Collective Intelligence: Generating the Semantic Web

Hello hello.

I’ve been contemplating the potential of semantically linked IPFS data. In this way, for instance, we could store a timeline of historical events linked to texts of all kinds including works of independent journalism, quotations, articles, books, personal testimonies recorded on video.

If all those content-addressed “texts” could be semantically linked together as in a graph database, and if we could layer critical analysis over and in-between them . . and if we can implement an open identity system by which we democratically moderated the quality (read as, ‘coherence’) of dialog . . we can create a permanent, pedagogical space for what one day we might feel satisfied calling Collective Intelligence. This could be the left-brained hemisphere of the Age of Imagination and could generate immense political power, if such a space could be formalized.

I’m more capable of writing philosophically than working technically, but I think advocacy will have a great role to play in the movement. Tell me, does The Graph Protocol ( have anything to do with semantically linking files stored by IPFS?


The graph is quite interesting! However it’s use case (my understanding anyways) is that it’s intended to be used by Dapps as a query method for data needed by smart contracts, and using IPFS to store this data. Although I think a solution like that for searching for data on IPFS, and categorizing it would be very cool!

There is an IPFS search engine however due to the nature of how IPFS stores content, and shares content, it doesn’t contain much, and is also hard to find known content due to the cryptic hashes. This is a hurdle which needs to be gotten over at somepoint, and using a graph style protocol like the one you linked to could definitely be a solution.

Thanks for the clarification postables. I had a feeling that it was a technology with totally different capabilities, but I wasn’t sure. What about ArangoDB?

Does it make sense theoretically? In order to break up the complexity of the problems at hand, couldn’t one begin testing methods of storing content addresses and semantic relations between them with a graph database limited to a single server? Perhaps I can work on some graphics to illustrate some of my more nuanced thoughts about how this kind of system might function.

I have been busy writing about these ideas from a non-technical standpoint, but I couldn’t wait to share the idea here. I’ve been very curious if more people in the community are thinking along similar lines.

No problem. I’m not too familiar with ArangoDB but there’s no reason GraphQL couldn’t be used just like The Graph Protocol that you linked to is using.

This area of the conversation starts to get out of my in-depth technical knowledger however I don’t see why it wouldn’t be possible.

I really like this particular community I find there’s a lot of healthy discussion despite the “small” userbase.

So I’ve published a draft of an article explaining my position:

And here’s an excerpt:

Decentralizing technology offers a path forward for humanity by making possible the creation of a collective instrument for elaborating and amplifying truth, an open interface giving each person the power to contribute to a coherent, democratic vision of what we believe to be true and actually desire for the future. Web3 is a renewed opportunity to raise our collective voice and to take control of our governments by defragmenting the human capacities that we channel through the internet. IPFS grants us the possibility of an intertextual, semantic web like Tim Berners-Lee could only dream of, but not for the purposes of machine readability. We should be less fixated on data and artificial intelligence and more interested in meaning and the consequence of the web’s human readability on collective intelligence. The interface that I am envisioning will connect individuals to a deliberative framework operating on an archival substratem to intersubjectively reveal semantic relationships between texts of all kinds. This semantic web provides a mechanism of democratically elaborating meaning, formulating constructive forms of critical analysis, protecting journalistic works from cultural amnesia, engaging our intellectual inheritance, holding political actors accountable for their integrity, and giving the testimony of common people a sociological and political weight.

If anyone is interested in contributing their thoughts, please let me know.

As I am interested in building bridges between Radical Intellectuals & members of the Web3 community, I am beginning to advocate for IPFS on twitter. Please link me in and join the conversation: @JYarrow4

Hi j.yarrow

i found your idea intresting. In my side i try to think a new approach on :

so i 'm very intresting by your approach.


1 Like

Thank you so much for reaching out, Josselin!

It’s great to see that you are developing GUI tools for IPFS. The only way to make the technology useful to the larger community is to create interfaces.

I’m writing up documentation of an approach to modeling semantic relationships between content-addressed ‘texts’. I’ve got some basic ideas down on paper, and I’ll try to clean it up and create some graphics soon to demonstrate my thoughts.

