400 JOURNAL OF LAW AND POLICY
The lack of significance in the difference between the actual
and randomized dialogues in the direction taken as of interest
here suggests that the dialogues might as well have had the turns
randomly reordered to obtain the same overall effect of
engagement. Certainly, it is clear that the repetitions of the
Paxman questions makes self-repetition high for Paxman but not
substantially different in the ten randomized reorderings.
Similarly, Howard’s repetition of terms used by Paxman is
relatively impervious to reordering. The statistical effects reveal
that at the level of textual content, there is little engagement
exhibited. Accordingly, this leaves open that an analysis
including gesture, timing of utterances, or overall energy
measurable in the scene during the flow of the dialogue could
still detect involvement and engagement at that level,^40 just as
speakers of mutually unintelligible languages may interact with
engagement but without full understanding. The lack of
significance in the contrasts of interest here implies that,
although it seems that willful avoidance might be at issue, one
cannot say for certain that Paxman and Howard reached an
understanding of each other.
It is interesting to note that it is relatively difficult to find
transcripts of naturally, publicly occurring dialogues in which
third party observers conclude that the interlocutors do not
understand each other. Fabrications of such are the stuff of
comedy, such as the “Who’s on first” routine of Abbott and
Costello. The Paxman-Howard example was selected as a
relatively famous example of failure to communicate. With more
successful dialogues, it is generally not true that in all cases all
parties will support measurements according to the methods used
here in which the actual dialogue differs in the hypothesized
direction from the ten turn-randomized alternatives. Where there
is significant self-repetition but no significant allo-repetition for
any of the participants in a dialogue, it would seem that there is
evidence of persistence but not of linguistic engagement.
Conversely, where allo-repetition effects are significant, but not
self-repetition, there is evidence of understanding and
engagement. Where neither of the effects is visible for any of
(^40) Ramseyer & Tschacher, supra note 25.