Positive Emotional Language in the Final Words Spoken Directly Before Execution

This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

Associated Data

Abstract

How do individuals emotionally cope with the imminent real-world salience of mortality? DeWall and Baumeister as well as Kashdan and colleagues previously provided support that an increased use of positive emotion words serves as a way to protect and defend against mortality salience of one’s own contemplated death. Although these studies provide important insights into the psychological dynamics of mortality salience, it remains an open question how individuals cope with the immense threat of mortality prior to their imminent actual death. In the present research, we therefore analyzed positivity in the final words spoken immediately before execution by 407 death row inmates in Texas. By using computerized quantitative text analysis as an objective measure of emotional language use, our results showed that the final words contained a significantly higher proportion of positive than negative emotion words. This emotional positivity was significantly higher than (a) positive emotion word usage base rates in spoken and written materials and (b) positive emotional language use with regard to contemplated death and attempted or actual suicide. Additional analyses showed that emotional positivity in final statements was associated with a greater frequency of language use that was indicative of self-references, social orientation, and present-oriented time focus as well as with fewer instances of cognitive-processing, past-oriented, and death-related word use. Taken together, our findings offer new insights into how individuals cope with the imminent real-world salience of mortality.

Keywords: mortality salience, language use, quantitative text analysis, emotional positivity, terror management

Introduction

Final words written or spoken shortly before death have fascinated people and have been collected in writing for a long time (e.g., Marvin, 1901; Brahms, 2010). Anything but banal word use when confronting death can offer researchers valuable insights into how people cope with existential threats and the human psyche in general (Pennebaker et al., 2003). Hence, final words in the form of rehearsed or impromptu sayings spoken by a dying person can reveal how individuals emotionally regulate the salience of imminent mortality. Here, we analyzed emotional language use in the final statements of executed Texas death row inmates and compared our findings with rates of positive emotion word use in general and in contexts involving contemplated death and attempted or actual death by suicide.

Intuitively, one might imagine that thoughts of one’s own death should evoke fear and anxiety as death may be associated with a broad range of frightening aspects (i.e., pain, loss of loved ones, unfulfilled goals; e.g., Niemeyer and Moore, 1994; Florian and Mikulincer, 1997). Psychological denial of death to escape the anxious awareness of our mortality constitutes one of the most basic drives in individual behavior (Becker, 1973). According to Terror Management Theory (TMT; Greenberg et al., 1986), individuals employ a wide range of cognitive and behavioral efforts to regulate the anxiety that mortality salience evokes (Greenberg et al., 1997; Pyszczynski et al., 2004). These psychological defense mechanisms are aimed at maintaining self-esteem and acquiring meaning in life (Pyszczynski et al., 1999).

Empirical findings by DeWall and Baumeister (2007) suggest that an automatic orienting toward emotionally positive information and associations serves as a way to protect and defend against mortality salience. In a series of studies, the authors showed that thinking about death (vs. thinking about dental pain) activated a non-conscious emotional coping response that was counterintuitive to the overt emotional distress one might expect, such that participants completed ambiguous word stems with relatively more positive emotion words and favored positive emotional associations in judgments of word similarity. Also, findings by Kashdan et al. (2014) indicated that such a shift toward the use of positive emotion words may be involved in regulating the fear of death: writings of individuals contemplating death (vs. people thinking about dental pain, uncertainty, or meaninglessness) showed an increased use of positive emotional language.

Although these studies offer important insights into the psychological dynamics of mortality salience, they are limited by the fact that the situation created by letting university students contemplate death in a standardized lab situation necessarily differs in several aspects from real life and real death. Consequently, in an attempt to better understand the feelings of individuals prior to their imminent actual death, researchers have conducted content analyses of suicide notes (Tuckman et al., 1959; Handelman and Lester, 2007) and other writings left behind after suicides (e.g., diaries, poetry; Stirman and Pennebaker, 2001; Pennebaker and Stone, 2004). Tuckman et al. (1959) reported that the emotional content of 165 analyzed suicide notes was surprisingly positive and contained expressions of gratitude and affection. Similarly, research has shown that notes from completed suicides contained more positive emotional language than notes from attempted suicides (Handelman and Lester, 2007) and that writings that occurred temporally closer to death were higher in positive emotional language use (Pennebaker and Stone, 2004).

Without any doubt, the psychological “terror” felt in the situation of self-decided death by suicide is extreme. However, there may be one situation where individuals face an even greater amount of terror: directly before death by execution. This situation is characterized by a complete absence of controllability and a maximal subjection to powerful others who have the right to end one’s life. Consequently, last words as part of the execution process, visible as early as 1388 in England (Howell, 1809), provide a unique opportunity for exploring the predictions of TMT.

An additional theoretical perspective that can help shed light on emotional regulation as reflected in final statements before execution is provided by socioemotional selectivity theory (SST; Carstensen et al., 1999). According to SST, the perception that one has limited future time (e.g., caused by a temporal proximity to the end of one’s life) is postulated to increase individuals’ motivation to derive emotional meaning by prioritizing close social relationships over instrumental or knowledge-related goals (e.g., Carstensen et al., 1999; Carstensen, 2006). Research has provided evidence that individuals who perceive their future time as limited or who have limited actual life expectancies strongly value emotionally close social partners (e.g., Carstensen and Fredrickson, 1998; Fung et al., 1999, 2001; Fung and Carstensen, 2006). Moreover, as individuals recognize the inevitable constraints on time that are imposed by one’s mortality, which is not necessarily associated with chronological age for death row inmates, the focus on the present moment is likely to increase (Carstensen et al., 1999).

