Abstract: 　　 Although theoretical work has identified a number of factors that may influence trust, only few studies investigated affective influences on trust and didn’t reach a consistent conclusion. According to Appraisal-Tendency Approach to Affect and Judgment, emotions are associated with specific appraisals reflecting the core meaning of the event that elicits each emotion and determining the influence of specific emotions on social judgment. The certainty appraisal, which means understanding what is happening in the current situation, and feeling able to predict what will happen next, is indicated by various literatures being related to processing. Emotions characterized by certainty appraisals promote heuristic processing, whereas emotions characterized by uncertainty appraisals result in systematic processing. The relationship between certainty and information processing implies that certainty may mediate the relationship between trust and specific emotions. Thus, we hypothesize that individuals in emotions associated with high certainty may rely more on “trustworthy cues” to make trust decisions compared with individuals in emotions associated with low certainty. 　　 In Experiment 1, 133 participants (66 men and 67 women, Mage=9.44±1.42) were randomly assigned to one of four conditions in a 3 (emotions: happiness (positive, high certainty) vs anger (negative, high certainty) vs sadness (negative, low certainty)
× 2 (settings: trustworthy vs untrustworthy) between-subjects factorial design. Participants watched a video to induce one type of emotion before finishing the Trust Game. Target trust was manipulated by providing participants with the results of a “trustworthiness scale” that the target had purportedly just completed. In Experiment 2, group boundaries were made to activate people’s trust and distrust schemas based on the recognition that ingroup members were generally perceived as more trustworthy than outgroup members. 153 participants (78 men and 75 women, Mage=19.55±1.27) in happy, angry or sad emotions completed the Trust Game with either ingroup partners or outgroup partners. The group boundary was set depending on whether the partner in the Trust Game was from the same A / non-A province (ingroup or outgroup). The results of Experiment 1 showed that participants experiencing happy and angry emotions provided more money to a trustworthy partner than an untrustworthy partner (thappiness（39）=2.65，p<.05；tanger（42）=3.32，p<.01). The participants experiencing sad emotion showed no significant difference in their trust (tsadness（46）=1.11, p>.05). Besides, participants in happy and angry emotion provided more money to a trustworthy partner than did participants in sad emotion(thappiness-sadness（45）=2.46，p<.05 , tanger-sadness（46）=3.14，p<0.01;thappiness-anger（39）=-0.41, p>.05 ), but not to an untrustworthy partners(F（2,63）=1.25，p>.05).The results of Experiment 2 showed that participants experiencing happy and angry mood provided more money in an ingroup setting compared to an outgroup setting (thappy（52）=2.32，p<.05; tanger（46）=2.43，p<.05). In contrast, the amount of money given by people in sad emotion did not differ(tsadness（49）=-1.77, p>.05). Happy and angry participants gave more money to a trustworthy partner than sad participants (F（2,73）=4.07, p<.05；thappiness-sadness（49）=2.61, p<.05 , tanger-sadness（47）=2.46, p<.01;thappiness-anger（50）=.06, p>.1) but not to an untrustworthy partners(F（2,74）=1.81, p>.05). 　　The present research suggested that when there are available cues about the trustworthiness of the target, the certainty appraisal of emotion has an effect on individuals’ trust behavior . In situations with available cues that promoted trust, people in emotions associated with certainty appraisals increased their trust; but the trust behavior of individuals in emotions associated with uncertainty appraisals was not affected.have different influences on trust on two occasions: When there is information associated with trustworthiness of the trustee, individuals misattribute their mood to trust judgment at hand and misattributed emotions influence trust in the direction of the emotion’s control dimension. However, when there are available cues suggesting the trustworthiness of the trustee, individuals experiencing emotions characterized by high certainty may rely more on “trustworthy cues” to make trust decisions compared with individuals experiencing emotions characterized by low certainty because emotions characterized by certainty appraisals promote heuristic processing, whereas emotions characterized by uncertainty appraisals result in systematic processing. 　　Two experiments were conducted which were designed with two conditions—whether there are cues associated with trust or distrust. In Experiment 1, 105 participants were recruited and randomly assigned to the two conditions. Before playing the trust game in which participants would allocate the money they owned freely to the other player and the money they allocated were used as the measurement of trusting behavior. Participants’ emotions characterized by other-person control (gratitude) or personal control (pride) were induced by recalling past experiences. In Experiment 2, 142 participants were recruited and were randomly assigned to one of the conditions from a 3（emotion：happiness, sadness, anger）× 2 （cues：high trustworthy, low trustworthy）design. Happiness and anger are both characterized by high certainty but differ in valence while sadness is characterized by low certainty. Participants were asked to complete a “trustworthiness” inventory and were told the computer would compute their scores of the inventory. During the waiting period participants watched a video aiming at inducing an emotion. Before playing the trust game, participants in the trustworthy and untrustworthy conditions were informed that their partner either had a high or low score on the trustworthiness scale. 　　Results of Experiment 1 showed that emotions characterized by other-person control (gratitude) influenced trust significantly more than emotions characterized by personal control (pride). Result of Experiment 2 indicated that compared with cues promoting distrust, when cues promoting trust about the trustee were available, participants in emotions characterized by high certainty (happy and anger) showed higher trust; while participants in an emotion characterized by low certainty (sadness) showed no difference in trust. 　　These findings shed light on the relation between specific emotions and trust. It suggests that emotions with different characteristics have different influence on trust, when cues associated with trust are not available, affect as information models works, while heuristic-reliance models takes effect when cues about the trustee are available.

Key words: emotion, trust, certainty

摘要： 本研究通过信任游戏的实验范式探讨了在与“受信任者”高/低可信赖性有关的信任线索时，具体情绪的确定性维度对信任行为的影响。实验一发现，当被试被告知“受信任者”在可信赖量表上的得分（高/低）时，个体在高确定性情绪（开心和愤怒）下的信任判断比低确定性情绪（悲伤）下的信任判断上更容易被受信任者的“可信赖性”水平的高低所影响；实验二发现，当告知被试“受信任者”的群体身份（内/外群）时，个体在高确定性情绪（开心和愤怒）下的信任判断比低确定性情绪（悲伤）下的信任判断更容易被受信任者的“内外群”身份所影响。上述结果表明，高确定性的情绪比低确定性的情绪更容易使被试的信任判断受到与“受信任者”是否值得信赖有关的线索所影响。

关键词: 情绪, 信任, 确定性