Research Goals:

Go here. [or PDF]


The Cognitive Agent: Overcoming informational limits


Vakarelov, O., forthcoming, Adaptive Behavior

The paper provides an answer to the question: What is the function of cognition? By answering this question it becomes possible to investigate what are the simplest cognitive systems. It addresses the question by treating cognition as a solution to a design problem. It defines a nested sequence of design problems: (1) How can a system persist? (2) How can a system affect its environment to improve its persistence? (3) How can a system utilize better information from the environment to select better actions? And, (4) How can a system reduce its inherent informational limitations to achieve more successful behavior? This provides a corresponding nested sequence of system classes: (1) autonomous systems, (2) (re)active autonomous systems, (3) informationally-controlled autonomous systems (autonomous agents), and (4) cognitive systems.

The paper provides the following characterization of cognition: The cognitive system is the set of mechanisms of an autonomous agent that: (1) allow increase of the correlation and integration between the environment and the information system of the agent, (2) so that the agent can improve the selection of actions and thereby produce more successful behavior.

Finally, it shows that common cognitive capacities satisfy the characterization: learning, memory, representation, decision-making, reasoning, attention and communication.


The Information Medium


Vakarelov, O., forthcoming, Philosophy & Technology

The paper offers the foundations of the theory of information media. Information media are dynamical systems with additional macro-structure of information-carrying states and information preserving transformations. The paper also defines the notion of information media network, as a system of information media connected by information transformations. It is demonstrated that many standard examples of information containing and processing systems are captured by the general notion of information medium. The paper uses the theory (and informal discussion) of information media to motivate a structural approach to the information in media. The idea is that the notion of information transformation should be regarded as more primitive than the notion of informational state. Thus in information systems, especially in the context of information technology, information is secondary, information transformation is primary.

Pre-cognitive Semantic Information


Vakarelov, O., 2010, Knowledge, Technology and Policy, 23

This paper addresses one of the fundamental problems of the philosophy of information: How does semantic information emerge within the underlying dynamics of the world? --- dynamical semantic information problem. It suggests that the canonical approach to semantic information that defines data before meaning and meaning before use is inadequate for pre-cognitive information media. Instead, we should follow a pragmatic approach to information where one defines the notion of information system as a special kind of purposeful system emerging within the underlying dynamics of the world, and define semantic information as the currency of the system. In this way, systems operating with semantic information can be viewed as patterns in the dynamics – semantic information is a dynamical system phenomenon of highly organized systems. In the simplest information systems the syntax, semantics and pragmatics of the information medium are co-defined. It proposes a new more general theory of information semantics that focuses on the interface role of the information states in the information system – the interface theory of meaning. Finally, with the new framework it addresses the debate between weakly semantic and strongly semantic accounts of information, siding with the strongly semantic view because the pragmatic account developed here is a better generalization of it.


Work in progress:

The Historical Necessity of Life for Cognition


I argue that life --- metabolic systems satisfying the property of autopoiesis --- is structurally historically necessary for cognition. That is: (1) the first place where cognition is needed and emerges naturally is within living (metabolic) agents; (2) non-metabolic cognitive agents can emerge only as artifacts of other cognitive agents, the earliest of which must be living. In order to make this precise I define the notion of priority hierarchy of complexity, which tracks the construction dependencies of structure formation. This last concept is based on a theory of propagation for order due to Stuart Kauffman.



Doctoral Dissertation Summary:

Title: General Situated Cognition

[PDF] Chapter by chapter detailed version of the summary.

