by Danila Medvedev
Context
I am an [applied] futurologist and transhumanist. I needed tools to augment my ability to think more complex thoughts, to store and organize knowledge and to develop complex solutions to complex problems collaboratively. Turned out there was no simple way to solve that.
Reinvention of language was required. Thus after a decade of thinking I have created a new fractal (2,5D) nonlinear scalable, mental-model compatible language, supported by a novel GUI.
It's conceptually hard to speak about language evolution, because so much of linguistics is limited to natural spoken and written languages. Even constructed languages are mostly similar to natural ones. However, I am making a claim for creation of a next-generation language that incorporates existing languages, but extends towards visual language, formal modeling, semi-formal cognitively easy modeling and other forms of communications.
There are countless theoretical and philosophical issues to be solved. But my personal focus is on exploring and expanding the possibilities, because this solution (called NeyroKod) exists as real software, it is being selectively deployed and used to a great effect.
Problem
Humans evolved only one type of language - linear.
A hypothesis states that evolution reused motor memory mechanism (that allowed us to make Oldovian tools). This led to linear oral language. After a brief attempt at pictographic writing (independently invented just 2 or 3 times, primarily in Egypt) humans quickly evolved phonograms in Sumer and then alphabet in Phoenicia. All modern written languages are descendants of that Alphabet with the exception of Chinese logograms (the second text lineage).
Illustrations: metaphors, such as "train of thought", "stream of consciousness", "line of argument".
Yes, speech and written language are different (some, like James C. Stalker, argue that written English is a dialect). But they are awfully close. I claim that modern text (including hypertext) is essentially the same in structure as the earliest spoken/gestural language circa 100,000 years ago.
We need language that handles complexity much better. Linear text is bad for that, even when it's not boustrophedonic. Documents are bad. These old tools do not allow collective thinking about complex problems (wicked problems).
Argumentation for problem
As Frode Hegland writes, our ability to work with text was not augmented in the digital era. I would argue that it barely evolved at all during the past 5,000 years. Digital text today devolves towards "fingered speech" (John McWhorter). We need more than reimagine text for digital era, we need to rethink language!
State of understanding
The transdisciplinary research problem of language 2.0 for humans in the 21st century is not defined yet. There are multiple research fields potentially converging, but there is no coherent research community and no one has solutions.
There is no speculative linguistics, so linguists mostly study linear text. Frode Hegland says that text is flat, but it's worse than that - text currently is linear! Semiotics is no better. Robert Horn staked out visual language, but has little to say about structure. There are a few potentially useful ideas such as spatial hypertext from Frank Shipman or structured writing from Robert Horn, but all of them don't lead very far.
There is no theory of thinking/knowledge (there are people working on "prototype theory", "concept theory" and "theory theory", but it's but a tiny portion of a proper theory of thinking). There is currently very poor understanding of mental models. No syntax to encode them.
Computer science has no theory of UI. Not even basic ideas of quantifying "transaction costs" or mental cost of interactions have been proposed. There is no science of human computer interaction and nothing has been done to establish Douglas Engelbart's research agenda for intelligence augmentation.
There is no research agenda and there are no solutions. There are many individual ideas, such as importance of addressability and interaction (Frode Hegland), but it's not enough to build an alternative to text systems as they are today.
Space of solutions
Interaction requires structure and structure requires turning text into something entirely different.
Today paragraphs are our units of discourse, but they are not clearly defined and no one (beyond school teachers) is responsible for optimizing and improving them. Robert Horn has rightly criticized the status quo and has made a right step with "structured writing".
Shipping containers are a good analogy. And we need to rebuild our entire system of discourse too. We need a new "symbol system". The famous art teacher Betty Edwards argued that most adults have a simple "symbol system" that limits what they can visually conceive. "It requires extraordinary imagination to conceive new forms of visual meaning – i.e., new visual cognitive technologies" (Michael Nielsen), such as "cubism" or "diagrams" for everyone to use.
My vision is that, treating language as a tech and building on our conceptual work with NeyroKod, humans over the next decades can go from rigid and simple to flexible, but complex thinking solutions.
