Abstract
Is it possible to interact with the dead? Belief in such encounters is more widespread than we might think. Yet sociologists, unlike other disciplines, have not fully engaged the question. Here, I review both long-standing theoretical objections to such research and recent theories that encourage attention to the issue. Leaning on the latter, I use closed- and open-ended survey data collected from 535 Americans to explore what I call “living-deceased perceived interaction.” My data show that nearly half of my study participants report meaningful and regular interactions with deceased relatives and friends who were important in their lives. I examine the characteristics of such interactions—how and when they are performed and what these experiences mean to respondents. I also investigate the role of one’s social location in initiating interactions with the dead. Finally, I explore the social benefits, if any, these interactions provide for individuals who engage in them.
Introduction
In the Japanese town of Otsuchi stands a very unique telephone. Callers pick up a seemingly disconnected handset and send messages “carried by the wind.” These messages are intended for very special people: deceased loved ones (BBC News 2019). A similar phone was recently installed in Priest Point Park, Washington, USA, offering hope of the same magical contact to those in North America. Lori Provoe, one of the phone’s regular users said, “When you’re grieving, you look for any avenue to try to connect . . . to make that emotional connection.” She feels the phone accomplishes this for her because she talks and believes she receives a response (CBS News 2021).
In the contemporary world, connecting with the dead is not confined to phone calls “carried by the wind.” Long-standing methods such as mediums are more popular than ever (Eaton 2015), with some like John Edward or the Long Island Medium televising their skills. New technologies offer additional possibilities. Amazon recently revealed an experimental version of Alexa—one that mimics the voices of users’ dead relatives. In a promotional video, a child asks Alexa to read a bedtime story in the voice of his dead grandmother—and so she does (“Amazon Alexa Mimicking Dead People’s Voices 2022). Rohit Prasad, head scientist for Alexa AI, explains “While AI can’t eliminate that pain of loss . . . it enables lasting personal relationships” (Vincent 2022). Google has gone even further in this area. A lifelike robot built and programmed using a deceased individual’s Internet trail and postings received a U.S. patent for “methods and systems of robot personality development” (Tung 2015; also see Sisto 2020). For Google, the dead can come alive as digital clones.
Belief in living-deceased interaction is more widespread than we might think. A 2023 Pew Research Center poll found that 44 percent of Americans believe one can communicate with the dead in dreams, in conversation, or by feeling the deceased’s presence (Tevington and Corichi 2023). Similarly, a recent OnePoll survey showed that 42 percent of respondents believe the dead can contact the living, 63 percent believe they have seen signs from deceased loved ones, and 55 percent believe deceased loved ones never leave our presence (Associated Press 2022). These numbers are roughly double those reported in surveys appearing over the past 30 years (e.g., Greeley 1987; MacDonald 1992; Moore 2005). 1
Are living-deceased exchanges actually possible? Are they examples of true interaction? Historically, sociological theories have argued against the claim. Yet recent work in psychology and sociology challenges such objections. Building on this recent work, this article explores the concept of “living-deceased perceived interaction”—that is, a living individual’s conviction that specific deceased people with whom they regularly interacted in life are hearing and responding to them, seeing and advising them, or regularly reacting to their requests. In essence, these individuals believe the deceased are an active part of their social networks.
This article begins by briefly summarizing long-standing theoretical objections that exclude exchanges with the dead as examples of interaction. I then discuss more recent theories—both psychological and sociological—that provide conceptual tools by which to place living-deceased perceived interaction on the sociological map. I test these recent theoretical ideas via an exploratory study—one using both closed- and open-ended survey data collected from 535 U.S. respondents. My data show that nearly half of my study participants report regular, meaningful, and impactful interactions with deceased relatives and friends. After documenting this phenomenon, I examine the characteristics of such perceived interactions—how and when they are performed and what these experiences mean to individuals in the study. I also investigate the role of one’s social location in initiating interactions with the dead. Finally, I explore what social benefits, if any, living-deceased perceived interactions provide for the living. Can these exchanges mirror the benefits sometimes found in human-to-human interactions? If so, should sociologists think of interaction in broader terms?
What Is Interaction? Can the Living and the Dead Do It?
Historically, many sociologists have promoted a rather strict definition of interaction. It involves reciprocal influence between individuals and usually happens during face-to-face encounters. Moreover, some theorists forward certain capabilities essential to all interactants. Weber (1947), for example, contended that consciousness—engaged awareness—is critical to interaction. Without it, actions are merely reactive and, consequently, nonsocial. By Weber’s standards, human exchanges with animals, objects, spirits, or the deceased are not true interaction because only humans are discernably engaged. In fact, Weber’s (1947:94) criteria also exclude many human-to-human exchanges: those involving habituating, fatigued, sleeping, or comatose humans or humans in the grips of euphoria.
Some theorists also promote intention as critical to true interaction. Intention is the ability to target one’s actions in ways that alter situations and maximize the achievement of goals. Like consciousness, intention is typically a human-only quality. Parsons (1937:47) was unequivocal on this point, arguing that behavior in nonhuman organisms was genetically programmed. Thus, animals lack the essential mindedness required for goal-oriented interaction. Similarly, Goffman (1959) viewed nonhuman entities as mere props utilized by humans to enhance one’s desired performance. One might pray to a spirit or the deceased to fill a void or help combat a negative emotion. But for Goffman, these entities are projections and do not represent true interactants.
For some theorists, self-identity is also vital to true interaction. Self-identity involves reflexiveness—an understanding of who one is with reference to biography and situation. For theorists such as Mead or Goffman, self-identity is necessarily restricted to humans because entities such as animals, objects, spirits, or the deceased lack the “I” so essential to reflexiveness. Only humans, via the body’s cognitive and affective capabilities, can see themselves as capable of social interaction.