I attempt to develop a theory of General Situated Cognition. The project has several goals: (1) to unify existing foundational approaches to cognition, such as the classical symbolic approach, connectionism, dynamism, as well as minimalist A-life based approaches; (2) to characterize the function of cognition in autonomous systems and suggest a general (meta-)framework for cognitive architecture; (3) to provide sufficiency conditions and classification of representation mechanisms in natural and artificial cognitive systems. I argue that cognition is essentially a situated/embodied phenomenon; however it emerges when systems can interact with the environment partly through informational mechanisms. To justify the insistence of informational description of cognition, I provide an account of information and informational mechanisms that is based on dynamics. Thus, informational processes are seen as emerging on top of and interacting with dynamical processes, accommodating the useful ideas of the dynamisist program, while allowing the functional models of connectionism and symbolic processing. I argue that the function of cognition is overcoming informational limitations of embodied autonomous agents in order to improve the success of their interactions with the environment. I propose that cognitive mechanisms be modeled as informational medium networks, where informational media integrate informational and causal/dynamical properties. The informational medium network model allows the definition of a rich family of concepts, including information processing operations, information management operations, etc. It allows a multi-dimensional space of evaluable characteristics for informational media which allows apparently “engineering” considerations to have central theoretic significance for cognitive architecture. Both connectionist and symbolic architectures can be modeled with the informational medium network framework, but as degenerate special cases. The true richness of the framework is exhibited when considerations of network topology are included. For example, one can make precise distinctions between distributed and modular architectures. With the help of the framework I also provide accounts of different kinds of representational systems, including a solution to the “symbol grounding problem”, and hints for naturalizing “intentionality”.



Master thesis:

Title: Accessible Domains from a Category Theoretic Perspective


The goal of the thesis is to develop an abstract, structural, category-theoretic presentation of accessible domains. Accessible domains are (freely generated) inductively defined structures on which there are an inductive proof principle and a recursion principle, guaranteed by the (free) construction. Examples of accessible domains include: the natural numbers, constructive ordinals, initial segments of ZFC; regular ordinals and the complete cumulative hierarchy. Accessible domains are interesting because of their intuitive, constructive method of conception -- the intuition of generating from below -- and as such, can be used, among other things, as concrete mathematical structures to which the consistency of other, nonconstructive structures can be reduced. The category theoretic formulation of accessible domain is interesting because it demonstrates that it is possible to obtain a formulation of the notion in a purely structural fashion: accessible domains are initial algebras of endofunctors in appropriate categories.



Research Goals:

Broad Research Program

My philosophical research is aimed at understanding the natural phenomena of science and mathematics, and the interaction between the two, as cognitive practices. A practice within a community of cognitive agents is cognitive if details of the cognitive architecture of the agents place a major constraint on the organization of the practice, and the practice serves to extend the cognitive capacities of the agents/community. I take the fundamental problem of the project to be: how physical systems – e.g., organisms, including human agents – can extract information about the world from within the world and utilize the information in their constructive interactions with the world.

A philosophical account of science and mathematics as cognitive practices, and a general account of embodied epistemic agents, requires several consecutive steps: (1) an account of information focusing on the emergence of information using systems; (2) an account of the phenomenon of cognition – an account of (a) how cognitive systems emerge as systems in the world, (b) how they can extract and utilize information, and (c) how they can organize the information into systematic behavior-guiding models of the world; (3) an account of the emergence and nature of modeling and representational systems, how such systems can be integrated in active cognitive agents, and how and what kinds of systems can be used for the generation of abstract scientific or mathematical models; (4) an account of how abstract models interact and change, and how they facilitate the production of technological artifacts, including scientific instruments, which conversely facilitate the changes of the models.

Dissertation Description

My current work addresses the first two steps of the program. In my dissertation titled “General Situated Cognition” I develop an account of pre-cognitive semantic information based on pragmatics, and use it as a base for offering a general account of cognitive systems as complex systems emerging within and operating in a complex environment. The dissertation is based on four papers that achieve four distinct tasks: (1) offering a pragmatic account of information systems; (2) describing a general minimal function of cognition that an information system must satisfy to be cognitive; (3) offering a general theory of information media; and (4) developing an architectural framework for modeling cognitive systems based on networks of information media.
(1) The concept of information plays a central role in my account of cognition. I take a pragmatic approach to semantic information, inspired by a work of Nauta from the 1960s. The main idea is to define a notion of information system, as an open physical system interacting with an environment, and define information as the “currency” of the information system. Special effort is made to demonstrate that the notion of information system can, in principle, be defined within the framework of dynamical systems theory, without any semantic or intentional notions. I offer a theory of meaning – the interface theory of meaning – where the content of an information state is specified by the interface role between the parts of the environment to which the medium is correlated, and the control mechanisms of the information system. This conception of meaning generalizes many different theories of meaning/content. Particularly, it collapses the externalist and internalist approaches to content, demonstrating that both offer viable theories of content, depending on the role of the medium. This theory is presented in the paper “Pre-cognitive Semantic Information” published in a special issue of Knowledge, Technology & Policy. 

