On epistemology

Definition of Epistemology

Recent advances in brain sciences, ethology, computer science (including AI), and logics impose a revision of our theories in epistemology. There are many ways in which these sciences inform the study of knowledge itself, I will focus only on a few of them. My primary interest is to understand complexity in scientific modelling. Why are some models more complex than others? Is complexity inherent to the real object under study? Or is it relative to a point of view? What is the relation between complexity and the amount of information a model coveys about an object. Can complexity be reduced? And if so, what do we have to compromise?  

These are important practical questions. As computers penetrate almost every aspect of our lives, there is a growing need to model/simulate virtually everything. Computers have a limited capacity and their usefulness comes from their ability to process information and respond in effective time. Consequently, we need to find ways to design models which require less computational resources. This can be done by reducing the complexity of models that we implement in our machines. Thus, the importance of computer technology in our lives forces scientists to design models that are syntactically simple, but how are we going about designing the simplest model possible of a given reality that is to be treated in a given computer application? How do we know that we have found the simplest model one that suits our needs? What is it about the object being modelled, about the type of machine that we are using, and about the use of a specific application, that dictates how to build the most efficient model?

I have to mention here that this pragmatic argument in favour of syntactically simple models has its roots in an environment where computers didn't initially exist, but computing was part of the game. Engineers working in big industries, for practical reasons, have always favoured simple models, as long as they do the job of predicting phenomena within a given margin of error. This is the reason why car manufacturers don't use quantum mechanics, even though the underlying model would give them results with far superior accuracy, and according to the popular belief, it represents reality better. They stick to the old Newtonian models, because they are simpler, and they do the job.  

Therefore, the growing need of computer modelling and simulation puts pressure on scientists to design simpler models. But are these models real science? What is the difference in "scientificness" between a very efficient model built for a specific computer application, one that uses a strange ontology and an exotic logic, and a model which speaks to us, is intuitive, derives from very well known scientific theories, but is very complex? In terms of representational power, amount of conveyed information, and predictive power, they can be equivalent. The only difference is that the second one looks more intuitive than the first, and we can easily link the second to other models and theories. In my opinion, they are both as scientific as they can get. But this view is very controversial, because it leads to a change of paradigm in science, it forces as to review the way we understand and practice science. It implies ontological relativity. 

The articles in this section tackle these issues. Some of them are written in French. If you are interested but you don't understand this language contact me, and we can translate them. 

Two scientific values, two distinct approaches to reality

We can extract two important values from our scientific activities: technological development and the increase of our overall understanding of the Universe. We can all agree on the fact that technology contributes to the development of mankind, and that this development ultimately bears on the efforts of the scientific community. Engineers use scientific models to make accurate predictions about the systems that they design. Alternatively, through Physics we aspire to acquire a coherent understanding of the world, a unique picture that encompasses everything that we know exists., from the smallest particle to the Universe at large. In my opinion, this aspiration is unrealistic but, nevertheless, this seems to be another role that we give to science. Servicing technology and building a broader comprehension of the world are two very different goals that require different sets of skills and resources.

As machines are progressively taking over humans in industrial plants, our need for rapid data analysis, real-time automatic decision-making, and simulations grows exponentially. Technologists are increasingly looking for models that are syntactically simple. As they are normally rewarded for accuracy and efficiency; the model that they use must not necessarily be general (all encompassing), but specialized and effective. We now know that an empirical domain can be modelled in many ways. In principle, it is possible to build very efficient models by choosing a proper ontology and logic, ones that are suitable for the particular object under study, and for the intended use. Physicists could provide engineers with the theoretical material suited to their needs. A coherent theory of everything is in itself a worthy goal, but the product of this effort cannot be something simple, and it is not necessarily suitable for all applications. The complexity of the model grows with the extent of the empirical domain to be modelled. 

The scientific method needs to be expanded

In order to accommodate two distinct scientific approaches, the scientific method itself needs to be readjusted. This constitutes an expansion of the scientific method by the addition of a new step: The scientist will be advised to choose rationally between different formal systems, and to adopt the proper one, justified based on the nature of the object to be modelled, and of the nature of the use intended for the model. This adjustment will create no contradiction with widely accepted bodies of knowledge in epistemology. The character of scientificity of new theories and models that will be generated using this extended method will be preserved. However, this expanded method will give us the tools to judge if one model is more suitable for a particular application with respect to another one. 

The landscape of theoretical science will be greatly affected. If we take Physics for example, this change in method will generate for each of its domains a variety of theories and models, all equally scientific, based on a different set of ontological, logical (qualitative), and quantitative formal systems. Each model will engender its own theories, or interpretations, which will be used by specialists to talk about the objects of study. Dictionaries will be created to translate each specific theory into another one that describes the same reality, and perhaps to translate the consequences of every specific model into the language of a more general theory, one that would also be perceived as more intuitive. 

The first task of my project is to show that the two aforementioned scientific values are distinct and incompatible. Secondly, I will describe each of the two scientific approaches and their own method. And finally, I will show examples and propose concrete applications of these new concepts in the field of Physics. (see the article The Structure of Physics)

Contribution to the advancement of knowledge

I think that my project will contribute significantly to the advancement of knowledge and to the technological development. First, I propose a distinction between two scientific approaches, which fulfill two different purposes: the development of technology, on the one hand, and our overall understanding of the Universe on the other. From a philosophical point of view, this distinction leads to a new definition of science, to an expansion of the scientific method. 

Furthermore, my project brings us to a non-anthropocentric conception of scientific modelling. Adopting a pragmatic attitude, and approaching logics, ontology, and metaphysics from neurosciences, in the context of the theory of evolutionary biology, we can rethink modelling on a new basis that transcends the human condition. It has become common practice for engineers to seek their inspiration in biology, to create functionality by imitating very efficient mechanisms "designed" by nature through millions of years of evolution. Experts in robotics have built very efficient navigational systems after studying the vision of flies, for example. Getting from the physical mechanisms through which interaction and action become physically possible, to the cognitive processes behind expectation, planning and optimization, decision making, etc., physicists will be able to get their inspiration from other intelligent creatures to formulate simple, but very effective, specialized scientific models. How can one transcend human imagination? By relying on the mechanics of formal systems; the same way Einstein arrived at talking about black holes, strange things in four dimensions.

• Bibliography

• Layout 

• Resources 

Audio

audible.com

A search engine for audio books

Websites

Philosopedia.org

Theory and History of Ontology. A Resource Guide for Philosophers - by Raul Corazzon

Philosophytalk.org - Philosophy Talk is a weekly, one-hour radio series produced by Ben Manilla. The hosts' down-to-earth and no-nonsense approach brings the richness of philosophic thought to everyday subjects. Topics are lofty (Truth, Beauty, Justice), arresting (Terrorism, Intelligent Design, Suicide), and engaging (Baseball, Love, Happiness). This is not a lecture or a college course, it's philosophy in action! Philosophy Talk is a fun opportunity to explore issues of importance to your audience in a thoughtful, friendly fashion, where thinking is encouraged.

Spoken Word (from Galsgow Caledonian University) - Repository of audio and video content (account: email + my classic password)

Video

Videolectures: The main purpose of the project Videolectures.Net is to provide free and open access of a high quality video lectures presented by distinguished scholars and scientists at the most important and prominent events like conferences, summer schools, workshops and science promotional events from many fields of Science.

TED: Annual conferences in Long Beach and Oxford bring together the world's most fascinating thinkers and doers, who are challenged to give the talk of their lives (in 18 minutes).