‘The Sum Total of All Human Knowledge’, Part III

Finding a schema

The quest to find a systematic knowledge organising schema—roughly aligned with the Cosmic Evolution timeline or Big History through-line—arose from the idea to go ‘full Zettelkasten’ on the many hundreds of notes I’ve accumulated over the years that are scattered about in various notebooks, electronic and physical, and scraps of paper filed in manilla folders languishing in various filing cabinet drawers. This is not only a useful and fun way to exercise one’s mind to try to keep it active, but is also a quite interesting exploratory research project to see just how far this wonderfully preposterous idea can be pushed. And it might even be of use to anyone else looking to use the Zettelkasten method for organising their research notes along the general lines being described in this series.

Zettelkasten as research partner

As noted in an earlier post, the German sociologist Niklas Luhmann utilised a system of A6-sized slips of paper (zettel—singular and plural; the connotation is one of fairly thin pieces of paper, rather than thicker card stock) organised into a cabinet or box (kasten) having filing drawers into which the slips could be filed. This was his Zettelkasten (ZK). He developed an intricate technique of numbering and cross-indexing the zettels1 so that he effectively created a hypertext-style system, implemented in an analogue way, long before the Web was invented or developed. Interestingly, one of the zettels in a section devoted to ad hoc notes about using a Zettelkasten (ZKII:9/8,2) contains this observation: “The microprocessors are announced but not really available yet”. That same zettel also implies that the ZK could be considered as a “second memory”, while elsewhere he even referred to it as a (junior) research partner (Luhmann 1981).

Luhmann actually had two distinct collections, what researchers studying Luhmann’s legacy and methods (‘Luhmannologists’?) call ZK I and ZK II. The second was developed to support his reply to the question put to him of what research he would pursue as the first professor appointed at Beilefeld University in 1968. His famous response: “Research project: Theory of society. Duration: 30 years. Costs: none.” Apparently, this is even available on a T-shirt these days. The earlier, first Zettelkasten, ZK I, by contrast, was not geared to a specific research theme, but rather was used to support his wider and more general reading, undertaken in the decade or so before he became a professor. As a result, it contains around 10 times more top-level categories than ZK II (Schmidt 2016).

My motivation for utilising the ‘Zettelkasten method’ is nowhere near as specific as Luhmann’s ZK II, but much more like the more general purpose of the first one. Regarded as a generic use case that other researchers might also have, that motivation is, well, as we’ve seen, trying to find a way to organise what few drops we individually manage to extract from the planet-spanning ocean that is the totality of knowledge.

The two main types of ZK one finds discussed in various online communities (e.g., on Reddit) are the Digital ZK (DZK), and the Physical ZK (PZK), sometimes also described as an Analogue ZK (AZK), or more recently even as an “Antinet” (Scheper 2022). There is considerably more discussion about the Digital form than the Physical/Analogue one, and many if not most people start by implementing a DZK through various software systems either dedicated to it, or which can be adapted to do so. This is indeed what I started trying to do, as mentioned in that earlier post. However, the following advice from the blogger abramdemski seems most apposite to this endeavour (abramdemski 2019):

I strongly recommend trying out Zettelkasten on actual note-cards, even if you end up implementing it on a computer. There’s something good about the note-card version that I don’t fully understand. As such, I would advise against trusting other people’s attempts to distill what makes Zettelkasten good into a digital format – better to try it yourself, so that you can then judge whether alternate versions are improvements for you. [emphasis in original]

Because a physical implementation does not allow for easy electronic hyperlinking or searching on keyword tags, some careful thought needs to be given to the form of the indexing system that will be used to place, or ‘install’, zettels into the ZK. Since these cards or slips of paper are physical objects, they will not be able to be in more than one place at a time, something that ‘virtual’ zettels can effectively be, via search queries on tags or special strings of characters. The indexing used needs to not only be able to file these sensibly, according to a logical system, but also allow them to be found again when needed. At least, ideally so!

