Knowledge arks – CONSTRAINT N°2: ACCESSIBILITY (7/9)

SUMMARY:

Hardware and software obsolescence can make it difficult for our distant descendants to read the information stored in a knowledge archive, especially if the technologies used are complex

A balance must be struck between storage capacity and ease of use, which can be achieved through various strategies: limiting the amount of information, using simpler technologies to complement or even replace mass storage media, actively maintaining these media, using media that facilitate reverse engineering, etc.

Languages regularly disappear or naturally evolve over time, which can pose a challenge for our descendants’ understanding of our knowledge

Creating new Rosetta Stones can reduce the risk that a knowledge ark will be indecipherable to its intended audience

We must not lose sight of the fact that all kinds of conventions may also disappear in the future (units of measurement, writing systems, reading directions, mathematical symbols, etc.)

The oral transmission of some knowledge can make it easier for our descendants to decipher our knowledge, while also benefiting learners in the very short term. An intangible knowledge ark can be a relevant complement to a tangible one, particularly for preserving knowledge deemed essential.

Knowledge passed down orally cannot be censored. On the other hand, it can be distorted (intentionally or unintentionally); therefore, oral transmission is much more meaningful when accompanied by instruction in critical thinking and open-mindedness.

No obsolescence of storage media is a prerequisite for accessibility

The other major challenge facing knowledge repositories is that of accessibility: we want to ensure that our distant descendants (or any other recipient) can access the knowledge we wish to pass on to them. This issue is all the more pressing given the immense amount of information that needs to be gathered. Indeed, if we are content to preserve only a selection of our knowledge, a relatively universal medium can be used, such as engraved tablets. But if the volume of knowledge to be preserved is large, as we have seen, only mass digitization can get the job done, which requires technologically complex storage media—electronic, nanotechnological, or otherwise. These are media that are highly susceptible to obsolescence, which our descendants may simply be unable to read if they do not (or no longer) have the appropriate hardware. This a major difficulty, that adds to other well identified challenges such as copyright (refusal to allow the digitization of a given work or article without financial compensation), the fragility of certain artifacts (some items are in such poor condition that the digitization process could destroy them), the cost of the equipment (which can be a particular problem in less economically developed countries, or in smaller institutions that do not maintain a huge collection), or the considerable time needed to digitize the contents of museums, archives, or libraries.

The biggest problem with digitization, is that everything that has ever been digitized—the web, the internet, private archives—is stored in reality on a multitude of servers, in computer memory banks. Unfortunately, computer storage technologies are constantly evolving: the first generations of physical media were cards and punched tapes. Then magnetic media became widespread (magnetic tapes, then hard drives and floppy disks). Subsequently, optical media became the norm (CDs, DVDs, Blu-Ray), sometimes in networks. Nowadays, media such as USB keys, SD cards and SSDs (as well as all kinds of variants) have become widely available. Added to this is the multitude of formats and programming languages appearing and disappearing on a regular basis. As a result, computer media must be constantly updated, either by converting the content from an obsolete medium to a newer one (if applicable) or by re-digitizing a given item. This last option may occur if previous technologies are no longer as common, if time has caused significant mechanical damage, or if considerably better technologies are emerging (better storage capacity, faster writing and reading speeds, better resolutions etc.). In addition to the various forms of hardware obsolescence, there is also software obsolescence. Sometimes, it is the application that allows access to information that is no longer compatible, the operating system may change, or programs may be completely redesigned.

Between 1984 and 1986, a project called the “BBC Domesday Project” was launched with the aim of digitizing all kinds of studies, maps, testimonials, videos, and statistics on the British population at the time. However, the medium chosen at the time to store all this information, the Laserdisc, is no longer in use, and no company produces machines capable of reading it. Furthermore, the code used at the time (in BCPL language) poses serious backward compatibility issues, so it is difficult to emulate the content of the project’s laserdiscs with current technologies, even when laserdisc players are available. As a result, new projects had to be created to convert the disc content into a more durable format and medium, sometimes involving reverse engineering, i.e., analyzing obsolete technology in order to “recreate” it long after production had ended and the engineers who designed it had passed away. A similar problem arose at NASA when it suddenly became impossible to read magnetic tape archives from certain past missions (Apollo missions, Viking probes, etc.) because the machines that could read these tapes had been scrapped, and that the programming languages of that era were mastered only by people who are now either deceased or retired. Once again, reverse engineering was the solution to the problem. These two examples illustrate the need to properly manage any large-scale digitization project, which means carefully planning ahead to ensure the long-term viability of the hardware and any software. Digital preservation therefore requires taking into account the evolution of physical media as well as that of the associated computer programs.

