Jay on the human origin story 2: Denying structure

This is to address the question ChatGPT proposed: “How much of what we think we know about human ancestors comes from the data, and how much comes from the assumptions we bring to it?”

To recognize some responses to my previous post, one could maybe replace “assumptions” with “Bayesian prior belief”. And one could replace “atheism” with “atheist ideology” (leads to a quagmire), “atheist propaganda” (overstates the explicit commitment) or with “the atheist gaze”.

As I understand it, the established narrative of human evolution is basically about an ape with the smarts gradually getting more intelligent. It’s obviously “adaptive” to be more intelligent, so that progress doesn’t seem to need more explanation. Here is a standard graphic of that progress, showing a gradual increase in brain size becoming steeper, according to a logarithmic trend.

Figure 1. Standard graph of brain increase.

I added on that graph three no-longer-so-recent discoveries, with their dates: Au. sediba, Ho. naledi and Ho. floresiensis, highlighting the “Flores Hobbit”, discovered more than 20 years ago. They look like outliers to that smooth progress, but “outliers” can be just an opinion, so I tried to put some numbers to it. I asked ChatGPT to choose prominent fossil discoveries, order them by discovery date, and give their brain volumes. For some of those data it gave ranges, so I adopted mid-range values. Then I used Excel to fit logarithmic curves for the data available at successive dates and plotted their goodness of fit to the data using the r-squared measure. Here is the graph.

Figure 2

This graph shows that up till the 21st century, available evidence supported a single logarithmic progress very well, even after discovering the small brained Taung child. But the Floris hobbit called that model into question and Homo naledi made it worse.

I need to emphasize that THIS ISN’T A SLAM DUNK. We’re just discussing something, right? But Wikipedia recounts curious attempts to explain away the Flores Hobbit’s small brain in terms of four different congenital disorders. That should raise a red flag. Wasn’t that objection raised by theologians against early fossil evidence of extinct animals? It’s embarrassing. I suspect that those explanations were raised by non-paleontologist scientists who swallowed a common perception amongst scientists that there was something wrong with the hobbits, which their own specialized knowledge could solve. But the accepted explanation of island dwarfism doesn’t seem much better. With a 400cc brain, could the hobbits talk? Wikipedia doesn’t discuss that problem, or discuss the simplest prospect that cousins of Homo naledi just gradually spread from Africa and then boated to the Flores island.

So I’m proposing that a weird denial about human origins is at work, and it’s common amongst scientists. Here is some evidence that what they are denying is structure. The data on hominoid brain sizes can be divided into two simple linear trends: one includes the hobbits and extends back to Proconsul, long before the LCA with apes Their brain size could be called “flatlining” at around 15cc increase per million years. The rest, which includes by far most hominid fossils ever found, fit another linear trend about 25 times more rapid, starting abruptly around 2.5 mya:

Figure 3: Two-trend lines model of encephalization.

This isn’t advanced theory, it’s the territory of “The Emperor has no clothes” with me as the bystanding little boy. A good falsifying evidence could be a serious discussion of human encephalization in terms of Punctuated Equilibrium. Or a graph something like figure 3. Years of experience has taught me that professionals have already thought things out, and anyway other people know a lot more than I do. Being the little boy bystander can expose one to heavy traffic. But just suppose that figure 3 reflects real structure in human evolution. It implies two constraint systems and it implies ruling constraint systems in the first place. The other day I realized that it gives some color to ID, by this argument. In order to create something, you do this and then you do that. For example, a potter friend of mine knows how to create a salt-shaker that is closed at the top, with an inward funneling hole underneath. Apparently, only a few potters have worked out how to make this beautiful thing. I was fascinated but he didn’t want me to video his method in case it got out how he did this and then that.

I’m seeing a lot of prejudice in your posts and not a lot of evidence. Certainly none for a “bad influence” that can be attributed to atheism.

That’s a very simplified view. As in the case of the horse series it seems that the reality is far more “bushy”. The straight line is based on very limited evidence.

It really shouldn’t. The problem is that if Homo Floriensis is descended from Homo Erectus - and by geography Homo Erectus is the most likely ancestor, something must account for the decline in brain size. Early speculations on that point are to be expected.

I don’t see why the first is relevant - indeed the matter must be settled by evidence, and anatomy seems at least as relevant as brain size. As for the second Wikipedia does discuss the possibility of descent from Homo Habilis or Australopithecines. Homo Naledi is probably excluded based on the dates - it’s not clear that Homo Naledi is old enough.

