A People-as-Means Approach to Interpersonal Relationships

Interpersonal relationships and goal pursuit are intimately interconnected. In the present article, we present a people-as-means perspective on relationships. According to this perspective, people serve as means to goals—helping other people to reach their goals in a variety of ways, such as by contributing their time; lending their knowledge, skills, and resources; and providing emotional support and encouragement. Because people serve as means to goals, we propose that considering relationship processes in terms of the principles of goal pursuit can provide novel and important insights into the ways that people think, feel, and behave in these interpersonal contexts. We describe the principles of means-goals relations, review evidence for each principle involving people as means, and discuss implications of our approach for relationship formation, maintenance, and dissolution.

From birth through adulthood, interpersonal relationships are shaped by the way partners support each other's goal pursuit. Infants form social bonds with caregivers who meet their needs for food, comfort, and safety. Adolescents develop social ties with their peer group, teachers, and coaches who provide companionship, guidance, and advice. Adults establish social connections with romantic partners and friends who support them as they pursue personal and interpersonal goals, and they develop relationships with professors, colleagues, and mentors who facilitate academic and professional goals, among others. Throughout the life span, then, people construct relationships with others who facilitate their progress toward goals. By facilitating progress toward goals, relationship partners serve as means to those goals. What implications does recognizing that partners serve as means to goals have for evaluations of those partners, and for relationship processes? In the pages that follow, we present a people-as-means approach to interpersonal relationships, summarize evidence linking principles of goal pursuit to instances in which people are means, and discuss the novel insights gained from adopting a people-as-means perspective on relationships.

Theoretical Overview
Goal systems theory  serves as the foundation from which our perspective grows. It outlines the principles of means-goal relations and is assumed to apply to all goal contents and all means by which they are pursued. Goal systems theory was not developed to explain close relationships and did not originally consider people as means to goals. However, much research has implicitly or explicitly investigated the implications of considering people as means to goals.
Goal systems theory can be summarized according to eight principles. Here, we outline the principles and consider how they apply to cases in which people serve as means.

Principle 1: Top-down activation of means by goals
When a person has an important goal in mind, he or she directs attention toward means of attainment so that goal pursuit can be initiated. Means that have proven instrumental in past pursuits are most likely to be brought to mind and used in the future (Aarts & Dijksterhuis, 2000;Custers, Eitam, & Bargh, 2012;Danner, Aarts, & de Vries, 2007;Sheeran et al., 2005). Thus, as people move through 744522P PSXXX10.1177 their various pursuits during a day, they should think of and attend to people who are useful to their current concerns. For example, reminders of sociability and achievement goals (compared with presentation of goalneutral information) brought to mind people who are instrumental to that goal (Fitzsimons & Shah, 2008). The goal to have sex increases attention directed toward potential sexual partners (Maner, Gailliot, Rouby, & Miller, 2007). Threats of failure, death, or separation lead people to bring to mind attachment figures who serve as support providers (Cox et al., 2008;Mikulincer, Gillath, & Shaver, 2002). Thus, as people pursue their goals, they mentally summon the people who are integral to goal attainment.

Principle 2: Affective transfer
Means are evaluated according to their perceived instrumentality to the goals of the moment (Fishbach, Shah, & Kruglanski, 2004). When partners serve as means to a person's goals, the person evaluates those instrumental partners positively. For example, people report wanting to spend more time with potential new friends who are instrumental (compared with noninstrumental) to their current goal (e.g., an academic achievement or fitness goal; Slotter & Gardner, 2011). People feel interpersonally closer to an instrumental partner than a noninstrumental partner (Fitzsimons & Fishbach, 2010;Fitzsimons & Shah, 2008;Gomillion & Murray, 2014). People feel more gratitude and closeness to a helper when the task with which the helper is assisting is still in progress, compared with when the task has been completed (i.e., gratitude is greater when the helper is still instrumental; Converse & Fishbach, 2012;Fitzsimons & Fishbach, 2010). Thus, partners are evaluated more positively when they are instrumental to current goals.

Principle 3: Bottom-up activation of goals by means
Available means signal an opportunity for goal pursuit and bring to mind associated goals (Shah, 2005;Shah, Hall, & Leander, 2009;Shah & Kruglanski, 2003). The real or imagined presence of a person should therefore bring to mind associated goals. Indeed, exposure to a role model can increase hard work and inspiration for shared goals (Aarts, Gollwitzer, & Hassin, 2004;Lockwood & Kunda, 1997). A reminder of a partner increases commitment to associated goals (Shah, 2003), adoption of associated goal-pursuit strategies (Orehek, Vazeou-Nieuwenhuis, Quick, & Weaverling, 2017), motivation to engage in goal-relevant behaviors (Leander, Shah, & Chartrand, 2009), persistence in pursuit of associated goals (Shah, 2003;vanDellen & Hoyle, 2010), and better performance on goal-relevant tasks (Fitzsimons & Bargh, 2003). Just as goals can instigate a search for a person to serve as a means, the presence of a person who serves as a means can initiate goal pursuit afforded by the person.

Principle 4: Multifinality
Some means-referred to as multifinal means-can satisfy multiple goals simultaneously (Orehek & Vazeou-Nieuwenhuis, 2013). For example, eating fruit is multifinal when it serves as a means to both pleasant taste and health (Orehek & Vazeou-Nieuwenhuis, 2016). Multifinal means are evaluated more positively than unifinal means, which serve only one goal, because they can satisfy many goals, thus generating more value (Chun, Kruglanski, Sleeth-Keppler, & Friedman, 2011;Kruglanski et al., 2013;Orehek, Mauro, Kruglanski, & van der Bles, 2012). However, multifinal means are not always easy to find and adopt. As the number of goals increases, the set of possible means that can satisfy all of them is reduced (Köpetz, Faber, Fishbach, & Kruglanski, 2011). For example, finding a delicious meal is easy, but finding a delicious, healthy, convenient, and inexpensive meal can be difficult (Orehek & Vazeou-Nieuwenhuis, 2016). Partners who serve multiple goals should be evaluated more positively than partners who serve only one goal. However, partners who serve multiple goals should also be less commonthat is, more difficult to find-than partners who serve just a few goals. We have found in several studies that the more goals a person serves, the more favorably that person is evaluated on measures of interpersonal closeness, perceived social support, conversation quality, and perceived responsiveness (Orehek, Forest, & Wingrove, in press). In another study, multifinal job applicants were rated more favorably when evaluators held each of the goals the applicant served (Chun et al., 2011). In addition, we found that partners who are multifinal are less common in people's existing social networks (Orehek et al., in press). This suggests that partners who are multifinal are less common in the population, and thus should be difficult to find. However, such people should also be uniquely appreciated.