Among the various methods used to address this issue are (non-exhaustive list):

*Refreshing: this involves transferring the content from one medium to another of the same type.

*Migration: this involves transferring information from one system to another. For example, from one file format to another, or from one operating system to another.

*Replication: this simply involves creating as many copies as possible to improve the probability that one of them will last for the desired length of time.

*Emulation: this involves replicating the system surrounding the medium, in a way, to make the medium “believe” that the tool used for writing/reading is the same as the one used previously.

*Encapsulation: this consists of objects that contain information about how they can be decrypted, facilitating reverse engineering and backward compatibility.

One final persistent difficulty with all these techniques is identifying the blind spots of digital preservation, i.e., the difficulty of identifying items that have been forgotten or not adequately updated or rescanned. Digital preservation, which complements digitization, is therefore a long-term, ongoing process.

Our descendants will need the technical ability to read our information

We note that digital technologies evolve and disappear frequently, but so far, we have assumed implicitly that future generations will still be able to decode digital information, at least partially, using the methods of their time and continuously converting data from one format to another. However, we have overlooked an important scenario, that of the collapse of civilization (linked to industrial risks). In this case, technology would disappear, and future generations would be helpless in the face of our information media, as they would no longer have a command of computer science, signal processing, encryption, electronics, etc. To remedy this problem, a digital medium would need to be invented that is capable of storing large amounts of information and can be decoded relatively easily, even by a civilization that is not very technologically advanced. This is the major drawback of digital preservation: the media we have used to date are very specific and require very specific machines (DVD players, magnetic tape readers, projectors, etc.), which can limit the universality of digital mass storage. With molecular storage, this problem would be reduced, since in theory, any machine capable of detecting the structure of a molecule (DNA or other) would be able to read that molecule, whereas for a specific electronic medium, a specific reading machine is required. Furthermore, if computer bits are stored directly in the form of molecules or nucleotides, with one type of nucleotide associated with “1” and another with “0,” the program needed to read the molecule’s content may be less complex to recreate than one conceived for an electronic medium, in which “1”s and “0”s cannot be equated with simple elementary components.

Despite the advantages of DNA, or more broadly molecular storage, our descendants will still need to have a basic understanding of microscopy, algorithms, and macromolecular chemistry in order to decode the contents of a strand of DNA. So, even if we make some progress, we still haven’t completely solved the problem of the non-universality of digital media. There are two complementary solutions that will enable our descendants to cope:

*The first solution is to leave instructions for our descendants, a kind of “user manual.” Let’s say that a digital data storage center capable of withstanding the passage of time will exist, we could also store in a separate location, on an engraved tablet for example, a set of very specific technical details on the operation of the center as well as useful technical basics (genetics, mathematics, user manuals, etc.).

*The second solution is to use a less massive memory than conventional digital memory, which brings us back again to the issue of the degree of information selection that we discussed earlier (should we provide our descendants with a selective or exhaustive epistemic memory?).

There is one important thing these two solutions have in common: information must be stored on a medium that is easy to decrypt, even for a less advanced civilization, and that has a very long lifespan. One possible avenue is existing physical media, such as paper, clay, or basalt, which can have a long lifespan. It will indeed be less difficult for a culture with no knowledge in computing technology to understand instructions left on sturdy tablets. However, the amount of information that can be recorded on these media remains very limited, if only because the amount of information that can be written on a tablet is inversely proportional to the size of the characters used. By writing all our knowledge of chemistry, genetics, mathematics, and computer science on plates, on which the size of the characters does not exceed a few tenths of a millimeter (as is the case in the “Memory of Mankind” project in Austria), it is still possible to avoid using too much physical media. But for a more extensive knowledge ark, even if we select the information, this technique is unsatisfactory. Just to write the contents of an “ideal library” of a few thousand books (which is already a very restrictive selection), several thousand plates are needed, which must be stored in a sufficiently large space. A possible compromise between accessibility and mass digitization is to use optical disc technology, in which information is stored by means of small engravings, but can be read using only rudimentary optical tools, such as conventional microscopes, and polarizers. 5D optical data storage is a well-known example of a medium capable of holding a reasonable amount of information while requiring a relatively simple reading tool. It also conveys the advantage of an excellent lifespan. This technology has already been tested for the lunar library project we mentioned earlier.

