Well, both of the world’s most prestigious science journals have gone off the rails, a fact confirmed again when I read the Science editorial and letter below on AI and indigenous science, which then led me to a Perspective in Nature on “two eyed seeing” in neuroscience. As you know, the two-eyed metaphor refers to the claimed salubrious effects of blending “indigenous science” with “Western science” (the latter is now known as “science”.)
The point of the Science editorial above is simply this: because indigenous knowledge (and I define “knowledge” as “justified true belief”) is not widely represented in publications or writings that can be accessed by AI, so it is not represented very often in searches.. Thus, as Running Horse Collin says there “cannot be an indigenous AI within the current AI framework.”
There are problems, of translation, too:
Additionally, AI models would have to be trained in translating Indigenous science vocabulary between languages. This is not impossible, but Indigenous languages reflect multidimensional scientific principles and concepts and how they work together. Language and context would have to be translated into those of the linear world of Western science. That is not presently possible. Moreover, AI is trained to reference Western scientific data and concepts as the “control” for the interpretation of concepts. Thus, it would evaluate Indigenous concepts, data, and frameworks against Western methodologies to determine validity, accuracy, or meaning.
Finally, the biggest problem is one that reflects how modern science’s “ways of knowing” aren’t the same as indigenous “ways of knowing.” This is a problem that I’ve discussed extensively with reference to Mātauranga Māori (henceforth “MM”), the Māori “way of knowing” that is the subject of a battle in New Zealand about what counts as science. From the paper above:
The major, overarching problem is that in Western science, the purpose of research is to obtain and convey information, whereas in Indigenous science, information must come with a purpose. Within the Lakota scientific system, for example, the point of conducting innovation, experiments, or research must align with ethics and the foundational scientific purpose, which is to preserve all life and work within it. This is a very different way to understand the acquisition, protection, and use of knowledge. In its current form and design, AI is effectively the outcome of multiple permutations of statistical calculations conveying statistically weighted output and trained to generate practical outcomes. However, within Indigenous scientific systems, there is no “right answer” without an initial, ethically aligned purpose and process. Therefore, “what is the point” is the point.
There’s more to indigenous “ways of knowing” than just obtaining knowledge under an ethical system, though. As in MM, indigenous “science” nearly always incorporates things that modern science—it’s called “Western science” in the article—doesn’t deal with qua science. These are things like ethics, religious codes, spiritual matters (as in spiritual healing), folklore, tradition, and so on.
I’m not denying that indigenous people can acquire knowledge through trial and error, as they have with navigation in Polynesia or growing mussels in New Zealand. Nor do I deny that these observations can lead to substantial medical or scientific advances. Three examples are the use of willow bark to relieve pain (leading to the development of aspirin), the use of cinchona bark to relieve fevers (leading to the development of quinine to mitigate malarial symptoms), and the use of plant-derived curare as arrow poison (leading to advances in anesthesia and muscle relaxants. These all came from knowledge acquired by indigenous (or pre-Caucasion) people.
But it’s not just “colonized” and oppressed people (a theme of all three documents) who have folk wisdom. Remember that the development of the smallpox vaccine by Jenner was prompted by the observation that English milkmaids who had gotten cowpox were immune to smallpox. This led to Edward Jenner’s development of a smallpox vaccine (the first vaccine ever) and the rise of immunology. But Jenner was successful because he tested the folk observations using methods now more refined in modern science: hypothesis testing and repeated observations with the subjects being blinded.
The point I want to make here is that folk observations can be a foundation for modern science, or even be incorporated into modern science. But they only become part of modern science when they are tested. Jenner published his repeated tests on the vaccine in 1798, and the use of salicylic acid and quinine were isolated, tested and reported on in the mid-18th century. That is when the scientific advances occurred, although again, they were derived from folk observations.
But remember that there are many parts of indigenous “ways of knowing” that do not comprise knowledge in the modern scientific sense. There are many herbal cures that don’t work, and don’t forget all the ineffective spiritual cures that are part of MM and other indigenous “systems of knowledge”. There is a well-funded branch of the NIH, the National Center for Complementary and Integrative Health, that has tested folk and indigenous remedies using modern blind controlled trials. Almost without exception, “folk wisdom” was not substantiated scientifically. A few examples are given in the linked 2009 article:
Ten years ago the government set out to test herbal and other alternative health remedies to find the ones that work. After spending $2.5 billion, the disappointing answer seems to be that almost none of them do.