Principle 5: Dilution
Multifinal means have the benefit of being able to satisfy multiple goals simultaneously, but they come with the cost of decreased perceived instrumentality toward each goal (Zhang, Fishbach, & Kruglanski, 2007; see also Orehek et al., 2012). This dilution effect occurs when a means serves multiple goals that are dissimilar or incompatible with one another, but the effect is reduced or eliminated when a means serves multiple goals that are similar or compatible with one another (Zhang et al., 2007). Thus, the more goals a partner serves, the less instrumental he or she may be considered for each of those goals-at least when those goals are incompatible (Light & Fitzsimons, 2014). Preliminary evidence supports the notion that dilution can occur when people serve as means and that similarity and complementarity of actions reduces the dilution effect (Myers, O'Brien, Balota, & Toyofuku, 1984;Park, 1989;Smith, Adams, & Schorr, 1978). Note that the conditions that foster the dilution effects are often absent when evaluating close relationship partners. The more a person gets to know a partner, the more likely it is that the person will perceive the partner's actions and goals as complementary. However, dilution may be more likely in early stages of impression formation or when abstract representations (such as online-dating profiles or job résumés) guide evaluations (see Eastwick, Finkel, & Eagly, 2011). Contrary to our expectations, some writers have suggested that dilution may occur frequently in close relationships (Finkel, Hui, Carswell, & Larson, 2014;Light & Fitzsimons, 2014). Research is needed to investigate these possibilities.

Principle 6: Resource pulling
People often have a number of goals and tasks in mind, and when they are pursuing a focal task, they can become distracted when alternative tasks pull attention away from and reduce commitment to that task. Whether an alternative task pulls resources from the focal task is determined by compatibility of the two tasks. When an incompatible second task captures a person's attention, resources are pulled away, impeding pursuit of the focal task . When people serve as means, a similar process would occur if an alternative partner pulls attention and commitment away from a current partner. In one illustration of the pulling effect, the longer people in committed romantic relationships attended to pictures of attractive alternatives, the less satisfied, less committed, and less close they felt to their relationship partner (R. S. Miller, 1997). Attention to alternatives also predicted relationship termination 8 weeks later. In another study, romantically attached individuals' satisfaction with their current relationship decreased after exposure to a high-quality alternative (compared with a low-quality alternative; Kenrick, Neuberg, Zierk, & Krones, 1994). Participants in another study who received feedback from an attractive alternative partner stating that they had potential as a dating partner felt less relationship commitment and satisfaction following the feedback (Kavanagh, Fletcher, & Ellis, 2014). In contrast, participants who received feedback stating that they did not have potential as a dating partner were more committed and satisfied with their current partner following the feedback. The presence of alternative partners, or the perception that alternative partners are available, is a predictor of divorce (South & Lloyd, 1995;South, Trent, & Shen, 2001;Udry, 1981). Thus, the presence of alternative partners may produce decreased commitment to present partners, increasing the likelihood of romantic relationship dissolution.

Principle 7: Shielding
Resource pulling can be reduced by inhibiting thoughts of alternative tasks during focal-task pursuit. Inhibition of alternatives, or shielding, occurs when one's level of commitment to the focal task is stronger than commitment to the alternatives . In close relationships, the shielding principle translates into what we call relationship shielding. When a person is strongly committed to a romantic partner, then alternative partners are devalued and attention is directed away from them, thus protecting, or shielding, the current relationship from threats. People in a romantic relationship rate alternative partners as less attractive than do participants who are single (Karremans, Dotsch, & Corneille, 2011;S. L. Miller & Maner, 2010;Simpson, Gangestad, & Lerma, 1990). This effect is especially pronounced for people who are strongly satisfied with or committed to their current partner (relative to people who are weakly satisfied with their current partner; Cole, Trope, & Balcetis, 2016;Johnson & Rusbult, 1989;Lydon, Fitzsimons, & Naidoo, 2003;Lydon, Meana, Sepinwall, Richards, & Mayman, 1999; see also Meyer, Berkman, Karremans, & Lieberman, 2011). Another series of studies has demonstrated that people who are committed to their partner direct attention away from attractive alternatives (Linardatos & Lydon, 2011;Maner, Gailliot, & Miller, 2009).
Another strategy for shielding one's relationship from potential threats would be to make oneself seem less desirable to potentially interested others. Engaging in risky behavior is one way in which men try to stand out to potential partners. After interacting with an available woman (compared with interacting with an unavailable woman), single men displayed more risktaking behavior if the available woman was going to see their performance (Baker & Maner, 2009). No effects were found when the performance was private. This pattern of results was replicated in another study, which also found that committed men became less risky under the same conditions in which single men became riskier (Frankenhuis & Karremans, 2012), suggesting that men in committed romantic relationships protect their relationship by appearing less attractive to potential partners. Recent research has found that women whose commitment to a relationship partner is low tend to dress more attractively when they expect to meet a single man than when they expect to meet an unavailable man, but women committed to their relationships do not show this effect (Forest, Krueger, & Orehek, 2017). These studies show that people engage in a variety of relationship-shielding tactics: They devalue and direct their attention away from attractive alternatives and make themselves appear less desirable to alternative partners.

