Content uploaded by Tara Marshall
Author content
All content in this area was uploaded by Tara Marshall on Sep 17, 2018
Content may be subject to copyright.
Amae
1
Running head: AMAE IN JAPANESE RELATIONSHIPS
Day-to-Day Experiences of Amae in Japanese Romantic Relationships
Tara C. Marshall
Brunel University
Kim Chuong
Guelph University
Atsushi Aikawa
Tokyo Gakugei University
Author’s note: This research was supported by a Postdoctoral Fellowship grant
awarded by the Japanese Society for the Promotion of Science (JSPS) to the first author.
We are grateful to Susumu Yamaguchi and his students at the University of Tokyo for
their assistance with data collection and translations. We also thank Miki Tanaka for her
help with translations, and Rebecca Pinkus for providing statistical consultation.
Correspondence should be addressed to Tara C. Marshall, Department of Psychology,
School of Social Sciences, Brunel University, Uxbridge, Middlesex, United Kingdom,
UB8 3PH. Email: Tara.Marshall@brunel.ac.uk
Amae
2
Abstract
The cultural psychology of romantic relationships is relatively understudied. To redress
this lacunae, the present study examined the Japanese concept of amae – or, the state of
expecting a close other‟s indulgence when one behaves inappropriately – within the day-
to-day relationships of 30 Japanese undergraduate romantic couples. For two weeks, both
partners completed daily diaries that assessed their amae behavior (requesting, receiving,
and providing amae), relationship quality, conflict, and motivation to enhance closeness.
Results revealed that amae behavior was associated with greater relationship quality and
less conflict. The motivation to enhance closeness partially mediated the association of
amae with relationship quality.
Amae
3
Day-to-Day Experiences of Amae in Japanese Romantic Relationships
To engage in amae is “to depend and presume upon another‟s love or bask in
another‟s indulgence” (Doi, 1992, p. 8). Long considered a fundamental aspect of
Japanese folk psychology (Doi, 1973; Yamaguchi, 2004), amae has been described as an
“indigenous Japanese concept of relatedness” (Behrens, 2004, p. 1). Many questions have
been raised regarding the nature and function of amae in close relationships. For
example, does the expression of amae enhance or detract from the quality of close
relationships? Do men and women express amae differently, and does gender moderate
the association of amae behavior with relational quality? Because little empirical research
has examined these questions (Behrens, 2004), the present study was intended as a first
step toward gauging the day-to-day experience of amae in Japanese heterosexual
relationships and its implications for relationship well-being.
What is Amae?
Derived from the Japanese word for “sweet,” amae refers to the feelings of
warmth, security, and intimacy experienced when one knows that a counterpart will
provide unconditional regard. Yet amae is more than a feeling state – it also refers to
behaviors associated with those feelings (Yamaguchi, 2004). As such, it is proposed that
there are two behavioral components to amae: (a) an inappropriate request or behavior,
and (b) a presumption that the inappropriate request or behavior will be accepted by one‟s
counterpart (Yamaguchi, 2004). Behavior is considered inappropriate if it violates
expectations associated with one‟s age, status, or role. If a relationship is close, however,
the codes of appropriate social behavior are temporarily suspended (Taketomo, 1986); the
behavior may be perceived by both parties as excessive, but not necessarily transgressive,
Amae
4
as would be the case if the relationship was less close (Niiya, Ellsworth, & Yamaguchi,
2006). Amae has been variously translated into English as playful and babyish behavior
(Taketomo, 1986), as whining, sulking, and being spoiled or pampered (Johnson, 1993),
and as the British slang word “mardy” (Lewis & Ozaki, 2002). These English
translations, however, have negative connotations that miss its usually positive meaning
in Japanese.
Amae has also been likened to dependency and attachment, although Yamaguchi
(2004) has argued that these are separate constructs. For example, amae entails exerting
control over one‟s counterpart, whereas dependency involves relinquishing all personal
control. Indeed, the amae requester actively exerts control by seeking closeness or by
manipulating others (Yamaguchi, 2004), contrary to Doi‟s original conceptualization of
amae as passive dependency. Moreover, because amae tends to be reciprocal,
emphasizing the role of the amae requester and provider, it better reflects
interdependence rather than dependence, which tends to be unidirectional (Rothbaum,
Kakinuma, Nagaoka, & Azuma, 2007). Amae is also distinguishable from the attachment
bond that develops between infants and caregivers. Relative to attachment, amae tends to
emerge later in childhood, it is triggered by mild stress (e.g., feeling tired) rather than by
threats to one‟s safety, it is associated with sadness rather than with fear, and it is related
to interdependent rather than independent exploratory behavior (Rothbaum & Kakinuma,
2004). In short, neither dependency nor attachment, nor any one word in English, can
capture the complexity of the Japanese term.
That amae has no language equivalent led Doi (1973) to argue that amae reflects
the unique psychology of the Japanese people. This view is consistent with the genre of
Amae
5
nihonjinron – studies of Japanese culture and identity that emphasize its distinctiveness.
