Fortcoming Events |
Past Events |
|
|
Futures studies (FS) is a broad field that employs a wide variety of techniques, ranging from forecasting to simulation, from planning to trend extrapolation to scenarios. FS is also characterised by many different conceptualisations and formalisations. The focus of this workshop is on a specific, recent development in the field of futures studies, namely the idea of “anticipatory systems”.
Intuitively, an anticipatory system is one where the actions of a system are influenced by expectations regarding future, anticipated phenomena. In other words, the choice of the action to perform depends, in part, on the system’s anticipation of its own evolution and/or the environment in which it exists. Non-anticipatory systems, on the contrary, are exclusively reactive systems where subsequent states depends entirely from preceding states (usually, according to some rule). Anticipation comes in different guises: the simplest distinction is between systems aware of their anticipations and systems unaware of them. Anticipation can therefore lie low and work below the threshold of consciousness or it may emerge into conscious purpose. In the latter form it constitutes one of the distinctive qualities of causation within the psychological and the social realms.
The aim of this workshop is to explore the implications for Future Studies as a discipline of using an “anticipatory systems” perspective for organising the field. This workshop is meant to the first in a series and will therefore also aim to set out a research agenda with respect to “anticipatory systems”. As well the workshop will include a tutorial on the theory of anticipatory systems as developed by the late Robert Rosen.
9,00-9,30 Welcome Addresses
-- Davide Bassi, Rector of Trento University
-- Remo Job, Dean of the Faculty of Cognitive Science
9,30-10,20 Roberto Poli (Trento), The Many Nuances of Anticipation
10,20-11,10 Riel Miller (Paris), Changing the Conditions of Change: Futures Literacy
11,10-12,00 Aloisius Louie (Ottawa), Tutorial on Anticipatory Systems 1
14,30-15,20 Ted Fuller (Lincoln), Experiments in Anticipation: The Creative Destruction of Futures
15,20-16,10 Pierre Rossel (Lausanne), Ways of Making Anticipatory Systems More Robust
16,10-16,40 Coffee break
16,40-17,30 Overland and Karlsen (Stavanger), Sociological Imagination and Premises in Norwegian Foresight Studies
17,30-18,20 Carlo Scognamiglio (Roma), Anticipation and Future Vision in the Ontological Structure of Human Freedom
9,00-9,50 M. Aaltonen (Turku), The Third Lens – Why We See What We Do and Not Something Else
9,50-10,40 Liliana Albertazzi (Trento), The Micro-genesis of Anticipation
10,40-11,10 Coffee break
11,10-12,00 Inna Semetsky (Newcastle), Anticipation or Precognition? A Phenomenological Inquiry
14,30-15,20 Aloisius Louie (Ottawa), Tutorial on Anticipatory Systems 2
15,20-16,10 David Melcher (Trento), The Role of Prediction in Perception Across Saccades
16,10-16,40 Coffee break
16,40-17,30 Francesco Vespignani (Trento), Are There Multiple Anticipatory Mechanisms Involved in Language Comprehension?
17,30-18,20 Vesselin Petrov (Sofia), Towards a Perspective Ontology of Anticipatory Systems
9,00-9,50 Simone Arnaldi (Trieste), Futures Mindscapes Methods and the Inter-subjective Construction of Anticipations
9,50-10,40 Juan Ferret (El Paso), The Physics of Anticipatory Systems
10,40-11,10 Coffee break
11,10-12,00 Fiorella Battaglia (Berlin), The Anthropology of Anticipatory Systems
12,00-12,50 David Peat (Pari Centre for New Learning), Gentle Action
If you intend to contribute to FuMee 1 please send a two-page abstract to Roberto Poli (roberto.poli@soc.unitn.it) before January 31. Please note that papers must be readable within 35 minutes so as to leave time for discussion. Preference will be given to those whose subject is close to the focus of the conference, namely anticipatory systems.
The workshop is sponsored by the Causality and Motivation research group (one of the interest areas of SophiaEuropa, a project of Metanexus Institute (http://www.metanexus.net/metanexus_online/index.asp) in conjunction with leading universities in Europe, made possible by the support of the John Templeton Foundation (http://www.templeton.org/) and the Municipality of Rovereto.