Principle 8: Means substitution
More than one means is often available to attain a goal (Kruglanski, Pierro, & Sheveland, 2011). Means are substitutable for one another when they are each instrumental ways of attaining a goal Lewin, 1935). The strength of a person's preference for one means over another depends on the perceived instrumentality of each for the goal. In romantic relationships, partner substitution occurs when one permanently leaves a current partner for an alternative partner (i.e., mate switching; Buss, Goetz, Duntley, Asao, & Conroy-Beam, 2017). Forty-three percent of men and 49% of women reported having engaged in such permanent romantic partner substitution at least once (Schmitt & Buss, 2001). In cases in which a person does not leave the original relationship, but merely establishes a new short-term relationship, complete partner substitution would not be said to occur. The conditions under which partner substitution occurs should depend on the suitability of one's current partner compared with the suitability of alternative partners one can attract. Partner substitution should be more likely when one's current partner decreases in value and/or when a partner of higher value is available.
Consistent with this principle, people in romantic relationships are more satisfied with their relationships when their current partner is perceived to be of higher quality than alternative partners in the local mating pool (Conroy-Beam, Goetz, & Buss, 2016). As one's current partner decreases in perceived instrumentality, the likelihood of relationship dissolution increases (Fitzsimons, Finkel, & vanDellen, 2015;Weisfeld & Weisfeld, 2002). Infertility in relationships is associated with diminished relationship satisfaction (Leiblum, Aviv, & Hamer, 1998;Millheiser et al., 2010) and is predictive of divorce (Kjaer et al., 2014). Our people-as-means approach suggests one possible explanation for such effects: When two people would like to have children and they are not successful in doing so, it is likely to lead them to perceive each other as noninstrumental in an important domain. When a person perceives a partner as less instrumental than he or she did in the past, or when alternatives who are more instrumental than the partner are available, then partner substitution becomes more likely.

Summary
As shown in Figure 1, evidence involving instances in which people serve as means to goals provides strong support for most of the principles of goal systems theory, which were developed considering objects or activities as means (not people). The principle with the greatest need for additional investigation is Principle 5, the dilution principle. Understanding whether, when, and why dilution operates (or does not) in contexts in which people serve as means to goals represents an important and exciting area for future research. Our people-as-means framework suggests that people evaluate others according to their perceived instrumentality, form and maintain relationships with partners they perceive to be instrumental to important goals, and dissolve relationships when a partner is no longer instrumental or when a more instrumental alternative is available. In addition, people find relationships most fulfilling or rewarding when they perceive their partner as instrumental. Principles 1 to 5 have been investigated among a variety of relationship types, including romantic relationships, friendships, and acquaintances. Principles 6 to 8 have mostly been investigated among romantic relationships. We expect these processes to extend to other relationship types. For example, a person may shield a relationship with a babysitter so that he or she is available when needed, may feel less close and committed to that babysitter if a new one becomes available, and may replace the babysitter with the new one if the new one proves more reliable. We next turn our attention to consideration of the person serving as the means, and we propose a dyadic approach to people as means, highlighting implications that go beyond what can be gleaned from goal systems theory.

