The invention of selective breeding was unbelievably late (2019)

by barry-cotter • 3 years ago

51 24 3 years ago

mncharity• 3 years ago

I'm not sure what's bothering me about this article. Maybe... When was common-thing first done with X during the last 15 ky? Given that X was obvious once understood, and not hard to note? When the first existent written record mentions X. That... seems a fraught triple? I guess if "humanity [...] discover" has a scale or continuity requirement, that's clearly kept in mind. Eg, "the first clinical use of antibiotics was in the 1890's", the "first existent record of external cephalic version is from Hellenistic Greece".

anonymouskimmer2• 3 years ago

The focus on animals also ignores the obvious evidence of selective breeding in food plants for 10k+ years.

Specialized breeds of dogs have also existed for a very long time.

This may be an accurate critique of the written science of animal breeding, and gives good examples of the reification of bad practices, but some techniques were obviously working over long periods of times whether they were recorded, or not.

What ultimately bothers me about it is the same things about the "discovery" of America. Ignorance that plant cultivation and ecosystem engineering was prevalent in the Americas (it wasn't a savage, untamed, wild land at all, but had been carefully cultivated).

And, in general, the idea that things didn't exist until written down by those in power.

Inventors who lived in a neighborhood with a high per capita patent rate are themselves more likely to file a patent. Future inventors might correspond with their heroes (shades of the ‘college of letters’) and, if they then meet them in person, they are more likely to go on to innovate, even if it was only a single short meeting and the hero was in an entirely different field and so it is difficult to see what key fact, skill, or wealth/object they could have transmitted which might make any difference.

Conflating the credential with the skill (or here, the patent with the inventiveness).

Robotbeat2• 3 years ago

You’re missing a critical insight: Artificial selection can and does occur unintentionally, ie without methodical intent.

s1artibartfast• 3 years ago

I would expect selective breeding of animals to coincide with domestication with no lag. Ancestral humans were innately familiar with animal breeding because they practiced it with humans first. They would understand that strong parents give rise to strong children and weak parents often give rise to weak children.

On the other hand, the reproductive interactions of plants are less apparent.

082349872349872• 3 years ago

I think the article is not saying "selective breeding" in general was late, but that actually selecting for characteristics which correlate highly with ultimate yield, rather than for merely subjectively* impressive characteristics, occurred very late in the whole "selective" breeding fancy.

* eg, here bulls were evaluated on (among other things) muscling until very recently, whereas in the States bulls of the same race were already being evaluated solely upon how many litres their daughters produced. Moneyball with straws.

Edit: of course, if one is far from a (local) optimum, just about any breeding program will tend to produce improvements. It's only when one is so near to an optimum that most changes make things worse that one needs to get properly (statistically vs. optically?) selective...

taion2• 3 years ago

The premise of the argument seems wrong. Even in just Europe, there’s evidence for an increase in animal size during the Roman period, a decrease during the early middle ages, and then an increase into the high middle ages, e.g. https://link.springer.com/article/10.1007/s12520-021-01426-w

These correspond to economic changes which made different sizes of animal more appropriate. Size and productivity aren’t the only criteria anyway – disease resistance, hardiness, or ease of feeding seem like they’d be just as important to subsistence farmers.

Earlier techniques may have been less good than modern ones, but they quite evidently worked to some extent. Specialized horse breeds had gotten to the point where they needed to be specially fed way before the start of the period in the book reviewed here!

Instead of positing a lack of understanding, the observed phenomenon is likely more a factor of limitations in documentary evidence (how many people engaged in stockbreeding were actually writing books anyway?) and in the actual goals of the people involved around optimizing for robustness and practicality over pure productiveness.

anonymouskimmer2• 3 years ago

(how many people engaged in stockbreeding were actually writing books anyway?)

Yeah, I was thinking that a lot of this knowledge would have been kept secret as proprietary trade information.

taion2• 3 years ago

Or less kept secret than not written down, especially before the early modern period, especially outside of things like war horses, just based on who was engaged in that activity.

In fact a quick glance at the Wikipedia page for medieval horse breeding shows documentation for quite deliberate, centralized, and successful operations quite far back! https://en.m.wikipedia.org/wiki/Horses_in_the_Middle_Ages

Robotbeat2• 3 years ago

Artificial selection can and does happen without being intentional.

gumby• 3 years ago

Yeah, fishing limits on the size of fish caught (and legal for sale) in the Atlantic has led to smaller mature fish.

pixodaros4• 3 years ago

Yes, I think Gwern might want to look into practical animals such as sheep not luxury animals such as racehorses. Just getting those big, strong, docile racehorses required selective breeding!

