Chapter 15
Observing Collective Intelligence
The Hypercortex coordinated by the semantic sphere was conceived from the outset as an instrument for scientific observation of the collective intelligence of creative conversations. What applies to creative conversations in the digital medium applies equally to any system of symbolic cognition, whether real or fictitious, personal or social. The question at issue in the discussion that concludes this first volume is none other than the possibility of scientific self-observation of the mind in general: what type of image will the observation of collective intelligence generate in the mirror of the Hypercortex?
As shown in Figure 15.1, the main purpose of this chapter is to explain the structure of the reflection of the Hypercortex in the Cortex, and vice versa. But before I come to that, I will review the main stages in the intellectual journey we have taken in Part 2 of this book.
15.1. The semantic sphere as a mirror of concepts
15.1.1. Reflecting the world of ideas
In Part 1 of this book, I emphasized the fact that human cognition was not only conscious in the sense of having a subjective capacity to perceive and feel, but also in the sense of having a reflexivity capable of representing its own categories and mental operations to itself. Reflexivity is fundamental to human symbolic cognition. Human thought is not only part of nature but it also offers this nature an organized representation reflecting its inexhaustible variety – including the thought that reflects it. In other words, this thinking mirror has the capacity to accommodate a cosmos, because only for human intelligence can there be a cosmos – rather than an indescribable chaos or a limited space of interactions with a closed environment. Symbolic cognition produces as many cosmoses as it does cultures. It should be understood that the metaphor of the mirror and reflection in no way implies the solid, objective, external existence of the world as it is thought of by any one cognitive system. Cognition in general, and symbolic cognition in particular, is necessarily based on a creative interpretation informed by a cultural history. This interpretation navigates between two reefs: total arbitrariness (not everything is permitted, not all forms of cognition are “viable”) and the illusion of absolute truth (according to which any different interpretation is purely and simply an error).
Figure 15.1. Position of Chapter 15 on the conceptual map
The first – and most difficult – problem I had to solve was to characterize the basic space in which the processes of human symbolic cognition take place. Through what self-examination by the mind could this radical foundation be found? How can we think rigorously about the container, the all-enveloping “place” of the universe of cosmoses? Where can we seek the formula for the cosmic mirror?
I started with the observation that human symbolic cognition is marked by a general capacity to manipulate and determine intellectual essences, or formal symbols1. Once the unity of the symbolic faculty was recognized, I hypothesized that there had to be a corresponding universe of concepts or signifieds, whose unity was based on it being the object of the symbolic faculty, since concepts can only be manipulated through meaningful forms. In addition, just as symbols are only manipulated for purposes of manipulation of concepts, the manipulation of concepts is in turn only a means for the effective and affective manipulation of data, because it is only when concepts categorize sensory data, or percepts, that the world of ideas emerges. Then affects circulate in memory and mobilize intelligence.
IEML can be considered a “semantic machine”, an automatic writing that makes the conceptual addressing of the world of ideas scientifically possible. This machine controls a universal system of coordinates: the semantic sphere. The shortest description I can give of the semantic sphere is that it is a calculable topological structure in which each node functions as the identification code for a single concept and each connection between nodes represents an explicit semantic relationship. The concepts and their relationships are expressed simultaneously in all natural languages.
Besides its monadological unity, two properties were absolutely essential to the IEML semantic machine: first, the number of distinct nodes had to be practically unlimited and, second, it had to be able to automatically perform a maximum number of calculations on concepts and their relationships, using the IEML codes of those nodes. Clearly, the second condition was the more difficult one to meet. We could think of the relationship between concepts and the IEML codes that represent them as analogous to the relationship between numbers and the number system used to write them. Ideally, although (all) concepts are not numbers2, we should be able to perform automatic operations on concepts and their relationships using their semantic codes as easily as we carry out automatic operations on numbers using their binary notation. As we saw earlier in the book3, it is by a similar property that the philosopher and mathematician Leibniz defined his “universal characteristic”. However – unlike Leibniz, but taking his experience into account – I had to design a system for encoding concepts that was distinct from the one that works so well for the notation of numbers. That is why IEML syntax is based on the structure of natural languages, but without their irregularity.
