Contribution
The paper turns the old claim that being in touch with one’s true self makes life meaningful into a testable social-cognitive one. It does not measure the true self, which “might be difficult, if not impossible, to operationalize directly” (p. 475). It measures the true self-concept, “a cognitive schema representing those aspects of the self that are considered, by the person, to be most emblematic of his or her true nature”, and its accessibility: how readily it comes to mind. It then asks whether accessibility predicts, and whether manipulating it changes, reported meaning in life. The actual (everyday, public) self-concept is the control throughout, so that effects belong to the true self-concept and not to the self-concept in general.
Key insight
The true self is proposed as “a wellspring for meaning by exporting feelings of legitimacy, importance, and value to other aspects of life” (p. 474). If so, simply activating it should raise felt meaning, “even in the absence of the phenomenological experience or actual expression of one’s true self-concept” (p. 486). And the data say it does, without the activation making people like themselves more: the true self’s weight lies in its being taken as who one really is, “regardless of how much we like it” (p. 486).
Assumptions
- Accessibility as speed. Faster “me” responses to a trait mean the trait’s self-concept is more accessible (Bargh et al. 2002). Speed is residualized on each person’s speed for control traits they called “me”.
- The true/actual split is the participant’s. Studies 1–3 define the true self as traits one would like to express but can express only around those closest to one, and the actual self as traits often expressed socially (after Bargh et al. 2002 and Rogers 1951). Studies 4–5 define the true self as “who you believe you really are, even if you sometimes act in different ways” and the actual self as “who you are during most of your daily activities”.
- Meaning is presence of meaning, self-reported: the five-item MLQ Presence subscale, with four PIL items added in Studies 3 and 5.
- Subliminal primes are unseen. 40 ms (Study 3) or 35 ms parafoveal (Study 5) masked presentations; awareness was probed after the task.
Key results
- Study 1 (N = 56 after 3 suspicious participants were dropped; traits chosen from a 60-word list, true and actual sets forced to differ). True-self accessibility predicted meaning (β = .29, p < .05) with actual-self accessibility, affect, need satisfaction (β = .37) and the traits’ social desirability controlled; actual-self accessibility did not (p. 477). True-self traits were rated less socially desirable than actual-self traits, t(55) = 4.14 (p. 477).
- Study 2 (N = 68; self-generated traits; meaning measured about 1.5 months later). True-self accessibility β = .24, need satisfaction β = .32 (p. 479). No desirability difference between the two sets.
- Study 3 (N = 80; priming one month after trait listing). Participants liked their true self more than their actual self (5.81 vs 5.29, t(79) = 3.55). True-self primes gave higher meaning than actual-self primes (5.59 vs 4.72, t(78) = 4.16), and the difference held in an ANCOVA with Time 1 meaning, affect and both liking ratings controlled, F(1, 73) = 4.46, p < .05 (p. 481). No participant identified the primes.
- Study 4 (140 recruited, 123–124 analysed after 17 were excluded for not following instructions; overlap between the two sets allowed, mean 4.05 of 10 shared). In a regression with nine covariates (R² = .56), overlap (β = .21), AI-3 authenticity (β = .34) and true-self accessibility (β = .19) predicted meaning; actual-self accessibility did not (p. 483). Table 4 (N = 124): true- and actual-self accessibility correlate r = .70; true-self accessibility correlates .05 with authenticity, .19 with state self-esteem and .30 with meaning.
- Study 5 (155 recruited, 149 analysed; a 2 × 2 of true vs actual self by liked vs disliked traits). Estimated marginal means for meaning: positive true 5.16, negative true 5.23, positive actual 5.29, negative actual 4.91. The planned contrast of negative-actual against the other three was F(1, 139) = 5.07, p < .05, and negative-true beat negative-actual, F = 4.63. Priming affected neither state self-esteem (ps > .40) nor reported authenticity (ps > .34) (p. 485).
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Individual differences in true-self (not actual-self) accessibility predict presence of meaning in life | moderate: three samples, small effects (β ≈ .2–.3), consistent | Studies 1, 2, 4 |
| C2 | Making the true self-concept accessible raises reported meaning | moderate: two priming experiments, with baseline meaning controlled; Study 5’s omnibus test is misprinted (see Corrections), the planned contrasts carry it | Studies 3, 5 |
| C3 | The effect does not depend on the true self being liked or positive | moderate: Study 5 disliked true-self traits as effective as liked; Study 1 true-self traits less desirable | Studies 1, 5 |
| C4 | True-self accessibility is distinct from self-reported authenticity and from need satisfaction | moderate: each predicts meaning independently, r = .05 with AI-3 | Study 4; p. 487 |
| C5 | Priming the true self does not raise state self-esteem | moderate for the experimental claim (Study 5, ps > .40); the correlational claim for Study 4 is contradicted by its own Table 4 (see Corrections) | Study 5; Table 4 |
| C6 | The true self serves as a “hub of meaning” that exports value to relationships, goals and work | theoretical framing, not tested: only the accessibility → meaning link is | pp. 474, 486 |
Concepts
- true self-concept: the person’s avowed true self, a schema of the traits they take to be “most emblematic of his or her true nature” (p. 475); not the true self of personality theory.