Other-directedness also plays a role in human-only definitions of interaction. Some theorists contend that interactants must observe the world around them and plan to affect a response. For Mead (1934:154, Note 7), spirits or the dead represent little more than ideas or ruminations that individuals act toward or respond to while developing the self. Unable to take the role of the other, spirits, the dead—any nonhuman entity—are incapable of interaction.
Finally, some theorists cite participants’ ability to actively negotiate and define interactive situations as critical to true interaction. Interactants must develop shared interpretations of values, norms, or sentiments. Generally, language proves the primary tool in this task, although gestures and other symbolic forms can also play a role in the process (Mead 1934:124). Spirits or the dead are seen as incapable of such communication and cannot develop social agendas in concert with the living.
These five criteria permeate sociological approaches to interaction, encouraging a human-only approach in this area. 2 The theorists discussed here represent just a few of the thinkers who defend this position, including Berger and Luckmann, Blumer, Durkheim, Garfinkel, Schutz and Simmel, among others. For those who prioritize interaction as a vehicle to understanding the social, the majority define such exchange as a strictly human endeavor.
Changing Course
Several recent psychological and sociological theories have opened the door to serious considerations regarding human-nonhuman interactions—even those involving the deceased.
Mental Constructs and Continuing Bonds
The literature on bereavement provides insight regarding beliefs on living-deceased contact and communication. Such work is dominated by psychologists but also includes sociologists, religion scholars, and bioethicists. This research is occasionally intertwined with paranormal studies, but the majority of scholarship on living-deceased exchanges does not treat the practice as stigmatized or supernatural. Klugman (2006:258) has been the most vocal here, writing, “Whether a deceased person is actually reaching out or the [living] person is making a connection with an inner representation . . . of the deceased is irrelevant for these purposes. For the subject, these experiences are real without the need for reference to some objective, external reality” (also see Krause 2011.)
Communication with the dead is referred to by several terms: contact with an “internal representation” (Fairbairn 1952; Klass 1993; Rosenblatt and Meyer 1986), postdeath contact (Krause 2011), and most commonly, “continuing bonds” (CB; Klass 1993; Stroebe et al. 1992). CB occur when bereaved individuals report feeling the deceased’s presence—sometimes well after the loved one has died (Field and Friedrichs 2004; Klass 1993). In CB models, people do not sever themselves from memories of the deceased. Rather, people maintain some type of ongoing connection via mental constructions of the deceased. This may be done privately (e.g., mental reminiscing, revering the deceased’s possessions or photographs, visiting their grave, recalling their influence, embodying their beliefs and morals) or publicly (e.g., telling stories about the deceased, constructing physical memorials). Under certain circumstances, CB practices may include seeking guidance from the deceased or, less often, seeing their image.
The questions explored by CB scholars vary. Some examine the potential benefits of the practice. For example, Krause (2011; Krause and Bastida 2009) finds that individuals who engage in contacting the dead develop a heightened sense of general connectedness to the living, deeper feeling of religious meaning in life, and diminished death anxiety. Others study the practice as a coping technique to help individuals resolve grief (e.g., Bennett, Hughes, and Smith 2005; Carr 2020; Field and Friedrichs 2004; Stroebe et al. 1992). Still others see inner representations of the dead as models for action, helping the living plan their futures (Marwit and Klass 1995).
There are contingencies surrounding CB practices. Certain demographic characteristics are sometimes associated with such behaviors (Carr 2020; Kalish and Reynolds 1973; Klugman 2006). Others examine the persistence of CB, ranging from a lifelong phenomenon (Klugman 2006; Nowatzki and Kalischuk 2009) to one that decreases as individuals move further from the loved one’s death and coping has been somewhat mastered (e.g., Rando and Buzo 1993; Vickio 1999). Note, however, that some refute the coping benefits of the practice, arguing that CB can have maladaptive consequences for the bereaved depending on the circumstances of the deceased’s death, the strength of the relationship between living and deceased, and characteristics of the bereaved (e.g., Carr and Sharp 2014; Eisma and Nguyen 2023; Field 2006; Field and Filanosky 2009; Stroebe and Schut 2005; Stroebe, Schut, and Boerner 2010; Stroebe, Schut, and Stroebe 2005).
Although some of the encounters discussed in the CB literature present the interaction as potentially “real” (Klugman and Krause are most prominent here), most of these encounters are described as internal dialogues—something purposively constructed by the living based on what they assume the deceased would say or do. I argue that living-deceased perceived interactions are something quite different. These exchanges are always viewed as two-way interactions, situated in social realms appropriate to such exchange, guided by social patterns that govern human-to-human exchange, and resulting in outcomes that mirror living-to-living interaction. Two recent sociological theories help support the distinction I am drawing
Actor-Network Theory
Actor-network theory (ANT) addresses some of CB’s weaknesses and offers the broadest theoretical tool by which to explore living-deceased perceived interactions. ANT uses a relational lens to build a role for nonhumans in social interaction. The theory begins by redefining the social. In ANT, the social consists of patterned networks of heterogeneous materials called “actants” (Callon 1987:93; Latour [1984] 1988, 1997, 2005; Law 1987:111). Actants connect, forming alliances and associations that result in an “actor network.” Thus, in ANT, actants are any independent entities that at any time can acquire the ability to make things happen within the actor network. Actants need not express intention, experience consciousness, or be reflexive. Instead, the initiation of an action-reaction chain is key, not the motivations or intentions behind action. For example, your doorbell, while typically defined as an object or prop in other theories, becomes an actant in ANT. This is because the doorbell, once it rings, engages you; it requires you to answer or ignore it, to make a decision and respond. This idea is important because it means that actants are not defined simply by what they are—human versus nonhuman. Rather, actants are defined by what they do and with, through, or about whom or what they interact. In this way, actants can be human or collectives; they can also include nonhuman entities such as animals, objects, (i.e., machines, clothing, money), text and other symbols, or mental concepts (i.e., deities, spirits, memories, projections, ideas). “An actant can literally be anything provided it is granted to be the source of an action” (Latour 1997, 2005).
