In the following, I will present an idea for a new type of encyclopedia. Encyclopedias have traditionally been the work of comparatively small groups of people who aggregate knowledge and compile what seem to be the most accurate and truthful accounts of any number of topics. The reputation of an encyclopedia depended on the reliability of the information it provided in the eyes of its buyers—the educated public.
Online encyclopedias (à la Wikipedia) have taken this to another level by letting the public participate directly, the idea here being that a process of distillation of information toward truthfulness is the expected outcome of public debate—in the form of collaborative editing and transparent cross-checking.
The unifying theme remains the same. In an encyclopedia, one expects to find the single most truthful description of any given topic. This is the so-called view from nowhere,[1] or the god’s-eye view, which means the view is neutral, free from interest and passion, and theoretically unbiased.
I have and will continue to express my doubts about the possibility of such a neutral viewpoint, but for the sake of this document my approach will be to simply point out that the philosophical opposite of this take, in the form of an actual encyclopedia, is an unoccupied niche. Whatever the criticisms leveled against the existing systems—mainly, of course, Wikipedia—nobody has attempted to produce anything fundamentally different.[2]
Philosophy, as I have recently said,[3] can reveal different design alternatives. By philosophy I mean not necessarily the (historical) discipline, but rather the practical exercise, which you might otherwise call first principles thinking—thinking based on logical possibilities.
When one really asks the question of where knowledge is actually “stored,” there are two places. One is “in the world,” that is, in concretely and materially existing symbolic artefacts. True, symbolic artefacts depend on some tacit component; there must exist a culture that knows how to read those symbols. But that aside, diagrams, lists, numbers, images, and texts are what knowledge consists of, ultimately—as far as it is to be found in the world. [4]
The other place where knowledge “sits” is in our heads. Only in individuals’ minds does knowledge actually get synthesised. Only in individuals’ minds does knowledge come together in such a way that new connections are made, new insights are gained, and sense can be made of things.
However inadequate such subjective knowledge might seem, I think we can concede—however scientifically minded we consider ourselves, and whatever our scientific aspirations may be or the merits or deficiencies of the scientific method—that there exist no perfect artefacts. That is, on no topic (with the situation being better only in mathematics and logic) do we have anything resembling truth. Different people might claim to be in possession of the current best explanation or theory on given topics, but the fact is that you won’t find any two individuals (or groups, with a head typically representing and championing the opinion of the group) who actually agree on a certain take on things. Moreover, groups of people tend to not just disagree “at the edges,” but rather polarise over their disagreements.
Encyclopedias gloss over, and therefore don’t model, that simple fact. The intuition that there exists one reality, and therefore there must be one true description of it, prevails. The system I am proposing takes the radically different position of putting not the text first, but the mind.[5] The name Personalist is an allusion to mid-20th-century polymath Michael Polanyi’s work Personal Knowledge, which argues that entirely dispassionate, neutral science is impossible; all knowledge contains a tacit component that simply cannot be formulated and is grounded in our being human (and embodied, as—I should add—later cognitive science will argue).
Chateau Versailles Galerie des Glaces [commons]
The cornerstone of the system under design is simple definitions, as they are written by—and thus appear true for—a given user. To the user’s best understanding, what a description says spells out what a title (a name) implies. What a thing with a certain name is is laid out in its description.
The fundamental data model atom is called Identity in our system. This is first of all in contrast to a perhaps more obvious choice like Entity. An entity implies being, and I have no reason to rule out talking about impossible or non-existent things. Furthermore, I want to make clear that the fundamental atom can not only express material things (which Entity nudges one toward) but also events, processes, or (possibly) relations. In fact, I talked about relations recently[6] and I feel inclined to keep the fundamental data-holding structure as slim as possible, offloading data into relations. I will return to this below.
What I think makes Identity an excellent choice against the backdrop of the current system’s philosophical underpinnings is that people change their minds. Often there is such a distinction between an early and a late Ludwig Wittgenstein or Douglas Hofstadter.[7] I want the system to be able to track one’s evolution of thought. The literally only thing that keeps a thing “together,” across space and time, is not even its name. It is our deeply ingrained capacity to track objects[8] across time (and space)—nothing else. In child psychology, this issue is discussed as object permanence, a child’s developing ability to understand that objects don’t disappear when occluded and that different momentary experiences[9] of a thing might constitute one and the same thing—across time (and space). Identity means that—it’s the same.[10]
As far as relations between things are concerned, the relevant observation is that words are always defined in terms of other words. Apart perhaps from conjunctions like “and” and onomatopoeic expressions like “Oh!”, there are literally no words that contain meaning per se. For further illustration, have a look at any monolingual Oxford English dictionary.