With regard to the calculability of the relationships between the nodes of the IEML system of semantic coordinates, we may think of the correspondence between the geometry of three-dimensional space and algebraic calculus4: there is a correspondence of the same kind between the IEML semantic topology and algebraic calculus (in the most abstract sense of the term). It should be kept in mind that the huge hypercomplex graph of the semantic sphere corresponds to the algebra of a system of symmetric transformations. Without this property of transparency in the calculation of its basic topology, without the possibility of functional translation of the movements and transformations of its concepts, without the conservation of invariants across its variations, human symbolic cognition could not be modeled – and thus conceived – as an object of science5. In addition, without this property of computability, the immense automatic calculating power that is now available to us in the digital medium could not be optimally used to explore our new cosmos: a unique and infinite nature that includes a reflection of the human collective intelligence6.
15.1.2. The IEML semantic sphere
The nodes of the semantic sphere are called USLs. USLs are in fact all the different standard texts that can be produced mechanically using IEML syntax. A USL designates a collection of sets of sequences of six elementary symbols. The space of the IEML texts is a transformation group, because all its elements (USLs) are themselves sets of elements produced by the same combinatory mechanism. All set operations (union, intersection, symmetric difference, etc.) can be performed, inverted and combined on USLs. Operations of concatenation (triplication) and cuts can also be automated on the sequences of symbols. In addition, all the functions that transform one USL into another USL (and that therefore lead from one node to another of the great network of the semantic sphere) can be inverted and/or combined to form more complex calculable functions.
Staying at the level of linguistic utterances, the signified of an IEML text includes not only its translations into natural languages but also the set of its explicit semantic relationships (translated into natural languages) to the other texts. The semantic sphere contains all the paradigmatic and syntagmatic connections among texts7. The calculable operations carried out on sets of sequences (USLs) are at the same time operations carried out on the meanings those sets of sequences represent (concepts). The main idea to remember is that any path in the hypertextual space of the connections among USLs can be represented by a calculable function and that this function can have semantic relevance.
As I have frequently pointed out, while the data of the Web are opaquely addressed with URLs – which paved the way for a universal logical memory – the metadata of the IEML semantic sphere are transparently addressed with USLs – which paves the way for a universal hermeneutic memory. In fact, it is only by permitting the computation of data on the basis of their meanings (encoded as USLs) that the hermeneutic memory of the Hypercortex can become operational. The calculability of semantic metadata is not an end in itself: the practical goal is to bring the multimedia data of the Web into the world of calculability opened up by the IEML semantic sphere.
To categorize data, it was necessary to have a metalanguage that would express and differentiate meanings with precisely the same power as a natural language. The construction of a metalanguage that would meet the double requirement of computability of its semantics and unlimited openness of its expressive capacities has been no easy task. The problem was not so much designing a regular, and therefore calculable, language; there are already many examples in mathematics and computer science. The main problem was the requirement of a correspondence, or isomorphy, between the structure of this regular language (IEML) and the basic structure of the natural languages that are normally used to express complexities of meaning (but that do so irregularly). It is precisely this isomorphy between the regular language and natural languages – which will be studied in detail in Volume Two – that now makes the automation of the linguistic function possible. In other words, it enables the mechanical transformation between (i) any valid expression in IEML (a USL) and (ii) a circuit explicating the meaning of this expression and the semantic, grammatical and intertextual relationships of this USL with other USLs. I reiterate once again that in order to be readable, this explication of meaning and semantic relationships internal and external to the USL uses the words of a natural language (French, English, Arabic, Hebrew, Mandarin, etc.) chosen by the user.
The IEML semantic sphere thus functions as a system for encoding meaning that is designed to make the greatest possible number of operations on concepts and their semantic relationships automatically calculable. I note, finally, that all this is based, in practice, on the existence of a matrix of semantic circuits with predefined meanings (the STAR dictionary) that is used for translating IEML texts into semantic circuits tagged in natural languages, and vice versa.
Far from being closed, autarkic, opaque 8, USLs are points of view open to all other points of view, virtual centers where multitudes of semantic perspectives intersect. The USL is thus not only a code or a text. It is also the nucleus of a monad9 whose radiating rhizomes10 are generated by all the paradigmatic and syntagmatic functions that crisscross with it to interweave the semantic sphere.