A conference is sponsored by ISCE Publishing (http://isce.edu) and by SophiaEuropa (http://www.sophiaeuropa.net), Interest Area "Causality and Motivation"; SophiaEuropa is a research project supported by the Metanexus Institute (www.metanexus.net).
The theory of levels of reality is possibly the single most relevant and unresolved problem in science and philosophy. A fully-developed theory of levels requires both a properly generalized concept of causation, able to consider not only material causes but psychological and social causes as well, and a correspondingly generalized idea of agency, able to explain not only material dynamics but psychological and social dynamics as well. Most details of the links connecting together the various levels of reality are still unknown, because the various sciences had mainly been working on causal links internal to their regional phenomena.
The lack of a theory of levels of reality has been the major obstruction to the development of the needed theories. Theories and models concerning the architecture of levels and their links will improve our understanding of the world and its many dependencies. Furthermore, the theory of levels helps asking new questions. This highly multidisciplinary workshop will address both conceptual and modeling aspects of the theory of levels of reality. Since the theory of levels suffers for the lack of any widely-accepted and well-established conceptual framework, plenty of time will be left for discussion and exchange of ideas.
Progress in the theory of levels will have significant consequences for applications in the areas of biological, psychological and social sciences.
Please notice that only 35 scholars will be admitted to the workshop.
If you are interested in participating to the workshop, drop a line to Roberto Poli explaining your qualifications and why the topic is relevant for your research.
Heinrich Herre and Roberto Poli
I see levels from a categorical standpoint (in the philosophical sense of category; however the claim is not dismissed that levels should/could also be modeled by mathematical categories). The first needed step is to distinguish between universal categories (which apply to everything in reality) and level categories which apply to some aspects/dimensions/types of reality only (e.g., the physical or the biological or the mental or the social).
So understood, a level can be seen as a metacategory whose instances are interrelated groups of categories. The next step requires to distinguish between two different types of levels, which will be called _ strata _ and _ layers _ . the basic difference is that strata of reality are characterized by orthogonal groups of categories (apart the already mentioned universal categories which apply everywhere), while layers in the simplest case are connected by a relation such as the one between a subset and its containing set (from a smaller to a bigger group including the previous one).
I _ ll then propose to distinguish three (and not four) main realms of reality (the material, the psychological and the social).
The next step is to figuring out the possible architectonic of levels, i.e., the way in which levels are interrelated one another, or, if one prefers, the interfaces between/among levels. A few essential options are possible: say a linear one in which one level comes after the previous one or a more _ tangled _ pattern in which levels present a more complex organization. Finally I _ ll show that the internal organization of each of the three strata of reality is essentially different. Descriptively speaking, the above looks like the minimal architecture required by a putatively robust theory of levels of reality.
In the context of theories of reality I use _ reduction _ in two ways: 1) the claim that the entire cosmos is exclusively X in nature, or 2) the claim that the ontically basic segment of the cosmos is exclusively X in nature. What is common to both is the claim that something has only the X kind of properties and is governed by only the X kind of laws. I argue that this claim is literally absurd on the ground that we cannot so much as frame the idea of anything _ s being exclusively X in nature. We cannot make any sense, for example, of the claims that there are purely physical objects, purely sensory percepts, or purely logical concepts. I then outline what a systematically non-reductionist theory of reality could look like, based on the one proposed by the late Herman Dooyeweerd (Free University, Amsterdam).
This article explores several standard ways of understanding the ontological status of categories in general and of categories that are often regarded as ontological and linguistic in particular. It rejects these views and adopts instead a position dubbed neutralism, according to which categories are conceived as neutral with respect to ontological status. This view is then used to solve some of the traditional problems associated with categories and to relate categories to levels of reality.
