What does it mean to be (seen as) human? Ten studies explore this age-old question and show that gender is a critical feature of perceiving humanness, being more central to conceptions of humanness than other social categories (race, age, sexual orientation, religion, disability). Our first six studies induce humanization (i.e., anthropomorphism) and measure social-category ascription. Across different manipulations (e.g., having participants recall experiences, observe moving shapes, imagine nonhuman entities as people, and create a human form), we find that gender is the most strongly ascribed social category and the one that uniquely predicts humanization. To provide further evidence that gender is central to conceptions of personhood, and to examine the consequences of withholding it, we then demonstrate that removing gender from virtual humans (Study 5), human groups (Study 6), alien species (Study 7), and individuals (Study 8) leads them to be seen as less human. The diminished humanness ascribed to nongendered and genderless targets is due, at least in part, to the lack of a gender schema to guide facile and efficient sensemaking. The relative difficulty perceivers had in making sense of nongendered targets predicted diminished humanness ratings. Finally, we demonstrate downstream consequences of stripping a target of gender: Perceivers consider them less relatable and more socially distant (Study 8). These results have theoretical implications for research on gender, (de)humanization, anthropomorphism, and social cognition, more broadly.
成为（被视为）人类意味着什么？10 项研究探讨了这个古老的问题，并表明性别是感知人类性的关键特征，比其他社会类别（种族、年龄、性取向、宗教、残疾）在人类概念中更为重要。我们的前六项研究诱导人性化（即拟人化）并测量社会类别归属。通过不同的操作（例如，让参与者回忆经历、观察移动的形状、将非人类实体想象成人，并创造人类形态），我们发现性别是最强烈的社会类别，也是唯一预测人性化的类别。为了提供进一步的证据证明性别是人格概念的核心，并研究隐瞒性别的后果，我们随后证明，从虚拟人类（研究 5）、人类群体（研究 6）、外来物种（研究 7）和个体（研究 8）中去除性别会导致他们被视为不那么人性化。无性别和无性别目标所赋予的人性减弱，至少部分是由于缺乏性别图式来指导简单而有效的意义建构。感知者在理解非性别目标方面的相对困难预示着人类评分的降低。最后，我们展示了剥离性别目标的下游后果：感知者认为他们不太相关，而且在社交上更疏远（研究 8）。这些结果对更广泛的性别、（去）人化、拟人化和社会认知的研究具有理论意义。

In 2017, National Geographic magazine published a special issue on “The Changing Landscape of Gender,” which argued that we are in the midst of a gender revolution that could mark the end of gender as we know it (National Geographic, 2017). Indeed, the notion of genderlessness is catching on—from genderless bathrooms to genderless pronouns, fashion, and identities. From an identity perspective, it is clear why genderlessness has become increasingly popular: Placing importance on gender—the set of characteristics that distinguish male from female—results in an oversimplified, binary view of sex and pathologizes people who do not identify with this paradigm (Bem, 1981; Butler, 1990; Hyde et al., 2019; Morgenroth & Ryan, 2018). Thus, doing away with gender frees people from being confined to two narrow identities. However, from a cognitive perspective, the feasibility and implications of genderlessness are less clear. As a primary schema we use to understand and process the social world (Bem, 1981; Martin & Slepian, 2017, 2020), gender may be more than a lens through which we view a person; it may very well be considered a defining feature of personhood.
2017 年，《国家地理》杂志发表了一期关于“不断变化的性别景观”的特刊，其中认为我们正处于一场性别革命中，这可能标志着我们所知道的性别的终结（国家地理，2017 年）。事实上，无性别的概念正在流行起来——从无性别的浴室到无性别的代词、时尚和身份。从身份的角度来看，无性别化越来越流行的原因很明显：重视性别——区分男性和女性的一组特征——导致对性别的过度简化、二元化，并使不认同这种范式的人病态化（Bem，1981;Butler， 1990;Hyde等人，2019 年;Morgenroth & Ryan，2018 年）。因此，消除性别使人们摆脱了两种狭隘身份的束缚。然而，从认知的角度来看，无性别的可行性和影响尚不清楚。作为我们用来理解和处理社会世界的主要模式（Bem，1981 年;Martin & Slepian，2017年，2020年），性别可能不仅仅是我们看待一个人的镜头;它很可能被认为是人格的一个决定性特征。

We contend that gender is an important aspect of “seeing human,” one that is more tightly linked to ascriptions of personhood than the other social categories to which people belong (e.g., race, age, sexual orientation, disability, religion). In making this prediction, we build on myriad research in social cognition showing gender to be a salient social category that perceivers use to make sense of others’ behavior (e.g., Fiske, 2002; Ito & Urland, 2003; Macrae & Bodenhausen, 2000). We extend this literature by demonstrating that, beyond using gender in person perception, individuals use gender to define what constitutes a “person” in the first place. We then consider one implication of the tendency to see gender as a defining feature of humanity: removing it may lead perceivers to see a target as less than fully human.
我们认为，性别是“看到人类”的一个重要方面，与人格的归属相比，它与人格的归属关系更紧密（例如，种族、年龄、性取向、残疾、宗教）。在做出这一预测时，我们建立在社会认知的无数研究的基础上，表明性别是感知者用来理解他人行为的一个突出的社会类别（例如，Fiske，2002 年;Ito & Urland， 2003;Macrae & Bodenhausen， 2000）。我们通过证明，除了使用个人感知中的性别之外，个人首先使用性别来定义什么是“人”，从而扩展了这些文献。然后，我们考虑了将性别视为人类定义特征的倾向的一个含义：去除它可能会导致感知者将目标视为不完全的人。

We contend that deeming something human entails attributing gender to it. But why might this be? And why, when ascribing humanness, should gender take primacy over other social categories to which people belong? We suggest gender’s importance in reproduction and survival, as well as the interpersonal and cultural pressures that reinforce gender in cognition, render it a defining feature of people’s notion of personhood. As we will review, social cognition purportedly evolved to help the human species rapidly differentiate biological sex (Buss, 1995, 2015), but everyday experience in social interactions and social systems reinforce sex-differentiation, creating a culturally constructed, binary notion of gender that structures and guides how we see ourselves and others (see Diekman & Schmader, 2021). Thus, we begin from an evolutionary perspective, zooming out to consider the role that gender  ^{[  1  ]}  has played in the course of human history, and then move toward a social and cultural perspective on how gender is created and reinforced in the modern day. In reviewing this work, we lay the groundwork for our argument that gender is critical not just to navigating survival, reproduction, and social relationships, but also to what it means to be considered “human.”
我们认为，认为某物是人类的，就意味着将性别归因于它。但为什么会这样呢？为什么在赋予人性时，性别应该优先于人们所属的其他社会类别？我们认为性别在繁殖和生存中的重要性，以及加强认知性别的人际和文化压力，使其成为人们人格概念的一个决定性特征。正如我们将要回顾的，社会认知据称是为了帮助人类快速区分生物性别而进化的（Buss，1995,2015），但是在社交互动和社会系统中的日常经验强化了性别差异，创造了一种文化建构的、二元的性别概念，构建并指导我们如何看待自己和他人（见Diekman & Schmader，2021 ).因此，我们从进化的角度开始，缩小考虑性别  ^{[  1  ]}  在人类历史进程中的作用，然后从社会和文化的角度来看待性别在现代是如何创造和强化的。在回顾这项工作时，我们为我们的论点奠定了基础，即性别不仅对驾驭生存、繁殖和社会关系至关重要，而且对被视为“人类”的意义也至关重要。

From an evolutionary perspective, gender is critical to social cognition, as sex—in both senses of the word—is necessary for reproduction and, thus, human survival. Social cognition is purported to have developed to navigate evolutionary challenges, and sex differentiation is crucial to this process (Bodenhausen & Hugenberg, 2009; Buss, 1991, 1995, 2015; Schillaci, 2006). That is, if reproduction sets the stage for evolution, adaptive organisms should have developed cognitive mechanisms to navigate reproductive obstacles. To navigate such obstacles, the ability to distinguish males from females is key, as confirmed by research spanning many domains: People exhibit a remarkable capacity for visually discriminating between the two sexes. People can differentiate males (angular facial features, broad shoulders) and females (round facial features, wide hips) quickly and accurately (Ito & Urland, 2005; Johnson et al., 2007; Macrae et al., 2005), and from a very young age (Martin & Ruble, 2004; Miller, 1983; Quinn et al., 2002). Into adulthood, humans can differentiate the sexes from limited perceptual information, such as single-facial features (Cloutier et al., 2005; Macrae & Martin, 2007) or point-light animations of gait (Johnson et al., 2007; Lick et al., 2013).
从进化的角度来看，性别对社会认知至关重要，因为性——在两个词的意义上——都是繁殖所必需的，因此也是人类生存所必需的。据称，社会认知是为了应对进化挑战而发展起来的，而性别分化对这个过程至关重要（Bodenhausen & Hugenberg， 2009;Buss，1991 年、1995 年、2015 年;Schillaci，2006 年）。也就是说，如果繁殖为进化奠定了基础，那么适应性生物应该已经发展出认知机制来克服生殖障碍。要克服这些障碍，区分男性和女性的能力是关键，跨多个领域的研究证实了这一点：人们表现出非凡的视觉区分两性的能力。人们可以快速准确地区分男性（棱角分明的面部特征，宽阔的肩膀）和女性（圆润的面部特征，宽大的臀部）（Ito & Urland，2005年;Johnson et al.， 2007;Macrae等人，2005年），并且从很小的时候开始（Martin & Ruble，2004年;Miller， 1983;Quinn et al.， 2002）。成年后，人类可以从有限的感知信息中区分性别，例如单面部特征（Cloutier et al.， 2005;Macrae & Martin，2007年）或步态的点光动画（Johnson et al.， 2007;Lick et al.， 2013）。

The skill with which humans differentiate between the sexes is not specific to visual cues but extends to information that arrives via other sensory modalities, such as voice (Lass et al., 1976; Pernet & Belin, 2012), touch (Slepian et al., 2011), and scent (Thornhill & Gangestad, 1999). Infants almost immediately learn to differentiate the mother from father (DeCasper & Fifer, 1980; Montague & Walker-Andrews, 2002), and subsequently male from female (Miller, 1983; Quinn et al., 2002). The evolution of this capacity is thought to reflect the survival advantage conferred on species that can quickly parse the social world into those who are and are not mate-worthy (Andersson, 1994; Buss, 1995, 2015). Given the adaptive benefits of distinguishing males from females, gender became a schema through which we differentiate, and attach meaning to, human beings. Across history and from early in the lifespan, there is frequently a demarcation between men and women.
人类区分性别的技能并不仅限于视觉线索，而是延伸到通过其他感觉方式（例如声音）到达的信息（Lass et al.， 1976;Pernet & Belin， 2012）， 触摸 （Slepian et al.， 2011） 和气味 （Thornhill & Gangestad， 1999）。婴儿几乎立即学会了区分母亲和父亲（DeCasper & Fifer， 1980;Montague & Walker-Andrews， 2002），随后由女性变为男性（Miller， 1983;Quinn et al.， 2002）。这种能力的进化被认为反映了赋予物种的生存优势，这些物种可以迅速将社会世界解析为值得交配和不值得交配的物种（Andersson，1994 年;Buss，1995 年、2015 年）。鉴于区分男性和女性的适应性好处，性别成为我们区分人类并赋予人类意义的图式。纵观历史，从生命的早期开始，男性和女性之间经常存在分界线。

Once this categorization process occurs, where an individual is identified as “male” or “female,” meaning (i.e., gender) is imposed on this distinction, guiding facile and efficient sensemaking about what the target is like and how they will likely behave (i.e., stereotyping; Allport, 1954; Ellemers, 2018; Fiske, 1998). Much work has highlighted the importance of the categorization and stereotyping processes in person perception (e.g., Allport, 1954; Bodenhausen & Hugenberg, 2009; Fiske & Taylor, 1991; Macrae et al., 1996). Gender stereotypes are especially meaningful, as the gender schema is a primary organizing cognitive structure that relies on gender stereotypes to categorize and interpret information, providing a scaffolding with which to understand ourselves and others (Bem, 1981; Markus & Oyserman, 1989; Markus et al., 1982; Martin & Slepian, 2020); as a consequence, without gender cues, individuals lack a schema to structure incoming information and guide their interactions with others. Such gender stereotypes, which we review next, are culturally created, context dependent, and stem from the historical division of labor between men and women.
一旦这个分类过程发生，一个人被确定为“男性”或“女性”，含义（即性别）被强加于这种区别，指导对目标是什么样子以及他们可能的行为如何进行简单而有效的意义建构（即刻板印象;Allport，1954 年;Ellemers，2018 年;Fiske，1998 年）。许多工作都强调了分类和刻板印象过程在个人感知中的重要性（例如，Allport，1954 年;Bodenhausen & Hugenberg， 2009;Fiske & Taylor， 1991;Macrae et al.， 1996）。性别刻板印象特别有意义，因为性别图式是一种主要的组织认知结构，它依靠性别刻板印象来分类和解释信息，为理解自己和他人提供了一个支架（Bem，1981 年;Markus & Oyserman， 1989;Markus et al.， 1982;Martin & Slepian，2020 年）;因此，如果没有性别线索，个体就缺乏构建传入信息并指导他们与他人互动的模式。我们接下来将回顾这种性别刻板印象，它们是文化创造的，取决于环境，并源于历史上男女之间的分工。

Males and females have some inherent physical differences (e.g., sex organs, height, strength), which shaped the social roles men and women have historically occupied (Eagly & Wood, 2013; Wood & Eagly, 2012, 2013). Men’s greater average physical size and strength made them better suited than women to tasks that involved agency (e.g., aggressiveness, dominance), such as hunting and defending. Women’s unique childbearing ability made them better suited to tasks requiring communality (e.g., warmth, nurturance), such as caretaking and childrearing (Diekman & Eagly, 2000; Eagly et al., 2000). Critically, this division of labor results in gender differences in the frequency with which men and women express agentic (masculine: dominant, assertive) and communal (feminine: warm, nurturing) behavior, respectively. This behavioral difference forms the basis of generalized beliefs about what men and women are generally like (descriptive stereotypes), which in turn creates expectations about how men and women should behave (prescriptive stereotypes; Ellemers, 2018; Rudman & Glick, 2008).
男性和女性有一些固有的身体差异（例如，性器官、身高、力量），这塑造了男性和女性历史上所扮演的社会角色（Eagly & Wood，2013;Wood & Eagly，2012年，2013年）。男性更大的平均体型和力量使他们比女性更适合从事涉及能动性（例如侵略性、支配性）的任务，例如狩猎和防御。女性独特的生育能力使她们更适合需要公共性的任务（例如，温暖、养育），例如照顾和育儿（Diekman & Eagly，2000;Eagly 等人，2000 年）。至关重要的是，这种分工导致男性和女性分别表达代理性（男性：主导、自信）和公共（女性：温暖、养育）行为的频率存在性别差异。这种行为差异构成了关于男性和女性通常是什么样子的普遍信念的基础（描述性刻板印象），这反过来又产生了对男性和女性应该如何行为的期望（规定性刻板印象;Ellemers，2018 年;Rudman & Glick， 2008）。

Because the gender schema is based on biological sex (and the aforementioned physical differences), people are especially likely to misperceive gender differences as biologically determined and immutable (Meyer & Gelman, 2016; see Diekman & Schmader, 2021 for discussion). Attesting to the significance individuals and societies ascribe to biological sex, the first question expecting parents are asked when sharing their news is one of biological sex (“Boy or girl?”; Slepian & Galinsky, 2016). From this point forward, society imposes gender upon this distinction. Gendered cues, such as names, clothing, and colors, demarcate whether someone is male or female, and society reproduces and reinforces this distinction, encouraging males and females to play different sports, accumulate different skills, pursue different interests, and display different behavior, which only continue to polarize male from female (Bem, 1981; Martin & Ruble, 2004). These differences are internalized, spawning a gender identity that typically, but need not, take a binary (male, female) form. Even those who identify as nonbinary or gender fluid are expected to conform to binary gender roles—gender takes on a reality outside individuals, being reproduced at institutional, interpersonal, societal, and cultural levels (Diekman & Schmader, 2021; Martin & Slepian, 2020).
因为性别模式是基于生物性别（以及上述身体差异）的，人们特别容易将性别差异误解为生物决定的且不可改变的（Meyer & Gelman，2016; 参见Diekman & Schmader，2021年的讨论）。为了证明个人和社会对生理性别的重要性，准父母在分享他们的新闻时被问到的第一个问题是生理性别的问题（“男孩还是女孩？Slepian & Galinsky，2016 年）。从这时起，社会将性别强加于这种区别。性别线索，如姓名、服装和颜色，划定了某人是男性还是女性，社会复制并强化了这种区别，鼓励男性和女性参加不同的运动，积累不同的技能，追求不同的兴趣，并表现出不同的行为，这只会继续使男性与女性两极分化（Bem，1981 年;Martin & Ruble， 2004）。这些差异被内化，产生了一种性别认同，这种性别认同通常但不需要采取二元（男性、女性）的形式。即使是那些认同为非二元性别或性别流动的人也被期望符合二元性别角色——性别在个体之外呈现现实，在机构、人际、社会和文化层面被复制（Diekman & Schmader，2021;Martin & Slepian，2020 年）。

Indeed, many scholars conceive of gender not as a biological characteristic determined at birth but as a behavior that one expresses—one does rather than has gender (Butler, 1990; West & Zimmerman, 1987). Doing gender is likened to a performance to an audience that has expectations about the focal actor’s behavior that are rooted in the social roles men versus women play (Butler, 1990; Goffman, 1959; Morgenroth & Ryan, 2018, 2020). Men are expected to behave and present as masculine (e.g., strong, tough) and women as feminine (e.g., tender, gentle; Bem, 1981; Ellemers, 2018). These social roles place pressure on an individual (i.e., an actor) to behave in accordance with their gender, shaping their actions and how they present themselves. Importantly, gender roles also shape how interaction partners (i.e., the audience) engage with the focal actor, cocreating such gender performativity. Without gender as a backdrop for organizing and making meaning of incoming data, “audiences” struggle to understand, evaluate, and interact with those who populate their social world. Indeed, many would struggle with the idea of a genderless human: What would they be like? How would they behave? How should we behave when interacting with them? People seem at a loss for how to proceed in interactions without this particular detail: People solicit gender information about an individual they have not met (Slepian & Galinsky, 2016), find it difficult not to mention gender when describing someone (Critcher & Ferguson, 2014), and feel discomfort when they cannot discern another person’s gender (Stern & Rule, 2018; West & Zimmerman, 1987).
事实上，许多学者认为性别不是出生时决定的生物学特征，而是一个人表达的一种行为——一个人有性别而不是有性别（Butler， 1990;West & Zimmerman，1987年）。做性别被比作对观众的表演，观众对焦点演员的行为有期望，这些期望植根于男性与女性所扮演的社会角色（Butler， 1990;Goffman，1959 年;Morgenroth & Ryan，2018年，2020年）。男性的行为举止和表现应为男性化（例如，强壮、坚韧），而女性应表现为女性化（例如，温柔、温柔;Bem， 1981;Ellemers，2018 年）。这些社会角色给个人（即演员）施加压力，要求他们根据自己的性别行事，塑造他们的行为和表现自己。重要的是，性别角色还塑造了互动伙伴（即观众）如何与焦点演员互动，共同创造了这种性别表演性。如果没有性别作为组织和理解传入数据的背景，“受众”就难以理解、评估和与居住在他们社交世界中的人互动。事实上，许多人会为无性别人类的想法而苦苦挣扎：他们会是什么样子？他们会如何表现？与他们互动时我们应该如何表现？ 如果没有这个特定的细节，人们似乎对如何进行互动感到困惑：人们会询问他们未曾见过的人的性别信息（Slepian & Galinsky，2016，发现在描述某人时很难不提及性别（Critcher & Ferguson，当他们无法辨别他人的性别时感到不适： 2018 年;West & Zimmerman，1987年）。

The central role gender plays in our interactions with others extends to other aspects of our lives as well, shaping how we perceive and interact with many nonhuman entities (Bem, 1981; Guthrie, 2007; Martin & Slepian, 2017, 2018, 2020; Stroessner et al., 2020; Wilkie & Bodenhausen, 2012). Many physical properties (e.g., shape, color, movement, etc.) are interpreted as gendered, giving insight into what an object is and for whom it is appropriate (Aggarwal & McGill, 2007; Martin & Slepian, 2020; Sherman & Zurbriggen, 2014). Gender provides a great deal of information about who someone is and what role they might play in our lives and, though the generalizations may be ill-fitting to a target, they are often “good-enough” impressions for everyday interaction purposes (Fiske, 1998).
性别在我们与他人互动中发挥的核心作用也延伸到我们生活的其他方面，塑造了我们如何感知和与许多非人类实体互动（Bem， 1981;Guthrie， 2007;Martin & Slepian， 2017， 2018， 2020;Stroessner等人，2020 年;Wilkie & Bodenhausen，2012 年）。许多物理属性（例如，形状、颜色、运动等）被解释为性别化，从而洞察到物体是什么以及它适合谁（Aggarwal & McGill，2007;Martin & Slepian， 2020;Sherman & Zurbriggen，2014 年）。性别提供了大量关于某人是谁以及他们在我们的生活中可能扮演什么角色的信息，尽管这些概括可能不适合目标，但它们通常是日常互动目的的“足够好”的印象（Fiske，1998）。

As we have reviewed, gender is an organizing structure that the mind uses to streamline and make meaning of the social world. Sex differences are ascribed considerable meaning and are reinforced and accentuated through an acculturation process. And while we know that gender is important in social evaluation, detected quickly and accurately (Hehman et al., 2017; Miller, 1983; Quinn et al., 2002) and streamlining processing of the social world (Bem, 1981; Lorber, 1994), whether gender is presumed to be a defining feature of humans, and how perceivers respond to its absence, are questions that have yet to be tested. Lack of attention to these questions is surprising given the extent to which gender theorists have discussed sex and personhood (see Bem, 1993; Butler, 1990; Nussbaum, 1999). Gender theorists have suggested that sex and humanness are intertwined and that those who do not conform or perform their gender “correctly” are denied full humanness and punished by others (Butler, 1990; Nussbaum, 1999; West & Zimmerman, 1987); however, “human” in these cases refers to the freedoms, rights, and protections granted to human beings (e.g., Butler, 1990; Nussbaum, 1999), quite distinct and disconnected from the psychological literature on humanization and the central question of how people conceive of human. Thus, before examining whether gender is a defining feature of perceiving humanness, it is first necessary to briefly review the psychological literature on humanization and the inverse process of dehumanization.
正如我们所回顾的，性别是大脑用来简化和理解社会世界的一种组织结构。性别差异被赋予了相当大的意义，并通过文化适应过程得到强化和强调。虽然我们知道性别在社会评价中很重要，但可以快速准确地检测到（Hehman et al.， 2017;Miller， 1983;Quinn et al.， 2002）以及简化社会世界的处理（Bem，1981 年;Lorber， 1994），性别是否被假定为人类的一个决定性特征，以及感知者如何应对它的缺失，这些都是尚未得到检验的问题。考虑到性别理论家对性和人格的讨论程度，对这些问题缺乏关注是令人惊讶的（参见 Bem， 1993;Butler， 1990;Nussbaum，1999 年）。性别理论家认为，性和人性是交织在一起的，那些不符合或“正确”表现性别的人被剥夺了完全的人性，并受到他人的惩罚（Butler， 1990;Nussbaum，1999 年;West & Zimmerman，1987年）;然而，在这些情况下，“人类”是指授予人类的自由、权利和保护（例如，Butler，1990 年;Nussbaum，1999 年），与关于人性化的心理学文献以及人们如何构想人类的核心问题截然不同且脱节。 因此，在研究性别是否是感知人类的决定性特征之前，首先有必要简要回顾一下关于人性化和非人化的逆向过程的心理学文献。

