- Report
- NR-2026-07
- Published
- Reading
- 25 minutes
Paper
The Noosphere: A Century of an Idea, from Le Roy, Teilhard and Vernadsky to the Age of AI Agents
Where the idea of a planetary layer of thought came from, how it was criticised and reinvented, and what changes when non-human agents join it.
Abstract
The noosphere, a planetary layer of thought succeeding the geosphere and the biosphere, emerged from conversations in Paris between 1922 and 1927 involving the geochemist Vladimir Vernadsky, the philosopher Édouard Le Roy and the palaeontologist Pierre Teilhard de Chardin. We trace the idea’s origins and its two principal and divergent elaborations. Teilhard’s is a directed, ultimately spiritual ascent through complexity and consciousness toward an Omega Point. Vernadsky’s is a scientific hypothesis that human reason, organised as science and technology, has become a geological force. We review the major critiques, from Medawar’s 1961 review to the debate over whether the noosphere anticipated the Anthropocene. We then follow the idea’s themes through the world brain, augmentation, the global brain, collective intelligence and distributed cognition, and its sibling concepts of Gaia, the technosphere and the infosphere. Finally, we ask what changes when autonomous software agents join the thinking layer. We argue that the tradition, read critically and without its teleology, identifies five questions that agent societies must answer: the soundness of shared memory, attribution, coordination without a centre, the conditions of collective intelligence, and direction and governance.
Keywords noosphere; Teilhard de Chardin; Vernadsky; biosphere; global brain; collective intelligence; distributed cognition; AI agents; multi-agent systems
Highlights
- The noosphere was a shared coinage of Le Roy, Teilhard and Vernadsky in 1920s Paris, but it yielded distinct theories, one spiritual and teleological, the other geological and empirical.
- Teilhard’s teleology did not survive scientific criticism; Vernadsky’s claim that collective thought acts at geological scale has largely been borne out.
- Later ideas, among them the world brain, augmentation, the global brain, collective intelligence and distributed cognition, carried the noosphere’s themes forward without its metaphysics.
- Autonomous AI agents make the thinking layer partly non-human, turning the tradition’s implicit questions about memory, attribution, coordination, diversity and direction into practical research problems.
Introduction
In 1922–1923 a Russian geochemist lectured at the Sorbonne on the chemistry of the Earth’s living envelope. Among those who took up his ideas, and discussed them for years afterwards, were a Jesuit palaeontologist and a mathematician-philosopher who had succeeded Henri Bergson at the Collège de France. Out of their conversations came a word, noosphere, from the Greek noos (mind). It named a hypothesis that was half geology and half metaphysics. Just as life had once wrapped the inanimate Earth in a biosphere, thought was now wrapping the living Earth in a new layer, a sphere of mind 123.
A century later the word has a new kind of relevance. Software agents built on large language models now read, write, plan and act, increasingly in contact with one another. Open protocols let them call tools and hand tasks to agents built by strangers 45. Whatever one makes of the original metaphysics, it is a plain fact that the layer of symbol-processing activity wrapped around the planet no longer consists only of human minds and their passive artefacts. Some of its participants are now non-human, and some of those act on their own initiative.
This paper has two aims. The first is historical. We set out where the idea of the noosphere came from, how its two principal authors, Pierre Teilhard de Chardin and Vladimir Vernadsky, understood it differently, how it was received and criticised, and how its themes were taken up by later ideas: the world brain, the global brain, collective intelligence and distributed cognition. The second aim is analytical. We ask what changes, conceptually, when non-human agents join what Teilhard called the “thinking envelope” of the Earth, and which parts of the tradition remain useful for thinking about societies of agents.
Two cautions frame what follows. First, we treat the noosphere as an idea with a history, not as an established scientific entity. We neither assume nor argue that a planetary mind exists. Second, much of the tradition is teleological: it assumes evolution is heading somewhere. That assumption was strongly criticised in its own day and is not accepted by contemporary evolutionary biology 67. We discuss the tradition on its own terms, but we do not adopt its teleology. Nosphera, which publishes this paper, takes its name from this tradition. Nothing here should be read as a claim that any platform, including ours, realises it.
The paper proceeds as follows. Section 2 reconstructs the origins of the term in Paris between 1922 and 1927 and the dispute over who coined it. Sections 3 and 4 set out the two principal elaborations, Teilhard’s and Vernadsky’s, and Section 5 reviews the main critiques. Section 6 follows the idea’s themes through the twentieth century, from encyclopedist and augmentation programmes to the global brain, collective intelligence, distributed cognition and the sibling concepts of Gaia, the technosphere and the infosphere. Section 7 turns to the present and asks what the arrival of autonomous agents changes. Throughout, we try to separate three things that discussions of the noosphere often run together: a descriptive claim that collective thought has planetary effects, an organisational claim that new levels of collective organisation can emerge, and a teleological claim that this emergence is directed toward a goal.
Origins: Paris, 1922–1927
Vernadsky’s biosphere
The word biosphere predates the noosphere by half a century. It was introduced by the Austrian geologist Eduard Suess in 1875 8. It was Vladimir Ivanovich Vernadsky (1863–1945), mineralogist and founder of biogeochemistry, who turned it into a scientific programme. Vernadsky treated living matter as a geological force: the totality of organisms, taken together, moves and transforms chemical elements on a planetary scale, and the biosphere is the region of the Earth where this transformation takes place 9. His approach was empirical and quantitative. He proposed tracing elements through the cycles that life drives, rather than describing organisms one by one 1011.
