Arguments, Conflicts, and Decisions - Conflict and Multimodal Communication: Social Research and Machine Intelligence - page 122

the discussion, to shelve an issue or postpone debating it, and more generally to find
suitable ways of concluding the argumentative interaction before the matter under
discussion has been settled. The factors affecting such decision are, by and large,
the same that are responsible for entering the argument in the first place: whatever
reason makes it worthwhile to argue with someone, it is typically also the reason that
justifies prolonged discussion of that matter with the same interlocutor. However,
the history of the ongoing argument also has a rational influence on the arguers'
decision to either continue or interrupt it.
Two main aspects are relevant here: how the argument progressed so far ( quality )
and how long it has been going on ( duration ). Obviously, the quality of the argument
is likely to predict the arguer's intention to continue it: if an argumentative exchange
is perceived as productive, interesting, engaging, informative, etc., it stands to
reason that both parties will be willing to keep at it; if, on the contrary, the argument
is experienced as pointless, frustrating, circular, depressing, etc., the arguers will be
ready to abandon it at the first chance. Once again, here we see the double-edged
role played by conflict considerations in argumentative decisions. However, the
quality of an argumentative exchange needs to be assessed on a case-by-case basis,
since it depends on the specifics of the interaction, rather than on abstract criteria.
As such, considerations of quality are unlikely to yield general predictions on
argument termination, aside from the obvious — good quality discourages argument
termination; bad quality encourages it.
Argument duration is a more objective metric, one that has general and verifiable
effects on argument termination. 5 To see how, let us consider the arguer as an
expected utility maximizer (or any other type of rational decision maker, as
discussed below): whenever continuing to argue yields an expected utility higher
than terminating the argument, this is what the agent will (and should) do. Thinking
in terms of expected utility implies focusing on two aspects of the decision problem:
the payoffs and the perceived likelihood of securing such payoffs. In the case of
argumentation, the payoffs can be roughly characterized as the benefits of arguing
(whatever positive goal the arguer expects to obtain) minus the costs (whatever
resources are needed to carry out the argumentative activity) minus the dangers
(whatever undesired side effect is likely to occur, due to argumentation). As for
the perceived likelihood of all these outcomes, benefits and dangers of arguing
are typically uncertain (arguers can be more or less sure of reaching a satisfactory
conclusion and more or less afraid of incurring harmful side effects), whereas the
costs of argumentation tend to be fairly certain — all other things being equal, a long
5 The reason why the argument happens to be prolonged will also affect the arguer's choice: that is,
argument quality and argument duration are not entirely independent from each other. Consider for
instance counter argumentation , that is, arguments from the other party aimed at undermining or
undercutting one's own arguments (see Pollock 1992 ;Prakken 2000 ; Besnard and Hunter 2001 ).
Given extant evidence on how preexposure to counterarguments affects resistance to persuasion
(e.g., Petty and Cacioppo 1977 ), the amount of counter argumentation experienced so far is likely
to matter for deciding whether and how to terminate the exchange.