While the evolution, social perception, and sociology literature have outlined the importance of gender in various aspects of interpersonal, group, and cultural relationships, we now move toward gender’s importance in perceiving someone—or something—as human.
虽然进化论、社会认知和社会学文献已经概述了性别在人际关系、群体和文化关系各个方面的重要性，但我们现在正朝着性别在将某人或某物视为人类的重要性迈进。

A robust literature has examined the unique and distinct capacities fundamental to the human species, as well as those that differentiate us from animals (Gray et al., 2007; Haslam, 2006; Haslam & Loughnan, 2014; Waytz, Morewedge, et al., 2010). These include emotional experience, meta representation, explicit intentions, moral consideration, and the ability to reason—or, more simply, having “mind.” Existing scholarship has looked at both the process of humanization as well as its inverse: dehumanization (Epley et al., 2008; Haslam, 2006; Haslam & Loughnan, 2014; Kteily & Bruneau, 2017; Waytz, Morewedge, et al., 2010). This literature suggest people are especially likely to humanize targets that are capable of “experience” (e.g., hunger, pain, pleasure) and “agency” (e.g., self-control, planning, communication; Gray et al., 2007). They also speak to target characteristics that trigger attributions of mind, such as the presence of specific features (e.g., eyes) and their configuration (eyes-over-nose-over-mouth; Deska et al., 2017; Fincher et al., 2017; Hugenberg et al., 2016), as well as motion that is humanlike (Morewedge et al., 2007). In contrast, dehumanized individuals are seen similarly to animals or objects: denied the capacity to reason, behave with intention, and experience the complete range of human emotions (Epley, 2018; Gray et al., 2007; Haslam, 2006; Leyens et al., 2000; Waytz, Morewedge, et al., 2010).
大量文献研究了人类物种独特而独特的基本能力，以及将我们与动物区分开来的能力（Gray 等人，2007 年;Haslam，2006 年;Haslam & Loughnan， 2014;Waytz、Morewedge 等人，2010 年）。这些包括情感体验、元表征、明确的意图、道德考虑和推理能力——或者更简单地说，拥有“思想”。现有的学术研究既关注人性化的过程，也关注其反面：非人化（Epley et al.， 2008;Haslam，2006 年;Haslam & Loughnan， 2014;Kteily & Bruneau， 2017;Waytz、Morewedge 等人，2010 年）。这些文献表明，人们特别有可能将能够“体验”（例如，饥饿、痛苦、快乐）和“能动性”（例如，自我控制、计划、沟通;Gray et al.， 2007）。它们还涉及触发心理归因的目标特征，例如特定特征（例如，眼睛）的存在及其配置（眼睛在鼻子上;Deska et al.， 2017;Fincher等人，2017 年;Hugenberg et al.， 2016），以及类似人类的运动（Morewedge et al.， 2007）。相比之下，非人化的个体被视为类似于动物或物体：被剥夺了推理、有目的的行为和体验人类情感的全部范围的能力（Epley，2018 年;Gray 等人，2007 年;Haslam，2006 年;Leyens 等人。，2000 年;Waytz、Morewedge 等人，2010 年）。

In addition to identifying features of targets that trigger ascriptions of humanness, this literature speaks to top-down processes that determine when a target is especially likely to be humanized, such as the perceiver’s need to socially connect (Epley et al., 2008) or feel in control of their surroundings (Waytz, Cacioppo, et al., 2010; see Epley et al., 2007), as well as intergroup motives (Hackel et al., 2014), and motives to derogate others (Haslam, 2006). The body of work on (de)humanization reveals the extent to which a target is humanized lies both in the targets’ features and perceivers’ psychological needs.
除了识别触发人性归属的目标特征外，这些文献还谈到了自上而下的过程，这些过程决定了目标何时特别有可能被人性化，例如感知者需要社交联系（Epley 等人，2008 年）或感觉可以控制周围的环境（Waytz、Cacioppo 等人，2010 年;参见 Epley 等人， 2007 年），以及群体间动机（Hackel 等人，2014 年）和贬低他人的动机（Haslam，2006 年）。关于（去）人性化的工作揭示了目标人性化的程度取决于目标的特征和感知者的心理需求。

Although having a sex is certainly not sufficient to prompt the granting of “mind” to targets—it is not uniquely human—gender (the cultural attributes we impose upon biological sex) has been argued to be a defining feature of human life (Butler, 1990; Nussbaum, 1999). As reviewed above, gender is a cognitive backdrop that people use to structure their social world and is reinforced at multiple levels: intrapersonal, interpersonal, and societal (Diekman & Schmader, 2021), and as such, is likely implicated in “seeing human.” Indeed, many of the physical and psychological features of targets (e.g., appearance, movement, personality), as well as the desires and motives of perceivers (e.g., understand, connect, control), operate through a gendered lens. We investigate the special role gender plays in seeing human by examining whether (a) ascribing something with mind leads to attributing it a gender (more so than other social categories) and (b) stripping an entity of gender leads to seeing it as less human.
尽管发生性行为肯定不足以促使目标获得“思想”——它不是人类独有的——但性别（我们强加于生理性别的文化属性）一直被认为是人类生活的一个决定性特征（Butler，1990;Nussbaum，1999 年）。如上所述，性别是人们用来构建其社交世界的认知背景，并在多个层面上得到强化：个人内部、人际和社会层面（Diekman & Schmader，因此，可能与“看到人类”有关。事实上，目标的许多生理和心理特征（例如，外表、运动、个性）以及感知者的欲望和动机（例如，理解、连接、控制）都是通过性别化的镜头运作的。我们通过研究 （a） 将某物与思想归为某物是否会导致将其归为性别（比其他社会类别更重要）和 （b） 剥夺一个实体的性别会导致将其视为不那么人性化，从而研究性别在看待人类方面所发挥的特殊作用。

In demonstrating the central role of gender in humanization, we make several theoretical and practical contributions to research on (de)humanization, anthropomorphism, and social cognition. While much research has demonstrated the importance of gender in stereotyping and evaluating human beings (Fiske & Neuberg, 1990; Macrae & Bodenhausen, 2000), seeing someone—or something—as human precedes the application of stereotypes (Epley & Waytz, 2010; Harris & Fiske, 2009, 2011). Our findings suggest the application of gender may not solely be the result of evaluating human beings: It may be a precursor to it. Whereas most work in psychology assumes removing gendered information (e.g., gender-neutral names, unspecified targets) allows for a controlled, precise, and independent understanding of “humans” (Hester & Gray, 2020), our results suggest participants often attribute a gender when none is explicitly provided, and to the extent that they do not, the target is likely to be seen as less than fully human. Thus, this work indicates that greater consideration of gender is warranted by all psychologists, not just those who study it.
在展示性别在人性化中的核心作用时，我们为（去）人化、拟人化和社会认知的研究做出了多项理论和实践贡献。尽管许多研究表明性别在刻板印象和评估人类中的重要性（Fiske & Neuberg，1990;Macrae & Bodenhausen， 2000），将某人或某物视为人类，这先于刻板印象的应用（Epley & Waytz， 2010;Harris & Fiske，2009年，2011年）。我们的研究结果表明，性别的应用可能不仅仅是评估人类的结果：它可能是它的前兆。尽管心理学中的大多数工作都假设删除性别信息（例如，性别中立的名字，未指定的目标）可以对“人类”进行受控、精确和独立的理解（Hester & Gray，2020年，我们的结果表明，参与者通常会在没有明确提供的情况下归因性别，而且在他们没有提供的情况下，目标可能会被视为不完全的人类。因此，这项工作表明，所有心理学家都有必要更多地考虑性别，而不仅仅是研究性别的人。

Further, past gender research has theorized that gender plays a role in humanization (Butler, 1990; Guthrie, 2007; Martin & Slepian, 2020), yet this work is purely theoretical, and few (if any) of these ideas have been empirically tested (see Morgenroth & Ryan, 2018, p. 6). To our knowledge, we are the first to provide empirical evidence that having (or doing) gender is central to conceptions of humanness and to treat gender as a consequence (i.e., an outcome) of humanization. In doing so, we heed recent calls to study gender as a dependent variable that arises from social context rather than using a gender binary to predict differences in attitudes and behavior (Morgenroth & Ryan, 2018).
此外，过去的性别研究从理论上表明，性别在人性化中起着一定的作用（Butler，1990 年;Guthrie， 2007;Martin & Slepian，2020），然而这项工作纯粹是理论性的，这些想法中很少（如果有的话）被实证检验过（见Morgenroth & Ryan，2018年，第6页）。据我们所知，我们是第一个提供实证证据的人，证明拥有（或正在做）性别是人类概念的核心，并将性别视为人性化的结果（即结果）。在此过程中，我们响应了最近的呼吁，将性别作为由社会背景产生的因变量进行研究，而不是使用性别二元来预测态度和行为的差异（Morgenroth & Ryan），2018）。

We also add to the dehumanization literature by offering a social-cognitive account of dehumanization (also see Fincher et al., 2017; Hugenberg et al., 2016). Scholars have historically offered motivational accounts of dehumanization, presuming it arises from power and from intergroup motives (see Cameron et al., 2016; Gruenfeld et al., 2008; Waytz, Cacioppo, et al., 2010). Likewise, the growing literature on the discrimination experienced by transgender individuals supposes that the mistreatment arises from an ideological place: to keep the gender hierarchy intact (Prusaczyk & Hodson, 2020; Rye et al., 2019). We show that this discrimination can also arise from a schematic place, such that the absence of a gender schema to organize one’s thinking about a target leads people to grant them less than full humanness. Consistent with our argument that dehumanization need not arise from the aforementioned motives, but instead can be a consequence of social-cognitive processes, is recent evidence that the experience of processing faces configurally leads to the ascription of humanness (Deska et al., 2017; Fincher et al., 2017; Hugenberg et al., 2016). While this social-cognitive account is concerned with bottom-up visual processes that lead to ascriptions of humanness, ours focuses on how people conceive of humans; it is a top-down process rooted in memory rather than visual perception.
我们还通过提供非人化的社会认知解释来补充非人化文献（另见 Fincher 等人，2017 年;Hugenberg et al.， 2016）。学者们历来对非人化提供了动机性的描述，假设它源于权力和群体间的动机（参见 Cameron 等人，2016 年;Gruenfeld 等人，2008 年;Waytz、Cacioppo 等人，2010 年）。同样，关于跨性别个体所经历的歧视的文献越来越多，认为这种虐待源于一个意识形态的地方：保持性别等级的完整性（Prusaczyk & Hodson，2020;Rye等人，2019 年）。我们表明，这种歧视也可以从一个示意图的地方产生，以至于缺乏性别图式来组织一个人对目标的思考，导致人们赋予他们不完全的人性。与我们的论点一致，即非人化不一定源于上述动机，而是可能是社会认知过程的结果，最近的证据表明，处理面孔的体验有配置地导致了人性的归属（Deska 等人，2017 年;Fincher等人，2017 年;Hugenberg et al.， 2016）。虽然这个社会认知账户关注的是导致人性归类的自下而上的视觉过程，但我们的账户则关注人们如何构想人类;这是一个植根于记忆而不是视觉感知的自上而下的过程。

Finally, we make an important practical contribution: We add to the small but growing dialogue on the experience of gender-non-conforming individuals, especially those who identify as “genderless” (14% of the gender-non-conforming population, a population representing over 1.4 million individuals; Flores et al., 2016; Gender Census, 2020). We show that recent societal changes that grant people greater freedom in gender identity may run afoul of deep-seated beliefs people hold about essential features of humans (i.e., that they are in possession of a [binary] gender), perhaps leading perceivers to withhold full personhood from nonconformers, particularly those who identify as “genderless.” Dehumanization can set off a chain of negative consequences: targets are denied secondary emotions (Leyens et al., 2000), elicit feelings of disgust (Hodson & Costello, 2007), and are widely discriminated against (see Ames & Mason, 2012, for review). Notably, these are all outcomes faced by gender-nonconforming individuals in the modern world (see Stotzer, 2009, for review), with profound consequences for mental health, well-being, and beyond (see Valentine & Shipherd, 2018, for review). This work calls for more consideration of the cognitive processes (i.e., gendered cognition) and language (i.e., “genderless”) involved in the evolving conversations on, and changing notions of, gender in contemporary society.
最后，我们做出了一项重要的实际贡献：我们增加了关于性别不一致个体经历的规模虽小但不断增长的对话，尤其是那些被认定为“无性别”个体的个体（占性别不一致人口的 14%，人口超过 140 万人;Flores et al.， 2016;性别普查，2020 年）。我们表明，最近赋予人们更大性别认同自由的社会变化可能与人们对人类基本特征的根深蒂固的信念（即他们拥有 [二元]性别）相冲突，这可能会导致感知者对不合规者，特别是那些认同为“无性别”的人隐瞒完整的人格。非人化可能会引发一系列负面后果：目标被剥夺次要情绪（Leyens等人，2000年，引发厌恶感，并被广泛歧视（参见Ames & Mason，2012年，以供评论）。值得注意的是，这些都是现代世界中性别不一致的个体所面临的结果（参见Stotzer，2009年，对心理健康、福祉等产生深远影响（参见Valentine & Shipherd，2018年，查看）。这项工作呼吁更多地考虑当代社会中关于性别不断发展和不断变化的对话所涉及的认知过程（即性别认知）和语言（即“无性别”）。

Across 10 studies, we test the prediction that gender is a primary aspect of perceiving humanness, one that is more strongly tied to the process of humanization than other social categories to which people belong. It is important to note that our argument is that gender is critical in “seeing” human, not in “being” human; we are interested in conceptions of humanness—how people think about humans—and the implications for person perception; not in how humanity is experienced by individuals who do or do not identify with traditional binary gender. To examine our hypothesis, we induce anthropomorphism by having participants recall experiences, observe moving shapes, imagine nonhuman entities as people, and create a human form, and then measure social-category ascription. This empirical approach provides a strong test of the prediction that gender is more closely linked to conceptions of humanity than other social categories; in these studies, participants grant humanity and then report on social-category membership where these social categories technically do not exist.
在 10 项研究中，我们检验了性别是感知人类的主要方面的预测，与人们所属的其他社会类别相比，性别与人性化过程的联系更紧密。重要的是要注意，我们的论点是，性别在“看到”人类方面至关重要，而不是在“存在”人类方面;我们对人性的概念感兴趣——人们如何看待人类——以及对人感知的影响;而不是认同或不认同传统二元性别的个体如何体验人性。为了检验我们的假设，我们通过让参与者回忆经验、观察移动的形状、将非人类实体想象成人、创造人类形态，然后测量社会类别归属来诱导拟人化。这种实证方法为性别与其他社会类别相比与人性概念更紧密相关的预测提供了有力的检验;在这些研究中，参与者授予人性，然后报告这些社会类别在技术上不存在的社会类别成员资格。

To provide further evidence that gender is central to perceiving humanness, and to examine the consequences of withholding it, we then demonstrate that removing it from nonhuman entities (Study 5), groups of people (Study 6), alien species (Study 7), and individuals (Study 8) leads them to be seen as less human. Studies 6 and 7 also establish a causal mechanism: Removing gender from a target results in dehumanization because it strips the cognitive backdrop (i.e., schema) and concomitant stereotypic knowledge that guides perceivers’ sensemaking of the target. Finally, Study 8 illustrates some downstream consequences of stripping a target of gender: because such people are seen as less than fully human, perceivers consider them less relatable and more socially distant.
为了提供进一步的证据证明性别是感知人类的核心，并研究隐瞒它的后果，我们然后证明，将其从非人类实体（研究 5）、人群（研究 6）、外来物种（研究 7）和个体（研究 8）中去除会导致他们被视为不那么人性。研究 6 和 7 还建立了一个因果机制：从目标中删除性别会导致非人化，因为它剥夺了指导感知者对目标进行意义构建的认知背景（即图式）和随之而来的刻板印象知识。最后，研究 8 说明了剥夺目标性别的一些下游后果：因为这些人被视为不完全的人，所以感知者认为他们不太相关，而且在社交上更疏远。

Across all studies, our findings hold when we include participants removed for failing to follow instructions and/or when we control for demographic variables (e.g., participant gender, age, etc.). Data, materials, and preregistrations can be found at: https://osf.io/nu9d5/?view_only=103feff2079548b89407bb14f646186b (Martin & Mason, 2021). In Study 1, we did not have strong a priori predictions about the effect size of humanizing on gender ascription and collected 500 responses. As preregistered in Study 2a, we collected as many responses as possible in a week-long recruiting period. The effect sizes observed in Studies 1 and 2a (r = .30) suggested that at least 82 participants were needed to detect our within-participant effect of humanization on gendering (using G*Power 3.1; 1 − β = .80, α = .05). In each study, we collected responses from at least 100 participants to test our hypotheses. Sensitivity analyses across studies revealed that there was 80% power to detect effect sizes (r) from .12 to .27. The effect sizes found in our studies ranged from .24 to .41. All participants were recruited from the United States and all studies were approved by the Institutional Review Board at Stanford University (IRB 48342).
在所有研究中，当我们纳入因未遵循指示而被移除的参与者和/或当我们控制人口统计变量（例如，参与者的性别、年龄等）时，我们的研究结果成立。数据、材料和预注册可以在以下网址找到：https://osf.io/nu9d5/?view_only=103feff2079548b89407bb14f646186b（Martin & Mason，2021）。在研究 1 中，我们没有关于人性化对性别归属的影响大小的强有力的先验预测，并收集了 500 份回复。正如在研究 2a 中预先注册的那样，我们在为期一周的招募期内收集了尽可能多的回复。在研究 1 和 2a 中观察到的效应量 （r = .30） 表明，至少需要 82 名参与者来检测人源化对性别化的参与者内部影响（使用 G*Power 3.1;1 − β = .80，α = .05）。在每项研究中，我们收集了至少 100 名参与者的回答来检验我们的假设。跨研究的敏感性分析显示，有 80% 的功效可以检测从 .12 到 .27 的效应大小 （r）。我们研究中发现的效应量从 .24 到 .41 不等。所有参与者均从美国招募，所有研究均已获得斯坦福大学机构审查委员会 （IRB 48342） 的批准。

In our first four studies, we tested our primary hypothesis that gender and humanization go hand-in-hand. Specifically, we had participants ascribe humanness to a nonhuman entity and then measured the extent to which different social categories were attributed to the entity.
在我们的前四项研究中，我们检验了性别和人性化齐头并进的主要假设。具体来说，我们让参与者将人性归因于非人类实体，然后测量不同社会类别归因于该实体的程度。

Study 1 sought evidence for our hypothesis that gender is more central to conceptions of personhood than other social categories. To test this hypothesis, participants were asked to assign an entity with humanness and then indicated the extent to which the entity possessed a gender (as well as race, age, sexual orientation, religion, and disability).
研究 1 为我们的假设寻找证据，即性别比其他社会类别在人格概念中更重要。为了检验这一假设，参与者被要求分配一个具有人性的实体，然后指出该实体拥有性别的程度（以及种族、年龄、性取向、宗教和残疾）。

Study 1 participants (N = 501) were recruited from Amazon’s Mechanical Turk (MTurk) to take part in a study on “anthropomorphizing entities” in exchange for monetary compensation. Participants were given a definition of anthropomorphism, asked to describe something in their life that they anthropomorphize, and prompted to indicate whether the target item could be described as having a gender as well as membership in other social categories.
研究 1 参与者 （N = 501） 是从亚马逊的 Mechanical Turk （MTurk） 招募的，参加一项关于“拟人化实体”的研究，以换取金钱补偿。参与者被赋予拟人化的定义，要求描述他们生活中被拟人化的事物，并提示说明目标项目是否可以被描述为具有性别以及其他社会类别的成员身份。

Participants who failed an attention-check question asking them to identify examples of anthropomorphism were excluded from the analyses, as were participants who appeared to be bots (i.e., wrote nonsense answers to open-ended questions; see Supplementary Online Material; SOM; N = 70). Because we were interested in the extent to which people assign nonhuman entities properties that they do not technically have, we also excluded participants who recalled entities with a human form (e.g., dolls; N = 47) or a biological sex (e.g., animals; N = 101). Retaining these data increases the significance of the predicted effect but is a less stringent test of our hypothesis. This exclusion left a sample of 283 participants for analysis (48% men, 52% women, and 0.4% nonbinary individuals). The sample was made up of 75% White, 7% Asian, 9% Black, 8% Hispanic, and 1% Other-Race participants, with a mean age of 37.84 (SD = 12.02).
未通过要求他们识别拟人化示例的注意力检查问题的参与者被排除在分析之外，似乎是机器人的参与者也被排除在分析之外（即，为开放式问题写下无意义的答案;见补充在线材料;索姆;N = 70）。因为我们对人们在多大程度上赋予非人类实体在技术上没有的属性感兴趣，所以我们还排除了回忆具有人类形态的实体（例如，玩偶;N = 47）或生理性别（例如，动物;N = 101）。保留这些数据会增加预测效果的重要性，但对我们的假设的检验不那么严格。这种排除留下了 283 名参与者的样本进行分析（48% 男性，52% 女性和 0.4% 非二元性别个体）。样本由 75% 的白人、7% 的亚洲人、9% 的黑人、8% 的西班牙裔和 1% 的其他种族参与者组成，平均年龄为 37.84 （SD = 12.02）。

After providing informed consent, participants were given a definition of anthropomorphism (Epley et al., 2007; Waytz, Morewedge, et al., 2010; see Appendix) and then asked to think about something in their lives that they anthropomorphize (or could be anthropomorphized), describe the entity, and provide an example of the ways in which they treat it like it is human. We then measured the extent to which they ascribed the entity with a number of social categories, as well as the extent to which they saw their entity as human.
在提供知情同意后，参与者被赋予了拟人化的定义（Epley 等人，2007 年;Waytz、Morewedge 等人，2010 年;见附录），然后被要求思考他们生活中被拟人化（或可以拟人化）的事情，描述该实体，并提供一个例子，说明他们如何像对待人类一样对待它。然后，我们测量了他们将实体归类为多个社会类别的程度，以及他们将自己的实体视为人类的程度。

To ensure our manipulation prompted humanization and to capture individual variance in humanization, we used a modified version of the blatant dehumanization scale (Kteily et al., 2015), which captures the continuity of how “human-like” people—or in this case, nonhuman entities—seem. This scale is a well-validated measure of overt dehumanization (Kteily & Bruneau, 2017). Participants were told, “the following scale represents humanness levels. Zero represents a very low degree of humanness and 100 represents a very high degree of humanness. Choose a number that represents the humanness of the entity you just described” on a scale from 0 (not at all human) to 100 (fully human).  ^{[  2  ]} 
为了确保我们的操作促进了人性化并捕捉人源化中的个体差异，我们使用了公然非人化量表的修改版本（Kteily 等人，2015 年），它捕捉了“类人”人——或者在这种情况下，非人类实体——看起来的连续性。这个量表是公认的非人化的衡量标准（Kteily & Bruneau，2017）。参与者被告知，“以下量表代表人类水平。0 代表非常低的人性程度，100 代表非常高的人性程度。选择一个代表您刚刚描述的实体的人性的数字“，范围从 0（根本不是人类）到 100（完全人类）。    ^{[ 2  ]} 

As an alternative measure to capture the extent to which the entity was attributed distinctly human traits, we used Waytz, Morewedge, et al.’s (2010) anthropomorphism measure. Specifically, participants were asked to what extent they see the entity as possessing: (a) intentions, (b) free will, (c) emotions, (d) consciousness, and (e) a mind of its own (1 = not at all to 5 = very much; α = .88).
作为捕捉实体被明显归因于人类特征的程度的替代措施，我们使用了 Waytz、Morewedge 等人 （2010） 的拟人化测量。具体来说，参与者被问及他们在多大程度上认为该实体拥有：（a） 意图，（b） 自由意志，（c） 情绪，（d） 意识，以及 （e） 自己的思想（1 = 完全没有到 5 = 非常多;α = .88）。