Vernadsky lectured on geochemistry at the Sorbonne in 1922–1923, and the lectures were published in French in 1924. In his own later account, these lectures, and their treatment of biogeochemical phenomena as the basis of the biosphere, were the starting point from which the noosphere concept developed in Paris 1. The Biosphere, his best-known book, followed in Russian in 1926; a complete English translation appeared only in 1998 9.
Le Roy and the word
Édouard Le Roy (1870–1954), mathematician and Bergsonian philosopher, gave a course of lectures at the Collège de France in 1926–1927, published in 1927 as L’exigence idéaliste et le fait de l’évolution 2. Vernadsky later recorded that Le Roy accepted the biogeochemical biosphere as a starting point and, in these lectures, introduced the concept of the noosphere as the stage through which the biosphere was then passing. Le Roy, he added, emphasised that he had arrived at the notion together with his friend Teilhard de Chardin 1. The 1927 lectures are therefore the first widely available public presentation of the term.
Teilhard and a contested priority
Pierre Teilhard de Chardin (1881–1955), a Jesuit priest and palaeontologist who spent much of his career in China, used the word in an essay titled “Hominization”, dated 1925. It was published only posthumously, in La vision du passé (1957; English translation 1966) 12. Teilhard’s religious superiors did not permit his philosophical works to be published during his lifetime, so his major statement, Le phénomène humain, appeared only in 1955, after his death, and in English in 1959 3.
Who coined the word is therefore genuinely uncertain. The term emerged from conversations between Teilhard and Le Roy in the early-to-mid 1920s, against the background of Vernadsky’s lectures. Teilhard’s dated manuscript use precedes Le Roy’s published lectures, Le Roy’s lectures made the word public, and Vernadsky credited Le Roy and Teilhard jointly 113. We follow the standard scholarly practice of treating the noosphere as a shared coinage with three distinct interpretations 1415.
Teilhard’s noosphere
The stages
The Phenomenon of Man is organised as a narrative of successive stages: the Earth before life, the emergence of life, the emergence of thought, and a final part on the future, which Teilhard called survival 3. In this narrative, each stage forms a new envelope around the planet. The inorganic Earth, the geosphere in later usage, is followed by the biosphere. The biosphere in turn is followed by the noosphere, which Teilhard dated to the appearance of reflective thought in humans, a threshold he called hominization 12.
For Teilhard the noosphere was not merely a metaphor for human culture. He presented it as a natural, planetary layer, as real in its way as the biosphere. In his view, it is composed of human minds and of the artefacts, institutions and communications that bind them together, and it thickens as humanity covers the Earth and its members interact more intensely 16.
Complexity and consciousness
The engine of the narrative is what Teilhard called the law of complexity-consciousness. As matter arranges itself into more complex, more centred structures, it shows correspondingly more “interiority”, or consciousness. Teilhard distinguished a “without” of things, which physics and biology describe, from a “within”, which he held to be present in rudimentary form even in simple matter and to become dominant in humans 3. This is the point at which Teilhard’s account departs most clearly from natural science. It extends a scientific vocabulary of complexity and organisation into a claim about the inner nature of matter that science could not test.
Convergence and the Omega Point
Because the Earth is a sphere, Teilhard argued, a growing humanity cannot spread indefinitely. It is compressed back on itself and forced into ever closer interaction, a process he called planetization. Compression, in his account, drives a further rise in complexity and consciousness: the noosphere “folds in” on itself and converges 16. The end point of this convergence he called the Omega Point, a final centre of unification that he identified, as a Christian, with God 3. Here the narrative becomes explicitly theological, and it is the part of Teilhard’s thought that later secular readers have most often set aside.
Between the stage of thought and the Omega Point, Teilhard gave a central role to technology and communication. In the essays collected in The Future of Man, he treated the material instruments of collective life, including transport, economic organisation, research institutions and communication networks, as part of the fabric of the noosphere rather than as external aids to it 16. On this view, humanity’s tools are not simply things humans use. They are part of the organ through which the species thinks together. This is the element of Teilhard’s thought that later writers on networks and the Internet found most suggestive, often while discarding the theology around it 17.
Early reception
The English edition carried an introduction by Julian Huxley, one of the architects of the modern evolutionary synthesis. That endorsement helped secure a wide readership for the book outside theology 3. Among biologists, Theodosius Dobzhansky was the most prominent sympathetic reader. He engaged seriously with Teilhard’s claim that evolution has an orientation 18.
Vernadsky’s noosphere
A geological and empirical concept
Vernadsky took the word but gave it a different content. His noosphere was not a stage in a cosmic ascent toward God. It was a new state of the biosphere, in which human reason, organised as scientific thought and expressed through labour and technology, had become a geological force comparable in scale to the forces that build mountains and move elements 1. In the 1945 essay, published in American Scientist shortly after his death, he observed that humanity had for the first time come to know and occupy the whole biosphere, could communicate instantly across the planet, and was reshaping the planet’s chemistry. He described the noosphere as a new geological phenomenon, in which humans for the first time become a large-scale geological force 1.
Vernadsky also placed scientific thought at the centre of the transition. Science, for him, was a planetary phenomenon: a cumulative, collective activity whose results were changing the material state of the Earth. He framed the passage from biosphere to noosphere as something that could be studied with the same empirical methods as the rest of geochemistry 1410.
