Dominance, hostility and autonomy are interpersonal phenomena that emerge from the complex dyadic interplay of two individuals reciprocally influencing each other. Assessing the complexity of interpersonal interactions usually involves its reduction, for example with self-report and observer-rated measures informed by the structural analysis of social behavior (SASB). In contrast, letting individuals generate a complex stream of interpersonal experience and behavior from moment to moment is an empirical approach not yet usual in interpersonal theory. In the present study, we developed and evaluated an interpersonal, generative paradigm that allows participants to interact nonverbally and spontaneously with a computer-controlled other player in real-time without the need for introspection or the capacity to verbalize potentially implicit interpersonal processes. In the game-like paradigm, participants use the keyboard to take over objects such as a handcar to move autonomously around and encounter interfering or freedom granting artificial other players. We expected that participants (1) experience the nonverbal so-called game mechanics of the paradigm as interpersonal in nature, (2) behave towards the other players in an interpersonally complementary way, and (3) are influenced by their own trait interpersonal expectations. During the paradigm, 40 participants appraised the majority of the game mechanics and computer-controlled other players as intended. Also, interpersonal traits affected the spontaneous behavior towards artificial characters. These findings corroborate the feasibility and validity of a generative assessment of interpersonal dynamics beyond self-reports and observer ratings. The paradigm paves the way for the empirical testing of formal, computational models of dyadic interaction.

支配、敌意和自主是人际交往现象，它们产生于两个相互影响对方的复杂的二元相互作用。评估人际交往的复杂性通常涉及其减少，例如通过社会行为结构分析 (SASB) 提供的自我报告和观察者评定的措施。相比之下，让个人时刻产生复杂的人际体验和行为流是一种在人际理论中尚不常见的经验方法。在本研究中，我们开发并评估了一种人际关系，生成范式允许参与者与计算机控制的其他玩家实时进行非语言和自发的互动，而无需内省或语言表达潜在的隐含人际过程的能力。在类似游戏的范例中，参与者使用键盘来接管诸如手车之类的物体以自主移动并遇到干扰或给予自由的人造其他玩家。我们预计参与者 (1) 将范式的非语言所谓游戏机制体验为本质上的人际交往，(2) 以人际互补的方式对待其他玩家，以及 (3) 受到他们自身特质的人际交往期望的影响。在范式中，40 名参与者按预期评估了大多数游戏机制和计算机控制的其他玩家。此外，人际关系特征影响对人造角色的自发行为。这些发现证实了超越自我报告和观察者评级的人际动态生成评估的可行性和有效性。该范式为对二元交互的正式计算模型进行实证测试铺平了道路。