So how to start? One obvious starting question is: what are the options available for use as a numbering schema to organise human knowledge? But the more important qualifying criterion for any serious consideration of any of these becomes: what schemas exist that could be used to not only systematically organise knowledge, but to do so in a way that is in some way aligned with the intuitive ordering sequence of the timelines of Cosmic Evolution or Big History? That is, which of those available would be suitable for using something like the Cosmic Evolution Time-Line or the Big History Through-Line as an ordering index parameter?2

“We shall not cease from exploration”3

Well, libraries have used numbering systems for classifying knowledge for well over a century and a half. The Dewey Decimal Classification (DDC) is perhaps the most widely known, but two others have also attained very wide international use: the Library of Congress Classification (LCC), and the Universal Decimal Classification (UDC). So, these were the ‘first approximation’ with which to commence the exploration of options.

However, the qualifying criterion—alignment with the through-line—immediately rules out the LCC. That numbering system is quite chaotic, since it has grown organically over time as new materials and themes showed up (rather as Luhmann’s did), instead of having been designed with an overarching unifying structure at the outset. It is not at all suited to the use case being considered in these posts.

This leaves DDC and UDC as candidates. As Batley (2014, 87) notes:

The original purpose [of UDC] was to use a modified and expanded version of DDC to organise a universal bibliography of everything that had been written throughout history. Thus its original purpose was for documentation, not shelf arrangement. [emphasis added]

In other words, what is striking is that UDC was specifically designed precisely for the purpose of organising physical catalogue cards (UDC Consortium 2022a; “Universal Decimal Classification” 2022). This is clearly a very positive point from the perspective of the design goals for a physical ZK.

The UDC (like the DDC) utilises 10 top-level categories for organising knowledge, with the numbering system starting at ‘zero’ (UDC Consortium 2022c):

  1. Science & knowledge, organization (with numerous related categories also contained in this section)
  2. Philosophy, Psychology
  3. Religion, Theology
  4. Social Sciences
  5. [currently vacant]
  6. Mathematics & Natural Science
  7. Applied Science, Medicine, Technology
  8. The Arts, Recreation, etc
  9. Language, Linguistics, Literature
  10. Geography, Biography, History.

This had an instant kind of ‘ooh, almost!’ feeling to it. Especially appealing is the fact that the process of science and creating knowledge has its own specific top-level category, into which other related topics and activities such as Librarianship, Information, Documentation etc are placed. One even finds Futures located in there (specifically at 001.18 Future of knowledge, although futurists do tend to baulk at the term ‘futurology’ which is also used). But on a closer look, it seemed that many categories would need to be shifted around quite a bit to even roughly approximate the time-line that is intended to act as the primary index parameter.

Consequently, a fair bit of time was spent on shuffling things around to try to cook up a suitable arrangement of knowledge disciplines roughly aligned to the through-line. For example, 159.9 Psychology seems a bit out of place with Philosophy, so it was moved to be in with Medicine. If you visit https://udcsummary.info/ and take a closer look, you’ll see a ‘gap’ at 619, at the end of 61 Medical sciences. The thought was that perhaps Psychology would better fit there. But that was only the start.

Then the thought occurred that perhaps 3 Social Sciences and 5 Mathematics & Natural Sciences should be swapped, to better reflect the temporal ordering of the time-line. And also that 91 Geography & Exploration of the Earth etc should perhaps have gone together with 55 Earth Sciences, Geological Sciences (before the 3 and 5 sections were swapped). Then it seemed that 57 Biological sciences, 58 Botany, 59 Zoology should go into a new top-level category, with the finer details of some of the biological sciences that are ‘buried’ a level or two deeper brought ‘up’ a level or two of decimal places. So, a new 6 Biological sciences was born, incorporating Botany and Zoology. You can probably see by now how this was guided by attempting to re-construct in outline form the ‘Cosmos, Earth, Life, Humanity’ sequence from Table 1 from the previous post.

Then Futures Studies needed to go in somewhere, too, of course, most sensibly at the end of History (obviously; see Figure 1 in the previous post), so it got its own spot as 99 Futures Studies. And so on. This quickly became a highly convoluted jigsaw puzzle, with pieces being moved about here and there, which had the unfortunate side-effect of making the existing cross-referencing already built into the UDC effectively unusable without a major re-jigging of the cross-refs. That looked to be something like an order of magnitude more complex and time-consuming than the cooking up of the new ordering of topics in the first place.