To capture social-category ascription, we used the most common essentialized social categories—or those seen as informative in defining people and groups (Haslam et al., 2000; Prentice & Miller, 2007). Participants were asked, “Do any of the following apply to the entity you just described?” and indicated the extent to which the entity had a (a) gender (e.g., man, woman), (b) race (e.g., African American, Asian), (c) generational cohort (e.g., millennial, baby-boomer),  ^{[  3  ]}  (d) religion (e.g., Catholic, Jewish), (e) disability (e.g., physical, mental health), and (e) sexual orientation (e.g., gay, bisexual), in randomized order on a scale from 1 = not at all applicable to 5 = very applicable.
为了捕捉社会类别归属，我们使用了最常见的本质化社会类别——或者那些被认为在定义人和群体方面提供信息量的社会类别（Haslam et al.， 2000;Prentice & Miller，2007 年）。参与者被问到，“以下任何一项是否适用于您刚刚描述的实体？”，并指出该实体具有 （a） 性别（例如，男性、女性）、（b） 种族（例如，非裔美国人、亚洲人）、（c） 代际群体（例如，千禧一代、婴儿潮一代）、 ^{[  3  ]}  （d） 宗教（例如，天主教、犹太教）、（e） 残疾（例如，身体、 心理健康）和 （e） 性取向（例如，男同性恋、双性恋），以 1 = 完全不适用到 5 = 非常适用的等级随机排序。

The most common entities anthropomorphized were: vehicles (N = 116), technology (N = 79), nature (N = 25), household objects (N = 15), vacuums (N = 7), and furniture (N = 6). To confirm that participants followed instructions, we looked at the means for our general humanization and anthropomorphism scales. Participants ascribed their entities humanness as evidenced by both the humanization (M = 61.08, SD = 25.21) and anthropomorphism (M = 3.28, SD = .98) scores being significantly higher than the scale midpoints (both ps < .001; see Table 1 and SOM, for exact descriptive and inferential statistics for all studies).
最常见的拟人化实体是：车辆 （N = 116）、技术 （N = 79）、自然 （N = 25）、家居用品 （N = 15）、吸尘器 （N = 7） 和家具 （N = 6）。为了确认参与者遵循了指示，我们查看了一般人性化和拟人化量表的平均值。参与者将他们的实体归因于人性，人化 （M = 61.08， SD = 25.21） 和拟人化 （M = 3.28， SD = .98） 分数明显高于量表中点（均 ps < .001;见表 1 和 SOM，了解所有研究的确切描述性和推理统计数据）。

 放大此图像。

Ratings of social-category membership were submitted to a repeated-measures analysis with a single, six-level factor (social group: gender, race, generational cohort, religion, disability, and sexual orientation). Results revealed a significant difference in ratings across the six social groups, Wilks λ = .36, F(5, 278) = 100.41, p < .001, η _p ²  = .64. In line with our hypothesis, participants were significantly more likely to ascribe the entity a gender (M = 3.45, SD = 1.54) as compared to a race (M = 1.60, SD = 1.03), generational cohort (M = 2.02, SD = 1.29), sexual orientation (M = 1.71, SD = 1.13), disability (M = 1.89, SD = 1.21), and religion (M = 1.33, SD = .73). See Table 1.
将社会类别成员的评级提交给具有单个六级因素（社会群体：性别、种族、代际群体、宗教、残疾和性取向）的重复测量分析。结果显示，六个社会群体的评分存在显著差异，Wilks λ = .36，F（5， 278） = 100.41，p < .001，η _p ²  = .64。根据我们的假设，与种族 （M = 1.60， SD = 1.03）、代际队列 （M = 2.02， SD = 1.29）、性取向 （M = 1.71， SD = 1.13）、残疾 （M = 1.89， SD = 1.21） 和宗教 （M = 1.33， SD = .73） 相比，参与者更有可能将实体归因于性别 （M = 3.45， SD = 1.54） 和宗教 （M = 1.33， SD = .73）。请参阅 表 1。

According to ratings on the humanization and anthropomorphism measures, participants varied in the extent to which they humanized their entity. As a secondary test of our prediction that anthropomorphizing leads people to see gender, we tested the extent to which participants’ scores on these scales predicted social-category ascription using a multivariate regression with the social categories as simultaneous predictors. Gender was the only social category that uniquely predicted humanization, b = 3.78, SE = 1.01, t(276) = 3.73, p < .001, CI₉₅ = 1.78, 5.77, and anthropomorphism, b = .14, SE = .04, t(276) = 3.61, p < .001, CI₉₅ = .064, .22 (all other social categories, ps > .06; see SOM, for all regression analyses). Thus, gender appears to be more central to conceptions of humanness than other social categories to which people belong.
根据对人源化和拟人化措施的评级，参与者对实体人源化的程度各不相同。作为我们预测拟人化导致人们看到性别的二次测试，我们使用多元回归测试了参与者在这些量表上的分数在多大程度上预测了社会类别归属，并将社会类别作为同时预测因子。性别是唯一唯一能预测人性化的社会类别，b = 3.78，SE = 1.01，t（276） = 3.73，p < .001，CI₉₅ = 1.78,5.77，拟人化，b = .14，SE = .04，t （276） = 3.61，p < .001，CI ₉₅ = .064，.22（所有其他社会类别，ps > .06;有关所有回归分析，请参阅 SOM）。因此，性别似乎比人们所属的其他社会类别在人类概念中更重要。

Study 1 provided an initial test of our hypothesis, demonstrating that ascribing a nonhuman entity humanness leads the majority of people to see it as having some degree of gender. In line with past work that speaks to the ubiquity of the tendency to “see” humanness, nearly all participants recalled something in their life that they anthropomorphized. Building on this work, we show that assigning a target humanness leads people to conceive it as having a gender, more so than membership in other social categories. Most participants ascribed gender to their entity, and the more humanlike they rated the entity, the more they gendered it. This latter effect did not occur for the other social categories. That is, when accounting for shared variance among social-category ascription, only gender predicted the tendency to conceptualize an anthropomorphized entity as human.
研究 1 对我们的假设进行了初步测试，表明将非人类实体归类为人类会导致大多数人认为它具有某种程度的性别。与过去的作品一致，这些工作说明了“看到”人性的趋势无处不在，几乎所有参与者都回忆起了他们生活中被拟人化的东西。在这项工作的基础上，我们表明，分配目标人类性会导致人们将其视为具有性别，这比其他社会类别的成员身份更重要。大多数参与者将性别归因于他们的实体，他们对实体的评价越像人类，他们就越能将其性别化。后一种影响没有发生在其他社会类别中。也就是说，当考虑社会类别归属之间的共同差异时，只有性别预测了将拟人化实体概念化为人类的趋势。

Study 1 provided initial support that gender is more central to conceptions of personhood than other social categories are. Study 2a tests if this effect emerges even if participants are not explicitly instructed to anthropomorphize the target and instead passively observe stimuli that induce anthropomorphism.
研究 1 提供了初步支持，即性别比其他社会类别在人格概念中更重要。研究 2a 测试这种效果是否出现，即使没有明确指示参与者将目标拟人化，而是被动观察诱发拟人化的刺激。

Study 2a was preregistered (https://aspredicted.org/blind.php?x=4b3qv5) and examined whether people ascribe gender, more than other social categories, onto spontaneously anthropomorphized nonhuman entities. Students (N = 152) at a West Coast university participated for monetary compensation. Participants who failed the attention check (identify shapes in the video; N = 5) were excluded. The final sample included 33% men, 65% women, and 1% nonbinary individuals; 29% were White, 35% were Asian, 9% were Black, 15% were Hispanic, and 8% were of another race. The mean age was 23.16 (SD = 5.64). Participants watched an animated video of Heider and Simmel’s (1944) stimuli. The video depicts three geometric figures moving in various directions and at various speeds (see https://www.youtube.com/watch?v=VTNmLt7QX8E), and has been shown to be a subtle induction of anthropomorphism. Although our participants were not explicitly instructed to anthropomorphize the shapes, most people tend to do so (Heider & Simmel, 1944). After viewing the animation, participants were prompted to write four-to-five sentences describing what happened in the video and were given an attention check and the same measures as Study 1.
研究 2a 是预先注册的 （https://aspredicted.org/blind.php?x=4b3qv5），并检查了人们是否比其他社会类别更多地将性别归因于自发拟人化的非人类实体。西海岸一所大学的学生 （N = 152） 参加了金钱补偿。未通过注意力检查的参加者（识别视频中的形状;N = 5） 被排除在外。最终样本包括 33% 的男性、65% 的女性和 1% 的非二元性别个体;29% 是白人，35% 是亚裔，9% 是黑人，15% 是西班牙裔，8% 是其他种族。平均年龄为 23.16 （SD = 5.64）。参与者观看了 Heider 和 Simmel （1944） 刺激的动画视频。该视频描绘了三个几何图形以不同的方向和不同的速度移动（见 https://www.youtube.com/watch?v=VTNmLt7QX8E），并被证明是拟人化的微妙诱导。尽管我们的参与者没有被明确指示将形状拟人化，但大多数人倾向于这样做（Heider & Simmel， 1944）。观看动画后，参与者被提示写四到五个句子来描述视频中发生的事情，并接受注意力检查和与研究 1 相同的措施。

As in Study 1, participants were told, “the following scale represents humanness levels. Choose a number that represents the humanness of the shapes” (0 = not at all human to 100 = fully human; Kteily et al., 2015); they were administered Waytz, Morewedge, et al. (2010) anthropomorphism measure and instructed to rate the extent to which they saw the shapes as possessing five human qualities (1 = not at all to 5 = very much; α = .90).
与研究 1 一样，参与者被告知，“以下量表代表人类水平。选择一个代表形状人性的数字“（0 = 完全不是人类到 100 = 完全人类;Kteily et al.， 2015）;他们接受了 Waytz、Morewedge 等人 （2010） 拟人化测量，并被指示对他们认为这些形状具有五种人类品质的程度进行评分（1 = 完全没有到 5 = 非常多;α = .90）。

Participants were asked: “Do any of the following apply to the shapes?” and instructed to rate the same social categories used in Study 1 (gender, race, generational cohort, religion, disability, sexual orientation) on a scale that ranged from 1 = not at all applicable to 5 = very applicable. The order in which the social categories were presented was randomized.
参与者被问到：“以下任何一项是否适用于这些形状？”，并指示对研究 1 中使用的相同社会类别（性别、种族、代际群体、宗教、残疾、性取向）进行评分，范围从 1 = 完全适用到 5 = 非常适用。社会类别的呈现顺序是随机的。

Participants ascribed their entities humanity; ratings on both the humanization (M = 61.95, SD = 29.44) and anthropomorphism (M = 3.84, SD = 1.03) measures were significantly higher than the scale midpoints; both ps < .001. Consistent with the hypothesis that people assign humanized entities a gender more than membership in other social categories, we find significant differences across the social-category measures, Wilks λ = .59, F(5, 139) = 19.69, p < .001, η _p ²  = .42, with gender ascription (M = 2.69, SD = 1.59) being larger than any other social-category ascription (M _Race = 1.45, SD = 0.92; M _GenCohort = 1.77, SD = 1.13; M _Religion = 1.50, SD = 0.94; M _Disability = 1.65, SD = 1.04; M _SexOrient = 1.81, SD = 1.18). See Table 1.
参与者将他们的实体赋予了人性;人化 （M = 61.95， SD = 29.44） 和拟人化 （M = 3.84， SD = 1.03） 测量的评分均显著高于量表中点;两个 PS < .001。与人们比其他社会类别的成员更多地赋予人源化实体性别的假设一致，我们发现社会类别测量之间存在显着差异，Wilks λ = .59，F（5,139） = 19.69，p < .001，η _p ²  = .42，性别归属 （M = 2.69，SD = 1.59） 大于任何其他社会类别归属 （M_种族 = 1.45，SD = 0.92; M_GenCohort = 1.77，SD = 1.13;M_宗教 = 1.50，SD = 0.94;M_残疾 = 1.65，SD = 1.04;M_SexOrient = 1.81，SD = 1.18）。 请参阅 表 1。

As a secondary test of our prediction that gender is primary in humanization, we examined the extent to which participants’ scores on these scales predict social-category ascription using a multivariate regression with the social categories as simultaneous predictors. Gender was the only category that remained significant, b = 6.35, SE = 1.85, t(141) = 3.44, p < .001, CI₉₅ = 2.70, 10.00 (other ps > .15; see SOM, for additional social category analyses). The same pattern emerged when we used the anthropomorphism scores, b = .21, SE = .062, t(141) = 3.44, p < .001, CI₉₅ = .09, .33, suggesting that gender was the only social category that shared unique variance with humanization.
作为我们对性别是人性化主要预测的二次测试，我们使用多元回归（以社会类别作为同时预测因子）检查了参与者在这些量表上的分数在多大程度上预测社会类别归属。性别是唯一仍然显著的类别，b = 6.35，SE = 1.85，t（141） = 3.44，p < .001，CI₉₅ = 2.70,10.00（其他 ps > .15;参见 SOM，有关其他社会类别分析）。 当我们使用拟人化分数时，出现了相同的模式，b = .21，SE = .062，t（141） = 3.44，p < .001，CI₉₅ = .09，.33，这表明性别是唯一与人性化共享独特差异的社会类别。

Thus far, our studies have measured anthropomorphism and tested whether this process is associated with ascriptions of gender. So that we can make a causal argument, Study 2b manipulates anthropomorphism and then measures its effect on social-category ascription. In doing so, we provide an important boundary of our effect: humanization. That is, if gender is central to humanization, it should arise specifically from ascribing humanness to an entity and not from just describing or observing an entity more generally. Further, Study 2b included additional humanization measures to show that the results hold irrespective of how we measure humanization. As noted, we are most interested in overall conceptions of humanness; yet, as reviewed, humanization is multifaceted, ranging from characteristics that are seen as essential to being human (human nature) to those unique to being human (human uniqueness) and emotional capacities that only humans experience (e.g., guilt, pride). In Study 2b, we include these measures to shed light on the specific characteristics that may be implicated in the gender-humanization link.
到目前为止，我们的研究已经测量了拟人化，并测试了这个过程是否与性别归属有关。为了我们可以做出因果论证，研究 2b 操纵拟人化，然后测量其对社会类别归属的影响。在此过程中，我们提供了我们效果的一个重要边界：人性化。也就是说，如果性别是人性化的核心，那么它应该特别源于将人性赋予一个实体，而不仅仅是更普遍地描述或观察一个实体。此外，研究 2b 包括额外的人源化措施，以表明无论我们如何衡量人性化，结果都成立。如前所述，我们对人类的整体概念最感兴趣;然而，正如所评论的，人性化是多方面的，从被视为作为人类所必需的特征（人性）到作为人类的独特特征（人类独特性）和只有人类才能体验的情感能力（例如，内疚、自豪）。在研究 2b 中，我们纳入了这些措施，以阐明性别与人性化联系可能涉及的具体特征。

Participants (N = 322) from MTurk were recruited to take part in a study on “perceiving moving objects.” We used the same criteria as Study 2a to identify participants who should be excluded (N = 46); the final sample consisted of 276 participants with a mean age of 37.58 (SD = 11.50), consisting of 60% men, 40% women, and 0% nonbinary participants, and including 76% White, 8% Asian, 9% Black, 5% Hispanic, and 1% Other-Race participants. The study had a single between-participant factor (process: anthropomorphize or describe). After providing consent, participants watched the Heider and Simmel (1944) animation either forwards (anthropomorphism condition) or in reverse (describe condition). Participants in the anthropomorphism condition were provided a definition of what it entails (see Appendix) and instructed to think about each “person” in the video: what they were like and how they behaved. By contrast, participants in the describe condition were shown the Heider and Simmel’s (1944) animation in reverse, which has been shown to not induce anthropomorphism (see Schweitzer & Waytz, 2020), and asked to think about each “shape” in the video: How they moved and what impressions the participants had of them. See Appendix. All participants then completed the dependent measures and were thanked and paid.
来自 MTurk 的参与者 （N = 322） 被招募参加一项关于“感知移动物体”的研究。我们使用与研究 2a 相同的标准来确定应排除的受试者 （N = 46）;最终样本由 276 名参与者组成，平均年龄为 37.58 （SD = 11.50），包括 60% 男性、40% 女性和 0% 非二元性别参与者，包括 76% 白人、8% 亚裔、9% 黑人、5% 西班牙裔和 1% 其他种族参与者。该研究有一个单一的参与者间因素（过程：拟人化或描述）。在提供同意后，参与者观看了 Heider 和 Simmel （1944） 的动画，要么是正向的（拟人化条件），要么是反向的（描述条件）。向拟人化条件的参与者提供了其含义的定义（见附录），并指示他们思考视频中的每个“人”：他们是什么样的，他们的行为如何。相比之下，描述条件中的参与者被反向展示了 Heider 和 Simmel （1944） 的动画，这已被证明不会引起拟人化（参见 Schweitzer & Waytz，2020 年），并要求思考视频中的每个“形状”：他们如何移动以及参与者对他们的印象如何。见附录。然后，所有参与者都完成了相关措施，并获得了感谢和报酬。

As with the past studies, participants rated the shapes’ level of humanness on a scale from 0 = not at all human to 100 = fully human. Participants also rated the extent to which the shapes were anthropomorphized, using the same five uniquely human traits, on a scale from 1 = not at all to 5 = very much (α = .92; Waytz, Morewedge, et al., 2010).
与过去的研究一样，参与者以从 0 = 根本不是人类到 100 = 完全人类的等级对形状的人类水平进行评分。参与者还使用相同的五个独特的人类特征对形状拟人化的程度进行了评分，范围从 1 = 完全没有到 5 = 非常多 （α = .92;Waytz、Morewedge 等人，2010 年）。

To test if specific aspects of “humanness” more strongly predict the ascription of gender than the general humanization process, we administered three additional scales, which measure the ascription of essential human qualities (human nature), unique human qualities (human uniqueness), and human emotions. To measure human nature and uniqueness, we adapted 12 items from Bastian and Haslam (2010), 6 related to human nature (e.g., “the shapes had interpersonal warmth,” α = .84), and 6 related to human uniqueness (e.g., “the shapes were refined and cultured,” α = .62; 1 = not at all to 7 = very much so). To capture the ascription (or denial) of secondary emotions (e.g., hope, guilt; Leyens et al., 2000) to the shapes, participants rated the extent to which the shapes could experience the following emotions on a scale from 1 = not at all to 10 = very much: hope, compassion, optimism, bitterness, contempt, and guilt (Costello & Hodson, 2011; Kteily et al., 2015; Leyens et al., 2000; α = .92).
为了测试“人性”的特定方面是否比一般的人性化过程更能预测性别的归属，我们进行了三个额外的量表，它们衡量基本人类品质（人性）、独特人类品质（人类独特性）和人类情感的归属。为了衡量人性和独特性，我们改编了 Bastian 和 Haslam （2010） 的 12 个项目，其中 6 个与人性有关（例如，“形状具有人际温暖”，α = .84），6 个与人类独特性有关（例如，“形状经过提炼和培养”，α = .62;1 = 完全没有，7 = 非常如此）。捕捉次要情绪的归属（或否认）（例如，希望、内疚;Leyens et al.， 2000）对于形状，参与者对形状能够体验到以下情绪的程度进行了评分，范围从1 = 完全没有到10 = 非常多：希望、同情、乐观、苦涩、轻蔑和内疚（Costello & Hodson，2011;Kteily等人，2015 年;Leyens et al.， 2000;α = .92).

As with Studies 1 and 2a, participants were asked to rate the extent to which the following social categories applied to the shapes on a scale from 1 = not at all applicable to 5 = very applicable: gender, race, generational cohort, religion, disability, and sexual orientation. Social categories were presented in randomized order.
与研究 1 和 2a 一样，参与者被要求以 1 = 完全不适用到 5 = 非常适用的等级对以下社会类别适用于形状的程度进行评分：性别、种族、代际群体、宗教、残疾和性取向。社交类别以随机顺序呈现。

The results reveal that our manipulation induced humanization: The shapes were rated more “human” in the anthropomorphism condition (M = 76.18, SD = 22.82) as compared to the describe (control) condition (M = 48.26, SD = 33.16), F(1, 274) = 66.20, p < .001, η _p ²  = .20. The manipulation also had the same effect on other measures of humanization: anthropomorphism (M _anthro = 4.22, SD = .80; M _control = 3.35, SD = 1.11), human nature (M _anthro = 4.93, SD = 1.21; M _control = 3.94, SD = 1.30), human uniqueness (M _anthro = 4.43, SD = 1.03; M _control = 3.92, SD = .92), and human emotions (M _anthro = 6.26, SD = 2.56; M _control = 3.23, SD = 2.93); all ps < .001.
结果显示，我们的操作诱导了人性化：与描述（控制）条件 （M = 48.26， SD = 33.16） 相比，形状在拟人化条件 （M = 76.18， SD = 22.82） 中被评为更“人性化”，F（1， 274） = 66.20，p < .001，η _p ²  = .20。这种操纵对人性化的其他衡量标准也产生了相同的影响：拟人化（M_anthro = 4.22，SD = .80; M_对照 = 3.35，SD = 1.11），人性 （M_anthro = 4.93，SD = 1.21; M_对照 = 3.94，SD = 1.30），人类独特性 （M_anthro = 4.43，SD = 1.03; M_对照 = 3.92，SD = .92）和人类情感 （M_anthro = 6.26，SD = 2.56; M_对照 = 3.23，SD = 2.93）; 所有 PS < .001。

To test whether gender is more central to conceptions of humanity than other social categories, and that its ascription is specific to targets that have been humanized rather than described, we used a mixed-model ANOVA with one between-subject (process: anthropomorphize, describe) and one within-subject (social category: gender, race, generational cohort, religion, disability, sexual orientation) factor. We find the predicted interaction, Wilks λ = .83, F(5, 270) = 11.07, p < .001, η _p ²  = .17. Gender was more likely to be ascribed by participants in the anthropomorphize condition as compared to the describe condition, F(1, 274) = 68.38, p < .001. While anthropomorphizing relative to describing significantly increased ratings for other social categories (η _ps ²  < .05), it did not do so to the same extent as it did for gender (η _p ²  = .20). Further, when we restrict our analyses to participants in the anthropomorphism condition, we observe that participants were more likely to ascribe the shapes a gender (M = 3.02, SD = 1.59) as compared to other social categories (Ms < 2.18; ps < .001 for all pairwise comparisons; see Table 1 and SOM). This was not the case for participants in the describe (control) condition, who were unlikely to ascribe any social categories to the shapes, including gender (Ms < 1.67; see Table 1, for all means).
为了测试性别是否比其他社会类别在人类概念中更重要，以及其归属特定于被人性化而不是描述的目标，我们使用了一个混合模型方差分析，其中包含一个主体间（过程：拟人化、描述）和一个主体内（社会类别：性别、种族、代际群体、宗教、残疾、性取向）因素。我们找到预测的交互作用，Wilks λ = .83，F（5， 270） = 11.07，p < .001，η _p ²  = .17。 与描述条件相比，拟人化条件下的参与者更有可能归因于性别，F（1， 274） = 68.38，p < .001。虽然拟人化 相对于描述其他社会类别的评分显着增加 （η _ps ²  < .05），但它的程度与性别 （η _p ² = .20）。此外，当我们将分析限制在拟人化条件下的参与者时，我们观察到与其他社会类别（M = 3.02，SD = 1.59）相比（<M = 3.02，SD = 1.59;所有成对比较的 ps < .001;见表 1 和 SOM）。对于描述（对照）条件的参与者来说，情况并非如此，他们不太可能为形状赋予任何社会类别，包括性别（Ms < 1.67;见表 1，了解所有方式）。