Mutual Perceived Instrumentality
One important way that people as means differ from the kinds of means-objects or activities-that have been traditionally studied in the goals literature is that people have their own minds. Thus, whereas previous work involving means unrelated to people has ignored the implications of goal-pursuit principles for the means-which makes sense-turning our focus to people as means introduces the need and exciting opportunity to consider the implications of these principles When an alternative means is more instrumental than a current means, and the alternative means replaces the original When an alternative partner is more instrumental than a current partner, and the alternative partner replaces the original When a new romantic partner is perceived as more instrumental than the original partner, and the person dissolves the original relationship in favor of the new relationship for the person serving as the means. We suggest that people often like to serve as means to goals. They derive a sense of personal importance, esteem, competence, control, value, and belongingness from their ability to help others (Inagaki & Orehek, 2017;Orehek, 2018;. Thus, people evaluate themselves in the same way they evaluate others: They evaluate themselves according to their perceived instrumentality to others. People should be motivated to form and maintain relationships with partners they can be instrumental toward and to dissolve relationships when they can no longer be instrumental to their partner or when they find an alternative to whom they can be more instrumental. In addition, people should find relationships most fulfilling or rewarding when they perceive themselves as instrumental to their partner. These possibilities have not yet been empirically investigated, and warrant investigation. When considering relationship partners as means to goals, it is important to take into account both partners' contributions and perspectives and the processes that unfold across time. We propose that relationships are likely to thrive when partners experience mutual perceived instrumentality-when each person perceives the partner to be instrumental to his or her important goals and perceives him-or herself as instrumental to his or her partner's important goals (Orehek & Forest, 2016). When such opportunities are absent, people may feel under benefited, unimportant, or unfulfilled. This emphasis on mutual perceived instrumentality leads to a number of novel predictions, and we present these in the remaining sections of the article.
We postulate that the value gained from serving as a means and the value gained from having another person serve as a means is additive. Thus, both having the opportunity to serve as a means for a partner and having a partner who is instrumental should contribute to strong relationship satisfaction. Having a partner with whom one is unable to be instrumental or who is not instrumental to one's own goals should reduce relationship satisfaction.
One might expect that (a) relationships are maximally satisfying when the instrumentality that one gives to and receives from a partner are balanced and (b) that imbalances in instrumentality would result in conflict and dissatisfaction in one or both members of the dyad (Hatfield & Rapson, 2011;Hatfield, Walster, & Berscheid, 1978). However, we postulate that apparent imbalances in instrumentality will not always be problematic, for at least three reasons. First, there are times when a relationship may appear, at first glance, to be imbalanced, but in fact, mutual perceived instrumentality exists. For example, it may be obvious that a parent feeds, clothes, and cares for his or her child. It may be less obvious that the child provides the parent with a sense of social connection and generativity. Second, in close relationships, partners often track each other's needs even when there is no opportunity for reciprocation (Clark, Mills, & Powell, 1986) and tend not to track each other's inputs into joint tasks (Clark, 1984;Clark, Mills, & Corcoran, 1989). Thus, we suspect that temporary imbalances in instrumentality may sometimes go unnoticed in close relationships and, if noticed, may be tolerated insofar as the person who is disproportionately serving as a means feels that the partner would be concerned and responsive (and willing to serve as a means) if the need arose (Clark & Aragon, 2014). Third, even if partners have some awareness of an imbalance, the balance of instrumentality between relationship partners should be less important than the total value derived from having a partner serve as a means to one's goals and from serving as a means to that partner's goals; to the extent that a person gains utility through either or both of these avenues, he or she should feel positively about the relationship.
Adopting a people-as-means perspective on relationships has important implications for understanding a variety of relationship processes. We outline these implications in the sections that follow and highlight some of the novel predictions that our perspective enables us to generate. First, we consider the perspective of the person serving as the means.
We postulate that serving as an instrumental means is just as important as having instrumental others. Being instrumental to others should increase both the closeness one feels toward the person whom one has helped and one's own sense of self-worth. Serving as an instrumental means to other people provides benefits to the support provider and enhances the relationship between the support provider and recipient (Inagaki & Orehek, 2017). For example, participants who performed a favor for someone they previously had a negative or neutral impression of came to like that person more ( Jecker & Landy, 1969). Giving affection to a romantic partner is associated with commitment to the relationship (Horan & Booth-Butterfield, 2010). Giving support to others increases self-esteem (Piferi & Lawler, 2006), self-worth (Gruenewald, Liao, & Seeman, 2012), and the social connection felt with that person (Inagaki & Eisenberger, 2012). People who help others benefit most if they experience the choice to help as autonomous (Weinstein & Ryan, 2010; see also Williamson & Clark, 1989, 1992. Thus, serving as an instrumental means is important to people: They want to feel valuable, useful, and helpful. Doing so allows them to feel close to their relationship partners and to experience a heightened sense of self-worth. Because serving as a means to others offers these benefits, people should be especially likely to offer to serve as a means to others or to go out of their way to do so when these needs are heightened. For example, if someone loses his or her job or his or her children move away, the reduced sense of being needed and useful that may accompany such events could lead people to seek out ways to serve others' goals. Of course, people do not always reap the potential benefits of serving as means to others. Our approach postulates that two conditions must be met for the person to experience the benefits of serving as a means (see also Inagaki & Orehek, 2017;Orehek, 2018;. First, serving as a means should benefit the helper if he or she feels instrumental as a result. This can occur by observing progress or improvement in the goal pursuer's situation, such as when a child grows or develops or when the goal pursuer expresses gratitude for one's help. Support efforts in which improvements are not made, such as when an ill person deteriorates or symptoms are merely maintained despite a person's devoted caretaking efforts, should be frustrating and should not lead to increased feelings of self-worth in the caretaker. People seem to adjust their behavior accordingly. For example, people are disinclined to provide support to partners who have a track record of consistent negativity-at least in part because they feel that their support efforts are unlikely to be instrumental in creating change (Forest, Kille, Wood, & Holmes, 2014). Our approach postulates that experiences of burnout or fatigue are due to feelings of noninstrumentality rather than workload per se.
Second, the person serving as a means must feel that he or she has freely chosen to adopt the role of means and therefore wants to serve as a means to the goal. Outcomes are likely to be suboptimal if the person feels coerced or serves as a means to a goal that he or she does not wish to serve (Inagaki & Orehek, 2017;Orehek, 2018;. Research has confirmed that people experience a boost in mood when helping others when that help is freely chosen, but not when it feels as if they did not have a choice (Weinstein & Ryan, 2010). Likewise, unwanted praise or compliments can leave a person feeling uncomfortable or upset because the person delivering the positive message is placing the recipient in the position of serving as a means to goals they do not want to serve. When a person does not want to be a means to a particular goal, but is treated as a means to that goal anyway, that person is likely to feel objectified and to experience the negative consequences of objectification .
Thus, a people-as-means approach postulates that to keep one's relationship partners happy and fulfilled, one should give those partners the opportunity to be a means to one's goals. We suspect that people realize that their partners like it when they serve as means to their partner's goals (e.g., help their partners out, do nice things for their partners), but it is probably less obvious that people should also give their partners a chance to do things for them. This should enable the partners to feel useful, valuable, and efficacious. To the extent that people acknowledge the help they have received or express gratitude for it, the benefits for the instrumental partner and for the relationship may be heightened (e.g., Algoe, Gable, & Maisel, 2010;Algoe, Haidt, & Gable, 2008;Kubacka, Finkenauer, Rusbult, & Keijsers, 2011), at least in part because such acknowledgment signals that the help was instrumental.
In addition, just as people evaluate partners more positively when those partners serve many goals (Principle 4), our approach postulates that people experience greater feelings of self-worth and social connection when they themselves serve as a means to multiple goals for partners. Serving as a means to multiple goals should (a) make people feel that their partners are more dependent on them, thus enhancing feelings of security (e.g., Murray, Aloni et al., 2009); (b) make people confident that, even if they lose the ability to be instrumental in one domain or if their partner no longer pursues a particular kind of goal, they will still be able to offer their partner other things; and (c) make it harder for the partner to find an alternative partner who could do all of the things that they do for the partner.

Implications for Interpersonal Relationships
Our people-as-means account has important implications for understanding relationship dynamics. Below we examine some of these implications at different stages of relationships, including relationship initiation, maintenance, dissolution, and loss.