This was made even worse by discovering—part-way through this re-arrangement process—an extremely suggestive intuitive schema, namely the “10 Pillars of Knowledge” (10PK) system developed by Chaim Zins (2011; 2022). A quick look at that structure on the web site shows that it very much more closely aligns with the time-lines of Big History and Cosmic Evolution. So, some more time went into trying to re-re-arrange the UDC based on a broad use of the 10PK top-level structure shown in that suggestive graphic, and the more detailed knowledge tree that can also be found there. This meant creating another new top-level category ‘The Human Being’ based on the 10PK’s 6th pillar ‘Body & Mind’. Here at least was an independent rationale to put Medicine and Psychology together, which seemed like a promising conceptual convergence. But the by-now already highly spaghetti-like classification schema became ‘über-spaghetti-fied’ even further…

“And the end of all our exploring”

By this stage abandoning the UDC entirely, foregoing any more rearranging, and just using the 10PK system ‘as is’ was looking like the sensible option, but for one major problem: 10PK is still a work-in-progress, so we would not be able to rely on the positions of the categories remaining fixed, as Zins elaborates and fills in the structure further. Since what we are intending to do is utilise a long-term filing system with fixed numbers for cross-referencing of physical research notes (so any re-indexing would be not at all straightforward), the filing numbers for the topic categories need to be unchanging from the outset. This means we need to find something else that is much more settled and not as potentially volatile in terms of structure and ordering.

Reading around a number of journal articles and book chapters dealing with knowledge classification systems in the emerging (and highly fascinating!) field of ‘knowledge organisation’ (e.g., Furner 2021) revealed a new knockout candidate: the Outline of Knowledge (OoK) laid out in the Propædia volume of the 15th edition of the Encyclopedia Britannica (Propædia: Outline of Knowledge & Guide to the Britannica 1994). The OoK, which was new in the 15th edition first published in 1974, took 8 years of development led by Mortimer J. Adler and a team of experts to design. Adler was well-known, among other things, as a philosopher and popular author, especially for his legendary How to Read a Book (1972). Here at last was something that finally looked like it was purpose-designed to fit the bill.

“Will be to arrive where we started”

If you go to the Wikipedia web page that describes the OoK, you’ll see this remarkable sequence of 10 top-level categories, or Parts:

  1. Matter and Energy
  2. The Earth
  3. Life
  4. Human Life
  5. Society
  6. Art
  7. Technology
  8. Religion
  9. History
  10. Branches of Knowledge.

Now, keep in mind the timelines and temporal sequences from the previous post, and then look more closely at both the second-level Divisions and third-level Sections given on the Wikipedia page. A very strong parallel with the Cosmic Evolution and Big History timelines is even more evident; in fact, it is almost as perfect a rendering of these timelines as one could hope for. Furthermore, this ordering has been fixed by posterity—the fixity of the published versions means that the potential volatility of subject placements is not an issue. So, while the 10PK framework is more recent and potentially more up-to-date and flexible (and the ‘family resemblance’ to OoK is certainly quite marked), the degree of certainty for the ordering and placements required for our use case is still not there yet. The advantage of a very widely known and stable schema, such as the OoK, is that researchers using it for their categorisations can compare notes (literally) as easily as asking the question, e.g., “what do you have at 131.D.2?”

There is also an advantage to the OoK over the UDC, apart from the ordering of knowledge disciplines, namely the relative ‘looseness’ of the structure. While it is possible in OoK to go down seven levels of sectioning, these are nowhere near as rigidly-defined as the decimalisations of the DDC or UDC. Part of the design thinking for a Zettelkasten is to allow for some degree of relative flexibility, to allow for instances of serendipity (Schmidt 2018). This includes allowing the possibility for the knowledge structure to also grow somewhat organically ‘bottom-up’ from a few defined starting points (as Luhmann certainly did with ZK II), even if there exists an overarching ‘top-down’ structure that enfolds it. But equally, part of the utility of the framework structure of the OoK is that the more extensive indexing and cross-indexing required in an organically-growing ZK that does not possess such an overarching knowledge structure is not quite as strongly necessary. The structure itself already provides a good deal of that indexing, as well as the pre-defined cross-referencing already built into the OoK from the start.