Because some participants anthropomorphized the shapes more than others, we then conducted supplementary analyses to determine whether the variance in humanizing positively predicts ascriptions of gender across each of the five humanization measures administered. Consistent with Studies 1 and 2a, when we restrict our analyses to participants who were instructed to anthropomorphize (vs. describe) the target entity, and include all social-category ascriptions as simultaneous predictors of humanization, only gender emerges as significant, b = 5.97, SE = 1.34, t(130) = 4.44, p < .001, CI₉₅ = 3.30, 8.61 (other ps > .47). This result similarly emerged for anthropomorphism, b = .19, SE = .047, t(130) = 4.06, p < .001, CI₉₅ = .10, .29, human nature, b = .35, SE = .07, t(130) = 5.02, p < .001, CI₉₅ = .21, .49, human uniqueness, b = .28, SE = .060, t(130) = 4.63, p < .001, CI₉₅ = .16, .39, and human emotions, b = .67, SE = .15, t(130) = 4.46, p < .001, CI₉₅ = .37, .97.
因为一些参与者比其他人更多地拟人化了形状，所以我们随后进行了补充分析，以确定人性化的差异是否积极预测了所施用的五种人格化措施中的每一种的性别归属。与研究 1 和 2a 一致，当我们将分析限制在被指示拟人化（而不是描述）目标实体的参与者，并将所有社会类别归类作为人化的同时预测因子时，只有性别出现显着，b = 5.97，SE = 1.34，t（130） = 4.44，p < .001，CI ₉₅ = 3.30,8.61 （其他 ps > .47）。 这个结果同样出现在拟人化中，b = .19，SE = .047，t（130） = 4.06，p < .001，CI₉₅ = .10，.29，人性，b = .35，SE = .07，t （130） = 5.02，p < .001，CI ₉₅ = .21，.49，人类独特性，b = .28，SE = .060，t（130） = 4.63，p < .001，CI ₉₅ = .16，.39，以及人类情感，b = .67，SE = .15，t（130） = 4.46，p < .001，CI₉₅ = .37，.97。

Study 2b builds on Studies 1 and 2a by providing evidence that the relationship between “seeing” human and ascribing gender is a causal one, and that the ascription of gender specifically arises from “seeing human” rather than from describing entities more generally. Participants prompted to anthropomorphize shapes also assigned them a gender, an effect that did not occur in the describe (control) condition. Confirming our argument that gender is more central to conceptions of personhood, gender was ascribed to anthropomorphized targets to a significantly greater extent than the other social categories. Further, Study 2b demonstrated that the relationship between ascribing humanness and attributing gender occurred across a variety of established humanization measures: Whether thinking about overall humanization, innately or uniquely human qualities, or specific human emotions, gender is the only social category that uniquely predicted humanizing, suggesting that gender takes primacy in humanization in a way that cannot be said of other social categories.
研究 2b 建立在研究 1 和 2a 的基础上，提供了证据，证明“看到”人类和归因于性别之间的关系是因果关系，并且性别的归属具体源于“看到人类”，而不是更普遍地描述实体。被提示拟人化形状的参与者还为他们分配了性别，这种效果在描述（控制）条件下没有发生。证实了我们的论点，即性别在人格概念中更为重要，性别被归类为拟人化目标的程度明显大于其他社会类别。此外，研究 2b 表明，归因于人类性和归因于性别之间的关系发生在各种已建立的人性化措施中：无论是考虑整体人性化、天生的或独特的人类品质，还是特定的人类情感，性别是唯一唯一预测人性化的社会类别，这表明性别在人性化中占据首要地位，这是其他社会类别无法说的。

Study 3a sought to build on these results by establishing that the relationship between gender and humanization occurs in contexts commonly used in anthropomorphism research. Study 3a also examined whether the assignation of gender to anthropomorphized entities happens in the absence of any explicit intention by including an incidental measure of gendering. Specifically, we coded participants’ responses to an open-ended question for gender-specific references (e.g., pronouns, titles). This open-ended question, which asked participants to write about the ways in which they would anthropomorphize the product (see Appendix), was administered prior to the measures we used to explicitly assess social-category ascriptions.
研究 3a 试图通过确定性别与人化之间的关系发生在拟人化研究中常用的背景下来建立这些结果。研究 3a 还通过包括性别化的附带措施，研究了在没有任何明确意图的情况下，是否在没有任何明确意图的情况下将性别分配给拟人化实体。具体来说，我们对参与者对开放式问题的回答进行了编码，以获得针对性别的参考（例如，代词、标题）。这个开放式问题要求参与者写下他们将产品拟人化的方式（见附录），在我们用于明确评估社会类别归属的措施之前进行。

As in Study 2b, this study had a single between-participant factor (process: anthropomorphize or describe). Participants (N = 401) were recruited from MTurk to take part in a study on “Anthropomorphizing Products” or “Describing Products” in the anthropomorphism and control condition, respectively. We excluded 98 participants who failed our manipulation check (asking participants to select examples of [anthropomorphism] [description]) and/or responses that appeared to be from bots (i.e., responses that included nonsense answers to open-ended questions). However, our results do not change when including all participants. This left a final sample of 303 participants (56% men, 44% women, and 0% nonbinary participants). The sample included 77% White, 8% Asian, 5% Black, 7% Hispanic, and 3% Other-Race participants (M _age = 35.78; SD = 10.55). Also, see SOM for a preregistered replication of this study, which includes several robustness checks (https://aspredicted.org/blind.php?x=ve398p).
与研究 2b 一样，这项研究只有一个参与者之间的因素（过程：拟人化或描述）。参与者 （N = 401） 是从 MTurk 招募的，分别在拟人化和控制条件下参加“拟人化产品”或“描述产品”的研究。我们排除了 98 名未通过操作检查（要求参与者选择 [拟人化] [描述] 的示例）和/或似乎来自机器人的回答（即，包括对开放式问题的无意义回答的回答）的参与者。但是，当包括所有参与者时，我们的结果不会改变。这留下了 303 名参与者的最终样本（56% 男性、44% 女性和 0% 非二元参与者）。样本包括 77% 的白人、8% 的亚裔、5% 的黑人、7% 的西班牙裔和 3% 的其他种族参与者 （M_年龄 = 35.78;SD = 10.55）。此外，请参阅 SOM 以获取本研究的预注册复制，其中包括多项稳健性检查 （https://aspredicted.org/blind.php?x=ve398p）。

After providing consent, participants were given either the anthropomorphism or the description (i.e., control) instructions. In the anthropomorphism condition, participants were given the same definition of anthropomorphism used in Study 1 and told “on the next page we will give you a product to anthropomorphize.” In the description (control) condition, participants were told:
在提供同意后，参与者被给予拟人化或描述（即控制）指令。在拟人化条件下，参与者被赋予了与研究 1 中使用的拟人化相同的定义，并被告知“在下一页，我们将为您提供一个拟人化产品”。在描述（对照）条件下，参与者被告知：

We’re interested in your experience with describing products. Description is the process of listing qualities or characteristics about a product. This can be related to many different observations about the product. On the next few pages, we will give you a product to describe.
我们很想知道您描述商品的体验。描述是列出商品质量或特征的过程。这可能与对产品的许多不同观察有关。在接下来的几页中，我们将为您提供一个产品来描述。

Participants in both conditions were then given a photo of a robotic vacuum (see Appendix). In the anthropomorphism condition, participants were instructed to “take a minute to anthropomorphize this robotic vacuum. Think about how you would use the product, what personality characteristics it might have, how it might ‘behave,’ and what emotions might you feel towards it.” In the describe (control) condition, they were told to “take a minute to describe this robotic vacuum. Think about the predominant characteristics of the product, what qualities you might notice, what this product is meant to do and how it functions, how you would you use the product, what you like and dislike about the product.” In both conditions, participants were asked to write a five-to-six sentence response.
然后，两种情况的参与者都获得了机器人真空吸尘器的照片（见附录）。在拟人化条件下，参与者被指示“花一分钟时间将这个机器人真空拟人化。想想你会如何使用这个产品，它可能有什么个性特征，它可能如何'行为'，以及你对它的感受是什么。在描述（控制）条件下，他们被告知“花一分钟时间描述这个机器人吸尘器。想想产品的主要特性，你可能会注意到什么品质，这个产品的目的是什么，它是如何工作的，你会如何使用这个产品，你喜欢和不喜欢这个产品的什么。在这两种情况下，参与者都被要求写出 5 到 6 个句子的回答。

To gather preliminary evidence that the attribution of gender to anthropomorphized targets is spontaneous, we first coded open-ended responses (asking participants to “anthropomorphize” or “describe” the robotic vacuum) for gender references. This open-ended reflection preceded our measure of (explicit) social-category ascription. To code these responses for gender references, we used Linguistic Inquiry Word Count (LIWC; Pennebaker et al., 2015), a computerized text analysis that captures the frequency of specific words in open-ended text (in this case, the categories “she/he,” “male,” and “female”). We then had a research assistant confirm that the gender referent was present and applied to the robotic vacuum (and not a third party or the participant). Responses were coded such that 1 = gendered referent and 0 = nongendered referent. Participants were then administered the two scales that measure humanization (Kteily et al., 2015) and anthropomorphism (Waytz, Morewedge, et al., 2010; α = .94) of the robotic vacuum. Next, they were asked to indicate the extent to which the vacuum they anthropomorphized [described] could be described as having membership in different social categories (gender, race, generational cohort, religion, disability, and sexual orientation), on a scale from 1 = not at all applicable to 5 = very applicable, as measured in Studies 1, 2a, and 2b.
为了收集初步证据，证明将性别归因于拟人化目标是自发的，我们首先编写了开放式反应（要求参与者“拟人化”或“描述”机器人真空）以进行性别参考。这种开放式的反思先于我们对（明确的）社会类别归类的测量。为了对这些回答进行性别参考编码，我们使用了语言探究字数 （LIWC;Pennebaker et al.， 2015），这是一种计算机化的文本分析，可捕获开放式文本中特定单词的频率（在本例中为“她/他”、“男性”和“女性”类别）。然后，我们让一名研究助理确认性别指涉存在并应用于机器人真空吸尘器（而不是第三方或参与者）。对回答进行编码，使 1 = 有性别的指涉，0 = 无性别的指涉。然后，参与者接受了两个衡量机器人真空的人源化（Kteily 等人，2015 年）和拟人化（Waytz、Morewedge 等人，2010 年;α = .94）的量表。接下来，他们被要求指出他们拟人化的真空在多大程度上可以被描述为具有不同社会类别（性别、种族、代际群体、宗教、残疾和性取向）的成员身份，范围从 1 = 完全不适用到 5 = 非常适用，如研究 1、2a 和 2b 中所衡量的那样。

First, we confirmed that our manipulation had its intended effect: Participants ascribed humanness significantly more in the anthropomorphism condition (M _anthro = 57.95, SD = 29.39; M _anthro = 3.35, SD = 1.09) as compared to the control condition (M _control = 20.54, SD = 23.90; M _control = 1.81, SD = .96); both ps < .001.
首先，我们确认我们的操作达到了预期的效果：参与者在拟人化条件下明显更多地归因于人类性（M_anthro = 57.95，SD = 29.39; M_anthro = 3.35，SD = 1.09）与对照条件 （M_对照 = 20.54，SD = 23.90; M_对照 = 1.81，SD = .96）; 两个 PS < .001。

Next, we consider our incidental measure of gender ascription—that is, whether or not participants used a gendered reference when describing the target entity in their open-ended text response. Confirming our prediction, 69% (N = 121) of participants in the anthropomorphism condition used a gendered reference to describe the robotic vacuum as compared to less than 1% (N = 1) in the control (describe) condition, χ² (1, N = 303) = 143.64, p < .001, suggesting that even unprompted, the majority of participants saw their anthropomorphized entity as having gender.
接下来，我们考虑我们对性别归属的附带测量——也就是说，参与者在他们的开放式文本响应中描述目标实体时是否使用了性别参考。证实了我们的预测，在拟人化条件下，69% （N = 121） 的参与者使用性别参考来描述机器人真空吸尘器，而在对照（描述）条件下，只有不到 1% （N = 1），χ² （1， N = 303） = 143.64，p < .001，这表明即使 没有提示，大多数参与者也认为他们的拟人化实体具有性别。

To test whether gender is more central to conceptions of humanity than other social categories, and that its ascription is specific to targets that have been humanized rather than described, we submitted the data to a mixed-model ANOVA with one between-subject (process: anthropomorphism, control) and one within-subject factor (social category: gender, race, generational cohort, religion, disability, sexual orientation). Replicating Study 2b, the results revealed the predicted interaction, Wilks λ = .41, F(5, 295) = 83.97, p < .001, η _p ²  = .59. Gender was more likely to be ascribed by participants in the anthropomorphism condition as compared to the describe (control) condition, F(1, 299) = 435.46, p < .001; and while anthropomorphizing, relative to describing, significantly increased ratings for other social categories (η _ps ²  < .082), it did not do so to the same extent as it did for gender (η _p ²  = .59; see Table 2, for all means and effect sizes). Further, participants in the anthropomorphism condition were more likely to ascribe the robotic vacuum a gender (M = 4.03, SD = 1.35) as compared to other social categories (Ms < 2.09; ps < .001 for all pairwise comparisons; see Table 2, for more detail). This pattern was not observed among participants in the describe (control) condition, who were unlikely to ascribe any of the social categories to the entity (all Ms < 1.65), including gender (M = 1.28, SE = .73). See Table 2.
为了测试性别是否比其他社会类别在人类概念中更重要，以及其归属特定于已被人性化而不是描述的目标，我们将数据提交给一个混合模型方差分析，其中包含一个主体间因素（过程：拟人化、控制）和一个主体内因素（社会类别：性别、种族、代际群体、宗教、残疾、 性取向）。复制研究 2b，结果揭示了预测的相互作用，Wilks λ = .41，F（5， 295） = 83.97，p < .001，η _p ²  = .59。 与描述（对照）条件相比，拟人化条件下的参与者更有可能归因于性别，F（1， 299） = 435.46，p < .001;虽然拟人化 相对于描述显着提高了其他社会类别的评分 （η _ps ²  < .082），但它的程度与性别的程度不同 （η _p ²  = .59;有关所有均值和效应大小，请参阅表 2）。此外，与其他社会类别相比，拟人化条件的参与者更有可能将机器人扫地机器人归因于性别（M = 4.03，SD = 1.35）（Ms < 2.09;所有成对比较的 ps < .001;更多详细信息见表 2）。在描述（对照）条件下的参与者中没有观察到这种模式，他们不太可能将任何社会类别归因于实体（所有 Ms < 1.65），包括性别（M = 1.28，SE = .73）。 请参阅 表 2。

 放大此图像。

As a secondary test of our prediction that gender is a stronger predictor of humanization than are other social categories to which people belong, we tested the extent to which participants’ scores on these scales were associated with social-category ascription using a multivariate regression with the social categories as simultaneous predictors. Consistent with Studies 2a and 2b, when we restrict our analyses to participants who were instructed to anthropomorphize (vs. describe) the target entity, gender was the only social category that uniquely predicted humanization, b = 6.55, SE = 1.65, t(166) = 3.97, p < .001, CI₉₅ = 3.29, 9.81 (for all other categories ps > .08; see SOM, for more details), and anthropomorphism, b = .16, SE = .062, t(167) = 2.51, p = .013, CI₉₅ = .03, .28. This pattern not only occurred with our explicit measure of gendering, but also with our incidental one (i.e., gender reference). Those who described their entity using at least one gender reference were more likely to ascribe the entity personhood, using both our measure of humanization, b = 13.00, SE = 4.73, t(172) = 2.75, p = .007, CI₉₅ = 3.67, 22.33, and anthropomorphism, b = .45, SE = .18, t(173) = 2.60, p = .010, CI₉₅ = .11, .80.
作为我们预测性别比人们所属的其他社会类别更强地预测人性化的二次测试，我们使用多元回归测试了参与者在这些量表上的分数与社会类别归属的关联程度，社会类别作为同时预测因子。与研究 2a 和 2b 一致，当我们将分析限制在被指示拟人化（而不是描述）目标实体的参与者时，性别是唯一唯一预测人性化的社会类别，b = 6.55，SE = 1.65，t（166） = 3.97，p < .001，CI₉₅ = 3.29,9.81（对于所有其他类别 ps > .08;有关详细信息，请参阅 SOM）和拟人化， b = .16，SE = .062，t（167） = 2.51，p = .013，CI₉₅ = .03，.28。 这种模式不仅发生在我们对性别的明确测量中，也发生在我们的偶然测量（即性别参考）中。那些使用至少一个性别参考来描述其实体的人更有可能归因于实体人格，同时使用我们的人性化测量 b = 13.00，SE = 4.73，t（172） = 2.75，p = .007，CI ₉₅ = 3.67,22.33，和拟人化，b = . 45，SE = .18，t（173 ） = 2.60，p = .010，CI₉₅ = . 11，.80。

Study 3b sought to replicate Study 3a and show that the effects generalize beyond a robotic vacuum to other stimuli used in past work on anthropomorphism (Aggarwal & McGill, 2007; De Visser et al., 2016; Im Shin & Kim, 2020; Mourey et al., 2017; Tam, 2014; Waytz et al., 2014).
研究3b试图复制研究3a，并表明其效果超越了机器人真空，推广到过去拟人化工作中使用的其他刺激（Aggarwal & McGill，2007;De Visser et al.， 2016;Im Shin & Kim， 2020;Mourey et al.， 2017;Tam， 2014;Waytz et al.， 2014）。

Participants (N = 1,207) were recruited from Amazon’s MTurk. We excluded 286 participants using the criteria from Study 3a. However, the results do not appreciably change when including all participants. Excluding these participants left a final sample of 921 participants (45% men, 55% women, and 0% nonbinary participants; 74% White, 8% Asian, 9% Black, 6% Hispanic, 3% Other-Race participants) with a mean age of 39.06 (SD = 12.69). Study 3b procedures were identical to Study 3a with a single exception: Instead of having participants anthropomorphize or describe a robotic vacuum, they were randomly assigned to anthropomorphize or describe one of four different stimuli (a car, a plant, a bicycle, or a computer; see Appendix). After providing informed consent, participants were administered either the anthropomorphism or the control instructions (see Study 3a).
参与者 （N = 1,207） 是从 Amazon 的 MTurk 招募的。我们使用研究 3a 的标准排除了 286 名受试者。但是，当包括所有参与者时，结果不会明显改变。排除这些参与者后，留下了 921 名参与者的最终样本（45% 男性、55% 女性和 0% 非二元参与者;74% 白人、8% 亚裔、9% 黑人、6% 西班牙裔、3% 其他种族参与者），平均年龄为 39.06 （SD = 12.69）。研究 3b 程序与研究 3a 相同，但有一个例外：他们不是让参与者拟人化或描述机器人吸尘器，而是被随机分配拟人化或描述四种不同刺激（汽车、植物、自行车或计算机;见附录）中的一种。在提供知情同意后，参与者接受了拟人化或对照指令（参见研究 3a）。

As with Study 3a, we first captured an incidental measure of gender ascription by coding participants’ open-ended reflections. Again, LIWC analysis was conducted, using the “he/she,” “male,” and “female” categories, and gender reference was confirmed by an independent coder. Participants were asked to indicate the extent to which the entities they [anthropomorphized] [described] could be seen as having membership in different social categories (gender, race, generational cohort, religion, disability, and sexual orientation) on a scale from 1 = not at all applicable to 5 = very applicable, as in Studies 1–3a. Participants were also administered the two scales that measure the extent to which they humanized (Kteily et al., 2015) and anthropomorphized (Waytz, Morewedge, et al., 2010; α = .95) the different products.
与研究 3a 一样，我们首先通过对参与者的开放式反思进行编码来捕捉性别归属的偶然衡量标准。再次，使用 “他/她”、“男性” 和 “女性” 类别进行了 LIWC 分析，并由独立编码员确认了性别参考。参与者被要求指出他们 [拟人化] [描述] 的实体在多大程度上可以被视为具有不同社会类别（性别、种族、代际群体、宗教、残疾和性取向）的成员身份，范围从 1 = 完全适用到 5 = 非常适用，如研究 1-3a。参与者还接受了两个量表，用于衡量他们人性化（Kteily 等人，2015 年）和拟人化（Waytz、Morewedge 等人，2010 年;α = .95）不同产品的程度。

The manipulation had its intended effect as evidenced by the fact that ratings of the target using both the humanization (M = 62.93, SD = 26.41) and anthropomorphism (M = 3.41, SD = 1.07) scales were significantly higher among participants in the anthropomorphize versus the describe (control) condition (M _human = 19.59, SD = 26.41; M _anthro = 1.81, SD = 1.07), both ps < .001. This pattern was observed for each of the four entities (all ps < .001; see Table 2).
该操作达到了预期的效果，事实证明，使用人化 （M = 62.93， SD = 26.41） 和拟人化 （M = 3.41， SD = 1.07） 量表的目标评分在拟人化参与者中明显高于描述（对照）条件 （M_human = 19.59， SD = 26.41;M_anthro = 1.81，SD = 1.07），两者均 ps < .001。对四个实体中的每一个都观察到了这种模式（所有 ps < .001;见表 2）。

To support our hypothesis that gender would be implicated in humanization even when participants are not asked about it, we first examined our incidental measure of gender and found that the majority of participants in the anthropomorphism condition (70%; N = 351) used a gendered reference to describe their entity; however, no participants (N = 0) used a gender reference when describing their entity in the describe (control) condition, χ² (1, N = 417) = 469.24, p < .001.
为了支持我们的假设，即使参与者没有被问及性别，性别也会与人性化有关，我们首先检查了我们对性别的偶然测量，发现拟人化条件下的大多数参与者 （70%;N = 351）使用性别参考来描述他们的实体;然而，没有参与者 （N = 0） 在描述（控制）条件下描述他们的实体时使用性别参考，χ² （1， N = 417） = 469.24，p < .001。

To test whether gender is more central to conceptions of humanity than other social categories, and whether its ascription is specific to targets that have been humanized (versus described), we submitted the data to a mixed-model ANOVA with one between-subject factor (process: anthropomorphism, control) and one within-subject factor (social-category ascription: gender, race, generational cohort, religion, disability, sexual orientation). The results revealed the predicted interaction effect, Wilks λ = .63, F(5, 912) = 105.05, p < .001, η _p ²  = .36. Gender was more likely to be ascribed by participants in the anthropomorphism condition as compared to the describe condition, F(1, 916) = 688.20, p < .001. Though anthropomorphizing significantly increased all social-category ascriptions as compared to describing (all η _ps ²  < .07), gender was the most affected by the anthropomorphizing process (η _p ²  = .43)—six times greater than the largest comparable social category (see Table 2, for all means and effect sizes). Among participants in the anthropomorphism condition, gender was more strongly ascribed (M = 4.06, SD = 1.34) than the other social categories (Ms < 2.39; ps < .001 for all pairwise comparisons). In the describe (control) condition, gender (M = 1.85, SD = 1.23) was both less (as compared to age) and more (as compared to the remaining social categories) likely to be ascribed (Ms range from 1.23 to 1.91). See Figure 1 and Table 2.
为了测试性别是否比其他社会类别在人类概念中更重要，以及其归属是否特定于已人性化的目标（与描述），我们将数据提交给一个混合模型方差分析，其中包含一个主体间因素（过程：拟人化、控制）和一个主体内因素（社会类别归属：性别、种族、代际群体、宗教、残疾、 性取向）。结果揭示了预测的交互效应，Wilks λ = .63，F（5， 912） = 105.05，p < .001，η _p ²  = .36。与描述条件相比，拟人化条件下的参与者更有可能归因于性别，F（1， 916） = 688.20，p < .001。尽管与描述相比 ，拟人化显着增加了所有社会类别的归属（所有 _{η ps} ²  < .07），但性别受拟人化过程的影响最大 （η _p ² = .43）—比最大的可比社交类别大六倍（参见表 2，了解所有均值和效应大小）。在拟人化条件的参与者中，性别归因（M = 4.06，SD = 1.34） 比其他社会类别（Ms < 2.39;ps < .001）更强烈。在描述（对照）条件下，性别（M = 1.85，SD = 1.23）被归因的可能性较小（与年龄相比） 和较高（与其余社会类别相比）（Ms 范围从 1.23 到 1.91）。请参阅 图 1 和 表 2。