Language evolution must be taken into account for transmission

In order for our knowledge to be usable by our descendants, we will need to use technologies that are sufficiently simple and standardized. If software components are necessary to read the information media, an easy-to-decipher computer language will also need to be developed. But there remains another major challenge: finding a natural language that future generations will understand. Regardless of the expected longevity of an epistemic memory project, and regardless of the recipients of such a project, the problem of language barriers will have to be addressed. Indeed, it cannot be ruled out that the languages we speak today will no longer exist by the time our descendants rediscover traces of our civilization. There have been many instances where a people outnumbered by another in a given territory has adopted the language of the dominant people, first as a lingua franca, then as their mother tongue. This assimilation can be forced (colonization, military conquest, etc.) or more “natural” (cultural influence, migration, the importance of trade, etc.), but in all cases, it can lead to the extinction of entire languages. Currently, it is estimated that of the 7,000 languages spoken on Earth, between 50% and 90% of them are threatened with extinction during this century. While the languages most at risk today are often those spoken by a small number of speakers, it is not impossible that the “major” languages (English, French, Mandarin, Arabic, etc.) will one day be endangered themselves. Incidentally, language can also be a heritage in itself, to be preserved for posterity and for the great joy of future linguists who strive to understand the structure, diversity, and dynamics of language. Efforts are sometimes made to preserve as much as possible of endangered languages through sound recordings, the creation of dictionaries that are as exhaustive as possible, or the most accurate documentation possible of the language’s usage. One example is the UNESCO Collection of Representative Works. However, this work is often complex to implement for many languages that are too isolated.

But languages do not necessarily have to disappear for the obstacle of translation to arise for our descendants; it is enough for current languages to change radically. Many languages have not disappeared because they have become extinct, strictly speaking, but simply because they have gradually evolved into new languages, without any real break during this evolution being identifiable. Latin evolved into French, Spanish, Italian, and other languages; Coptic evolved from ancient Egyptian; Hindustani and many Indian languages evolved from Sanskrit, etc. However, the consequences remain the same: over the centuries, our writing will become increasingly difficult for our descendants to decipher. One study estimates that in 20,000 years, 99 out of 100 basic vocabulary words in the Swedish language will have disappeared by then, and this observation can probably be extrapolated to other languages around the world. While it is not possible to prevent languages from evolving (nor is this actually desirable), and, as a corollary, while it is not possible to remove the language barrier, workarounds are possible. To ensure that a repository of knowledge remains accessible to future generations, the following strategies can be employed:

*The use of multiple languages: if archaeologists were able to decipher ancient Egyptian, this is mostly because legal texts in several languages were found: this is the case for the famous Rosetta Stone, written in Egyptian and ancient Greek, the latter helped to decipher the former. The more languages in which our knowledge is transcribed, the greater the likelihood that future archaeologists will be able to decipher it. Differences between the languages used can also be a decisive factor: by using languages that are not closely related to one another, we can potentially make the translation process easier. In any case, using a single language for a knowledge base can be a major risk.

*The use of multiple writing systems: in the same vein as above, we could consider using several different writing systems, including for a given language (transliteration), in order to increase the chances that at least one of them will be accessible to our descendants. Writing systems also evolve, and it is not unreasonable to think that the current systems (Latin, Cyrillic, Sinographic, Devanagari, etc.) could disappear in a few millennia.

*The dispersion of “Rosetta stones” across the globe: in parallel with knowledge arks, “instructions for use” could be created in several languages and writing systems, which can be duplicated and spread on several different locations across the globe, again to maximize the probability that future linguistic communities will be able to understand our information. We must not forget that with the significant population movements that can sometimes occur, languages can also move and end up being used far from their territory of origin (Portuguese in America, French in Africa, Turkish in Europe, etc.).

*The use of images: text alone is not enough to preserve many elements of our heritage: photographs, films, visual arts, etc., and much of our literature (including scientific and encyclopedic articles) illustrate their subject matter with schematics anyway, so it would be a shame to preserve only the body text. In line with the previous tactics, it is entirely feasible to create simplified dictionaries, mainly containing basic terms and words referring to less abstract concepts, illustrated with photographs, diagrams, and pictograms.