“And know the place for the first time”

Adler himself observed (1994) that Part 10 of the OoK has a special character—it is a kind of ‘hub’ surrounded by the other 9 knowledge sectors. Although he didn’t use those exact terms, he did use a circular diagram showing this arrangement (p.7), which seems to imply exactly this analogy. Elaborating, he says:

The reason for this special placement of Part Ten stems from the one organizing principle to which the Editors were explicitly committed in planning and producing this new Britannica. Briefly stated, that principle involves a distinction between (a) what we know about the world of nature, of man and society, and of human institutions by means of the various branches of learning or departments of scholarship; and (b) what we know about the branches of learning or departments of scholarship—the various academic disciplines themselves. [emphasis in original]

In clarifying this distinction further, he says:

For the most part—there are a few exceptions—Parts One through Nine represent the knowledge of nature, of human society, of human institutions, and their history. In clear contradistinction, Part Ten mainly covers the disciplines themselves—the branches of knowledge or fields of scholarship—by which one inquires into, thinks about, or comes to have knowledge of the world in which he lives. Part Ten examines the nature, methods, problems, and history of the various branches of knowledge or scholarly disciplines, the actual content of which is set forth in Parts One through Nine.

Let us take this two-fold organising principle as foundational then.

If we change ‘Part 10’ to ‘Part 0’, then a convenient numbering system results that has quite interesting resonances with the UDC. Recall that the UDC treats class 0 in a somewhat related manner, as concerning science and knowledge in general, with sections on how science is done and how knowledge is created, stored, curated and disseminated. So, fortuitously, and without too many conceptual gymnastics, we can choose to use the OoK as the broad schema, with enhancements added to that, where needed, from the UDC and its own design philosophy, at least for Part 0. Therefore, the time spent investigating the UDC has not been wasted; indeed, rather the contrary. It seems to have borne fruit, since the UDC approach provides a ready way to add extra useful categories especially suited to the central ‘hub’ aspect of the OoK, which is concerned with a kind of ‘meta-knowledge’—i.e., knowledge about how we develop knowledge about the Universe.

Adler also suggests (p.7):

Because it is constructed in this manner, the Propædia provides the reader who wishes to pursue the study of a whole field of knowledge with an easily used guide.

To this we might add, and not without some justification, that the Propædia provides a way to pursue the study of not only “a whole field of knowledge”, but—by dint of its overall design philosophy—to organise the totality of human knowledge itself (at least in principle).

One also finds the following intriguing related statement on the UDC web site (UDC Consortium 2022b):

In UDC, the universe of information (all recorded knowledge) is treated as a coherent system, built of related parts, in contrast to a specialised classification, in which related subjects are treated as subsidiary even though in their own right they may be of major importance. Thus specialists may often be led to related information of which they would otherwise have been unaware.

And so…

What has emerged from our explorations of the options available to make sense of and organise human knowledge using the temporal sequence of Big History and Cosmic Evolution as an index parameter is the following possibility: to use a combination of both the OoK and the UDC, since they both contain elements that would seem to be useful for this purpose. This is a rather nice and conceptually satisfying confluence of ideas from these two different knowledge-organisation systems.

In closing, it is worth noting one key caution that we might bear in mind as we proceed.

Luhmann himself (1981) warned against a fixed order of topics (as indeed do some influential figures in online Zettelkasten communities of interest). One might suspect this may be because he was conducting his researches and investigations growing outwards from a very specific set of starting points in ZKII—i.e., for his theory of society—rather than reading with the much broader generality which accompanied the creation of ZKI (which we recall had ten times more top-level categories). Nonetheless, it is well to keep this caution in mind as we progress to the next stage of this process—namely, implementing the proposed numbering schema in an actual real-world physical Zettelkasten, rather than merely in the immaculate perfection of a conceptual idealisation.

At any rate, both the design philosophy of the UDC (as quoted above) as well as the so-named ‘auxiliary’ elements of the numbering system it uses (which will be examined in the next post) allow for some quite flexible cross-linking between topic areas, which perhaps goes some way to mitigate the perceived dangers of fixed-ordering filing that Luhmann cautioned against.

We’ll pick up that story anew in the next post.