The original basic terrain of symbolic manipulation is one. It is a semantic continuum in which languages translate each other as best they can; in which the metaphors, correspondences and resonances of literary traditions can be woven; in which human thought can carry its models from one discipline to another and make connections across registers, genres, traditions, paradigms and epistemes.
15.2. The structure of the cognitive image
If the Hypercortex is a mirror, it is a mirror of cognitive functions rather than material bodies. The first thing this mirror had to be able to reflect was the universe of concepts characteristic of human symbolic cognition. Now we have to move from the reflection of concepts to the reflection of the dynamics of relationships among ideas.
15.2.1. The integration of data into calculable cognitive models
The Hypercortex combines:
– a set of distributed data in a logical memory, the data of the Web, addressed with URLs;
– a set of distributed metadata in the IEML semantic sphere, which are addressed with USLs.
Using these data and metadata, collective interpretation games assemble ideas and connect them in noumenal circuits. As we have seen, ideas are represented in the Hypercortex by semantic information units (USL, C, URL) in which the URL – the address of the data – formalizes the percept, the semantic current C formalizes the affect and the USL formalizes the concept. The collective interpretation games that produce the information units are made up of two types of functions, which we will review using Figures 13.3, 9.3 and 7.5:
– Functions of perception construct semantic information units from data flows in real time. They can be divided into functions of categorization, which link USLs to URLs, and functions of production of current, which inject a semantic current C into the semantic circuit corresponding to the USL. This current formalizes the affective or emotional dimension of cognition.
– Functions of thought link semantic information units (data categorized and evaluated), creating narratives, theories and models that transform information into knowledge.
Collective interpretation games thus integrate data into models of cognitive systems in which both the qualitative dimensions (the circuits of USLs) and the quantitative dimensions (polarized intensive values of semantic current) belong to calculable transformation groups.
15.2.2. The ternary structure of the cognitive image S/B/T
15.2.2.1. The ternary structure of the semantic information unit
The structure of semantic information units (see Figure 11.5) imposes a ternary structure on the images of the cortical functions in the mirror of the Hypercortex. We have seen that these units (USL, C, URL) are made up of three parts: (i) USLs, encoded addresses of concepts in the IEML semantic sphere; (ii) URLs, Web addresses of multimedia data categorized by USLs; and (iii) C, the semantic current propagated in the circuits defined by USLs. The semantic current indicates the intensity (a cardinal number) and the polarity (an ordinal number11) of the energy that links the data (URLs) to the metadata (USLs).
The referential data – the URL – can be associated with the thing T of the IEML ternary dialectic (corresponding to the referent of the semiotic triad)12. The USL can be associated with the sign S of the IEML ternary dialectic, and the signifier of the semiotic triad. The semantic current, the binding energy that connects the sign to the referent, can be associated with the being B of the IEML ternary dialectic, and the interpreter of the semiotic triad.
15.2.2.2. The topological image: semantic circuits S
The USL designates a semantic circuit, i.e. a topological form that stands out against the background of the IEML semantic sphere. A cognitive system can manipulate a large number of semantic information units. The set of USLs of the information units manipulated by the cognitive system forms the imprint or topological image of this cognitive system on the semantic sphere. The topological image can be associated with the semantic shadow or profile of the cognitive system. It delimits its universe of discourse in the form of a hypercomplex fractaloid circuit, each channel and junction of which has a determinate meaning readable in natural languages. This topological image, or semantic profile, can be transformed over time. In Figure 15.2, the topological image corresponds to the two poles on the left (S).
Figure 15.2. Computing image of a cognitive system in the IEML-based Hypercortex
15.2.2.3. The energy image: semantic currents B
The semantic currents of the set of information units manipulated by a cognitive system from its energy image. These currents follow the circuits that define the universe of discourse of the cognitive system. Just as the topological image traces a figure on the background of the semantic sphere, the energy image traces cycles, oscillations of intensities and values from which a dynamic figure emerges against the background of the topological image. The energy image of a cognitive system represents its intensive (force) and affective (attractive or repulsive) dimensions, with everything controlled by an axiology (criteria for measurement and rules of evaluation). The dynamics of distribution of the value and intensity of the current – the energy images – “animate” the topological images from within. In Figure 15.2, the energy image corresponds to the two poles in the center (B).
