Technically Lost
When and how were innovations lost to history?
Part of a series on the history of technological contingency & the future of AI.
In the far prehistory of 2011, Kevin Kelly made online waves when he claimed on NPR that ‘there is no species of technology that have ever gone globally extinct on this planet.’ In his What Technology Wants, he had expanded on this notion, writing that:
A close examination of a supposedly extinct bygone technology almost always shows that somewhere on the planet someone is still producing it. A technique or artifact may be rare in the modern urban world but quite common in the developing rural world. For instance, Burma is full of oxcart technology; basketry is ubiquitous in most of Africa; hand spinning is still thriving in Bolivia. A supposedly dead technology may be enthusiastically embraced by a heritage-based minority in modern society, if only for ritual satisfaction. Consider the traditional ways of the Amish, or modern tribal communities or fanatical vinyl record collectors. Often old technology is obsolete, that is, it is not very ubiquitous or is second rate, but it still may be in small-time use.
Like any challenge made to the internet, this sparked a determined search to prove him wrong. Yet, the resulting proposals for extinct technologies, such as 8-track players; Radium Suppositories, the Roman corvus (e.g. planks used to board enemy ships), or the idiosyncratic memory technologies of 1950s-era jukeboxes, all were found to still be extant in some way.
Indeed, as Étienne Fortier-Dubois (of Historical Tech Tree fame) more recently argued, there appear to be relatively few examples of technology-branches that once flourished, but which were lost outright and then never recovered or equated; many common candidates for such lost technologies are not actually lost; and confirmed lost technologies, he claims, are arguably quite marginal.
Is this true? Are once-achieved tech tree branches so resilient to loss? How do we square this with the perception of widespread cultural, artistic and knowledge loss in our fragmented historical record? Are tech tree branches more resilient to loss today than in the past? And what does it actually mean if no technology has been entirely lost, but many have fallen from use so long that they cannot be readily produced anew?
1. The Great Sieve
The idea that technology is almost persistently un-lose-able is contrary to the popular view of the exciting mystery and lost craft of vanished forerunner civilizations. More prosaically, the idea should be somewhat surprising on its face, given that we have, provably, lost a tremendous amount of work, craft and knowledge to history.1
War; civilizational collapse; disaster; decay; simple generational churn and disinterest. Call this the Great Sieve:2 a past-focused Great Filter that determines which of our ancestors’ practices, artefacts, achievements and knowledge make it out of history and into the onrushing present. The Sieve catches all those things that our ancestors were not able (or simply not motivated) to readily preserve. It determines which technological practices remain alive; which thrive in our collective memory; which manage to cling on in some archive awaiting a future raider of the lost tech; and which are forgotten.
What has the Sieve taken? Let’s momentarily put aside (1) losses of the natural world (i.e. of animal species or stocks or of local ecosystems that were once-critical to local civilizations and which, once-exhausted, could not be revived);3 (2) key materials or non-organic natural resources necessary for particular technologies (though see below); (3) the more obviously esoteric claims to lost knowledge (e.g. ‘Atlantean geoengineering, alchemical knowledge, Lemurian stone cutting, the Giza power plant...’). Even then, the (er) record of loss is extensive. In his Inaugural Regress Studies symposium, Santi Ruiz writes:
We’ve lost vast libraries of human knowledge, like Alexandria,4 of course, but also more recent cases: thousands of texts during the English dissolution of the monasteries, Leuven in 1914, the great Polish Krasinski and Zaluski troves in 1944. We know many of the texts we’ve lost: Aristotle’s dialogues, Pindar and Sappho’s poetry, Pliny’s Naturalis Historia, works of Sulla, Cicero, Caesar, the vast majority of Maya and Aztec literature, Incan quipus, an epistle to Corinth, the Yongle Encyclopedia, Shakespeare’s Cardenio. Two of Bach’s Passions, George and Martha Washington’s letters, Carlyle’s first draft of The French Revolution, all of Hemingway’s work pre-1922, Walter Benjamin’s final manuscript. And there are others we’ve lost which we can’t know, of course: countless oral traditions, for instance, including works of Homer [...] Then there are losses of craft, of techne, and of ability: Dhaka muslin, or bardic chant, or poetic navigation. Greek fire, whatever the antikythera mechanism slotted into...