Echinacea for colds. Ginkgo biloba for memory. Glucosamine and chondroitin for arthritis. Black cohosh for menopausal hot flashes. Saw palmetto for prostate problems. Shark cartilage for cancer. All proved no better than dummy pills in big studies funded by the National Center for Complementary and Alternative Medicine. The lone exception: ginger capsules may help chemotherapy nausea.
As for therapies, acupuncture has been shown to help certain conditions, and yoga, massage, meditation and other relaxation methods may relieve symptoms like pain, anxiety and fatigue.
However, the government also is funding studies of purported energy fields, distance healing and other approaches that have little if any biological plausibility or scientific evidence.
The point is not to denigrate indigenous knowledge, but to show that any folk “knowledge” not tested with the methods of modern science cannot be regarded as “knowledge.”
And that brings me to the main point: there is not “indigenous science” and “Western science” that stand as separate entities, much less coequal entities. There is scientific knowledge and there are folk assertions, and that is all there is. It is telling that Running Horse Collin, a member of the Oglala Lakota Nation, does not tell us what indigenous science comprises beyond pointing out its differences from “Western science” (see below)
And that is the second point: indigenous science is said to be different from “Western science” in ways that make the former, in several ways, a complementary and salubrious way of knowing. That feeds into the metaphor of “two-eyed seeing”: the claim that you find out a lot more things if you simultaneously look at nature from the “Western” perspective and the indigenous perspective. Here is a figure from the Nature paper below implying a coequality (or at least a substantial contribution) of both eyes, even though in reality one eye sees a lot wider landscape than the other:
Another point: examples of major contributions to human knowledge made by indigenous groups are rarely given. The paper by Running Horse Collin a case in point. She cites only one example (my bolding)
Yet, these differences do not mean that Indigenous knowledge is irrelevant to the problems that Western science addresses or immaterial to the development of robust AI models. For example, a recent study that explicitly integrated Indigenous science with Western scientific methods showed that the unhindered mobility of wildlife and ecological connectivity are vital for megafauna survival, a framework that could guide modern conservation.
Developing new, nonpolluting forms of AI with an aligned purpose for sustainability of all life within the relational frameworks of Indigenous knowledge could build a technology that serves the planet. Such a shift would reposition a future form of AI as a critical mechanism for the stewardship of all life on Earth.
What is the example? No citation was given, but I’d like to at least read about the contributions of both “ways of knowing”.
More: indigenous science’s contribution to knowledge rests on claims that it is more holistic, nonlinear, nonreductionist, and multidimensional than is modern science. But these differences are never explained. Here again from Running Horse Collin:
Indigenous science embodies knowledge and problem-solving methods developed over hundreds to thousands of years through experiential learning and long-term implementation. Information is tested and refined over generations, and knowledge is passed down and applied through strict protocol-based transfer processes. This system is also multidimensional in nature in that it acknowledges the existence of all parts of life and manages their interdependence, interrelation, and interconnection. In Lakota science, there are four parts—matter, energy, relationality, and a unifying life force—for which the reductionist approach of Western science, which breaks down complex systems into smaller parts for independent analysis, has limited application.Today’s AI models are trained on Western scientific datasets and constructs that are consistent with the linearly recorded nature of Western science, mathematics, and technology. Thus, knowledge is framed chronologically within a single dimension, in contrast to knowledge that is framed relationally and multi-dimensionally.
I really don’t know what these claims means, save that the “four parts of Lakota science” sound awfully close to teleology or spirutuality than to modern science. And in what wasys is modern science “unidimensional” or “linear” or “not interdependent on other relevant areas”, save that knowledge advances with time, experiment by experiment.