The language tech evolved (it was not consciously developed) in our ancestors:
1. Pictograms - representations of objects/concepts through visual images. They were interpreted.
2-3. Linear speech/text - representations through visual expression of audio expression (speech) of objects/concepts. Text was read. Text acquired grammar allowing representing thought structure.
I think we now can switch to non-linear structures of complex objects that semi-formally represent objects/concepts and allow interaction instead of reading/interpretation. Computers are a good potential tool, but without a conceptual framework for next-gen language they are as powerful as a bunch of emojis.
Figure 1. Language evolution
A very high-level abstract idea of what we have achieved with NeyroKod can be understood in terms of elements, their position and containers for elements that can all be fixed or flexible.
Consider the text technologies of woodblock and movable type.
· Woodblock printing: fixed elements, fixed position, fixed container.
· Movable type: fixed elements, (somewhat) flexible position, fixed container.
· NeyroKod movable blocks: mostly flexible elements, flexible position, flexible container.
Flexibility allows working with arbitrary size units of meaning - not just with letters or entire pages, but paragraph-sized or book-sized blocks too.
We can use digital computers to enable those interactions, but we first need to rethink computer as not just communications devices (Licklider), but as thinking devices. If we pull this off (and I believe we can) we can get to rich and smooth interactions with thoughts/knowledge. To borrow a metaphor from complexity science, we can go from rigid solid text and beyond "liquid" (liquid water has no structure). Your thoughts can be fluid, but not too fluid, keeping you on the "edge of chaos" where the magic of complexity happens!
What doesn't work
May people have argued for simple solutions. I don't believe they can work to enable a qualitative transition in human collective intelligence. Here is what won't work:
· Addons and plugins over web/html (band-aids)
· Graphs/mindmaps
· Hypertext linking. Linking, citation and pointers are not enough. Explicit structure and transclusion are needed.
· Better documents
· Machine readable data for documents
· Scribing and other primitive forms of visual language
Inspirations (analogies)
My solution (NeyroKod) evolved in complex ways. I can't say that it was inspired by a particular idea, but it can be understood better through analogy with some ideas.
· Open hyperdocument system (OHS) by Engelbart.
· Heptapod B language from "Arrival" movie.
Complex solution
Thinking, communicating, knowing, remembering and writing and reading are parts of a whole.
Rearranging knowledge is key. editing text is not that. Engelbart argued in framework that its just the first step. My work is a logical continuation (higher levels from the Framework).
Many understand the need to go there, but no one knows how to do it (except me). "Historically, lasting cognitive technologies have been invented only rarely. Modern computers are a meta-medium enabling the rapid invention of many new cognitive technologies." (Michael Nielsen) But despite the potential, real results are underwhelming.
Many things need to be invented to enable a breakthrough. Methodology is just as important as an actual tool. Format is important, it's connected to what is the format of thoughts/knowledge. Several conceptual inventions enable progress here.
First is: two-way transformation of conceptions and models into a 2D interactive representation. This is taking Claude Shannon breakthrough in logic and applying it to ontologics.
Second: the entire information management system should minimize the need to communications, because they prevent scaling and are costly in a system.
Extended thinking requires intermediate storage of information and only visual thinking can provide random direct access to large volumes of information. So the key issue becomes not how it is stored in the brain, but how it's stored on paper/screens and transmitted in/out of the brain(s).
New modality that supports this process, compatible with both brains and computers, was invented and implemented in NeyroKod.
Four modalities
I claim that there are four basic modalities for representing complex knowledge:
1. text
2. "gestalt" images
3. formal databases
4. semi-formal visual models
1-3 are widespread. Number 4 needed adequate tools.
A flexible visual modeler such as NeyroKod combines the best elements of images, text and DBs. It supports structure, complexity, is backwards compatible with text and supports spatial navigation and visual thinking.
Implementation
NeyroKod
I became interested in intelligence augmentation and personal information management in 1990s. In 2000 I expressed the ambition to solve the problem to my colleagues at Trigon Capital. In 2013 I solved it and started development.