For our system, that means first of all that Relations between definitions will be an integral part. There is also a consideration of actually treating relations as full-on first-class citizens and allowing definitions of the same “rank” for relations as for what I called Identities (up to having only definitions in Relations). This allows for modelling aspects of things. For example, “functions” in mathematics are something related yet slightly different from functions in programming. For purposes of exposition and curation, this could turn out to be a very good choice.
Finally, we keep a history of Relations in the same way as we do for Identities. This will enable us to time travel and trace a use’s development of their whole “web of thought.” As said, things are defined in terms of other things. The understanding of one thing informs the understanding of another. One’s understanding of Liberalism influences one’s understanding of Fascism as well as of Communism, and so on. But once new data “comes in”—for example, when one reads another book on any of those topics—one’s understanding of the others is affected as well. To reflect the latest understanding, all definitions would have to be updated to form a new coherent whole for that specific point in time. We will be able to follow the user along that intellectual journey across time.
The things said so far concentrate on a single person’s worldview. While that is true, Truth is—if never fully objective, because that is impossible—intersubjective. The integrated universe of all persons’ individual knowledge is the body of knowledge that ultimately should interest us and take us out of our own “bubbles,” and we will model this by allowing users to reference definitions from other users, to express their approval of viewpoints that have not been formulated by themselves.[11] However, users will only be able to modify their own definitions. But allowing users to “follow” others aims at mirroring sociological phenomena like schools of thought, lineages, traditions, and entire cultures. Polanyi’s relevant expression is that of a Society of Explorers,[12] who together make sense of the world. This is a polycentric view of knowledge.
In the domain model, users will be called Persona. First of all, on an abstraction level this fits well with Identity, since it generalises over multiple things. In allusion to David Deutsch’s conception, over humans and other intelligences. And then, over one or multiple people (for example, to represent “the voice” of a collective, like Bourbaki, a group of mathematicians, or, say, that of a music band). Furthermore, to protect oneself in political climates that change over decades, and taking the long memory of the internet into account, one might want to publish first of all (semi-)anonymously, and related to that, possibly under multiple personas (just as people do in real life, when they present themselves differently on LinkedIn than in private chat groups).
You can visit a live Beta version of this project at personalist.org. Its GitHub repository can be found here.
The ideas developed here are in part a derivative from ideas developed from an earlier system of mine. See Rhizome - A “total recall” note-taking and content-management and -archival system for Superhuman Memory [Whitepaper] (link). Another source of inspiration, and the basis for the implementation, is the XTDB database; as so often, the mere availability of certain technologies can spark new ideas.
Footnotes
To use Thomas Nagel’s phrase.
The most recent contender has been Grokipedia. All of the contenders concentrate on information quality and fact-checking, while keeping the fundamental premise.
Here: Relations All the Way Down - link.
Often, if not here then elsewhere, or the sake of simplicity, I will simply speak of texts.
A minor allusion in the name Personalist is also to David Deutsch’s concept of a person: an intelligent mind explaining the world. “Person” abstract away the embodied part and as such doesn’t necessarily refer only to humans.
Here: Relations All the Way Down - link.
I always marvel at their similar trajectories, from Tractatus Logico-Philosophicus to Philosophical Investigations. From Gödel, Escher, Bach: an Eternal Golden Braid to Surfaces and Essences: Analogy as the Fuel and Fire of Thinking.
First of all, this extends to classes of objects, not to individuals. We are able to track the meaning of Zebra. But then also, this is not meant to be restricted to objects; objects here can stand for processes or events as well. Language shows us that “Cooking is fun” is as valid a sentence as “Roses are red.” And I mean roughly the same thing by “cooking” a year from now as I do now. “Cooking” and “roses” are both concepts whose meaning I track over time.
Reminiscent of W.V.O. Quine’s contrived “momentary rabbit experiences” in his famous Gavagai example from Word and Object.
Ship of theseus: solved! ;)
Again, this will naturally support time travel.
As an aside, there is also Marvin Minsky’s similarly named Society of Mind, which seeks to explain the mind as made up of interacting agents.