Yes, there is a difference between indigenous science and modern science, but I would claim that the difference rests on the toolkit of modern science that truly has led to development, refinement, and advances in knowledge: things like hypothesis testing, pervasive doubt, designed experiments and blind controlled trials, the use of statistical tests, and so on. Without these tools, folk “knowledge” is limited and stalled, and, indeed, I would say that I have seen no examples of a truly indigenous science. Certainly MM is not one, and I doubt whether Lakota science is anything comparable to modern (“Western”) science.
Before I end, I want to critique two other claims often seen in articles like these:
a. Indigenous people own their “knowledge”, i.e. it is their right to determine what is done with it. Here’s one quote from the letter of Velarde et al.:
We propose epistemic auditing as a practical step. Indigenous Peoples, exercising their right to self-determination and to maintain, control and protect their traditional knowledge (3), would co-create and audit AI models and outputs. This would retain and return wisdom to the interpretation of large datasets and counter the narrow perspectives of an AI built in a handful of knowledge-production hubs. As active partners in this context, Indigenous Peoples remain guardians of their lands, their knowledge, and data governance principles (4).
This is confusing. Surely the organic material or unpublished data used by non-indigenous people cannot be used without their permission. You can’t take blood from Native Americans for a scientific purpose without getting permission, using it only in the ways agreed on, and, when indigenous people make a substantive contribution to a research effort, make them coauthors on the resulting publications. But once indigenous people have published their data, it sits in the public domain and does not require permission to be used by anybody else. That is the way all science has worked and continue to work. I am not a guardian of the data on speciation in Drosophila that I’ve produced. Publication is intended to advance science by allowing others to build on your results.
b. Indigenous people are not invariably “wise stewards of their land”. The Māori, for example, not only caused extinction of New Zealand’s dodos (and the predatory Hast’s eagle), but also burned off much more of the native forest than did British “colonizers”. I’ve given examples from the AI bot Grok below the fold, but here you see the claim from the Iles et al perspective in Nature:
Regardless of nomenclature, the impact of environmental despair for Indigenous people is compounded with historical transgenerational trauma. The harm faced by the anxiety around the helplessness for a situation is conversely at odds with Indigenous practices and ways of being. Indigenous cultures around the world are renowned for their respectful cohabitation with the land, not an exploitative one
That last statement is simply false; see the data below the fold. Of course some groups tied to one place try to conserve and preserve it, but that is not true of all “indigenous cultures around the world.”
***********
I conclude that these two articles and the letter below represent attempts to sacralize people seen as oppressed or colonized, accomplished by trying to show that they have a form of science that is in any way comparable to modern science. Yes, the Maori and the Lakota were treated badly by colonialists, and may be even be due forms of “reparation” via educational outreach (but not money). But we shouldn’t give those reparations via performative things like land acknowledgments or, in this case, pretenses about the “science” done by such groups. We don’t need two-eyed seeing, but clear-eyed seeing.
Here’s the letter that directly follows the article above, and I’ve already quoted from it.
Three of the authors on the “eLtter” above are also authors on the Nature Perspective below (click to see an extract). To me it seems a long-winded attempt to show that neuroscience has been greatly enriched by infusions from indigenous cultures. But it seems more like an exercise in scientific anthropology or sociology than of science itself. I showed it to U of C neuroscientist Peggy Mason. She pronounced it, without any input from me, as “intellectually empty” and “a word salad that has only the patina of neuroscience.” (I quote her with permission.). But don’t rely on just our take: have a look for yourself if you can plow through it:
Click “continue reading” to find examples of indigenous people not being “wise stewards” of their habitat taken from Grok:
Here is what Grok gave me (with references) when I asked it “Give examples of indigenous people destroying species or their habitat instead of being ‘responsible stewards'”:
Archaeological, paleontological, and genetic evidence shows multiple cases where indigenous peoples caused or substantially contributed to species extinctions and habitat alteration—often rapidly after arrival, especially on islands with naive fauna. The idea of indigenous groups as inherently perfect “responsible stewards” is a modern romanticization that does not match the record. Humans everywhere modify environments; outcomes depend on population density, technology, ecology (especially islands vs. continents), and time.New Zealand (Māori / Polynesians). Polynesians arrived around the late 13th century. All nine species of moa (large flightless birds, some up to ~250 kg) were hunted to extinction within roughly 100–200 years. A very low human population density (estimated under 2,000 people initially, ~0.01 per km²) was sufficient. They also burned large areas of forest (up to ~40% of native forests in places) for hunting access, agriculture (kūmara, fern-root), and other uses. The Polynesian rat they introduced devastated insects, lizards, and birds that had evolved without mammalian predators. Haast’s eagle (the largest known eagle, dependent on moa) also went extinct. This is among the best-documented and fastest human-driven megafauna extinctions.