Others have pointed out, however, that Japanese culture is neither homogeneous nor
ahistorical (Gjerde, 2004), and that the Japanese language, rather than an age-old
embodiment of the national psyche, did not become a standard national language until the
Meiji period (1886-1912). This relatively recent historical development challenges the
essentialist view that amae terminology and its corresponding psychology have always
been present in Japanese culture. Moreover, recent research has found manifestations of
amae in non-Japanese cultures that have no vernacular for amae (Niiya et al., 2006;
Rothbaum et al., 2007). Amae, then, may be best considered a fluid cultural
representation that has different meanings and salience for different people. As such,
Japanese are likely to show considerable individual variation in their amae behavior.
Motivations of Amae Requester and Provider
Amae relationships are characterized by role-taking behavior – one person is the
requester and the other is the provider of amae. In verb form, the act of requesting amae
is called amaeru, and the act of providing amae is called amayakasu (Lebra, 1976).
Either interactant can initiate amae; the person who amayakasu is not always the passive
recipient of an amae request, but may actively encourage the other to amaeru. The
requester is usually motivated by a need to feel positively regarded, understood, accepted,
and intimate with the other (Rothbaum & Kakinuma, 2004), or by a desire to further
one‟s own goals. These motives are referred to as noninstrumental and instrumental
amae, respectively (Behrens, 2004). The provider is more likely to feel amae if he or she
imputes a noninstrumental rather than instrumental motive from the other‟s amae request
(Niiya et al., 2006). Even if the request inconveniences the provider, it may still stimulate
Amae
6
positive feelings in the provider if it signifies an affective, intimate, unconditional
relationship between the parties. It may arouse negative feelings if the provider feels
manipulated by the requester acting in his or her own instrumental self-interest rather
than in the interest of the relationship (Niiya, Yamaguchi, Murakami, & Harihara, 2001).
Amae behavior may enhance relationship quality, then, when the requester
and the provider are motivated by the desire to increase intimacy. Individuals with
stronger intimacy goals tend to be more likely to engage in activities with their partner, to
provide social support, and to perceive their partner as possessing similarly strong
intimacy goals, which in turn may enhance relationship satisfaction (Sanderson & Cantor,
2001). Intimacy motivation is also associated with more positive emotion and greater
cultivation of relationships through conversation (McAdams & Constantian, 1983). When
intimacy goals are blocked, partners may experience more negative emotion; indeed,
Japanese adults report feeling “lonely and sad” when their desire to request or provide
amae is thwarted (Kim & Yamaguchi, 1995). To the extent that these negative feelings
are directed at one‟s partner for failing to adequately provide amae or to accept one‟s
indulgence, the likelihood of conflict may escalate.
Amae in Romantic Relationships
From Doi‟s (1973) psychoanalytical perspective, adult amae relationships are
expressions of the mother-child prototype internalized in childhood. Kim and Yamaguchi
(1995) found that Japanese adults reported experiencing amae most frequently with
spouses, followed by mothers and friends. Indeed, adults may behave helplessly,
playfully, or coquettishly with their partners to increase intimacy (Behrens, 2004).
Imagine, for example, that a wife is tired after a long day at work, and phones her
Amae
7
husband to ask if he will drive her home, even though she usually takes the train and it is
inconvenient for him to drive so far. She knows it is an inappropriate request, but makes
it regardless because she feels confident that he will oblige her out of love and devotion.
Extrapolating from research on friendships suggests that amae may be viewed
positively by romantic partners because it signals that the relationship is intimate (Niiya
et al., 2006), and can facilitate relationship initiation and maintenance (Kato, 1996), but
little research to date has tested this hypothesis within actual relationships. Studies of
adult amae tend to test participants‟ reactions to hypothetical vignettes instead of to real-
life relationship events (Kato, 1996; Niiya et al., 2006). To the authors‟ knowledge, the
present study is the first to use a daily diary methodology to examine the day-to-day
experience of amae in real-life romantic relationships. In addition, we examined the ways
that this experience may be moderated by gender.
Gender. Are there gender differences in adult amae behavior? Although some
studies have found that women tend to express more amae in romantic relationships than
men (Ohsako & Takahashi, 1994), such as requesting amae through coquettish behavior
(Taketomo, 1986), others have observed that Japanese husbands request amae more often
than do wives because they expect their wives to “mother” them (Lebra, 1976). Some
Japanese wives may even refer to their husband as their “oldest son” because men tend to
show helplessness in the domestic realm (Lebra, 1984), and request amae to foster
intimacy or to manipulate their wives into doing things for them (Behrens, 2004).
Similarly, Allison (1994) argued that, in heterosexual relationships, Japanese women are
more likely to be the caregivers and men the care receivers, replicating the mother-son
relationship from childhood. Lebra (1984, p. 267) argued that such role-taking behavior
Amae
8
in amae relationships sustains “an extreme gender hierarchy with male master and female
servant.” Evidence of gender asymmetry in requesting and providing amae was found by
Onishi and Gjerde (2002) in a sample of Japanese married couples. They found that, for
securely-attached husbands, adherence to gender asymmetry was positively related to
marital attachment and security, whereas for wives, asymmetry was negatively related to
marital attachment and security. Thus, men appeared to reap more benefits from the
gendered role-taking of amae than did women, who perhaps resented their husbands for
requesting too much amae and wished to be indulged themselves. Based on these
findings, we predicted in the present study that men would be more likely to request
amae, whereas women would be more likely to provide amae to their partners. Moreover,
we predicted that requesting and receiving amae would be more strongly associated with
men‟s relationship quality than with women‟s, and further, that providing amae would be
related to women‟s lower relationship quality.