I don’t think we need to worry about identity management. is developing an open identity system, and we’ll all meet in the middle.

you’r welcome instresting project i will see that do you know also that : and that use IPFS


Thank you for sharing.

I must admit: My efforts and insights are completely inadequate. I don’t have a plan . . I have a few ideas that will appear completely amateurish when I do formulate them, and I will.

But I have a vision of possibility: the transcendence of the hyperlink. This I think is bound to that unspeakable mission to locate meaning which generated both IPFS & HTTP. I have a vision of limitation. I don’t know how we will build an alternative, but I can feel it.

I believe that in our information-age struggle to comprehend and formulate meaning we have disregarded something: . . we must remember that philosophical value cannot carry the Truth with capital ‘T’ . . every perspective generates a truth that carries some level of validity. I think we can be sure: There can be no singular apparatus for holding onto meaning/truth, and likewise no artificial intelligence operating on the machine readability of information will be capable of encapsulating the hyper-dimensionality of human meaning. We need to hang up the pursuit of artificial intelligence to solve all our problems.

Look at the experience of ‘surfing’ the web: Every link is a temptation to follow, because if we fail to double-click and take the right course and complete the circuit of the pursuit of meaning, we fear that it may forever be lost, that the connection is broken. This is true in a way. Surfing the internet is like rolling dice, because hyperlinks are almost apparitional . . the connections they forge are fragile because we may never again discover them, and the connective significance they hold may be lost to us.

Furthermore, it is a failure in the design of the internet (or at least the one required by Collective Intelligence) that hyperlinks are essentially singular: They take the explorer distinctly from one text to another, and they can point in no other direction. However it seems to me that their pointification is deserving of innumerable linkages . . In the midst of reading a certain text, every single user may think to themselves of a new connection from this text to another, and these connections will likely be outside of the awareness of the original author and their intent.

If we could somehow offer each and every internet explorer a unique record of the semantic linkages that came to mind for them specifically between all the texts within their reading experience . . perhaps we could build a web space that offers something beyond what is now possible.

Hi j.yarrow,

intresting, very intresting ! Do you know Biomimetics ? Human brain create to understand tangible world. this french / swiss startup try to found a new approach to modelize hyperlink under somes ressources : when a read you i think your approach can be close to that.

So here the goal is to modelize link between different ressource with a milestone on island. each milestone represente a document / web page etc. According by this creators this system is more efficient for our brain.

technically it can possible to adapt this system for http / IPFS :
rest api


josselin, this is an excellent direction to be thinking in! YES. How else can we create an interface for individuals to explore such multi-dimensional space but through generative visualization? Thinking visually is exactly how I began to imagine all of this . . how a system could work to organize streams of inter-subjective expression, these ‘texts’ and the beliefs that people hold in relation to them.

It is a multi-dimensional space. On a higher level, people gravitate together into liquid thought-communities based on the relation of their revisable beliefs. These communities collaborate intellectually to challenge their beliefs by analyzing texts: drawing semantic links between existing texts and generating texts anew by writing analysis on and in-between texts. It is a liquid community process of deliberating on material in order to generate open-source intellectual work. The system requires high levels of coherence and organization. This coherent organization is essential for the sake of human readability and collaborative productivity. How else can someone read a massive collective of expression but for it to be highly processed into synthetic forms? This is exactly our issue with HTTP (as a means for collective intelligence) --it lacks a process of organization, coherence, and synthesis.

Within a mass conversation, people often share very similar thoughts (we see this in social media), and these thoughts should be organized together. These shared sentiments generate something like a micro-thought-community that carries a kind of power and identity of its own. Thought-communities are holarchical because of this kind of differentiating branching, and this differentiation allows for a different kind of deliberation: that of opposition. This opposition, of course, is key, because it is the basis of debate and rhetorical process that generates pedagogy and synthesis.