But how can nonhuman entities be the source of action? Callon (1986) contends that this happens when a set of key actants—called “focal actants”—initiate a four-stage process called “translation.” In stage one, focal actants in the network execute three tasks: They negotiate and define the problem at hand, identify the actants that are relevant to its solution, and make themselves indispensable to all solution-oriented strategies. In stage two, focal actants diffuse through the network the issues, actants, and strategies just articulated and work to make these things acceptable to all network members. Stage three involves the creation of stable alliances built around the newly defined reality. In translation’s final stage, network consensus ensues around the newly defined reality. The network moves beyond acceptance and toward enactment. When the process of translation is complete, human and nonhuman alike become interlocking units—parts that, together, form a coherent whole. From that point forward, one’s legitimacy in social interaction derives from relationships—from actants (human or nonhuman) “doing” in concert and creating results based on their functional positions in the network.
In ANT, nonhumans—such as animals, deities, the deceased, objects, avatars, and so on—mediate interaction in ways as significant as those of humans. Thus, when scientists study test tube contents or observe cells through microscopes, writers link to the works of their predecessors, manufacturers create and disperse goods adopted by community members, or living individuals call on deceased loved ones, all actants form connections that function as critical parts of a network. In these scenarios, not all members of the network are conscious or intentional, not all have a sense of self, but all beckon or preclude interaction. All members—human and nonhuman—can make things happen (Law 1992).
ANT helps us overcome some historical objections to living-deceased perceived interaction. Focal actants allow the deceased to make things happen. But how? A second theory, doing mind (DM), helps answer that question.
Doing Mind
Some feel that the ANT concept of translation, in part, suggests nothing more than projection (e.g., Amsterdamska 1990). DM, another recent theory, forwards a deeper look into the process of human-nonhuman interaction, suggesting more active input from nonhuman interactants. Clinton Sanders centralized the concept in his studies of human-animal interaction. He defined DM as “meaning or intention to behavior [which] hinges crucially on the extent to which such behavior is considered meaningful within the context of the social relationship” (Sanders 2007:324). Sanders (2003:407) concedes that humans and animals may not contribute equally to the interaction: “Where both parties in an interaction cannot equally share the burden of this work, one may redistribute the load by doing mind for the other.” 3 Irvine (2003:49) adopts a similar approach, arguing, “The Other and I will act toward each other as two beings who can orchestrate their own conduct.” The animals “must be subjective others and not just the objects of anthropomorphic projection.” This is possible because “several dimensions of subjectivity appear among dogs and cats, constituting a ‘core’ self-consisting of agency, coherence, affectivity, and history” (Irvine 2004:3). Andrea Laurent-Simpson makes similar claims in studying families who elevate pets to the status of children. She argues, “[T]he formation of role identity in actors can also occur in relationship to nonhuman actors, if they are perceived as minded” (Laurent-Simpson 2017:610).
Animals are not the only nonhumans sociologists may wish to theorize and analyze. Thus, Owens (2007) pushes DM further, allowing us to apply the concept to a variety of nonhuman entities, including objects, deities, spirits, the deceased, and so on. Like ANT scholars, Owens acknowledges a break with prior theoretical claims, writing: “Interaction does not require that both parties willingly and purposefully engage with one another. All that is required is the assumption by one party that this is so.” In addition, she both supports and expands the original DM models by defining some interaction as the result of a perceptual shift, one where an interactant is a member of a “pairing or group who is perceived as being able to act independently of the other party(ies), regardless of whether this behavior is minded, sustained, or meaningful in a conventional sense” (Owens 2007:568–9). In DM, interaction is not merely an internal dialogue with an imaginative construction. Rather, Owens brings the pairing or group and the two-way nature of interaction to the foreground in ways that CB does not.
Unlike ANT, Owens’s (2007) version of DM is not universal. She contends that four contingencies make this perceptual shift possible. First, potential nonhuman interactants must appear capable of action that is independent of direct human manipulation. Thus, your dog may seem like a legitimate co-interactant, but a favorite toy may not. Second, humans must perceive nonhumans’ ability to act independently in situations that preclude humans from achieving their desired goals. In essence, something is stopping the human from acting, but the nonhuman is capable of filling in. Third, humans must experience a certain level of urgency, making interaction with nonhumans necessary to goal achievement. Thus, we consider interacting with nonhumans only in the face of pressing circumstances or emergencies. Finally, nonhuman targets must be viewed as necessary to completing a task or achieving a goal.
Owens’s (2007) version of DM suggests that our definition of interaction must be expanded. For her, humans can form and understand self through interaction with other humans. Then, when attempting to engage in previously defined noninteractional behaviors such as human-nonhuman interaction, humans “use an understanding of social interaction as a sort of thought template” (Owens 2007:578).
