA to become as fancy as it did from the original RNA molecules that could produce more of themselves without any enzymes or DNA having developed yet. Neurons already re-edit and splice their chromosomal DNA, so the mechanisms are already there for at least part of it.