In respect to the debate about a line of instruction from the very small world of genes to the very large world of organisms, scientists have held to a core belief in causal processes as the single driver behind realisation of the complex from the simple. This view of the living world, mapped out in a single causal thread, has been put in question by a computer model by Cancho and Sole on the evolution of language and signalling processes that highlights the tension between information and signification, or words and their meanings in a given context. Applied to the evolution and development of organisms, this implies a central role for ambiguity rather than reductive causation, with the creativity and adaptability of evolution arising from the multiple meanings that are constantly held as possibilities which are then resolved by context. This deposing of a single factor and a primary level of causal explanation leads in its place to a fractal view of the world, where form is seen to resolve globally in space and time from the permitted local geometric transformations, creating different types of order at many different levels of reality.
Biology as seen from an organicist perspective have typically been seen as comprising several levels of live, from within the single cell, to organisms and ecosystems, yet with the molecular level being the most fundamental one out of witch all other components could be built in a bottom-up manner. Much discussion of emergence and reduction in biology have taken departure from this fixed picture of entities at various levels of organization and their dynamics of interaction. Often a substance ontology have been presupposed as fitting for the most "basic" levels, and notions of "decoupling", "chunking" have been applied to support this view. However, in recent philosophy of biology, this picture has been questioned. Epistemically, attention has been drawn to the patchy and "interlevel" character of attempts to provide reductive explanations of biosystems (Schaffner). Furthermore, process metaphysics have been suggested as a better starting point to view component systems of biology. Biosemiotics has been suggested as providing a better framework to conceptualize the crucial informational or communicational aspects of living systems, thus, biosystems are conceived as sign systems composed of processes and mechanisms dealing with sign action. In this paper, some of the specific problems in the attempt to use semiotic concepts for describing the processual interlevel character of organisms and other biosystems will be described.
After fifty years of research in cognitive science, there is no consensus about the role of logic and ontology in relating language to the world. Some people consider lexical resources, such as WordNet, as a kind of ontology, and others consider the informal concepts of natural languages to be too vague and imprecise to serve as a basis for the precise formalisms of science. Yet scientists, mathematicians, and logicians use ordinary language to analyze and explain the most abstruse theories ever conceived. This article reviews the ongoing research, relates it to the historical developments in logic and philosophy, and proposes a synthesis that can accommodate the full range of language use from the most casual conversation to the most precise technical jargon. Such a synthesis is required for computer systems to analyze natural languages, to reason about the concepts they express, and to relate those concepts to the images and actions that relate language and logic to the world. Any hypothesis about the mapping between language and the world must explain how an infant can learn that mapping at a primitive stage and develop it into the most sophisticated systems of mathematics, science, business, politics, and the arts. A semiotic hypothesis would suggest that each child develops a mapping through the process of learning and using language in social interactions. With this hypothesis, nothing is assumed to be innate except the biological needs and the ability to form and transform patterns of patterns of signs. Everything else, including logic and ontology, can develop through experience.
I will outline a model of how the social level of reality emerges out of the psychological level of agentive animals. This occurs primarily, though not exclusively, among humans. Human sociality differs from that of, for example, insects in that, while there may be a hive or nest level emergence of sociality among social insects, there is an individual by individual emergence in human beings of social persons as co-constitutive participants in society and culture. Language is a co-evolved emergent that is central to the ontology of both sociality and of persons.
The quest at the center of evolutionary computation has been to understand the emergence of emergence in nature and how to implement it in simulations. It is not difficult to simulate the emergence of global patterns of behavior from local primitive rules of causation, and thus understand and explain the emergence of one level of complexity, the first order of emergence. But how can we build simulations which then capture the emergent patterns of behavior at one level and use them as the local primitives to build yet a higher level of global patterning, a second order of emergence. This is the problem which has been the focus of recent workshops on _ Dynamical Hierarchies _ in Australia, _ Dynamic Ontology _ in Italy, and _ Computational Synthesis _ in California. It is, as stated in the invitation to the meeting on _ Levels of Reality, _ _ possibly the single most relevant and unresolved problem in science and philosophy. _ Significant progress is likely to result from the realization that the _ capture _ of emergences often takes place by the transference of information from one medium to another, a process I refer to as INTERMEDIATION. In the process of representing reality in computer code materiality is often lost. Thus intermediation is often overlooked in computer simulations of social and cultural multiagent (multicausal) models. Information is always instantiated on a marker, and the materiality of that marker changes the character of the information it conveys. Each marker, information carrier, or medium, consequently has a life of its own: its own audience and capacity, its own costs of inscription, maintenance and presentation. Each has its own vagueness and ambiguity, its own accuracy, precision and repeatability, and ultimately its own robustness, durability, failure and decay. The world is re-presented in the medium of mind as concepts and ideas, re-presented as speech or written text, and embodied in a myriad of material contrivances, signs, symbols and suggestions of behavior, in a world of technological artifacts and ways of doing things - a material cultural environment that is, in many ways, more immersive, durable and compelling than the ideas themselves. It is these artifacts of technology which constitute distributed material cultural cognition, the processes which mediate our daily lives.