The Sieve has most obviously claimed numerous non-material cultural products or crafts,5 including many oral traditions, as well as gestural languages and traditions; but it has also erased masses of material, documentary artefacts, including perhaps 98-99% of all ancient Greco-Roman literature produced in the millennium after the 7th century BCE; the vast majority of everyday letters and correspondence throughout history,6 and virtually all child art.7
This means that, in practice, the Sieve has erased not just the seminal works of some of the greatest scholars and artists of history, but also the prosaic records of most humans’ identities and lives.8 In some cases, this means we have lost functional access to basic historical facts: for instance, when did the famous Hammurabi actually reign? We can cross-reference shattered Babylonian archives against astronomical records of eclipses, as well as carbon-dated wood artefacts, to get estimates. But we ultimately don’t know, and possibly never will, as the information has been lost, and not enough adjacent or indirect reflective records exist to triangulate or reconstruct it.9
In the limit, deep history ends up grinding down most informational and technological signals within a few hundred thousand to million years. For instance, the ‘Silurian Hypothesis‘ is a theoretical idea amidst astrobiologists, which asks: if an (industrial) civilization existed on Earth millions of years ago, would we be able to know? Given that the likely geological fingerprint of our Anthropocene is not necessarily greatly distinct from that of many other known events in the geological record, we likely would not be able to detect such societies.
However, to what extent does the Sieve catch technologies? Or are branches on the tech tree always or reliably more resilient?
2. Techne Lost?
I think it is worth revisiting when, and how, we actually have ever lost technologies--and whether our susceptibility to lose them may have changed over time. So it’s worth revisiting many of the old and new examples of lost (practical) technologies, to understand which are serious or genuine cases of loss, and overall where and how loss works on the tech tree.
Some clarification before we start (feel free to scroll down to the cases):
Types of technology loss: there are different ways in which we might speak of ‘losing’ technology. Any given technology can be subject to any and all types of loss at once.
Artefactual loss: Loss of all (original) instances of the technology
Epistemic loss: Lost awareness that a technology once existed; or uncertainty over whether it did10
Understanding or design loss: Lost recipe or technique for a technology in question
Semantic loss: Lost understanding of function (or meaning) of remaining artefacts (or, in the special case of information storage mediums or scripts, the loss of decoding ciphers for the stored information)
Production base loss: e.g. lost ability to manufacture a (known and understood) technology, whether because of resource loss (critical enabling resource exhausted), tacit knowledge loss, or mothballing of industrial capacity
Degrees of loss: as a result, the degree or way in which we have ‘lost’ a technology has various degrees or depth of loss (across the above forms). For instance, going from (roughly) lesser to greater types of loss:
(-1) memory mirages: technologies we never lost, because we never had them (e.g. Archimedes’ solar heat ray; Starlite);
(0) memetically fragile tech: technology in production and use, but where the production supply chains (or the underlying tacit knowledge) could be easily shattered and then hard to reconstitute;
(1) obsolete tech: technologies that are mostly superseded by newer versions or alternatives; but are still occasionally used, produced, and understood;
(2) ‘senescent‘ tech: no longer actively produced, but still occasionally used and workings still understood, and production of new examples could be trivially restarted;
(3) ‘withered’ tech: no longer in use or production; workings are theoretically understood but tacit knowledge lost; resuming production would take lengthy rebuilding of human or industrial capital;
(4) prodigal tech: technologies we once (locally or globally) lost, but which societies later rediscovered (e.g. Roman concrete; Mayan rubber);
(5) obsolete enigmas: lost technologies whose precise workings we still don’t understand, but which have been since equated (e.g. ‘Greek Fire’; Damascus Steel)
(6) Exhausted technology-enabling resources: technologies that are functionally hard to (re)produce because some basic material resource has been exhausted;
(7) genuine losses: technologies whose workings have been lost, and which we do not know how to produce (e.g. Segato’s techniques for cadaver petrification);
(8) ‘dark matter tech losses’: technologies we don’t remember were once extant (???)
I’ll discuss some of these below, in a different order.