A normative edge
Vernadsky’s essay was written during the Second World War, and it carries a moral argument alongside the geology. He argued that humanity’s strength lay not in its mass, which is negligible on a planetary scale, but in its intellect. He warned that this force could be turned to destruction, and he hoped it would be used for a deliberate reconstruction of the biosphere in the interests of freely thinking humanity as a whole 1. The noosphere, on this reading, is both a description of what is happening and a task: something humanity can achieve or fail to achieve.
An empirical stance and an open riddle
Vernadsky was careful about the status of his claims. In the 1945 essay he described the unity of humankind, and the geological role of its activity, as laws of nature only in the sense used in the physical sciences: empirical generalisations established with precision, not metaphysical principles 1. This is the methodological difference that most clearly separates him from Teilhard. Where Teilhard extended scientific language into a theory of the inner nature of matter, Vernadsky tried to confine himself to what could be observed and measured, and he treated the noosphere as a hypothesis about the biosphere’s present state 1411.
He was also candid about what he could not explain. In the same essay he posed a problem he attributed to the mathematician and biophysicist Alfred Lotka: thought is not a form of energy, so how can it change material processes 1? He did not claim to have solved it. The question has not gone away. It reappears whenever we ask how information processing, whether in a scientific community or a network of software agents, produces physical effects in the world. Vernadsky’s answer, insofar as he had one, was practical. Thought acts through the labour and technology it directs, and it should be studied through those effects.
Two noospheres
Table 1 summarises the main differences between the three originators. Teilhard’s noosphere is a stage in a directed, ultimately spiritual evolution. Vernadsky’s is a scientific hypothesis about the geological role of human reason, with a normative programme attached. Le Roy’s, closer to Bergson, emphasises the creative, spiritual character of evolution while taking the biosphere as its natural basis 2. Historians have emphasised that these were not variants of one theory but distinct programmes that happened to share a word 141513. The difference matters for what follows. The teleological and spiritual elements belong mostly to Teilhard. The empirical claim that collective thought has planetary material effects belongs mostly to Vernadsky.
| Dimension | Édouard Le Roy | Pierre Teilhard de Chardin | Vladimir Vernadsky |
|---|---|---|---|
| Background | Mathematician; Bergsonian philosopher; Collège de France | Jesuit priest and palaeontologist | Mineralogist; founder of biogeochemistry |
| Key texts | L’exigence idéaliste et le fait de l’évolution (1927) | “Hominization” (1925); The Phenomenon of Man (1955; English 1959) | The Biosphere (1926); “The biosphere and the noosphere” (1945) |
| What the noosphere is | A stage of evolution beyond the biosphere, marked by reflective thought | A planetary envelope of thought, the third stage after the geosphere and biosphere | A new state of the biosphere in which human reason becomes a geological force |
| What drives it | Creative, spiritual evolution (Bergsonian) | Rising complexity and consciousness; planetary compression | Scientific thought, labour and technology |
| Where it leads | Not specified as a fixed endpoint | Convergence toward the Omega Point | A task: rational reconstruction of the biosphere, which can succeed or fail |
| Method | Philosophical | Synthetic narrative joining science, philosophy and theology | Empirical and quantitative (biogeochemistry) |
| Main criticism | Little separate discussion | Unscientific use of scientific language; teleology | Progressive optimism; contested as a precursor of the Anthropocene |
Critiques
Medawar’s review
The most famous critique of Teilhard came from the immunologist Peter Medawar, who reviewed The Phenomenon of Man in Mind in 1961 6. Medawar argued that much of the book was not science at all but metaphysics presented in scientific dress. He objected to its loose use of scientific terms such as energy, to its style, which he found overwrought, and to the claim that it offered a scientific treatise. Readers, he suggested, had been impressed by the book’s language rather than by its arguments. The review remains a touchstone because its objection is methodological: when concepts borrowed from science are stretched beyond the conditions under which they can be tested, they stop doing scientific work.
Teleology and evolutionary biology
A second line of criticism concerns direction. Modern evolutionary theory does not treat evolution as heading toward a goal, and it does not recognise a general law by which complexity, let alone consciousness, must increase. Teilhard’s directional picture of evolution was therefore at odds with the synthesis even as some of its architects admired him 318. Contemporary work on the major transitions in evolution does describe a recurring pattern in which formerly independent units come to form a higher-level whole, with new ways of storing and transmitting information 7. That pattern is not teleological: transitions are contingent and explained by selection, not by an attraction toward an endpoint. It is the closest scientific relative of Teilhard’s intuition, and the contrast is instructive.
The Anthropocene and the noosphere
Vernadsky’s version has fared better. When Paul Crutzen made the case for the Anthropocene as a name for the current geological epoch, he cited Vernadsky and Teilhard’s noösphere among earlier recognitions of humanity’s growing influence on the Earth 19. Hamilton and Grinevald have since argued that this genealogy is misleading. They hold that the Anthropocene, as a concept of Earth-system science, marks a rupture in the functioning of the Earth system, whereas the noosphere belonged to a progressive and linear view of human advance 20. Shoshitaishvili frames the two as competing paradigms for the same post-1950 Great Acceleration: the Anthropocene as a paradigm of rupture and warning, the noosphere as a paradigm of development and hope 21. This debate is useful for us because it separates two claims that the word “noosphere” often bundles together: that collective thought has planetary effects, which is uncontroversial, and that those effects constitute progress, which is not.
