The poetic computer
John Bjarne Grover
My PhD dissertation of 1997 (rejected in 1998) built on the work I did for my master thesis, including the idea (which was not in the thesis but was part of my work for it in 1991) that computation of information flux from eternity to history is crucial for understanding aspects of signification and necessary for constructing a new technological level. The commutability of conditional subset entropies can be summed up in the following formula (for the derivation see the end of this file, the part called 'The holocaust and the poetic computer'):
H(A) + H(BŠA) = H(AB) means that
|- ∑i∑j||p(aibj) |
|=||- ∑i∑j||p(aibj) |
which applies when every p(BŠai) is identical to p(BŠA), which again is when every symbol is always loyal to its category. Which sounds over-trivial to mathematicians who are down to earth on a symbolic level. If the opposite obtains, that a symbol ai can be illoyal to its category A, as is a normal thing in natural language (when an item sometimes is not what it normally is, or perhaps carries a poetic load), then the commutability does not obtain and one gets what can be defined as 'technical Time' - a linear ordering which can be considered a hallmark of natural language.
The formula can therefore be considered an equation for the condition of commutability of these subsets - which is a definition of when they are non-commutable. That is when the probabilities of a subset add up to more than 1.0. Or, rather, when the hypothesis is not yet fully settled. This probabilistically oriented definition of Time relates individual cognition to the poetic space - a vertical function. (Or is it horizontal?)
Perception transforms sensory signals from some time metric to some sort of frequency spectrum (or at least a perception transform of some sort), and there must be some sort of identity function which lets the inner model recognize an outer signal. This must clearly also affect categorization issues (cp. the equation above) if we are to model the outer space of signals in our inner cognitive space. If music is or at least knows about the harmony of the inner and the outer, identity must expectedly be found on 'pythagorean' ratio points. That is a simple model of cognition: Take the input signal and its (optimized) transform and record when the signal value is identical to the transform value on some Pythagorean ratio point - such as e.g. 1/2 through the analyzed 'window' (clearly the signal and the transform will intersect all the time if continuous so this must be some measurement which involves in some way or other the values in the various parts of the ratio). Make a mark in the window sill where they are identical - and that creates a series of notches (of various kinds) which can be called 'cuneiform'. This function computes in a similar way on the social space as the above function on sensory signals and is a horizontal function. (I would say horizontal rather than vertical). Or is this the poetic function and the above is social?
It is in the intersection between these two functions that one probably can find a new start for continuous linguistic theory: If the 'vertical' function computes on technical Time, this 'horizontal' function could compute on technical Place or nominality. The conjunction of these two functions could solve not only the 'adjectival' problems of current knowledge theory but could perhaps also create a sort of poetic-logical computer.
It is likely that the computations in such a computer would be very complex: Given a slice of a signal and a notch telling that here is some identity with the transform, what could the continuation of the signal look like for producing another notch a little later? 500 milliseconds later? For finding wellformedness conditions on a series of such notches one would have to carry out very complex math which I suppose largely would be about differential equations. There exist analog computers, working on continuous voltage metering rather than on digital 0/1 thresholds, which can solve some such equations which cannot by any means be solved by digital computation. One can guess that this is where secrets on linguistic wellformedness are hidden.
Goodness-driven economy - solving the problems which Lenin's revolution could not at that time for making the poetic revolution - will have a fundamental parametre in economic value created by way of the opening of the subsets of humanity. That means that the work which contributes to the development in the direction towards the classless and stateless society - or category-less knowledge - is what creates value. The economy will be anti-racist, anti-nationalist, anti-class-divided. Anti nazi. But that means that when the subsets are opened, they will have the character of these expanding subsets. The 'alphabetic' misunderstanding is that technical Time arises from this expansion while the state wherein all elements are loyal to their categories, such as when all employees are loyal in all and everything to the employer or chief of the department, then a Timeless state can be obtained which amounts to something close to eternity. But that is a serious misunderstanding. It is rather when the subsets open - and hence 'technical Time' arises as seen from the 'alphabetic' viewpoint - that poetic logic can apply instead and without chaos. This also seems to be the core of the new political dimensionality today - it is not primarily about the left/right issues of economic exploitation but about this dimensionality of relation between eternity beyond historic time and the symbolic level of representation.
Cuneiform is perhaps not solved at all. Could be it is a code of the kind described here - the old scribes recording the voices and visions in the clay by skilled spacing. I studied some tablets during my work with the PhD dissertation in the second half of the nineties and noticed a peculiar mirror symmetry in the series of symbols on many tablets. That could suggest that the semantic function of cuneiform could be precisely something like this.
© John Bjarne Grover
On the web 5 December 2007
Last updated 18 December 2008
Earlier version with most of the same material on 29 May 2005