Memory mirages: technologies that never really were
Right off the bat, there are some often-discussed examples of ‘lost’ technologies that we should put aside, because they are:
Semi-mythical, and their existence or function is likely misremembered, or based on mis-interpretation of genuine accounts: E.g. Roman ‘flexible glass‘ (vitrum flexile) that would dent rather than shatter when struck (very likely a myth); accounts suggesting that the Roman engineer Hero invented motion pictures and/or that the Romans had telescopes (most likely based on mis-reading of genuine evidence); Archimedes’ ‘solar heat ray (more likely a steam cannon) and (debatably) the ‘Iron Hand/Claw of Archimedes‘ (alleged anti-ship grappling crane, of uncertain historicity); the Mithridate ‘universal antidote’; or the ‘Philadelphia Experiment’ (alleged 1943 Navy experiment in rendering a destroyer invisible and teleporting it).
Existed in some form, but were likely fraudulent to begin with: e.g. the Sloot Digital Coding System; the Starlite extreme-heat-resistant material (unclear if its properties we really as good as promised); (arguably) Stradivarius violins (in terms of their claimed unique quality, as attributed to lost secrets of craft)
genuinely lost, but unclear what was their exact functioning (e.g. what is the actual thing that was lost):11 e.g. the Baghdad battery (hypothesized galvanic cell, dating to the Parthian or Sasanian empires; most likely just a scroll vessel);12 the Panjagan archery weapon/technique for firing five arrows in short succession (supposedly used by Sassanian horse archers; though it is unclear what this was exactly); Roman dodecahedra (small copper alloy hollow objects of unclear role)
(arguably) rare artefacts that appear to be one-offs cases rather than evidence of a consistent, repeatable and widely practiced technique: e.g. the iron pillar of Delhi which has a higly rust-resistant composition; the Lycurgus Cup, a 4th-century Roman cup made from dichroic glass).
Senescent tech: technologies that were never really lost, just retired from widespread use
In addition, there are other technologies that don’t really qualify as ‘lost’ technologies, since they are:
not really forgotten, just no longer actively produced, (since the need or market they served has evaporated): e.g. Egyptian mummy embalming tools; ‘grave torpedoes‘ (1880s coffin explosives to deter grave-robbers); the ‘automatic (bowler) hat tipper’; etc.
were not, in fact, entirely forgotten or lost, but merely saw their use decline:
Roman hypocaust heat storage systems (preserved even today in e.g. Castillian Gloria systems); many ‘ancient’ Windows computer operating systems, especially those used in older infrastructures or large specialized hardware; etc.
could plausibly be easily reconstructed with today’s knowledge, even if no one has seen reason to:13 Roman artillery weapons can be reconstructed, even if their actual operating procedures remain mostly lost to us; the Helepolis super-large siege tower; gargantuan wooden ‘stupa‘ towers of 2nd century Pakistan;
have been reconstructed in modern times, e.g. out of curiosity: e.g. Athenian triremes; traction trebuchets; WW-II-era ‘Upkeep’ bouncing bombs; many other products of experimental archeology, including entire medieval-style castles.
Prodigal Technologies: once lost but then found
Putting these aside; there are however a wide range of technologies that were genuinely lost after their invention, often locally, and sometimes globally. There is in fact a long history of technological forgetfulness, even for remarkably influential and useful technologies. As Beatrice Heuser discusses in one recent survey of historical military technology:
Until the late 18th century, technical inventions were repeatedly forgotten and not rediscovered until centuries later, or applied for a long time only in a limited area or among some peoples, leaving other parts of the world unaffected. Instances of the reinvention of older technology include the gear, which existed in antiquity and would become so crucial for the harnessing of wind and water power to so many human endeavours, from grinding corn and draining wetlands to generating electricity.2 Other instances also include weapons systems, such as the trebuchet, that could project anything from stones to fire brands onto enemy forces, and even into enemy fortifications or walled towns and cities. It was used by the Romans but seems to have been forgotten during the West European Dark Ages.14 The crossbow seems to have been reinvented independently in different parts of the world and periodically forgotten again: while it existed in Ancient Greece for example, it seems to have been considered a new ‘barbarian’ innovation by the Byzantines in the 12th century.3 Breech-loading rifles were invented on separate occasions in the 16th and 18th centuries,4 before being produced on an industrial scale and becoming widely used in the 19th century, when they were adopted by the Prussian Army and first used effectively in the Austro-Prussian War of 1866, in the form of the Dreyse needle-gun. Hot-air balloons were used for military reconnaissance purposes by the French at the very end of the 18th century, then not again until the mid-19th century, and only came into their own in new shapes in the First World War. Submarines were first used in 1776, then reinvented independently in 1797 and 1879, while percussion-fused hand grenades were first invented in 1861 before being reinvented in 1905;5 and the list goes on
Premodern prodigal technologies
We can add some further cases to this list of technologies that were once (locally) forgotten, but were eventually recovered (or independently reinvented):
Ancestral Polynesians lost the craft of Waka hourua oceangoing canoes; as well as (in some contexts) various other practical technologies (e.g. canoes, pottery, the bow and arrow, in some cases fire),
the Indus Valley Civilization utilized and then abandoned large-scale urban sanitation, a technology not developed or rediscovered until the Romans 2,500 years later;
the ancient Greeks may have come at least somewhat close to working theories of gravity: although Aristotle’s theory of gravity (relying on the idea that objects fall at rates proportional to their weight, driven by an innate ‘Lightness’ or ‘Heaviness’ seeking its ‘natural place’) was already empirically refuted in antiquity, a rival atomist tradition got substantially closer, but was lost to the indifference of Christian scribes.