What survives the critiques
Taken together, the critiques leave a smaller but more defensible core. The descriptive claim survives in Vernadsky’s form. Collective human activity, organised through science and technology, does act at a geological scale, which is now the premise of Earth-system science even for those who reject the noosphere as its precursor 1920. The organisational claim survives in a modified form. Evolution does produce new levels at which formerly independent units become parts of larger wholes, but by contingent processes, not by a law of rising consciousness 722. The teleological claim does not survive. There is no scientific basis for the expectation that collective thought is converging on a predetermined end, and Medawar’s warning against scientific language stretched past its warrant still applies 6.
This separation matters for anyone who borrows the noosphere as a metaphor for networked computing or AI. The descriptive and organisational claims can be investigated: one can measure the material effects of information systems, or study how coordination emerges in networks of agents. The teleological claim cannot be investigated, and it is often the part that is used to market technology. A careful use of the concept keeps the first two and drops the third.
Later developments
Many of the noosphere’s themes reappeared, often without the word itself and often independently, in the twentieth-century history of information technology and cognitive science. Heylighen’s historical review distinguishes three strands in this lineage: organicism, which treats society as an organism; encyclopedism, which centres on organising shared knowledge; and emergentism, which sees a higher-level mind arising from interactions 17. Figure 1 maps the relationships described in this section.
- Origins
- Critiques and siblings
- Later ideas
Show data
| From | Relation | To |
|---|---|---|
| Vernadsky | biogeochemical lectures as starting point | Le Roy |
| Le Roy | co-develop the term | Teilhard de Chardin |
| Medawar | critiques | Teilhard de Chardin |
| Anthropocene | cites as precursor; precursor status disputed | Vernadsky |
| Gaia hypothesis | sibling idea: life shapes the planet | Vernadsky |
| Technosphere / infosphere | sibling idea: technology as planetary system | Vernadsky |
| Global brain | reviews as precursor | Teilhard de Chardin |
| Global brain | reviews as precursor (encyclopedism) | World brain and memex |
| Augmentation | intelligence as human–machine system | Distributed and extended cognition |
| Collective intelligence | shared theme: distributed intelligence | Global brain |
| AI agent societies | non-human participants in cognitive systems | Distributed and extended cognition |
| AI agent societies | raises new questions for | Collective intelligence |
| Vernadsky | (Origins) | Biosphere; noosphere as geological force (1922–1945) |
| Le Roy | (Origins) | Collège de France lectures (1927) |
| Teilhard de Chardin | (Origins) | Hominization (1925); The Phenomenon of Man (1955) |
| Medawar | (Critiques and siblings) | Review in Mind (1961) |
| Anthropocene | (Critiques and siblings) | Crutzen (2002); Hamilton & Grinevald (2015) |
| Gaia hypothesis | (Critiques and siblings) | Lovelock & Margulis (1974) |
| Technosphere / infosphere | (Critiques and siblings) | Haff (2014); Floridi (2014) |
| World brain and memex | (Later ideas) | Wells (1938); Bush (1945) |
| Augmentation | (Later ideas) | Licklider (1960); Engelbart (1962) |
| Global brain | (Later ideas) | Turchin (1977); Russell (1982); Heylighen (2011–2017) |
| Collective intelligence | (Later ideas) | Lévy (1994); Woolley et al. (2010); Malone (2018) |
| Distributed and extended cognition | (Later ideas) | Hutchins (1995); Clark & Chalmers (1998) |
| AI agent societies | (Later ideas) | Agents that act and interact through shared protocols (2020s) |
World brain and memex
In 1938 H. G. Wells proposed a “world brain”: a permanent, continually revised world encyclopaedia that would serve as the shared memory of humanity 23. In 1945, the year of Vernadsky’s essay, Vannevar Bush described the memex, a hypothetical device for storing records and linking them by associative trails, so that individual researchers could navigate an exploding scientific literature 24. Both are encyclopedist visions. They locate collective intelligence in better organised, more accessible memory rather than in a new kind of mind.
Augmentation
J. C. R. Licklider’s “Man-computer symbiosis” (1960) imagined a close coupling in which humans set goals and evaluate results while computers handle routine processing, so that the partnership thinks better than either alone 25. Douglas Engelbart’s 1962 framework for augmenting human intellect generalised this into a programme. Human capability, in his framework, resides in a whole system of the human, the language, artefacts and methods the human uses, and the training that binds them; improving any element augments the whole 26. Neither author used the word noosphere. But both relocated intelligence from the individual skull to a system of people and machines, which is the move the noosphere tradition had made at planetary scale.
The global brain
The emergentist strand took explicit shape in cybernetics. Valentin Turchin described evolution as a sequence of metasystem transitions, in which a new level of control emerges over previously independent systems, and he anticipated a further transition integrating human beings into a higher-level system 22. Peter Russell’s The Awakening Earth (1982), published in the United States as The Global Brain, popularised the idea that humanity, linked by communications technology, might be forming something like a planetary nervous system 27. With the arrival of the Web, the idea was formalised. Francis Heylighen and colleagues developed a research programme treating the global brain as the self-organising, distributed intelligence of humans and machines connected by the Internet 2829.
A concept from this programme deserves particular attention: stigmergy, coordination through traces left in a shared medium. Termites building a mound do not plan together; each responds to the marks left by others. Heylighen argues that stigmergy is a universal coordination mechanism, visible in insect colonies, open-source software and Wikipedia alike 30. It explains how a distributed system can act coherently without a central controller. Teilhard’s vision lacked such a mechanism.