15.2.2.4. The referential image: multimedia data T
The set of multimedia data (including fictional or imagined data) addressed by the information units manipulated by a cognitive system constitutes its reference corpus. When this corpus is projected on the topological/energy image of the cognitive system, it becomes the referential image of the cognitive system. Like the topological and energy images, the referential image is dynamic. It fills the representation of a cognitive system with sensory texture and documentary materiality. In Figure 15.2, the referential image corresponds to the two poles on the right (T).
15.2.3. The dual structure of the cognitive image U/A
The source of the ternary structure of the cognitive image is the composition of the semantic information unit (sign–USL/being–C/thing–URL). Its dual structure, on the other hand, comes from the distinction between the functions of perception (actual) and the functions of thought (virtual) of collective interpretation games.
The application of functions of perception to input data flows in real time produces phenomenal ideas, and these ideas together make up a phenomenal image of the cognitive system. The phenomenal image varies with the data and evolves with the refinement of the functions of perception.
In contrast, the application of functions of thought to phenomenal and noumenal ideas produces a noumenal reflection of the cognitive system. The information units of the noumenal image have exactly the same composition as those of the phenomenal image. The only difference is the fact that in the phenomenal image, it is actual input data that go into the production of the information units, while in the noumenal image, the multimedia data mobilized by the semantic circulations are remembered or simulated (they are virtual, i.e. imagined). Narration, theory and thought (modeled by functions of association) imagine relationships among ideas, whether these ideas are noumenal or phenomenal. The noumenal image varies with the phenomenal image and evolves with the refinement of the functions of thought. In Figure 15.2, the phenomenal image is in the area of the actual A (the bottom half), while the noumenal image is in the area of the virtual U (the top half).
We need to be clear on the concept of phenomenon. Etymologically, the word phenomenon comes from a Greek verb meaning appear. In a sense, all ideas and all connections among ideas are phenomena of the mind, and for human beings there are no phenomena other than what appears in the mind. How could it be otherwise? When I say that all ideas are phenomena, the word phenomenon is meant in an absolute sense. I do, nevertheless, make a distinction between phenomena and noumena. Phenomena are our immediate interpretations of empirical data or percepts that arise over sequential time. Noumena are relationships among phenomena that we establish on the basis of narrative or theoretical patterns (these are chains of thoughts that organize or contextualize perceptions and other thoughts). When I contrast phenomena and noumena, the word phenomenon is meant in a relative sense. This opposition between empirical phenomena (in the actual realm) and theoretical or fabulating thought (in the virtual realm) is traditional, and it is very useful in practice. That is why I am using it. This should not, however, hide the fact that, on one hand, even empirical phenomena are interpretations – since they are categorized and evaluated – and on the other hand, even noumena are phenomena (in the absolute sense) – since they arise in the mind: like all ideas and connections among ideas, they result from cognitive operations.
15.3. The two eyes of reflexive observation
I have described the structure of the image of a cognitive system as it results from the combination of semantic information units (which formalize ideas) and the hermeneutic functions that produce these information units. This image corresponds to the syntactic or computing dimension of the IEML model of the mind. In Figure 15.1, at the beginning of the chapter, this is represented as an upward arrow going from the Hypercortex to the Cortex. Creative conversations can organize the image of their own cognitive functioning as they wish, so that it reflects their universe of discourse. Let us now suppose that the universe of discourse of a creative conversation is focused on the theme of human development, as in Figure 15.3. In this case, the image corresponds to the specifically semantic or humanities dimension of our model of social cognition. In Figure 15.1, this is represented as a downward arrow going from the Cortex to the Hypercortex.