 放大此图像。

As a secondary test of our prediction that humanizing leads people to ascribe gender, we tested the extent to which participants’ scores on the humanization and anthropomorphism scales were associated with social-category ascription using a multivariate regression with the social categories as simultaneous predictors. Among participants in the anthropomorphism condition, gender ascription was most strongly associated with humanization, b = 4.98, SE = 0.92, t(494) = 5.40, p < .001, CI₉₅ = 3.17, 6.79 (all other categories ps > .02), and anthropomorphism, b = .17, SE = 0.04, t(494) = 4.56, p < .001, CI₉₅ = .09, .24 (all other categories ps > .012; see SOM, for more details). Again, this effect was observed not only with our explicit measure of gendering but also with our incidental one (i.e., gender references): those who used at least one gender reference were more likely to humanize their products, b = 10.36, SE = 2.64, t(502) = 3.92, p < .001, CI₉₅ = 5.17, 15.55, though the relationship between using a gender reference and the anthropomorphism measure was marginal, b = .17, SE = .10, t(502) = 1.67, p = .097, CI₉₅ = .03, .38.
作为我们预测人性化导致人们归因性别的二次测试，我们使用多变量回归测试了参与者在人化和拟人化量表上的分数与社会类别归属的关联程度，并将社会类别作为同时预测因子。在拟人化条件的参与者中，性别归类与人化最密切相关，b = 4.98，SE = 0.92，t（494） = 5.40，p < .001，CI₉₅ = 3.17,6.79（所有其他类别 ps > .02），拟人化，b = .17，SE = 0.04，t（494） = 4.56，p < .001，CI₉₅ = .09， .24（所有其他类别 ps > .012;有关详细信息，请参阅 SOM）。同样，这种影响不仅在我们的明确性别测量中观察到，而且在我们的偶然测量（即性别参考）中也观察到：那些使用至少一个性别参考的人更有可能将他们的产品人性化，b = 10.36，SE = 2.64，t（502） = 3.92，p < .001，CI₉₅ = 5.17,15.55，尽管使用性别参考和拟人化测量之间的关系很小， b = .17，SE = .10，t（502） = 1.67，p = .097，CI₉₅ = .03，.38。

Study 4 sought to assess which social categories are deemed essential enough to consistently include when conceiving of a human. Inspired by the pet-rock phenomenon popularized in the 1970s (Weston, 2000), Study 4 adopted a “blank slate” approach by having participants create a human from scratch (i.e., from memory or imagination) rather than anthropomorphizing an entity that already exists.
研究 4 试图评估哪些社会类别被认为足够重要，在构想人类时需要始终如一地包括。受 1970 年代流行的宠物摇滚现象的启发（Weston，2000 年），研究 4 采用了“白纸”方法，让参与者从头开始（即，从记忆或想象中）创造一个人类，而不是将已经存在的实体拟人化。

Participants (N = 202; 32% men, 68% women; 40% White, 34% Asian, 6% Black, 13% Hispanic, and 7% Other-Race participants; M _age = 24.75; SD = 8.63) were recruited at an American university on the West Coast. The study had a single between-participant factor (task: anthropomorphize or decorate). In the anthropomorphism condition, participants were told:
参与者（N = 202;32% 男性，68% 女性;40% 白人，34% 亚裔，6% 黑人，13% 西班牙裔和 7% 其他种族参与者;M_年龄 = 24.75;SD = 8.63）是在西海岸的一所美国大学招募的。该研究只有一个参与者之间的因素（任务：拟人化或装饰）。在拟人化条件下，参与者被告知：

We are interested in how people make objects ‘come alive.’ We would like you to anthropomorphize the rock in front of you. Anthropomorphism is the process of ascribing uniquely human qualities to this rock. Anthropomorphism entails attributing humanness—such as personality, emotions, and intentions—to your rock.
我们对人们如何使物体“栩栩如生”感兴趣。我们希望您将您面前的岩石拟人化。拟人化是赋予这块岩石独特的人类品质的过程。拟人化需要将人性（例如个性、情感和意图）归因于您的岩石。

In the describe (control) condition, participants were told:
在描述（控制）条件下，参与者被告知：

We are interested in how people decorate objects. We would like you to decorate the object in front of you. Decoration is the process of ascribing a unique and creative design to this rock. It involves attributing beauty, such as color, texture, and shapes to your rock.
我们对人们如何装饰物品感兴趣。我们希望您装饰您面前的物品。装饰是为这块岩石赋予独特而富有创意的设计的过程。它涉及将颜色、质地和形状等美感归因于您的岩石。

Participants in both conditions received a basket with materials and a rock and were given 15 min to do the task. Participants were further told that their rock would be reviewed by two independent judges and that the one that received the highest ratings would receive $100. See Figure 2 for examples of rocks.
两种情况下的参与者都收到了一个装有材料和石头的篮子，并有 15 分钟的时间来完成任务。参与者进一步被告知，他们的岩石将由两名独立评委进行审查，获得最高评分的那块将获得 100 美元。有关岩石的示例，请参见图 2。

 放大此图像。

Participants were asked to indicate the extent to which the rock could be described as having membership in different social categories (gender, race, generational cohort, religion, disability, and sexual orientation) on a scale from 1 = not at all applicable to 5 = very applicable, and we administered the humanization measure used in Studies 1–3 (0 = not at all human to 100 = very human).
参与者被要求指出岩石在多大程度上可以被描述为具有不同社会类别（性别、种族、代际群体、宗教、残疾和性取向）的成员身份，范围从 1 = 完全不适用到 5 = 非常适用，我们进行了研究 1-3 中使用的人性化措施（0 = 完全不是人类到 100 = 非常人类）。

We also had independent coders rate the rock on several dimensions to ensure several confounds were not at play. To rule out the possibility that the study suffers from a common-method bias, we had two independent coders, blind to the hypotheses, rate the “humanness” of the rock (1 = not at all humanlike to 7 = very humanlike; r = .87) as well as the extent to which they could accurately tell the social-category memberships of the rock (1 = very difficult to 7 = very easy).  ^{[  4  ]}  These data allowed us to examine whether an independent observer more readily perceived gender as compared to other social categories in a humanized rock. They also allowed us to examine whether coders’ social-category ascription ratings predicted participants’ humanness ratings, thereby allowing us to mitigate any common-method bias. Finally, independent coders rated the quality of the pet rocks to rule out the possibility that the effect of humanization on gender ascription is confounded by artistic quality or effort.
我们还让独立的编码人员对岩石的多个维度进行评级，以确保没有几个混杂因素。为了排除该研究存在通用方法偏差的可能性，我们让两名独立的编码员对岩石的“人类性”进行评分，他们对这些假设视而不见（ 7 = 非常像人类; r = .87），以及他们能够准确判断岩石的社会类别成员身份的程度（1 = 非常困难到 7 = 非常容易）。    ^{[ 4  ]}  这些数据使我们能够检查与人性化岩石中的其他社会类别相比，独立观察者是否更容易感知性别。他们还允许我们检查编码员的社会类别归类评级是否能预测参与者的人性评级，从而使我们能够减轻任何常用方法偏差。最后，独立编码员对宠物石头的质量进行了评分，以排除人性化对性别归属的影响被艺术质量或努力混淆的可能性。

Our manipulation was successful: Participants in the anthropomorphism condition rated their rock as more human (M = 64.37, SD = 25.72) than did those in the decorate (control) condition (M = 47.42, SD = 34.18), F(1, 200) = 15.81, p < .001, η _p ²  = .073. To test the prediction that gender is more central to conceptions of humanity than other social categories, and that its ascription is specific to targets that have been humanized rather than decorated, we submitted the data to a mixed-model ANOVA with one between-subject factor (condition: anthropomorphism, control) and one within-subject factor (social-category ascription: gender, race, generational cohort, religion, disability, sexual orientation) and found a significant interaction, Wilks λ = .79, F(5, 196) = 10.27, p < .001, η _p ²  = .21. Participants in the anthropomorphism condition were significantly more likely to ascribe gender to the rock than participants in the decorate (control) condition, F(1, 200) = 50.71, p < .001. Anthropomorphizing significantly increased all social-category ascriptions (all η _ps ²  < .139); however, as predicted, gender was the most strongly affected by the anthropomorphizing process (η _p ²  = .20). Moreover, participants in the anthropomorphism condition were significantly more likely to ascribe gender (M = 3.80, SD = 1.45) than any other social category (all Ms < 2.99; ps < .001 for pairwise comparisons; see Table 1 and SOM). Participants in the decorate (control) condition were also more likely to ascribe the rock a gender (M = 2.31, SD = 1.49) than any other social category (p < .001), but this difference was less pronounced (all Ms < 1.91) and was largely driven by a handful of participants who spontaneously decorated their rock to resemble a person. See Table 1 for means and descriptive statistics.
我们的操作是成功的：拟人化条件下的参与者认为他们的岩石比装饰（对照）条件下的参与者（M = 47.42，SD = 34.18），F（1,200） = 15.81，p < .001，η _p ²  = .073）更像人。 为了检验性别比其他社会类别在人类概念中更核心的预测，并且其归属特定于被人性化而不是装饰的目标，我们将数据提交给一个混合模型方差分析，其中包含一个主体间因素（条件：拟人化、对照）和一个主体内因素（社会类别归属：性别、种族、代际队列、 宗教、残疾、性取向）并发现显着交互作用，Wilks λ = .79，F（5， 196） = 10.27，p < .001，η _p ²  = .21。 拟人化条件下的参与者明显比装饰（对照）条件下的参与者更有可能将性别归因于岩石，F（1， 200） = 50.71，p < .001。拟人化 显着增加了所有社会类别归类（全部η _ps ²  < .139）;然而，正如预测的那样，性别受拟人化过程的影响最大 （η _p ²  = .20）。此外，与任何其他社会类别相比，拟人化条件下的参与者明显更有可能归因于性别（M = 3.80，SD = 1.45）（所有 Ms < 2.99;成对比较的 ps < .001;见表 1 和 SOM）。 与任何其他社会类别 （p < .001） 相比，装饰（对照）条件下的参与者也更有可能将岩石归为性别 （M = 2.31， SD = 1.49），但这种差异不太明显 （所有 Ms < 1.91），并且主要是由少数参与者驱动的，他们自发地装饰他们的岩石以使其看起来像一个人。有关方法和描述性统计，请参见表 1。

When considering all potential social categories as simultaneous predictors of humanity ascription among participants in the anthropomorphism condition, gender was the only one that mattered for seeing human, b = 5.08, SE = 1.85, t(93) = 2.74, p = .007, CI₉₅ = 1.40, 8.75 (for all other categories ps > .12). We observed the same pattern among the coders, such that gender ascription most strongly predicted humanness ratings, b = .37, SE = .09, t(93) = 4.02, p < .001, CI₉₅ = .19, .55, as compared to other social-category ascriptions (ps > .016). Further, to circumvent common-method bias, we also looked at the extent to which participant ratings of social-category ascription predicted coder ratings of humanness. Using a multivariate regression, we find that ascription of gender was the strongest predictor of humanness, b = .31, SE = .13, t(93) = 3.67, p < .001, CI₉₅ = .20, .68 (all ps > .018). These findings suggest that gender is more tightly linked to humanization than are the other social categories to which people belong.
当将所有潜在的社会类别视为拟人化条件下参与者人类归属的同时预测因子时，性别是看待人类的唯一重要因素，b = 5.08，SE = 1.85，t（93） = 2.74，p = .007，CI₉₅ = 1.40,8.75（对于所有其他类别 ps > .12）。 我们在编码员中观察到相同的模式，因此性别归类最强烈地预测了人类评级，b = .37，SE = .09，t（93） = 4.02，p < .001，CI₉₅ = .19，.55，与其他社会类别归类 （ps > .016）。 此外，为了规避通用方法偏倚，我们还研究了参与者对社会类别归类的评分在多大程度上预测了编码员对人类的评分。使用多元回归，我们发现性别归因是人类的最强预测因子，b = .31，SE = .13，t（93 ） = 3.67，p < .001，CI₉₅ = .20，.68（所有 ps > .018）。 这些发现表明，与人们所属的其他社会类别相比，性别与人性化的联系更紧密。

To assuage concerns that ratings of humanness are conflated with art quality and that the effect is spurious, we had the same two independent coders rate the quality of each rock in the study (1 = low quality to 7 = high quality) and perceived participant effort (1 = none at all to 7 = very much). These items were combined, as they were consistent within (rs > .78, p < .001) and across (r = .49, p < .001) coder ratings.  ^{[  5  ]}  When examining these results controlling for rock quality, we found the relationship between humanness and gender ascription remained significant, b = 5.04, SE = 1.87, t(93) = 2.69, p = .008, CI₉₅ = 1.33, 8.76.
为了缓解人们对人类评级与艺术质量混为一谈以及效果是虚假的担忧，我们让相同的两名独立编码员对研究中每块石头的质量进行评分（1 = 低质量到 7 = 高质量）和感知参与者的努力（1 = 完全没有到 7 = 非常多）。这些项目被组合在一起，因为它们在 （rs > .78， p < .001） 和跨 （r = .49， p < .001） 编码器评级中是一致的。 ^{[  5  ]}  在检查这些控制岩石质量的结果时，我们发现人类与性别归属之间的关系仍然显着，b = 5.04，SE = 1.87，t（93） = 2.69，p = .008，CI₉₅ = 1.33,8.76。

Together, these results demonstrated the link between gender and humanness by using a novel paradigm in which participants constructed a human from “scratch.” Bringing participants’ imagination to “life” allowed us to test our central question: What social category characteristics do people generate when conceiving of a human? The approach provided a window into which characteristics are deemed essential to being human. As predicted, and in convergence with the results of the other studies, gender was more strongly associated with personhood than other social categories were.
总之，这些结果通过使用一种新的范式证明了性别与人类之间的联系，在该范式中，参与者从“零开始”构建了一个人。将参与者的想象力带入“生活”使我们能够测试我们的中心问题：人们在构想一个人时会产生哪些社会类别特征？这种方法提供了一个窗口，让人们了解哪些特征被认为是人类必不可少的。正如预测的那样，与其他研究的结果一致，性别与其他社会类别相比，与人格的相关性更强。

In Studies 1–4, we demonstrated that when attributing nonhuman entities humanness, the majority of participants also ascribed them gender, which was not true for the other social categories. In our next four studies, we tested the reciprocal hypothesis, examining whether removing gender leads to dehumanization. We also measured the downstream consequences of the gender-humanization link. As we have suggested, without a gender schema (e.g., stereotypes) to streamline processing, perceivers may struggle to efficiently understand the people who populate their social world. We predict that difficulty with the sensemaking process ultimately leads perceivers to see targets as less human.
在研究 1-4 中，我们证明，当将非人类实体归类为人类时，大多数参与者也将其归为性别，而其他社会类别则并非如此。在接下来的四项研究中，我们检验了互惠假设，研究了消除性别是否会导致非人化。我们还测量了性别-人性化联系的下游后果。正如我们所建议的，如果没有性别图式（例如刻板印象）来简化处理，感知者可能难以有效地理解居住在他们社交世界中的人。我们预测，意义构建过程的困难最终会导致感知者将目标视为不那么人性化的目标。

Gender scholars have long argued that people perceive gender to be a central feature of personhood, but evidence in support of this assertion has been severely lacking (see Morgenroth & Ryan, 2018). Studies 1–4 address this gap by showing that conceiving of something as human leads to ascribing it gender, more so than the other social categories. If gender is core to perceiving humanness, as this finding implies, then stripping targets of gender should lead them to be seen as less human. Study 5 seeks preliminary evidence that targets lacking gender are dehumanized by comparing perceptions of technology—virtual-human assistants—that do and do not have a gendered voice.  ^{[  6  ]} 
性别学者长期以来一直认为，人们认为性别是人格的核心特征，但支持这一说法的证据严重缺乏（见Morgenroth & Ryan，2018）。研究 1-4 通过表明将某物视为人类会导致将其归为性别来解决这一差距，这比其他社会类别更重要。如果正如这一发现所暗示的那样，性别是感知人类的核心，那么剥离性别目标应该会导致他们被视为不那么人性化。研究 5 通过比较对技术（虚拟人类助手）的看法，寻求初步证据，证明缺乏性别的目标被非人化，这些技术有和没有性别的声音。    ^{[ 6  ]} 

In this preregistered study (https://aspredicted.org/blind.php?x=gw8vp2), participants (N = 300) were recruited via MTurk for a study on evaluating products. After removing participants who failed the manipulation check (recalling the gender of the voice assistant) and/or nonsensical responses that appeared to be from bots (N = 66), we were left with a final sample of 244 individuals (62% men and 38% women; 75% White, 6% Asian, 7% Black, 11% Hispanic, 2% Other-Race participants; M _age = 34.27, SD = 10.45). Participants were randomly assigned to one of three conditions: describing a (a) male, (b) female, or (c) genderless  ^{[  7  ]}  voice assistant. They were told, “We are interested in your thoughts about digital-voice assistants” and to “Please take a minute to write about a [male] [female] or [genderless] voice assistant.” It should be noted that there were no actual voices used in the study; rather, participants were prompted to write about the following: “What are features of this product? What would it do? How would it interact and behave? What would it sound like? What would it look like? What would its distinctive features be?” Participants wrote five-to-six sentences and responded to the dependent measures.
在这项预先注册的研究 （https://aspredicted.org/blind.php?x=gw8vp2） 中，参与者 （N = 300） 是通过 MTurk 招募的，用于评估产品的研究。在删除未通过操作检查（回忆语音助手的性别）和/或似乎来自机器人的荒谬回答 （N = 66） 的参与者后，我们留下了 244 人的最终样本（62% 男性和 38% 女性;75% 白人，6% 亚洲人，7% 黑人，11% 西班牙裔，2% 其他种族参与者;M_年龄 = 34.27，SD = 10.45）。 参与者被随机分配到以下三种情况之一：描述 （a） 男性、（b） 女性或 （c） 无 ^{性别 [  7  ]}  语音助手。他们被告知，“我们对您对数字语音助手的想法很感兴趣”，并且“请花一点时间写一篇关于 [男性] [女性] 或 [无性别] 语音助手的文章。应该注意的是，研究中没有使用实际的声音;相反，参与者被提示写下以下内容：“这个产品有什么特点？它会做什么？它将如何交互和行为？听起来会是什么样子？它会是什么样子？它的独特之处会是什么？参与者写了 5 到 6 个句子并回答相关的措施。

Participants were given scales that measure humanization and anthropomorphism (α = .94). We also collected measures of human uniqueness and human nature (reported in SOM).
参与者获得了衡量人性化和拟人化的量表 （α = .94）。我们还收集了人类独特性和人性的测量（在 SOM 中报告）。

As expected, ascriptions of humanness varied across the three conditions, F(2, 241) = 11.33, p < .001, η _p ²  = .086. Participants were less likely to ascribe humanness to those in the genderless condition (M = 39.88, SD = 29.94) as compared to those in the male, (M = 53.94, SD = 28.97), F(1, 244) = 9.27, p = .003, η _p ²  = .037, CI₉₅ = −23.16, −4.96, and female (M = 61.62, SD = 29.54), F(1, 244) = 22.02, p < .001, η _p ²  = .084, CI₉₅ = −30.07, −12,62, conditions. There was no significant difference in the ascription of humanness between the male and female conditions (p = .098). The same pattern was found for anthropomorphism, F(2, 241) = 5.29, p = .006, η _p ²  = .042. Participants in the genderless condition (M = 2.21, SD = 1.14) were significantly less likely to anthropomorphize as compared to those in the female condition (M = 2.84, SD = 1.32), F(1, 244) = 10.53, p = .001, η _p ²  = .042, CI₉₅ = −10.53, −.25, and directionally less likely to do so in the male condition (M = 2.52, SD = 1.26), F(1, 241) = 2.44, p = .11, η _p ²  = .01, CI₉₅ = −.69, .08. The female voice assistant was marginally more anthropomorphized than the male voice assistant, F(1, 244) = 2.75, p = .099, η _p ²  = .011, CI₉₅ = −.06, .71.
正如预期的那样，人类的归类在三个条件下有所不同，F（2， 241） = 11.33，p < .001，η _p ²  = .086。与男性相比，参与者不太可能将无性别状态的人（M = 39.88，SD = 29.94）归因于无性别状态的人（M = 53.94，SD = 28.97）， F（1,244）= 9.27，p = .003，η _p ²  = .037，CI₉₅ = -23.16，-4.96，和女性（M = 61.62，SD = 29.54），F（1,244）= 22.02，p <.001，η _p ²  = .084，CI₉₅ = −30.07，−12,62，条件。男性和女性条件之间的人性归属没有显着差异 （p = .098）。拟人化也发现了相同的模式，F（2， 241） = 5.29， p = .006， η _p ²  = .042。与女性 （M = 2.84， SD = 1.32）， F（1， 244） = 10.53， p = .001， η _p ²  = .042， CI₉₅ = -10.53， -.25 相比，无性别状态 （M = 2.21， SD = 1.14） 的参与者明显不太可能拟人化 （M = 2.52， SD = 1.26）， F（1， 241） = 2.44， p = .11， η _p ²  = .01， CI₉₅ = −.69， .08。女性语音助手的拟人化程度略高于男性语音助手，F（1， 244） = 2.75，p = .099，η _p ²  = 。 011，CI₉₅ = −.06，.71。

Study 5 demonstrated that withholding gender from a target (virtual assistant) led to it being seen as less human. This finding is consistent with evidence that adding gendered features to robots increases the application of human traits and humanization (Nass & Moon, 2000; Stroessner & Benitez, 2019) but also implies that the current move toward genderlessness may have unintended consequences for the very targets these policies are designed to protect. One weakness of Study 5 it that it used (nonhuman) voice assistants as targets; thus, caution should be used in interpreting implications for human beings. To address this weakness, Study 6 uses “human” targets: men, women, and people.
研究 5 表明，对目标（虚拟助手）隐瞒性别会导致其被视为不那么人性化。这一发现与证据一致，即为机器人添加性别特征会增加人类特征的应用和人性化（Nass & Moon，2000;Stroessner & Benitez，2019 年）但也意味着当前走向无性别的趋势可能会对这些政策旨在保护的目标产生意想不到的后果。研究 5 的一个弱点是它使用（非人类）语音助手作为目标;因此，在解释对人类的影响时应谨慎。为了解决这一弱点，研究 6 使用了 “人类” 目标：男性、女性和人。

Study 6 had a twofold aim: First, it sought to replicate Study 5’s finding that targets stripped of gender are seen as less human using actual-human rather than virtual-human targets. Second, it sought evidence that this devaluation emerges because genderless targets lack a schema that aids in efficient sensemaking (e.g., Stern & Rule, 2018). That is, we expected that perceivers would attribute less humanness to a genderless target because they lack a readily accessible framework for organizing their thinking about the target (i.e., gender stereotypes, gendered traits; Haines et al., 2016) and thus struggle to make sense of them. We accomplished these aims by comparing perceptions of men, women, and people. We show that people (a nongendered group) are seen as significantly less human than men and women, and that the effect operates via attributions of masculinity and femininity, respectively (i.e., gender schemas).
研究 6 有两个目标：首先，它试图复制研究 5 的发现，即使用真实人类而不是虚拟人类目标，被剥夺性别的目标被视为不那么人性化。其次，它寻求证据证明这种贬值的出现是因为无性别目标缺乏有助于有效意义构建的模式（例如，Stern & Rule，2018）。也就是说，我们预计感知者会将较少的人性归因于无性别目标，因为他们缺乏一个易于访问的框架来组织他们对目标的思考（即，性别刻板印象、性别特征;Haines等人，2016 年）因此很难理解它们。我们通过比较对男性、女性和人的看法来实现这些目标。我们表明，人们（一个无性别群体）被认为比男性和女性明显更不人性化，并且这种效应分别通过男性气质和女性气质的归因（即性别图式）来运作。