Relationship initiation and evaluation
The evidence we have reviewed supports the ideas that people perceived as high-quality means to a goal are attended to, evaluated positively, and readily approached while the goal to which they are instrumental is being pursued (Principles 1 and 3). Whether evaluating a job candidate via her résumé, a romantic prospect on a dating website, or one's best friend, people evaluate others according to their perceived instrumentality to current or future goals. Our people-as-means approach postulates that central relationship evaluations, such as liking, relationship satisfaction, relationship commitment, interpersonal closeness, perceived support, and perceived responsiveness should be derived from perceived instrumentality of the person being evaluated.
Our approach postulates that the number of goals a partner serves should also be associated with people's evaluations of that partner: Partners who are instrumental to multiple goals (i.e., are multifinal) should be evaluated more positively than those who serve fewer goals because they can facilitate more goal pursuit (Principle 4). Thus, we expect that people will find others who are instrumental to a greater number of goals (rather than fewer goals) to be more supportive, responsive, and likeable, and to feel interpersonally closer to them. However, such multifinal partners should be more difficult to find than unifinal, or task-specific, partners. It may be easy to find another person with whom you enjoy playing tennis. There are far fewer people with whom you enjoy playing tennis, studying for exams, attending concerts, and going shopping for clothes. The scarcity of such multifinal partners may make them all the more desirable and appreciated.
From our people-as-means perspective, one reason why partner responsiveness is so important (Reis, Clark, & Holmes, 2004;Reis & Shaver, 1988) is that it serves as a signal that a partner is likely to be willing and available to be a means to many goals that one might pursue. Responsive partners should be better at tracking goal shifts and needs across time, and they should also be more motivated to meet those needs. Likewise, partners who are perceived as supportive are expected to be available to support future pursuits, whatever they may be (e.g., Lakey & Orehek, 2011). Partners who are instrumental to a variety of goals are especially valuable because they are also the people who are likely to adapt in a way that they are instrumental to future unknown goals. Thus, responsiveness and the motivation to be supportive are characteristics of relationship partners that are likely to produce sustained instrumentality over time. Although having a particular skill or expertise in a goal domain can help someone be instrumental, such skill is not sufficient in the absence of motivation to help the other person. In addition, a person can be instrumental to a goal in many ways. If a person's goal is running a marathon, having a partner who is an expert marathoner and can accompany him or her on training runs could be an asset, but a nonexpert partner could do many other things to facilitate pursuit of this goal (e.g., providing childcare so there is time to train, laundering running clothes, helping learn the course topography, or providing support and encouragement on race day). Responsive and supportive partners may have the will, or the motivation, to help their partner (Winczewski, Bowen, & Collins, 2016) and will therefore find a way to do so.
However, one intriguing possibility suggested by our approach is that the perceived instrumentality of a relationship partner in relation to one goal may become diluted if that partner serves as a means to additional goals (Principle 5). If this happens, then it suggests that the global high evaluation of a multifinal partner comes at the expense of perceiving the partner as less instrumental to each particular goal that he or she serves. For example, consider a case in which a romantic partner is first merely perceived as an instrumental means to enjoyable nights out. Over time, that partner demonstrates that he or she is instrumental for contributing to household tasks, for having enjoyable intellectual conversations, for helping with one's finances, and eventually for being a skilled and caring coparent. Although the multifinality of this partner should increase one's felt closeness to him or her (along with perceived support and other positive evaluations), the dilution principle suggests that one's perception of the partner as a fun date may decrease. Thus, there may be a trade-off between perceptions of overall instrumentality (and associated relational outcomes) and instrumentality toward any particular goal. If this happens, then the person may stop looking to the partner to fulfill the diluted goal and/or may turn to alternative partners to fulfill that goal.
Yet, as reviewed earlier, each of the factors that mitigate the dilution effect tend to be present in cases in which people serve as means. Because the goals people serve tend to be compatible with one another, we do not expect dilution to be common among well-known others. In the example above, a partner who is a good coparent can make date night more fun by having enjoyable conversations about their children. However, situations in which the partner is not well known, such as when one learns about a partner on an online-dating website, may be more likely to produce dilution because the compatibility of the various goals may be less obvious. Research has demonstrated that early impressions from online-dating sites, speed dating, and other first encounters can differ markedly from impressions formed over time (e.g., Hunt, Eastwick, & Finkel, 2015). An interesting possibility is that dilution may represent a form of evaluation error. If people in lasting relationships do not experience dilution, but people forming first impressions perceive dilution, then early impressions may be overly pessimistic with respect to a partner's ability to serve many goals. These possibilities have important implications for the utility of platforms that attempt to aid in match-making, such as online-dating websites (Finkel, Eastwick, Karney, Reis, & Sprecher, 2012).
Although instrumental others are appreciated for the support they provide, an important caveat is that such appreciation may be fleeting. Once the goal has been attained and the actor has moved on to other pursuits, appreciation for the instrumental partner tends to wane (Converse & Fishbach, 2012;Fitzsimons & Fishbach, 2010;Fitzsimons, Friesen, Orehek, & Kruglanski, 2009;Fitzsimons & Shah, 2008). One challenge for maintaining interpersonal closeness, then, is to transform these momentary experiences of appreciation into lasting sentiments. The present perspective offers two novel predictions regarding avenues by which people can do so. First, a partner can serve as a means to goals that are fundamental to the person pursuing them and to goals that are unlikely to be (fully) satiated by other partners. In other words, striving to serve goals that the goal pursuer is unlikely to ever cease pursuing may enable a person to remain instrumental and close to a partner over time. The need for social connection and belonging are goals that are likely meet these criteria (e.g., Baumeister & Leary, 1995). Second, a partner can strive to be instrumental to multiple goals, so that even if one of the goal pursuer's goals is satisfied, the partner remains instrumental to the next goal that is pursued.
One's closest others are likely to be those who serve a variety of goals, making it easy to spend substantial amounts of time engaged in a variety of mutually rewarding tasks together. If a close other satisfies fundamental needs and serves a large set of goals, then appreciation for that person should be relatively constant. If, however, a person serves a smaller set of less grand goals (e.g., one's course instructor, who serves only a limited number of goals related to the academic domain), then appreciation for that person is likely to be limited both to times and contexts in which that person is instrumental.
Because goals change as a person moves through life, people likely seek partners who serve as means to their new goals and stop relying on people who served as means to previous goals but who are no longer useful. In order for a person to continue to be close to a partner, he or she must remain instrumental to the goals of the moment (for a similar argument related to attraction, see . For example, parents' roles shift as their children develop, and the behaviors they enact to remain instrumental to their children should shift as a result. In romantic relationships, newlyweds may serve as instrumental means to having children, setting up a home, and helping each other find suitable career paths. However, once the children seem to be doing well, the house is settled, and one's career is stable, these goals may no longer garner as much attention. Thus, one's partner who was instrumental to realizing their dream life may be viewed as dull because he or she does not seem to be serving goal striving. The well-known benefits for relationships of engaging in novel, exciting activities (Aron, Norman, Aron, McKenna, & Heyman, 2000) may emerge, at least in part, because such activities provide an avenue for partners to serve as means to shared goals. Thus, it is important that one's partner is instrumental to the goals of the moment to maintain interpersonal closeness and relationship satisfaction.
When considered in light of a people-as-means account, it is not surprising that people often attempt to be a unique and irreplaceable means to remain important to a partner and in their own eyes Orehek & Forest, 2016;. To do this, they may serve as exclusive means to particular goals (e.g., by serving a goal no one else currently serves or by actively trying to prevent other people from serving as means to goals that one serves) or may be multifinal in a way that is unlikely to be reproduced by another person. In monogamous romantic relationships, partners serve as the exclusive means to sex. In friendships, one friend may relish being the only person in on a secret or the only person who shares a particular interest. For goals to which the partner is the exclusive (possible or actual) means, being instrumental may afford greater relational benefit than would instrumentality to goals that are also served by other people. However, a lack of instrumentality to a goal to which the partner is the exclusive means (e.g., lack of enjoyable sex in a monogamous relationship) would pose a more serious challenge to the relationship.
Each of the foregoing ideas also applies to the person serving as the means. When someone is instrumental to his or her partner's goals, then he or she feels interpersonally closer and derives a sense of self-worth (Inagaki & Orehek, 2017). Our people-as-means approach makes the novel prediction that being instrumental to multiple goals for a particular partner should increase feelings of connection to that partner and should enhance self-worth compared with serving fewer goals. This suggests that people may be better off looking for partners whom they can assist in many life domains rather than finding someone whom they can help with one focal goal. In addition, it highlights the fact that fitting a partner's multiple goals is just as important as having a partner fit one's own set of goals. The person serving as the means' willingness and ability to notice when a partner has new goals and to become a means to them should contribute to relationship quality. People serving as means who want to be supportive and who are responsive to their partners should be better suited to shifting their efforts dynamically in a way that suits their partners' needs (Fitzsimons et al., 2015;Lakey & Orehek, 2011). Future research could explore how these features predict sustained instrumentality.
Considering partners as means to goals allows for a new understanding of the propinquity effect. Research has revealed the power of physical closeness in determining social ties (Festinger, Schachter, & Back, 1950;Segal, 1974). Close physical distance and frequent encounters may promote social connection because each occasion on which a person is encountered presents the opportunity for that partner to be perceived as instrumental to one's goals (and for one to be instrumental to that person's goals). This can lead a person to adopt new goals to fit with the people who are present as well. Smart goal pursuit in an interpersonal context is about not only what is already most important to the actor but also what is possible and what partners are available to help. Thus, people often observe the partners who are available and then align action to fit with the context. If a person frequently encounters a particular partner, then it becomes more likely that the person will eventually figure out ways that partner may be instrumental and begin a relationship with him or her.

