Figure 1.
(a) Set-up of the experiment with a volunteer participating in the study on the right and a confederate of the experimenter acting as a putative interaction partner on the left. (b) One exemplar interaction block of the experiment consisting of six interaction events.
Figure 2.
Hypotheses of humanness ascription under changing situational demands are depicted here as simple models.
(1) Naive interaction: The ascription of humanness is based on maximally congruent reactions (solid line). (2) Cooperative interaction: The ascription of humanness is based on the mere contingency of reactions (dotted line). (3) Competitive interaction: The ascription of humanness is neither based on congruency nor on contingency of gaze reactions (dashed line).
Figure 3.
Experiments 1, 2, and 3: The ascription of humanness to a virtual character during interaction with an interactor that is (a) supposedly naïve to the participants’ task, (b) introduced as cooperative, (c) or as competitive.
Figure 4.
Experiments 4 and 5: The ascription of humanness to a virtual character while concurrently solving a cognitive load task.
A median split separated participants with high and low scores in the cognitive load task. Solid lines represent the mean humanness ratings of high performers, whereas dashed lines represent low performers. (a) During naïve interaction cognitive load performance had an effect on humanness ascription (p = 0.49). High performers show a stronger congruency-based response pattern compared to low performers. (b) In cooperative interaction there was no effect of load performance on the ascription of humanness.
Figure 5.
Reaction times of humanness decisions split by median.
Grey bars indicate mean ratings. Mean ratings of fast responders (reaction time below median) are indicated by green scatter plot, mean ratings of slow responders are indicated by red scatter plot. Effects of response time are indicated in brackets. (a) Naïve interactor (p = .034): In fast responders humanness ascription is driven more strongly by congruency than in slow responders. (b) Cooperative interactor (n.s.). (c) Competitive interactor (n.s.). (d) Naïve interactor + cognitive load (p = .004): Fast responders show stronger congruency-based response patterns compared to slow responders. (e) Cooperative interactor + cognitive Load (n.s.). (f) Mean reaction times for all experiments (see text for details).