Not really. You seem to be speculating that there is a parallel line of evolution with a slower increase in brain size. If you can come up with better evidence - because your graph really isn’t much - then your idea may be accepted.

3 Likes

@Jay This figure makes no sense. You are plotting date of discovery versus r-squared?
Also, it is not possible to calculate r-square for single data points.

Multiple problems here, and really with the whole analysis. Brain size cannot stay on a logarithmic growth curve forever, because it implies that brains will eventually become exponentially large. This isn’t physically possible.

Fitting lines in Excel is a nice way to describe data, but not a great way to do analysis (I do this for a living! :wink: ).

It’s not clear which data points are used to fit each line. It appears you may have selected points to intentionally create diverging lines.

I think you are fitting linear relationships when that is clearly inappropriate. Clearly there is divergence, but I don’t think you are representing it correctly. To be honest, I don’t understand what you think you are repesenting.

I hope I’m not being overly critical, but as a statistician I can’t help but notice these things. :slight_smile:

4 Likes

what does this have to do with atheism?

5 Likes

Among the public, this may be a common perception but among scientists this is not the “established narrative”. This framing is ‘orthogenic’, as it depicts human evolution as single-line trend towards some apparent goal. Historically, this is due to the fact that early palaeontologists were orthogenesists. The classic example is the orthogenic depiction horse evolution as one sequence of successive forms (left image), but now we know that horse evolution concisted of many divergent lineages leaving along side one another, and some took very different ‘turns’ when it comes to their ecological adaptations (right). For example, at first glance horse evolution appears to “advance” (emphasis on the quotes) from browse-feeding to grazzing in accordance with the expansion of grassland, but this is not always the case. Some horses went back from grazzing to browsing or mixed feeding, e.g. Hippidion (aka Onohippidium) from South America.

The same for human evolution, and anthropologists are well aware of this. It’s not one line of descent. As you later pointed out, this is also true for brain size evolution. It wasn’t linear. Many sizes living alongside one another. In fact, the trend of brain size in the last 100 or 3 thousand years was negative.

Also, to be more intelligent is not always adaptive. There is a huge cost to have big brains. Most life (even those with brains) are able to survive very well without our level of intelligence. So increased brain sizes and intelligence during human evolution can’t be taken for granted, and hypotheses for why and how our brains got so big are highly varied and are continuously discussed. Some examples include these Nature papers from 2018 and 2024.

The claim that Homo floresiensis was microcephalic has been loudly rejected. However, prior to the research questioning whether one specific specimen with very unusual features might have been due to a pathology that is associated with said features is not unreasonable. Pathologies DO happen from time to time of course. What is unreasonable though is claiming that, for example, Neanderthals were just normal humans deformed by old age despite the fact that (1) none of the proposed pathologies matches the features we observe in the specimens, (2) we possess hundreds of specimens including children, and (3) we even have their DNA.

The two claims are not the same. Comparing the two is embarrassing.

This section I find the most garbled. How are brain size or the capacity of language relevant to plausibility of insular dwarfism as an explanation? I find insular dwarfism the best explanation, especially considering the fauna it is associated with, which include also other dwarf animals like a dwarf Stegodon, and correspondingly insular giant rats, storks and varanids.

Lastly, wikipedia DOES discuss the oceanic dispersal of hominins to the island:

Because of the deep neighbouring Lombok Strait, Flores remained an isolated island during episodes of low sea level. Therefore, the ancestors of H. floresiensis could only have reached the island by oceanic dispersal, most likely by rafting.[29]

??? Okay I take back what I said earlier. This is the most garbled segment.

3 Likes

@Jay can correct me, but how I read it is that he is calculating how R^2 would change over time with the addition of each new data point. So the first value reading down the column is the R^2 for a logarithmic curve fit to the first four data points, aka all the data available in 1856. The next value down includes the first five data points. And so on. The chart is then a representation of how the R^2 of the model fit changed over time as more data became available.

1 Like

OK, I see it now.

@Jay I see how you are using R^2 now, to make the case that those points don’t fit an exponential curve. This is true!

Normally when using R^2, we keep all the same data and compare different models or curves. :wink:

With respect, you could be advertising your prejudgement of what I did present.