- actual self-concept: in Studies 1–3, traits “often able to express to others in social settings”; in Studies 4–5, the “everyday self”. The authors note it “might best be thought of as a public self” (n. 1).
- true self-concept accessibility (TSCA): reversed standardized residual of mean log reaction time to true-self traits, regressed on reaction time to control “me” traits.
- meaning in life: here only presence of meaning (MLQ), not search for it.
Connections
- Kernis and Goldman (LIT-574). The paper describes their programme as SDT-based and defines authenticity, after them, as “the unimpeded functioning of one’s true self in daily life” (p. 474). It measures their AI-3 (Studies 4–5) and finds it unrelated to TSCA (C4). It reads its disliked-trait result as fitting their claim that authenticity includes “awareness and acceptance of both one’s positive and negative attributes” (p. 487).
- SDT (LIT-559). Need satisfaction is the standing rival predictor of meaning, and the autonomy items of the Basic Psychological Needs Scale include “I feel like I can pretty much be myself in my daily situations” (p. 476), an authenticity item inside the autonomy measure.
- THEORY-054 (autonomy as graded integration). The overlap between true and actual self-concepts predicted meaning on its own (β = .21) and correlated with authenticity, need satisfaction and true-self accessibility (Table 4). Overlap is a crude, trait-level index of how far who one is every day matches who one takes oneself to be, which is close to what SDT means by integration. It is a better candidate for a non-endorsement measure of integration than the self-reported reasons SDT uses, though the paper does not frame it that way.
- THEORY-029 (ownership cannot come from order alone). The paper’s “true self” is whatever the participant says it is; it measures the psychological weight of avowed ownership, not whether the avowal is correct. Its authors say so: the person’s “assessment of the contents of the true self may not coincide with a person’s ‘real’ true self” (p. 475). So it neither supports nor tests the philosophical question.
- Sheldon & Elliot (LIT-589). The same idea, goals that express the true self, reached through reasons for goals rather than through a self-concept; Sheldon is thanked in the acknowledgements.
- Meaning (LIT-183). Subjective meaning only; Wolf’s objective condition is outside the measure.
- Taylor (LIT-594). The paper’s own historical framing (the “existential vacuum” after shared value bases eroded, pp. 474, 487) matches Taylor’s account of why authenticity became an ideal; it cites Baumeister, Bellah, Becker, Frankl and Fromm for it.
Bearing on the record
- Candidate THEORY, Proposed: “Activating what a person takes to be their true self raises their felt meaning in life, independently of whether they like those traits”, sourced to this paper. It should promote on an independent, preregistered replication of the priming effect, given that it rests on two subliminal-priming experiments of the 2005–2009 period, a literature whose replication record has since been widely questioned (my judgement, not the paper’s), and one of them with a misprinted omnibus test. It would be refuted by a well-powered failure to replicate Study 5’s negative-true vs negative-actual contrast. Not filed as stated: the record filed C4 instead, as THEORY-074, whose promote_when counts a replication of the priming effect as bearing on it.
- Against conflating authenticity measures. C4 is the record’s first evidence that implicit true-self accessibility and self-reported authenticity come apart. Any THEORY that treats “authenticity” as one construct should cite it. THEORY-074 now states it.
- No ML instruction. Nothing here belongs in the anthology.
Limitations
- Self-reported meaning only, which the authors acknowledge (p. 488).
- Students in one lab, one point in life.
- Small effects and modest samples (N = 56, 68 in Studies 1–2) for regressions with six or seven covariates.
- Subliminal priming effects of this vintage have a questioned replication record (my judgement); no replication is reported here, and Study 5’s main effect of self type was null (ps > .18), with the result resting on the interaction contrasts.
- The control is not neutral. There is no no-prime condition, so footnote 9 uses change scores to argue that true-self primes raised meaning and negative actual-self primes lowered it, but conditions differed at baseline.
- Study 4’s true and actual accessibility correlate .70 where Studies 1–2 found them unrelated; allowing the two trait sets to overlap (4 of 10 words shared on average) makes the two scores partly the same reaction times. The regression still separates them, but the independence the earlier studies claimed does not hold in this design.
Open questions
- Does TSCA predict anything other than self-reported meaning, for instance behaviour or persistence?
- Is overlap between true and actual self-concepts a usable measure of integration in SDT’s sense?
- Would a neutral-prime control show that true-self primes raise meaning, rather than negative actual-self primes lowering it?
Corrections
- none to a seeded skim (there was no seed)
- Study 5 interaction (p. 485). The page image prints “a significant interaction between type of self-concept and valence of self-concept, F(1, 139) = 1.63, p < .05”. An F of 1.63 on (1, 139) degrees of freedom has p ≈ .20. Either the F or the p is misprinted, and the stated significance of the omnibus interaction cannot be taken from the paper. The planned contrasts (F = 5.07 and 4.63) are reported separately and are internally consistent.
- State self-esteem in Study 4. The General Discussion says Study 4 showed that true-self accessibility was “unrelated to state self-esteem” (p. 486), but Table 4 gives r = .19, marked p < .05. The experimental claim for Study 5 is unaffected.
- Study 3 table reference. The Study 3 results refer to “Table 3 … in Study 5” (p. 480); Table 3 is Study 3’s.
- Study 4 N. 140 recruited minus 17 excluded is 123; Table 4 says 124.