What are the odds
of creating a computational theory of argument?

This post is an abstract of a more extensive text to be posted later. In the moment, I wish to explain the intention of the above title, hoping that even such a concise message may become thought-provoking for those interested in the subject in question.

In the envisaged  post I am to discuss some ideas of Marcin Koszowy’s paper
“Argument and Computation, and the Legacy of the Lvov-Warsaw School”,  as published recently  in Our Pub Library. There are at least two points to be considered,  both mentioned in the quoted title. Its second line hints at  the need of inquiring into LWS contributions  to the computational approach to argument. This is a challenging task for historical studies which should be undertaken, esp. by Polish scholars.

Nevertheless, I am to focus on what is signaled in the first part of the title – the question of how the computation theory may enhance our knowledge of arguments. There are at least two possible understandings of this relation. (1) We ask which logical theories can be used to create a computational  model of logically correct argumentation; an illuminating example  is to be found in Paul Lorenzen’s dialogical logic,  in paraconsistent Jaśkowski-style logics, etc. (very useful remarks on this point are found in Max Urchs’ paper Discursive Logic. Towards a Logic of Rational Discourse, “Studia Logica”  vol.54, 231-249,  Springer 1995).  (2) While the former is a normative approach, the second one is partly descriptive and psychological. It would consist in digital simulation of real processes of arguing in which a proponent addresses her/his audience in order (i) to convince it to accept a view, (ii)  to motivate it to some action. In  more complex issues, there appears the need of (iii) introducing concepts with which the audience has not been familiar so far.

Point (iii)  is most difficult,  as demanding a creative effort both from the proponent and the audience.  Such creativity  does not seem likely to be reproduced  with digital simulation; I am to exemplify this point with some experiences in the field of automated theorem proving,  e.g. those obtained with an attempt to mechanize Boolos’ “curious inference” as  formalized in the second-order logic.  In the conclusion, as having been motivated by such considerations,  I am  to express a doubt concerning the feasibility of an adequate  digital simulation of argumentation processes. This is why the odds do not seem to be in the favour of a computational theory of argument.

This entry was posted in Rational Discourse, Rationalism in Science. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *