The discussion so far
Involved were :
I have lightly edited the tweets to include readable link and format them nicer, no changes were made to the content.
RM : "Coming from an engineering background, I have identified memetics / generalized Darwinism / Universal Darwinism as relevant fields and topics, but have not found a singe talk about it. (...) I understood that these views are considered obsolete and wrong."
JW : The core of this analysis, the absence of both material culture and of any idea of evolution in CES, is I think excellent. The direction, sustainability, is more towards Darwin's pigeon breeding than natural selection, and in itself admirable; purposive rather than descriptive
IN: Do we have models/theories within CE that describe how this reflexivity influences evolution, while co-occuring with natural selection? Can we even meaningfully talk about "natural" selection in culture? Tech is normative and purposive, even if side effects are unintentional
JW “Reflexivity” is part of a dualist analysis with a mystical agent "mind". An evolutionary analysis, based on eg
- Causal understanding is not necessary for the improvement of culturally evolving technology
- The evolution of an ancient technology
- The evolution of air resonance power efficiency in the violin and its ancestors
would look only at the environment at whatever scale to identify factors of selection.
The case is simple. A webweaving spider is in obligate symbiosis with its web, its extended phenotype. Homo is the same, its obligate symbiont being the technosphere. The evo environment of the homo organism is the technosphere, and of the technosphere is the hom organism.
RM Ok, everything evolves in some environment. And what next? What if you're interested in real-life phenomena where there are many spiders, one big web, many ways of improving or destroying the web? And you're interested in the prediction of the specific situation? Not simple.
JW Its usually one spider, one web, iterated a million times with infinitesimal variation, selection by environment outside spider and web, and prediction plays no part in it
RM E.g. state policy. It consists of many documents that are replicated (because of the continuity of law most of regulations are preserved in subsequent laws). The set of all law rules written in all enactments is like the state DNA. It's constantly evolving. Thousands of spiders.
JW Trying to account for the evo if "state policy" is like trying to account for the evo of the biome of the African continent. Best start with the evo of a tilapia species, or the holes in violins
RM But if you're interested in the explanation of sociotechnical systems evolution now, then it's natural that you'll start with a theory which uses coarse-grained concepts and leaves places for future explanations of lower-level mechanisms ("black boxes").
JW A blogpost Evolution as a black box The blogpost summarizes the rest of the discussion between JW and MW, no need to reproduce it here
IN excellent! Great and very interesting discussion, thanks for the summary. I'm writing up a reaction...stay tuned :)
So, here is the promised reaction
I am heavily influenced by Steven Pinkers paper The cognitive niche: Coevolution of intelligence, sociality, and language The auto-catalytic process of information moving through media physical reality / technology -> human thought & language -> physical reality / technology makes a lot of sense to me.
Natural / artificial is I believe a artificial split, as technology is something that human "naturally do". So
Coming from Complex Adaptive System perspective, there is a continuum of embedded emergent processes. By emergent I mean many entities interacting with each other over time creating/growing novel system level patterns. An emergent pattern can not be caused by / is not ontained in properties of a single entity alone. As Kevin Kelly likes to say, "More is different" This pattern can itself be considered an entity, that interacts with its constituent elements, creating reflexive downward causation or other with other patterns creating, over time, yet more emergent patterns.
Sufficient interactions of "quantum physics" emerges into atoms. Sufficient interaction of atoms emerges into "chemistry". Sufficient interaction of chemistry self organizes and emerges biochemistry / biology. Biochemistry is the first moment we have sufficiently complex structures and mechanisms to allow information to be stored, varied and replicated. Evolution is fundamentally an informational process, as Dennet argues in "Darwins Dangerous idea". Lots of biology happened to evolve minds, which collectively emerged culture. Culture (through social learning etc) emerges technology, as a way to increase its own fitness, by modifying its own physical and social environment, being its own self-directed, purposeful and normative fitness landscape. Of course, this coupled fitness landscape changes continuously.
A auto-catalytic process start when culture produces technology that allows us to store, replicate and vary social, technical and biological information. Technology here is very broadly conceived, being physical, such as making fire or a CPU, or "social" technology such as language or flood insurance.
Technology allows us to continually increase the efficiency of mass /energy flows from/to nature to and within technosphere, and information flows within society, creating more "culture" by growing more population(biomass), culture(ehm... memespehre?) and civilisation (technomass), which in turn increase the speed / efficiency and so on. This, I would argue, is why so many technologist are so obsessed with the notion of Singularity, as it is a logical consequence of this thought. I also do believe that they suffer from techno-fix, and forget pesky notions like second law of thermodynamics and planetary boundaries, which are becoming very clear in the form of global warming and the impending resource crisis.
Throghts on The evolution of an ancient technology paper
After this rant, I went and read the excellent The evolution of an ancient technology paper. Hoping that it will move the discussion further, here are some of my notes and thoughts on how it relates to the discussion above.