Overview of the Present Study
The present study explored the associations of amae behavior with relationship
quality and conflict in Japanese heterosexual relationships. We also examined whether
these associations were mediated by the need for intimacy, and moderated by gender. To
date, no study to our knowledge has examined amae in romantic relationships by asking
both partners to complete daily diaries over a period of time. The dyadic nature of this
data meant that the Actor-Partner Interdependence Model (APIM; Kashy & Kenny, 2000)
could be used to examine the influence of the actor‟s and the partner‟s independent
variables on the actor‟s dependent variable (referred to as actor and partner effects,
respectively). For example, the actor‟s relationship quality may be a function not only of
Amae
9
the actor‟s own amae behavior (actor effect), but also of the partner’s amae behavior
(partner effect). We tested the following hypotheses:
Hypothesis 1: Actor‟s and partner‟s amae behavior will be positively associated
with actor‟s relationship quality and negatively associated with actor‟s
perceptions of conflict.
Hypothesis 2: Traditional role-taking behavior, in which the male partner is more
likely to request amae, and the female partner to provide amae, will be
positively associated with relationship quality for men, and negatively
associated with relationship quality for women.
Hypothesis 3: Actor‟s motivation to seek intimacy will mediate the association of
actor‟s amae behavior with actor‟s relationship quality and conflict.
Method
Participants
Thirty-four heterosexual dating couples were recruited from two large universities
in Japan. The study was advertised through announcements in a psychology class and
through an email sent to psychology students. Both partners participated in the study;
each received 3000 yen as an incentive. All participants were born in Japan. Ten
participants (15%) had lived outside of Japan for an average of 1.9 years; of these
participants, 70% indicated that they had lived in the United States. Women and men did
not significantly differ in age (Ms = 20.58 and 21.16, respectively; p = .21). The average
length of the participants‟ relationship was one year (SD = 1.17) and ranged from 2
months to 5 years; 77% of relationships were between 3 and 12 months in length.
Additionally, 77% of participants indicated that it was a dating relationship, 16%
Amae
10
indicated that they cohabitated with their partner, 2% were engaged, and 5% did not
specify their relationship status. Finally, 23% of men and 42% of women lived with their
parents, χ2(1, N = 62) = 2.66, p = .10.
Procedure
This study consisted of an intake phase and a diary phase. For the intake phase,
partners arrived at the laboratory together and were placed in separate rooms to complete
questionnaires assessing demographic indices (the intake questionnaires also consisted of
items assessing aspects of the relationship and personality styles, but these will not be
discussed here). When both partners were finished, they were given oral and written
instructions about the diary phase of the study. They were informed that every day for the
next 14 days, beginning later that day, they would be emailed a diary record. They were
asked to complete it at night and email it back to the experimenter. At the end of the two-
week period, they would then receive a full debriefing and their payment.
At this point, participants were asked if they wished to proceed with the diary
phase of the study. 31 couples went on to complete the diary phase of the study, but one
of these couples was eliminated because one partner did not provide adequate data. To
encourage participants to remain in the study, couples who agreed at the end of the intake
session to complete the diary phase were asked to sign a form that indicated their
commitment to finishing the study. During the diary phase, participants who did not
complete a daily record that day were emailed a reminder to complete their diaries on
time. Finally, participants who agreed to complete the diary phase would not receive their
full payment until the end of the 14-day period. Therefore, 30 couples comprised the
final sample size. The couples were told that they would receive the diary record via
Amae
11
email every day for 14 days. All participants indicated that they had access to email on a
daily basis. Each participant was given 5 paper copies of the diary record and 5 envelopes
in case they did not have access to email at any point during the diary phase and had to
mail their responses. If they forgot to complete a survey one day, they were asked to skip
that day rather than to complete the diary by memory the following day.
At the end of the intake session, participants were again informed that they would
receive an email that day containing the diary record and were asked to reply to it before
going to bed that night. They were asked to complete the diary record in private to ensure
that partners did not see each others‟ responses, and to contact the experimenter at any
time throughout the two-week diary phase if they had any questions or concerns.
Participants were given an information sheet consisting of instructions, definitions of key
terms, and contact information for the experimenters. At the end of the diary phase,
participants were emailed a debriefing form, and either obtained their payment directly
from the third author or received it in the mail.
Materials
All materials were first translated from English to Japanese. A second translator,
unfamiliar with the hypotheses of the study, back-translated the materials from Japanese
to English. The original and back-translated English versions were then compared, and
changes were made to the Japanese materials to reduce differences while preserving
cultural appropriateness in meaning.
Demographic questions. During the intake session, participants answered
questions about their gender, age, place of birth, whether they had ever lived outside
Japan (if yes, where and how long), the number of relationships lasting six months or
Amae
12
longer they had been involved in prior to their current relationship, length of their current
relationship, the status of their current relationship (dating, cohabitating, engaged,
married, or other), their parents‟ employment status, level of education, and marital
status, and whether they were currently living with their parents.