NeyroKod is not just a new concept for text/language, it's an existing innovative platform for working with information in a new way. So it's reinventing IT ecosystem too, it's not just an addon, a new document format or a new app.
Features
NeyroKod is a cross-platform multi-device native application that functions as a general-purpose multi-user knowledge work environment.
Interactivity (reading and writing are combined) is a basic mode. But it's unimportant, compared to everything else.
Links to elements of structure give freedom to rearrange and remix text.
There is always an explicit outline/table of contents. Structure and content are integrated in one view. There can be multiple dynamic views. Transclusion is fully supported.
Argumentation for solution
NeyroKod can be both more and less formal than text. Its fractal nature allows unfinished fragments, while its innate structure supports formalisms.
Higher (estimated to be ~2,5) dimensionality allows more content per unit of area ("page") than regular text or lists.
Fractal nature allows better scaling, leads to resilience. It can be used to represent complex fractals such as organizations and information flows. NeyroKod interface is space filling and self-similar, leading to a more organic evolution of knowledge content than in traditional CMS-like systems.
Evidence and testing
We are deploying NeyroKod is Rosatom, the Russian Atomic Energy Corporation, with our ultimate goal - to support all energy R&D and communications about complex technical projects. But the power users right now are members of our core project group.
Users need to make explicit and change their thinking patterns. And to see the biggest benefits they need to collaborate intensively (with other power users).
NeyroKod scales to large knowledge bases (millions of elements per power-user now, hundreds of millions of elements conceivably).
Use cases
There are countless use cases, but we focus on applied complexity. For example NeyroKod can potentially be used to design industries, manage global climate change, integrate national innovation systems. This shapes both our overall strategy and our development priorities.
Our strategy
In the short term we will continue working in stealth mode with closed development. The goal is to maintain control over architecture and its conceptual integrity (as Fred Brooks implored). We will focus on serving extreme power users (system engineers, futurologists, ontologists, etc.).
The medium-term plans are about a David vs. Goliath competitive strategy. The communication and knowledge paradigm is controlled by the digital monopolies who prescribe tools and services. Microsoft controls productivity and desktop publishing, thus it controls the text paradigm. We can't change the text without destroying Microsoft monopoly. Google controls information search and email (partially). Facebook controls communications and messaging.
We need to defeat them too, which is obviously a really huge challenge. But we need to liberate text. We are ready for a long struggle fought on the ideological and cultural battlefield, not just in the marketplace.
Only in the long term, after we defeat the incumbents, we will be able to make standards, protocols and the entire tech stack open and free in all senses.
Implications
"New representations of thought — written language, numerals, mathematical notation, data graphics — have been responsible for some of the most significant leaps in the progress of civilization, by expanding humanity’s collectively-thinkable territory." (Bret Victor) My claim that NeyroKod may enable the next historic leap in the 21st century, the one, arguably envisioned by Engelbart 50 years ago.
Interpersonal communications are replaced by working with models, which scale much better. Consensus may emerge in such system in a more organic way, leading through Man-Computer Symbiosis to better organizations.
We will be able to design complex things, including architecture for digital transformation of national economies, architecture for next generation IT systems, for innovation management systems and for entire economies. Our tools determine the limits of the possible. Our design and management tools even more so.
Impact on productivity: NeyroKod may serve as "a mechanism for speeding-up the process of technological diffusion" (Andrew Haldane) because it enables building and transferring more complex models than traditional documents or presentations. There are many implications for management, R&D, ICT, etc. Overall NeyroKod may lead to significant quantitative effects as well as radical leaps.
Next steps
There are many research challenges ahead, too many to list here. Fortunately, NeyroKod may serve as a foundational stone for bootstrapping and for directed co-evolution of human thinking and computer tools. We can better manage R&D with NeyroKod, which will help us develop it further, use it better and become smarter and more capable, leading to more R&D breakthroughs. I believe that the next wave of technological development is starting now.
Appendix
NK link: nk://id/p2cTmR55TkqYUSY4F7MpDA==