Applying Recent Theories to the Analysis of Living-Deceased Perceived Interaction
CB, ANT, and DM present ideas that challenge long-standing human-only definitions of interaction. CB models provide an empirical starting point for the study of living-deceased exchange, documenting the breadth of such occurrences. However, CB theories typically (but not universally) frame such interactions as one way, temporary, and internally experienced. Both ANT and DM help us move to a more social experience. The theories place perception on equal footing with physical copresence and action. First, ANT’s concept of actant redefines social networks and their components. This allows us to include nonhuman entities—including the dead—in a broader network of interactants. DM presents the idea of perception change, allowing for a more flexible definition of interaction. Both theories also provides guidelines by which human-nonhuman interaction may ensue (DM) and the role of nonhumans (ANT).
I use these theories to center the notion of living-deceased perceived interaction in sociological studies. Specifically, I present four multifaceted research questions to help us better understand the phenomenon. My first question is designed to document the frequency of and meaning attributed to living-deceased perceived interactions.
Research Question 1: What percentage of respondents report perceived interaction with deceased relatives and friends? Do these individuals equate these exchanges with those experienced with living relatives and friends?
Because living-deceased perceived interaction has been understudied in sociology, I build on issues raised by ANT and DM and ask:
Research Question 2: By what processes do living-deceased perceived interactions occur? Are the processes similar to those typically associated with human-to-human interaction? And do such interactions follow the premises presented in ANT and DM?
In response to calls from CB scholars for increased information on the demographics of perceived interaction with the dead, I ask:
Research Question 3: What is the demographic profile of those who engage in living-deceased perceived interaction? Are certain groups more likely than others to report such exchanges?
Finally, both CB and DM suggest that living-deceased perceived interaction can have positive consequences for the living. Knowing this, I ask:
Research Question 4: What advantages, if any, do living-deceased perceived interactions provide for the living? Can these experiences furnish any of the benefits associated with human-to-human interactions?
Methods
I obtained data for this exploratory study of living-deceased perceived interactions via a survey containing both closed- and open-ended questions. Like most work in this area, I focused on living-deceased exchanges between relatives and friends. This allowed me to tap relationships that promised some depth of experience as opposed to cursory exchanges with strangers.
Some of the survey questions on living-deceased perceived interaction were inspired by the Changing Lives of Older Couples data (Nesse et al. 2006; also see Carr, Nesse, and Wortman 2006; Cerulo 2009:542.) Others were original questions pretested before administering the survey. Finally, I posed several questions regarding physical and mental health, social connectivity, and feelings of happiness modeled after those included in the General Social Survey.
Qualtrics administered the survey using a sample of 535 respondents drawn from panels available to the company. Qualtrics builds panels based on years of recruiting individuals. This includes intercepts 4 from offer walls, affiliate networks, social media, or other platforms, all of which drive traffic to a survey. My sample is not representative of the American population. However, sampling criteria stated minimum proportions of people from certain age groups, education levels, genders, marital statuses, races, religions, and social classes. This maximized the heterogeneity of respondents. Table 1 gives a demographic breakdown of the sample. 5
Sample Demographics.
Closed-ended questions were analyzed in two ways. First, I calculated and reported frequencies of occurrence for questions addressing participation in and characteristics of living-deceased perceived interactions—such as, “Have you engaged in meaningful interaction with a deceased relative or friend—the kind of communication you might have with another human being?” or “How do interactions with deceased relatives and friends make you feel?” Where possible, I compared the frequencies reported in my study to those found in other studies and reliable surveys/polls. Second, when exploring the association between variables—including associations between gender or religion and living-deceased perceived interaction—I used cross-tabs along with appropriate measures of association. However, because I am not dealing with a representative sample, the statistics presented here are strictly descriptive.
I also included open-ended questions in my survey to gain insight into the processes, feelings, and purposes involved in pursuing living-deceased perceived interactions. These quotes capture the elements that distinguished my respondents’ experiences from the one-way, imaginary constructions reported in many CB studies. In dealing with answers to open-ended questions, I content analyzed all responses and classified them thematically. A second coder—one not involved in the study—repeated this process using 15 percent of randomly selected open-ended responses. Intercoder reliability was 90.5 percent.
Findings
Who Perceives Living-Deceased Interactions?
Respondents were asked, “Have you engaged in meaningful interaction with a deceased relative or friend—the kind of communication you might have with another human being?” Forty-seven percent answered yes. This proportion is nearly identical to the Pew Research and OnePoll surveys discussed earlier and an Omnibus/Yougov (2012) poll.
Respondents answering yes to my first question were then asked, “How often do you interact with a deceased relative or friend? Most indicated an ongoing exchange, with 20 percent answering daily, an additional 20 percent answering a few times per week, and 18 percent saying a few times per month; 34 percent answered a few times, per year and 8 percent said rarely. Similar to some works on CB (Klugman 2006; Nowatzki and Kalischuk 2009), these answers suggest fairly regular and persistent contact.
By What Processes?
Did respondents view their communication as one-way imaginings or an internal dialogue as suggested by much of the CB research, or were such exchanges experienced as a two-way exchange as DM proposes? Did respondents hear or see the dead? Did living-deceased perceived interactions follow patterns similar to living-to-living interaction? Were the dead, as ANT suggests, able to make things happen?
Several survey questions addressed these issues. Respondents were asked: “When you interact with the deceased, what are the circumstances?” (Multiple answers were permitted.) Here, 63 percent said they felt the deceased’s presence in their physical space. Answers to open-ended questions showed that such feelings typically occurred when the living summoned the deceased. Sixty-one percent identified reciprocal communication in thought—conversations described as similar to those respondents had with the deceased when they were alive. The same number, 61 percent, reported speaking aloud to the deceased, with many saying they then heard (30 percent) or saw (22 percent) the deceased. Among those who perceived living-deceased interaction, 100 percent said they saw the communication as a two-way exchange—a back-and-forth interaction between living and deceased. Here, too, many of these results were quite similar to those found in the polls discussed earlier. In addition, respondents’ answers reflected the processes described in DM research. Respondents viewed the deceased as capable of interaction and able to act independently.