there is evidence that indigenous Australians lost bows and arrows15
The Roman Republic was capable of constructing very large polyreme warships, of a type and size unequalled by the later Roman empire which by 200AD focused more on ‘liburnian’ patrol vessels (debatable, since this may be more reflective of changes in their strategic position)
the medieval Middle East saw the loss of wheel and oxcart technologies between the 5th and 19th centuries; a severe decline in the use of labor-saving capital goods (water-mills, windmills, cranes); as well as a near-complete loss of a technically sophisticated 9th-14th century block tin-plate casting printing industry, so completely forgotten that its existence was only recently rediscovered16
Zhang’s seismoscope (the “earthquake weathervane”) was repeatedly lost and reinvented or reconstructed in various forms
early Mesoamerican societies developed a range of technologies, such as workeable and flexible rubber, or Aztec cultivation of spirulina, which were subsequently lost for centuries until they were rediscovered
the once-valuable Dhaka muslin fabric; i.e. the last decade has seen projects that have aimed to restore or approach the original, after the rediscovery of the rare phuti karpa cotton plant
(arguably) the highly stable Maya Blue pigment; similarly, the Tyrian purple dye
Damascus - or (Wootz) steel: although the chemical composition (in terms of carbon content and carbide network) is today reasonably well understood, this steel still cannot be reliably reproduced, in part because the original ore supplies (containing traces of vanadium, molybdenum, chromium) were exhausted;
Vitamin-C based cures to scurvy, discovered in 1747; institutionalized in the Royal Navy by 1799; but then somehow institutionally forgotten by the Royal Navy during the mid-1800s (especially after it replaced lemon juice with an ineffective substitute in 1860), before eventually rediscovered in the early 1900s.
I think this shows that there is a fair amount of reinvention happening, though loss can often persist for centuries.
Modern prodigal technologies
But surely we are doing better today, in our large complex civilization with its mass data recording technologies? Well--maybe? Mostly? However, interestingly, in some cases, even high technology is not immune to being forgotten, even if temporarily. For instance:
FOGBANK was an extremely classified material used in nuclear warheads, the synthesis method for which was forgotten during the late 1980s such that it had to be subsequently re-engineered during the early 2000s (see also Wellerstein on ‘Forgetting the bomb‘ ).
the NASA F1 moon rocket engine used in the Saturn V rocket had to be almost entirely re-engineered in 2013, even though all the original design documents and blueprints remained on file. The problem was that the original design process that produced the F-1 engine was iterative and experimental; some aspects of the engine’s performance were not well understood even by its original designers; and many of the components or assembly tools were extremely bespoke or even unique, such that they functionally had to be redesigned,17 since many ‘off-the-shelf’ components wouldn’t work.18
some trade secrets have been kept so closely they were accidentally lost; e.g. McDonald’s for some years lost track of its secret ‘special sauce‘ recipe in 1991, when it tweaked the formula, with the result that it had to work in-depth with one of its suppliers to reverse-engineer the sauce in 2004 (the recipe has since been released);
when Polaroid stopped production in 2008, their precise instant film reagent chemistry was functionally lost, though it was re-engineered by the ‘Impossible Project’.
These suggest that temporary loss of a complex, high technology can frequently occur, especially in conditions where (1) a technology is highly classified or secret; (2) the technology’s original design was dependent on iterative, theory-free engineering, such that the engineers may not have an explicit understanding of all its parts and their contribution to the final design; (3) the technology is heavily dependent on tacit knowledge or on many bespoke production components that themselves have been discontinued, and/or (4) the technology is abandoned or phased out for some time. In such cases, a technology either needs to be reinvented, or the path is left withered on the vine. What strikes me is the relative speed at which tacit knowledge or industrial base can atrophy once production lines close, which in some cases appears to be barely more than a decade.
Obsolete Enigmas: lost but since equated or surpassed
In addition, there are some interesting cases of technologies that are genuinely lost, in that the precise processes or techniques remain lost to us; but where the inventions (or the functions performed by them) have since been approximated, equated or exceeded through other techniques. For instance:
(arguably) Roman waterproof concrete (although there is recent work that seems to have cracked this)
Roman techniques for producing dichroic glass (e.g. the Lycurgus cup, above);
the Polybolos repeating ballista (precise workings unknown)
the Byzantine Empire’s ‘Greek Fire‘ weapon (lost because of secrecy, with ongoing debate over its composition);
extremely precise Egyptian and Incan stonecutting techniques
(arguably) Roman wines.
Exhausted technology-enabling resources
Interestingly, there appear to not be many (and possibly not any) truly lost materials or natural resources which have been globally mined or used to exhaustion, especially in ways that we have not been able to substitute for. The only (prospective) candidates I’ve found at a skim:
Malachite, a crystalline copper ore, may have been exhaustively utilized;
Low-background steel (i.e. pre-atomic steel, produced prior to the detonation of the first atomic bombs) has specific applications as the shielding material for particle detectors and whole-body counting, but is intrinsically limited in supply, with sunk fleets of pre-1950s battleships one of the main repositories. Idem for ‘ancient lead’ (i.e. lead that has lower radiation levels than freshly refined lead because it has had time to decay), which is sourced from Roman shipwrecks.
Silphium, a prized Cyrene / Roman plant with diverse applications; lost for millennia, though there are some recent claims that Ferula drudeana is a surviving form of the plant.
helium may one day qualify, as a nonrenewable resource that has seen recurring global shortages of helium since 2006 (most recently in 2022); at the current rate of use it might be exhausted within a century.
In addition, there are suggestions that humanity may before long run out of various metals such as cryolite, phosphorus, or “rare earths” (e.g. europium, indium, rhenium, tellurium, terbium, dysprosium, and neodymium) (see also Belfield 72).
the phenomenon of disappearing ‘polymorphs‘ provides a strange, special case; as it involves metastable materials with a particular crystal structure being transformed into a different crystal structure as a result of contamination with a microscopic seed crystal. This phenomenon has led to the functional loss of some drugs: for instance, the HIV/AIDS medication Ritonavir became impossible to be produced in its original form I, after all production facilities were contaminated with form II. However, in practice, researchers were able to compensate for these changes; in this case, they managed to recover the effects downstream, by replacing the capsule with a refrigerated gelcap for drug delivery.
There is also the possibility that certain critical resources, such as easily accessible fossil fuels, would no longer be functionally accessible to a lower-tech industrial economy recovering after a global collapse, with the result that key energetic rungs of the technological and industrial ladder would have been lost to us even if the resources themselves strictly remain.
Memetically fragile technologies: technology-enabling data resources
Or, for a more contentious case of loss, consider modern losses of historical data or datasets:
in the 1980s, NASA lost the original, high-quality tape recordings of the Apollo 11 moon landings; (likely) because they were taped over and reused for its Landsat program during a shortage of data tapes.
millions of early 2000s photos lost during the transition from film to digital photography, in a brief window where we lacked reliable storage for the new file format; or where file storage turned out to be dependent on single servers (e.g. the mass internet data lost in MySpace’s 2019 server crash, which deleted over 50 million files);
a decade of early Arabic-language internet forum material lost when Yahoo acquired the Arab web portal Maktoob in 2009, and migrated its data badly to a new portal, functionally losing nearly all archival record of the first (pre-Arab-Spring, pre-internet repression) wave of young Arabs online activity.
