Source: Graphic generated using the free version of the Stable Diffusion latent text-to-image diffusion model (https://stablediffusionweb.com/)

When I deconverted from Fundamentalist Evangelical Christianity in 2017, one of the things I found hardest was reckoning with the fact I was no longer going to live forever. That might sound crazy, but I had been raised to believe that I would. In the autumn of 2017 I became the worst person to talk to at a party because all I wanted to talk about was the fact we were all going to die one day.

Then I read Sheldon Solomon’s interview with Sophie Roell on fivebooks.com. In the interview Solomon contends that Ernest Becker achieved an important development to our understanding of fear of death. Becker was an American anthropologist who died in 1974 and wrote the books The Denial of Death and Escape From Evil. Solomon says that Becker’s ideas about the fear of death being a massive unconscious influence on us was one of the main inspirations for him choosing to study death anxiety. Solomon and his co-authors Tom Pyszczynski and Jeff Greenberg have gone on to see if they can empirically prove Becker’s theories, calling their ideas Terror Management Theory (TMT). They have also written a book about TMT called The Worm at the Core.

What I found compelling about Becker’s writing, and TMT, was the idea that anxiety about fear of death was not benign. TMT holds that human beings knowing we are gong to die produces existential terror (Greenburg, Pyszczynski, & Solomon, 1986). The theory says people seek to counter this unconscious fear through symbolic immortality. This means things like, for example, living on after death through the things we achieve professionally, or by associating our identity with culture that will live on after we die.

I was intrigued and captivated. Becker’s theories made intuitive sense to me. And Solomon was saying that “[t]here are now upwards of 1000 of these studies [that have been] done“. I was afraid of dying and I was afraid of the nonexistence that came after. And I didn’t like the idea of being unconsciously controlled by fear of death. I was very interested that TMT held the promise of both understanding my death anxiety better and potentially finding more peace in the face of my future expiration.

One of the main ideas that has sprung out of TMT is the mortality salience (MS) hypothesis. This hypothesis suggests that an outcome of feeling death anxiety is that people are more likely to defend their cultural worldviews. For example, if nationalism is one way a person is trying to achieve symbolic immortality (“I am part of this great nation which will live on after I die”), when encountering reminders of death the same person may be inclined to unconsciously increase their defence of their nationalism and also possibly disparage other countries/cultural worldviews.

There was something about the “underpinning of everything”-ness of TMT that was deeply attractive to me. It felt like the dark shadow of humanism; yes we had the gift of being alive, but not only did the end always lie in wait, there was this evidence that unconsciously fear of death troubled us more than we wanted to admit and was negatively affecting how we were living. I thought TMT could help me to live a better life less afraid of death so that I might die a better death.

At the same time though I was worried I was falling for TMT too hard, too quickly. Was I looking for a worldview that explained everything to replace my Fundamentalist Christianity? There were times when I felt like all I could see was mortality salience playing out in every aspect of our lives. I was worried I might have reclaimed myself from one ideology only to find myself giving it over to another.

“While some of the foundational studies on which TMT is based have failed to replicate, thereby drawing criticism within the field of psychology, the framework continues to resonate for many.”

Source: https://www.psychologytoday.com/gb/basics/terror-management-theory

I googled “Terror Management Theory” and found myself on psychologytoday.com which says “[w]hile some of the foundational studies on which TMT is based have failed to replicate, thereby drawing criticism within the field of psychology, the framework continues to resonate for many” (retrieved on December 7, 2022, archive.ph capture | Google cache capture). What did ‘failed to replicate’ mean? Was none of it true? I sat at my computer trying to think of ways to justify ignoring this new information about failure to replicate so I could keep believing in TMT.

I decided to look up critical reviews of The Worm at the Core on goodreads.com, the book review website, secretly hoping to find only positive reviews. One review said: ‘I felt the book suffered from too many claims and assertions that, while not being necessary for the integrity of the rather convincing underlying thesis, seem to be somewhat suspect or at least poorly grounded.’ Another review: ‘Too much of this premise is taking place behind the scenes, which is a risk you run in all social psychology experimentation, but that doesn’t mean your extrapolations qualify as proof.’ A person named B. Rule said that ‘[t]his book is really a restatement of Ernest Becker’s ideas, and the value-added, if you will, is that the team of authors have supposedly provided empirical support for his theories. Several of their experiments are described throughout the text. While several were interesting, there were many where I was highly dubious that the lines of causation were as clear as the authors proposed. They just didn’t seem very rigorously designed to eliminate other variables or preclude other interpretations’. 

The critical reviews for The Denial of Death were no better. Someone called Tyler said that Becker’s claims to scientific proof were ‘pseudoscience’. Another review called Becker’s work ‘outdated psychobabble’ and ‘closer to medieval scholasticism’. A person called Gary Beauregard Bottomley summed up their review of Denial of Death this way: 

‘I can highly recommend this book since it gives such an interesting window that psychoanalysis mistakenly provided to human understanding in 1973. It clearly gives a great peak [sic] into how psychiatry got off the rails. I would highly recommend reading “Shrinks: The Untold Story of Psychiatry” before attempting this pseudo-scientific book. “Shrinks” documents how psychiatry got so far off the rails and how it found itself by becoming a real science by including the empirical. This book, “Denial of Death”, marks the start of the beginning from which a new era for human understanding began to finally find itself and jettison junk like this book contains.’

Psychobabble and pseudoscience. I worried I had been too quick to sign on to TMT. And to make matters worse, I had started this very blog about fear of death. I groaned internally, how embarrassing.