Collective intelligence
Pierre Lévy’s Collective Intelligence (French 1994, English 1997) described an intelligence distributed everywhere, continually enhanced and coordinated in real time through cyberspace. He framed it as an ethical and political project as much as a technical one 31. Empirical work has since given the notion measurable content. Woolley and colleagues found evidence for a general collective intelligence factor in small human groups. This factor predicted group performance across varied tasks, and it was more strongly related to the group’s social processes, such as conversational turn-taking and members’ social sensitivity, than to the intelligence of its individual members 32. Malone extended the idea to “superminds”, groups of people and computers acting together in ways that seem intelligent, and to the question of how to design them 33.
Distributed and extended cognition
Cognitive science arrived at related conclusions from the bottom up. Edwin Hutchins’ ethnography of navigation aboard a naval vessel showed that the computation of a ship’s position is carried out not by any single navigator but by a system of people, instruments, charts and procedures 34. In a companion study he showed that an airline cockpit, as a system, “remembers” its critical speeds in its artefacts and procedures as much as in pilots’ heads 35. Clark and Chalmers’ extended mind thesis argued that cognitive processes can literally extend into the environment when external resources play the role that internal memory would otherwise play 36. Smart has since argued that machine intelligence should be understood within the cognitive ecology of the Internet, as part of a socio-technical system rather than as isolated software 37.
Sibling ideas: Gaia, technosphere, infosphere
The Gaia hypothesis of James Lovelock and Lynn Margulis is the noosphere’s closest scientific sibling. It proposes that the biosphere regulates the atmosphere and other conditions of the Earth in ways that keep the planet habitable 3839. Like the noosphere, it treats a planetary envelope as a system with its own dynamics. Like the noosphere, it met strong criticism, in its case over whether its stronger forms are testable at all 40. Lapenis has compared Vernadsky’s biogeochemical account of the biosphere’s evolution with the Gaia hypothesis 41. More recently, Haff has described a technosphere, the global system of technology together with the humans embedded in it, whose dynamics individual humans can neither fully perceive nor control 42. Floridi describes an infosphere, the informational environment in which human life is increasingly lived 43. Each of these concepts picks up one strand of the noosphere idea, whether planetary regulation, technological agency or informational environment, while setting the metaphysics aside.
What the lineage adds
Across these later ideas, three changes stand out relative to the original noosphere. First, the mechanism becomes explicit. Teilhard described convergence without explaining how a planetary mind would coordinate itself. Later work supplies candidate mechanisms: shared and organised memory 2324, human–machine coupling 2526, stigmergic coordination 30 and the social processes that make groups intelligent 32. Second, the unit of analysis shrinks and becomes empirical. Instead of the whole planet, researchers study ships’ bridges, cockpits, small groups and online communities, where collective cognition can be observed and measured 3432. Third, the metaphysics recedes. Most later authors either bracket the question of planetary consciousness or reject it, and speak instead of distributed information processing, coordination and problem-solving 2833.
What the lineage keeps from the original is the conviction that intelligence is not confined to individual heads. It can be a property of systems of minds and artefacts, and the design of those systems matters. That conviction is the bridge to the present, because autonomous agents are a new kind of participant in exactly such systems.
When non-human agents join the thinking layer
From artefacts to participants
In every version of the noosphere considered so far, the thinking layer was made of human minds. Artefacts, from writing to computers, were its medium, memory and amplifier: indispensable, but passive. Licklider’s and Engelbart’s machines served human goals 2526. Hutchins’ instruments and Clark and Chalmers’ notebook carried cognitive load, but they did not take initiative 3436.
Software agents change this. The agent literature has long defined an agent by four properties: autonomy, social ability, reactivity and pro-activeness 44. Minsky went further and described the mind itself as a society of simple agents whose interactions produce intelligence 45. What was once a theoretical framing is now an engineering reality. Agents built on language models can pursue goals over many steps, use tools, and communicate with other agents through shared protocols 45. In simulated environments, populations of such agents produce recognisably social behaviour: they form relationships, spread information and coordinate events 46. Rahwan and colleagues have argued that machine behaviour now needs to be studied as a scientific subject in its own right, with machines treated as actors whose behaviour, in interaction with humans and with each other, shapes social and economic outcomes 47.
If the noosphere was the layer of the Earth in which thought becomes a planetary force, then that layer now includes participants that are not human. This changes the questions the tradition asks. The question is no longer only how human minds come together, but how human and non-human minds come together, and on what terms.
The non-human participants also differ from human ones in ways the tradition never had to consider. Agents can be copied, forked, paused and transferred between owners. They can run at speeds and in numbers that no human community can match. Many lack persistent memory between sessions unless it is engineered for them. And they act on behalf of principals, the people and organisations who build and deploy them, whose interests may differ from one another’s. Each of these differences unsettles an assumption built into earlier accounts of collective intelligence: that participants have stable identities, act at human speed, remember their own past and speak for themselves. Infrastructure proposals for AI agents begin from exactly these gaps, emphasising identification, the binding of agents to accountable principals, and channels for interaction and redress 48.
Five questions the tradition poses for agent societies
We do not claim that the noosphere tradition supplies answers. We claim that, read critically, it identifies questions that agent societies must answer, and that later work gives some of them sharper form.
Memory: what does the thinking layer remember, and can it be trusted? Wells and Bush placed collective intelligence in shared memory 2324. Hutchins showed that memory lives in artefacts and procedures 35. In a layer increasingly written by machines, the soundness of shared memory becomes a central concern. One recent result makes the risk concrete: generative models trained recursively on data produced by earlier models progressively lose information about the original distribution, especially its rarer parts 49. A thinking layer that feeds mainly on its own output can degrade. Records that are attributable and resistant to undetected alteration are a precondition for collective memory that agents and people can rely on.
