Past Tech Tree Divergence was not (just) Environmental or Extraneous
Possible objections to premodern tech variety #1
Part of a series on the Historical Mechanics of Technological Contingency (and the future of AI).
In my previous post I presented an obsessively long survey of cases of technological divergence and variety amongst premodern societies. But what does this tell us?
It’s worth considering a range of counter-arguments to apparent cases of technological divergence in premodern societies. After all, maybe apparent differences between the technologies used by different societies shouldn’t really count as evidence for understanding the degree of latitude or contingency in tech development. This could be because these cases…
…only occurred in ‘marginal’ (or ‘frivolous’) technologies, not critical, core ones;
…are fully explained by variation in the available material resources
…are fully explained by the variation in the environment (and the needs it creates);
…were (or would have been) temporary dead-end detours before those civilizations (assuming they would have avoided endogenous collapse) would have hit upon the ‘right’ solution by themselves;
Let’s consider some of these, in turn.
I. ‘Any variation occurs only in frivolous technologies, not key ones’
Maybe. However, the survey of technological divergence I’ve offered identifies significant technological variation in the very categories of technology–agriculture, urban design, mathematics, textiles, currency, architecture, bureaucracy, metallurgy, the wheel, writing–on which BBE base their argument for independent technological convergence, along with many further categories besides.
More to the point, I don’t think we have an easy, robust or principled way to rank technologies as critical or irrelevant, that isn’t subject to strong hindsight bias. Our judgment of how important technologies are will be colored by our judgment of how important they seem today, or how important they appear in retrospect, given the later key technologies that were derived from, or inspired by them (consider: limited Greek experiments with early steam engines, which remained very marginal in usage at the time, but may loom large to us).
In other cases, when we try to judge the relatively importance of a technology by considering whether that technology appeared to be marginal to, or key to, the operation of those societies that we already judge as high civilizations, leads us to understate the degree to which a technology might rank as ‘critical’ for one society but not another. For instance:
oceangoing canoes may have been relatively irrelevant to (and indeed largely absent from) the Roman Empire’s naval infrastructure, yet they were a critical technological mainstay to some (but not all) Polynesian societies
in contrast, the Roman Empire strongly depended on centralized taxation and bureaucracy;Creating an system of resource extraction that, some argue, eventually became so complex and burdensome that it yielded declining marginal returns (requiring further bureaucracy to enforce compliance) and resulting in continuously increasing costs, societal fragility, and eventually, coupled with over-extension, a collapse. but these were absent from other long-lasting societies such as the Free Icelandic Commonwealth; and they existed in the complex, expansive and large-scale Indus Valley Civilization, but appear to have taken far more equitable forms in that remarkably egalitarian society
Aztec (social and artefactual) technologies and infrastructures related to mass ritual human sacrifice didn’t have clear parallels in other societiesEven though, as per a Girardian reading of history, many other societies had a social need for channelled collective violence to keep uncontrolled violence at bay; though these took different, and much less bloody forms in Greek, Chinese and Jewish civilizations. –and in fact the practice was so distinct it became a severe source of conflict with the Spanish.Ruiz: ‘[Cortes] and his men were also genuinely motivated to end the barbarism of human sacrifice and cannibalism. In one tense political negotiation, Cortes killed a man who had offered him a bowl of human blood.’ Sam Kris: “in November 8, 1519, the first Europeans to walk into the Aztec capital of Tenochtitlán saw something none of them had seen before. “All these buildings,” the conquistador Bernal Díaz del Castillo later recalled, resembled “fairy castles . . . so high, majestic, and splendid.” The city was a garden filled with flowers. In the evenings, the air smelled of perfumed resin. The squares were “kept so very clean that there was not the smallest particle of dust or straw to be seen anywhere.” And then Díaz del Castillo climbed the one hundred and fourteen steps of the great temple and saw the “altar encrusted with blood grown black,” and the bowls in which human flesh was cooked, and the endless piles of skulls.” Yet, while this means you could try and write the practice and technology off as a particularly violent but random cultural quirk, it is generally held that the practice played a significant and powerful, if terrible, legitimizing political role for the Aztecs, one that was potentially critical to their structure and their domination and the preservation of their empire, every bit as much as the actual military weapons deployed.
II. ‘Any variation is just due to different available resources’
Perhaps any apparent tech divergence should not be counted as ‘real’ variance, since it was just entirely determined by differing available resources?
For instance, Barnett, Besiroglu and Erdil argue that
“Specifics varied with local circumstances: writing emerged on clay tablets in Mesopotamia, papyrus in Egypt, bamboo strips in China, and bark paper in Mesoamerica. Each civilization worked within their local constraints, utilizing the materials that were available to them. Yet each came to possess similar technologies when faced with similar problems.”