To further explore living-deceased perceived interaction, I asked: “How do interactions with deceased relatives and friends make you feel? Are the feelings similar to those experienced when your loved one was alive?” Seventy-eight percent of respondents said these interactions made them feel connected to their loved one, and their exchanges were similar to those had when the dead were alive. Similarly, 65 percent said the interactions made them feel as if they still had a friend or supporter. Respondents also reported that living-deceased perceived interactions made them feel calm (75 percent), loved (73 percent), and happy (71 percent). For most, living-deceased perceived interactions had a reported positive effect on mood and emotion. No one offered feelings of fear, resentment, or annoyance. Thus, my findings did not reveal the maladaptive consequences noted in some CB studies (e.g., Carr and Sharp 2014; Eisma and Nguyen 2023; Field 2006; Field and Filanosky 2009; Stroebe et al. 2010). This may be because my respondents typically initiated the interaction with a positive purpose in mind.
As in Krause’s CB work (Krause 2011; Krause and Bastida 2009), my respondents reported that perceived living-deceased interaction had other effects that influenced their daily lives: better coping skills (69 percent), an increased sense of well-being (65 percent), increased tolerance of others (54 percent), greater productivity (50 percent), and more trust toward others (43 percent). Here, too, many respondents (59 percent) said these benefits were the same as those garnered from living-to-living interactions.
Finally, respondents were asked: “When you interact with the deceased, what are the circumstances?” (Again, multiple answers were permitted.) The most frequent site for living-deceased perceived interactions occurred in the privacy of respondents’ homes (80 percent). Forty-seven percent named a place the deceased once lived; the same percentage said it happened in a place they and the deceased frequented on a regular basis; 45 percent said they interacted with the deceased at the person’s gravesite, and 31 percent said they were in a sacred place like a church or shrine. No one reported perceived interactions involving mediums, séances, or third parties. Many said that a third party would make the exchange seem artificial. Thus, people most often reported living-deceased perceived interactions in contexts where they had experienced living-to-living interactions. In the case of gravesites, many said that they chose this place because the deceased, while living, had engaged in similar gravesite behavior. In essence, all of these places seemed “natural” sites for “reconnecting” with their deceased loved one because these places were where their living-to-living interactions occurred (Jonsson and Walter 2017). Thus, in line with ANT, the dead remained an active element of respondents’ social networks (albeit in a different form from the living) and able to make things happen within the living’s social space.
More on the Process of Perceived Interaction
Earlier, I defined living-deceased perceived interaction as a living individual’s conviction that specific deceased people with whom they regularly interacted in life are frequently hearing and responding to them, seeing and advising them, or regularly reacting to their requests. In essence, these individuals believe the deceased is an active part of their social network. Closed-ended questions helped chart the similarities between living-to-living and living-deceased interactions. But open-ended question told me much more about the feelings behind these perceived interactions.
As in living-to-living interaction, many respondents perceived the dead to be conscious and involved in two-way communication, possessing the capacity to answer respondents’ questions or pleas for guidance or support, sustain a conversation, or answer their requests and make things happen. They saw the dead as exhibiting the same identity and sense of connection they had when alive. For example, an Asian man in his 20s said, “I think they [the dead] can guide and protect . . . they send signs that make it clear that they are watching over you.” A White woman in her 40s noted, “When I am trying to do something and I can’t, I pray to my Dad and always I am able to accomplish what I am doing; he hears me and helps. . . . He tells me the very same things he did when he was alive.” A White man in his 20s said: “I always talk to my deceased parents especially when I’m trying to make a life changing decision and I somehow am at peace in whatever decision I make. . . . They help me see I am making the right decision.”
Unlike simply reminiscing or thinking of the deceased, respondents reported that living-deceased perceived interaction provided a support system—one similar to that experienced with other living individuals. As ANT and DM suggest, the dead remain part of the living’s social network and act when summoned by the living. A Black woman in her 60s said: “[These talks] help me carry on. Many times, I just want to give up and let loose without a care but after I talk to my deceased best friend, I feel better and stronger.” A Latinx man in his 40s reported: “When I really need help and no one else is there, interactions with deceased relatives strengthens me, makes me feel I am listened to and understood. They come to the rescue.”
The connections reported by respondents were not internal dialogues, as many CB studies suggest. These connections were, for many, as strong as those they maintained with the living. The dead, as ANT suggests, simply change their role in the living’s social network but are, nevertheless, meaningful actants in the network. A Black male in his 40s offered, “I feel my friend is still with me. I feel like I am having the same relationship as if the person were here.” A White woman in her 60s noted, “It’s very important because I miss my parents and siblings, so any chance to reconnect is a blessing. I feel as if though they are gone, they are soooo with me.” And as a White man in his 20s said, “It helps me keep their memory and makes me feel like they never left. By talking to them and hearing them, I never feel alone.”
To underscore the authentic nature of living-deceased perceived interactions, a sizeable portion of participants (roughly 30 percent) described how their deceased loved ones explicitly answered the requests stated in their interactions. An Asian man in his 40s said: I’ve been in touch with relatives—mother and brother, and it’s a time when I’m going to be involved in a situation dealing with a decision. Depending on the situation, I’ve been told by my brother to get my act together, or, out of nowhere, monies arrive to give me cash I need.