Next Time: Part IV: Refining the schema

Notes

  1. Strictly speaking, the plural should also be zettel, however an influential German-speaking proponent of the Zettelkasten technique has suggested that when the word is carried over into English, it is OK to use the plural ‘s’ as we would for English words. Who am I to argue with a native German-speaking Zettelkasteneer? https://zettelkasten.de/posts/qna-6-plural-zettel/.
  2. Henceforth, further references will be made simply to the ‘time-line’ or ‘through-line’, used somewhat interchangeably, of the overall temporal sequence of Cosmic Evolution and Big History, without necessarily qualifying them as being Cosmic Evolution or Big History. It should be understood that the reference is being made to either or both the Cosmic Evolution Time-Line and/or the Big History Through-Line and their commonalities.
  3. From TS Eliot, ‘Little Gidding’, Four Quartets (Gardners Books; 2001). Orig. publ. 1943.

References

abramdemski. 2019. “The Zettelkasten Method.” LessWrong (blog). September 20, 2019. https://www.lesswrong.com/posts/NfdHG6oHBJ8Qxc26s/the-zettelkasten-method-1.

Adler, Mortimer J. 1994. “The Circle of Learning.” In Propædia: Outline of Knowledge & Guide to the Britannica, 15th edn, 30:5–8. The New Encyclopaedia Britannica. Chicago: Encyclopaedia Britannica Inc.

Adler, Mortimer J., and Charles van Doren. 1972. How to Read a Book: The Classic Guide to Intelligent Reading. Revised and updated edn. New York: Touchstone / Simon & Schuster.

Batley, Sue. 2014. “Classification Schemes for Specialist Collections.” In Classification in Theory and Practice, edited by Sue Batley, 2nd ed., 87–142. Oxford: Chandos Publishing. doi:10.1533/9781780634661.87.

Furner, Jonathan. 2021. “Knowledge Organization in the Wild: The Propædia, Roget’s, and the DDC.” Information & Culture 56 (1): 1–31. doi:10.7560/IC56101.

Luhmann, Niklas. 1981. “Kommunikation mit Zettelkästen: Ein Erfahrungsbericht [Communication with slip boxes: A field report].” In Öffentliche Meinung und sozialer Wandel [Public opinion and social change], edited by Horst Baier, Hans Mathias Kepplinger, and Kurt Reumann, 222–28. Wiesbaden: VS Verlag für Sozialwissenschaften. doi:10.1007/978-3-322-87749-9_19.

Propædia: Outline of Knowledge & Guide to the Britannica. 1994. 15th edn. Vol. 30. 32 vols. The New Encyclopaedia Britannica. Chicago: Encyclopaedia Britannica Inc.

Scheper, Scott P. 2022. The Anatomy of an Antinet Zettelkasten. Scott P. Scheper YouTube channel. https://youtu.be/UV7vaqElPHk.

Schmidt, Johannes F.K. 2016. “Niklas Luhmann’s Card Index: Thinking Tool, Communication Partner, Publication Machine.” In Forgetting Machines: Knowledge Management Evolution in Early Modern Europe, edited by Alberto Cevolini, 289–311. Library of the Written Word 53. Leiden & Boston: Brill. doi:10.1163/9789004325258_014.

———. 2018. “Niklas Luhmann’s Card Index: The Fabrication of Serendipity.” Sociologica 12 (1): 53–60. doi:10.6092/issn.1971-8853/8350.

UDC Consortium. 2022a. “UDC History.” UDC Consortium. 2022. https://udcc.org/index.php/site/page?view=about_history.

———. 2022b. “UDC Structure and Tables.” 2022. https://udcc.org/index.php/site/page?view=about_structure.

———. 2022c. “Universal Decimal Classification (UDC) – Summary.” UDC Consortium. 2022. https://www.udcsummary.info/.

“Universal Decimal Classification.” 2022. In Wikipedia. https://en.wikipedia.org/wiki/Universal_Decimal_Classification.

Zins, Chaim. 2022. “10 Pillars of Knowledge: Map of Human Knowledge.” 2022. https://www.success.co.il/projects/10pk/map.

Zins, Chaim, and Plácida L.V.A.C. Santos. 2011. “Mapping the Knowledge Covered by Library Classification Systems.” Journal of the American Society for Information Science and Technology 62 (5): 877–901. doi:10.1002/asi.21481.

Leave a comment