Isn’t my figure 1 standard graph of brain size against time is a commonly accepted way that human evolution is presented to the public? “Darwin’s Bush” of horse evolution is an another way of viewing the past as genealigy, where the second dimension on the page doesn’t mean anything intrinsic. A graph of hominid brain size versus time is more like a graphic of toe loss in horse evolution, it’s about causation, in that case maybe recounting an arms race between kicking horse ancestors and their predators.

What I’m trying to talk about is the story of human evolution as actually told, sometimes graphically, aren’t I right that figure 1 is such a story in good standing? Like the picture of a succession of less crouched apes, that is so recognizable as to be used in endless jokes?

Figure 4

Yes, speculation can be expected, but was it valid to speculate in terms of congenital disorders? Surely you see the red flag there? Why should geography have made Homo erectus the most likely ancestor? Homo species spread repeatedly and widely over the world, isn’t it the physical signature of a fossil that tells paleontologists what the animal was?

2025-02-24T01:31:01Z

Brain size is highly relevant if it’s accepted as “derived” i.e. a feature that doesn’t reverse itself later (true?) The hobbit had three possible small, small brained, hunched shoulder ancestors. I suppose that in a while, Naledi will be included in a cladistic comparison with Floresiens but in the meantime the Australian Museum article on them several times implies close similarities with Australopithecus.

What I’m basically on about here is to allege a Siren Call of an acausal origin story that the Flores hobbits didn’t fit, so scientists (maybe from other fields) tried shoe-horning them.

I’m actually presenting evidence to support that for the long period of 10 million years, animals with our body plan but ape-sized brains prospered and spread widely over the earth. About 2.5 million years ago there was a dramatic change, when the brain size of some started increasing rapidly, culminating in their heads hosting LLMs. Not all hominins were caught up in that creativity and those that didn’t dwindled and died out.

You judge that my graph (figure 3) “isn’t much” but I believe that the data is robust: if you add more data points, the story stays the same.

Hi Dan, Figure 1 does make sense, for every date of discovery of a new fossil the vertical axis value gives how well all the currently available data fitted a logarithmic function. Also, the goodness of fit only starts when four data points were available.

Honestly, I do think you are finding fault here. Whatever app would be better than Excel at fitting lines to data, Excel is what is available to me as a layman, and what it shows deserves to be looked at.

I’m not the first person to fit a logarithmic curve to human encephalization without worrying about its extrapolation.

The species selected to fit the line of increasing brain size were all indeed increasing their brain size as shown in this graph from the geneticist Oppenheimer’s wonderful book "Out from Africa’s Eden:

Figure 5: Homo Encephalization(Oppenheimer Added 2)

I added three data points for later discoveries, again you can see that two of them are outliers. Ther green lines are also added, I used them to scale off trend lines from the Oppenheimer’s non-linear axes.

I think I’m representing in Figure 3 that over about 10 million years, primates with the human body plan but ape-sized brains survived in a variety of habitats. About 2.5 million years ago there was an abrupt change in what was determining the form of most but not all of them. My interest is in the story about that.

I might have to edit this post if it is accepted but is muddled in its attributions. Is there a way to review a draft post before submitting it?

There is a review panel on the right if you are on desktop.

When I need to edit a post I put a note at the top or bottom “EDITING, PLEASE WAIT TO RESPOND”,
then save it so others can see it, then edit my corrections, removing the note when I’m done. :cowboy_hat_face:

I must say, it is a pleasure to read the posts on this forum.

1 Like

This is a technical complaint as a statistician. R^2 is a measure of model fit (curve fit) of a given model to the same data set. By adding points and calculating R^2 for each, you are changing the data each time. The R^2 values generated this way are not comparable because it’s not all the same data.

There are better ways to show outliers, like just showing a plot, but R^2 is the wrong tool for this job. (A correct use would be to compare R^2 for logarithmic and linear curve fits.)

I agree there are outliers. :smiley:

Excel is good for some tasks, and I see a lot of people using it. I have a friend who is a world class Excel user who does some crazy things. It’s a clumsy tool for statistical analysis, but it does make nice graphs. I have occasionally pushed data back to Excel to generate graphs.

If you are serious about doing analysis I recommend R, and a lot of people are working Python now too.

I have to take part of that back. I sometimes built spreadsheet tools for specific task, especially when I need to see the calculations at each step. 99% of what I do needs a program (usually in SAS) as a record of the statistical analysis. With that program I can easily go back to revise an analysis if needed.