A Black woman in her 20s noted, “I can receive messages that I need to get and relay messages to them. It’s a gift you receive,” while a White man in his 60s said, “I know my spouse and siblings hear me and I feel better talking to them. Sometimes I know they can help and sometimes I know they helped and responded to my request.” In essence, ongoing, two-way exchange was key among these respondents. As a White man in his 20s said: “I know they know what I am asking and why. I don’t ask for anything that would get them in trouble and most of the time they come through and I am thankful for that.”
As the living describe the processes involved in living-deceased perceived interaction, we see something more than inner dialogues, constructed or imaginary connections, or temporary bonds so often portrayed in the CB literature. Indeed, we can see the premises of ANT and DM come to life. As ANT notes, the dead are fair game in interaction because for the living, the dead make things happen. Humans become the focal actants that draw the dead into interaction. After doing so, the dead become viable members of an action network. The dead can make the living feel socially connected, supported, positive, tolerant, and trusting. Furthermore, the living experience these feelings in places where living-to-living interactions occurred. Thus, the substance of the dead individual may have changed, but their contribution to the social network remains, in many ways, the same. DM theory argues that certain contexts and conditions—such as a need for guidance, strength, help, or connection—create the circumstances by which the dead can become viable participants in social interaction. The quotes previously presented illustrate this. When do the living reach out to the dead? “When I am trying to do something and I can’t, I pray to my Dad” or “Many times, I just want to give up and let loose without a care but after I talk to my deceased best friend, I feel better and stronger.” Here, the living cannot deal with a situation on their own, so they “construct then forget” that the target of their communication is not physically copresent. Individuals define the dead as capable of providing help when they themselves cannot meet the demands of the moment. And often, as DM suggests, those moments involve a sense of urgency (Owens 2007:577).
Social Location and Living-Deceased Perceived Interactions
Are certain characteristics of one’s social location more or less likely to encourage living-deceased perceived interactions? CB studies show that women, the less educated, and the unmarried are most likely to engage in living-deceased communication. Why? Some argue that men see such contact as maladaptive or triggering unwelcomed reminders of their loss. Low levels of education are often linked to high religiosity, with scholars arguing that religiosity makes living-deceased contact more likely. And many argue that married individuals are not as isolated as the unmarried, giving them more avenues for living-to-living support (e.g., Klugman 2006; Shermer 2000; Simon-Buller, Christopherson, and Jones 1989; Sochos and Bone 2012; Stroebe and Stroebe 1991). Some CB scholars (Field, Hockey, and Small 1997) have called for additional research regarding the impact of age, race and ethnicity, religion, and social class on living-deceased communication. Because these things matter in prayer practices with deities, saints, angels, and so on (e.g., Baker 2008; Cerulo and Barra 2008; Levin and Taylor 1997; Stark 2008:chapter 7), such factors may be important to living-deceased contact. I pursued all of these elements in my study. I also explored what I call experiential factors.
Demographic factors
In my data, I found that gender and one’s current religion were significantly associated with living-deceased perceived interactions. A greater proportion of females (51 percent) report meaningful living-deceased perceived interaction than do males (40 percent). (These findings are similar to those of CB studies and the most recent Pew survey on this topic.) My study also tapped transgender individuals; 57 percent reported living-deceased perceived interactions. However, because the transgender group constitutes such a small proportion of the overall sample (N = 21), we must review this finding with caution (χ2 = 9.39, p = .025; Cramer’s V = .132, p = .025). 6
The relationship between one’s current religion (if any) and living-deceased perceived interactions also proved significantly associated to living-deceased perceived interactions. The groups most likely to report such exchanges were Buddhists (79 percent), Latter Day Saints (75 percent), and other (including Odonists, Wiccans, and other spiritualist groups—70 percent). Muslims (60 percent) Catholics (59 percent), and Hindus (53 percent) followed. All of these religions hold some belief in an afterlife. Thus, for members of these religions, interacting with the deceased could be viewed as a viable possibility. Of the remaining groups, less than half reported living-deceased perceived interactions, with Protestants at 40 percent (similar to the percentage reported in the most recent Pew poll), Jews at 29 percent and “nones” at 25 percent. Because Jews and nones emphasize the here and now, the lack of living-deceased perceived interaction with the dead is not unexpected. But the low numbers of Protestants in this category is surprising because these groups maintain a belief in the afterlife. However, Protestants, unlike Catholics, discourage communication with the dead, referencing Old Testament warnings against it (e.g., Staples 2013; Williams 2023). In sum, the majority of religions emphasizing some form of afterlife and the ability to rejoin loved ones beyond death are most likely to engage in living-deceased perceived interactions (χ2 = 20.613, p = .008; Cramer’s V = .196, p = .008). However, because some religions constitute a small proportion of the overall sample, we must view this finding with caution.
In my data, no other demographic characteristics—including age, level of education, marital status, race, or social class—proved significantly associated to living-deceased perceived interaction.
Experiential factors
The broader project from which these data are derived dealt with other human-nonhuman perceived interactions: humans and animals, spiritual entities, avatars, and computerized office assistants. Do some individuals conform to tenets of ANT and DM, recognizing all nonhuman entities as potential interactants?
I found that 51 percent of those who perceived living-deceased interactions in their lives were also like to report perceived interactions with their pets (χ2 = 11.76, p < .001; Φ = .148, p < .001). Fifty-seven percent of that group reported perceived interactions with spiritual entities (e.g., God, Allah, Jesus, saints, and angels; χ2 = 42.24, p < .001; Φ = .281, p < .001), 64% reported perceived interactions with avatars used in gaming and other virtual life sites (χ2 = 24.60, p < .001; Φ = .214, p < .001), and 60% reported perceived interactions with computerized office assistants such as Alexa or Siri (χ2 = 28.75, p < .001; Φ = .232, p < .001). The associations are weak yet significant.