Then I read this review of The Worm at the Core from 2020 by someone called Travis Rebello. Rebello talks about how they had once given a presentation about denial of death ideas including those of Terror Management Theory. Rebello said one of the professors had said the problem with TMT was “replicability”. Rebello then goes on to list examples of studies failing to replicate the results of experiments originally used to empirically support the validity of TMT.

I felt puzzled. What was going on here? How could ‘upwards of 1000’ studies be getting carried out on a theory that couldn’t even be replicated?

One study was particularly interesting. It was called Many Labs 4 (ML4), first published in 2019, and was specifically about trying to replicate an original TMT study from 1994 whose co-authors included the three authors behind Worm at the Core: Solomon, Greenburg, and Pyszczynski. Many Labs was so-called because of the participation of 17 labs around the world with 1578 participants, meaning a decent sample size to replicate or not replicate the original findings. Not only that, a group of half the labs involved with ML4 would have expert input from one of the original authors of the 1994 study (Tom Pyszczynski) in order to test whether the “secret sauce” of the original researchers had an outsized effect on results.

The result? From one of the lead authors of ML4: “ultimately neither group successfully replicated the effect”. Pretty open and shut. But if this was true, I still didn’t understand why there were thousands of studies as Solomon claimed? I didn’t know enough about academia to know how common it was for an official handbook to be written for a theory that had essentially been debunked. I glumly listened to the ML4 episode of the Black Goat podcast where one of the hosts declared TMT “dead in the water”. It seemed the “replication crisis” for psychology studies (retrieved on December 7, 2022, archive.org capture) I had heard about was real.

I took to Twitter to see if there was any commentary on the ML4 paper. It turned out in 2020 a paper called ‘A Word of Caution about Many Labs 4’ had been published which reanalysed the ML4 data and concluded the studies had not been well-conducted and, most importantly from my point of view, the conclusions of ML4 had been wrong: an effect had been found!

Source: https://twitter.com/G_Hirschberger/status/1225737078291468288

So where did things stand now? A successful TMT study from 1994 had been chosen to attempt to be replicated by a large modern study. ML4 couldn’t replicate the 1994 findings. But the authors of a reanalysis said that ML4’s own data showed findings of a ‘modest effect size’. Modest effect size didn’t sound like “dead in the water” but it also didn’t sound like a slam dunk either.

Still, I thought modest effect size was certainly different from the ‘neither group successfully replicated the effect’ that ML4 was concluding. I could see online that ML4 had been republished in November 2020 but when I compared the changes to the 2020 version to the 2019 version the conclusions were still the same: ’we observed little evidence that priming mortality salience increased worldview defence compared to a control condition’ (p. 30)  and ‘with these protocols, in the context of these Labs and time in history, we find little support for this key finding of TMT’ (p. 30).

For a non-academic like myself, it was hard to know which conclusion from the data I should agree with.

---

I had read earlier this year about how meta-analysis studies can be a good barometer of the state of an academic field. Meta-analysis is essentially a study of a group of studies. I had come across a recent meta-analysis of TMT literature out of the University of British Columbia from January 2022. Here’s what they concluded:

‘[O]ur main results suggest that tests of the MS [Mortality Salience] hypothesis are generally underpowered in the literature’ (p. 23).

Translation: These TMT test results are not strong.

‘[W]e conclude there is evidential value in the overall literature, and there are non-zero effects underlying the studies on a whole.’ (p. 23)

Translation: But, there is something there.

‘The pattern of increasing power after the field began to take QRPs [Questionable Research Practices] seriously bodes well for TMT’ (p. 30)

Translation: No more cheating you guys.

‘Our findings are therefore more optimistic about the MS [mortality salience] hypothesis than failed replications pointing to a null effect.’ (p. 38)

Translation: E.g. ML4.

‘[T]he overall weak effects found in studies that test the MS hypothesis would seem to call into question early articulations of TMT as an organising principle that explains the origins of such foundational human characteristics as culture and self-esteem (see Greenburg et al. 1986; Solomon et al.,1991). If effect sizes can serve as a guide for interpreting the field’s conclusions about which psychological phenomena are most foundational, the evidence is mounting against these early claims that many behaviours emerged as a product of efforts to suppress death thoughts.’ (p. 41)

Translation. TMT may not be as complete an explanatory theory as Dave Underwood is hoping it is.

The study concludes:

‘It is worth noting that our p-curve analysis focuses on the MS hypothesis, derived from TMT, but is not itself a test of the overarching theory. For this reason, many phenomena may yet have an existential origin that are not adequately tested through the MS hypothesis.’ (p. .42)

‘While the recent publication of several failed replications of TMT studies may have led some to expect that our p-curve analysis of the TMT literature would result in an autopsy of the field, we have instead detected a clear pulse. The overall state of the field may not be as healthy as its most enthusiastic champions have suggested, but it is alive, and we hope future researchers design their studies with the results of this p-curve analysis in mind’. (p. 42-43)

---

When a layperson such as myself wanders into an area of academia he knows nothing about, I think it is easy to want to draw a line under things. I think this is especially true when, as in my case, someone is associating their identity and understanding of the world with a theory or area of research.

Can you have “hope” in a theory? Or is that too much like having “faith” in a theory?

The other day I watched a video about how we had Newton’s theory of gravity, then Einstein’s theory of relativity, and now we have Quantum Mechanics, each theory improving on the one that came before it. Is TMT pseudoscience? I don’t think so, but let the evidence fall where it will. Could TMT be superseded by a better theory one day? Yes. In fact, I hope it is. May the march of progress, and the scientific method, march ever forward.

And as for me? I am keen to continue thinking & writing about the fear of death. But I feel like I’ve learnt an important lesson about treating books about theories as sacred texts of quasi-belief systems. They are not the same, and thank Zeus for that.