What is it?
Dreamlogs produce streams of idea associations on the Web. Starting from
a discursive position you can for instance try to reach the opposite discursive
position within a number of steps. To do so, choose a word or a piece
of sentence as a starting point: "I love life", "I'm sad",
etc. whatever you want (it works better with a simple one). When you press
"Go", the program searches related text fragments among all
the texts containing the input. Then choose another word or sentence among
the texts displayed (by clicking on it) and reiterate the process. At
the end, save your journey through the space of discourse and you will
see the dreamlog that has been generated. The initial idea of the project
was to obtain paths as large and as "non-local" as possible.
Otherwise you can just enjoy wandering in the space of discourse.
Why is it non-local?
When you perform a search request in a search engine, you reach the set
of all texts that contain the searched string. It is a local action, since
you stay in the neigbourhood of your input. Dreamlogs allow you to iterate
the process and to deviate little by little from your starting point.
How is it related to the OuLiPo?
After I started working on this project, a friend of mine (Gilles Esposito-Farese)
drew my attention to a project by G. Perec and M. Benabou. The PALF (Production
Automatique de Littérature Française) which they submitted
in 1964 to enter the OuLiPo (Ouvroir de Littérature Potentielle)
founded in 1960 by Raymond Queneau et François Le Lionnais. The
idea is that you consider a text and you replace each word with its definition
(in some dictionary for instance). Then you reiterate the process and
you observe the evolution of the text. Then you take an other text, as
different as possible from the first one, and you do the same. After a
serie of iterations, both have to converge towards a single huge intermediate
text. This of course produces very beautiful and funny texts (in French).
The project opens with the sentence : "On ne commande au langage
qu'en lui obéissant" ("you can only command language
by obeying it"). The full extract reads: "On ne commande au
langage qu'en lui obéissant. Il est grand temps pour la littérature
de prendre le langage au sérieux. Il faut déclarer hautement
que le langage n'est plus le petit enfant que lon mène par la main
et que l'on force à décrire en ses balbutiements le monde
qui l'entoure. Il est assez vieux pour causer tout seul, le langage. Ce
qu'il peut dire dans sa totale liberté vaut mieux que nos puériles
Here is a representation in terms of graphs:
If instead of using
definition from the dictionnary, you use the idea of proximity of two
elements of discourse, and instead of looking vertically, you try to follow
the paths... you generate what I call Dreamlogs. The elementary vertex
looks like this :
So Dreamlogs are a generalization of the PALF, by relaxing its constraints
and considering weaker ones. Just as General Relativity involves more
general a symmetry than Restreint Relativity, or as Topological Field
Theory involves more general a symmetry than General Relativity etc.
What about the virtual identities that emerge from these Dreamlogs?
Here, individuals are not the elementary bricks. You define individuals,
by adding up elements of discourse, close to each other. This somehow
looks like a kind of ergodic hypothesis: imagine for instance that the
whole set of texts of the Web had been written by only one person but
at different moments of his/her life, instead of many persons at some
arbitrary moments. In a Dreamlog, you start from one discursive position,
and little by little, you deviates from its original position, and at
the end you find yourself in opposition with what you were in the past.
It's a little bit like real life. And from this final discursive position,
you have enough perspective so that somehow you can see yourself as you
were, through a kind of global mirror.
to be continued...