(2) I analyze the question of the function of cognition. I ask: what is the design problem that cognition solves in a system; and for that matter, what systems need a solution to the design problem in the first place? I specify the design problem as follows: I observe that all natural information systems (which I call agents) that interact with the environment by extracting information and utilizing the information to control more effectively their behavior are severely informationally deprived. The complexity of the environment is vastly too great for an agent to cope with directly. This demands that the agent possesses specialized internal mechanisms for a more effective use of information to compensate for the informational limitation. I claim that those mechanisms make up the cognitive system of the agent. I justify this characterization by demonstrating that core cognitive capacities – learning, memory, feature detection, representation, and reasoning, all broadly construed – can be viewed as strategies for a solution to the design problem. The problem of describing the function of cognition is the subject of a paper under review titled “The Cognitive Agent: Overcoming Informational Limits”.

(3) I offer a framework for investigating cognitive architectures based on the theory of information media networks. An informational medium is a dynamical system with sets of dynamical states interpreted as information-carrying states and a collection of information preserving transformations. An informational network is a collection of information media with information processing operations on each medium, and information management transformations among the media. By insisting that information media are dynamical systems, we can model information systems both at the functional level of information and at the dynamical/causal level of a system interacting with the world. This allows us to model the interactions between information and physical dynamics. The theory of information media is presented in a paper titled “The information medium” forthcoming in Philosophy & Technology. The paper also addresses an independent problem in the philosophy of information related to the proper ontological place of (non-semantic) information. It endorses a structural principle of information whereby information transformations are more basic, and information is defined as invariance under information transformations.

(4) Information media networks offer a powerful tool for modeling cognitive systems. The architecture is general enough to capture as special cases both symbol processing and connectionist cognitive architectures. The dynamical root of information media accommodate some of the insights of the dynamicist approach to cognition – the approach insisting that cognitive systems are best (or only) modeled with the machinery of dynamical systems theory – while interfacing the dynamical processes with (the sub-class) of informational processes. It also naturally encompasses modeling distributed and modular cognitive architectures. The architecture is developed in a paper titled “Information Networks: A Meta-architecture for Situated Cognition”.

Future Research Goals

The programmatic nature of my research implies that I cannot exhaust the topic in a length of a dissertation. Many problems at stages 1 and 2 are still left open. They will occupy significant portion of my future research.  One problem that I am working on currently in a paper, The Historical Necessity of Life for Cognition, explores the connection between the phenomena of life and cognition. I argue that if we view cognition as a phenomenon of complex systems, as I think we should, then life is a prerequisite for the existence of the first cognitive systems. However, complex cognition may produce artifacts that are cognitive but are not living. Thus life is historically necessary for cognition, but it is not necessary for a cognitive system to be living. As similar open problems are resolved, I plan, in the next several years, to extend the dissertation into a book developing a general theory of cognition.

My longer term research goals will focus on modeling the nature of scientific and mathematical practice. I view scientific model/theory construction as an extension of cognition. In cognition, models play only an intermediary control role in the interaction of the organism with the environment. Similarly, in science, model building and modification activity play an intermediary control role in the coordination between measurement/experimentation and technological development. Mathematics is viewed as a limiting case of this activity, where the role of external world is eliminated. Mathematical model building is characterized by maintaining representational invariance among representational systems, and not by a representational connection to an external world. Within this picture of mathematics, it is not difficult to see how a mathematical model can be integrated in an informational network where (e.g.) a physical model is developed, and can be used as an informational interface to formulate a physical theory in a mathematical language.

Back to the top.