However, is it the case that, within the same environment, with the same material and preceding tech-tree resources, societies end up hitting upon essentially the same solutions?
When different material resources did condition technology
To be sure, the environment can be an important conditioning force to technological development. For instance, in antiquity, tin bronze (much superior to arsenical bronze) was geographically rare, and required sophisticated long-distance trade networks, deliberate alloying techniques, and guild-like metallurgical knowledge. Bronze was the high-end, incumbent technology backed by complex supply chains; by contrast, iron ore was widely abundant, but early iron was inferior to good tin bronze. It therefore took time for iron wielding peripheral peoples to overturn bronze incumbents.
However, even here, differences can occur: for instance, Sub-Saharan Africa did not progress through a Bronze to an Iron age, but rather saw the more-or-less simultaneous adoption of iron and bronze metallurgy (even though the region simultaneously had fewer, though not no, copper deposits).
When different material resources created options that not all took
In other cases, the differential availability of local resources merely created additional technological options that were adopted by some but not all: plentiful and controllable sources of nutrients (e.g. rivers with salmon runs) often created both political and technological opportunities for resident societies that could control those choke-points (Kemp).
When similar material resources still led to different tech
There are a range of cases of similar material resources resulting in different technologies:
Old-vs-New World metallurgy: as mentioned in my previous post, even though both European and Andean (and meso-American) cultures had developed metallurgy, and practiced this on identical raw materials (copper, tin, gold, silver), Andean and meso-American cultures applied the craft to the development of decorative and ritual objects, never replacing utilitarian stone tools. In the Old World, the same metals were developed for tools, weapons, and armor
Stoneworking arches: while a very wide range of societies have built with stones, only the Sumerians invented the ‘true’ stone arch (which from there spread through cultural contact), whereas other societies (Maya, Inca) focused on perfecting the ‘corbel arch‘.
While many societies worked with similar fibers, Andean societies such as the Inca empire continued to scale up the use of backstrap looms, while the Old World developed larger frame looms.
Many societies had access to optical-quality transparent spar crystals: e.g. Iceland spar, calcite, or alternate ones such as cordierite – have deposits across the world, and would have sufficed to enable polarization-based optical compasses; but only the Vikings seem to have developed this technique.
III. ‘Any variation is just due to the environmental differences’
Perhaps any seeming tech divergence should not be counted as ‘real’ variance, since it was just entirely determined and conditioned by differing environmental needs?
However, while similarities or differences in environment must surely explain some of the difference in technology (e.g. societies in hot environments explored more technologies aimed towards cooling and air conditioning, rather than heat-preservation), there are still nonetheless significant situations where even similar environments led to, or afforded (independent or co-located) variation in technologies. We can consider three cases:
Within-environment tech variety
For one, there is variation even amongst societies operating in the very same biomes, alongside one another, and in full or plausible awareness of one another’s technological practices. I won’t aim to be exhaustive, but consider:
The Calusa (mentioned in my last post) were in the same Florida biome as their farming neighbours, yet actively rejected it, showing a case of within-environment difference
Prehispanic societies in southwestern New Mexico differed significantly, with Chaco Canyon Ancestral Puebloans adopting large, monumental and socially stratified dwellings, whereas the Mimbres culture in the same region pursuing an ‘insular’ strategy of limited external contact and household-ritual sovereignty
Various reviews have found that even amongst nearby polities there can be very significant differences in precolonial political centralization and form of state; in their governmental organizational structure and its management of knowledge (e.g. between Athens, Sparta, and Macedon) or in childcare practice
Similar-environment tech variety: a ‘natural experiment of history’
In addition, we see enough technological (and institutional) variation even amongst societies operating in similar (or even the same) biomes–even when they derive from the same ancestral culture.