Edit: And amusingly he does talk briefly about sheep but is not interested in where the types of sheep in 18th century England came from because the people who bred them did not leave records and get praised in a book on early modern history :)

aidenn0• 3 years ago

This is the second post on HN to Gwern's site recently with fairly dubious assertions about the late development of something.

If he reads the comments here, that's good, but I'm going to have to take everything he writes now as "X doesn't exist" to "I don't know about X, but people generally familiar with the field might"

pixodaros4• 3 years ago

To be fair, he does notice that Roman historians are saying things about antiquity which are different than what this specialist in 18th/19th century said in 1986 (footnote 11). There is a famous passage by one of the ancient Greek poets that cobbler competes with cobbler and singer with singer and this leads to excellence.

Many medivalists suspect that monasteries and abbeys drove scientific agriculture in medieval Europe because they had capital, literacy, and education. Some berries seem to have been first domesticated in monastery gardens with nets to keep the birds out. Medieval people may not have had good theories of how to breed animals for traits but they could make it work with enough time and effort.

pixodaros4• 3 years ago

Turning to a Hacker News friendly theme, Mr. Watt and Herr Doktor Haber und Bosch have a lot to do with why livestock today are so big and produce so much milk and wool or so many eggs. Internal combustion engines and artificial nitrates let you feed animals more energy-rich food which let them grow bigger and produce more milk and wool and eggs. Animals also grow bigger if they are vaccinated against diseases, medicated against parasites, protected from pests by cheap steel, and watered with filtered water. Animal breeding was just one of the aspects of British agriculture that started to rapidly improve around the 18th century.

jmount• 3 years ago

American hemisphere corn, tomatoes, potatoes, and peppers were clearly selectively bread very far away from their wild cultivars a very long time ago.

neovialogistics• 3 years ago

The article is about the selective breeding of animals specifically, but there seems to be an issue here where at least four different strategy/theory combinations are all referred to as selective breeding. I don't see why there shouldn't be more precise terminology to help clear things up.

vajrabum• 3 years ago

Can you give those four strategy/theory combinations referred to as selective breeding? I read the linked doc but didn't see it in there.

jvanderbot3• 3 years ago

This article features some of the longest sentences I've read since Meditations.

jvanderbot3• 3 years ago

There's two themes of comments here. First is the observation of earlier evidence of selective pressure on plants and animals that humans used (arguably domesticated) long before TFA claims selective breeding was "invented". Second is the (I guess) counter claim that selection can happen without intent by human farmers.

Both are obviously true, so I don't see the need for back and forth about that.

And whatever side you fall on I think the claim by TFA is almost certainly false because it posits that the first definitive evidence we have of a theory and practice of selective breeding is in fact the first instance. Its crazy to think that's true.

contingencies• 3 years ago

The act of agriculture itself is by definition selective breeding.

For plants, taking them out of their natural environment with its co-occuring species, mycorrhizal networks, mineral composition and distributive species and sewing them by hand in a novel environment constitutes in itself a hugely genetically significant act. Even first-round seeds that germinate at all could show substantially different aggregate characteristics than those of the original wild population.

For animals, those that may be caught for raising are in all probability the children of less aggressive parents. Feeding over time will cause more domestication and the removal of environmental stressors that had maintained levels of fitness for the natural pressures and begin to change them instantly.

Both cases would only accelerate over generations, as it is a rare human who will later go and obtain seed or animals from the wild once a conveniently captive population is already available for breeding.

I've come to the conclusion that humans, in writing their own histories, frequently give themselves far too much credit. We would do well to consider ourselves more humbly.

jackcosgrove• 3 years ago

If this is an example of an "invention" merely being the documentation of long-known folk practices, as some have argued, it makes me wonder how many inventions out there remain to be documented.

How many construction sites, workshops, factories, studios, etc have craftsmen passing along techniques that have never been formally described?

I know this type of science and discovery was much more prevalent in the 19th century when these undocumented practices were more common, but there must still be some and they must still be evolving.

I'm reminded of a talk I attended in school where a guest speaker from industry came in to describe the ideation aspects of industrial automation. The faculty were asking about all these cutting edge technology and management frameworks. The speaker replied that the most effective meta-tool they had for process improvement was a wiki that the assembly line workers could edit.

I am guilty of ruminating on the new hotness in SaaS apps alone at my computer when trying to come up with business ideas. I should just go out and talk to people!

pixodaros4• 3 years ago

Aside from the massive and overwhelming evidence that people have been selectively breeding plants since the Stone Age, Gwern needs to read Carolyn Willekes on ancient horses. One of her observations is that the ancients seemed to have valued horses which were well adapted to specific environments, whereas modern horse and dog breeding often leads to animals which are poor as animals but good on one attribute which the breeders are optimizing for. The Hacker News crowd will say "oh, Goodhart's Law!" and they may have a point.

The invention of selective breeding was unbelievably late (2019)

Related Stories