Hacker News new | past | comments | ask | show | jobs | submit login

The core problem of taxonomic classification of species is that they're semi-arbitrary, constantly-drifting centroids of similarity, not static, walled-off groups with definitive boundaries.

For those who aren't scientists, we are sometimes referred to as H. sapiens sapiens.

Neanderthals / Neandertals are sometimes grouped as a subspecies of H. sapiens as H. sapiens neanderthalensis.




The problem of classification is deeply philosophical. The core problem is that the taxonomic classification of anything is arbitrary, proven only by consensus.


I've had a series of conversations elsewhere over the past several months on the question. My own view being that classifications are based on usefulness to the classifiers, and we have terms for things which we both 1) experience (there's much of the Universe we've yet failed to classify) and 2) based on our own interactions with those things.

Ideas are interfaces. And interfaces provide utility.

The very notion of "biological species" is based on the underlying notion of philosophical species, that specific classifications. Evolution turned this on its head by noting that species evolve (that is, change), over time. There is no such thing as an absolute species in the earlier philosophical sense.

https://www.etymonline.com/word/species

So the question becomes "where is it useful to define boundaries between species?", with various arbitrary boundaries proposed. The question isn't of their arbitraryness, but of their utility.

And those definitions are ultimately a relationship between us and the universe we're attempting to define.

You can drag in Wittgenstein and a whole bunch else. Knowing some of the philosophical background is useful, less so because a particular view is right, but more because virtually any argument or viewpoint proposed has been offered before, and many of the strengths and weaknesses discussed. Usually to great length.


If I understand you correctly, biological taxonomy, a thing we've had for hundreds of years, and over which huge intellectual battles have been fought, is basically all about making scientists happy, and has little to do with the actual reality of a vast population of living things that is constantly changing over time.

Because, if I did understand you correctly, I completely agree with that.


Not entirely, though there's a fair element of that.

With gene sequencing and techniques such as molecular clocks we can determine evolutionary trees and the approximate time between mutations. This gives us a fairly clear understanding of what current species are placed where on the evolutionary tree --- where you draw the boundaries still has some ambiguity, but it's usually possible to at least say which of two specific samples is closer to or further from a third.

We've also realised that genetic inheritance and transfer is not strictly along descendent lineages --- there's swapping of genetic material, espeically between living organisms and viruses, as well as amongst complex organisms and bacteria. Your mitochondria --- common to nearly all eukaruyotes --- is itself an entirely distinct organism which has become entirely symbiotic with you.

So there's nearer and further, and indirect inheritance. But there remains a great deal of cases --- remember, 99.9% of all life forms on Earth have gone extinct --- where we simply do not know. Or where knowledge isn't sufficient to make a clear or unambiguous distinction or classification.

Jainism has the concept of Anekantavada, "many-sidedness", exemplified in the fable of the blind men and the elephant. It's useful to recognise that our perceptions are arbitrary --- they're dependent on both the particular subset of reality we're observing, as well as the limited channels through which we're making that observation, and the conceptual frames by which we organise it. That said, many sidedness is not any sidedness --- all interpretations are not equivalently valid. That's the gist of a utility model of epistemology: usefulness is context-dependent, but also exists on a spectrum, from high usefulness (greater truth value) to low (likewise).

Arbitrary is not indifferent.

Mind as is usually the case I'm speaking of an area in which I have largely lay knowledge, though I think I'm not going too badly astray.


I also read it as utility an usefulness in a more practical sense. It's useful to differentiate between horses and donkeys, so we call them different species.


Well, we have names for that: "horse" and "donkey".

I get that having formal names is useful, and basing those names on ancestral characteristics seems sensible. But it doesn't seem to take into account that this is a moving target: everything is evolving, and populations are not going to stay the same over time. Trying to classify every living thing into neat categories is like herding a million very slow cats. Especially with palaeontology, where the timespans are large enough for evolution to really matter.


Other cases might be the various organisms we classify as "trees" or "fish" or "crabs", which are valid from a functional or role-based sense, but which aren't robust when considered strictly based on genetic evolution. That is, there have been multiple convertent evolutionary paths which have arrived at similar forms, roles, and niches, despite having hearer relatives or ancestors which don't fall into that category.