Pacific Islands, especially Hawaii (Polynesians). Human colonization of remote Pacific islands caused the extinction of close to 1,000 nonpasserine landbird species alone (estimates of total bird losses, including seabirds and passerines, often range 1,300–2,000). This represented roughly 10–20% of the world’s bird species at the time. Many were flightless rails and other island endemics. In Hawaii specifically, Polynesians (arriving roughly 1,000+ years ago) drove dozens of endemic bird species extinct through hunting, habitat clearing/burning for agriculture, and introductions of rats, pigs, and dogs. Roughly half or more of the original avifauna disappeared before European contact. Similar patterns appear across Eastern Polynesia and other groups.
Easter Island / Rapa Nui (Polynesians). Settlers transformed a palm-dominated forest (estimated 15–19 million palms) into grassland and shrubland between roughly 1200–1650 CE. Human slash-and-burn and resource use played a role; introduced Polynesian rats were a major accelerator by consuming palm seeds and preventing regeneration. The island lost its original forest cover, associated birds, and other species. (Later European impacts finished some remaining vegetation; the classic “total societal collapse from ecocide” narrative has been nuanced by recent work, but the ecological transformation is clear.)
sci.newsMadagascar (early Austronesian/Polynesian-related settlers). Humans arrived roughly 2,000–2,500 years ago. The entire megafauna disappeared: elephant birds (up to ~3 m tall and hundreds of kg), multiple giant lemur species (including gorilla-sized ones), dwarf hippos, and giant tortoises. Causes were primarily hunting plus habitat alteration (burning and conversion for agriculture/herding). Some survived longer in remote areas, but the overall loss of megafauna was complete. These extinctions removed key seed-dispersers and altered ecosystems.
North America (Paleo-Indians / early indigenous Americans). Arrival of humans around 13–15,000 years ago coincides with the extinction of ~38 genera of large mammals (mammoths, mastodons, ground sloths, horses, camels, etc.). Early groups specialized in hunting megafauna (one study estimated megafauna provided 83–88% of meat/fat). The “overkill” hypothesis is debated—climate change was also a major factor, and direct kill-site evidence is limited for many species—but multiple lines of evidence indicate human hunting pressure contributed, particularly for certain taxa and regions. This was part of a broader Late Pleistocene extinction pulse.
sci.newsAustralia (early Aboriginal Australians). Humans arrived ~50–65,000 years ago (debated exact timing). Megafauna extinctions (giant marsupials, birds like Genyornis, reptiles) clustered around 40–46,000 years ago. Many researchers attribute a substantial role to hunting and fire-driven habitat change; some evidence (e.g., cooking marks on eggshells, vegetation shifts after extinctions) supports human involvement. Climate was also changing. Later Aboriginal fire management practices (cultural burning) shaped landscapes and reduced fuel loads in some areas, but the initial megafauna loss is widely linked to human presence. Debate continues on the relative weight of climate versus humans.
Broader patterns. Island ecosystems were especially vulnerable because endemic species often lacked defenses against novel predators or habitat change. Introduced commensals (especially rats) amplified impacts. On continents the picture is more mixed and often synergistic with climate, but human arrival repeatedly correlates with losses of large, slow-reproducing species. Slash-and-burn agriculture, overhunting of preferred game, and resource extraction for tools/fuel/housing appear repeatedly.This does not mean every indigenous group everywhere acted the same way, or that all practices were destructive. Many developed local knowledge and management systems that sustained populations for long periods after initial impacts (or in less fragile environments). Low technology and low densities limited the scale compared with industrial societies. The key point is empirical: indigenous peoples are human, not a separate category of automatic stewards. Environmental outcomes vary by context, not by identity. The archaeological record rejects the pure “noble savage” framing.