In the present research, we analyzed the actual final statements made by executed death row inmates in the state of Texas. This unique data set provided the opportunity to analyze how people cope with immediate death in a standardized, uncontrollable situation in a large sample. Prior research found some indications of positivity by using qualitative content coding, implemented by humans, of the prevalent thematic categories found in Texas death row inmates’ last statements, the most prominent themes being the expression of love or appreciation (Heflick, 2005; Schuck and Ward, 2008). Further themes that were identified included addressing others; seeking forgiveness; expressing self-comfort, wishes, or hopes; and religious references (Schuck and Ward, 2008). In contrast to these previous studies, we used computerized quantitative text analysis (Mehl, 2006) as a more objective measure of emotional language use. In the following, we outline the important advantages of the methodology we applied and the relevant research questions that we were thus able to address.

Quantitative text analysis can be differentiated with regard to “what” (i.e., content) a person is saying and “how” (i.e., style) the person is saying it, whereby “how” the person says something can reveal more subtle aspects of communication (Mehl, 2006). Issues that occur with human coders consist of a hastiness to ascribe meaning and a general inability to monitor word choice (e.g., Hart, 2001). Thus, given these issues and the fact that different content themes (e.g., appreciation/love; expressing self-comfort or wishes) may be expressed through the same linguistic style (e.g., emotion-related words), a computerized word-count-based text analysis allows researchers to objectively and reliably assess (i.e., count) the linguistic features of individuals’ statements and detect subtle linguistic styles (Pennebaker and King, 1999; Pennebaker et al., 2003; Mehl, 2006; Tausczik and Pennebaker, 2010). Subsequently, the linguistic features that are assessed (e.g., percentage of positive emotion words in death row inmates’ final statements) can be (a) compared with word usage rates in samples contemplating death or experiencing a limited time horizon due to a decision to commit suicide, (b) related to the demographic characteristics of the sample of death row inmates (e.g., age at execution, years on death row, or educational level), and (c) associated with other linguistic markers (e.g., pronoun use, verb tense, word use indicative of psychological processes such as social orientation or cognitive processing).

We hypothesized that if tuning in to emotional positivity acts as a psychological mechanism aimed at coping with the threat of mortality (DeWall and Baumeister, 2007; Kashdan et al., 2014), it would be reflected in the emotional word use in death row inmates’ final statements such that death row inmates would use a higher proportion of positive than negative emotion words. We further hypothesized that the immense existential threat that real-life executions evoke would lead to a higher proportion of positive emotion words compared with word usage base rates (Pennebaker et al., 2007b) as well as compared with the words of individuals contemplating death (Kashdan et al., 2014) or individuals attempting or committing suicide (Handelman and Lester, 2007). Drawing on postulations by SST (Carstensen et al., 1999), we further aimed to explore the relations between positive emotional word use and language indicative of self-references, social orientation, cognitive processing, time orientation, and personal concerns with religion and death. We expected that emotional positivity in executed death row inmates’ final statements would be associated with a greater use of social-orientation words and present-tense verbs. Furthermore, we conducted an exploratory investigation of associations between positive emotional language use and self-references, cognitive-processing word use, past-tense and future-tense verb use, as well as religion and death-related word use.

Materials and Methods

Sample

To test our hypotheses, we analyzed death row inmates’ final statements, spoken minutes before their executions in the US state of Texas. The final statement transcripts as well as death row inmates’ demographic information were provided publicly available on the Texas Department of Criminal Justice (2015) website. Information about whether death row inmates abandoned their appeals and volunteered for execution was provided publicly available on the Death Penalty Information Center (2015b) website. As we only reanalyzed these publicly available data on a de-identified and anonymized aggregate level, a particular ethical approval for our study was not required. Of the 527 death row inmates executed between December 1982 and June 2015, transcriptions of spoken last statements were not available for 119 death row inmates. Of these 119 death row inmates, 108 were listed as having declined a last statement, eight had no statements provided, and three had only written statements provided. Table ​ Table1 1 presents demographic information consisting of age at time of execution, age at incarceration, years spent on death row, and educational level (i.e., highest grade completed) for death row inmates with and without transcribed spoken final statements. Of the 408 death row inmates with transcribed spoken statements, one was excluded from analyses due to a reported higher age at incarceration compared to age at execution on the Texas Department of Criminal Justice (2015) website. The analyzed sample of death row inmates for whom spoken last statements were provided who were executed between 1982 and 2015 consisted of N = 407 (404 male, three female) individuals; 178 (43.7%) death row inmates were reported as White, 150 (36.9%) as Black, 77 (18.9%) as Hispanic, and two (0.5%) as another ethnicity. Death row inmates with no transcribed spoken last statement on record did not differ from the analyzed sample of 407 death row inmates in ethnic background, χ 2 (3) = 1.25, p = 0.740, Cramer’s V = 0.05, or on the other demographic variables, all ps > 0.09 (see Table ​ Table1 1 ).

Table 1

Demographic information on death row inmates executed between 1982 and 2015 (June 30) in Texas.