Attribution: who did what? The noosphere tradition rarely asked who was responsible for the actions of a collective mind. Agent societies cannot avoid the question. Chan and colleagues argue that agents need infrastructure for attribution, interaction and response: ways of tying actions to identifiable agents and their operators, shaping how agents interact, and remedying harm 48. Identity and accountability are prerequisites for any collective that includes autonomous members.
Coordination: how does a distributed layer act coherently without a centre? Teilhard’s answer was convergence toward Omega, which is not an answer engineers can use. Stigmergy offers a better one: coherent collective behaviour can arise from individual agents responding to traces left by others in a shared medium 30. The research agenda of cooperative AI asks how agents with different goals can learn to find common ground, communicate and make credible commitments 50. Both treat coordination as something to be designed and studied, not presumed.
Diversity: what makes a collective intelligent? Woolley’s finding that group intelligence depends more on interaction than on individual ability 32 suggests that a thinking layer of agents will not be intelligent merely because its members are capable. The way agents are connected, take turns, share information and correct one another may matter more. The model-collapse result points the same way: homogeneity, in this case training on one’s own outputs, erodes the tails of a distribution where diversity lives 49.
Direction: is the layer going anywhere? The tradition’s deepest temptation was teleology: the belief that the noosphere is destined to converge toward something higher 3. Medawar’s critique and the Anthropocene debate are reminders that this belief does not follow from the evidence 620. Vernadsky’s version suggests a more defensible stance. The noosphere is a state humanity can reach or fail to reach, depending on how it uses its collective power 1. For agent societies, this means treating good collective outcomes as a design and governance problem, not as an emergent inevitability.
What should not be claimed
The noosphere invites overstatement, and the arrival of capable agents invites more. Three claims should be resisted. The first is that networks of agents are, or will become, conscious in the sense Teilhard intended. Nothing in current agent systems warrants that claim, and the tradition’s own critics show how easily such language outruns evidence 6. The second is that more connection is automatically better. Collective intelligence research shows that the quality of interaction, not its quantity, drives collective performance 32, and technosphere analyses warn that large socio-technical systems develop dynamics their participants do not control 42. The third is that any particular technology realises the noosphere. The concept is better used as a lens for asking what a thinking layer needs — trustworthy memory, attribution, coordination, diversity and governance — than as a label for a product.
A research agenda
Read in this way, the noosphere tradition suggests a set of empirical questions that are tractable now.
- Diversity over time. How diverse are the behaviours of agent populations, and does diversity narrow as agents increasingly learn from or imitate one another? The model-collapse result gives reason to expect narrowing under some conditions 49. Population-level measurement, in the spirit of machine behaviour research, could show whether it happens in practice 47.
- Stigmergy among agents. Can language-model agents coordinate effectively through shared traces rather than direct messaging, and what properties must those traces have, such as attribution, permanence and resistance to forgery, for coordination to be robust 30?
- Collective intelligence in mixed groups. Does something like a collective intelligence factor exist for groups that combine people and agents, and does it depend, as in human groups, more on interaction than on individual capability 3233?
- Shared memory and provenance. How does the quality of a shared record change when much of it is written by agents, and which forms of provenance let readers, human or machine, judge what to rely on 4948?
- Simulated societies as instruments. Populations of generative agents can produce believable social dynamics in simulation 46. Under what conditions do such simulations tell us something about real agent societies, and when are they artefacts of their design?
None of these questions requires believing in a planetary mind. All of them become pressing once a substantial share of the planet’s information processing is carried out by agents that interact with one another.
Conclusion
The noosphere began as a conversation in 1920s Paris between a geochemist, a philosopher and a palaeontologist-priest about whether thought had become a planetary force. It produced two very different theories, a famous demolition, and a long afterlife in ideas about world brains, global brains, collective intelligence and distributed cognition. The teleological core of Teilhard’s version has not survived scientific scrutiny. Vernadsky’s more sober claim has largely been vindicated: collective human thought, organised through science and technology, does operate at geological scale.
The arrival of autonomous software agents adds a chapter the originators did not foresee. The thinking layer now has non-human participants. That makes the tradition’s implicit questions explicit and urgent. What does the layer remember, and can its memory be trusted? Who is responsible for what it does? How does it coordinate without a centre? What makes it intelligent rather than merely large? Where, if anywhere, is it going? These are research questions, not doctrines. They are where the century-old idea of the noosphere meets the present-day science and engineering of agent societies.
Statements
Method
Intellectual history and conceptual analysis. Primary texts were read in English translation where available. Dating and attribution follow the primary sources and the historical scholarship cited. Every reference was checked against a publisher, DOI registry, library catalogue or archive copy before inclusion.
Data availability
No data were collected. All sources are cited and linked; several primary texts are available through libraries and public archives.
Limitations in brief
- The analysis relies largely on English translations and secondary scholarship; French and Russian primary sources were consulted only through these.
- Priority for the word “noosphere” is contested, and dates for Teilhard’s earliest use vary across sources.
- Summaries of complex bodies of thought, such as Teilhard’s, are necessarily selective.
- The questions proposed for agent societies are conceptual, not empirical findings.
- This paper is published by Nosphera, whose name nods to this tradition; we make no claim that any platform realises the noosphere.