The two types of image, computer and humanistic, will be able to be explored interactively, with the possibility of zooming in on details or obtaining composite views. Even better, the computer image and the humanities image contain each other. As we can see in Figures 15.2 and 15.3, the six poles are represented by hexagons. In Figure 15.2, each of the six hexagons represents Figure 15.3, but analyzed from six different points of view: phenomenal circuits, noumenal circuits, phenomenal currents, noumenal currents, phenomenal corpus and noumenal corpus. Symmetrically, in Figure 15.3 each of the six hexagons represents Figure 15.2, broken down into six images: epistemic capital, ethical capital, practical capital, biophysical capital, social capital and communication capital.
Figure 15.3. Humanities image of a cognitive system in an IEML-based Hypercortex
For a creative conversation, the image of its cognition in the Hypercortex functions as a tool for representation of both the self (since it depicts the self’s own cognitive processes) and its environment (since it categorizes, measures and contextualizes the data it processes). Again, this cognitive image is dynamic, first because it is transformed according to the input data flows and second because the functions that determine it can be modified and evolve over time. We can now imagine that a good part of the work of conception and refinement of collective interpretation games will consist of organizing resonances and coherences between the virtual and actual (or noumenal and phenomenal) dimensions of cognitive systems, and exploring different forms of flexible, productive alignment of their dimensions as sign, being and thing (or their topological, energy and referential dimensions).
The nature of the mind is one, but it is also explorable, open and infinitely complex. If we want to produce scientific images of the mind, we have to construct an observation instrument that can channel this inexhaustible complexity. In other words: to reflect a hypercomplex universe, we need a hypercomplex mirror of cosmic dimensions. It is precisely this role of a mirror adapted to its object that the Hypercortex will play. In Kantian language, the Hypercortex resembles the “transcendental subject” of human knowledge. The images of collective intelligence reflected back by the symmetrical surface of the IEML semantic sphere must, however, be conceived in an open, plural, emergent and fractal way. As complex as it is, the scientific image of a cognitive system in the Hypercortex will necessarily be limited. A particular cognitive system will only be projected on a subset of the semantic sphere, and will only organize part of the available data. Its finite representation will clearly show that it is contingent, that it is the result of a (more or less controlled) choice from among an infinity of other possible cognitive systems and myriad of other real cognitive systems, all organized differently. The existence of a common framework of calculable modeling will thus not preclude this framework being used differently at each level and according to each distinct point of view. The accommodating unity of the cosmic mirror will reflect back an indefinitely open multiplicity of cognitive images.
The Hypercortex will serve as a scientific observatory, enabling cognitive systems, whether individual or collective, to empirically observe and compare their own processes of knowledge production and management. Creative conversations will thus be able to use the Hypercortex as a dynamic medium for the modeling and self-observation of their collective intelligence. On the other side of the mirror, researchers and engineers will organize this observatory using concrete technical and scientific methods. They will be able to assemble, dismantle, dissect and criticize the perfectible mechanisms of the Hypercortex.
Now that the cultural purpose and the scientific and technical functions of the semantic sphere have been fully explained and justified, Volume 2 of this work will be devoted to the linguistic and mathematical description of IEML from a practical engineering perspective.
1 See Chapter 3 and Figure 11.4 and the related discussion, below.
2 Numbers are a special case of concepts, and not the reverse.
3 See section 10.1.
4 Philosopher and mathematician René Descartes is usually credited with the invention of algebraic geometry.
5 In the words of Galileo, one of the founders of modern science, “The great book of nature is written in the language of mathematics”. For a historical and epistemological study on this point, see the interesting book by Georges Lochak, La Géométrisation de la Physique [LOC 1994]. See also Jean-Marc Lévy-Leblond’s comments in note 8 in Chapter 2.
7 On syntagmatic and paradigmatic connections, see section 1.3.1, and Volume 2. I am talking here about meaning at the level of language and utterances. For meaning in the context of enunciation and narration, see Chapter 13 and the general conclusion of Volume 2.
8 Unlike the URLs of the web of data.
9 On Leibniz’s monads, see his little masterpiece, The Monadology [LEI 1714a].
10 I recall that the concept of the rhizome was developed philosophically by Gilles Deleuze and Félix Guattari in the introduction to A Thousand Plateaus [DEL 1987b]. As we will see in Volume 2, the circuits of the semantic sphere are rhizomatic graphs.
11 This ordinal number represents a value on a scale between a positive pole and a negative pole.