Diary record: Part A. The diary record consisted of two parts. In Part A,
participants were asked to describe “the last interaction you had with your partner today.”
Interactions were defined as any verbal exchange between the participant and his or her
partner that lasted 10 minutes or longer. Talking on the telephone therefore qualified as
an interaction, but email, instant messaging, and text messaging did not. Moreover,
activities wherein the partners were together but did not talk (e.g., sitting beside one
another in class or watching a movie together) did not qualify as discrete interactions. In
this section of the diary record, participants indicated the nature of the last interaction
with their partner, the time that the last interaction with their partner began, the
approximate length of the last interaction, and the approximate length of the total
interaction with their partner that day. Expressions of amae in the last interaction were
measured by four items, which were each rated on a 5-point Likert scale anchored with
Very little (1) and A great deal (5). These questions were, “How much did you request
amae from your partner?” “How much amae did your partner provide for you?” “How
much did your partner request amae from you?” and “How much amae did you provide
for your partner?” The variable assessing how much the partner requested amae did not
significantly contribute to variance in any dependent variable over and above the other
amae variables, and was therefore dropped from the following analyses. Finally, Part A
assessed perceived conflict with one item (“How much conflict did you experience with
Amae
13
your partner?”) and the motivation to seek intimacy with one item (“How much did you
try to maintain or increase closeness?”). Both items were rated on a 5-point Likert scale
anchored with Very little (1) and A great deal (5).
Diary record: Part B. In this section, participants were asked to describe “how
you felt in your relationship today.” If participants did not interact with their partner that
day, they were instructed to skip Part A and proceed to Part B. Thus, participants were to
complete at least Part B every day. Three questions asked participants how they felt in
their relationship today, regardless of whether they had interacted with their partner:
“How much intimacy did you experience in your relationship today?” (anchored 1 = Very
little, 5 = A great deal), “How satisfied do you feel in your relationship today?”
(anchored 1 = Not at all satisfied, 5 = Very satisfied), and “How committed do you feel to
your relationship today?” (anchored 1 = Not at all committed, 5 = Very committed). It
was explained to participants during the intake session and reiterated in the information
sheet that intimacy referred to how emotionally close a participant felt to his or her
partner; it did not have to be sexual or experienced only through conversation. Because
intimacy, satisfaction, and commitment were highly related to each other (α = .88 and .86
for women and men, respectively), we standardized each of these three variables across
days and individuals, then summed them together for each day to form a single index of
daily feelings of relationship quality.
Results
Data Analytic Strategy
In this hierarchically nested data set, daily diary responses were nested within
person, and each person‟s responses were nested within dyad. As such, there were three
Amae
14
sources of interdependence in the data: an individual‟s own day-to-day responses were
interdependent, such that scores on Day 1 may be related to scores on Day 2; an
individual‟s day-to-day responses were interdependent with his or her partner‟s day-to-
day responses, such that an individual‟s responses on Day 1 may be related to his or her
partner‟s responses on Day 1; and a participant‟s responses on the intake scales were
interdependent with his or her partner‟s responses on these scales (Kenny, Kashy, &
Cook, 2006). We used multilevel modeling to control for this interdependence, following
Campbell and Kashy‟s (2002) procedures for using PROC MIXED in SAS. Actor and
partner effects were examined for each day that both partners provided data. Gender of
the participant was effect coded (1 for men, -1 for women) and all continuous variables
were centered on the grand mean prior to analysis. Means and standard deviations for
continuous variables across days and individuals are reported in Table 1. On the whole,
there were no significant gender differences in the aggregated means, nor was there any
evidence of gender asymmetry in the extent of amae requested, received, or provided
when within-person and within-dyad interdependence was accounted for in a multilevel
analysis. Correlations between average daily variables are presented separately for men
and women in Table 2.
Descriptive details of the interaction. Out of a maximum of 840 observations (60
participants*14 diary observations each), there were 425 completed records for the last
interaction with one‟s partner (Part A), and 760 completed records of feelings about the
relationship today (Part B). Of the diaries completed, then, 56% contained information on
both the extent of amae experienced in the last interaction and on relationship quality that
day. These observations formed the basis of most of the following regression analyses.
Amae
15
Details of the nature of the last interaction are described in Table 3. Female and
male participants highly agreed on the nature of the last interaction, χ2(1, N = 417) =
3.51, p = .74. Additionally, women reported that 61% of last interactions began between
9pm – 12am, and 10% between 6-9pm, whereas men reported that 50% and 14% of last
interactions began at these respective times, χ2(1, N = 415) = 5.47, p = .36. Women
reported that 40% lasted more than 6 hours (only 10% lasted 10-30 minutes); the
respective figures for men were 43% and 7%, χ2(1, N = 415) = 4.73, p = .45. For total
length of all interactions that day, women reported that 50% of interactions lasted more
than 6 hours, and only 8% were 10-30 minutes in total length; the respective figures for
men were 48% and 6%, χ2(1, N = 411) = 5.25, p = .39. Overall, then, men and women did
not significantly differ in their descriptive details of their last interaction.