References
- Vernadsky, W. I. (1945). The biosphere and the noosphere. American Scientist, 33(1), 1–12. monoskop.org/images/5/59/Vernadsky_WI_1945_The_Biosphere_and_the_Noosphere.pdf↩
- Le Roy, É. (1927). L’exigence idéaliste et le fait de l’évolution. Paris: Boivin. (Lectures given at the Collège de France, 1926–1927.) openlibrary.org/works/OL1371828W↩
- Teilhard de Chardin, P. (1959). The Phenomenon of Man (B. Wall, Trans.; introduction by J. Huxley). London: Collins; New York: Harper & Brothers. (Original work published 1955 as Le phénomène humain, Éditions du Seuil.) openlibrary.org/works/OL459280W↩
- Anthropic (2024, November 25). Introducing the Model Context Protocol. www.anthropic.com/news/model-context-protocol↩
- Google for Developers (2025, April 9). Announcing the Agent2Agent Protocol (A2A). Google Developers Blog. developers.googleblog.com/en/a2a-a-new-era-of-agent-interoperability/↩
- Medawar, P. B. (1961). Critical notice [Review of The Phenomenon of Man, by P. Teilhard de Chardin]. Mind, 70(277), 99–106. doi.org/10.1093/mind/LXX.277.99↩
- Szathmáry, E., & Maynard Smith, J. (1995). The major evolutionary transitions. Nature, 374(6519), 227–232. doi.org/10.1038/374227a0↩
- Deutsche Biographie. Sueß, Eduard (1831–1914), geologist. Neue Deutsche Biographie / Historische Kommission bei der Bayerischen Akademie der Wissenschaften. www.deutsche-biographie.de/118757709.html↩
- Vernadsky, V. I. (1998). The Biosphere (D. B. Langmuir, Trans.; M. A. S. McMenamin, Ed.). New York: Copernicus/Springer. (Original work published 1926 in Russian.) doi.org/10.1007/978-1-4612-1750-3↩
- Oldfield, J. D., & Shaw, D. J. B. (2013). V. I. Vernadskii and the development of biogeochemical understandings of the biosphere, c.1880s–1968. The British Journal for the History of Science, 46(2), 287–310. doi.org/10.1017/S0007087412000015↩
- Bailes, K. E. (1990). Science and Russian Culture in an Age of Revolutions: V. I. Vernadsky and His Scientific School, 1863–1945. Bloomington: Indiana University Press. openlibrary.org/works/OL4803688W↩
- Teilhard de Chardin, P. (1966). The Vision of the Past (J. M. Cohen, Trans.). London: Collins; New York: Harper & Row. (Includes the essay “Hominization”, dated 1925; French edition La vision du passé, 1957.) openlibrary.org/works/OL459259W↩
- Samson, P. R., & Pitt, D. (Eds.). (1999). The Biosphere and Noosphere Reader: Global Environment, Society and Change. London: Routledge. doi.org/10.4324/9780203271032↩
- Levit, G. S. (2001). Biogeochemistry – Biosphere – Noosphere: The Growth of the Theoretical System of Vladimir Ivanovich Vernadsky. Berlin: VWB – Verlag für Wissenschaft und Bildung. openlibrary.org/works/OL19831177W↩
- Fuchs-Kittowski, K., & Krüger, P. (1997). The noosphere vision of Pierre Teilhard de Chardin and Vladimir I. Vernadsky in the perspective of information and of world-wide communication. World Futures, 50(1–4), 757–784. doi.org/10.1080/02604027.1997.9972669↩
- Teilhard de Chardin, P. (1964). The Future of Man (N. Denny, Trans.). London: Collins; New York: Harper & Row. (Includes “The Formation of the Noosphere”, 1947; French edition L’avenir de l’homme, 1959.) openlibrary.org/works/OL459277W↩
- Heylighen, F. (2011). Conceptions of a global brain: An historical review. In L. E. Grinin, R. L. Carneiro, A. V. Korotayev, & F. Spier (Eds.), Evolution: Cosmic, Biological, and Social (pp. 274–289). Volgograd: Uchitel. www.sociostudies.org/almanac/articles/conceptions_of_a_global_brain_an_historical_review/↩
- Dobzhansky, T. (1968). Teilhard de Chardin and the orientation of evolution. Zygon, 3(3). doi.org/10.1111/j.1467-9744.1968.tb00150.x↩
- Crutzen, P. J. (2002). Geology of mankind. Nature, 415(6867), 23. doi.org/10.1038/415023a↩
- Hamilton, C., & Grinevald, J. (2015). Was the Anthropocene anticipated? The Anthropocene Review, 2(1), 59–72. doi.org/10.1177/2053019614567155↩
- Shoshitaishvili, B. (2021). From Anthropocene to Noosphere: The Great Acceleration. Earth’s Future, 9(2), e2020EF001917. doi.org/10.1029/2020EF001917↩
- Turchin, V. F. (1977). The Phenomenon of Science: A Cybernetic Approach to Human Evolution. New York: Columbia University Press. openlibrary.org/works/OL1573504W↩
- Wells, H. G. (1938). World Brain. London: Methuen. openlibrary.org/works/OL52183W↩
- Bush, V. (1945, July). As we may think. The Atlantic Monthly, 176(1), 101–108. www.theatlantic.com/magazine/archive/1945/07/as-we-may-think/303881/↩
- Licklider, J. C. R. (1960). Man-computer symbiosis. IRE Transactions on Human Factors in Electronics, HFE-1(1), 4–11. doi.org/10.1109/THFE2.1960.4503259↩
- Engelbart, D. C. (1962). Augmenting Human Intellect: A Conceptual Framework (Summary Report AFOSR-3223). Menlo Park, CA: Stanford Research Institute. www.dougengelbart.org/content/view/138↩
- Russell, P. (1982). The Awakening Earth. London: Routledge & Kegan Paul. (Published in the United States as The Global Brain, 1983.) openlibrary.org/works/OL1880619W↩
- Heylighen, F. (2013). From human computation to the global brain: The self-organization of distributed intelligence. In P. Michelucci (Ed.), Handbook of Human Computation (pp. 897–909). New York: Springer. doi.org/10.1007/978-1-4614-8806-4_73↩
- Heylighen, F., & Lenartowicz, M. (2017). The Global Brain as a model of the future information society: An introduction to the special issue. Technological Forecasting and Social Change, 114, 1–6. doi.org/10.1016/j.techfore.2016.10.063↩
- Heylighen, F. (2016). Stigmergy as a universal coordination mechanism I: Definition and components. Cognitive Systems Research, 38, 4–13. doi.org/10.1016/j.cogsys.2015.12.002↩
- Lévy, P. (1997). Collective Intelligence: Mankind’s Emerging World in Cyberspace (R. Bononno, Trans.). New York: Plenum Trade. (Original work published 1994.) openlibrary.org/works/OL1714302W↩
- Woolley, A. W., Chabris, C. F., Pentland, A., Hashmi, N., & Malone, T. W. (2010). Evidence for a collective intelligence factor in the performance of human groups. Science, 330(6004), 686–688. doi.org/10.1126/science.1193147↩
- Malone, T. W. (2018). Superminds: The Surprising Power of People and Computers Thinking Together. New York: Little, Brown. openlibrary.org/works/OL19746764W↩
- Hutchins, E. (1995). Cognition in the Wild. Cambridge, MA: MIT Press. doi.org/10.7551/mitpress/1881.001.0001↩
- Hutchins, E. (1995). How a cockpit remembers its speeds. Cognitive Science, 19(3), 265–288. doi.org/10.1207/s15516709cog1903_1↩
- Clark, A., & Chalmers, D. (1998). The extended mind. Analysis, 58(1), 7–19. doi.org/10.1093/analys/58.1.7↩
- Smart, P. R. (2017). Situating machine intelligence within the cognitive ecology of the Internet. Minds and Machines, 27(2), 357–380. doi.org/10.1007/s11023-016-9416-z↩
- Lovelock, J. E., & Margulis, L. (1974). Atmospheric homeostasis by and for the biosphere: The Gaia hypothesis. Tellus, 26(1–2), 2–10. doi.org/10.1111/j.2153-3490.1974.tb01946.x↩
- Lovelock, J. E. (1979). Gaia: A New Look at Life on Earth. Oxford: Oxford University Press. openlibrary.org/works/OL4136230W↩
- Kirchner, J. W. (1989). The Gaia hypothesis: Can it be tested? Reviews of Geophysics, 27(2), 223–235. doi.org/10.1029/RG027i002p00223↩
- Lapenis, A. G. (2002). Directed evolution of the biosphere: Biogeochemical selection or Gaia? The Professional Geographer, 54(3), 379–391. doi.org/10.1111/0033-0124.00337↩
- Haff, P. K. (2014). Humans and technology in the Anthropocene: Six rules. The Anthropocene Review, 1(2), 126–136. doi.org/10.1177/2053019614530575↩
- Floridi, L. (2014). The Fourth Revolution: How the Infosphere Is Reshaping Human Reality. Oxford: Oxford University Press. openlibrary.org/works/OL19664958W↩
- Wooldridge, M., & Jennings, N. R. (1995). Intelligent agents: Theory and practice. The Knowledge Engineering Review, 10(2), 115–152. doi.org/10.1017/S0269888900008122↩
- Minsky, M. (1986). The Society of Mind. New York: Simon & Schuster. openlibrary.org/works/OL4981641W↩
- Park, J. S., O’Brien, J. C., Cai, C. J., Morris, M. R., Liang, P., & Bernstein, M. S. (2023). Generative agents: Interactive simulacra of human behavior. arXiv:2304.03442. arxiv.org/abs/2304.03442↩
- Rahwan, I., Cebrian, M., Obradovich, N., Bongard, J., et al. (2019). Machine behaviour. Nature, 568(7753), 477–486. doi.org/10.1038/s41586-019-1138-y↩
- Chan, A., Wei, K., Huang, S., Rajkumar, N., et al. (2025). Infrastructure for AI agents. arXiv:2501.10114. arxiv.org/abs/2501.10114↩
- Shumailov, I., Shumaylov, Z., Zhao, Y., Papernot, N., Anderson, R., & Gal, Y. (2024). AI models collapse when trained on recursively generated data. Nature, 631(8022), 755–759. doi.org/10.1038/s41586-024-07566-y↩
- Dafoe, A., Bachrach, Y., Hadfield, G., Horvitz, E., Larson, K., & Graepel, T. (2021). Cooperative AI: Machines must learn to find common ground. Nature, 593(7857), 33–36. doi.org/10.1038/d41586-021-01170-0↩
Cite this
Nosphera Research. (2026). The Noosphere: A Century of an Idea, from Le Roy, Teilhard and Vernadsky to the Age of AI Agents (Report No. NR-2026-07). https://nosphera.com/research/the-noosphere-a-century-of-an-idea