I like to half-jokingly claim that people who use the internet are a different species from people that don’t use the internet, since the two groups don’t interbreed.


In the long run, you can become right


neurons have multiple dendritic input sources but can only sum these to a single axon output signal. we are obligate compressors at any scale.


While you are of course right, and much terminology is endlessly challengeable, we can also drift too far in this direction and end up denying that words mean anything. I do find words useful when there is consensus on their meaning, and I think it is still valid to attempt to define classifications even where they must inevitably be imperfect.


Well, they are tripping on their own imprecise definitions.

Using the most prevalent definition of species, the one listed on Wikipedia, we don't have "too many species lumped together into Homo Sapiens", we actually have too few. We know all or almost all "species" since and including Homo Erectus were able capable of interbreeding, meaning they were actually single species.

The trouble is they are trying to use definition that has been developed on observation of the current state that is not suitable for classifying evolving organisms over long periods of time.

That might not have been a huge problem when discussing other animals, but for some reason we pay much closer attention to precision of resolving our own parentage and that is when the definition falls short of being able to serve the need.


> We know all or almost all "species" since and including Homo Erectus were able capable of interbreeding, meaning they were actually single species.

While this is what we were taught in grade school, actual biologists don’t use this children’s definition anymore than actual quantum physicists say atoms are like little solar systems.

There are plenty of separate species that are physically capable of producing fertile hybrid offspring. Dogs and wolves and polar and grizzly bears are a couple obvious examples. Also plenty of species don’t reproduce sexually, so defining species in terms of sexual compatibility isn’t even well defined.


> actual biologists don’t use this children’s definition anymore

Can you link me to the definition biologists use? I googled for a minute and didn't get very far.

Wikipedia (https://en.wikipedia.org/wiki/Species): "A species is often defined as the largest group of organisms in which any two individuals of the appropriate sexes or mating types can produce fertile offspring, typically by sexual reproduction."

What is the correct & concise definition?


The whole problem is that there isn't a single correct definition. For example:

https://evolution.berkeley.edu/biological-species-concept/

https://evolution.berkeley.edu/other-species-concepts/


Yes and I think we're stuck with what's already happening: modifications to existing classifications by genetic discoveries as they occur. There's no good way to please everyone, have a trillion special cases, or throw it all away because it's imperfect.


"We know it when we see it."


> What is the correct & concise definition?

IMO it's that "species" is a term coined by older, less-informed biologists who didn't yet understand all the nuances of genetics, and created a taxonomic system based, erroneously, on superficial differences in appearance. It's like how old scientists believed in "ether" to describe the mysteries of space and "humor" to describe the mysteries of the body. Species is what old scientists used to describe the mysteries of genetics.


Older, less-informed biologists also gave us scientific racism.

  Polar bears and Asiatic black bears live in different places and have different appearances and behaviors. They called those different species.

  Norwegian humans and Chinese humans live in different places and have different appearances and behaviors. They called those different races.


If we're gonna go down that route, probably our genetics gave us racism. Many species are "racist" towards other populations that they might've otherwise bred with. Extant primates also engage in what would be called genocide if performed by humans. All just because the other group lived in some other place and might've looked/smelled/talked a bit different. Nature is cruel and unforgiving and does not give two damns about egalitarianism.

Equality is a human cultural construct, a social contract that attempts to minimize in-group violence. It is itself a cultural evolution, a relatively recent one in the West, longer-lived than that in other cultures, but altogether very limited and mostly only applied to some convenient in-group du jour, not humans around the world and certainly not to hominids at large.

Equality has no real evolutionary basis. Evolution depends on inequality in the face of selection pressures. May fit individuals breed, may adaptable populations persist, and may the sexiest genome win.

Scientific racism failed both because of changing social norms and because taxonomy-by-appearance was doomed to fail upon closer inspection, not because humans are inherently equal... we're anything but, and those "self-evident truths" are believed by no one in particular, not even grade school kids. It's ironic that the nation who says that on one hand has, on the other hand, produced one of the most unequal societies in human history.