To control for nature of the last interaction, time that the last interaction began,
and length of the last interaction in the following analyses, they were dichotomized and
effect coded (nature of the interaction: 1 = spent private time together, -1 = all other
categories; time that the last interaction began: 1 = after 9pm, -1 = before 9pm; length of
the last interaction: 1 = length of the last interaction lasted 6 hours or longer, -1 = total
length lasted less than 6 hours). Length of the total interaction was not included because
it was strongly correlated with length of the last interaction (b = .89, t(722) = 50.84, p <
.0001). Because more serious couples were likely to report longer interactions than less
serious couples (those who cohabitate, for example, may spend more than 6 hours a day
together), the following regression models also controlled for status of the relationship (1
= cohabitating or engaged, -1 = dating or other) and length of the relationship. Finally, to
test Hypothesis 2 – that gender asymmetry in amae behavior will differentially influence
Amae
16
associations with relationship quality for men and women – gender and its interactions
with the amae variables were included in all models.
Relationship quality. Actor‟s and partner‟s amae variables, along with gender,
interactions of gender with the amae variables, and the control variables, were entered
into a multilevel model to predict actor‟s relationship quality. Two actor effects were
significant: actor‟s perceptions of receiving amae (b = .65, t(347) = 6.01 p < .0001) and
providing amae (b = .55, t(330) = 5.91, p < .0001) were associated with actor‟s greater
relationship quality. Two partner effects were also significant: partner‟s amae request (b
= .34, t(338) = 3.18, p =.002), and the interaction of gender with partner‟s perception of
receiving amae (b = .22, t(341) = 2.09, p =.04) were positively associated with actor‟s
relationship quality. Analysis of this interaction revealed that partner‟s perception of
receiving amae was related to men‟s relationship quality (b = .29, t(320) = 1.95, p =.05),
but not to women‟s (p = .31). In sum, these results support Hypothesis 1: to the extent
that actors and partners engaged in various amae behaviors, actors reported enhanced
relationship quality on a day-to-day basis. Additionally, these results offered no support
for Hypothesis 2: there was no evidence of traditional gender asymmetry in the
requesting and providing of amae, and men and women reaped equal relationship benefits
from their amae behavior. Furthermore, men reported enhanced relationship quality when
their partners felt that they had received amae – a partner-pleasing orientation not
traditionally associated with males.
Conflict. To further test Hypothesis 1 – that actor‟s and partner‟s amae behavior
will be associated with actor‟s lower perceived conflict – the same predictor variables
included in the preceding analysis were entered in a multilevel model to predict actor‟s
Amae
17
perceived conflict. One actor effect was significant: actor‟s perceptions of receiving amae
was negatively related to conflict (b = -.32, t(348) = 3.56, p < .001). In parallel, one
partner effect was significant: the interaction of gender with partner‟s perception of
receiving amae (b = -.23, t(351) = 2.58, p = .01). Analysis of this interaction revealed that
partner‟s perception of receiving amae was significantly associated with men‟s
perception of lower conflict (b = -.31, t(322) = 2.54, p = .01) but not with women‟s
perception of conflict (p = .26). Overall, these results further buttress Hypothesis 1: men
and women perceived less conflict when they felt that they had received amae. Again,
there was no support for Hypothesis 2: rather than demonstrating traditional gender
asymmetry, with women instead of men reporting lower conflict when their partners
perceived receiving amae, we found the reverse. Therefore, it appears to be in men‟s best
relationship interests – both in terms of lower conflict and higher relationship quality – to
ensure that their female partners receive amae.
Motivation to seek closeness. According to Hypothesis 3, the need to seek
closeness will mediate the association of amae behavior with relationship quality and
conflict. In line with Baron and Kenny‟s (1986) steps for testing mediation, it was first
necessary to demonstrate that the independent variable (actor‟s amae behavior) predicted
the mediator (actor‟s need to maintain or increase closeness). Accordingly, the same
predictor variables as in the preceding analysis were entered into a multilevel model with
the need to maintain or increase closeness as the dependent variable. Actor effects for
requesting amae (b = .16, t(350) = 2.74, p = .01) and providing amae to one‟s partner (b =
.28, t(337) = 5.58, p < .0001) were significant. The latter effect was qualified by a
significant interaction with gender (b = .13, t(338) = 2.58, p = .01); decomposition of this
Amae
18
interaction revealed that providing amae to one‟s partner was more strongly related to
men‟s need to increase or maintain intimacy (b = .41, t(273) = 5.28, p < .0001) than to
women‟s need (b = .15, t(327) = 2.37, p = .02).
Because previous analyses established that requesting amae was not significantly
related to relationship quality or conflict, this left only actor‟s provision of amae as a
plausible independent variable in the meditational model (see Figure 1). Moreover,
because providing amae was not related to conflict, only relationship quality was tested
as a dependent variable. Results had already demonstrated that the independent variable
(actor‟s provision of amae) was related to the mediator (actor‟s need to maintain or
increase intimacy) and to the dependent variable (actor‟s relationship quality), therefore it
remained to be tested whether the mediator was related to the dependent variable when
controlling for the independent variable. As such, actor‟s and partner‟s need to maintain
or increase intimacy, actor‟s and partner‟s provision of amae, gender, interactions of
gender with the actor and partner effects, and the control variables were entered in a
multilevel model to predict actor‟s relationship quality. The actor effect of need to
maintain or increase intimacy was significant (b = .56, t(333) = 5.67, p < .0001).
