Peirce’s Semiotics and General Intelligence
There is a natural evolution from the ideas that deep learning has empirical revealed to a theory of general intelligence. A common criticism of deep learning is its lack of good theory. Deep learning is like the supercolliders in high energy physics. It reveals the inner behavior of an artificial intuitive process. It reveals to us patterns of what does work.
To build up that theory we must walk back into the ideas of past thinkers. Thinkers who have never seen the empirical evidence. What will they conclude about their ideas if they had been exposed to evidence in deep learning?
Let’s go back into the past to see how it informs the future. How far in the past do I need to go to find an informative theory? Perhaps I could go back enough to the precursors of AI. Back to the 1940s before computers where invented. Back to exploring Nobert Wiener’s cybernetics. After all, Deep Learning is a modern rendition of cybernetics.
But there is a problem here, it is related to the reason why Wiener was disinvited from the infamous Dartmouth conference that coined the term ‘Artificial Intelligence’. In the 1950s with the emergence of computers, there was a belief that artificial LOGICal system was the key to human intelligence. After all, isn’t logic what separates us from the brutes? This unfortunately led to decades of exploration in Artificial Logic that never achieved any semblance of general intelligence. But it was the discovery of Artificial Intuition (Deep Learning) that finally revealed a path to general intelligence.
The problem however of Wiener and is a problem also inherited by Deep Learning is that it treats cognition as a dynamical system. It employs all the tools that we have inherited from centuries of doing physics.
We have today two competing ideas about how the brain works, one is based on formulas describing dynamical systems, both cybernetics and deep learning share this commonality. We also have a discrete computational kind that originates from the 1950s conference on AI (GOFAI).
Furthermore, we have what is known as the ‘symbolic grounding’ problem or the semantic gap. How does language which is discrete in nature achieve semantic grounding? How can we merge the connectionist and symbolist worlds?
We know intuitively that the solution for semantic grounding is related to the solution of general intelligence. But again, what thinkers of the past have thought of this problem? Apparently, there was one thinker, an American thinker who died impoverished in the early 1900s. A thinker who barely published any of his works but left behind an outstanding amount of theory. This American thinker in fact actually conceived of the universal gates used in computers today. His ideas on this were never noticed, until the ideas were reinvented to create computers in 1950s. The same American thinker was credited by Heisenberg himself for the idea of the uncertainty principle in Quantum Mechanics. This thinker was Charles Sanders Peirce.
In my study of Peirce’s work on semiotics, it occurred to me that the common understanding of signs, that is one confined to the triad of icon, index and symbols, is an incomplete understanding of Pierce’s formulation. Here I will use Peirce framework of 10 genuine signs and map them into other common notions of information.
The diagram above is difficult to understand because Pierce used uncommon words to describe characteristics of signs. Qualisign, Sinsign and Legisign can be mapped to the words Tone, Token and Type. Rhematic, Dicent and Argument can be mapped to the more common idea of Term, Proposition and Argument. Each sign is defined in a combination of three kinds of characteristics, there are 10 valid kinds as described by the diagram above. The Tone, Token and Type is associated to the sign itself. Icon, Index and Symbol is related to how as sign is related to the object it refers to. Term, Proposition and Argument is how a sign and its object is related to the interpretant.
Aboutness
Preparation and measurement characterize all that is communication in correspondence to Shannon’s framework. However, when we begin to introduce agents with an intentional stance, we find a new kind of information. This is information aboutness. It goes beyond how information is codified and into the semantics of information. We know this as icons in semiotics and references or pointers in computation. Information aboutness is of value only for intentional agents with memory. These are signs that capture similarity of what has been observed. Micmicry is an example of a procedural form of information aboutness. Information aboutness makes possible the recall of information that previously was encoded in memory.
Aboutness is captured by the icon-index-symbol triad in characterizing signs.
Self
CT does not cover in its exploration of life is that characteristic of living things (or agents) to have what is known as an intentional stance. The intentional stance is a term coined by Daniel Dennett to refer to agents with behavior that are due to having cognitive capabilities. I use the term in the broadest of senses, which also includes the most primitive of cognitive capabilities (i.e., stimulus-response).
Physics has the law of conservation of energy, which is basically an invariant property with respect to time. Analogously, biological agents have the intentional stance of preserving self or the conservation of self (aka survival and replication). Thus, information about self (see: AGI using self-models) is essential for all biological life.