Relationship maintenance
When a person is committed to a relationship partner (regardless of relationship type), maintaining the relationship becomes a goal in its own right. Although many of the processes reviewed in the previous section may occur passively when a person has a partner who serves as a means or when they themselves serve as a means, the person can also actively attempt to maintain a relationship once that relationship has become a goal. Just as a person who rides a bike as transportation to work must maintain and care for the bike (e.g., repairing it, securing it with a lock), a person who begins a relationship with someone who facilitates their goal pursuits must also maintain and care for the relationship with the person (Orehek, 2018). However, even when maintaining a means-bike or relationship partner-becomes a goal, it continues to be evaluated according to its instrumentality as a means. The person still evaluates the bike according to its usefulness for transportation and a relationship partner according to his or her instrumentality, even after a maintenance goal is formed.
Engaging in relationship maintenance behaviors is an important part of partner retention (Buss, 1988;Buss & Shackelford, 1997). Our people-as-means offers unique insights into tactics available to retain a partner, which are applicable across relationship types. One may protect one's relationship with a current partner by engaging in the following tactics: (a) becoming a more instrumental means to one's partner, (b) allowing or providing opportunities for one's partner to become more instrumental to oneself, (c) preventing rivals from becoming instrumental means to one's partner, (d) preventing one's partner from becoming instrumental to rivals, (e) preventing oneself from becoming an instrumental means to alternative partners, or (f) preventing alternative partners from becoming an instrumental means to oneself. We review each of these in turn.
The first two tactics involve increasing interpersonal closeness with one's partner. This can be accomplished by enacting behaviors that increase felt closeness and commitment on the part of one's partner. A person can become more instrumental to a partner and/or can facilitate their partner's instrumentality. One way to do this is to increase the number of goals each person serves for the other person. Other possibilities include increasing the frequency or competence with which one person serves the other's goals, ensuring that partners are serving each other's most important goals of the moment and that each person recognizes and appreciates all of the ways the other person is instrumental.
Tactics 3 and 4 involve keeping rivals away from one's partner, referred to as mate guarding in romantic relationships. These tactics can be used in other relationships as well, such as when siblings compete for a parent's affection, when employees vie for recognition from a supervisor, or when several close friends angle for the honor of serving as a bridesmaid. Rivals can be pushed away from one's partner by blocking access to the partner or by engaging in competition with rivals (Arnocky, Ribout, Mirza, & Knack, 2014). Blocking a rival's access to the partner prevents the rival from becoming instrumental to the partner and prevents the partner from becoming instrumental to the rival (a role that could increase felt closeness). Competing with rivals is an attempt to be more instrumental than the rival or at least to convince the partner that one is more instrumental than the rival. This could be accomplished by actually outperforming the rival in terms of instrumentality or trying to alter the partner's subjective perceptions of relative instrumentality by glamorizing one's own abilities or contributions and/or disparaging the rival on these dimensions (Buss & Dedden, 1990;Schmitt & Buss, 1996).
Tactics 5 and 6 involve keeping oneself away from alternative partners, a process that we have termed relationship shielding. This can be accomplished by preventing oneself from being a means to others and/ or by preventing others from being instrumental to oneself. For example, a person can limit his or her appeal or attractiveness to potential partners (Forest et al., 2017;Frankenhuis & Karremans, 2012) and can avoid opportunities to help others progress toward their goals. Preventing alternatives from seeming instrumental can be executed by devaluing alternative partners Lydon et al., 1999Lydon et al., , 2003 and by directing attention away from alternatives (Linardatos & Lydon, 2011;Maner et al., 2009).
Whereas previous research involving romantic relationships has focused on devaluing the physical attractiveness of alternatives, the present perspective postulates that these effects should occur across goal domains. For example, a partner may want to avoid emotional infidelity, which results from mutual perceived instrumentality with someone that exceeds an appropriateness threshold. People can protect romantic relationships by having others (aside from the focal relationship partner) serve as means to goals that do not threaten the relationship (i.e., safe goals). For example, having a friend serve as a means to playing tennis might be a safer goal than having that friend serve as a means to back massages. Another way to protect romantic relationships would be to have close others who are nonthreatening to the relationship, such as close kin or friends of a different gender than one's romantic partner (i.e., safe targets), serve as means to important goals. Finally, members of one's social network can serve as means to retaining one's romantic partner by making it more likely that the relationship will last (termed coalitional mate retention; Barbaro, Shackelford, & Weekes-Shackelford, 2016;Pham, Barbaro, Mogliski, & Shackelford, 2015;. For example, friends may praise a person to his or her partner, may help keep rivals away from the partner, or may provide information about a gift the partner would cherish. Moreover, our perspective postulates that shielding tactics should also be enacted in nonromantic relationships.