generally, internet ‘link rot‘ is a well-known and growing problem; one that affects perhaps two-thirds of links to websites over the last decade or so (including e.g. information referenced in US Supreme Court cases and academic materials). In many fields it’s prominence is growing over time; and it takes very active and deliberate effort to avert;
(at least) 20% of all existing 18.5 million Bitcoin, lost or otherwise stranded because their owners have lost access to wallets or private keys
There is also a distinct class of ‘important data we forgot to collect, and now cannot recover’. For instance, even though Google Search has been perhaps the core epistemic infrastructure shaping belief formation for most of the world for the past quarter-century, social scientists failed to collect detailed and longitudinal empirical data on the relative quality and composition of Google Search queries over time, and now will no longer be able to do so.19
Genuinely lost technologies
So what’s left? What’s genuinely lost and unrecovered and (arguably) unequalled since?
the Electronium, an early combined electronic synthesizer and algorithmic generative music machine, implemented as an analog electronic machine;
This is... a surprisingly short list. That appears to suggest that even many seemingly ‘obsolete’. technologies do still stick around, even after they are nominally superseded; that once an invention has been made and memory of it is retained, it is easier for others to create it even if they lack direct designs, or to reinvent it later.20
In my next post, I’ll add some reflections to some of the patterns we find here, such as: why do we (seem) to lose art but not technology? Is the modern world more impervious to technology loss, or less? And when does technological persistence actually matter?
Of course, perhaps not all of these are entirely beyond our reach, especially given trends in the use of AI systems for the rediscovery and restoration of inscriptions or the interpretation of documents.
Or, if you prefer a fictional videogame reference, The Pale. (Joyce Messier: “Others argue that pale somehow consists of past information, that’s degrading. That it’s rarefied past, not rarefied matter”; see also ‘Introductory Entroponetics’).
Santi Ruiz: “Losses of the natural world come readily to mind for Americans: bison on the Great Plains, or beavers great and small, or the passenger pigeon, the mammoth, or perhaps the dodo bird. In 1497, John Cabot's crew reported that the Newfoundland sea "is full of fish that can be taken not only with nets but with fishing baskets." Oysters were the size of dinner plates in the Chesapeake Bay, said the first European explorers. Salmon in the Pacific Northwest was so plentiful that even during the Great Depression it sold for the contemporary equivalent of a dollar per pound: poor man’s food.”
‘Can we estimate how many megabytes were contained in the lost library of Alexandria?’ is one of those quirkily misconceived-but-still-revealing questions fit for an at-least-impressionistic AI answer. Claude back-of-the-envelopes it as (40,000-700,000 scrolls lost) x (10,000-15,000 words per scroll, at 6 bytes/word = 60-90kb per scroll) = 2,4GB - 64GB: likely it is much lower since it is generally believed that ancient writers may have inflated the scroll count significantly; and because much of the collection consisted of duplicates. Of course, in a cultural perspective, its loss is much more about the irreplaceability of those scrolls. Though pour one out also for all the possible valuable scrolls left unwritten.
See also XKCD on ‘Spider Paleontology’.
i.e. ‘Darlings of Oblivion’.
Benjamin Breen: “If we assume this is a human universal, and I think we can, then it is a bit staggering to imagine just how much child artwork has ever been created. Here’s a back of the envelope calculation: there have been around 100 billion humans, and the average kid makes about 1-2 drawings (or other forms of artistic mark-making) each day for a period of, say, 10 years. That adds up to over 500 trillion individual artworks by children. Surely that’s not accurate, at all. But it’s likely roughly the right order of magnitude.” I’m a bit more skeptical that kids achieved this rate of ‘artistic mark-making’ per se, but even so find it hard to knock the rate down by more than 2-3 OOMs.
Is this a bad thing? Beyond a cultural-heritage perspective, some ethical views would see the loss of information around individual lives as a tremendous ethical disaster in its own right: in antiquity, the Romans would fear the effects, on your shade in the afterlife, of your life passing from living memory. But in our more enlightened times we know that this is nonsense—the loss of 99.999% of all life-records instead is obviously an information-theoretic moral catastrophe, from a (neo)-Fyodorovist perspective. Consider, after all, Nikolai Fyodorovich Fyodorov‘s ‘Common Task‘--the future resurrection of the dead through technological means, on the basis of their recorded data traces and personal archives. As recently discussed by Philosophy Bear, “Our best theories suggest that information is never really destroyed, however they give basically no hope for local and practically feasible reconstruction of the past. ... reconstructing the soul of an anonymous peasant who died in 1750 let alone a Paleolithic hunter gatherer circa 30,000 years ago with any fidelity is probably impossible. In the case of many 30,000-year-old hunter gatherers we may never even be able to lock in a target: an approximate sense that a particular person existed and we’re trying to replicate them. ... The problem is that while the information technically persists, it is not in a practically recoverable form, and in many cases is not locally available because it has dispersed at the speed of light. ... There are caveats on this- game changers we can’t yet rule out. One is the discovery of some unknown form of trace- new physics”. To such a view, the eternal unreachability of the past would stand as a (perhaps the) moral wound in the world.