Literature on human-nonhuman interactions suggests that these findings are not surprising. Work on human-animal interactions shows that increasingly, people treat their pets as friends or members of the family (e.g., Carter and Charles 2018; Irvine 2003, 2004, 2008; Jerolmack 2009; Laurent-Simpson 2017, 2021; Stoltz, Van Ness, and Bjerre 2020). Similarly, cultures that believe in the existence of spirits (e.g., Asian, American, European, Native American) often attribute agency, intention, and consciousness to the deceased, making interactions with them as normative as interacting with the living (e.g., Bader, Baker, and Mencken 2017; Baker 2008; Cerulo and Barra 2008; Stark 2008). Researchers show that avatars and computerized assistants establish presence with humans, providing a sense of intimacy, understanding, enjoyment, involvement, and social support. And the more isolated the individual is, the more likely they are to perceive interactions with these entities (e.g., Felton and Jackson 2022; Ki, Cho, and Lee 2020; Riedl et al. 2011; Suzuki et al. 2023).
Taken together, these findings support the tenets of ANT and DM. Some people and some circumstances bring new, seemingly viable players into interactions. As such, the continued assessment of new ideas and theories on human-nonhuman interaction is warranted.
Living-Deceased Perceived Interactions and Health and Well-Being
Scholars have long argued that social contact can increase one’s quality of life, including better physical and mental health, satisfaction with social life, and overall happiness (e.g., Burleson, Albrecht, and Sarason 1994; Cohen 2004; Goldsmith and Albrecht 2011; Helliwell and Putnam 2004; House, Landis, and Umberson 1988; Thoits 2011). Some argue the same for certain human-nonhuman interactions, including interactions with pets (e.g., Blouin 2012; Charles 2014; Fox and Ray 2019; Laurent-Simpson 2021; Owens and Grauerholz 2019), spiritual entities (e.g., Çoruh et al. 2005; Jantos and Kiat 2007; Sharp 2012; Simão, Caldeira, and De Carvalho 2016; Townsend et al. 2002), and computerized office assistants and avatars (e.g., Choi and Drumwright 2021; Lambert et al. 2020; Lopatovska et al. 2019; Ostrowski et al. 2019; Pitardi and Marriott 2021; Pu et al. 2019; Sciuto et al. 2018). If people perceive living-deceased interactions as viable and important, can such exchanges provide benefits similar to those gained from human-to-human or other human-nonhuman interactions?
I explored the association between respondents who reported living-deceased perceived interactions and their self-reported physical health, mental health, satisfaction with social contact, and happiness. I measured these elements using questions modeled after those used in the General Social Survey.
Respondents rated their own physical health as very good, good, fair, or poor. The association between engaging in living-deceased perceived interactions and physical health was mixed. Interactors were more likely than noninteractors to report very good (18 percent vs. 13 percent) or fair health (27 percent vs. 23 percent) and less likely than noninteractors to report good (52 percent vs. 58 percent) or poor health (3 percent vs. 7 percent). Although the association is significant, the findings do not suggest a clear trend (χ2 = 7.66, p = .05; Cramer’s V = .150, p = .05).
The story for people’s self-reported mental health was different. Participants were asked: “Think about your mental health, which includes stress, depression, and problems with emotions. During the past month, how many days did you experience such issues?” Respondents indicating no perceived interaction with the dead were twice as likely than those who had such interactions to report zero mental health days (35 percent vs. 17 percent). Those who perceived living-deceased interactions were more likely to report 1 to 10 mental health days (58 percent vs. 44 percent) or more than 10 mental health days per month (25 percent vs. 22 percent). Thus, living-deceased perceived interaction was negatively associated with good mental health (χ2 = 21.01, p < .001; Cramer’s V = .198, p < .001).
Respondents reported their satisfaction with their social activities and relationships as being very satisfied, somewhat satisfied, somewhat dissatisfied, or very dissatisfied. Those who reported living-deceased perceived interactions were more likely than others to report being very or somewhat satisfied with their social activities and relationships (78 percent vs. 66 percent; χ2 = 8.99, p = .05; Cramer’s V = .18, p = .05). This finding reflects reports from open-ended questions, answers such as: “I always have someone to talk to” or “By talking to them and hearing them, I never feel alone.” Those with satisfactory social lives seem to consider their living-deceased perceived interactions part of their larger social support system.
Finally, I asked people to rate their general level of happiness using responses very happy, pretty happy, somewhat unhappy, and not at all happy. Those who reported living-deceased perceived interactions were more likely to be very or pretty happy than those who did not (74 percent vs. 66 percent; χ2 = 8.19, p = .05; Cramer’s V = .16, p = .05). This association confirms what answers to our open-ended questions show. Living-deceased perceived interactions made people feel more content and cheerful. Like exchanges that occur between copresent living beings, living-deceased perceived interactions appear to contribute, to some degree, to overall feelings of happiness.
With regard to happiness, my findings contradict those of Moor and de Graaf (2016). They found that people who lost a parent or child experienced more long-term unhappiness than those who suffered no such loss. My study was not confined to the loss of a parent or child but included the loss of spouses, partners, and friends. Moreover, I measured happiness at one point in time and not across the life course. Finally, my subjects were drawn from the U.S. population—a culture that values optimism and happiness more than European countries—Moor and de Graaf’s domain. These factors may explain the differences in the two studies.