The final step for testing mediation demonstrated that the association of the
independent variable with the dependent variable dropped from b = .97 (t(313) = 11.73, p
< .0001) to b = .73 (t(346) = 7.99, p < .0001) when the mediator was included in the
model (partner effects and the control variables were included as covariates in these
multilevel models). The Sobel test of mediation was significant, (Z = 4.88, p < .0001),
thereby lending support to Hypothesis 3: the association of actor‟s provision of amae
Amae
19
with actor‟s relationship quality was partially mediated by actor‟s motivation to seek
intimacy.
Discussion
The goal of this study was to explore the influence of amae in Japanese romantic
relationships. Results indicated that this influence was positive: when actors and partners
engaged in amae behavior, they rated their relationship quality as higher and conflict as
lower than when they engaged less in such behavior. To our knowledge, this is the first
study that demonstrates the positive influence of amae in the day-to-day relationship
perceptions of Japanese couples. Next, we discuss the specific associations of actor‟s and
partner‟s amae behavior – requesting, perceptions of receiving, and provision of amae –
with relationship quality, conflict, and motivation to seek intimacy.
First, actors reported greater relationship quality when their partners requested
amae, which may signal to the actor that the relationship is close. When actors
themselves requested amae, they reported greater need to maintain or increase closeness,
supporting claims that amae requests are often motivated by affective, non-instrumental
desires to seek emotional proximity with close others (Niiya et al., 2006).
Second, whereas actor‟s amae request was not directly related to his or her own
relationship quality or conflict, the perception that he or she had received amae was
significantly related to both. Perceiving that one has received amae may lead one to feel
unconditionally loved, indulged, cared for, and close with one‟s partner – feelings that are
central to the concept of amae (Doi, 1973). Furthermore, men reported greater
relationship quality and less conflict to the extent that their partners perceived receiving
amae, suggesting that men may bask in the reflected glow of women‟s amae satisfaction.
Amae
20
It is also noteworthy that perceptions of receiving amae was the only amae variable
related to perceiving less conflict. Feeling indulged may satisfy one‟s intimacy goals, and
remove any potential for conflict with a partner who does not adequately provide amae.
Third, providing amae to one‟s partner was related to greater relationship quality
and to the need to increase or maintain intimacy (particularly for men), suggesting that it
may function as a vehicle for demonstrating one‟s love and devotion to one‟s partner.
Providing amae signals that the relationship is intimate and the partner is loved
unconditionally (Doi, 1973) – a point not lost on the recipient, who will likely feel
happier in the relationship when provided with amae. Although providing amae usually
follows from a partner‟s amae request, some actors may proactively offer amae to
increase intimacy with a partner (Lebra, 1976). Importantly, the present findings
suggested that actors provided amae to partners – even if they did not request it – at least
in part to maintain or increase intimacy, which in turn was associated with greater
intimacy, satisfaction, and commitment.
Gender Asymmetry in Amae
We did not find any evidence of traditional gender asymmetry in amae behavior:
men were not any more likely to request and perceive receiving amae, and women were
not any more likely to provide amae. Moreover, there was no evidence that requesting
and receiving amae was more strongly associated with men‟s than with women‟s
relationship quality, nor was providing amae associated with women‟s lower relationship
quality, as found elsewhere (Onishi & Gjerde, 2002). If anything, partner effects were
more strongly associated with men‟s relationship quality and perceptions of conflict,
suggesting that women‟s perceptions of receiving amae had a stronger impact on men
Amae
21
than vice versa. Furthermore, the association of providing amae with the need to increase
or maintain intimacy was stronger for men than for women. It appears, then, that the men
in this sample were aware that providing their female partners with amae would bring the
couple closer together, and produce relational dividends for both partners. Thus, not only
was there no evidence of traditional role-taking behavior, with women more eager to
please their partners than vice versa – even at the expense of women‟s own need
satisfaction – but the present results suggested that it was the men who were particularly
attuned to their partner‟s amae needs, perhaps because relationships were more intimate,
satisfying, committed, and conflict-free for both partners when women felt indulged.
These results, however, may be unique to this particular sample of highly-
educated, urban, young dating couples, who may be more egalitarian in their gender role
behavior than older, married, or rural Japanese participants. Indeed, Lebra (1984) and
Onishi and Gjerde (2002) found gender asymmetry in married couples, who may be
especially prone to traditional role-taking behavior to the extent that the wife stays at
home to care for children while the husband works. These studies also detected gender
asymmetry through detailed interviews and a more extensive scale, in contrast to the
daily diary methodology used here.
Limitations and Future Directions
As with any study that utilizes measures that were developed and validated in a
different cultural context, it was challenging to ensure fidelity in our translations and to
establish cultural equivalence in meaning. It was also problematic to rely on a few single-
item measures. In particular, the daily diary item that asked participants to rate how much
they tried to maintain or increase intimacy in the last interaction with their partner may
Amae
22
have confounded two different motivational constructs. Maintenance and attainment
motives are linked to prevention- and promotion-focused concerns, respectively (Higgins,
1997). Future research would do well to examine the separate associations of maintaining
and increasing intimacy with amae behavior and with relationship quality.