Intermediation, in simulation and in nature is often conceptualized as hierarchical, with level building upon level of complexity. However, emergence takes place not only upward, but laterally and downward with significant feedback taking place through heterarchical circuits. We must also take into account at least two temporally different meanings of levels of emergence which might best be described as evolutionary and instantaneous. Not only do we see complexity arise through evolutionary time, we see it maintained at every instant. If the causal regularities that govern the world of physics should change instantaneously, all higher level processes would stop as quickly. With regard to computer models of social and other processes, we must begin to represent the materiality of the real world in the medium of simulation. In other words, we must imbue our simulations with simulations of the media of materiality within the media of computation.
As an aside, this question is insightfully pursued by Greg Egan in his science fiction book PERMUTATION CITY, and the entailments discussed in his description of the relative robustness of the _ Lambertians _ versus the _ Copies. _
Dynamics play a major role in many phenomena, addressed in a variety of disciplines. As a central principle, the temporal factorisation principle is discussed. This principle expresses, that every temporal relationship of the form _ past pattern implies future pattern _ can be factorised into a relationship of the form _ past pattern implies present state _ and a relationship of the form _ present state implies future pattern _ . To enable this, for every _ past pattern implies future pattern _ relationship, the principle postulates the existence of certain mediating state properties in the present state. Using Kim _ s notion of relational specification of representational content of mental state properties, it is described how such mediating state properties relate to patterns of state properties in past and future.
The question is addressed whether and how a postulated mediating state property relates to other state properties in the (present) state in which they occur. In particular, the situation is analysed that realisers exist: other state properties or combinations thereof that co-occur with the mediating state property in states. This is analysed from Kim _ s perspective on functional reduction, thus giving mediating state properties a second-order status.
The analyses discussed provides a conceptual framework covering various concepts and themes that usually are considered totally different and unrelated, such as, the notion of differential equations in Mathematics, the notions of transition system and rule-based system in Computer Science, Cognitive Science and Artificial Intelligence, and the notion of reduction in Cognitive Science and Philosophy of Science.
An ontology of a domain D is based on a system Cat(D) of categories that are associated to D. We stipulate that a domain D = (Obj, V, CP) is determined by a set Obj(D) of objects, by a view V, and by a classification principle CP for Obj(D). The notion of view is used in an informal, intuitive sense, whereas the classification principle CP can be made, in many cases, more precise. In understanding, acquiring and representing the knowledge about a domain D we must use categories from Cat(D) and relations Rel(D) between them. Then, a conceptualization of a domain is presented by a system Concept(D) = (Obj(D), Cat(D), Rel(D)). A core ontology Core(D) of a domain D is to be understood as a theory, based on a set PrincCat(D) of principal categories, that describes what the domain D is about. We present criteria for core ontologies of domains and investigate their relations to levels of reality.
ISO 15926 is designed to support engineering applications and covers physical and functional levels of reality. Shell's Downstream Data Model uses ISO 15926 as a foundation and includes some intentionally constructed objects. Both are 4 dimemsionalist. The levels of reality in these ontologies are presented, and the 4 dimensionalist analysis of levels of reality is examined. A problem faced with 3 dimensional analysis is the coincidence of objects, and whether apparently coincident objects really are different or the same. A 4 dimensional analysis sees objects as being different if they have different spatio-temporal extents. Objects at different levels of reality can be seen to have different spatio-temporal extents, with objects at higher levels of reality being states of objects at lower levels of reality, with it being possible to construct a complete chain from the lowest level of reality to the highest that applies. Interestingly this gives a non-reductionist account of reality, whereas 4 dimensionalism is sometimes claimed to be reductionist in nature.