Discussion and Conclusion
This study began by posing one question: Is it possible to interact with the dead? Roughly half of my respondents believe they can—a proportion similar to findings in recent national polls, surveys, and scholarly studies. In sociology, long-standing theories argue that living-deceased perceived interactions are not “real” because they lack important criteria. However, CB models and premises from two contemporary sociological theories—ANT and DM—suggest how and why such perceived interactions are formed and should be considered. I used these theories to guide my research.
My respondents displayed some of the practices reported by subjects in CB studies. However, my respondents saw their contacts as more than just internal dialogues, imaginative constructions, or one-way petitions. They described their encounters as two-way interactions similar to those they had with the deceased when they were alive. In concert with ANT, respondents see the deceased as capable of making things happen. Many respondents believed that the deceased heard or saw them, provided guidance, connected with them, or granted specific pleas. As DM theory argues, the living both construct a presence for the dead and forget the restrictions that discourage this (Owens 2007:577). And under certain circumstances, particularly those that present great need or urgency, the living experience the deceased as true interactants—independent and capable of response and action, be it guidance, strength, comfort, connection, or answering specific requests.
Participants reported benefits from living-perceived interactions, finding them emotionally positive and helpful; they felt they received guidance, strength, and true connection. Unlike those reporting no living-deceased perceived interactions, these respondents described higher satisfaction with their social lives and greater levels of happiness. However, living-deceased perceived interaction had no clear effect on respondents’ physical health and a negative effect on their mental health.
The data also suggest something about the types of people who perceive living-deceased interaction. Females and transgender individuals and those involved in religions that recognize an afterlife are most likely to engage in such interactions. In addition, those who engaged in living-deceased perceived interactions were also most likely to report interactions with pets, spirits and deities, avatars, and computer office assistants. Thus, for some, the idea of interaction is broad and can involve many types of nonhuman entities.
Of course, this study has limitations. For example, the data were collected via Qualtrics—a company that relies on self-selection and familiarity with the Internet. Despite my efforts to create a sample that reflects the characteristics of the U.S. population, certain groups are overlooked or over/undersampled because of Qualtrics’s methods. Second, although the over-65 group in my sample reflects the U.S. population at large, my sample includes a very small proportion of widows and widowers—much fewer than in the population at large. One might expect this group to engage more frequently in living-deceased perceived interaction. Therefore, the effects of widowhood on such exchanges may be underestimated in my study. Finally, I was unable to document the exact relationship between living-deceased interactants, nor did I discover the circumstances surrounding the deceased’s death. In CB studies, these factors sometimes influenced people’s inner dialogue with the dead (see e.g., Klass 1993, 2006; Neimeyer, Baldwin, and Gillies 2006) and, thus, deserve further attention.
Despite these limitations, these data suggest that nonhumans may play a more prominent role in social interaction than previously acknowledged and may deserve a more central place in sociologists’ analytic frame. The now renowned W. I. Thomas quote, “If men define situations as real, they are real in their consequences” (Thomas and Thomas 1928:571–72), can guide our thinking here. When are nonhuman interactants viewed as viable to exchange? Do human-to-human and human-to-nonhuman interactions differ in terms of norms, narratives, and processes? If individuals report real effects from living-deceased perceived interactions (or exchanges with other nonhuman entities), what patterns and rules govern the process? How are shared goals and understandings reached in human-nonhuman situations? One can imagine a variety of other elements that might deem the dead, or any other nonhuman entity, more or less viable in social interaction: the relative power of interactants, their relative command of resources, interactants’ social differences and similarities. In essence, these questions speak to the negotiation of meaning in varying social contexts—an area ripe for further research.
There are pragmatic issues involved as well. How do human-nonhuman interactions alter overall support systems or coping mechanisms? What are the full array of positive and negative effects that derive from such exchanges? We may also wish to more systematically explore whether this is a uniquely American phenomenon (Klass 2001; Lalande and Bonanno 2006). Currently, I am engaged in a replication of this study using Mexican and South Korean samples.
Overall, the findings on living-deceased perceived interaction and other forms of human-nonhuman interaction encourage us to rethink our definition of social interaction in general. What is the relationship between perception, physicality, and action? Are all three needed for true interaction, or does interaction exist on a continuum of some sort, making perceived interactions meaningful yet different from interactions occurring between copresent humans? This exercise seems especially important as artificial intelligence makes its way into everyday exchanges. We may also wish to consider that within interactions, as ANT and DM suggest, humans and nonhumans contribute to greater or lesser degrees—and that status may be dynamic rather than static. This makes the parameters of interaction more complex.
Despite its limitations, my study and the questions it raises suggest that we may wish to amend our definition of social interaction to realms beyond human-to-human exchange. By further exploring human-nonhuman exchanges, we may gain potentially important insights that have long been blocked from view. Broadening our lens in this way will keep our analysis of the social comprehensive and alive.
Footnotes
1
Seeking contact with the dead has deep historical roots. Consider the centuries of “bone washing” that occur on Dia de los Muertos (the Day of the Dead) or the spiritualism of the midnineteenth century with its séances and mediums. Recall the Victorian tradition that involved posing the dead with the living for one last family portrait (a practice still upheld in many hospitals when babies die at birth).
2
For more on classical theory assumptions, see Cerulo 2011.
3
In quoting authors or respondents, italics represent my emphasis.
4
Intercepts involve asking potential participants to take a survey for a reward while engaged in another activity—such as playing online games, reading news, and so on.
5
Only those who completed the entire survey were included in the study.
6
Again, my sample is not representative of the population at large. Thus, I offer statistical test results merely as descriptive information. All tests met the minimum requirement for the expected values of cells.