For instance, it has been argued that Polynesia offers a very strong “grand natural experiment of history”, since it saw [a single Ancestral Polynesian](a single ancestral population (Ancestral Polynesian Society, ~500 BC–500 AD) disperses across Pacific islands, many with broadly similar tropical/subtropical environments.) disperse across the Pacific islands, many with broadly similar tropical or subtropical environments. Its descendants differed widely in both political and institutional form, as well as technological practices, even amongst (largely or entirely) similar environments:
in the Marquesas Islands, there was extreme political fragmentation amongst tribal groups of 200-800 people; the Marquesas are archaeologically identified as the immediate homeland of Hawaii’s first settlers;
Hawaii developed an archaic state, with the divine kapu kingship, a formal warrior class, intensive fishponds supported by a pondfield irrigation system;
The Tongan state developed a long-distance yet integrated maritime polity that stretched across an archipelago, with a hereditary sacred kingship and an expansionist ‘empire’ that relied on the importation of exotic lithic tools, imported from far away, which served as a form of prestige political capital
Samoa, closely related to Tonga and in a near-identical environment, went the opposite direction, with power distributed across thousands of matai chiefs responsible for extended families, and without a paramount hereditary king;
Mangaia (small, resource-limited) collapsed ecologically and devolved into endemic militarism with rotating war-chief leadership
Rapa Nui (Easter Island) invested massively in monumental stone architecture; by contrast, Rapa Iti (settled by a common ancestral population at roughly the same time) gradually militarized and built fortified hilltop refuges instead, in a development that is believed to be partially related to declining resources
Environmental differences without clear causal link to technological divergence
Sometimes, societies pursuing divergent technological paths did in fact operate in different environments; and yet it seems unwarranted to point to those environmental differences as the sole and exhaustive cause–or even as at all explanatory–of the observed technological divergence.
After all, even if environmental differences have a large explanatory role in accounting for e.g. why paddy rice agriculture flourished in China (but not in the Peruvian Desert); why the main draft animal was the water buffalo in highly-population dense East Asia (but the horse in less population-dense regions); or why the tidal water mill saw take-up by peoples living on the coast or near river deltas (but not in Himalaya)… can you point to any plausible causal story for why the (moderate to large) environmental differences that existed would have played any role at all in:
the Greek rejection of the idea of zero, when it was developed by Indian mathematicians in an environmentally different (though arguably socially similar–i.e. literate, urban, agricultural) contexts
the Mesoamerican non-adoption of the wheel (the usual explanation–a lack of draft animals and mountainous terrain–offers only a partial explanation, given that–as shown by the Chinese wheelbarrow, wheels are useful for transport even in very local (e.g. village-level) contexts, and even without draft animals);
the Chinese retention of the xylography woodblock-printing (even after they were aware of the European movable-type press); or their adoption of the moldboard plow (when the Europeans lagged by perhaps a millennium); or their stagnation in firearms development in spite of an early lead;
Number base variation across societies in every kind of environment: the environmental objection has almost zero purchase
In all of these cases, there is no clear, plausible story for why the existing environmental differences (large or small) should have exerted a clear effect on the variation in technologies–whether in an enabling or restrictive direction.
IV. ‘Variation was extraneous; an ephemeral dead-end’
Perhaps any unusual tech track was realy ephemeral: just a temporary detour or track down an eventual dead-end branch, which would eventually have been superseded and displaced by the ‘optimal’ solution (for a given tech level)? Maybe all societies would have discovered the ‘right’ solutions–even if they had never come in contact with other societies that had found those solutions–if only they had been given more time to develop by themselves?
I mean, this is always possible but it is, in some sense, unfalsifiable or just very difficult to evaluate. At the very least, it seems that our situation is one where (1) there are many cases of societies pursuing different technological branches for centuries or even millennia on end, suggesting that they were at least minimally functional for a complex society, and (2) that there are many cases where, given their prevailing tech level and available raw materials, the different (eventually prevailing) path was one that could have been explored at any point throughout those centuries, but it wasn’t.(this relates to a rather large track record of surprisingly delayed technologies–the subject of a future post). It’s not clear how much further or more sophisticated or powerful these other technology paths could have become, if only one of these other societies had gotten to kick off its own version of an industrial revolution;
This is not that I think that there are no overarching determinative factors that shape, channel and constrain the development of technologies, or the course of societal development more broadly. As I’ll argue in future posts, inter-societal competition can play such a structuring role; as can large external environmental shocks or asymmetric disease transmission rates. This therefore doesn’t mean I think there are no good counter-arguments to cases of premodern technological pluralism.
However, many of these intuitions essentially trade on sloppy or difficult-to-defend analytical choices (e.g. over what technologies count as frivolous or crucial) that dismisses inconvenient evidence on incoherent grounds, in ways that stack the deck in favour of finding a great degree of technological convergence. Others (such as environmental pressures or material availability offer some a lot of explanatory force but–as just argued–still don’t account for the degree of variation we see; or simply discard the evidence of centuries of supposed ‘dead-end’ tech tree branches seeing fairly successful development and usage.
If we are going to develop an understanding of the dynamics of technological determinism and -contingency, we can and should do better.
Acknowledgements: thanks to Luke Kemp and Jacob Slosser for discussion.
Comment on Substack · Subscribe to Critical Maas · All essays