The rise of ontology engineering as a distinct field in computer science over the past 15 or so years has had great impact in a relatively short amount of time on information technology. The field has borrowed much from formal ontology in philosophy, formal semantics in linguistics and philosophy of language, knowledge representation from artificial intelligence, and aspects from logic and theoretical computer science. Many issues have been addressed and continue to be. The rise of Semantic Web technology (from the W3C: RDF [Resource Description Framework], OWL [Web Ontology Language], RIF [Rule Interoperability Framework], but there are many more emerging standards, including OWL-S (OWL web services ontology), etc., has introduced the notions of ontologies and semantics to a much larger community of developers and potential users. However, there is a perceived growing bifurcation between the ontology engineering community and the Semantic Web community, largely over the presumed complexity of the task of mapping among ontologies, to address semantic interoperability. The Semantic Web community tends to under-estimate the complexity of ontology mapping and assumes that the growing semantic islands emerging bottom-up will be easily mapped. The ontology engineering community, to the contrary, thinks that ontology mapping is very hard and for semantic interoperability to be realistically addressed, that domain ontologies need to be embedded in or linked to common middle and upper ontologies (typically strongly axiomatized), in order for their semantics to be ascertainable and commensurable. This talk will discuss the issues confronting the two communities, their discordances, and prospective paths to a common resolution. It will especially be pointed out that the two communities diverge radically on their perspectives about which levels of reality they are addressing when they develop machine-interpretable semantics.
In "Three obstructions: forms of causations, chronotopoids, and levels of reality", R. Poli suggests the development of a "generalized theory of causal connections" taking account of the multiplicity of levels of reality. We propose to model this by a 'Hierarchical Evolutive System' (HES): the various levels are interconnected through links (modeling relations) between their objects in any direction: bottom-up, intralevels, top-down, so that the hierarchy is "tangled" (as proposed by Poli). The characteristic property is that an object of a given level is an aggregate (modelled by the colimit operation) of patterns (or interconnected assemblies) of objects of lower levels.
The HES would correspond to a global view, but a given observer can only have a partial view, This is modelled by the consideration of a net of 'Co-regulators' (or CRs). Each CR has its own level of interpretation (modelled by its "landscape"), depending on its complexity level, its time scale (to be compared to the "chrono-topoids" of Poli), its own characteristics determined by its admissible procedures, and a differential access to a central 'Memory'. The HES, with its CRs and its Memory becomes a 'Memory Evolutive System' (MES), which corresponds to a processual model whose dynamics is modulated by the cooperation/competition between the various CRs, the Memory conferring it some anticipatory behaviour (by allowing to recall the results of similar former experiences). The evolution of the MES is modelled by a sequence of 'complexification processes'. A main result identifies and discusses the conditions allowing for the emergence of more complex structures and properties, namely the 'Multiplicity principle' (a kind of "degeneracy principle" in the sense of Edelman): a higher structure admits several non-isomorphic decompositions in lower components. We develop the theory of MES in a book "Memory Evolutive Systems: Hierarchy, Emergence, Cognition" (Elsevier, to appear in May); the last part of the book gives an application of MES to the development of semantics and of higher cognitive processes up to consciousness.
The nonstandard mathematics (infinitesimal and Boolean) constitute the clearest case where levels of reality are present in mathematics. These are essentially based on the Protean nature of the concept of _ point _ (see, [CD]). The other important case is Algebraic Geometry, especially in the form of Grothendieck. These in turn lead to Category and Topos theory.