All these tactics can also facilitate the translation of formal languages (mathematics and computer science), but also and above all the instructions that will need to be passed on to our descendants in order to decipher the digital storage media used for mass preservation. We must also remember that many of the conventions used in mathematical language (symbols in equations, the decimal system, arrows, etc.) and even in other sciences (SI units, the periodic table of elements, electronic diagrams, etc.) are arbitrary, and therefore temporary, and must therefore be given special attention.

Knowledge ark can also be immaterial

The transmission of language is therefore essential for future generations to be able to decipher our knowledge, and this is possible with the creation of a dedicated memory, made of simple material supports scattered throughout the world. But language can also be transmitted through other means that do not require any physical media. Historically, humanity has not always mastered writing, which has actually only existed for 5,000 years, whereas the human species is much older. However, peoples have been able to pass on not only their language, but also stories, anecdotes, knowledge, moral codes, and traditions without the use of writing. This transmission was essentially oral, with little or no use of writing medium, at most, mnemonic devices such as carved bones to count the days of the year, or visual representations such as the paintings in the Lascaux cave). This form of transmission relies on three factors for success: oral elaboration, individual memory, and ritualization.

For an idea or story to be passed down from generation to generation, it must first be shaped by the storyteller in such a way that it is more impactful, so that those who hear it are more motivated to pass it on in turn. This can be achieved through all kinds of techniques such as rhetoric (the art of speech), stylistic devices (alliteration, assonance, proverbs, etc.), rhythm (verses, rhymes, repetitions, etc.), music, oral performance, and the gestures associated with it. This construction also makes oral teaching more memorable for the recipient, which brings us to individual memory. In each generation, once the storyteller can no longer tell the story, the listeners must be able to recall the story perfectly, down to the smallest detail, so that they, in turn, can become storytellers for future generations. This is, in fact, the whole point of oral teaching: to ensure that stories survive beyond death, despite the cognitive limitations of the individual. Thus, successful oral transmission depends not only on the storyteller but also on the listener; the former must employ techniques to facilitate learning, while the latter must have the mental capacity to memorize and must use mnemonic devices. Finally, to further ensure the successful transmission of an oral story, it is preferable that it constitute a kind of “ritual,” a tradition that, if not sacred, is at least considered important to the group within which this transmission takes place. Teaching of stories then becomes not only simple learning, but a cultural heritage, knowledge from ancestors, something that is part of the group’s identity.

Thus, many peoples who did not master writing managed to preserve tales, legends, philosophical ideas, ethical teachings, and even knowledge of very specific historical events such as volcanic eruptions or floods for thousands of years, with relatively little alteration. Even in our contemporary societies, where most learning requires material support (textbooks, guides, tutorials, etc.), part of it still takes place orally, whether through radio programs, political speeches, public debates, documentaries, or even school. Admittedly, the limited cognitive abilities of individuals, even taking into account the size of the world’s population, make it impossible to transmit the entirety of our knowledge, especially the most complex or abstract aspects. However, relying on intergenerational oral memory would have certain advantages: knowledge learned orally benefits the individual throughout their life and is immediately available to them. Moreover, the more fundamental knowledge is known by the greatest number of people, the more resistant it will be to attempts at censorship or any form of unintentional or collateral destruction. Thus, if tomorrow civilization collapses or a particularly autocratic regime takes hold, what survivors remember today will have a decisive impact on what will be available to future generations, including key information for understanding the world (natural selection, atomism, the Big Bang theory, critical thinking, history, etc.). In other words, passing down certain parts of our knowledge orally would have two advantages: first, our distant descendants would be better equipped to interpret our written knowledge because they would be on familiar ground, and second, our more immediate descendants would always have access to fundamental knowledge, even in the event of a societal collapse.