All genuine signs in Peirce’s framework involve an aspect that concerns the intepretant. The interprentant is the self in the context of a sign. In other words, Peirce framework is unusual in the sense that it introduces subjectivity in the the interpretation of signs. There’s more to discuss about this later.
Entropy
Entropy is a measure of information that finds its way into the vocabulary of macroscopic phenomena and in information communication capacity. Both Boltzmann and Shannon (years later) defined the use of entropy (a measure of a kind of information) in different fields but with similar equations. Boltzmann in his pioneering work in Statistical Mechanics defined entropy as a measure of statistical disorder. Shannon was unaware of the similarity of his measure of information with that of Boltzmann’s thermodynamic entropy. However, when searching for a name for his measure, Von Neumann recommended to Shannon that “nobody knows what entropy is, so in a debate you will always have the advantage.” Thus these two separate ideas of measures of information were eternally linked.
Entropy is Term-Icon-Type. That is a characteristic of a sign, its form has a similarity with uncertain, disordered or useless information and it is defined by convention.
Replication
Replication is a key characteristic of life and is made possible via digitalization. Symbols are the embodiment of the digital in Semiotics. Digital information can be identified as Term-Symbol-Type. This differs from Entropy in that it a symbol and not an icon.
Efference Copy
Intentional agents learn the aboutness of information through environmental interaction. That is, learning and model building is achieved through intervention with the world. The expectations that an organism acquires is achieved by testing conjectures and predictions. Biological organisms are known to involve an efference copy or efferent copy. This is information about an organism’s own movement. This information explains why we can’t tickle ourselves or why we rub ourselves when we get hurt. Information about our own movements reduces the sensitivity of our sensors that are caused by our own actions. It allows us to maintain the stability of what we see despite the movement of our heads and eyes.
Efference copies are classified as Term-Index-Token. This is a signal that an organism is aware of that refers to something that is caused by its own actions.
Affordances
The next tier of information is what Gibson would describe as ecological affordances and what semiotics would describe as indexical signs. The value of this information is that it conveys to an intentional agent the possibilities and impossibilities that are available in a context. Constructor Theory revolves around the identification of possibilities and impossibilities of transformational tasks. In the realm of intentional agents, the recognition of possible and impossible potential actions is ideal for the preservation of self (i.e., survival). To learn about affordances requires the development and query of internal mental models of the world. This allows an organism to “see” what is possible or impossible without actual interaction with the world.
Affordances are Proposition-Index-Type. The reason it is type is that affordances are innate index information that is shared by members of its species. An organism is born with innate cognitive capabilities that allows it to recognize affordances that its organism has evolved to be useful.
Self Reference
Let’s now delve deeper into the concept of information aboutness and usefulness with respect to self-models. That is, what can we say about information about the self that is referential and ultimately useful? Self-referential information involves higher information abstractions such as autonomy, introspection, and reflection. Brian Cantwell-Smith describes these kinds of information are based on notions of the self as unity, self as a complicated system and self as an independent agent.
Autonomy is information on the self that recognizes its self-direction and agency. Information about autonomy is under Proposition-Index-Token sign. It is not of a certain type because these are signs that are of the individual and not the collective. The proposition is about itself and the signs of the environment.
Introspection is observations of a self’s cognitive processes. It allows reasoning about our own thoughts. Information about introspection is where the object of the sign is the thoughts of the interpreter. It is classified the same as autonomy but the proposition is about itself and the signs generated by itself.
Reflection is a detached perspective of a self’s cognitive process and reasoning from the perspective beyond the self. It is classified the same as autonomy but the proposition is about itself and the signs generated by itself and the environment.
There is ever-increasing complexity with different kinds of information required in the self-referential exploration of self. The object in that the sign is about is the same as that of the interpreter.
Coordination
But we are not yet done with our emerging ontology of information. No man is an island, and no organism is independent of its ecology. In the Pi-calculus of distributed computation, information aboutness is known as a channel or vocative name. The Pi-calculus employs aboutness information as information coordination. Human civilization employs information coordination as a means of resource allocation. Money is an example of this kind of information. Money is essentially information about obligations and ultimately related to trust. But what is trust from the perspective of information? Trust is what I would fall into the same category known as information usefulness. However, trust is information about the self that is conveyed in interaction and communication with other-selves.
A coordination sign or a vocative name is a Term-Index-Type sign. Resource signs like Money (a decentralized coordination mechanism) is like an affordance, however with a sign based on convention. That is, a Proposition-Symbol-Type.