Partner substitution
Our perspective uniquely postulates that partner substitution-replacing one partner with anotheroccurs for one of two reasons. First, partner substitution can occur when available alternatives are perceived to be more instrumental than one's current partner. An alternative should be perceived as more instrumental when he or she serves more goals or when he or she is more instrumental to one's current goals than one's current partner is. Such a relative valuation of an alternative over one's partner can occur simply because a highly instrumental alternative has become available or because one's partner has decreased in perceived instrumentality over time. A decrease in perceived instrumentality can occur because the partner has ceased to be instrumental to goals he or she used to serve or because one's goals have changed.
Second, partner substitution can occur when an available alternative affords the person greater opportunity to be instrumental themselves. This could occur when the person perceives him-or herself as more instrumental in helping the alternative partner (compared with the current partner) to reach his or her goals. Because serving as an instrumental means increases feelings of self-worth and social connection (e.g., Inagaki & Orehek, 2017), an alternative partner who makes this possible is desirable. An alternative partner can appear desirable because he or she allows the person to serve more goals than the current partner or because he or she allows the person to serve different goals than the current partner.
Within monogamous romantic relationships, emotional or sexual infidelity-in which a person engages in a relationship with an alternative partner but does not want to leave the long-term relationship-should occur for different reasons than permanent partner substitution. Infidelity is likely to occur because the person perceives the alternative partner to be instrumental for some goal(s) the long-term partner is not fulfilling or because the alternative affords the opportunity for the person to serve as a means to one of the alternative's goals that the person is not able to serve for the long-term partner. Alternative partners should often be less instrumental and/or allow oneself to be less instrumental overall compared with one's long-term partner and may be quite different from one's long-term partner precisely because they serve different goals. A person's long-term partner may be very instrumental overall and may satisfy many of the person's most important goals, such as having children, invested parenting, social connection, high status, and financial security. However, because one's current partner has satisfied many important goals, momentary considerations and particular contexts may increase the attractiveness of alternative partners who serve unfulfilled goals-even if those goals are not as important as the ones served by the long-term partner. Likewise, a person's long-term partner may afford the person the opportunity to be instrumental to many of the partner's goals. However, because the person has satisfied many of their partner's goals, momentary considerations and particular contexts may increase the attractiveness of alternative partners who allow one to serve new or shared goals-even if those goals are not as important as the ones served for the long-term partner.
Such a process may partially explain why people sometimes commit infidelity with partners whom they perceive to be of lower value than their long-term partner. If the person had higher overall value, then the person would be more likely to permanently leave a current partner in favor of the alternative. Each of the notions above suggests that changing contexts, such as starting a new career or moving to a new city, should increase the likelihood of relationship dissolution (Fitzsimons et al., 2015). As person's goals change, a partner may no longer appear to be the best candidate. As a person encounters new people and discovers new ways to be instrumental to others, he or she may wish to take on those new roles.

Relationship loss
People can lose a partner either because one person decides to end the relationship, the partners become more distant over time, or the partner dies. From our perspective, when a person loses a partner, he or she loses a means to goals. He or she also loses a partner to whom he or she served as a means. The previous partner who served as a means is no longer available to serve as a reminder to pursue certain goals (Principle 2). The more goals the partner served, the more the person should feel the loss (Principle 4). In addition, the more goals a person served for their partner, the more the loss should be felt. The loss of one partner should relax attempts at shielding alternative partners (Principle 7). Because of this, alternative partners may intuitively step up in an attempt to fill the role of the lost partner. Losing a means to one's goals can be disruptive to one's goal pursuit. Among widowed individuals, those who were highly dependent on their spouse experience more anxiety than those who were less dependent (Carr et al., 2000). It is possible that some of this anxiety may be the result of lost instrumentality. Indeed, when people go through a break-up with a partner who was instrumental to their goals, those people experience disruptions in goal progress (Gomillion, Murray, & Lamarche, 2015). As a result, people must decide whether to locate new means to those goals, to lean more heavily on existing means, or to abandon pursuit of the goal altogether, perhaps replacing it with a new goal to which means are available. When people derive gratification and a sense of self-worth from serving a partner's goals, they should suffer losses on these dimensions when they lose the partner.
When someone loses a relationship partner, either through relationship termination, gradual distancing, or death, other partners can be helpful by serving as a means to goals previously filled by the lost partner and allowing the person who has suffered the loss to serve as a means to their goals-either goals the person previously filled for the lost partner, or new goals. Romantic break-ups have been shown to lead to reductions in self-concept clarity, which predict postdissolution distress (Slotter, Gardner, & Finkel, 2010). When a person can regain means to his or her important goals and feel that he or she is instrumental to other people's goals, we expect that this should help restore selfconcept clarity, minimize the disruption posed by the loss, and facilitate coping.