Signore Galilei : “One thing that fascinates me about archaeology compared to other sciences is that there’s a finite amount of data out there, and the questions you have may literally be unanswerable. We do still find new major sites every now and again, as well as new techniques for parsing our existing data. Maybe somewhere buried in the middle east, there’s the smoking gun to confirm exactly when Hammurabi ruled his empire. But unless and until we find it, everyone – historians, archaeologists, astronomers, and more – will keep working to solve what’s become quite a puzzle.”
E.g. Lost Futures discusses a quote from the 13th century medieval monk Roger Bacon in his Epistola de Secretis Operibus, that seems to confidently assert the contemporary existence of self-rowing galleys, animal-less wagons, or submarines. The historicity is questionable, and likely he was relying on flawed sources, describing toy automata, or just using stylistic convention. But we don’t quite know.
See also Étienne’s discussion of ‘mysterious artefacts‘.
Ironically doubly lost today, since ‘The current whereabouts of the artifact are unknown, since the 2003 US-led invasion of Iraq.’
I, for one, see reason to, and I would love to see these created. Can someone set up a Make Siege Towers Great Again fund?
ED: this seems slightly wrong: both the Chinese and the Byzantines had the traction trebuchet; the European Dark Ages discovered the counterweight trebuchet. However, Roman torsion artillery may have been lost for a time.
From Luke Kemp’s Goliath’s Curse: ‘we need to bear in mind that technological development is not a straight, rising line across history. Nor is the loss of innovation innately tied to collapse. Even the most basic tools have sometimes been forgotten. When Europeans arrived in Australia the indigenous people didn’t have a single bow or arrow, despite both having been invented up to 70,000 years ago in Africa. Instead, they had invented and relied on the boomerang rather than the bow (alongside spears and throwing clubs). Somehow during the migration from Africa to Australia the technology had been lost.’
See Bulliet’s chapter on ‘Determinism and Pre-Industrial Technology’.
(WIRED): “Why was Nasa working with ancient engines instead of building a new F-1 or a full Saturn V? One urban legend holds that key “plans” or “blueprints” were disposed of long ago through carelessness or bureaucratic oversight. Nothing could be further from the truth; every scrap of documentation produced during Project Apollo, including the design documents for the Saturn V and the F-1 engines, remains on file. If re-creating the F-1 engine were simply a matter of cribbing from some 1960s blueprints, Nasa would have already done so. A typical design document for something like the F-1, though, was produced under intense deadline pressure and lacked even the barest forms of computerized design aids. Such a document simply cannot tell the entire story of the hardware. Each F-1 engine was uniquely built by hand, and each has its own undocumented quirks. In addition, the design process used in the 1960s was necessarily iterative: engineers would design a component, fabricate it, test it, and see how it performed. Then they would modify the design, build the new version, and test it again. This would continue until the design was “good enough.” Further, although the principles behind the F-1 are well known, some aspects of its operation simply weren’t fully understood at the time. The thrust instability problem is a perfect example. [...] The exterior scan was therefore used to develop the specialised tooling needed to fit the F-1’s nuts, bolts, and fasteners. Some of the bolts were annoyingly unique - Betts noted that at least one high-torque bolt in the turbopump assembly required its own special torque adapter to remove.’
As aerospace historian Scott Lowther writes: “If you try to rebuild the Saturn V based on a complete and pristine set of fifty-year-old blueprints, one of many problems you’ll discover is that a lot of the off-the-shelf stuff meant to go in it… doesn’t exist anymore.”
Though, as Gavin notes, the second-best time is now.
For instance, it is generally held that the Soviet thermonuclear bomb project got a lot of (political, but especially scientific) ‘stimulus’ from the US’s 1952 Ivy MIKE test, as this showed that it was possible to work in the first place.