Participants (N = 475) were recruited from Amazon’s MTurk to take part in a study on “evaluation of groups in society.” Participants (N = 44) who failed a manipulation check (which group they were asked about) were removed, leaving a final sample of 431 participants (54% men, 45% women, and 1% gender nonbinary participants, including 75% White, 9% Asian, 7% Black, 6% Hispanic, 3% Other-Race participants, M _age = 39.41, SD = 12.61). Participants were told that we were interested in their perception of groups and that they would receive a “random” group to evaluate. Specifically, they were told, “We are interested in your perceptions of [men] [women] [people] in the year 2020. Please take a second to think about what [men] [women] [people] in the year 2020 are like.” They were then asked a number of questions about the group they received and were thanked and paid.
参与者 （N = 475） 是从亚马逊的 MTurk 招募的，参加一项关于“社会群体评估”的研究。未通过操作检查的参与者 （N = 44） 被删除（他们被询问的是哪组）被删除，留下 431 名参与者的最终样本（54% 男性、45% 女性和 1% 性别非二元参与者，包括 75% 白人、9% 亚洲人、7% 黑人、6% 西班牙裔、3% 其他种族参与者，M_年龄 = 39.41，SD = 12.61）。 参与者被告知，我们对他们对群体的看法感兴趣，并且他们将接受一个“随机”组进行评估。具体来说，他们被告知，“我们对您在 2020 年对 [男性] [女性] [人] 的看法感兴趣。请花点时间想一想 2020 年的 [男人] [女人] [人] 是什么样的。然后，他们被问到一些关于他们收到的小组的问题，并得到了感谢和报酬。

As with other studies, participants were administered our “humanness” scale, being told “the following scale represents humanness levels. 0 represents very low degree of humanness and 100 represent very high degree of humanness. Choose a number that represents how easy it is to ascribe the group you were asked about with humanness” (0–100).
与其他研究一样，参与者接受了我们的“人类性”量表，并被告知“以下量表代表人类水平。0 表示人类的程度非常低，100 代表人类的程度非常高。选择一个数字，代表将你被问到的群体归类为人性是多么容易“（0-100）。

If removing gender from a target leads to it being dehumanized because perceivers lack stereotypic beliefs (i.e., a gender schema) to structure their thinking about the target, we should find the effect of genderlessness on humanness perceptions is mediated by reduced ascriptions of gender stereotypes (i.e., gendered traits). Participants were, therefore, asked the extent to which the target groups possess six masculine (dominant, daring, logical, assertive, risk-taking, analytical; a = .82) and six feminine (affectionate, sympathetic, supportive, warm, gentle, nurturing; a = .94) traits on a scale from 1 = not at all characteristic to 7 = very characteristic (Diekman & Eagly, 2000).
如果从目标中删除性别导致它被非人化，因为感知者缺乏刻板印象的信念（即性别图式）来构建他们对目标的思考，那么我们应该发现无性别对人类感知的影响是通过减少对性别刻板印象的归属（即性别特征）来介导的。因此，参与者被问及目标群体在多大程度上拥有六种男性特征（主导性、大胆、逻辑性、自信、冒险、分析性;a = .82）和六个女性（深情、同情、支持、温暖、温柔、养育;a = .94） 特征，范围从 1 = 完全没有特征到 7 = 非常有特征（Diekman & Eagly，2000 年）。

As expected, there was a significant effect of condition on humanness ascription, F(2, 428) = 32.78, p < .001, η _p ²  = .13. People were seen as significantly less human than both women, F(1, 428) = 63.70, p < .001, η _p ²  = .13, CI₉₅ = 14.81, 24.48, and men, F(1, 428) = 24.95, p < .001, η _p ²  = .055, CI₉₅ = 7.26, 16.68. Though not predicted, humanness ratings of women were significantly higher than men, F(1, 428) = 9.50, p = .002, η _p ²  = .022, CI₉₅ = 2.78, 12.56; See Table 3 for means.
正如预期的那样，条件对人类归属有显着影响，F（2， 428） = 32.78，p < .001，η _p ²  = .13。 人们被认为明显少于女性，F（1， 428） = 63.70，p < .001，η _p ²  = .13，CI₉₅ = 14.81,24.48，男性，F（1， 428） = 24.95，p < .001，η _p ²  = .055，CI₉₅ = 7.26,16.68。 虽然没有预测，但女性的人类评分明显高于男性，F（1， 428） = 9.50，p = . 002，η _p ²  = .022，CI₉₅ = 2.78,12.56;有关平均值，请参见表 3。

 放大此图像。

For masculine traits, there was a significant effect of condition, F(2, 428) = 16.32, p < .001, η _p ²  = .071. Men were ascribed more masculine traits than were women, F(1, 428) = 22.23, p < .001, η _p ²  = .05, CI₉₅ = .33, .81, and people, F(1, 428) = 26.34, p < .001, η _p ²  = .058, CI₉₅ = .37, .83. There was no difference between the women and people condition (p = .82). For feminine traits, there was a significant effect of condition, F(2, 428) = 38.79, p < .001, η _p ²  = .15. Women were ascribed more feminine traits than were men, F(1, 428) = 49.65, p < .001, η _p ²  = .10, CI₉₅ = .71, 1.26, and people, F(1, 428) = 67.57 p < .001, η _p ²  = .14, CI₉₅ = .87, 1.41. There was no difference in masculinity ascription between women and people (p = .26).
对于男性性状，条件有显着影响，F（2， 428） = 16.32，p < .001，η _p ²  = .071。 男性比女性被赋予更多的男性特征，F（1， 428） = 22.23，p < .001，η _p ²  = .05，CI₉₅ = .33，.81，以及人，F（1， 428） = 26.34，p < .001，η _p ²  = .058，CI₉₅ = .37，.83。 女性和人的情况之间没有差异 （p = .82）。对于女性特征，条件有显着影响，F（2， 428） = 38.79，p < .001，η _p ²  = .15。 女性比男性被赋予更多的女性特征，F（1， 428） = 49.65，p < .001，η _p ²  = .10，CI₉₅ = .71,1.26，以及人，F（1， 428） = 67.57 p < .001，η _p ²  = .14，CI₉₅ = .87,1.41。 女性和人之间的男子气概归属没有差异 （p = .26）。

To test if removing gender from a target leads the target to be dehumanized because perceivers lack a rich (gender) schema to guide their sensemaking, we conducted two separate mediation analyses using PROCESS Model 4 (5,000 bootstraps; Hayes, 2017). To compare conditions, we created three dummy variables, each capturing an experimental condition: one represented the men condition (1 = men, 0 = women, 0 = people), one representing the women condition (1 = women, 0 = men, 0 = people), and one representing the people condition (1 = people, 0 = men, 0 = women). When entering any two dummy variables, they are independent of each other and relative to nonincluded dummy variable; thus, all three levels of condition are represented when entering two of the dummy variables (see Hayes, 2017; Hayes & Preacher, 2014). We expected that the lesser humanness ascribed to people, as compared to men and women, would be due to the lesser masculine and feminine traits (i.e., a gender schema) attributed to people, compared to men and women, respectively.
为了测试从目标中删除性别是否会导致目标非人化，因为感知者缺乏丰富的（性别）图式来指导他们的意义建构，我们使用 PROCESS 模型 4（5,000 个引导;Hayes，2017 年）。为了比较条件，我们创建了三个虚拟变量，每个变量捕获一个实验条件：一个代表男性条件（1 = 男性，0 = 女性，0 = 人），一个代表女性条件（1 = 女性，0 = 男性，0 = 人），一个代表人条件（1 = 人，0 = 男性，0 = 女性).输入任意两个虚拟变量时，它们彼此独立，并且相对于未包含的虚拟变量;因此，在输入两个虚拟变量时，所有三个级别的条件都表示出来（参见 Hayes，2017 年;Hayes & Preacher，2014 年）。我们预计，与男性和女性相比，人们被赋予的人类性较低，这将是由于与男性和女性相比，人们的男性和女性特征（即性别图式）较小。

We compared the “people” condition to the “men” condition, hypothesizing that perceptions of masculinity should mediate the greater humanness ascribed to men. To do so we entered the dummy variables for people (1 = people, 0 = women, 0 = men) and women (1 = women, 0 = men, 0 = people), which allows us to compare “people” to a baseline condition (men) and test for a mediational path through masculinity. As predicted, we find a significant indirect effect, indirect effect = − 2.38, SE = .87, CI₉₅ = − 4.28, −.87, such that “people” were seen as less human than “men,” in part because they were ascribed fewer masculine traits.
我们将“人”状况与“男人”状况进行了比较，假设对男子气概的感知应该调解赋予男性的更深层次的人性。为此，我们输入了人（1 = 人，0 = 女性，0 = 男性）和女性（1 = 女性，0 = 男性，0 = 人）的虚拟变量，这使我们能够将“人”与基线条件（男性）进行比较，并测试通过男性气质的中介路径。正如预测的那样，我们发现了一个显著的间接效应，间接效应 = − 2.38，SE = .87，CI₉₅ = − 4.28，−.87，使得“人”被认为不如“男性”人性化，部分原因是他们被赋予的男性特征较少。

We next compared the “people” condition to the “women” condition by entering the dummy variables for people (1 = people, 0 = women, 0 = men) and women (1 = men, 0 = women, 0 = people), which allows us to compare “people” to a baseline condition (women) and test for a mediational path through femininity. We again find a significant indirect effect = − 6.35, SE = 1.43, CI₉₅ = − 9.42, − 3.72, such that “people” were seen as less human, compared to “women,” because they were ascribed fewer feminine traits. Full regression tables and mediation models are reported in SOM.
接下来，我们通过输入人（1 = 人，0 = 女性，0 = 男性）和女性（1 = 男性，0 = 女性，0 = 人）的虚拟变量，将“人”条件与“女性”条件进行比较，这使我们能够将“人”与基线条件（女性）进行比较，并测试通过女性气质的中介路径。我们再次发现一个显著的间接效应 = − 6.35， SE = 1.43， CI₉₅ = − 9.42， − 3.72，因此与“女性”相比，“人”被视为不那么人性化，因为他们被赋予的女性特征较少。SOM 中报告了完整的回归表和中介模型。

Study 6 showed that “people” were rated less human than “men” and “women.” The results suggest that when the target group was “men” or “women,” participants had a rich gender schema to guide their sensemaking. Gender provides a backdrop for organizing information about a target and an entry point for predicting how the individual or group will behave (i.e., men will be agentic, women will be communal). While decades of social–psychological research document the downside of gender stereotypes (e.g., Heilman, 2001; Rudman & Phelan, 2008), here we establish that gender helps humanize targets and that targets described without a gender are granted less than full personhood.
研究 6 显示，“人”的评价低于“男性”和“女性”。结果表明，当目标群体是 “男性 ”或 “女性 ”时，参与者有丰富的性别图式来指导他们的意义建构。性别为组织有关目标的信息提供了背景，并为预测个人或群体的行为方式提供了切入点（即，男性将是代理性的，女性将是公共的）。虽然几十年的社会心理学研究记录了性别刻板印象的负面影响（例如，Heilman，2001 年;Rudman & Phelan，2008），在这里我们确定性别有助于使目标人性化，而没有性别描述的目标被赋予的不是完整的人格。

Notably, women were more humanized than were men and people. This result is surprising given evidence that women are more dehumanized than are men (e.g., Nussbaum, 1999), but consistent with recent research to suggest that communal traits (i.e., warmth, kindness) take primacy in the process of humanization (Chu & Martin, 2021); the primacy of communality also existed in this data, as well, where the relationship between femininity (i.e., communality) and humanization was twice as large (r = .33, p < .001) as the relationship between masculinity (i.e., agency) and humanization (r = .15, p = .002). Additionally, results of this study converge with work on the attribution of mind to groups (e.g., Cooley et al., 2017; Morewedge et al., 2013; Waytz & Young, 2012). People see groups as having mind, and subtle linguistic changes in descriptions of them can shape how strongly groups are attributed mind (i.e., “people in groups” have more mind than “groups of people”). Here, we show that groups described as “men” or “women” are attributed greater humanity than those that are described as “people.”
值得注意的是，女性比男性和人更人性化。鉴于女性比男性更非人化的证据（例如，Nussbaum，1999），这个结果令人惊讶，但与最近的研究一致，即在人性化过程中，社区特征（即温暖、善良）占据主导地位（Chu & Martin，2021）;公共性的首要地位也存在于这些数据中，其中女性气质（即公共性）和人性化之间的关系是男性气质（即能动性）和人性化（r = .15，p = .002）之间关系的两倍 （r = .33， p < .001）。 此外，这项研究的结果与将心智归因于群体的工作趋同（例如，Cooley 等人，2017 年;Morewedge等人，2013 年;Waytz & Young，2012 年）。人们将群体视为有思想，对他们描述中的细微语言变化可以塑造群体归因于思想的强度（即，“群体中的人”比“群体”有更多的思想”）。在这里，我们表明，被描述为“男性”或“女性”的群体比被描述为“人”的群体被赋予了更大的人性。

Study 7 had a twofold aim. First, it sought to replicate Study 6’s finding that targets stripped of gender are seen as less human because perceivers lack a schema to guide their sensemaking. Second, it sought more direct evidence of the causal pathway we propose: Gendered targets are associated with a gender schema (i.e., stereotypes, gendered traits), which facilitates sensemaking (i.e., predictability, understanding), which in turn predicts attributions of humanness.
研究 7 有双重目标。首先，它试图复制研究 6 的发现，即被剥夺性别的目标被视为不那么人性化，因为感知者缺乏指导他们意义建构的图式。其次，它寻求我们提出的因果途径的更直接证据：性别目标与性别图式（即刻板印象、性别特征）相关联，这促进了意义建构（即可预测性、理解），进而预测了人类的归因。

In this preregistered study (https://aspredicted.org/blind.php?x=xg3xf7), participants (N = 304) from Amazon’s MTurk were recruited to take part in a study on “meeting alien life.” Participants who provided nonsensical responses and those who failed a manipulation check (the gender of the alien) were excluded (N = 13), leaving a final sample of 291 participants for analysis (55% men and 45% women; 69% White, 7% Black, 16% Asian, 5% Hispanic, and 2% Other-Race participants, with a mean age of 40.86 [SD = 12.34]).
在这项预先注册的研究 （https://aspredicted.org/blind.php?x=xg3xf7） 中，来自亚马逊 MTurk 的参与者 （N = 304） 被招募参加一项关于“遇见外星生命”的研究。提供无意义回答的参与者和未通过操纵检查的参与者（外国人的性别）被排除在外（N = 13），留下 291 名参与者的最终样本进行分析（55% 男性和 45% 女性;69% 白人，7% 黑人，16% 亚洲人，5% 西班牙裔和 2% 其他种族参与者，平均年龄 40.86 [SD = 12.34]）。

Participants were told to “imagine that they were a crew of astronauts that has landed on Freon, a planet just beyond Neptune, whose inhabitants include an alien species called ‘the Orinthians’” (used in Hoffman & Hurst, 1990). They were randomly assigned to one of three conditions: male, female, nongendered. In all conditions, the participant was told to imagine that they see an Orinthian in the distance and that it did not seem dangerous. In the male condition, they were told they could tell it was a male member of the species. In the female condition, they were told they could tell it was a female member of the species. In the nongendered condition, this information was left out. Participants were then asked to write a few sentences about the alien and their interaction with it. Participants then completed several dependent variables, were thanked for their participation, and were paid.
参与者被告知“想象他们是一群宇航员，他们已经降落在海王星以外的氟利昂星球上，那里的居民包括一种叫做'奥林斯人'的外星物种“（在Hoffman & Hurst， 1990中使用）。他们被随机分配到以下三种情况之一：男性、女性、无性别。在所有情况下，参与者都被告知想象他们在远处看到一只奥林斯猫，而且它似乎并不危险。在雄性情况下，他们被告知他们可以看出这是该物种的雄性成员。在雌性情况下，他们被告知他们可以看出这是该物种的雌性成员。在无性别条件下，这些信息被遗漏了。然后要求参与者写几句关于外星人以及他们与外星人互动的句子。然后，参与者完成了几个因变量，感谢他们的参与，并获得了报酬。

Participants rated the humanness of the Orinthian using the “humanness” measure used in our other studies (0 = not at all human to 100 = very human).
参与者使用我们其他研究中使用的“人性”衡量标准（0 = 完全没有人性到 100 = 非常人性化）对奥林斯人的人性进行评分。

Participants were asked the extent to which they believed a number of traits were characteristic of the Orinthian (1 = not at all to 5 = very much). Twelve items measured masculinity (active, confident, efficient, skilled, masculine, independent; a = .75) and femininity (warm, kind, nurturing, feminine, affectionate, friendly; a = .88).
参与者被问及他们认为许多特征是奥林斯特征的程度（1 = 完全没有到 5 = 非常多）。12 个项目测量了男子气概（积极、自信、高效、熟练、阳刚、独立;a = .75）和女性气质（温暖、善良、养育、女性化、深情、友好;a = .88）。

To assess the extent to which gender eased participant sensemaking by providing a framework to organize their thinking about the target, we used five items: “I could predict how the alien would act and behave,” “I felt like I knew how to interact with the alien,” “I felt like I knew the alien’s intentions,” “I felt like I could predict how the alien would interact with me,” and “I would have some idea of the alien’s likes and dislikes” (a = .93).
为了评估性别通过提供一个框架来组织参与者对目标的思考，在多大程度上缓解了参与者的意义建构，我们使用了五个项目：“我可以预测外星人会如何行动和行为”，“我觉得我知道如何与外星人互动”，“我觉得我知道外星人的意图”，“我觉得我可以预测外星人将如何与我互动，”“和”我会对外星人的好恶有所了解“（a = .93）。

As expected, there was a main effect of condition, F(2, 288) = 4.92, p = .008, η _p ²  = .033. The nongendered alien was seen as less human as compared to both the male, F(1, 288) = 8.60, p = .004, η _p ²  = .029, CI₉₅ = 3.44, 17.48, and female, F(1, 288) = 6.26, p = .013, η _p ²  = .021, CI₉₅ = 1.89, 15.80, alien. Ratings of the male and female alien were not significantly different (p = .64). See Table 4 for means.
正如预期的那样，存在条件的主效应，F（2， 288） = 4.92，p = .008，η _p ²  = .033。与男性 F（1， 288） = 8.60， p = .004， η _p ²  = .029， CI₉₅ = 3.44， 17.48 和女性相比，无性别的外国人被认为不那么人性化，F（1， 288） = 6.26， p = .013， η _p ²  = .021， CI₉₅ = 1.89， 15.80， 外星人。男性和女性外国人的评分没有显著差异 （p = .64）。有关平均值，请参见表 4。

 放大此图像。

In terms of masculine traits, there was a significant effect of condition, F(2, 288) = 13.19, p < .001, η _p ²  = .084, such that the male alien was ascribed more masculine traits than the female, F(1, 288) = 20.83, p < .001, η _p ²  = .067, CI₉₅ = .25, .63, and nongendered (control) alien, F(1, 288) = 18.42, p < .001, η _p ²  = .060, CI₉₅ = .23, .63. Ratings of the female and nongendered aliens were not significantly different (p = .92). By contrast, for feminine traits there was a significant effect of condition, F(1, 288) = 16.29, p < .001, η _p ²  = .10; participants assigned the female alien more feminine traits than both the male, F(1, 288) = 25.50, p < .001, η _p ²  = .081, CI₉₅ = .36, .83, and nongendered aliens, F(1, 288) = 22.52, p < .001, η _p ²  = .073, CI₉₅ = .34, .81. Ratings of the male and nongendered aliens were not significantly different (p = .87). See Table 4 for means.
在男性特征方面，条件有显着影响，F（2， 288） = 13.19，p < .001，η _p ²  = .084，因此男性外来者比女性被赋予更多的男性特征，F（1， 288） = 20.83，p < .001，η _p ²  = .067，CI ₉₅ = .25，.63，无性别（对照）外来者，F（1， 288） = 18.42，p < .001，η _p ²  = .060，CI₉₅ = .23，.63。女性和非性别外国人的评分没有显著差异 （p = .92）。相比之下，对于女性特征，条件有显着影响，F（1， 288） = 16.29，p < .001，η _p ²  = .10; 参与者为女性外星人分配的女性特征比男性多，F（1， 288） = 25.50，p < .001，η _p ²  = .081，CI₉₅ = .36，.83，和非性别外星人，F（1， 288） = 22.52，p < .001，η _p ²  = .073，CI₉₅ = .34，.81。 男性和非性别外国人的评分没有显著差异 （p = .87）。有关平均值，请参见表 4。

As predicted, there was a significant effect of condition, F(2, 288) = 4.95, p = .008, η _p ²  = .033. On average, participants had greater difficulty making sense of the nongendered alien as compared to the male, F(1, 288) = 4.45, p = .036, η _p ²  = .015, CI₉₅ = .03, .90, and female aliens, F(1, 288) = 9.60, p = .002, η _p ²  = .032, CI₉₅ = .25, 1.11. Ratings of male and female aliens were not significantly different (p = .32). See Table 4.
正如预测的那样，条件有显着影响，F（2， 288） = 4.95，p = . 008，η _p ²  = .033。平均而言，与男性相比，参与者更难理解无性别的外国人，F（1， 288） = 4.45，p = .036，η _p ²  = .015，CI₉₅ = .03，.90，女性外国人，F（1,288） = 9.60，p = .002，η _p ²  = . 032，CI₉₅ = .25,1.11。 男性和女性外国人的评分没有显著差异 （p = .32）。请参阅 表 4。

We expected a serial mediation where the diminished humanness ascribed to the nongendered alien (as compared to the male and female aliens) was due to the lack of a gender schema to guide sensemaking (Med₁ = gendered traits), which in turn predicted more subjective difficulty understanding the alien and predicting its behavior (Med₂ = sensemaking), which in turn led it to be seen as less human (DV). To establish this pathway, we conducted a serial-mediation analysis using PROCESS Model 6 (5,000 bootstraps), with a three-level categorical variable, computing two dummy variables that when both entered represent all three levels (as in Study 6). Of note, all regression tables, mediation models, and alternative mediations are reported in SOM.
我们预计会出现一系列中介，其中无性别的外星人（与男性和女性外星人相比）的人性减弱是由于缺乏指导意义构建的性别图式（Med₁ = 性别特征），这反过来又预测了理解外星人和预测其行为的更多主观困难（Med₂= 意义建构），这反过来又导致它被视为不那么人性化 （DV）。为了建立这条通路，我们使用 PROCESS 模型 6（5,000 个引导程序）进行了序列中介分析，使用三级分类变量，计算两个虚拟变量，当两个虚拟变量都进入时代表所有三个水平（如研究 6 所示）。值得注意的是，所有回归表、中介模型和替代中介都在 SOM 中报告。

We compared the control condition to the male condition, hypothesizing that perceptions of masculinity should provide more information about the alien, which in turn guides people’s sensemaking, and leads the alien to be seen as more human. We again entered the dummy variables for the control (1 = control, 0 = female, 0 = male) and female (1 = female, 0 = male, 0 = control) conditions, allowing us to compare the control (nongendered) alien to a baseline condition (male) and test for a mediational path through masculine traits (Med₁) and sensemaking (Med₂). Indeed, this pathway was supported, where there was a significant serial mediation, indirect effect = 1.22, SE = .46, CI₉₅ = .44, 2.26. See SOM for more detail.
我们将控制条件与男性条件进行了比较，假设对男性气质的感知应该提供有关外星人的更多信息，这反过来又指导人们的意义建构，并导致外星人被视为更人性化。我们再次输入了对照（1 = 对照，0 = 女性，0 = 男性）和女性（1 = 女性，0 = 男性，0 = 对照）条件的虚拟变量，使我们能够将对照（无性别）外来者与基线条件（男性）进行比较，并通过男性特征（Med₁）和意义建构（Med₂）测试中介路径。事实上，这条途径得到了支持，在存在显着的连续中介的情况下，间接效应 = 1.22，SE = .46，CI₉₅ = .44,2.26。有关更多详细信息，请参阅 SOM。

We next compared the nongendered alien to the female one by entering the dummy variables for the control (1 = control, 0 = female, 0 = male) and female (1 = male, 0 = female, 0 = control) conditions, which allows us to compare the control (nongendered) alien to a baseline condition (female) and test for a mediational path through femininity (Med₁) and sensemaking (Med₂). As expected, a significant proportion of the variance explaining the effect of the female (compared to the nongendered) condition (IV) on humanness (DV) was accounted for through feminine traits (Med₁) and sensemaking (Med₂) pathway, indirect effect = 2.22, SE = .77, CI₉₅ = .91, 3.93. See SOM for more detail.
接下来，我们通过输入对照（1 = 对照，0 = 女性，0 = 男性）和女性（1 = 男性，0 = 女性，0 = 对照）条件的虚拟变量，将无性别的外星人和女性进行比较，这使我们能够将对照（无性别）外星人与基线条件（女性）进行比较，并测试通过女性气质（Med₁）和意义建构（Med₂）的中介路径).正如预期的那样，解释女性（与非性别相比）条件 （IV） 对人类性 （DV） 影响的方差的很大一部分是通过女性特征 （Med₁） 和意义建构 （Med₂） 途径来解释的，间接影响 = 2.22，SE = .77，CI₉₅ = . 91,3.93。有关更多详细信息，请参阅 SOM。