"Considered in the most general sense, technology consists of knowledge about how to modify our environment, passed from one generation to the next." Nice, comment to my heart as an engineer. We also like to thik of engineering as "Action under uncertainty, while solving areal problem" That problem is usually a modification of the physical world.
"All six of the studies that we drew upon for our analysis report the same basic pattern of transmission: weaving-related practices are passed inter-generationally (‘vertically’), primarily from mother to daughter, and to a lesser extent between other female weavers within the community." This to me seems like a nice example of field bias, from anthropologists/archaeologists perspective. Nothing wrong with it, but it does mean that this can be studied this way in "primitive" societies with "primitive" technologies, that can be understood/conceptualised within a single person. My (and probably authors) assumption here is that a weaver also knows how to make a loom. Is that so?
Also, there seems to be a mix of both technology and its design (i.e. weaving as a capability, and associated loom design) vs products that are made by it, changes is designs, patterns, colors. Of course some patterns are only possible on some looms, and not all materials / pigments are available everywhere.
What happens when nobody knows or even can know, how to build the technology alone? E.g. the (I pencil)[https://en.wikipedia.org/wiki/I,_Pencil], or The Toaster Project When technology under study is manufactured by the global supply chains what would be horizontal and what would be vertical transfer? I see no reason why it should not be there, but how do we talk about it?
"The phylogeny of looms that we have deduced here does not support the notion that technological change has an inherent ‘direction’. Technological progress (increased complexity) is not inevitable. "
I would think that this is just an artefact of the limited scope ( mainly in time) as they consider societies in relatively stable periods, The speed of cultural and technological change in "old" societies was much more slower than today, so that the tech did not have to change to support the rest / much more limited co-evolution. Today is very different, internet is 30~40 year old, depending on how you count)
"Some of the issues related to the transmission of cultural information were discussed by Pocklington and Best [40, p. 81], who argued that the ‘units of selection’ of culture are ‘the largest units of socially transmitted information that reliably and repeatedly withstand transmission’ and that these must be ascertained empirically rather than deduced a priori, a conclusion that our work supports."
Very interesting, as authors note, when we do not have noise channels, or at least channels with low noise, we can copy blueprints, CAD files, operational parameters perfectly, and in theory anyone could make a naptha cracker. However, the resources needed are staggering,what is takes a corporation, an "AI", to do successfully.
Also, they nice make a good point about simplification, caused by lack of resources! Note here the difference between functional and structural simplicity. In tech, many functionally simple devices require a highly complex system to operate (light switch, ordering a pizza with an app)
"The degree to which ‘content bias’ operates, and hence the degree to which innovations from outside the community are adopted, seems to depend on the type of information and its place and importance within the weaving tradition. As noted, weavers are reluctant to make changes in looms or technique, which seems to be linked to the complexity of this information and the difficulty of transmitting it intact. Errors are costly, and the rewards of invention are uncertain. By contrast, new types of commercial yarn and dyes, particularly for daily use clothing, can be adopted relatively easily, since they tend to simplify the weaving process. Their adoption may also be linked to a different mode of transmission (peer-to-peer) and to competitive pressures within the community, particularly between young, unmarried women, since clothing is highly visible and communicates status."
Biases, from Boyd and Richerson are an relevant notion to encode the normative /purposefulness of cultural evolution. Especially if we can identify the mechanisms that drive the evolution of biases themselves. This to me directly signifies the fucntional vs structural simplicity. Innovations that provide the same or better functional simplicity ( modern yarn, rather than a existing yarn) are picked up / selected for easily. Structural complexity is hard, new loom, or new yarn making process. So the more fundamental a technology is, the more difficult it is to innovate ? As it imacts so many things, or as it requires so many complex inputs?
Finally, authors present a interesting classification of parts of technology. 1. the skills and knowledge of weavers; 2. tools (looms and other equipment); 3. templates (heirloom textiles, pattern samplers, pattern storage devices used on some looms); and 4. domesticates (plants and animals, used as sources of fibres and dyes)._
Interesting. I would classify this slightly differently, and maybe more generically : 1. skill and knowledge needed to operate technology 2. skill and knowledge needed to design/build technology ( we don't need the templates) 3. tool - the physical assets /device 4. resources - both for producing the tech and its consumables
My question, related to sustainability transitions, CE etc would be, how various functions /needs are fulfilled, and why one options is selected above the other, when they seem equivalent, e.g. For example, a choice between a naphta fired naphta cracker or electricity powered naphta cracker. It is "of course" an economics, business decisions, but what is cheap, what is acceptable is a societal decision, and evolves over time. Of course, various alternatives must be invented, and available at scale, before they can be taken up in the transition. I suspect that the general mechanisms are the same between traditional/primitive tech, but the way it works today is (at least superficially) so different than the apprenticeship and rituals of weaving.
I very much welcome your comments and inputs, espeially if I am wrong or have missed existing literature! Please contact me via twitter or send me an email