Connections and Contrasts With Existing Relationship Theories
Relationship researchers have long acknowledged the important role that role that close relationship partners play in each other's goal pursuits. For example, attachment theory posits that partners serve as a secure base from which to engage in exploration (Bowlby, 1988;Feeney, 2004;Hazan & Shaver, 1990). Interdependence theory (Kelley & Thibaut, 1978) recognizes that partners can provide each other with rewards and costs and that they are dependent on each other to the extent that they influence each other's outcomes. The present perspective builds on previous theories by detailing the principles that govern how people serve as means and the specific implications for relationship initiation, maintenance, and dissolution. In addition, our perspective enables us to make predictions about relationship processes and outcomes in a variety of relationship types, including close relationships (e.g., romantic relationships, family relationships, and friendships) and relationships with more distant partners (e.g., acquaintances, colleagues, service providers, and service recipients).
A people-as-means perspective offers new insights into classic theoretical perspectives and principles in the area of close relationships. We consider some relevant examples here. According to the investment model (an extension of interdependence theory; Rusbult, 1980Rusbult, , 1983Rusbult, Agnew, & Arriaga, 2012), people's commitment to their relationships is a function of three factors: their relationship satisfaction, the investments they have made in the relationship, and the quality of their alternatives to the relationship. From our perspective, (a) relationship satisfaction should be felt when a partner is instrumental, (b) investments reflect dependence on a partner for continued instrumentality, and (c) alternatives should be evaluated according to their instrumentality. In addition, we have suggested that (a) one should experience relationship satisfaction when a partner allows one to be instrumental to that partner, (b) investments reflect dependence on the partner for continuing to allow oneself to be instrumental, and (c) alternative partners should be evaluated according to the degree to which they afford opportunities for one to be an instrumental agent. Although the investment model was especially novel because it could explain why people stay in harmful relationships (e.g., abusive) or unsatisfying relationships (e.g., due to investments and/or poor alternatives; Rusbult & Martz, 1995), our model is novel because it explains why some people may relish relationships in which they have especially needy partners and how those people can benefit from caretaking.
The Michelangelo phenomenon concerns how relationship partners help each other strive for and achieve their goals through affirmation (Drigotas, Rusbult, Wieselquist, & Whitton, 1999;Rusbult, Finkel, & Kumashiro, 2009). We suggest that affirming a partner's goal strivings is one way in which people serve as means, and people should benefit (via an increased sense of self-worth and felt connection with the partner) when they affirm a partner's strivings. Of course, people can also be instrumental to partner's goal pursuit in many other ways.
Theorists have made a distinction between communal and exchange relationships, emphasizing that benefits are given freely in response to needs in communal relationships and that reciprocity is expected in exchange relationships (Clark & Mills, 1979Mills & Clark, 1982). This perspective assumes that the person giving a benefit to another person incurs some cost (and will provide the benefit only up to some "cost threshold"). We do not make this assumption and instead emphasize the benefits of helping others (see also Williamson & Clark, 1989). A communal/exchange perspective would predict that people want to be instrumental to a partner if they care about the partner's welfare (i.e., when a communal relationship exists or is desired). Communal/ exchange theorists have nicely demonstrated that caring for others is a cause of instrumental behavior. To this, we add the proposition that people's care for others can also be a result of instrumentality.
Most recently, transactive goal dynamics theory (TGD; Fitzsimons et al., 2015) was presented as an innovative approach to understanding how relationship partners form a single self-regulating system. TGD is focused on understanding the antecedents and consequences of partners exhibiting transactive density-a new TGD construct defined as "the extent to which dyads have numerous and strong links among member's goals, pursuits, and outcomes" (Fitzsimons et al., 2015, p. 650). Our approach shares with TGD a general emphasis on the interpersonal nature of goal pursuit. However, our approach focuses on how specific principles of goal pursuit, drawn from goal systems theory , influence relationship processes. Our approach and TGD are complementary perspectives. From our standpoint, the most important prediction TGD makes about relationships is that people are more likely to remain in relationships in which partners experience better goal outcomes while working together toward goals than they would alone-an effect due in part to perceived instrumentality of the partner. We importantly add specific goal systemic principles that detail when partners will be perceived as instrumental (including an emphasis on multifinality). Our approach also adds the observation that people evaluate relationships on the basis of the opportunity afforded for them to be instrumental to a partner. In addition, we expect that people will make comparisons between one's partner as a means and alternative partners who could serve as means (compared with the TGD comparison to pursuing the goal alone).
There is another noteworthy difference between our approach and TGD: We emphasize that a partner who serves as a means to a particular goal will be replaced by an alternative partner who serves as a more instrumental means to that goal. Thus, rather than predicting relationship duration on the basis of the sum of one's contributions to goal pursuit overall compared with goal pursuit alone (the comparison emphasized in TGD), we predict continuation of a person serving a role as a means to a specific goal on the basis of their instrumentality to that goal compared with alternative partners. Once a partner is no longer serving as a means to any goals, then the relationship with that partner should be discontinued. Therefore, we expect relationships to continue when a partner continues to be instrumental (and allows oneself to be instrumental) and to end when a partner ceases to be instrumental (and the person is no longer instrumental to the partner).

Conclusions
Our people-as-means framework suggests that people form and maintain relationships with partners who are instrumental to important goals and dissolve relationships when a partner is no longer instrumental or when a more instrumental alternative is available. In addition, people often like to serve as means to goals. They derive a sense of personal importance, esteem, and value from their ability to help others. They also feel closer and more connected to the people they help. A consideration of the two people in the dyad-both of whom are goal pursuers and means-leads to our principle of mutual perceived instrumentality. We suggest that when both people in a relationship feel that they and their partner are instrumental to each other, relationships are more likely to endure and to thrive.

Declaration of Conflicting Interests
The author(s) declared that there were no conflicts of interest with respect to the authorship or the publication of this article.