If society is ever to actually become egalitarian, it must be because we -- by social contract or force of mythology -- place equal value on each individual, not because our genetics are the same. We're not clones and we're all better and worse at some things, and most of us are completely mediocre at everything except, maybe, being an adaptable hairless ape between ice ages in the temperate zones of a wet rock floating around a star.

In other words, we can't morphologically evolve our way out of racism, but perhaps -- someday, if we survive these next few decades -- we can culturally evolve more egalitarianism. Genomes aren't the only unit of information reproduction anymore. We can only hope culture will succeed where nature failed.


This is the correct answer imo. There's a huge rift in this area of biology on how to organise things, with multiple different ideas.


> What is the correct & concise definition?

There isn’t one. “Species” is an abstraction, vague when you look too closely, and increasingly meaningless when you move from animals to plants to fungi to bacteria and archaea.

There’s just no way to decisively say exactly when two closely related organisms transition from one species to two.

You will find debate and varying methods but in the end for the professional scientist, it’s mostly an unimportant question for most because it’s obviously just vague and never possible to be perfect, and you have to ask “what’s the point?” of the really fine grained labeling which seems to only be done for the purpose of labeling.


There isn’t a correct and concise definition, any more than than there is a correct and concise definition of good art, good code, mathematical proof, or sandwiches. https://plato.stanford.edu/entries/species/

Even the term planet has recently given us trouble. There we’ve adopted a definition, it’s just not fully agreed upon.


Wait, mathematical proof?

> A mathematical proof is an inferential argument for a mathematical statement, showing that the stated assumptions logically guarantee the conclusion.

Unlike all the other sciences, mathematics is abstract enough to allow for things being fundamentally simple and true.


As a mathematical object, a proof is a series of derivations from axioms to a conclusion. That is well defined in various systems, can be studied as a branch of mathematics.

In fact, most mathematics isn't done this way (though there's interesting work in the Lean community). Mathematical proof as it is actually done is far messier. Here's the paper that first came to mind (Probabilistic Proofs and Transferability, by Kenny Easwaran), though I'm sure there is a more canonical source: https://ucfc6eb8f695030deb8332de441a.dl.dropboxusercontent.c...

P.S. You're free to adopt a way of speaking in which you say the majority of graduate mathematics textbooks do not involve any proofs, but I think that's a revision of the term.


I think it's fair to say that the state of the art of proofs is the equational form. Recorde revolutionized mathematical writing with a couple parallel line segments. Generalizing that to proofs using the equivalence, consequence, and implication is an equally (hoho) great leap forward.


As a formalist, I would like to agree with you. As someone who knows most mathematicians are not formalists, I think you may have some problems.


Mathematical proof actually has very precise definition. You can take something and answer precisely whether it is or is not a mathematical proof.


That Wikipedia article does cite other ways of defining species, just after your quote.

The bottom line is that there is no one correct and concise definition.

It's also obvious that the fertile offspring definition is not sufficient of you look at plants (pomelos, mandarins and citrons are rather obviously different species, but they are interfertile and most modern citrus species are their hybrids) or even more obviously, at bacteria or yeasts that don't even reproduce sexually.


iirc dogs and wolves are classed as the same species.


Wolves are canis lupus, while dogs are canis familiaris (and coyotes are canis latrans). All are capable of interbreeding and producing non-sterile offspring.


Philogenetic taxonomies may suffer from this, yes. Ontological classification is a different ball game. Any "rational animal" in the universe qualifies as "human".


I think the term "people" is better. Neanderthals were clearly people, even if they were a different species.


They're (i.e., H. habilis) already classified as "Archaic humans." To call every H. "human" or "people" is ridiculous.


Any entity of genus Homo is (or was) human. Just not necessarily culturally or physiologically modern.


That's not what the word means. Trying to redefine language for political correctness reasons is absurd. Smh.


I'm usually the first person to complain about political correctness. I've just always considered humanity to be at the genus level, not the species level.




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

Search: