Resolving the Technosphere

Galbraith, Eric; Faisal, Abdullah-Al; Matitia, Tanya; Fajzel, William; Hatton, Ian; Haberl, Helmut; Krausmann, Fridolin; Wiedenhofer, Dominik

doi:https://doi.org/10.5194/egusphere-2024-1133

Preprints

https://doi.org/10.5194/egusphere-2024-1133

Preprints

08 May 2024

| 08 May 2024

Resolving the Technosphere

Eric Galbraith, Abdullah-Al Faisal, Tanya Matitia, William Fajzel, Ian Hatton, Helmut Haberl, Fridolin Krausmann, and Dominik Wiedenhofer

Abstract. The global assemblage of human-created buildings, infrastructure, machinery and other artifacts has been called the `technosphere', and plays a major role in the present-day dynamics of the Earth system. It enables the rapid extraction and processing of materials from other spheres, combusts fossil fuels causing climate change, and transports materials and people across the planet surface. It provides a vast range of services to humans, such as supporting the production of food, shelter, long-distance communication, and entertainment. However, Earth system science has been slow to explicitly incorporate the technosphere as an integrated part of its conceptual and quantitative frameworks. Here we propose a refined definition of the technosphere, intended to assist in developing functional integration with other Earth system spheres, and an End-Use Technosphere categorization, EUTEC, that is theoretically aligned with human activities and wellbeing. The formal definition and resolved categorization enable basic attributes of the technosphere to be delineated, including its mass distribution among components and in space, as well as its temporal dynamics. In particular, of the roughly 1 Tt of technosphere mass, we estimate that one third is comprised of residential buildings and one third by the transportation system, both of which we map at one-degree resolution. Moreover, we show that reconstructions of technosphere mass since 1900 follow exponential growth with long-run growth rates of >3 % y^-1, consistent with autocatalytic behaviour, allowing it to become an ever-more dominant component of the Earth system. The quantitative understanding of the technosphere remains rudimentary, and is in great need of further work to better integrate it with Earth system science.

Received: 15 Apr 2024 – Discussion started: 08 May 2024

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.

Download & links

Preprint (PDF, 1304 KB)

Supplement (72 KB)

Download & links

Eric Galbraith, Abdullah-Al Faisal, Tanya Matitia, William Fajzel, Ian Hatton, Helmut Haberl, Fridolin Krausmann, and Dominik Wiedenhofer

Status: final response (author comments only)

RC1:
'Comment on egusphere-2024-1133', Anonymous Referee #1, 25 May 2024
General comments
The authors provide a very valuable attempt at refining and constricting the definition and scope of the still somewhat loose concept of the ‚technosphere‘. Through devising a classification scheme and appropriate metrics they aim to prepare it for integration into an Earth system analysis framework. This by itself is already applaudable as such an inclusion is still largely missing. In order to better delineate and understand the system-wide impacts of anthropogenic incursions into the ‚natural‘ Earth system metabolism and the dynamic coupling between the technosphere and the classical Earth spheres, the composition and quantification of technosphere entities, their respective properties and an overall growth function appear to be very meaningful for needed quantitative and qualitative assessments. As is obvious, and made also very clear in the paper itself, the value of the paper therefore lies not not so much in the originality of the approach but in the provision of such a conceptual and classificatory basis for singling out crucial features, attributes and evolutionary dynamics of technosphere components, as it is important for making the concept amenable for modeling the past and future (co-)evolution of the techno-Earth system. By doing so the paper largely follows a tradition in industrial ecology and socio-ecological metabolism research, with which the referee is only faintly acquainted and therefore can’t properly judge the potential of the paper within that school of analysis. Instead the value of the paper is seen in providing, through the lens of and the expertise in material stock & flow analysis, a careful examination and thereby a helpful stimulus to advance the discussion about the mentioned practical ends for a joint assessment of technosphere and general Earth system dynamics.
Specific Comments
1) Classifying the entities of which the technosphere is comprised of is a bit like catching clouds with a butterfly net, or like an attempt to compress water in an air balloon: It will always escape and can never be arrested in a stable state. The approach the authors take is to guide their classification scheme via a number of defined end uses and derived general categories. In line with my general remarks above, this approach should certainly be granted publication, but the authors should be aware of – and possibly express in the paper – the taxonomic limitations of such an approach. A knife can be used to spread butter on bread, to kill a fish, deer or human, to open a letter, to decorate a wall, to repair a gadget, and countless other ways. Similarly, material entities that fall under the rubric (in parlance of the authors) of food preparation could also be registered as sports equipment, furnishing, artifact creation and maintenance, domestic appliance, weapon, deliberate neural restructuring (through religious/sacred objects), etc.
Also, the categories/terms offered in the paper are somewhat disputable. Aren’t heating/cooling devices more a means for survival in bitter cold or inhumane heat (or heat waves, hence a matter of public health) than just providing “somatic comfort”? The authors briefly mention physiological limits (line 192) but the Level 2 designation of the Ambient Context primary category in Table 2 and general definition in Figure A1 do not incorporate such considerations of fundamental life support systems and thereby seem ill-named and/or ill-defined. A further example is the designation of “Non-informational equipment and buildings specialized for teaching, religious activities and research” in Table 2. It seems unclear why a white board or a bible is non-informational and would not fall into the Information category as material media that communicate, store, and process information. Equally ambiguous is the filing of buildings into different categories. Each architectural structure on this planet is a complex assemblage of multiple functions or end uses, houses a myriad of different technosphere entities, is created out of myriads of different lithospheric materials, processes a variety of energy conversions. Thereby a classification into residential buildings (Ambient Context) and buildings for growing food or extracting materials (Activity-specific) seems quite arbitrary.
Moreover, the technosphere mutates over time, and to some extent also end uses mutate. Although the paper acknowledges, and even attempts to estimate, the specific dynamics of technospheric growth over time, its categorization effort cannot otherwise than result in somewhat artificially arresting technospheric entities in terms of contemporary end uses which might be outdated very soon and might have not existed in previous periods of technosphere evolution. As the cited Chris Otter expresses: „Homo sapiens and the technosphere coevolved“, yet such a coevolution can’t be captured with a static and fixed classification entirely focused on the here and now (especially if it something superhuman as the technosphere).
In sum: A classification of the diversity, variety and variability of the technosphere assemblage cannot do other than amounting to an exercise in futility. This is not the authors fault; such ambiguities are fundamental and likely can’t be resolved to general satisfaction. However, attempts such as the one made in the paper can still help to refine our understanding of the technosphere composition and shed light on its main drivers. The referee only thinks that such more modest and limited aims should be addressed more openly in the paper. Just in general, the paper would profit from a bit more reflexive stance on the approach chosen. The made claim that “we are now in a position to characterize the basic features of the basic attributes of the technosphere in a holistic, exhaustive and exclusive manner,” (line 223) is hardly defendable.
2) The focus on mass analysis of technospheric components (and not, for instance, structural function) necessarily leads to rather modest gains in knowledge, however complex and challenging such an assessment already is. This was also the problem of the cited Zalasiewicz et al. 2017 paper, although the authors of the present paper address and resolve some of the issues that came with it, e.g. by excluding human-modified soils.
An example here is the class of the Informational, a category extremely important for the (autocatalytic) growth of the technosphere (e.g. Felix Creutzig et al. 2022). As the authors rightly mention in line 244 “The Informational category, despite its prominence in human affairs and experience, accounts for only a small portion of the total mass”. That by itself should be a warning sign for the authors. Weight clearly doesn’t equal importance. It doesn’t tell you what drives, accelerates or decelerates the technosphere’s overall power to establish itself as a new Earth sphere. It tell’s you how much heavy material is now installed (and in use) on Earth’s surface, or better: how much material is converted from the lithosphere to the technosphere, but it does not assess – or “weight”, for that matter – the structural, functional and dynamical nature of the latter.
Another example is the discussion of the geographical distribution in the mapping subsection 4.1.2 and Fig. 3 which lacks not only insight but sophistication. As Henri Lefebvre has already shown in the 1970s, the planet is already largely ‘urbanized’ as it is a necessary background resource for sustaining what once was called cities or agglomerations. Again, a focus on simple mass accumulation does not help to explain how agglomerations and the planetary “hinterland” crucially interact to form a planetary-scale agricultural/extractive/consumptive anthrome (see e.g. Neil Brenner, 2014).
3) The paragraph discussing the choice of focusing the analysis on components 'in use' (starting at line 98) might merit from briefly addressing the concept of the ‘technofossil’ as a sort of counterpart (e.g. Zalasiewicz et al. 2014, or https://www.anthropocene-curriculum.org/anthropogenic-markers/novel-materials-and-technofossils/contribution/the-technofossil-record-where-archaeology-and-paleontology-meet).
Technical comments
The paper misses consistency in terminology used across tables, text and figures. These are essential, however, for the reader to decode, for instance, Figure 2, and align it with what is listed in Table 2 and Figure A1.

Line 110: correct spelling is foraminifera

The sentence starting in line 219 should be moved up to the opening paragraph of subsection 3.1. Thereby the reader is better prepared to what is to come and is helped to navigate the categories.

Suddenly, in the last sentence (line 334), the “trillion-ton technosphere” pops up, probably a remnant of the earlier analysis by Zalasiewicz et al. That number doesn’t appear earlier in the paper, if I am correct.
Citation: https://doi.org/10.5194/egusphere-2024-1133-RC1
- AC1: 'Reply on RC1', Eric Galbraith, 31 May 2024
  
  We thank Reviewer 1 for the thoughtful and constructive comments on our manuscript. We look forward to revising the manuscript accordingly, at the end of the discussion period.
  In response to the General Comments and Specific Comment 1:
  We very much appreciate the viewpoint regarding the difficulty of classifying the components of the technosphere. Indeed, this is a problem that we have pondered for years. It is notable that the reviewer chose the example of a knife to illustrate the difficulty – this is also an example that we considered as being particularly tricky to classify (a swiss army knife is even worse). There’s no question that classifying such a diverse array of things presents a challenge.
  However, we would suggest that, rather than the exercise being futile, it should simply be seen as having a non-unique solution. We would maintain that the exercise is potentially very useful, because it can provide frameworks within which functional aspects of the technosphere can be better understood, including its composition, spatial distribution, internal processes, links to the Earth system, and consequences for human wellbeing.
  We would therefore propose a revision of the text to better indicate that our end-use orientation is just one possible approach to categorization, and to suggest that other categorizations might be more appropriate for different research questions. In addition, we propose to improve the methodological description of the classification, including an expanded priority scheme that reduces arbitrariness in categorization. (For example, white boards and bibles would be unambiguously classified as informational given that the information category has a higher priority than alternative uses.)
  We empathize with the reviewer’s dissatisfaction regarding the blurry division between Ambient Context and Activity-specific buildings. However we hesitate to remove the distinction, since the obvious alternative would be to lump all buildings together, which would obscure the functional roles of a very large portion of the technosphere. Factories, refineries and smelters are highly distinct from residential buildings, and we feel that it is useful to be able to identify them as such. We will attempt to address this by considering ways to better demarcate the division, and to reduce ambiguity.
  We would also note that this is intended to be a first version of the end use classification, and we would welcome revisions that could improve it, as well as alternative bases for categorization. The ongoing evolution of the technosphere, which the reviewer highlights, is a very important point that we will also mention in the revision. We attempted to design categories that could be expected to remain relatively stable over time, but it is unavoidable that innovative end uses will continue to arise (as they have in the past), and we will indicate that these could be addressed with future versions. In addition, we do not think that the classification should be limited to what can be discerned using presently-available data, but should also provide a foundation that can be populated with improved data in future (e.g. better resolved data on non-residential buildings). We plan to add these details to the revised manuscript.
  In summary, we very much appreciate the specific comments as well as the suggestion to make the language more reflexive, and plan to revise the manuscript accordingly.
  Regarding Specific Comment 2, we also entirely agree that mass, on its own, is insufficient for a thorough understanding the technosphere. We will clarify that the mass distribution shown here is intended to provide an overview of this classification and framework, and that many interesting aspects remain to be resolved next, such as energy conversion rates, metal composition, toxin production, lifetimes, etc. We also acknowledge that the maps simply quantify the mass per unit area of the technosphere components over the planetary surface, as a birds’ eye perspective, and do not reflect the impact of human activities on the hinterland that supports urban environments. We will make this more clear in the revision, as well adding further insight regarding features of the observed distributions.
  In regard to Specific Comment 3, we appreciate the suggestion to address the concept of technofossils, and will include this in the revised manuscript.
  Thanks also for the Technical Comments, we will address these. Please note that the size of the technosphere appears at a few points in the paper as being on the order of 1,000 Gt (e.g. Fig. 1, text line 227, Fig. 5), which is equal to a trillion tons.
  Again, we appreciate the reviewer’s time and insights, and are confident that they will help to produce a significantly improved revised manuscript.
  
  Citation: https://doi.org/10.5194/egusphere-2024-1133-AC1
RC2:
'Comment on egusphere-2024-1133', Anonymous Referee #2, 07 Jun 2024

The paper aims to refine the definition of the technosphere, introduce the End-Use Technosphere Classification (EUTEC) for systematic categorization, quantitatively assess the mass and spatial distribution of the technosphere, and understand its dynamic growth patterns. In addition, it seeks to establish an interdisciplinary foundation linking the material composition of the technosphere to its functionality and impact on human well-being, while emphasizing the need for further integration with Earth system science. I think these are all important and necessary goals, and the authors make substantial progress toward some of them. However, I am very ambivalent about the manuscript as a whole. I very much appreciate the enormous effort the authors have made to quantify all aspects of the technosphere. This is already an important and substantial contribution, despite the fact that there is a lot of uncertainty in some aspects of this quantification.

It is the analytical part that I struggle with. I have some disagreements with some of the definitions of the technosphere and the proposed dynamics of the technosphere as described by the authors. I also do not see how the authors suggest that the analytical part contributes to the inclusion of the technosphere in Earth system models.
My main issues concern: (1) the suggested autocatalytic nature of the technosphere and the implications of these assumptions for the potential modeling of the technosphere in an Earth system model, (2) the boundary definitions of the technosphere, (3) the overemphasis on infrastructure growth to satisfy end-user demand, to the exclusion of all other factors that shape the technosphere, and (4) the suggested intrinsic dynamic of the technosphere as described by equation 2.

In addition, I agree with the first reviewer on all their major points so will not reiterate them here.
(1) I would disagree that the technosphere as defined here (i.e., excluding „human activities or mental constructs such as institutions, corporations, or social norms“) can usefully be described as autocatalytic. I agree that the quality and quantity of artifacts can determine the quality and quantity of artifacts that can be produced, but I would argue that the technosphere as defined here does not possess the inherent properties required for self-sustaining, self-enhancing growth and evolution. Instead, it is mostly shaped and driven by external societal and human factors (besides the availability of material inputs).
The development and evolution of technological systems is highly dependent on human innovation, decision-making, and societal needs, and is driven by societal goals, economic factors, regulatory environments, and cultural values (all of these factors often shaped by whoever holds power in these societies). Without the input of human creativity and direction, technological systems do not inherently create or improve themselves (the current AI hype notwithstanding).

In my understanding, for a system to be autocatalytic, it must have self-sustaining mechanisms that promote its own growth and complexity. In the absence of human and societal inputs, the technosphere (as defined here) lacks the intrinsic capacity to innovate or evolve autonomously. Technological systems require maintenance, updates, and guidance from human agents. While there are certainly feedback loops within technological systems (e.g., improved machinery leading to more efficient production of new machinery), these loops are initiated and maintained by human intervention and societal demands.
I admit that some of my disagreements may be more a matter of semantics. For example, I do not find the arguments in Herrmann-Pillath 2018 based on Stuart Kauffman's notion of autonomous agents very convincing, nor am I sure that the definitions of the boundaries of the technosphere used there and here are consistent. If one understands industrial metabolism as a property of industrial societies rather than a property of industrial systems (as originally defined by Aryres), I would agree with the characterization of this coupled system as autocatalytic, where flows of energy and material are mediated by both physical infrastructures, symbolic structures, and human labor. Of course, I could very well be wrong about this, or misunderstand the intent of why it is appropriate to call the technosphere as defined in the manuscript autocatalytic. My main concern, and the reason I am spending so much time on this issue, is what follows from this assertion. I will return to this below.
(2) My next issue is with the third definition paragraph of the technosphere regarding when an „in-use component ceases to be fit to serve an intended end-use“. You concede that this „boundary is often subject to social characteristics“ but then you claim that it „is nonetheless observationally quite easy to identify.“ I am not convinced that it is so easy, and I am not sure how significant the term "intended" is in this statement. As such, I find this argument perhaps overly simplistic and potentially biased toward capitalist or Western perspectives that do not take into account the nuanced and subjective nature of determining when an object is no longer fit for its intended end use. For example, I live next to a dilapidated building that is being used as an impromptu skate park and a place for people to take drugs. It is not being maintained in any relevant sense, but I would still consider it part of the technosphere. Or, for that matter, the Pyramids of Giza probably left the technosphere by your definition before they were repurposed as a tourist attraction. I would argue that many structures and objects retain their place in the technosphere even though they no longer serve their original or intended purpose. I think the definition overlooks the different ways that different societies might use and maintain objects, reflecting a bias that doesn't apply universally to all cultural contexts. The reason you give for excluding waste is that it is difficult to determine when waste ceases to be waste. I would argue that for many parts of the technosphere it is difficult to judge when they begin to be waste, so we are simply shifting the conceptual uncertainty from one place to another.
(3) I also struggle with the justification for the end-use categorization, which says that "human end-use outcomes are what motivates [sic] the existence of the technosphere - every component of the technosphere was motivated by at least one type of intended end use". For example, you say that buildings exist or are constructed primarily to "provide humans with a more comfortable, attractive, or otherwise desirable immediate environment". There is, of course, some truth to this (while acknowledging the first reviewer's criticism that this is often more a matter of survival than comfort and beauty). However, much of the construction of buildings is actually driven by capital looking for investment opportunities. This can be seen everywhere, perhaps most dramatically in the infamous Chinese "ghost towns", where the end use of providing shelter for people is not even a secondary use, as it would devalue the investment. The same could be said for most types of infrastructure. Many additional lanes for roads are not built to make mobility more efficient and comfortable, as decades of transportation studies show that building additional lanes does not reduce congestion or make mobility more effective. So I, for one, would strongly disagree with the conclusion (l320) that the infrastructure that locks in car-dependency in may cities exists primarily for the purpose of making the „relocation of humans and materials faster and more convenient.“ Or, for example, Timothy Mitchell (Carbon Democracy) has written extensively on what motivated and shaped the development of the global oil infrastructure, which was only tangentially related to an existing demand for the end use of oil.
More generally, I think there is much to be said for the argument that infrastructure can very often be seen as a physical manifestation of power rather than simply a neutral response to a demanded end use. Political economy or science and technology studies would argue that technological systems and infrastructures are not neutral, but are imbued with the values and interests of those who design, build and control them. Infrastructure can thus be seen as a materialization of social order and power dynamics that shape how people interact with technology and each other (e.g. works by Foucault, Harvey, Castells, Levebre, and many others). This reference may also be helpful in the context of this manuscript [1]. This concern also comes back to the characterization of the technosphere as mainly autocatalytic, which seems to negate or at least disregard these concerns. I think it is not only wrong, but even dangerous, to ascribe a largely intrinsic and autonomous dynamic to the technosphere and to model it in an Earth system model in order to explore possible future trajectories. To be somewhat polemical, there is a danger of cementing a destructive status quo, based in no small part on a particular (symbolic!) ideology, and recasting it as a neutral and intrinsic property of inanimate matter.
(4) The whole focus of my critique on the autocatalytic (and thus autonomous) nature of the technosphere is driven by my concern about what the implied end use of equation (2) is. This is not clearly stated. However, you state that a major goal is to integrate the technosphere into Earth system models. There are few specifics on how to do this, but since you give this equation, I assume you are suggesting that it would be helpful. Equation 2 fits one or more coefficients to the historical growth function of the technosphere. This describes the empirical observations, but it has no explanatory power and is therefore not really suitable for predicting much about the future. On the one hand, we can already see a certain saturation in the throughput of non-metallic minerals in developed economies, and the exponential growth is coming from regions that are currently urbanizing/industrializing and thus accumulating large stocks. Also, similar levels of end-use in the US and Europe in terms of roads and buildings are achieved at very different levels of material use, and it is unlikely that Europe will grow these stocks to US levels.
Overall, since any kind of "sustainable" social metabolism is likely to require a reduction in material throughput, which is quite achievable in rich societies while increasing well-being, I do not think it makes sense to extend Earth system models with an "autonomous" technosphere module that exhibits intrinsic exponential growth.
Small things:

Table 2: why only fluids in pipelines and not also gas?

Figure A1: The category "Technosphere" should be "Technosphere Creation & Maintenance" to avoid throwing the reader into a recursive loop. =)
On the one hand, I appreciate the motivation and goals of the paper, and I think that the empirical part of the paper should definitely be published because it is a valuable contribution to knowledge. Here I would urge you to consider publishing the data separately in a data repository (like zenodo) and ideally also any software code used to transform the already published data into your results.

As may have become clear, I am less comfortable with the analytical part. My personal suggestion would be to simply remove that part altogether. I realize that this is probably not a suggestion you will want to follow. In that case, I think a fairly major revision of this part is needed, one that seriously addresses the issues I have raised. The questions I hope to have answered are:
Why does it matter whether the technosphere is autocatalytic or not? What implications does this have for how the technosphere would be implemented in Earth system models? How is such a characterization justified or operationalizable if all symbolic structures are excluded?
Why the focus on end use? How should dynamics of the technosphere that are not driven by end-use demand be included in a possible operationalization in Earth system models? What about the role of power shaping the trajectory of the technosphere (either political or economic)?
How is this definition and categorization of end use likely to hold in future societies that may have very different social and economic relationships with their material environment?
What is the role of Equation 2 with respect to the inclusion of the technosphere in Earth system models, and how is an intrinsic exponential growth dynamic useful for modeling potential post-growth, steady state, or other economic imaginaries where infrastructure stocks are saturated?
I apologize for the length of this review and hope that you will find some of the points I have raised interesting and helpful in improving the manuscript.
[1] Schaffartzik, A. et al. The transformation of provisioning systems from an integrated perspective of social metabolism and political economy: a conceptual framework. Sustain Sci 16, 1405–1421 (2021).

Citation: https://doi.org/10.5194/egusphere-2024-1133-RC2
- AC2: 'Reply on RC2', Eric Galbraith, 18 Jul 2024
  
  We are grateful to Reviewer 2 for the depth with which they have engaged with the manuscript, and for the wealth of thoughtful comments. We reply to these below, point-by-point, with the reviewer comments in italics.
  The paper aims to refine the definition of the technosphere, introduce the End-Use Technosphere Classification (EUTEC) for systematic categorization, quantitatively assess the mass and spatial distribution of the technosphere, and understand its dynamic growth patterns. In addition, it seeks to establish an interdisciplinary foundation linking the material composition of the technosphere to its functionality and impact on human well-being, while emphasizing the need for further integration with Earth system science. I think these are all important and necessary goals, and the authors make substantial progress toward some of them. However, I am very ambivalent about the manuscript as a whole. I very much appreciate the enormous effort the authors have made to quantify all aspects of the technosphere. This is already an important and substantial contribution, despite the fact that there is a lot of uncertainty in some aspects of this quantification.
  It is the analytical part that I struggle with. I have some disagreements with some of the definitions of the technosphere and the proposed dynamics of the technosphere as described by the authors. I also do not see how the authors suggest that the analytical part contributes to the inclusion of the technosphere in Earth system models.
  
  My main issues concern: (1) the suggested autocatalytic nature of the technosphere and the implications of these assumptions for the potential modeling of the technosphere in an Earth system model, (2) the boundary definitions of the technosphere, (3) the overemphasis on infrastructure growth to satisfy end-user demand, to the exclusion of all other factors that shape the technosphere, and (4) the suggested intrinsic dynamic of the technosphere as described by equation 2.
  In addition, I agree with the first reviewer on all their major points so will not reiterate them here.
  
  (1) I would disagree that the technosphere as defined here (i.e., excluding „human activities or mental constructs such as institutions, corporations, or social norms“) can usefully be described as autocatalytic. I agree that the quality and quantity of artifacts can determine the quality and quantity of artifacts that can be produced, but I would argue that the technosphere as defined here does not possess the inherent properties required for self-sustaining, self-enhancing growth and evolution. Instead, it is mostly shaped and driven by external societal and human factors (besides the availability of material inputs).
  
  The development and evolution of technological systems is highly dependent on human innovation, decision-making, and societal needs, and is driven by societal goals, economic factors, regulatory environments, and cultural values (all of these factors often shaped by whoever holds power in these societies). Without the input of human creativity and direction, technological systems do not inherently create or improve themselves (the current AI hype notwithstanding).
  In my understanding, for a system to be autocatalytic, it must have self-sustaining mechanisms that promote its own growth and complexity. In the absence of human and societal inputs, the technosphere (as defined here) lacks the intrinsic capacity to innovate or evolve autonomously. Technological systems require maintenance, updates, and guidance from human agents. While there are certainly feedback loops within technological systems (e.g., improved machinery leading to more efficient production of new machinery), these loops are initiated and maintained by human intervention and societal demands.
  
  I admit that some of my disagreements may be more a matter of semantics. For example, I do not find the arguments in Herrmann-Pillath 2018 based on Stuart Kauffman's notion of autonomous agents very convincing, nor am I sure that the definitions of the boundaries of the technosphere used there and here are consistent. If one understands industrial metabolism as a property of industrial societies rather than a property of industrial systems (as originally defined by Aryres), I would agree with the characterization of this coupled system as autocatalytic, where flows of energy and material are mediated by both physical infrastructures, symbolic structures, and human labor. Of course, I could very well be wrong about this, or misunderstand the intent of why it is appropriate to call the technosphere as defined in the manuscript autocatalytic. My main concern, and the reason I am spending so much time on this issue, is what follows from this assertion. I will return to this below.
  
  We appreciate these points, which have made it very clear to us that we need to provide more discussion regarding the term ‘autocatalytic’. We entirely agree that the technosphere is not self-sustaining, and is entirely shaped and driven by human factors: it is not, by any means, autonomous. (In fact, in thinking this over, we were at pains to come up with any autocatalytic systems that can sustain growth in isolation – they are always dependent on some other forms of input and interactions, including living organisms.) Strictly, we had been thinking of the term in the sense of the chemistry definition: a reaction is autocatalytic if one or more products of the reaction accelerate the reaction rate. Because exponential growth is a feature of autocatalysis, we find it a useful description.
  
  We would propose the addition of a definition of autocatalysis in order to prevent confusion, together with a discussion that clarifies that this does not imply autonomy of the technosphere – its creation is tightly coupled to human actions, and is therefore dependent on social dynamics. We are also experimenting with the idea of modifying the technosphere classification to include a ‘catalytic’ category of entities that accelerate/modify the outputs of human actions, which should help to better clarify the concept of autocatalysis in this context.
  (2) My next issue is with the third definition paragraph of the technosphere regarding when an „in-use component ceases to be fit to serve an intended end-use“. You concede that this „boundary is often subject to social characteristics“ but then you claim that it „is nonetheless observationally quite easy to identify.“ I am not convinced that it is so easy, and I am not sure how significant the term "intended" is in this statement. As such, I find this argument perhaps overly simplistic and potentially biased toward capitalist or Western perspectives that do not take into account the nuanced and subjective nature of determining when an object is no longer fit for its intended end use. For example, I live next to a dilapidated building that is being used as an impromptu skate park and a place for people to take drugs. It is not being maintained in any relevant sense, but I would still consider it part of the technosphere. Or, for that matter, the Pyramids of Giza probably left the technosphere by your definition before they were repurposed as a tourist attraction. I would argue that many structures and objects retain their place in the technosphere even though they no longer serve their original or intended purpose. I think the definition overlooks the different ways that different societies might use and maintain objects, reflecting a bias that doesn't apply universally to all cultural contexts. The reason you give for excluding waste is that it is difficult to determine when waste ceases to be waste. I would argue that for many parts of the technosphere it is difficult to judge when they begin to be waste, so we are simply shifting the conceptual uncertainty from one place to another.
  
  Thanks for the request to better define the waste-adjacent boundary of the technosphere. This is certainly among the most difficult parts to delineate, and we welcome the opportunity to improve it.
  It’s important to note that we primarily aim for definitions that are useful for estimating stocks and flows. Given this aim, it does not seem to us that uncertainty regarding a small fraction of stocks at the margin of what is clearly in-use - such as a recently vacated but still intact building – is on the same scale as the many billions of tons of waste that are distributed across the land surface and oceans in clearly unusable forms. On long timescales, mass that is clearly waste will keep accumulating and gradually be transformed into degraded states (at least the parts that are not fossilized), dwarfing the relatively small grey area at the margins of what is in-use. Thus, in our assessment, our boundary is not simply a question of kicking the can down the road, but one that constrains the fuzziness within a smaller part of the system.
  Still, we very much agree that it would be great to have a better solution for this! For example, we are considering redefining in-use as being maintained in an appropriate state to provide an end-use, since an entity which falls out of use will always decay towards an unusable state (at some rate). In this case, the abandoned building would be ‘in use’ since it remains capable of providing an end use as a skate park and drug injection site. Meanwhile the pyramids, having been utterly abandoned to decay for millennia, would indeed have left the in-use technosphere while they did not provide an end-use to humans. To us this seems as reasonable as any alternative, and would be similarly applied to myriad other ruins around the world in various states of decay. We could now refer to these as technofossils, following the suggestion of reviewer 1.
  We will consider these nuances further as we worked towards an improved boundary description in the revised manuscript (and would welcome any further suggestions!).
  (3) I also struggle with the justification for the end-use categorization, which says that "human end-use outcomes are what motivates [sic] the existence of the technosphere - every component of the technosphere was motivated by at least one type of intended end use". For example, you say that buildings exist or are constructed primarily to "provide humans with a more comfortable, attractive, or otherwise desirable immediate environment". There is, of course, some truth to this (while acknowledging the first reviewer's criticism that this is often more a matter of survival than comfort and beauty). However, much of the construction of buildings is actually driven by capital looking for investment opportunities. This can be seen everywhere, perhaps most dramatically in the infamous Chinese "ghost towns", where the end use of providing shelter for people is not even a secondary use, as it would devalue the investment. The same could be said for most types of infrastructure. Many additional lanes for roads are not built to make mobility more efficient and comfortable, as decades of transportation studies show that building additional lanes does not reduce congestion or make mobility more effective. So I, for one, would strongly disagree with the conclusion (l320) that the infrastructure that locks in car-dependency in may cities exists primarily for the purpose of making the „relocation of humans and materials faster and more convenient.“ Or, for example, Timothy Mitchell (Carbon Democracy) has written extensively on what motivated and shaped the development of the global oil infrastructure, which was only tangentially related to an existing demand for the end use of oil.
  
  More generally, I think there is much to be said for the argument that infrastructure can very often be seen as a physical manifestation of power rather than simply a neutral response to a demanded end use. Political economy or science and technology studies would argue that technological systems and infrastructures are not neutral, but are imbued with the values and interests of those who design, build and control them. Infrastructure can thus be seen as a materialization of social order and power dynamics that shape how people interact with technology and each other (e.g. works by Foucault, Harvey, Castells, Levebre, and many others). This reference may also be helpful in the context of this manuscript [1]. This concern also comes back to the characterization of the technosphere as mainly autocatalytic, which seems to negate or at least disregard these concerns. I think it is not only wrong, but even dangerous, to ascribe a largely intrinsic and autonomous dynamic to the technosphere and to model it in an Earth system model in order to explore possible future trajectories. To be somewhat polemical, there is a danger of cementing a destructive status quo, based in no small part on a particular (symbolic!) ideology, and recasting it as a neutral and intrinsic property of inanimate matter.
  
  We are glad this point was brought up, as we realize that our meaning of motivating end-uses was not sufficiently fleshed out here.
  We absolutely agree that infrastructure cannot be seen as a neutral response to demanded end uses. However, in the aim of providing a categorization system that is as unambiguous as possible, we wish to avoid being confounded by complex social factors. As stated by the reviewer, ‘…Much of the construction of buildings is actually driven by capital looking for investment opportunities’. Investment decisions are one component of what can be referred to as a social motivations, distinguished here from physical outcome-oriented motivations (e.g. Galbraith et al., PLoS One 2022). One reason humans can appear so bewildering is that we have many layers of socially-coordinated motivations that determine our behaviours in interacting ways. For example, while it is certainly true (and important) that capitalist investment decisions underlie the construction of many buildings at present, it is equally true that buildings are materially constructed by workers, in order to be paid. Or, that a large portion of building material footprints is motivated by the desire to raise the social status of the owners.
  Similarly, we agree it’s important that an increase in mass of a technosphere component does not linearly predict an increase in service provision. The reviewer makes excellent points about the dysfunction of some constructions in providing services, such as the ineffective addition of road lanes. One could add the point that building larger houses does not necessarily improve the lived experience of people living within them, or that having more clothes than one needs does not improve their ambient context, or more cars than one can simultaneously drive does not help in transportation. Changes in technology over time also introduce a lot of nonlinearity into the relationship between mass and function. So, one should definitely not take the conclusion that an increase in mass of a stock will lead to a corresponding increase in the provision of services. We will be sure to mention this.
  Power, values and social order can be seen as aspects of complex social systems that motivate the activity of technosphere construction, and determine who has access to which parts of it. But because these are multi-layered and depend on cultural and economic/political features of societies, we do not find them better suited as a basis for categorization, in an unambigious manner that would lend itself to long timescales. (The reviewer alludes to this in the comment regarding the difficulty of predicting social and economic relationships in the distant future.)
  Our focus on the physical end-uses of technosphere entities is intended to provide a globally-applicable and robust categorization, rather than capture the complexity of social dynamics that contribute to their creation. We will add text to explain this better, and also to point out that the social factors are very important for driving motivations, even though they are not used for this particular categorization. In addition, although we don’t know what distant future societies might look like, we do have an idea of what ancient hunter-gatherer societies looked like, so we will include a discussion of the EUTEC would have applied to their artifacts to help provide a sense of likely robustness over time.
  
  (4) The whole focus of my critique on the autocatalytic (and thus autonomous) nature of the technosphere is driven by my concern about what the implied end use of equation (2) is. This is not clearly stated. However, you state that a major goal is to integrate the technosphere into Earth system models. There are few specifics on how to do this, but since you give this equation, I assume you are suggesting that it would be helpful. Equation 2 fits one or more coefficients to the historical growth function of the technosphere. This describes the empirical observations, but it has no explanatory power and is therefore not really suitable for predicting much about the future. On the one hand, we can already see a certain saturation in the throughput of non-metallic minerals in developed economies, and the exponential growth is coming from regions that are currently urbanizing/industrializing and thus accumulating large stocks. Also, similar levels of end-use in the US and Europe in terms of roads and buildings are achieved at very different levels of material use, and it is unlikely that Europe will grow these stocks to US levels.
  
  Overall, since any kind of "sustainable" social metabolism is likely to require a reduction in material throughput, which is quite achievable in rich societies while increasing well-being, I do not think it makes sense to extend Earth system models with an "autonomous" technosphere module that exhibits intrinsic exponential growth.
  
  We were initially surprised that equation 2 came across as concerning, but in retrospect can see why, and are very glad the reviewer brought this up. As discussed above, we did not intend to imply autonomy of the technosphere, and will clarify this in revision. We also clearly did not do a good job of contextualizing the equation within the overall paper, or discussing how the technosphere might be incorporated in models.
  We had meant the equation primarily as a descriptive tool of the observed dynamics, rather than a means to predict the future. Its utility is that it shows remarkably simple behaviour despite being the emergent outcome of a fantastically complicated, globally-interacting system. It provides a dynamical birds’-eye perspective on what has happened to date, rather than being a sophisticated tool for prediction. We agree with the evidence that there is a great potential to live well, and sustainably, by slowing and ultimately halting the technosphere growth – indeed it is hoped that this work can ultimately contribute to achieving that aim in a way that is good for human wellbeing.
  It may be worth pointing out that, by changing the coefficient (c) of the autocatalytic equation, one could represent a scenario of steady-state (setting c=0) or degrowth (setting c <0). However, in most cases this would probably be too simplistic to capture the dynamics of interest.
  There are many more sophisticated ways the technosphere could be included in Earth system models, other than using equation 2 with a constant coefficient, and we will add text to make this very clear. To our understanding, any of these must start from a physical quantitative basis for technosphere stocks and fluxes, just as exists for all other parts of the Earth system. The technosphere is what enables humans to control large fluxes of matter and energy, by acting as a catalyst of human activities. Linking the technosphere to the processes it drives and facilitates, in combination with human actions, can provide a mechanistic basis for coupling human-and non-human components of Earth system models (e.g. formulation in Galbraith et al., ESD 2021). The categorical framework and mass estimates presented here provide one avenue to start thinking towards physically-based models, of which many configurations are possible. We will add further text to clarify this.
  Small things:
  Table 2: why only fluids in pipelines and not also gas?
  
  By ‘fluids’ we had meant both liquids and gases (in the physicist’s sense). We will clarify this.
  
  Figure A1: The category "Technosphere" should be "Technosphere Creation & Maintenance" to avoid throwing the reader into a recursive loop. =)
  Sorry about the recursive loop! Hope it did not cause a stack overflow. This will be corrected.
  
  On the one hand, I appreciate the motivation and goals of the paper, and I think that the empirical part of the paper should definitely be published because it is a valuable contribution to knowledge. Here I would urge you to consider publishing the data separately in a data repository (like zenodo) and ideally also any software code used to transform the already published data into your results.
  Thanks, we do plan to make the data and scripts publicly available via Zenodo.
  As may have become clear, I am less comfortable with the analytical part. My personal suggestion would be to simply remove that part altogether. I realize that this is probably not a suggestion you will want to follow. In that case, I think a fairly major revision of this part is needed, one that seriously addresses the issues I have raised.
  As anticipated by the reviewer, we would rather not remove the analytic part from the paper. We did initially consider removing the discussion of autocatalysis, but the comments of Reviewer 3 support keeping it. So, instead, we propose to revise and expand the discussion in accordance with the comments, which we feel can be done so in a thorough way that will significantly improve the manuscript.
  The questions I hope to have answered are:
  
  Why does it matter whether the technosphere is autocatalytic or not? What implications does this have for how the technosphere would be implemented in Earth system models? How is such a characterization justified or operationalizable if all symbolic structures are excluded?
  
  Why the focus on end use? How should dynamics of the technosphere that are not driven by end-use demand be included in a possible operationalization in Earth system models? What about the role of power shaping the trajectory of the technosphere (either political or economic)?
  
  How is this definition and categorization of end use likely to hold in future societies that may have very different social and economic relationships with their material environment?
  
  What is the role of Equation 2 with respect to the inclusion of the technosphere in Earth system models, and how is an intrinsic exponential growth dynamic useful for modeling potential post-growth, steady state, or other economic imaginaries where infrastructure stocks are saturated?
  
  I apologize for the length of this review and hope that you will find some of the points I have raised interesting and helpful in improving the manuscript.
  We hope the responses above have answered these questions, and again thank the reviewer for the thoughtful points.
  
  Citation: https://doi.org/10.5194/egusphere-2024-1133-AC2
RC3:
'Comment on egusphere-2024-1133', Anonymous Referee #3, 12 Jun 2024
General comments
The authors attempt to define and quantify the technosphere, to assist its integration into Earth system analysis. Their approach differs from prior work largely by centering their definition on human-intended end uses of technosphere elements. This is an ambitious, challenging and, in my view, important task. I think the authors’ approach has many strengths and is a substantial step towards their stated goals. However, there remains work to do before they can defend their claim that they are able to exhaustively and exclusively characterise the technosphere. Like the other reviewers I support publication of at least part of the text, on the condition that the authors revise it to be more explicit about the limitations of their taxonomy and more clearly identify the areas where more nuanced development is needed.
Specific comments
Reviewers 1 and 2 raise many interesting lines of commentary that I think stand well on their own. I would like to extend one line that runs through both of their reviews: that, however convenient it may be for accounting purposes, identifying the technosphere with material composition, detached from the cognition within the social processes that coevolve with the technosphere, does not achieve the authors' goal of exhaustive characterisation. For want of a better analogy, this seems like trying to describe how and why a computer is useful for its user by describing its material composition while ignoring the software installed. I agree with reviewer 2, that to understand the autocatalytic behaviour described by the authors, one needs to consider the sociotechnosphere (by which I mean something like 'technosphere as defined by the authors plus social cognition'), not just the technosphere as defined. I also agree with reviewer 2 that much of the technosphere is better understood as resulting from emergent socioeconomic dynamics that can be quite disconnected from the needs and intentions of the individual humans who interact with the resulting technosphere elements.
However, I part ways somewhat with reviewer 2 on the value of the manuscript’s analytical section, again with the proviso that equation 2 describes the sociotechnosphere and not just the technosphere. The point here is not to include the biomass of humans but the cognitive processes that facilitate the autocatalytic dynamic. In my view, that the data demonstrate exponential growth of the mass of the sociotechnosphere does have explanatory power insofar as it places that system within a broader class of exponentially growing autocatalytic systems about which a lot has been learned. Although we may like to believe that by exercising our collective agency we can steer the system and convert exponential to logistic growth on the relevant timescale, it may be that the emergent system dynamics are now beyond our control and overshoot-and-collapse is inevitable. In my opinion, presenting and interpreting the century-worth of data of the system’s ’gravimetric growth’ is very worthwhile. To agree again with reviewer 2 though, I suggest presenting equation 2 together with a discussion of its potential limitations; future data may reveal a need to elaborate the model to accommodate saturation, collapse or other potential divergences from exponential growth.
I would also like to venture in a direction not explored by reviewers 1 or 2, but still on the theme of limitations due to grounding the taxonomy purely in material composition. I offer a few rather pedantic ’thought experiments’ that are intended to help the authors stress test their definitions by considering edge cases. I agree with reviewer 1 that aiming for a perfect taxonomy is a fool’s errand, so I’m not suggesting that the authors need to address everything I raise here in detail before publication. Rather, I suggest considering whether they can use my provocations to strengthen their taxonomy by refining definitions.
In short, the weakness I see is that where something fits in the authors’ taxonomy depends strongly on scale of analysis and context. From line 70, the authors state that they ’follow the Earth system science convention of defining a sphere in terms of the nature of the matter of which is it comprised, rather than in terms of specific processes or fluxes.’ They illustrate this with the example of the atmosphere, which is defined by its location above the Earth’s surface and its chemical composition rather than its internal dynamics. This seems to imply a kind of coarse-graining over the atmosphere at some large scale of analysis, such that only its average material composition over that scale is considered.
This approach may run into trouble when trying to identify boundaries between Earth system spheres, as the authors do later. If a thing is categorised according to its average material composition, why exclude a recently dead organism from the biosphere? If it decomposes through heterotrophic respiration, presumably the mass assimilated by the heterotrophs is still part of the biosphere. I would agree that the original organism is no longer part of the biosphere but if this is true while some of its material still is part of the biosphere, then the organism was categorised not by its material composition but rather by its functional organisation (the thing that make it alive). Presumably, individual metabolite molecules within the organism are not alive but are categorised in the biosphere because they are located inside a living organism. If this is the case, why isn’t a person’s pacemaker part of the biosphere? If the answer is that its average material composition differs from the rest of the host’s body, would the distinction still hold if it were possible to build a pacemaker from biomaterials? By analogy, why isn’t a person inside a building a part of the technosphere? If they are, is their biomaterial-based pacemaker now also part of the technosphere? If they step outside, does it become part of the biosphere? Conversely, given that human settlements are autocatalytic dissipative structures like organisms, and resulted from the same evolutionary process, are they really ’nonliving matter’? This seems clear at the scale of a brick but is perhaps less clear at the scale of a city. If a genetically modified sheep is part of the biosphere, is its DNA, which is no more alive than the brick, part of the biosphere or technosphere? In which sphere would a genetically modified virus be categorised (given that the living/nonliving status of viruses is controversial even without engineering)? What about a completely novel, artificial virus? Would the answer change if it were inside/outside an organism? Do these examples demonstrate that applying the authors’ taxonomy unambiguously may not be possible without first defining a scale of analysis and considering context? If the authors wish to connect their definitions to accounting systems that have different resolutions, these issues may become crucial. Perhaps the authors have faced such questions before in their work on sociometabolic accounting (contrasting MEFA with MuSIASEM, for example). I think they should make some effort to address these issues in their paper and justify any relevant simplifying assumptions underlying their taxonomy.
94-97: The issue of human agency arises here and while this seems to support my earlier comments about the necessity of including (social) cognition in an adequate definition of the technosphere, I would like to briefly also consider a less anthropocentric view. The existence of genuine agency continues to be controversial even in humans, and the existence of technologies with behaviour that is indistinguishable from human agency is becoming ever more plausible. Accordingly, technologies that make their own institutions and norms within systems of collective computation are also becoming increasingly plausible. Might it therefore be too restrictive to exclude phenomena like institutions and norms from the technosphere even if the human versions are excluded? Even in the absence of strong AI, some form of artificial life may be achieved. Would this be part of the technosphere or biosphere? Perhaps the taxonomy needs a tiebreaking rule for such cases?
Technical corrections
At first glance I found the title a bit confusing because the word ’resolving’ is ambiguous (Is the paper trying to fix the technosphere?). The abstract clarifies the meaning but the title alone does not clearly communicate what the paper is about.

In Table 1 caption, is ’of which they are comprised’ correct English? Presumably the Earth spheres comprise various kinds of matter rather than the other way around. Maybe use ’of which they are composed’?

33: It’s unclear whether the 20 TW refers to the technosphere or terrestrial above-ground NPP.

70-71, and also 178: Unusual use of ’comprised’ again.

259: Three million years ago seems to be too early for reference to a human population. Hominin population might be more accurate.

263-265: Before and after equation (1), you carefully specify units in which all the terms are measured. Shouldn’t the equation be insensitive to unit selection, provided all terms are consistent? I note that you specify kg for some terms and g for others; is this intentional?
Citation: https://doi.org/10.5194/egusphere-2024-1133-RC3
- AC3: 'Reply on RC3', Eric Galbraith, 18 Jul 2024
  
  Many thanks to Reviewer 3 for these comments, which are very helpful for preparing a revised manuscript. We copy the comments below (in italics) and provide a point-by-point response.
  
  General comments
  
  The authors attempt to define and quantify the technosphere, to assist its integration into Earth system analysis. Their approach differs from prior work largely by centering their definition on human-intended end uses of technosphere elements. This is an ambitious, challenging and, in my view, important task. I think the authors’ approach has many strengths and is a substantial step towards their stated goals. However, there remains work to do before they can defend their claim that they are able to exhaustively and exclusively characterise the technosphere. Like the other reviewers I support publication of at least part of the text, on the condition that the authors revise it to be more explicit about the limitations of their taxonomy and more clearly identify the areas where more nuanced development is needed.
  
  We are happy to be more explicit about limitations and add nuance, as we agree this would be a good idea.
  Specific comments
  
  Reviewers 1 and 2 raise many interesting lines of commentary that I think stand well on their own. I would like to extend one line that runs through both of their reviews: that, however convenient it may be for accounting purposes, identifying the technosphere with material composition, detached from the cognition within the social processes that coevolve with the technosphere, does not achieve the authors' goal of exhaustive characterisation. For want of a better analogy, this seems like trying to describe how and why a computer is useful for its user by describing its material composition while ignoring the software installed. I agree with reviewer 2, that to understand the autocatalytic behaviour described by the authors, one needs to consider the sociotechnosphere (by which I mean something like 'technosphere as defined by the authors plus social cognition'), not just the technosphere as defined. I also agree with reviewer 2 that much of the technosphere is better understood as resulting from emergent socioeconomic dynamics that can be quite disconnected from the needs and intentions of the individual humans who interact with the resulting technosphere elements.
  
  We absolutely agree that interaction between humans and the technosphere is essential for determining the dynamics, and now recognize that this was not sufficiently clear in the original manuscript. But it’s a complicated problem, and needs to be parsed. Here, the parsing is along a boundary that fits well with the understanding of the Earth system: there is a distinction between living organisms, including humans, and nonliving materials. The parsing offerred here does not prohibit the considerations of interactions between humans and technosphere, just as the identification of a living organism does not prevent considering its interactions with the nonliving environment. Organisms cannot survive in a vacuum, just as the technosphere cannot grow without human action (at least, at present). The parsing is intended to help think more lucidly about the interactions, rather than ignore them.
  We therefore plan to revise the text to make it clear that the dynamics of the technosphere are intimately linked with human motivations, coordinated through social mechanisms (including economic and political). Indeed, these must be what are responsible for the changes in the autocatalytic coefficient, resulting in changes that are most pronounced around 1950 and 1980.
  Because we feel the parsing is important, we would still maintain the distinction here of the technosphere as separate from the social, which – in physical terms – is couched in the neural structures of humans, coordinated by symbolic information exchange. From the Earth system perspective, the neurons are located within the biosphere. However there’s no problem with the autocatalytic nature of the technosphere being dependent on the social dynamics, in the same way that the growth of plankton in a culture depends on the nutrient content of the water in which they grow – autocatalysis does not imply autonomy. We propose to add text to better clarify this (see also response to Reviewer 2).
  
  However, I part ways somewhat with reviewer 2 on the value of the manuscript’s analytical section, again with the proviso that equation 2 describes the sociotechnosphere and not just the technosphere. The point here is not to include the biomass of humans but the cognitive processes that facilitate the autocatalytic dynamic. In my view, that the data demonstrate exponential growth of the mass of the sociotechnosphere does have explanatory power insofar as it places that system within a broader class of exponentially growing autocatalytic systems about which a lot has been learned. Although we may like to believe that by exercising our collective agency we can steer the system and convert exponential to logistic growth on the relevant timescale, it may be that the emergent system dynamics are now beyond our control and overshoot-and-collapse is inevitable. In my opinion, presenting and interpreting the century-worth of data of the system’s ’gravimetric growth’ is very worthwhile. To agree again with reviewer 2 though, I suggest presenting equation 2 together with a discussion of its potential limitations; future data may reveal a need to elaborate the model to accommodate saturation, collapse or other potential divergences from exponential growth.
  
  We appreciate the support of the autocatalytic discussion, which has convinced us to keep it within the paper, despite the hesitations of reviewer 2. We will add text to express that the equation does not imply perpetual growth – changes in the coefficient can equally lead to stasis or decay.
  I would also like to venture in a direction not explored by reviewers 1 or 2, but still on the theme of limitations due to grounding the taxonomy purely in material composition. I offer a few rather pedantic ’thought experiments’ that are intended to help the authors stress test their definitions by considering edge cases. I agree with reviewer 1 that aiming for a perfect taxonomy is a fool’s errand, so I’m not suggesting that the authors need to address everything I raise here in detail before publication. Rather, I suggest considering whether they can use my provocations to strengthen their taxonomy by refining definitions.
  
  Much appreciated – we will continue to try to improve the taxonomy in revision.
  In short, the weakness I see is that where something fits in the authors’ taxonomy depends strongly on scale of analysis and context. From line 70, the authors state that they ’follow the Earth system science convention of defining a sphere in terms of the nature of the matter of which is it comprised, rather than in terms of specific processes or fluxes.’ They illustrate this with the example of the atmosphere, which is defined by its location above the Earth’s surface and its chemical composition rather than its internal dynamics. This seems to imply a kind of coarse-graining over the atmosphere at some large scale of analysis, such that only its average material composition over that scale is considered.
  
  This approach may run into trouble when trying to identify boundaries between Earth system spheres, as the authors do later. If a thing is categorised according to its average material composition, why exclude a recently dead organism from the biosphere? If it decomposes through heterotrophic respiration, presumably the mass assimilated by the heterotrophs is still part of the biosphere. I would agree that the original organism is no longer part of the biosphere but if this is true while some of its material still is part of the biosphere, then the organism was categorised not by its material composition but rather by its functional organisation (the thing that make it alive).
  
  We agree that the aim to capture the global scale is integral to the approach, and that all of the spheres break down at some scale. In addition to the given example of a recently dead organism, one could ask whether bubbles of air mixed into the surface ocean by waves are atmosphere or hydrosphere? Do they similarly oscillate back and forth between one sphere and the other, as they are injected and subsequently outgas?
  As such, the sphere framework is not obviously suited to precise categorizations, at least not without a long list of instructions on how to treat edge cases. But in our mind that is not really the purpose, at least not here – the purpose is to be able to sketch in broad strokes the components of our planetary system, in a way that’s intuitive, within which to think more clearly about processes that extend up to the global scale. For accounting purposes, it may be necessary to develop edge case instructions, which may be difficult to base on first principles. But as long as this is just for a small fraction of the total matter, it does not undermine the utility of the overall approach for capturing the big picture.
  Perhaps the wording we chose for ‘material’ was not a good one, as the comment indicates that this could be interpreted as chemical composition. We were thinking more of a broader set of characteristics, which could include the processes that generated the material. Perhaps the distinction is that the sphere definition does not include the processes themselves, but only the material. We will work on improving this wording.
  Presumably, individual metabolite molecules within the organism are not alive but are categorised in the biosphere because they are located inside a living organism. If this is the case, why isn’t a person’s pacemaker part of the biosphere? If the answer is that its average material composition differs from the rest of the host’s body, would the distinction still hold if it were possible to build a pacemaker from biomaterials? By analogy, why isn’t a person inside a building a part of the technosphere? If they are, is their biomaterial-based pacemaker now also part of the technosphere? If they step outside, does it become part of the biosphere? Conversely, given that human settlements are autocatalytic dissipative structures like organisms, and resulted from the same evolutionary process, are they really ’nonliving matter’? This seems clear at the scale of a brick but is perhaps less clear at the scale of a city. If a genetically modified sheep is part of the biosphere, is its DNA, which is no more alive than the brick, part of the biosphere or technosphere?
  
  Many thanks for the provocative questions about the categorization. These have helped us to recognize some unwritten assumptions we’d been making, and the need to better define what an ‘in-use’ entity is.
  We would agree that the functional organization that makes something alive is key – but feel that this is still a characteristic of the matter. The matter is organized in a way that continues the living processes, if that makes sense.
  The pacemaker made of biomaterials is a good one. We would suggest that it could be distinguished because it is not generated through the production of proteins by ribosomes using information encoded by the organism’s DNA, which is perhaps the clearest hallmark of a living organism? But we will think about this a bit more!
  The person inside the building example points out an important distinction that absolutely should be improved in revision: the scale at which we distinguish an end-use entity. We refer to a recognizable entity as a physically coherent functional unit that serves an identifiable end use, without requiring further transformation. Many entities are assembled from multiple components, such as the building itself. We follow convention in taking the commonly recognized functional unit – the assembled building in this case. One could certainly design an alternative categorization that uses components as the scale - considering each brick, board, concrete slab, wire, pipe, switch, pane of glass, shingle, panel of wallboard, lightbulb, etc. separately in an alternative categorization. An entity may require activation by a human, and may require an energy conversion or another metabolite to generate the end use, but otherwise it can serve an end use on its own (though it could also be used in combination with other entitites). As such, a building does not include the furniture inside of it. This is consistent with the convention used by architects and builders, which can be approximated as everything that would not fall down if you turned a building upside-down. Thus, humans do not become part of the building by going inside of it. Similarly, clothes do not become part of the human while being worn, and a phone does not become part of clothing when you put it in your pocket.
  In which sphere would a genetically modified virus be categorised (given that the living/nonliving status of viruses is controversial even without engineering)? What about a completely novel, artificial virus? Would the answer change if it were inside/outside an organism?
  
  According to the definition that living organisms are constructed by proteins produced by their own ribosomes (above), viruses would not be living (since they don’t have ribosomes), whether natural or artificial. But this could be revisited if someone were to use a different definition of living organisms.
  Do these examples demonstrate that applying the authors’ taxonomy unambiguously may not be possible without first defining a scale of analysis and considering context? If the authors wish to connect their definitions to accounting systems that have different resolutions, these issues may become crucial. Perhaps the authors have faced such questions before in their work on sociometabolic accounting (contrasting MEFA with MuSIASEM, for example). I think they should make some effort to address these issues in their paper and justify any relevant simplifying assumptions underlying their taxonomy.
  
  Again, thanks for the provocative questions! They are very helpful in revising details of the taxonomy.
  94-97: The issue of human agency arises here and while this seems to support my earlier comments about the necessity of including (social) cognition in an adequate definition of the technosphere, I would like to briefly also consider a less anthropocentric view. The existence of genuine agency continues to be controversial even in humans, and the existence of technologies with behaviour that is indistinguishable from human agency is becoming ever more plausible. Accordingly, technologies that make their own institutions and norms within systems of collective computation are also becoming increasingly plausible. Might it therefore be too restrictive to exclude phenomena like institutions and norms from the technosphere even if the human versions are excluded? Even in the absence of strong AI, some form of artificial life may be achieved. Would this be part of the technosphere or biosphere? Perhaps the taxonomy needs a tiebreaking rule for such cases?
  
  The question of machine norms is interesting. Since these would be encoded within the matter of the technosphere, it would appear to make sense to include them within the technosphere, just as human norms – encoded in neurons – belong with the biosphere. However we would prefer not to delve into questions of agency in the current paper, as we feel it will be better handled by other investigators. We also do not see how it would alter the categorization, as it would appear to us that our categories are fairly agnostic as to the role of agency.
  Technical corrections
  
  • At first glance I found the title a bit confusing because the word ’resolving’ is ambiguous (Is the paper trying to fix the technosphere?). The abstract clarifies the meaning but the title alone does not clearly communicate what the paper is about.
  
  Thanks for the suggestion about the title. We had intended ‘resolving’ in the sense of improving the resolution of the currently-fuzzy technosphere concept.
  
  • In Table 1 caption, is ’of which they are comprised’ correct English? Presumably the Earth spheres comprise various kinds of matter rather than the other way around. Maybe use ’of which they are composed’?
  
  Thanks for correcting the incorrect use of ‘comprised’! Apparently the lead author has been making that mistake for many years.
  • 33: It’s unclear whether the 20 TW refers to the technosphere or terrestrial above-ground NPP.
  
  This should have been the technosphere – will be clarified.
  • 70-71, and also 178: Unusual use of ’comprised’ again.
  
  • 259: Three million years ago seems to be too early for reference to a human population. Hominin population might be more accurate.
  
  Thanks, good suggestion
  • 263-265: Before and after equation (1), you carefully specify units in which all the terms are measured. Shouldn’t the equation be insensitive to unit selection, provided all terms are consistent? I note that you specify kg for some terms and g for others; is this intentional?
  Thanks, these were intended as example units to help with understanding, this will be clarified and tidied up.
  
  Citation: https://doi.org/10.5194/egusphere-2024-1133-AC3

Eric Galbraith, Abdullah-Al Faisal, Tanya Matitia, William Fajzel, Ian Hatton, Helmut Haberl, Fridolin Krausmann, and Dominik Wiedenhofer

Supplement

https://doi.org/10.5194/egusphere-2024-1133-supplement

Eric Galbraith, Abdullah-Al Faisal, Tanya Matitia, William Fajzel, Ian Hatton, Helmut Haberl, Fridolin Krausmann, and Dominik Wiedenhofer

Viewed

Total article views: 631 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
389	194	48	631	38	32	25

HTML: 389
PDF: 194
XML: 48
Total: 631
Supplement: 38
BibTeX: 32
EndNote: 25

Views and downloads (calculated since 08 May 2024)

Month	HTML	PDF	XML	Total
May 2024	141	46	9	196
Jun 2024	110	62	9	181
Jul 2024	47	33	13	93
Aug 2024	36	14	13	63
Sep 2024	36	25	3	64
Oct 2024	19	14	1	34

Cumulative views and downloads (calculated since 08 May 2024)

Month	HTML	PDF	XML	Total
May 2024	141	46	9	196
Jun 2024	110	62	9	181
Jul 2024	47	33	13	93
Aug 2024	36	14	13	63
Sep 2024	36	25	3	64
Oct 2024	19	14	1	34

Viewed (geographical distribution)

Total article views: 675 (including HTML, PDF, and XML) Thereof 675 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 01 Nov 2024

Short summary

The technosphere, including buildings, infrastructure and all other non-living human creations, has become a major part of the Earth system. Here we provide a refined definition of the technosphere, and an end-use classification aligned with the physical outcomes of human activities. We use these definitions to describe the composition and spatial distribution of technosphere mass, and discuss the exponential character of its growth since 1900, which presents a challenge for sustainability.


Total:	0
HTML:	0
PDF:	0
XML:	0