Study 7 demonstrated that gender provides a schema that guides sensemaking about women (feminine) and men (masculine); consequently, its absence leads perceivers to struggle to make sense of a target (e.g., to ascribe it motives, predict its behavior, etc.) and, in turn, to see it as less human.
研究 7 表明，性别提供了一种模式，指导对女性（女性）和男性（男性）的意义构建;因此，它的缺失导致感知者难以理解目标（例如，赋予它动机，预测它的行为等），进而将其视为不那么人性化。

Studies 5–7 converge on the finding that targets described without gender are seen as less human, and Studies 6 and 7 provide evidence of the mechanism by which genderlessness affects ascriptions of personhood. However, these studies have either entailed nonhuman entities (Study 5) or withholding gender information rather than describing targets as “genderless” (Studies 6 and 7). This calls into question whether these findings have any bearing for current societal trends toward genderlessness. The first aim of our final study was to measure perceptions of male and female human beings relative to targets who were explicitly described as “genderless” and to examine whether they are dehumanized. The second aim of this study was to consider the downstream implications of dehumanization, namely people’s ability to relate and connect to genderless others. Building on our findings from Studies 6 to 7, if people cannot predict or understand others without a gender schema, they should be less likely to connect to genderless others.
研究 5-7 一致发现，没有性别描述的目标被视为不那么人性化，研究 6 和 7 提供了无性别影响人格归属的机制的证据。然而，这些研究要么涉及非人类实体（研究 5），要么隐瞒性别信息，而不是将目标描述为“无性别”（研究 6 和 7）。这让人质疑这些发现是否与当前无性别的社会趋势有任何关系。我们最终研究的第一个目标是衡量相对于被明确描述为“无性别”的目标对男性和女性的看法，并检查他们是否被非人化。本研究的第二个目的是考虑非人化的下游影响，即人们与无性别他人建立联系和联系的能力。基于我们研究 6 到 7 的发现，如果人们在没有性别图式的情况下无法预测或理解他人，那么他们应该不太可能与无性别他人建立联系。

In this preregistered study (https://aspredicted.org/blind.php?x=ua3ed4), participants (N = 450) were recruited via MTurk for a study on “impression formation.” Participants were randomly assigned to one of three impression-formation conditions: male target, female target, or genderless target. After removing participants who failed to recall the gender of the individual and/or provided nonsensical responses (N = 78), we were left with 372 participants (54% male, 46% female, and 0.3% nonidentified participants; 72% White, 7% Black, 8% Asian, 7% Hispanic, and 2% Other-Race participants, with a mean age of 37.91 [SD = 12.10]).
在这项预先注册的研究 （https://aspredicted.org/blind.php?x=ua3ed4） 中，参与者 （N = 450） 是通过 MTurk 招募的，用于一项关于 “印象形成” 的研究。参与者被随机分配到三种印象形成条件之一：男性目标、女性目标或无性别目标。在剔除未能回忆起个人性别和/或提供无意义回答的参与者 （N = 78） 后，我们只剩下 372 名参与者 （54% 男性、46% 女性和 0.3% 未识别参与者;72% 白人、7% 黑人、8% 亚洲、7% 西班牙裔和 2% 其他种族参与者，平均年龄为 37.91 [SD = 12.10]）。

Participants were told we were interested in their initial impressions of a hypothetical person, based on limited information. They were then provided one of three descriptions of an individual, depending on their condition (adapted from Martin et al., 2019):
参与者被告知，根据有限的信息，我们对他们对假设人物的初步印象感兴趣。然后，根据他们的状况，向他们提供对个人的三种描述之一（改编自 Martin 等人，2019 年）：

This person is a [man/woman/genderless individual]. They use [male/female/do not use] pronouns, wear [men’s/women’s/neutral] clothing, and look somewhat [masculine/feminine/androgynous]. [He/She/They] is/are from a suburb of a large city, and while [he/she/they] live comfortably, they are not wealthy. [He/She/They] enjoys going out on weekends with friends, watching Netflix, and listening to music.
此人是 [男性/女性/无性别个体]。他们使用 [男/女/不使用] 代词，穿着 [男/女/中性] 服装，看起来有点 [男/女/中性]。[他/她/他们] 来自大城市的郊区，虽然 [他/她/他们] 生活舒适，但他们并不富有。[他/她/他们] 喜欢周末和朋友出去玩、看 Netflix 和听音乐。

Participants then completed our dependent variables, which include humanization, human nature, human uniqueness, and social inclusion, and were then thanked and paid.  ^{[  8  ]} 
然后，参与者完成了我们的因变量，包括人性化、人性、人类独特性和社会包容性，然后获得感谢和报酬。    ^{[ 8  ]} 

To measure humanization, we administered the same “humanness” scale as used in past studies (as well as other measures of humanization, which can be found in SOM).
为了衡量人源化程度，我们采用了与过去研究相同的“人性”量表（以及人源化的其他衡量标准，可以在 SOM 中找到）。

We used a four-item measure of social connection. Items from this scale include “I feel very similar to this person,” “I think I would like this person,” “I think me and this person would get along,” and “I would feel a lot of respect for this person” (1 = strongly disagree to 7 = strongly agree; α = .90).
我们使用了四项社会联系测量方法。这个量表中的项目包括“我觉得和这个人很像”、“我想我喜欢这个人”、“我认为我和这个人会相处得很好”和“我会对这个人感到非常尊重”（1 = 非常不同意 7 = 非常同意;α = .90）。

Ascriptions of humanness varied across the three targets, F(2, 369) = 4.42, p = .013, η _p ²  = .023. Those who read about the genderless individual ascribed the target with less humanity (M = 83.16, SD = 22.15) than did those who read about either the male, (M = 89.19, SD = 16.91), F(1, 369) = 6.45, p = .011, η _p ²  = .017, CI₉₅ = − 10.69, − 1.36, or female targets (M = 89.17, SD = 15.43), F(1, 369) = 6.84, p = .009, η _p ²  = .018, CI₉₅ = − 10.53, − 1.49. There was no difference in humanity ascription to male and female targets (p = .995). We also find that ratings of social connectedness varied across the three conditions, F(2, 369) = 19.84, p < .001, η _p ²  = .097. Those who read about the genderless target felt less connected to the individual (M = 4.26, SD = 1.60) than did those who read about the male target (M = 5.21, SD = 1.25), F(1, 369) = 29.71, p < .001, η _p ²  = .075, CI₉₅ = − 1.30, −.61, or female target (M = 5.18, SD = 1.18), F(1, 369) = 30.00, p < .001, η _p ²  = .075, CI₉₅ = − 1.26, −.60; again, ratings of the latter two did not differ (p = .88).
人类性的归类在三个目标中有所不同，F（2， 369） = 4.42，p = .013，η _p ²  = .023。那些阅读无性别个体的人认为目标的人性较低 （M = 83.16， SD = 22.15） 比那些阅读男性 （M = 89.19， SD = 16.91）， F（1， 369） = 6.45， p = .011， η _p ²  = .017， CI₉₅ = − 10.69， − 1.36 或女性目标 （M = 89.17， SD = 15.43） 的人。 F（1， 369） = 6.84， p = .009， η _p ²  = .018， CI₉₅ = − 10.53， − 1.49.人类对男性和女性目标的归因没有差异 （p = .995）。我们还发现，社会联系的评级在三个条件下有所不同，F（2， 369） = 19.84，p < . 001，η _p ²  = .097。与阅读男性目标 （M = 5.21， SD = 1.25）， F（1 ， 369） = 29.71， p < .001， η _p ²  = .075， CI₉₅ = − 1.30， -.61 或女性目标 （M = 5.18， SD = 1.18）， F（1， 369） = 30.00， p < .001，η _p ²  = .075，CI₉₅ = − 1.26，−.60;同样，后两者的评分没有差异 （p = .88）。

Again, our hypothesis was that ascriptions of humanness have implications for interpersonal outcomes, such as social connection. To test whether the lesser social connection participants felt to the genderless individual, compared to the man and woman, we conducted two mediation analyses (Hayes, 2017), as in Studies 6 and 7, first entering the dummy variables for the male (1 = male, 0 = female, 0 = genderless) and female (1 = female, 0 = male, 0 = genderless) conditions, to compare the genderless target to the male target, and then entering the dummy variables for the female (1 = female, 0 = male, 0 = genderless) and male (1 = male, 0 = female, 0 = genderless) conditions, to compare the genderless target to the female target. We expected the lesser social connection felt by participants when evaluating the genderless target (compared to the gendered targets) would be due to the lesser humanness ascribed to them. Indeed, we find that the effects of condition on social connectedness were mediated by humanity ascription: The lack of social connection participants felt for the genderless target was due to the lower levels of humanity ascribed to them, compared to both the male (indirect effect = −.10, SE = .05, CI₉₅ = −1.16, −.50) and female (indirect effect = −.10, SE = .05, CI₉₅ = − 1.19, − .52) targets.
同样，我们的假设是，对人类性的归类对人际结果有影响，例如社会联系。为了测试与男性和女性相比，参与者是否对无性别个体感到社会联系较少，我们进行了两次中介分析（Hayes，2017），如研究 6 和 7，首先输入男性（1 = 男性，0 = 女性，0 = 无性别）和女性（1 = 女性，0 = 男性，0 = 无性别） 条件，将无性别目标与男性目标进行比较，然后输入女性（1 = 女性、0 = 男性、0 = 无性别）和男性（1 = 男性、0 = 女性、0 = 无性别）条件的虚拟变量，以将无性别目标与女性目标进行比较。我们预计参与者在评估无性别目标时（与有性别目标相比）感受到的社会联系较少，这将是由于赋予他们的人性较少。事实上，我们发现条件对社会联系的影响是由人性归因介导的：参与者对无性别目标缺乏社会联系是由于与男性（间接效应 = −.10，SE = .05，CI₉₅ = −1.16，−.50）和女性（间接效应 = −.10， SE = .05， CI₉₅ = − 1.19， − .52） 目标。

Study 8 found that people ascribed less humanness to a target who was lacking gender (i.e., “genderless”) as compared to those who were described in gendered terms (i.e., man and woman). Further, we find that the lower humanness ascription for those described as “genderless” had downstream consequences; participants felt less liking toward, respect for, and connection to the genderless target (as compared to the male and female target). Further, we find that these effects hold even when considering political ideology—an individual difference shown to predict discrimination toward those with gender-nonconforming identities (Stern & Rule, 2018). Thus, these findings suggest that the dehumanization of gender-nonconforming individuals purported in past research (see Prusaczyk & Hodson, 2020; Stern & Rule, 2018) may not only be due to disapproval of gender norm violation, but may also stem from an inability to conceptualize a human beyond the binary distinction of men and women. Further, it may be that participants coded the genderless target as gender-nonconforming (e.g., transgender, a-gender) rather than as someone without a gender (as our theory might suggest). Thus, understanding how both attitudes (e.g., transphobia) and cognitive complexity (e.g., gender schema) might explain the dehumanization of gender-nonconforming individuals is important to explore, as both traditional gender ideologies and desire for cognitive closure are more strongly held for political conservatives (Jost et al., 2003; Prusaczyk & Hodson, 2020; Stern & Rule, 2018).
研究 8 发现，与以性别术语描述的目标（即男性和女性）相比，人们将缺乏性别（即“无性别”）的人性归因于较少的人性。此外，我们发现，那些被描述为“无性别”的人的较低人类性归类产生了下游后果;参与者对无性别目标的喜欢、尊重和联系较少（与男性和女性目标相比）。此外，我们发现即使在考虑政治意识形态时，这些影响也成立——个体差异被证明可以预测对那些具有性别不一致身份的人的歧视（Stern & Rule，2018）。因此，这些发现表明，过去研究中声称的性别不一致个体的非人化（参见 Prusaczyk & Hodson，2020 年;Stern & Rule，2018 年）可能不仅是由于不赞成违反性别规范的行为，也可能是由于无法超越男性和女性的二元区别来概念化人类。此外，参与者可能将无性别目标编码为性别不一致（例如，跨性别、无性别），而不是没有性别的人（正如我们的理论所暗示的那样）。因此，了解态度（例如，跨性别恐惧症）和认知复杂性（例如，性别图式）如何解释性别不一致个体的非人化是很重要的，因为传统的性别意识形态和对认知闭合的渴望在政治保守派中更为强烈（Jost 等人，2003 年;Prusaczyk & Hodson， 2020;Stern & Rule，2018 年）。

Across 10 studies, we demonstrate that gender is most commonly used in, and strongly related to, ascribing humanness. Studies 1–4 show that granting a target personhood results in attributing gender to the target more so than other social categories. Whether in the form of recall, perceiving moving shapes, imbuing products with human traits, or creating a “human form,” we find that gender is more tightly linked to perceiving personhood than other social groups to which people belong (e.g., race, age). The effect size for gender ascription was over three times as large as any other social category, and in every study, the majority of participants ascribed an anthropomorphized entity with gender (which was not the case for other social categories). These findings provide support for the untested assumption that gender is a defining feature of humanity (Butler, 1990; Lorber, 1994; Nussbaum, 1999); targets that are granted humanity are presumed to possess a (binary) gender. Studies 5–8 converge on this finding, showing that those who perform gender—as clearly male or female—are ascribed more humanness, and those who do not are dehumanized. The diminished humanness ascribed to nongendered and genderless targets is due, at least in part, to the lack of a gender schema to guide facile and efficient sensemaking. The relative difficulty perceivers had in making sense of nongendered/genderless targets predicted diminished humanness ratings. Moreover, we show the downstream consequences for people who do not possess (binary) gender: They are more dehumanized, less liked, and seen as more socially distant.
在 10 项研究中，我们证明了性别最常用于归因于人类，并且与人类性密切相关。研究 1-4 表明，与其他社会类别相比，授予目标人格会导致将性别更多地归因于目标。无论是以回忆的形式、感知移动的形状、为产品注入人类特征，还是创造“人形”，我们发现性别与感知人格的联系比人们所属的其他社会群体（例如种族、年龄）更紧密。性别归属的效应量是任何其他社会类别的三倍多，在每项研究中，大多数参与者将性别归类为拟人化实体（其他社会类别并非如此）。这些发现为性别是人类的一个决定性特征这一未经检验的假设提供了支持（Butler， 1990;Lorber，1994 年;Nussbaum，1999 年）;被赋予人类性的目标被推定为具有（二进制）性别。研究 5-8 都集中在这一发现上，表明那些执行性别的人——显然是男性或女性——被赋予了更多的人性，而那些没有的人则被非人化了。无性别和无性别目标所赋予的人性减弱，至少部分是由于缺乏性别图式来指导简单而有效的意义建构。感知者在理解无性别/无性别目标方面的相对困难预测了人类评分的降低。此外，我们展示了不拥有（二元）性别的人的下游后果：他们更加非人化，更不受欢迎，并被视为更疏远。

These findings make several contributions to literature on gender and (de)humanization, and potentially speak to public dialogue and discourse around gender issues.
这些发现为关于性别和（非）人化的文献做出了多项贡献，并可能涉及围绕性别问题的公共对话和讨论。

These findings converge with the existing literature to suggest that gender is a lens through which we see our social realities, taking primacy in the understanding of ourselves (Bem, 1981; Markus & Oyserman, 1989; Markus et al., 1982; Martin & Ruble, 2004), others (Kenrick et al., 1994; Martin & Ruble, 2004; Prentice & Miller, 2007), and even other social categories (e.g., race, sexual orientation; Freeman et al., 2010; Hall et al., 2015; Johnson et al., 2007, 2012). This work suggests that gender is not only important for self and social perception, but also is core to what it means to be seen as human. None of the target stimuli used in Studies 1–4 have membership in any social category—they are not human—but as soon as they are ascribed humanness, they are presumed to have a gender, being categorized into two dimensions: “male” and “female” (see Study 3). Thus, although much work in social perception tries to control for the effect of gender by describing targets in a genderless way (see Hester & Gray, 2020), our findings suggest that gender is often ascribed to a target by default—even when no gender information is provided—and that when it is not, the target is seen as less than fully human. Our findings suggest that using genderless targets in psychological science may not control for statistical noise: It may create it.
这些发现与现有文献相吻合，表明性别是我们看待社会现实的镜头，在理解自身方面处于首要地位（Bem， 1981;Markus & Oyserman， 1989;Markus et al.， 1982;Martin & Ruble， 2004），其他（Kenrick等人，1994;Martin & Ruble， 2004;Prentice & Miller，2007），甚至其他社会类别（例如，种族、性取向;Freeman et al.， 2010;Hall et al.， 2015;Johnson et al.， 2007， 2012）。这项工作表明，性别不仅对自我和社会认知很重要，而且是被视为人类意味着的核心。研究 1-4 中使用的目标刺激都不属于任何社会类别——它们不是人类——但一旦它们被赋予人类性，它们就被假定具有性别，被分为两个维度：“男性”和“女性”（见研究 3）。因此，尽管社会认知中的许多工作试图通过以无性别的方式描述目标来控制性别的影响（参见Hester&Gray，2020年），但我们的研究结果表明，即使没有提供性别信息，性别通常被默认归因于目标——而当没有提供性别信息时，目标被视为不完全的人类。我们的研究结果表明，在心理科学中使用无性别目标可能无法控制统计噪声：它可能会产生统计噪声。

Our findings suggest that a reason gender may play an especially central role is a functional one: Gender is an elaborate and therefore efficient schema. Evidence suggests that nearly all human traits can be reliably sorted into male and female (Martin & Slepian, 2020). Ascribing gender to a target provides a wealth of (mis)information from which perceivers can build. Indeed, people seem to depend on it as its absence results in processing disfluency (i.e., impedes sensemaking; Stern & Rule, 2018) and gives rise to anxiety and discomfort (Lick & Johnson, 2015; West & Zimmerman, 1987).
我们的研究结果表明，性别可能发挥特别核心作用的一个原因是功能性的：性别是一个精心设计的，因此是有效的图式。有证据表明，几乎所有的人类特征都可以可靠地分为男性和女性（Martin & Slepian，2020）。将性别归因于目标提供了大量（错误）信息，感知者可以从中构建。事实上，人们似乎依赖它，因为它的缺失会导致处理不流畅（即，阻碍意义构建;Stern & Rule，2018 年）并引起焦虑和不适（Lick & Johnson，2015;West & Zimmerman，1987年）。

Further, while the present investigation has little to offer the hotly debated question of which social categories are most (visually) salient (Ito & Urland, 2005; Kite et al., 1991), it provides clear evidence that gender is more central than the others to conceiving of humanness. We are uniquely suited to make this argument because our “blank slate” methodological approach permits isolating how likely each social category is to “piggy back” on ascriptions of humanness. In our studies, participants granted humanity to nonhuman entities and then reported on social-category membership where these social categories technically did not exist.
此外，虽然目前的调查几乎没有提供激烈争论的问题，即哪些社会类别最（视觉上）突出（Ito & Urland，2005年;Kite et al.， 1991），它提供了明确的证据，表明性别在构想人类方面比其他性别更重要。我们特别适合提出这个论点，因为我们的 “白纸 ”方法论方法允许隔离每个社会类别 “背负 ”人性归类的可能性。在我们的研究中，参与者赋予非人类实体人性，然后报告这些社会类别在技术上不存在的社会类别成员身份。

Our work calls for more research to connect and bridge gender literature across disciplines. As recent work has argued, gender is an elaborate schema that not only exists at a psychological level but is created and reinforced at a societal, institutional, and cultural one, as well (Diekman & Schmader, 2021; Martin & Slepian, 2020). Psychology is particularly well positioned to answer questions and test hypotheses posited in other disciplines and domains. For example, by establishing that gender is a core aspect of perceiving personhood, this work contributes to research that adopts a sociological perspective to the study of gender (see Butler, 1990; Lorber, 1994; Morgenroth & Ryan, 2018, 2020). As discussed, this literature presumes gender plays a fundamental role in how human beings interact and behave but, to this point, this assumption has remained untested (see Morgenroth & Ryan, 2018, p. 6). Our empirical test bridges the sociological and psychological perspectives on gender and heeds calls to examine it as a psychological process that results from humanization rather than as an independent variable to predict differences between men and women (Morgenroth & Ryan, 2018).
我们的工作呼吁更多的研究来连接和桥接跨学科的性别文献。正如最近的工作所论证的，性别是一个精心设计的图式，不仅存在于心理层面，而且在社会、制度和文化层面也被创造和强化（Diekman & Schmader，2021;Martin & Slepian，2020 年）。心理学特别擅长回答其他学科和领域的问题和测试假设。例如，通过确定性别是感知人格的核心方面，这项工作有助于采用社会学视角进行性别研究的研究（参见 Butler，1990 年;Lorber，1994 年;Morgenroth & Ryan，2018年，2020年）。如前所述，这些文献假设性别在人类的互动和行为中起着重要的作用，但到目前为止，这个假设仍未得到验证（参见Morgenroth & Ryan，2018年，第6页）。我们的实证测试将社会学和心理学对性别的看法进行了桥梁化，并响应了将其视为人性化产生的心理过程的呼吁，而不是作为预测男女差异的自变量（Morgenroth & Ryan，2018）。

These results also have implications for research on (de)humanization. Existing scholarship has identified the target traits and capacities that lead perceivers to attribute mind (Gray et al., 2007; Haslam, 2006), perceivers’ motives for granting and denying humanity (Epley, 2018; Waytz, Morewedge, et al., 2010), and the physical properties and perceptual processes involved (Deska et al., 2017; Fincher et al., 2017; Hugenberg et al., 2016; Khalid et al., 2016). It has also documented that dehumanization and subsequent derogation are more likely directed at out-group members, highlighting that the assignation of social-category membership is often a part of the (de)humanization process. To our knowledge, this is the first paper to show that possessing membership in a social category (gender) is a defining feature of humanity and to consider the consequences of this phenomenon. We add to the literature by showing that, beyond assigning entities uniquely human traits, the process of humanization leads to the assignation of social identities, and gender in particular.
这些结果也对（去）人化的研究具有意义。现有的学术研究已经确定了导致感知者归因于心智的目标特征和能力（Gray et al.， 2007;Haslam， 2006）、感知者给予和否认人类的动机（Epley，2018 年;Waytz、Morewedge 等人，2010 年）以及所涉及的物理特性和感知过程（Deska 等人，2017 年;Fincher等人，2017 年;Hugenberg et al.， 2016;Khalid et al.， 2016）。它还记录了非人化和随后的克减更有可能针对外群体成员，强调社会类别成员的分配通常是（非）人化过程的一部分。据我们所知，这是第一篇表明拥有社会类别（性别）成员身份是人类的一个决定性特征并考虑这一现象后果的论文。我们通过表明，除了为实体分配独特的人类特征之外，人性化过程还导致了社会身份的分配，尤其是性别，从而对文献进行了补充。

It is of note that much of the research on humanization employs manipulations to induce anthropomorphism that include gender, without considering its role or importance. For example, Waytz et al. (2014) induced anthropomorphism by giving a car a name, gender, and human voice. Similarly, Tam et al. (2013) used an article that anthropomorphized nature by using honorifics (Mr. Nature) and personal pronouns (him). Likewise, Kim and McGill (2018) induced anthropomorphism through gendered names (e.g., Jasper). As Aggarwal and McGill (2007) note, “marketers encourage consumers to think of their products in human terms by referring to them with the personal pronouns ‘he’ or ‘she’ instead of ‘it’” (p. 470). Gender is often embedded in, or confounded with, humanization; however, until now, the extent to which gender plays a role in inducing humanness has not been explored. Research on gender and anthropomorphism should consider including and measuring gender, as outcomes measured in this literature (e.g., attachment, treatment, moral regard) may vary as a consequence of which gender is being ascribed (see Hester & Gray, 2020).
值得注意的是，许多关于人性化的研究都采用操纵来诱导包括性别在内的拟人化，而没有考虑其作用或重要性。例如，Waytz 等人（2014 年）通过给汽车起名字、性别和人声来诱导拟人化。同样，Tam et al. （2013） 使用了一篇文章，通过使用敬语 （Mr. Nature） 和人称代词 （him） 将自然拟人化。同样，Kim 和 McGill （2018） 通过性别名称（例如 Jasper）诱导拟人化。正如 Aggarwal 和 McGill （2007） 所指出的那样，“营销人员鼓励消费者用人称代词'他'或'她'而不是'它'来用人性化的术语来考虑他们的产品”（第 470 页）。性别通常嵌入人性化或与人化混淆;然而，直到现在，性别在诱导人类性方面的作用程度尚未得到探索。关于性别和拟人化的研究应考虑包括和测量性别，因为这些文献中测量的结果（例如，依恋、治疗、道德尊重）可能会因性别被归因的结果而变化（参见Hester&Gray，2020）。

This work also offers a new perspective on the causes of dehumanization. A wide variety of cognitive, emotional, motivational, and structural factors are purported to drive dehumanization (see Haslam & Loughnan, 2014, for review). People dehumanize when they experience power (Gwinn et al., 2013), are motivated to protect the in-group (Koval et al., 2012), or feel threatened (Maoz & McCauley, 2008), to highlight just a few reasons. To date, most accounts conceive of dehumanization as resulting from intergroup motives (see Haslam & Loughnan, 2014, for review). Prior research has paid little attention to the role basic social-cognitive forces might play in (de)humanization (see Deska et al., 2017; Fincher et al., 2017 for exception). Here, we demonstrate that understanding a target is so tied up in a gendered process that an inability to attribute gender leads to seeing them as less human.
这项工作还为非人化的原因提供了新的视角。各种各样的认知、情感、动机和结构因素被认为推动了非人化（参见Haslam & Loughnan，2014年，用于审查）。当人们体验到权力（Gwinn等人，2013年），有动力保护内部群体（Koval等人，2012年），或者感到受到威胁（Maoz & McCauley，2008年）时，他们就会失去人性，这只是几个原因。迄今为止，大多数账户认为非人化是由群体间的动机造成的（参见Haslam & Loughnan，2014年，以供审查）。先前的研究很少关注基本的社会认知力量在（去）人化中可能发挥的作用（参见 Deska 等人，2017 年;Fincher et al.， 2017 例外）。在这里，我们证明了理解目标与性别过程如此紧密相连，以至于无法归因性别会导致将他们视为不那么人性化。

These results have implications for the anthropomorphizing of consumer products and technology. Past research has shown that consumers feel more attached, more engaged with, and have greater trust in products, especially technology (e.g., robots, voice assistants), that are seen as having human features (Aggarwal & McGill, 2007; Hart et al., 2013; Stroessner & Benitez, 2019; Waytz et al., 2014). Interestingly, marketers often induce anthropomorphism by giving their products a gender: One need not look further than devices like Siri and Alexa as examples of (female) voice assistants or brands that use gendered mascots, such as “Mr. Peanut” or the “Pillsbury Dough Boy,” to market their products. Whether marketers appreciate the central role that gender plays in anthropomorphism is unknown; however, our results suggest that it is important for these branding tactics to consider gender. As one example, Study 5 demonstrated that people were more likely to humanize a gendered voice assistant, which may have implications for their propensity to take its advice or become attached to it.
这些结果对消费品和技术的拟人化具有影响。过去的研究表明，消费者对产品感到更依恋，更投入，并且对产品有更大的信任，尤其是那些被视为具有人类特征的技术（例如，机器人、语音助手（Aggarwal & McGill，2007;Hart et al.， 2013;Stroessner & Benitez， 2019;Waytz et al.， 2014）。有趣的是，营销人员经常通过给他们的产品赋予性别来诱导拟人化：只需看看 Siri 和 Alexa 等设备，就可以看出（女性）语音助手或使用性别吉祥物（如“花生先生”或“Pillsbury Dough Boy”）来推销他们的产品的品牌。营销人员是否欣赏性别在拟人化中发挥的核心作用尚不清楚;然而，我们的结果表明，这些品牌策略考虑性别很重要。举个例子，研究 5 表明，人们更有可能将性别化的语音助手人性化，这可能对他们接受其建议或依恋它的倾向产生影响。

At the same time, there is growing concern that assigning technology a gender reifies and reinforces gender norms (see Guthrie, 2007; Schiebinger, 2019). Given that seeing something as human appears to entail granting it gender, it may be difficult for marketers to enact genderless technology, as the two are tightly intertwined. It is imperative for future research to explore how to induce anthropomorphism without reinforcing gender norms or beliefs about binary gender.
与此同时，人们越来越担心为技术分配性别会强化和强化性别规范（参见 Guthrie，2007 年;Schiebinger，2019 年）。鉴于将某物视为人类似乎需要赋予它性别，营销人员可能很难实施无性别技术，因为两者紧密交织在一起。未来的研究必须探索如何在不强化性别规范或关于二元性别的信念的情况下诱导拟人化。

Additionally, over the past decade, the topic of de-gendering has captured public attention, and the notion of genderlessness has increased in popularity, both within scholarly research (e.g., Hyde et al., 2019; Lorber, 1994) and in popular media (e.g., National Geographic, 2017; Time Magazine, 2017). Studies 5 through 8 find that to the extent that an entity, person, species, or group is perceived as having a gender, it is also perceived as more human; this result raises questions about the unintended consequences associated with de-gendering (i.e., nongendered targets are seen as less than fully human). By showing that individuals feel less socially connected to a genderless individual, Study 8 speaks to the downstream consequences of removing binary gender from targets: They are seen as less than fully human, which in turn leads perceivers to feel they cannot relate to them. Whether these negative effects persist, given the increasing visibility of nonbinary individuals, remains to be seen; what is clear from the current findings is that, given current cultural norms, genderless individuals are at risk of being dehumanized.
此外，在过去的十年中，去性别化的话题引起了公众的关注，无性别的概念在学术研究中越来越受欢迎（例如，Hyde 等人，2019 年;Lorber，1994 年）以及流行媒体（例如，国家地理，2017 年;时代杂志，2017 年）。研究 5 到 8 发现，在某种程度上，一个实体、个人、物种或群体被认为具有性别，它也被认为更人性化;这一结果引发了人们对与去性别化相关的意外后果的疑问（即，无性别的目标被视为不完全是人类）。通过表明个体与无性别个体的社会联系较少，研究 8 说明了从目标中去除二元性别的下游后果：他们被视为不完全的人，这反过来又导致感知者觉得他们无法与他们产生共鸣。鉴于非二元个体的知名度不断提高，这些负面影响是否持续存在还有待观察;从目前的调查结果中可以清楚地看出，鉴于当前的文化规范，无性别个体面临被非人化的风险。

The present investigation establishes that gender is linked to conceptions of humanness more tightly than are other social categories. This finding raises a vexing question: Why? Consistent with a wealth of evidence, our results suggest gender provides a framework for perceivers to organize their thinking about a target. In its absence, people report a more difficult sensemaking process, which in turn predicts diminished ratings of humanness (see Studies 6 and 7). And while our effect appears to be rooted in a basic cognitive process, it is certainly possible that the relationship between personhood and gender is reified and reinforced via interpersonal and societal forces (Diekman & Schmader, 2021).
目前的调查确定，性别与其他社会类别相比，与人类概念的联系更紧密。这一发现提出了一个令人烦恼的问题：为什么？与大量证据一致，我们的结果表明，性别为感知者提供了一个框架来组织他们对目标的思考。在没有它的情况下，人们报告了一个更困难的意义建构过程，这反过来又预示着对人类的评价会降低（见研究 6 和 7）。虽然我们的效应似乎植根于一个基本的认知过程，但肯定有可能通过人际和社会力量来具体化和强化人格和性别之间的关系（Diekman & Schmader，2021）。

This research sparks many future research questions that are timely and important to explore. First, there is the question of whether the magnitude of the observed effect might diminish with increased societal acceptance of nonbinary gender. If the social-cognitive mechanism established here are rooted in or shaped by exposure and acculturalization processes, one might expect them to diminish with greater acceptance of gender neutrality and de-gendering, and as stereotypes continue to change over time (Eagly et al., 2020). What role acculturation plays in the link between binary gender and seeing human, and how the link might shift as nonbinary gender becomes more normative, are open questions. Our data cannot speak to this possibility, but one might reasonably expect gender and humanness are linked because binary gender is the norm—its absence is disturbing because it is counter-normative—which leaves open the possibility for progress in how gender nonbinary individuals are perceived.
这项研究激发了许多未来研究问题，这些问题值得及时且值得探索。首先，存在一个问题，即观察到的效果的大小是否会随着社会对非二元性别的接受度的提高而减弱。如果这里建立的社会认知机制植根于或由暴露和文化化过程塑造，那么人们可能会预期，随着对性别中立和去性别化的更大接受，以及刻板印象随着时间的推移不断变化，它们会减弱（Eagly et al.， 2020）。文化适应在二元性别和看到人类之间的联系中扮演什么角色，以及随着非二元性别变得更加规范，这种联系会如何变化，这些都是悬而未决的问题。我们的数据无法说明这种可能性，但人们可以合理地预期性别和人性是相互关联的，因为二元性别是常态——它的缺失令人不安，因为它是反规范的——这为如何看待性别非二元个体留下了进步的可能性。

Similarly, the role of learning and individuation is unknown. Our research suggests perceivers struggle to efficiently make sense of genderless targets because they lack a schema to guide their sensemaking. It is very likely that with repeated interaction perceivers develop a schema to guide interactions with genderless and nongendered targets. One could easily test if personal experience with nonbinary individuals moderates the link between gender and perceptions of humanness. Indeed, in subcultures where nonbinary status is visible and valued, the link between seeing human and gender may be greatly attenuated if it exists at all.
同样，学习和个性化的作用也是未知的。我们的研究表明，感知者难以有效地理解无性别目标，因为他们缺乏指导他们意义构建的模式。通过重复交互，感知者很可能会开发出一个图式来指导与无性别和无性别目标的交互。人们可以很容易地测试与非二元个体的个人经历是否调节了性别与人类认知之间的联系。事实上，在非二元身份可见和重视的亚文化中，如果存在的话，看到人类和性别之间的联系可能会大大减弱。

The findings raise several questions about the role of perceiver characteristics in the effect. It may be that characteristics of the perceiver affect the social categories ascribed to humanized entities, suggesting that the link between gender and humanization is more important to some than others. For example, gender schema theory would suggest that one’s identification with masculinity and/or femininity affects how gendered information is processed (Bem, 1981; Markus et al., 1982); thus, gender identification may similarly affect classification and humanization of people and entities. Likewise, race is often a more salient identity to racial minorities than is gender (Gay & Tate, 1998; Levin et al., 2002); thus, the tendency to ascribe gender versus race may be less extreme among racial minority participants. The effect may also be different—more pronounced or more attenuated—among gender-nonconforming individuals.
这些发现提出了几个关于感知者特征在效果中的作用的问题。感知者的特征可能影响了归属于人化实体的社会类别，这表明性别与人化之间的联系对某些人来说比其他人更重要。例如，性别图式理论会表明，一个人对男性气质和/或女性气质的认同会影响性别信息的处理方式（Bem，1981 年;Markus et al.， 1982）;因此，性别认同可能同样影响人和实体的分类和人性化。同样，种族往往比性别更突出地代表少数种族（Gay & Tate， 1998;Levin et al.， 2002）;因此，在少数族裔参与者中，将性别与种族归结的倾向可能不那么极端。这种影响也可能不同——在性别不一致的个体中更明显或更减弱。

It is also important to note that a number of our studies were conducted on Amazon’s MTurk. Although we prevented people from participating in multiple studies with the same experimental design, we did not prevent them from participating in studies that had different paradigms, which may have implications for our studies’ generalizability. While we have no reason to believe that the results depend on the characteristics of this population (Buhrmester et al., 2011; Gosling & Mason, 2015), future research should sample within other populations to ensure these results generalize broadly (Simons et al., 2017).
同样重要的是要注意，我们的许多研究都是在 Amazon 的 MTurk 上进行的。尽管我们阻止人们参与具有相同实验设计的多项研究，但我们并没有阻止他们参与具有不同范式的研究，这可能对我们研究的普遍性产生影响。虽然我们没有理由相信结果取决于这个人群的特征（Buhrmester et al.， 2011;Gosling & Mason，2015），未来的研究应该在其他人群中进行抽样，以确保这些结果广泛地推广（Simons等人，2017）。

Similarly, all of our studies used United States participants, and as such, we cannot speak to the cross-cultural nature of these effects. Gender stereotypes emerge across many cultures and persist over time (Eagly et al., 2020; Valian, 1999); thus, there is reason to believe these findings may emerge cross-culturally. Yet, there is great variability in the magnitude and content of gender differences across cultures (Costa et al., 2001), and as such, it is likely that the gender-humanization link would be attenuated in cultures where gender differences are less pronounced. Examining the gender-humanization link across cultures opens up a fruitful avenue for future research. For example, future research should test whether the effect is as pronounced in countries where the dominant language does not use gendered pronouns. Likewise, the effect may be attenuated in cultures where binary gender is less normative.
同样，我们所有的研究都使用了美国参与者，因此，我们无法谈论这些影响的跨文化性质。性别刻板印象出现在许多文化中，并随着时间的推移而持续存在（Eagly 等人，2020 年;Valian， 1999）;因此，有理由相信这些发现可能是跨文化的。然而，不同文化之间性别差异的大小和内容存在很大差异（Costa et al.， 2001），因此，在性别差异不那么明显的文化中，性别与人性化的联系很可能会减弱。研究跨文化的性别-人性化联系为未来的研究开辟了一条富有成效的途径。例如，未来的研究应该测试这种影响在主导语言不使用性别代词的国家是否如此明显。同样，这种影响在二元性别不太规范的文化中可能会减弱。

Although we provide compelling evidence that gender is central to conceptions of personhood, the question of which gender is more commonly ascribed, what contextual factors shape the ascription process, and the downstream consequences of assigning something “maleness” or “femaleness” remain unclear. An existing literature suggests that, at a default, people represent others as males (Bailey et al., 2019; Cheryan & Markus, 2020; Smith & Zarate, 1992; Van Fleet & Atwater, 1997) or exhibit a bias that entails the “privileging of male experience and the ‘otherizing’ of female experience” (i.e., androcentrism; Bem, 1981, p. 41). For instance, men are considered better examples of people than women and are more often used to represent gender-inclusive categories. Moreover, people more widely generalize information based on men than women (see Bailey et al., 2019, for discussion). We found no consistent evidence of androcentrism in our data—humanization was as likely to lead to “seeing male” as it was to “seeing female”; however, these studies were not designed to test for androcentrism in this process. Future research might consider doing so or articulate when we assign one gender over the other and the downstream implications of this outcome.
尽管我们提供了令人信服的证据，证明性别是人格概念的核心，但哪种性别更常见、哪些背景因素塑造了归类过程以及分配“男性”或“女性”的下游后果的问题仍不清楚。现有文献表明，默认情况下，人们将他人代表为男性（Bailey et al.， 2019;Cheryan & Markus， 2020;Smith & Zarate， 1992;Van Fleet & Atwater，1997年）或表现出一种偏见，导致“男性经验的特权化和女性经验的'他者化'”（即男性中心主义;Bem，1981 年，第 41 页）。例如，男性被认为是比女性更好的人的例子，并且更常被用来代表性别包容性的类别。此外，人们比女性更广泛地基于男性泛化信息（参见 Bailey et al.， 2019， for discussion）。我们在数据中没有发现雄心主义的一致证据——人性化导致 “看到男性 ”的可能性与导致 “看到女性 ”的可能性一样大;然而，这些研究并非旨在测试这一过程中的雄心主义。未来的研究可能会考虑这样做，或者阐明我们何时将一种性别置于另一种性别之上，以及这一结果的下游影响。

Finally, our studies examined the role of (binary) gender in perceptions of humanness and inferences in its absence. The findings do not speak to the experience of humanness among people who identify as nonbinary, gender fluid, etc. By no means do we intend to suggest that gender is required for self-humanization, though this question opens up an interesting avenue for future research to explore. For example, it is likely that those who view themselves in nonbinary ways have a clear schema through which they understand themselves—yet, our theory would suggest that it is one that is less culturally accessible, supported by norms, and reinforced by others. Thus, it is important to understand the experiences of nonbinary, gender fluid, and a-gendered people, as these insights can help advance theory about when breaking out of a binary schema is possible and, further, when it will be valued. Indeed, while breaking (out of) the binary is likely to be difficult, as it disrupts an established schema through which people understand themselves and others, such costs may be outweighed by the greater feelings of authenticity that transcending the gender binary brings. Further, it is possible that as cultures or local norms change to de-emphasize gender, schema accessibility may open up to more people, including those who neither identify as a man nor a woman. Our findings only allow us to speculate on these ideas, but they make clear that more research is needed to understand the role of gender in humanization, especially as gender identities continue to expand beyond the binary.
最后，我们的研究考察了（二元）性别在对人类的看法中的作用以及在人类不存在时进行的推断。这些发现并没有说明非二元性别、性别流动等人的人类体验。我们绝不打算说性别是自我人化所必需的，尽管这个问题为未来的研究探索开辟了一条有趣的途径。例如，那些以非二元方式看待自己的人很可能有一个清晰的图式来理解自己——然而，我们的理论会表明，这是一个在文化上不太容易接近的、由规范支持并得到他人强化的模式。因此，了解非二元、性别流动和无性别者的经历很重要，因为这些见解可以帮助推进关于何时可以打破二元模式以及何时会受到重视的理论。事实上，虽然打破（打破）二元论可能很困难，因为它破坏了人们理解自己和他人的既定模式，但这种成本可能会被超越性别二元论带来的更大的真实感所抵消。此外，随着文化或当地规范的变化以淡化性别，图式可访问性可能会向更多人开放，包括那些既不认同男性也不认同女性的人。我们的研究结果只能让我们对这些想法进行推测，但它们清楚地表明，需要更多的研究来了解性别在人性化中的作用，尤其是在性别认同不断超越二元的情况下。

In both research and public discourse, there has been an increased push toward genderlessness. The current research suggests that, more than any other social category, gender is a primary component in ascribing humanity. Thus, while genderlessness presents societal opportunities, it may also present unforeseen challenges. Gender may be not simply a cue for grouping humans into categories, but one that informs who is seen as “human” in the first place.
在研究和公共话语中，无性别主义的推动力度越来越大。目前的研究表明，与任何其他社会类别相比，性别是赋予人性的主要组成部分。因此，虽然无性别带来了社会机遇，但它也可能带来不可预见的挑战。性别可能不仅仅是将人类分为几类的提示，而是首先告知谁被视为“人类”。

附录 A

Instructions Used in Studies 1–3
研究 1-3 中使用的说明

The definition of anthropomorphism that participants were provided within Studies 1 and 3 was taken from Waytz, Morewedge, et al. (2010) and read:
研究 1 和 3 中提供的参与者拟人化的定义取自 Waytz、Morewedge 等人（2010 年），内容如下：

Anthropomorphism is the process of ascribing uniquely human qualities to non-human things, animals, or events. Anthropomorphism goes beyond behavioral descriptions or observations (e.g., the object is dark); it entails attributing humanness—such as conscious experience and intentions—to non-human objects, animals, events, or entities (e.g., the object is angry). It can include giving mental capacities that are uniquely human, like having conscious awareness (e.g., an understanding of the self) and possessing explicit intentions or desires (e.g., revenge, ambition). It can also include experience uniquely human emotions (e.g., pride, shame, guilt). Some examples include (a) giving a non-human thing a personality; (b) seeing faces or other human characteristics in objects; (c) feeling love, anger, or other emotions towards an entity (e.g., guilt for breaking it, angry when it doesn’t work); (d) assuming the entity can either react or feel emotions towards you (e.g., the object is happy); (e) assuming the object has intentions (e.g., talking to it, believing it is “acting up”).
拟人化是将独特的人类品质赋予非人类事物、动物或事件的过程。拟人化超越了行为描述或观察（例如，物体是黑暗的）;它需要将人性——例如有意识的体验和意图——归因于非人类的物体、动物、事件或实体（例如，物体是愤怒的）。它可以包括赋予人类独有的心理能力，例如拥有有意识的意识（例如，对自我的理解）和拥有明确的意图或欲望（例如，复仇、野心）。它还可以包括体验独特的人类情感（例如，骄傲、羞愧、内疚）。一些例子包括 （a） 赋予非人类事物个性;（b） 在物体中看到面孔或其他人类特征;（c） 对某个实体感到爱、愤怒或其他情绪（例如，破坏它的内疚、当它不起作用时感到愤怒）;（d） 假设该实体可以对你做出反应或感受到情绪（例如，该对象很快乐）;（e） 假设对象有意图（例如，与它交谈，相信它在“行动”）。

The definition of anthropomorphism participants used in Study 2b read:
研究 2b 中使用的拟人化参与者定义如下：

We would like you to anthropomorphize the shapes in the video below. That is, try to imagine this as a scene with human agents. Anthropomorphism is the process of ascribing uniquely human qualities to non-human things or events. It goes beyond behavioral descriptions or observations (e.g., the object is dark); it entails attributing humanness—such as conscious experience and intentions—to non-human objects or entities (e.g., the object is angry). It can include giving mental capacities that are uniquely human, like having conscious awareness (e.g., an understanding of the self) and possessing explicit intentions or desires (e.g., revenge, ambition). It can also include experiencing uniquely human emotions (e.g., pride, shame, guilt). Please think about each “person” in the video: what they were like, and how they behaved.
我们希望您将下面视频中的形状拟人化。也就是说，试着把它想象成一个有人类代理的场景。拟人化是将独特的人类品质赋予非人类事物或事件的过程。它超越了行为描述或观察（例如，物体是暗的）;它需要将人性——例如有意识的体验和意图——归因于非人类的物体或实体（例如，物体是愤怒的）。它可以包括赋予人类独有的心理能力，例如拥有有意识的意识（例如，对自我的理解）和拥有明确的意图或欲望（例如，复仇、野心）。它还可能包括体验独特的人类情感（例如，骄傲、羞愧、内疚）。请想想视频中的每个“人”：他们是什么样的，他们的行为举止。

The definition of description participants used in Study 2b read:
研究 2b 中使用的描述参与者定义如下：

We would like you to describe the shapes in the video below. That is, try to remember qualities, movements, and features of the video. Description is the process of recalling and listing information about objects, movement, and imagery. We would like you to focus on descriptions or observations (e.g., there were multiple objects, the objects were blue), movement (e.g., jagged movement, spinning), quality (e.g., the video was high resolution), and impressions (e.g., the video was interesting). You will be asked to describe your impression of the video: what you observed and whether you enjoyed it.
我们希望您在下面的视频中描述这些形状。也就是说，尽量记住视频的质量、动作和特征。描述是调用和列出有关对象、运动和图像的信息的过程。我们希望您专注于描述或观察（例如，有多个对象，对象是蓝色的）、运动（例如，锯齿状运动、旋转）、质量（例如，视频是高分辨率的）和印象（例如，视频很有趣）。您将被要求描述您对视频的印象：您观察到了什么以及您是否喜欢它。

The instructions given in Studies 3a and 3b read:
研究 3a 和 3b 中给出的说明如下：

Please take a minute to anthropomorphize this product. You may give it a name and/or personality. Think about how you would you would interact with the product, what human characteristics it might have, how would it “behave,” and what emotions would you feel towards it. Please write 5–6 sentences.
请花点时间将此产品拟人化。您可以给它起个名字和/或个性。想想你会如何与产品互动，它可能具有哪些人类特征，它会如何“行为”，以及你会对它有什么情绪。请写 5-6 个句子。