In this talk we are going to:
[CD] Costas Drossos, Structures, Points and Levels of Reality. In: G. Sica (ed.) Essays on the Foundations of Mathematics and Logic, pp. 84-114, Polimetrica International Scientific Publisher Monza/Italy, 2005
A Categorical Ontology of Space and Time is presented for emergent Biosystems, super-complex Dynamics, Evolution and Human Consciousness. Relational structures of organisms are represented by natural transformations of biomolecular reactions. The ascent of man and other organisms through adaptation, evolution and social co-evolution is viewed in categorical terms as variable biogroupoid representations of evolving species. The unifying theme of local-to-global approaches to organismic development, evolution and human consciousness leads to novel patterns of relations that emerge in super- and ultra- complex systems in terms of colimits of biogroupoids, and more generally, as compositions of local procedures to be defined in terms of locally Lie groupoids. Solutions to such local-to-global problems in highly complex systems with `broken symmetry' may be found with the help of higher homotopy theorems in algebraic topology such as the generalized van Kampen theorems (HHvKT). Primordial organism structures are predicted from the simplest Metabolic-Repair systems extended to self-replication through autocatalytic reactions. The intrinsic dynamic `asymmetry' of genetic networks in organismic development and evolution is investigated in terms of categories of many-valued, \L ukasiewicz-Moisil logic algebras and then compared with those obtained for (non-commutative) Quantum Logics. The claim is defended that human consciousness is unique and should be viewed as an ultra-complex, global process of processes. The emergence of consciousness and its existence seem dependent upon an extremely complex structural and functional unit with an asymmetric network topology and connectivities _ the human brain-- that developed through societal co-evolution, elaborate language/symbolic communication and `virtual', higher dimensional, non--commutative processes involving separate space and time perceptions. Anticipatory systems and complex causality at the top levels of reality are also discussed in the context of the ontological theory of levels with its complex/entangled/intertwined ramifications in psychology, sociology and ecology. The presence of strange attractors in modern society dynamics give rise to very serious concerns for the future of mankind and the continued persistence of a multi-stable Biosphere. A paradigm shift towards non-commutative, or non-Abelian, theories of highly complex dynamics is suggested to unfold now in physics , mathematics, life and cognitive sciences, thus leading to the realizations of higher dimensional algebras in neurosciences and psychology, as well as in human genomics, bioinformatics and interactomics.
An important aspect of Ontology is as the study of `objects' and their `ties'. Simple models of this, such as Chu spaces, can be put forward for consideration. They naturally involve use of spatial-type representations of the interrelationships between concepts. They have been use in formal concept analysis and artificial intelligence. They have their own logic which has a good transparent interpretation.
I will discuss the nature of these models and explore some possible relevance of them to levels of reality, logics of observing reality, and the refinement of knowledge.
We propose to see the theories of levels of reality and complexity as the two representation sides of the same reality: the former theories provide mainly a descriptive representation of levels and their interconnections. On the other hand, the latter aims at making explicit the ways in which levels emerge and interact or influence each other. The claim is further advanced that mainstream, physicalist theories of complexity are intrinsically unable to address the emergence of organisms, minds and societies. A novel approach is therefore required, which the authors have named "the higher-order complexity theory". A higher-order complex system is here defined as having at least one complex level of subsystems.
Thus far, the development of higher-order complexity parallels the latest developments in the fields of algebraic topology and higher dimensional algebra, including non-commutative (non-Abelian) frameworks endowed with both topological and algebraic structures, many-valued (non-commutative) Lukasiewicz-Moisil logics, and the paradigm shift from groups to groupoids, especially topological groupoids, such as Lie or holonomy groupoids that may represent symmetry breaking, phase transitions or emergent processes. Such recent developments are especially promising for both a deeper understanding of the brain's variable topology and the way in which brain and mind developments occur stimulated by social/environmental interactions. Two key questions will be considered: how neural systems are linked into hierarchical level structures that support human cognitive functions, and how such hierarchical cognitive dynamics can be related to observable human behaviors. Answering these questions will advance our understanding of how cognition emerges, for example, through social interactions and activation of mirror neuron nets, and how such dynamic bionetworks may change their topology/connectivity to generate inner speech, higher-order layers of dynamic complexity in the human brain, and ultimately, the ultra-complexity level of the unique human minds.