With respect so long as you make unlikely accusations without providing adequate evidence the appearance of prejudice will remain. Your references to “atheist ideology” and “atheist propaganda” are still completely unsupported.

Firstly if you are going to be restricting yourself to popular presentations you should say so. And stick to it, but you don’t. Secondly you miss the point that horse evolution - especially in popular presentations had exactly the same problem. Originally it was presented as a straight line but a “bushy” graph is more correct.

Speculation based on causes known to be possible are hardly a red flag. The bigger red flag is your jumping to conclusions.

If it’s a simple trait like size then of course it can reverse itself. Island dwarfism is an example.
Homo erectus is a geographically more likely ancestor since erectus remains were the only other hominid remains known from the region.

Trying to fit the evidence into the established picture is fundamental to science. And, of course, scientists would be working off the scientific picture not from popular presentations.

Doesn’t the existing scientific picture largely agree with that? Certainly it’s closer to it than the popular presentation. The main difference seems to be your ideas about ancestry but even there you chose to complain that Homo naledi was not named as a possible ancestor for Homo Floriensis - ignoring the fact that other small-brained ancestors were seriously proposed as ancestors.

So long as it only looks at a single trait it is grossly inadequate as evidence of descent. Adding more points can’t change that fact.

1 Like

That’s not what “derived” means, for one thing.

1 Like

Sometimes, but decreasing in frequency. Besides, you are moving the goalposts far from your initial claim:

That’s a very different claim and your understanding is simply wrong. Ironically, you could have corrected it by asking ChatGPT:

Hominin phylogeny is viewed as a tree, not a linear sequence. Modern evolutionary biology recognizes that species evolve through branching patterns, where different hominin species shared common ancestors and diverged over time. This branching tree model reflects the complex web of evolutionary relationships, rather than a single linear progression.

The idea of a “linear” progression often oversimplifies the evolutionary process, implying that one species directly evolves into another. In reality, hominin evolution is much more intricate, with multiple species coexisting at various points in time, some of which may have gone extinct without contributing directly to modern humans.

Wikipedia even shows a tree showing the current view in its Homo entry under “Phylogeny.”

I think it’s safe to assume that not a single one of the people you cite as denying anything in the context of brain size agrees with your foundational misconception, so this whole exercise has no real point that I can see.

Was there any push back from these early othogenesists? Say, a grumpy old professor who stuck to the way of thinking that he had been trained in? Or was there a consensus turnover? You show a big red cross over the older supposedly wrong way of thinking, but it seems to me as layman that your Darwin’s bush on the right doesn’t show what was wrong with the pic on the left which specifically correlates horse size with the number of toes touching the ground. The pic on the right shows small horse ancestors existing over 45 million years but it =doesn’t establish that later small species had fewer toes. For alI I know that might have been the case but that isn’t illustrated in your comparison. Nor does the right way of thinking pic show that some species adapted from fewer to more toes.

The “wrong” pic begs an explanation, in other words a deterministic origin story. It looks to me that you are depicting quite an authoritarian (“cross out the wrong”) consensus building, that denies deterministic structure/

Nesslig20 Figure

Figure 6

That does look a lot like my figure 3, except for the extrapolation towards Homo naledi and Floresiensis, that I have added as a dotted line. I would really appreciate a link to the article it comes from, especially to see why those data points are shown in a different color. The vertical axis uses a log scale, so I added green grid line equivalents for a linear brain size scale to help compare with figure 3. (it seems sensible seeing that number of neurons and energy consumption would be proportional to volume, not log (volume}?) . If you imagine squashing the bottom of the pic upwards to make the raw scale linear, then the goodness of fit to linear “flatlining” trend would look better, while the dispersal of modern data would look bigger. If Dan reads this maybe he can explain as a statistician whether that is just an artifact of a larger data sample for modern man?

Human encephalization is a bit like the horse toe issue: it’s about one parameter changing through time. And my figure 3, the Oppenheimer figure 5 and your punk-eek figure that I annotated above figure 6 all show the same phenomena. Like I said, the evidence is robust.

Oppenheimer’s graph shows the recent negative trend in brain size clearly, but it’s not comparable with the supposed decline in the ancestors of Floresiensis. The human decline looks to me like miniaturisation without loss of function. Once “Nature” had created a talking animal, “she” adapted it towards better energy efficiency. Is that outlandish? What is the established biologist’s explanation?

