Monday, April 18, 2011

The possibility of cognition

The most fundamental question raised by “The Blank Swan” may be that of the level of cognition of the market an individual can acquire, and the usefulness of such a cognition if it is, at all, possible. Such a matter is obviously paramount to the validity of Technical Analysis. The untotalization of possibilities Elie Ayache shows with regard to financial markets, seems to invalidate most of the current attempts at thinking this market in explicit terms, as most, if not all, of these attempts are ultimately based on probabilities computation (and therefore on unwarranted, even false, assumptions about the totalization of possible states), and this is indeed the case for Technical Analysis, though I believe that the fractal analysis I have endeavored to develop in this blog provides for an untotalization by means of an implicit multifractal model, where Hurst exponent keeps being recomputed (I however start thinking this model still falls short of being efficient at a theoretical point of view). In this post, I therefore intend to examine, from the standpoint of such a critic of probability theory, whether some kind of cognition is still possible as to what the market is going to be.


The best way to read a book before it is written is to write it, and that is, to some extent, what Elie Ayache is proposing us to do in The turning. There he shows how the market can be dealt with, not by predicting it by computing some probabilities artificially attached to possible states of the world, but rather by writing contingency, i.e. writing contingent claims. However, the book of the market is not written by any single individual (or even any single intentional entity), as is clearly said on page 43:

“The place of the contingent claim is nobody’s place in particular. It falls to no subject to assign a price to the contingent claim or to reflect it in his mind.”

Writing a contingent claim, therefore, does not quite amount to write the book of the market. It does amount, however, to protect one’s financial interest from the uncertainty of the market, from its contingency. In this sense of one’s direct financial interest, as being under the threat of contingency, writing of contingent claims indeed appears as the means to “mediate contingency”. The question which interests me, at the level of Technical Analysis, is whether we can mediate contingency beyond this direct financial interest, and still do that in a speculative manner (in the philosophical sense of the term “speculative”), in other terms, can we speculate (financially) speculatively?
As to read the book of the market before it is written, it obviously is not possible, as such a thing would clearly come down to write it, and as such, it would make it redundant, and therefore destroy it. If the book of the market was to be written by one subject (or if its writing could be seen as being the work of one subject), it would immediately cease to be a market, as a market can only be a place of exchange, necessarily supposing the presence of at least two independent subjects.
Nonetheless, speculative knowledge is not perfect knowledge of the phenomenon under inquiry, on the contrary, speculative knowledge is precisely imperfect, partial, fragmentary, as such a knowledge is rooted in the necessity of contingency, which implies the knowledge that perfect knowledge is illusory (not in an epistemological sense but in an ontological one).
As a consequence, we will not be able to read the book of the market before it is written, we will not be able to predict it in a deterministic way, nor will we be able to predict it in a probabilistic way, what we could endeavor to know however is the language of the market, and from knowing its grammar, we may be able to infer something about the market and its dynamics, just like a knowledge of a natural language allows us to expect a verb after a subject (or the reverse, depending on the language we consider). Such a knowledge may not be enough to diminish the absolute contingency of the market, but it should be sufficient to provide a basis for a speculative speculation, or, as Nishida calls it, an action-like intuition.


Robert Wilkinson presents the concept of Action-like Intuition, that he calls Action-Intuition, in the following manner:

“We must experience the world in order to act on it, and we learn to perceive the world better by acting on it. Just as he [Nishida] insists that practical reason is more profound than the theoretical, so he insists that our natural mode of being-in-the-world is not contemplative but active, an aspect of the constant mutual interaction between individual and the world. The idea that experience is a passive reflection of the world he regards as entirely false: ‘intuition, separated from action, is either merely an abstract idea, or mere illusion’(Intelligibility and the Philosophy of Consciousness, p.208). Action-intuition, like any other form of action in Nishida’s late thought, is a mutual relation of forming and being-formed: ‘Action-intuition means our forming of objects, while we are formed by the objects. Action-intuition means the unity of the opposites of seeing and acting.’(ibid, p.191)
[…], the philosophy of pure experience leads Nishida to take a view of concept formation diametrically opposed to that to be found, for example, in the classic empiricists, according to whom concepts are arrived at by some process of abstraction based on noting common elements in numerically disctinct perceptions. Concepts are not formed in this way in Nishida’s view. We form concepts in the course of action-intuition: ‘Conceiving something through action-intuition means: seeing it through formation, comprehending it through poiesis.’(ibid, p.210)
The basic thesis of the philosophy of pure experience is that the world is a construction from such pure experience, and manifestly such construction has to have some method: action-intuition is the basic formative operation by means of which this construction is carried out. […]. Cognition has to be understood as a form of dynamic, reciprocal expression”

[Nishida and Western Philosophy, Robert Wilkinson(2009),p.120-121]