Conversational Shared Experience
Sharing information context is an essential component for human cognitive development. Human eyes, specifically the white of our eyes, allow others to recognize what we are attending to. Human eyes have also evolved to understand the subtle changes in the color of our faces. These evolved capabilities reveal the importance of shared contextual information. A human’s ability to share their own experiences even without verbal language is an essential tool that accelerates cognitive development. One can even make the general assertion that the essence of being human is in the activity of sharing the human experience. One can therefore not comprehend human-compete intelligence without having a level of understanding of human experiential sharing. In art, there is a concept of the “beholder’s share.” That is, beauty is in the eye of the beholder, what is meant is that an interpretation of art is performed by its perceiver. However, good art is the kind of art where the artist is able to share an experience with its beholder. Da Vinci’s Mona Lisa’s smile, as an example, is sufficiently ambiguous such that it can morph to the preference of the beholder.
I’ve written several times that to achieve AGI, that achieving conversational cognition is required. In my capability model, conversational cognition is at the highest steps. Conversational cognition makes possible cultural evolution. The notion in psychology of dual-hereditary theories proposes that human cognitive development is both biological as well as cultural. A recent position paper from DeepMind addresses specifically this perspective in“Emergence of Innovation from Social Interaction”. What’s interesting though is that humans have been able to converse for thousands of years with barely any technological progress. That is, the same tendency for bureaucratic organization exits also in the development of human society and cognition.
Human conversations involve Argument-Symbol-Type signs. Conversations in general are defined as the exchange of signs.
Memes
The scaling of shared human experience is what Richard Dawkins described as memes. Memes are the basis of an evolutionary model for cultural information transfer. Analogous to a gene, the meme was conceived as a “unit of culture” (an idea, belief, pattern of behavior, etc.). Memes are “hosted” in the minds of many individuals, and reproduce themselves by transferring from the mind of a person to the mind of another. It can be regarded as an idea-replicator that replicates itself by influencing the adoption of new beliefs by many individuals. Analogous to genetics, the success of a meme is dependent on the ubiquitous use and replication of its host. Daniel Dennett proposes that the development of human languages is a consequence of the spreading of memes.
Memes, analogous to biological cells and viruses, have a two-level structure(i.e. A replicator-vehicle logic). Douglas Hofstadter in Metamagical Themas describes the structure of an effective meme:
System X:
-Begin
X1: Anyone who does not believe System X will burn in hell,
X2: It is your duty to save others from suffering.-End.
If you believed in System X, you would attempt to save others from hell by convincing them that System X is true. Thus System X has an implicit `hook’ that follows from its two explicit sentences, and so System X is a self- replicating idea system.
There is a bait that conceals the hook that allows a meme to propagate. Let’s frame this structure using the replicator-vehicle logic in CT. The vehicle, unlike in biology, is not physical but somewhat abstract. If we are to understand a vehicle as the mechanism that ensures sustainability, then it is the hook that is relevant here. The hook is the deception of information usefulness in the meme that encourages its own propagation. The concept of “unknown knowns” or willful ignorance has utility in the propagation of memes (and thus knowledge). It is indeed interesting that deception is a characteristic of information that originates not just in memes but in more primitive biological organisms. Camouflage and viruses are examples of deceptive strategies in biology. The physical replicator here would be the host mind that accepts the truthfulness and thus the utility of the original meme.
Memes therefore involve two kinds of signs. The bait is a Proposition-Index-Type sign (the same as affordance, one where the index refers to the interpreting self) and the hook is a Term-Symbol-Type sign (the same as replication).
Analogies
Douglas Hofstadter has proposed that analogies are the fuel and fire of thinking. Information affordances are what makes analogies possibles. These are indexical information that binds concepts together and permits the combination of new concepts. This is what I would call “ingenuity” and it is entirely lacking in current Deep Learning models.
The consequence of an analogy is the creation of signs with an Argument aspect, that is, one kind of genuine sign in Peirce framework. It is however involves the combination of Proposition based signs to form the Argument. Conventional analogies are formed in the identification of similarity (i.e. icons) or the identification of two similar indexes and thus leading to a more subtle connection between two concepts. Analogies do not work at the symbolic layer unless symbols refer to concepts which only happens when symbols are grounded. Said differently, the signs have meaning and that meaning either reveals a similarity or an index.
Hofstadter’s analogy making is a process that involves semiotics (aka semiosis). That is, a process that involves the interpretation of signs to create new signs. Can Peirce’s refined distinction of signs lend us a better understanding of general intelligence?