Seems to me that the distinctive biologically endowed human trait is the ability to speak. Children wo never learn to speak or sign are incapable of functioning in society and are excluded in one way or another. That happens through defects in brain function e.g. foetal alcohol syndrome, microcephaly. So human intelligence can’t be separated from the astonishing ability of an infant to learn the language it hears around it.

Neither articles you linked to demonstrate continual discussion in the stratosphere of evolutionary biology (Nature) , seem to foreground language. That’s a bit like explaining bird wings without mentioning flying.

I’m happy to accept that island dwarfism could explain why the hobbits were a bit smaller than even the small, gracile, australopiths. Modern humans in forests are sometimes much smaller than average, e.g. pygmies, But their brains are not proportionally shrunk. It would be a serious disadvantage to lose whatever proto-speech functionality an ancestor had.

Those were the only possible ways to reach the island, short of flying. The point I want to make is that the hobbits were the last survivors of a 10 million year “flatlining” trend of small brained hominids with the basic human body plan surviving till almost modern times. And the opposite interpretation is that No, they were descendants of large brained hominins whose brains shrank on the island. The vital difference is that the former interpretation begs an explanation for why, 2.5 mya, some hominins did not begin encephalization, while others did.

The most significant thing about the Flores hobbits was surely that they on the Australasian side of the Wallace line that very few placental mammals managed to cross. Homo naledi remains were found in a similarly hard-to-reach place. Maybe both stayed clear of the language-developing hominins.

Fair enough, I was maybe treading beyond the limit of reasonable discourse by talking about “doing this and then that”. I hadn’t thought that much about a connection between my thinking and Intelligent Design. I have long been leery of ID because it seems to me to be almost politically organized and to be blatantly pushing a line. My argument is that Evolutionary Biology has the same weakness and that’s more interesting and less known. But darn it, human evolution unexpectedly started to look to me like a manufacturing process. I don’t know what to do with that, maybe just park it for now.

Would it help if I shifted the goal posts to:
“As I understand it, the established narrative of human evolution is basically about BUNCHES OF APES with the smarts gradually getting more intelligent”?

Brain case size has some correlation, but it is important to realize that intelligence is not weighed by the pound. Neuron pruning, for instance, is vital in childhood development, and neural density, speed of signalling, and brain architecture are also parameters which define intellectual potential. To characterize a move to a smaller brain case as somehow a case of evolutionary regression misses that there is more to intelligence, and the human potential for speech, abstraction, orientation, and manipulation, than brain size. This is fundamental and germane to the problem, and cannot be disregarded for any convenience of physical anthropology.

1 Like

I am not following your line of reasoning here. It seems to me the exact phylogenetic relationship between the various hominid species is an issue that ought to be shared in common between evolutionists and creationists, in that both groups would view them as being descended from a common ancestor. The only difference is that the evolutionists would accept that this common ancestor was, itself, descended from earlier ancestors that eventually connect to all other life. Creationists, OTOH, would believe that this common ancestor was created by God.

Either way, the fact that one subpopulation of these descendants showed a reduction in brain size is no more problematic for evolutionists than it is for creationists. Which is to say, it is not problematic at all.

1 Like

Thanks for your kind response. I raised Homo floresiensis brain size for two reasons. First, to make the case that they were widely seen as not having the brains that they should have had according to a pre-existing origin story of gradually increasing brain size, which I’m suggesting is acausal, and that acausality is part of an atheist origin story. I’m arguing that “evo-devo” biologists are exploring acausal origin stories, for example in this post from above:

Here is the abstract from the 2024 article:

What is a layman to make of that? Looking up “preovulatory ovarian follicles” didn’t help me much. The abstract made me think of the apocryphal Schoolmen arguing about how many angels could dance on the head of a pin. But it seems to me that the human origin story is a lot stronger than that.

My second purpose in bringing up the Flores hobbits was to prepare the ground for an argument that goes well with what you told me. I want to portray them as the heirs of a 10 million-year success story of a chimp-brained bipedal short-fanged short-toed savanna herbivores, coevolving with wooden weapons in an arms race with their predators. In other words, pretty much the Robert Ardrey story, I’m not an original thinker.

Figure 3 repeated

So I’m trying to present Australopithecus and Floresiensis as having attributes we associate with human intelligence: close attention to making things, identification with things, commitment to defending the group. To locate Adam in the geological record, he would be Australopithecus, which would be to declare that they had souls and then, so do other animals.