In order for thought-communities to self-organize, we cannot rely on artificial intelligence --that would be a mistake. This is where democratic moderation comes into play and the idea of reputation which we are already seeing in similar collaborative projects like Colony. By building one’s personal body of work having written consistently and coherently (in a way that clearly resonates a certain perspective within a specific deliberative context) a contributor builds reputation, and also they build reputation by moderating effectively not merely banning this or that post but by contributing to the generative organization of the thought-community. I would imagine that no contribution is ever really banned or deleted; rather it’s just reorganized or dis-integrated, and its contributor may lose reputation/credibility/voice-power by fouling their integrity and/or coherence. By working in a way as to disrupt the deliberative resonances of the conversation (e.g. trolling), one loses reputation, and so their contribution is demoted and plays a lesser role in the development of the conversation --their voice loses power. The space is infinitely revisable, so each individual always has an opportunity to recast or refine their contribution hopefully to adapt to a developing perspective. Because thought-communities are liquid and generative, the revision of a contribution or belief may diminish or enhance the power of that community, but its removal should not break the form that the community has generated.

Notice, that the content of a communities beliefs is not at issue. The system will be as happy to provide space for communities that might consist of members who identify for example as neo-nazis or antifa or socialists or conservatives or progressives or anything in-between or otherwise. It’s not what the community believes or the texts that they employ to situate their beliefs that is important. The system itself is completely open and neutral. It merely provides place for differentiation. It does not attempt to anticipate the form of that differentiation before hand. This is why if I were given the opportunity to contribute a name to this system, I would call it Khôra after Plato’s Timaeus, because it is a platform that provides place for intertextuality. [^Derrida]

What is crucial to the generative process of the thought-community is the coherence of its deliberation. And anyway as I said, thought-communities are not identified by some formal association that gives title but by the content of their beliefs. In this way everyone is provided complete freedom to express their beliefs, however they are not allowed to destroy or disrupt others freedom of speech. They are not provided the opportunity to hide or rally simply behind names or slogans, because they self-organize according to the content. Most importantly, the form propels individuals into a critical pedagogical process of self-critique. When a text, lets say a video documenting a certain brutality, is evoked, it may come as a direct affront to someone’s belief. This is a problem-posing critical pedagogy like Fraire speaks of in Pedagogy of the Oppressed.

[ I will have to pick this up again tomorrow and write what I can about the underlying dimension of texts that the thought-communities have as their material and their product. This is really exciting, because this is where IPFS can really shine! Also this is how we can store all of our intellectual inheritance! All of the philosophy and science throughout time can be stored here, readily available to be evoked and employed in analysis of current events. Also this is where all of our journalism will be provided a permanent home!

Forgive me for leaving a mess. I will clean this up soon. ]

. . . For me, I wanted to know: How can people of like minds gravitate together into thought communities, where everyone is contributing little thoughts, little texts . . In my mind, it appears to me as something like inter-woven streams of beliefs, and each little expression of opinion is like another drop in that stream. This is a nebulous metaphor, but I think it may be essential for creating a way of interfacing with this collective intelligence. Through visualization.

[^Derrida]: _ Khôra is unpacked beautifully by Susanne Ludemann in Politics of Deconstruction

with method loci ( you can create things in this example also :

It’s really funny: What is so challenging about research in general and especially collaborative thinking, in doing what we are doing right now, pointing out and hoping to map all of these interrelated themes is exactly what people need serious help with: collaboration and semantic mapping.

. .
I will check out Method Loci too. And look what these guys are doing: Collaborative Book Authorship! Now we’re talking! I think the movement is happening on both sides: In the intellectual world and obviously here in the tech world. I’m really interested in riding the intersection, building bridges, and assisting cross-pollination and building the movement of decentralization!

I’ve been thinking of which currently available organizational & collaboration tools to work with just to move the collaboration forward in mapping the research . . Maybe we should start a Slack or a Discord, but I really like revisable tools . . Think it over. I don’t want to move away and abandon this thread, but it would be nice to link it to a platform where we can collaborate to sketch and organize ideas, themes, and sources.

Hi j.yarrow,

i see this presentation this year. For community split around the world i think this approach can be intresting : : Asynchronous Decision Making

bad sound sorry :

The Khôra Project:
Manifesto for Collective Intelligence

1 Like