Although we did gauge the general context of the amae behaviors by asking
participants to report the nature of the last interaction, we did not ask participants to
provide qualitative descriptions of the actual amae behaviors per se. We wanted to keep
the diary to a relatively short length to encourage daily completion and to reduce attrition.
Such additional information, however, could shed light on the specific amae behaviors
that are most associated with intimacy motivation, relationship quality, and conflict.
Furthermore, we did not include text messaging, instant messaging, or usage of
social networking sites in the definition of interaction, even though these may be popular
vehicles through which couples communicate in Japan, especially in light of the indirect
communication often favored here (Lebra, 1976). On the other hand, that we limited
interactions to direct exchanges, either face-to-face or (more rarely) over the telephone,
allowed for the influence of nonverbal communication cues such as tone of voice,
gestures, and facial expressions, which may be pivotal to the expression of amae in
Japanese relationships (Doi, 1973). Future research may help to determine the relative
weight that modern communication technology versus traditional verbal exchanges
contribute to amae behavior specifically and to relationship quality more generally, in
Japanese and in other cultural contexts.
These results raise several additional questions. For one, how accurate are people
at detecting a partner‟s amae request or provision of amae? One‟s partner may report
Amae
23
providing amae, but it is possible that one may not consciously perceive it. Alternatively,
one may feel that one has received amae, even if the partner has not reported providing it.
Research in the West that has used a quasi-signal detection paradigm to examine
perceived partner responsiveness in relationships may be of use here (e.g., Gable, Reis, &
Downey, 2003). For example, it is possible that a “miss” – when an actor does not
perceive receiving amae even though the partner has reported providing it – may still
boost the actor‟s relationship quality through the undetected influence of the partner‟s
responsiveness.
Future research might also devote greater attention to the motivations underlying
amae behavior. Not all individuals engage in amae to the same extent (Behrens, 2004);
accordingly, the link found here between amae behavior and the need to enhance
closeness may reflect individual differences in the strength of intimacy goals. It may also
reflect the influence of attachment styles, such that secure or anxious styles encourage
amae behaviors in order to enhance emotional proximity with partners, whereas avoidant
styles discourage amae in order to maintain interpersonal distance.
Concluding Remarks
Overall, these findings are the first to demonstrate that feelings of amae on a day-
to-day basis are associated with greater relationship quality, perceptions of less conflict,
and increased motivation to enhance closeness in a romantic relationship. The concept of
amae is often portrayed as unique and automatic within Japanese culture, but there is
growing recognition that the cultural ideology of amae needs to be teased apart from its
actual experience in the daily lives of Japanese citizens (Gjerde, 2001). Empirical studies
such as the present one seek to take additional steps toward this end.
Amae
24
References
Allison, A. (1994). Nightwork: Sexuality, pleasure, and corporate masculinity in a Tokyo
hostess club. Chicago: University of Chicago Press.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in
social psychological research: Conceptual, strategic, and statistical considerations.
Journal of Personality and Social Psychology, 51, 1173-1182.
Behrens, K. Y. (2004). A multifaceted view of the concept of amae: Reconsidering the
indigenous Japanese concept of relatedness. Human Development, 47, 1-27.
Campbell, L., & Kashy, D. A. (2002). Estimating actor, partner, and interaction effects
for dyadic data using PROC MIXED and HLM: A guided tour. Personal
Relationships, 9, 327-342.
Doi, T. (1973). The anatomy of dependence. New York: Kodansha.
Doi, T. (1992). On the concept of amae. Infant Mental Health Journal, 13, 7-11.
Gable, S. L., Reis, H. T., & Downey, G. (2003). He said, she said: A quasi-signal
detection analysis of daily interactions between close relationship partners.
Psychological Science, 14, 100-105.
Gjerde, P. F. (2001). Attachment, culture, and amae. American Psychologist, 56, 826-
827.
Gjerde, P. F. (2004). Culture, power, and experience: Toward a person-centered cultural
psychology. Human Development, 47, 138-157.
Higgins, E. T. (1997). Beyond pleasure and pain. American Psychologist, 52, 1280-1300.
Johnson, F. A. (1993). Dependency and Japanese socialization: Psychoanalytic and
anthropological investigations into amae. New York: New York University Press.
Amae
25
Kato, K. (1996). Cross-cultural studies of amae interactions in American and Japanese
adults: Constructing relational models and testing the hypothesis of universality.
Dissertation Abstracts International: Section B: The Sciences and Engineering, 56
(8-B), p. 4640.
Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). Dyadic data analysis. New York:
The Guilford Press.
Kim, U., & Yamaguchi, S. (1995). Conceptual and empirical analysis of amae:
Exploration into Japanese psychological space. In Proceedings of the 43rd Annual
Conference of the Japanese Group Dynamics Association (pp. 158-159). Tokyo:
Japanese Group Dynamics Association.