While Nishida obviously considers these remarks to apply to the whole of reality, and while such a stance may be argued against, I believe there is not much argument as to the relevance of his remarks when it comes to the market. Cognition, in this domain, can only “be understood as a form of dynamic, reciprocal expression”, and concepts formation, according to Nishida, can only occur within a poietic attitude, that is an active one, and not a detached, analytical one. This dimension is well-established by Elie Ayache in “The Blank Swan” with regards to the writing of contingent claims, and particularly with the logic of inverting dynamic replication with the view of implying volatility. When it comes to Technical Analysis, what Nishida is saying, also has an interesting consequence, in that it tells us, that, in order to grasp the market, we must grasp the grasping itself. We therefore need a Technical Analysis tool that is essentially self-referential, there is however a difficulty in understanding this sentence, that lies in the difference of velocity between the processes in historical reality, which are the ones Nishida is treating, and the processes in the market which are the ones interesting us.
The remarkable characteristic of the market is its proximity to the virtual (wherein speed is infinite), a consequence of this proximity is its very high speed, and its emancipation from causality. This high speed also accounts for the absence of a subject-object distinction because such a duality does not have the time to accrete. We are therefore confined, within the market, in a relatively unfriendly environment when it comes to scientific investigation (even a probabilistic one). In this context, self-reference itself becomes an ill-defined notion, since we don’t even know on which entity to apply such a self-reference. Of course, we may say that the market is self-referential, in some sense, but since we don’t know what the market is, since we can’t reduce it to a subject or an object, we have no direct way to comprehend such a self-referentiality in order to translate it in a cognition (be it a partial one) of how the market may evolve. This ambiguity is enough to invalidate a TA tool that would simply be self-referential since such a tool could only be efficient if every market-actors were to use this specific tool, which is obviously impossible. What we need is a tool that is self-referential in the way the market (whatever it is) is self-referential, we therefore need a TA tool that accounts for the very grammar the market is writing itself in.


What I call the grammar of the market, extending the analogy made by Elie Ayache between the market and a book, asks for a little precision here. As said above, the velocity of the virtual is infinite (because the virtual is not situated in time), and the market inherits some of this velocity more directly than history, as such it appears much faster than history and mundane life (this high speed also contaminates real history and accelerates it in some way, this is particularly visible in recent times). Natural languages also happens in history and as such, their grammar seems relatively constant to us, nonetheless, natural languages change, and so do their grammar, we must therefore expect the grammar of the market to change faster than the pace we are accustomed to with natural grammar.
In order to elucidate what we can know of this grammar (that can only amounts to some structure of it, and therefore to a meta-grammar), we must first look at the market globally and that leads us to recognize that it has fractal features. This, in itself, is already a very interesting finding, one from which I have tried to develop some TA tools , but many unknowns remain, such as the adequate period for calculation, the real meaning of fractal dimension, the scope of the probabilistic model (Fractional Brownian Motion) used,…,and the mathematics that sprang from the fractal theory seem relatively limited to clarify these unknowns. The holistic approach of Fractal Theory only provides a very global view of the price dynamics, and Mandelbrot himself even excluded its possible application either to investing or to trading; in his view, Fractal Theory only served to invalidate probabilistic and statistical inference from the market.
However, to obtain a model that would provide a higher interest in building TA tools, we need to start considering a reductionist approach at some level. Again here, I must insist, it would be absurd to look forward obtaining a precise account of the working of the market, when I am talking of reductionism, it must be clear that I mean a very partial one, that will inevitably fall short of elucidating the processes of the market. Reductionism may indeed not be the right word, what I am intending to look at, is something in between holism and reductionism. Despite such reserves, I believe there may be something valuable to find and to explicit about the market, and that this something may lead to a deeper understanding of the whole reality.


I said earlier that the fundamental properties I wish to look at are to be found at a topological level. One way to study such properties is to find a space homeomorphic to the one we wish to investigate, and that is simpler to manipulate.
When it comes to self-similar fractals, which are typically build by IFS (Iterated Function Systems), it is known that we can find a map ψ so as to assert the homeomorphism of some self-similar fractals with a space of p-adic integers:

From this map, we can obtain the fractal dimension of the constructed self-similar set:

For b=3 and p=2, we get:

Where this homeomorphism is actually mapping the ring of 2-adic integers onto the Cantor Set
Alain M. Robert provides a more detailed discussion of these maps in "A course in p-adic Analysis"(pp.8-17)

Of course the fractals we wish to investigate in Finance are not as simple as those built by IFS, in particular, the self-similarity is not strictly true. Nonetheless, I think such a direction may lead to some interesting results. The ideal objective would be to establish a general procedure to find a map between a set of arbitrary fractal dimension and a subset of the space of p-adic numbers. I believe such a question is still an open one, and I am not sure of the advancement of research in this area (or even whether there are any), as I am just starting to look at this question.
The fields of p-adic numbers also present another interesting feature when it comes to account for the process of decision-making at an atomic level. The market is clearly the product of multiple decision-making processes, and as such they are all, individually, rooted in a valuation of reality. While we are well-acquainted with the classical absolute value that leads to the intuitive definition of distance (metric), p-adic fields are equipped with an ultrametric that satisfies the strong triangle inequality.
Whereas a metric satisfies the following triangle inequality:

An ultrametric satisfies the following:

Such a feature leads to rather counter-intuitive results, when we try to visualize them in geometric terms, such as the following formula, known as "The strongest wins":

However, if we think in terms of decision making, we will indeed tend to ignore menial parameters to base our decision on the one parameter we consider as the most relevant. In that, we seem to be closer to an “ultrametric mode” of thinking.

These considerations are still far from exploitable intuitions, and I myself am not very sure whether they will lead anywhere. Once again, I am only in the process of learning about this problematic, and anybody is welcome to criticize or comment, either positively or negatively, on such ideas.