In category theory, an ontology can be formalized in modular form as an interrelated system of domain theories expressed in a formal logic, where a theory relationship expresses the incorporation of one theory in another. The relationships form a many-threaded hierarchy of specializations or, in the opposite direction, abstractions. The analysis of ontologies does not end there, however, for a categorical expression has many levels. Structure-preserving mappings relate theory categories to other categories within which theories and the possible worlds, or models, they describe also form categories. This expresses coherence within and between ontologies in a multi-level system of knowledge representation. This, in turn, provides an explicit, declarative semantic representation for computer software and system interoperability.
The main purpose of the talk is to compare and discuss various formulations of the’levels of reality thesis’. For this purpose I present in outline a new (non-Whiteheadian) process ontology called ‘General Process Theory’ (GPT) where forms of interactivity can be particularly clearly distinguished. In part I of the talk I introduce the notion of a ‘general process’ and the non-standard mereology used to defined compositional relationships between processes. In part II I explain how complex general processes may be further classified within the GPT typology of dynamics. In part III I put the resources of GPT to analytical use and relate various characteristic forms of interactivity (necessary and accidental production, functional dependency, circular and reciprocal dependencies, weak and strong entanglements) to claims about levels of reality to investigate how these relate to species of process composition.
Nowadays, our living environments already provide ubiquitous network connectivity and are populated by an increasing number of artefacts (objects enhanced with sensing, computation and networking abilities). Already, an increasing number of sensors are becoming embedded in the everyday objects or in the environment at a low cost. As a result of this continuing trend, an elementary Ambient Intelligence (AmI) infrastructure has become installed (though still fragmented), information appliances are commercially available, and Ubiquitous Computing (UbiComp) applications (currently in the form of games and informative services) are being deployed. In order to create, manage, communicate with, and reason about ubiquitous computing environments which involve hundreds of interacting artefacts and cooperating mobile devices, we propose to embed in these entities social memory, enhanced context memory, and shared experiences.
Within such an AmI space, people still have to realize their tasks, ranging from mundane everyday tasks (i.e. studying, cooking etc) to leisure or work related tasks, or even tasks that relate to emergency situations (i.e. home care, accident, unexpected guests etc). To do so, they have at their disposal the objects that surround them. These in fact, are new or improved versions of existing objects, which by using Information and Communication Technology (ICT) components (i.e. sensors, actuators, processor, memory, wireless communication modules) can receive, store, process and transmit information, thus allowing people to carry out new tasks or old tasks in new and better ways. In the following, we shall use the term “artefacts” for this type of augmented objects. By viewing the user as part of the system, one has to ensure an unobtrusive symbiosis between human and artificial entities and to establish the user acceptance and confidence towards the UbiComp applications. Our approach is based on a human-centric autonomic system constituting of self-managing ecologies which are diffused in the everyday living space. In such ecologies artificial entities coexist unobtrusively with humans and perform collaborative tasks through a continuous evolvable process concerning both their physical and social cognitive growth (we call them “ambient spheres”). The driving force behind the sphere formation is the selfishness of objects and services; they try to “survive” by operating within a given set of resources. The antagonism of the members is reflected on the spheres, which are formed in ad-hoc (emergent) way and they persistently try to reach the “success criterion”: user satisfaction.
As a consequence of the availability of new technologies, the nature of the human activities eventually assisted by artefacts is rapidly changing. People have to (consent to) build new task models or adapt the ones they have already been using, a task not trivial at all. Execution of new tasks may become difficult due to the inherent systemic complexity of UbiComp applications, which, among others, results from conceptual discrepancy, device incompatibility, fragmentation and the huge number of interactions among visible and non-visible actors.
The theory of Levels of Reality is a promising framework that could help us deal with the above issues. Artifacts, devices, agents, people and, in general, members of activity spheres exist in the same real world. A common low level ontology of terms and concepts can be developed to describe this world in varying levels of granularity. The material world constitutes the Virtual Machine on top of which software programs and applications execute. Software programs implement the local resource management and goal seeking policy of each non-human ecology member. UbiComp applications are modeled as compositions of autonomous artifacts and components; then, their global behavior emerges from the interaction among application components, as well as among applications and people.