Lebra, T. S. (1976). Japanese patterns of behavior. Honolulu: University of Hawaii
Press.
Lebra, T. S. (1984). Japanese women: Constraints and fulfillment. Honolulu: University
of Hawaii Press.
Lewis, J. R., & Ozaki, R. (2002). Amae in the UK. Paper presented at the XIV
International Association for Cross-Cultural Psychology Congress, Yogyakarta,
Indonesia.
McAdams, D. P., & Constantian, C. A. (1983). Intimacy and affiliation motives in daily
living: An experience sampling analysis. Journal of Personality and Social
Psychology, 45, 851-861.
Niiya, Y., Yamaguchi, S., Murakami, F., & Harihara, M. (2001). When being
inappropriate is appropriate: The acceptability of Amae in the Japanese context.
Amae
26
Paper presented at the Conference of Asian Association of Social Psychology,
Melbourne, Australia.
Niiya, Y., Ellsworth, P. C., & Yamaguchi, S. (2006). Amae in Japan and the United
States: An exploration of a “culturally unique” emotion. Emotion, 6, 279-295.
Ohsako, H., & Takahashi, S. (1994). Effects of „amae‟ on interpersonal emotions and
conflict-solution strategies in interpersonal conflict situations. The Japanese
Journal of Experimental Social Psychology, 34, 44-57.
Onishi, M., & Gjerde, P. F. (2002). Attachment strategies in Japanese urban middle-class
couples: A cultural theme analysis of asymmetry in marital relationships. Personal
Relationships, 9, 435-455.
Rothbaum, F., Kakinuma, M., Nagaoka, R., & Azuma, H. (2007). Attachment and amae:
Parent-child closeness in the United States and Japan. Journal of Cross-Cultural
Psychology, 38, 465-486.
Rothbaum, F., & Kakinuma, M. (2004). Amae and attachment: Security in cultural
context. Human Development, 47, 34-39.
Sanderson, C. A., & Cantor, N. (2001). The association of intimacy goals and marital
satisfaction: A test of four meditational hypotheses. Personality and Social
Psychology Bulletin, 27, 1567-1577.
Taketomo, Y. (1986). Amae as metalanguage: A critique of Doi‟s theory of amae.
Journal of the American Academy of Psychoanalysis, 14, 525-544.
Yamaguchi, S. (2004). Further clarifications of the concept of Amae in relation to
dependence and attachment. Human Development, 47, 28-33.
Amae
27
Table 1
Means and standard deviations (in parentheses) for continuous variables across days
and individuals
Men
Women
Amae requested
3.33 (1.35)
3.43 (1.37)
Amae received
3.51 (1.31)
3.57 (1.28)
Amae provided
3.62 (1.29)
3.38 (1.26)
Need for closeness
3.91 (1.26)
3.88 (1.15)
Perceived conflict
2.71 (1.36)
2.91 (1.37)
Intimacy
3.16 (1.36)
3.34 (1.28)
Satisfaction
3.21 (1.26)
3.22 (1.23)
Commitment
3.19 (1.36)
3.18 (1.16)
Amae
28
Table 2
Correlations among men’s and women’s average daily amae behavior, need for
closeness, conflict, and relationship quality
Variable
1
2
3
4
5
6
1. Amae requested
.41***
.79***
.67***
.59***
-.05
.59***
2. Amae received
0.73***
.40***
.55***
.51***
-.12†
.64***
3. Amae provided
.65***
.67***
.41***
.58***
-.08
.55***
4. Need for closeness
.54***
.50***
.71***
.27***
-.18**
.62***
5. Conflict
-.21**
-.23***
-.09
-.11
.18*
-.25***
6. Relationship quality
.65***
.70***
.74***
.66***
-.33***
.56***
Note. Men‟s data is presented below the diagonal, and women‟s data is presented above
the diagonal. Correlations along the diagonal are between dyad members.
†p < .10. *p < .05. **p < .01. ***p < .001.
Amae
29
Table 3
Descriptive information about the nature of the last interaction.
Nature of the Interaction
Percentage of Interactions
Females Males
Spending private time together at one of the partner‟s
residences (or shared residence if partners live
together)
50 %
47%
Going to a specific place together for pleasure or
entertainment (restaurant, movie theatre, bar, travel,
etc.)
18 %
20%
Studying, eating, or meeting together at the university
campus
12 %
15%
Socializing together with friends or family
6 %
4 %
Talking on the phone
3 %
4 %
Engaging in sports or physical recreation
2 %
1 %
Other
9 %
7 %
Amae
30
Figure Captions
Figure 1. Testing actor’s need to maintain or increase intimacy as a mediator of the actor
effect of provision of amae on actor’s relationship quality. Note: The coefficient in
parentheses refers to the association between actor’s provision of amae and actor’s
relationship quality after actor’s need to maintain or increase intimacy was introduced
into the model. The Sobel test indicated that the decrease in this coefficient was
significant (Z = 4.88, p < .0001).
Amae
31
*p< .05, **p<.01, ***p<.0001
.41***
.56***
0.97*** (.73***)
Actor‟s
Relationship
Quality
Actor‟s
Provision of
Amae
Actor‟s Need
to Maintain or
Increase
Intimacy