Integral transmission of knowledge will be successful with the help of critical thinking learning

To ensure that our knowledge is available to future generations, current generations must be able to store as much information as possible, so that the most important ones can survive a catastrophe that could damage our heritage and our science. This method is part of the broader issue of what is known as “collective memory,” i.e., all the information shared within a group, whether it be a family, an organization, a country, or even humanity as a whole. Collective memory is a concept that refers to the way in which knowledge, values, and subjective experiences are disseminated within a group. The concept is mainly used to refer to historical memory, the memory of the past, particularly traumatic events, which is similar to the concept of the duty to remember. But oral transmission can also, more broadly, be seen as part of epistemic memory—and thus of the very long-term preservation of ideas, key discoveries, theories, cultural elements, and, indeed, all of the more or less ephemeral intellectual heritage that civilization has produced. The challenges associated with oral transmission remain the same, and the successful transmission of knowledge across generations can only be achieved by overcoming the following obstacles:

*Error: collective memory is obviously not some kind of memory that “floats” above society and is completely immaterial; it is in fact a way of referring to all the individual memories surrounding a given subject at a given moment, distributed among each member of a society. Collective memory exists, in a sense, in a tangible form, through the bodies of individuals. This means that any information that can be memorized by an individual can be altered by that individual, sometimes unintentionally: errors can arise during learning due to cognitive biases, sensory illusions, paralogisms, or simply because the brain is unable to store information the way it is emitted. False memories and quid pro quos can therefore distort oral transmission.

*Appropriation: another way of altering knowledge is to transform it at will, to twist it so that it appears consistent within a pre-existing mental framework. We can give a new definition to a word, exaggerate a point of view with which we disagree, minimize a fact, omit a “detail” that is not really a detail, etc. This appropriation can be beneficial for the creation of new ideas, but can be problematic for the proper understanding and transmission of existing ideas.

*Communication: Knowledge can also be altered when information is shared between several members of a group, whether directly (verbally) or indirectly (typically via social media). The infamous “fake news,” false information about an event in the recent or distant past, is a well-known example of distortion occurring during learning, rather than after the fact due to poor memory. In this case, it is often the person sharing the false information who is at fault, rather than the person receiving it. Understanding of a philosophical or scientific subject can thus be reduced if the least reliable information on that subject circulates more widely than other information within a group. This is an even more significant problem given that certain types of information, particularly partial information and information associated with strong emotions (anger, joy, sadness, etc.), are transmitted more easily than impartial or emotionally neutral information.

*Omission: sometimes, someone who wants to pass on knowledge may deliberately alter it by deciding not to reveal everything. Omission is a form of misinformation, but in this case, the cause is not so much related to the nature of the information or to an actual modification, but simply to a selection of information. Whereas communication problems are most often errors or lies, omission is more of a half-error or half-truth, since the information conveyed is accurate, but simply not complete enough for the learner to have a good understanding of a given subject. This omission may be intentional or unintentional, but when it happens, this means that censorship and taboo are not far away.

*Dogmatism: conversely, knowledge may be considered so important to the group that the slightest alteration is unthinkable, which paradoxically undermines the preservation of that knowledge. Indeed, preservation is not for preservation’s sake; it must also help knowledge to evolve and accumulate over generations. Preservation is also a means for future generations to continue to progress, to create new ideas, to construct new theories, to discuss alternatives, to explore consequences, etc. Dogmatic knowledge, or doctrine, constitutes a form of “local” knowledge that is unalterable and resistant to the passage of time, but dangerous for all other knowledge that also deserves preservation. In other words, all knowledge must be treated equally so that it can continue to evolve and we do not become complacent with what we have already achieved.

All these challenges, which are already paramount for the memory of history, are also paramount for the proper transmission of our intellectual and cultural heritage to our descendants. Therefore, teaching as much knowledge as possible to as many people as possible is not a viable strategy on its own. If we want our heritage to survive through the ages, this strategy must be accompanied by learning methods that limit dogmatic thinking, selection or confirmation bias, and the sharing of unverified information. In other words, the intangible transmission of knowledge also involves the teaching of scientific methodology and critical thinking: knowledge of cognitive biases, the importance of sources, scientific experimentation, hypothesis testing, logic, rhetoric, media literacy, fact-value distinction, etc. In order for our descendants to be better able to welcome the wealth of information we wish to pass on to them, we must ensure that the rudiments of critical thinking are firmly rooted in the current population. Of course, an “immaterial” epistemic memory is not necessarily an alternative to the traditional material memory we discussed earlier. but it is a complementary form of memory, aimed at facilitating access to our knowledge for future generations in the event of any form of cultural amnesia. The critical oral transmission of knowledge aims to foster open-mindedness, curiosity, perspective, and rigor among future generations, so that the preservation of knowledge does not devolve into a kind of intellectual stagnation, scientism, or censorship of new theories, especially when they concern topics on which there is no scientific consensus.