The Hylomorphism Inventory (HI): Theoretical Foundations and Validation of a Scale Measuring Folk Beliefs Congruent with Hylomorphism

Abstract

The article introduces the Hylomorphism Inventory (HI), a new instrument designed to measure lay beliefs about the soul–body relationship that are congruent with the Aristotelian–Thomistic framework of hylomorphism. Although research on intuitive ontology has predominantly focused on dualist and monist models, the hylomorphic perspective—central to Catholic anthropology yet difficult to articulate in everyday cognition—remains largely unexplored. Drawing on research in intuitive anthropology, we conceptualize hylomorphic beliefs as endorsing the human person as a psychophysical unity in which the soul functions as the organizing form of the body. Using a theory-driven approach and expert evaluation, we developed an initial 10-item scale and tested it in a nationwide online sample of Polish adults (n = 407). Exploratory (EFA) and confirmatory factor analyses (CFA), supported by nonparametric Mokken scaling, converged on a primarily unidimensional 9-item solution with high internal consistency (α = 0.89, ordinal α = 0.91, ω ≈ 0.90). Validity analyses revealed that HI scores were strongly associated with beliefs emphasizing the integration of body, mind, and soul, but only weakly related to their mere endorsement as components. This pattern suggests that what distinguishes hylomorphism at the psychological level is not belief in the soul per se, but belief in the unity of the human person. The HI provides a parsimonious tool for differentiating lay anthropological models and enables empirical investigation of how hylomorphism-congruent beliefs relate to moral reasoning, spiritual practices, and broader psychological functioning.

1. Introduction

1.1. Folk Anthropology and Variability of Anthropological Intuitions

The study presented in this article introduces the Hylomorphism Inventory (HI), a scale designed to measure beliefs about the relationship between the soul and the body that are consistent with a hylomorphic understanding of the human person. The development of this tool constitutes part of a broader research program aimed at examining the psychological significance of beliefs about the soul (Fortuna and Wróblewski 2024).

Although the concept of the soul was largely removed from the domain of scientific psychology in the early twentieth century (Schlosberg 1961), it remains a robust and functionally relevant element of everyday cognition (Hillman 1978). As Kugelmann (2013, 2023) argues, the notion of the soul continues to structure human self-understanding, language, and cultural practices. This is reflected both in its linguistic persistence (Wierzbicka 1989) and in empirical findings showing that a substantial proportion of adults across cultures endorse belief in the soul (Anglin 2014; Gut et al. 2021; Halman et al. 2008; Richert and Harris 2008). Moreover, both children and adults spontaneously distinguish the soul from related constructs such as the mind (Harris and Corriveau 2020). Thus, the postulate of a “science without the soul” does not entail either a “world without the soul” or “human life without the soul.”

Within a genetic framework, three broad modes of conceptualizing the soul can be distinguished: (a) folk (D1), grounded in commonsense psychology and first-person experience; (b) religious (D2), rooted in mythic and theological traditions; and (c) philosophical (D3), developed within formal theoretical systems (Fortuna and Wróblewski 2024). These modes differ in their level of abstraction, internal coherence, and epistemic justification, and they need not align with one another in lay cognition.

Within commonsense cognition, beliefs corresponding to the folk level (D1) constitute part of a broader system of anthropological intuitions that together constitute folk anthropology (Cohen and Barrett 2011). This domain of commonsense cognition enables individuals to construct an intelligible understanding of human nature in ontological terms. Such beliefs, together with intuitions about other aspects of human structure, form a complex configuration referred to as an anthropological mindset (Fortuna et al. 2023b), which supports category-based reasoning about what it means to be human (cf. Weisman 2022). For instance, individuals may express intuitions such as: “The soul is ‘me’, because ‘I’ am not the body—when I die, the body will remain, but I will not be there” (Czaja 2006, p. 241).

These lay representations do not require conceptual precision or internal coherence from a philosophical standpoint and may differ substantially from their formal theoretical counterparts. Although they are not necessarily theoretically consistent, they remain psychologically functional and play a significant role in structuring human experience. References to the notion of the soul emerge particularly in contexts involving moral evaluation and existential threat, where they are used to describe experiences of dehumanization, suffering, and violations of dignity (Landry et al. 2024; Markowitz and Slovic 2021). They also inform moral judgments, such as the attribution of moral status or positions on bioethical issues (Richert and Harris 2008; Fortuna et al. 2023a, 2023b), and are associated with preferences for activities conducive to well-being (Burgmer and Forstmann 2018; McGhee et al. 2024).

At the same time, anthropological mindsets involving the soul exhibit substantial variability. While their developmental foundations are often linked to early dualistic intuitions, there is ongoing debate as to whether humans should be regarded as “natural born dualists” or whether their ontological intuitions are more flexible and culturally shaped (Bering 2006; Bloom 2004). In this respect, available evidence suggests that intuitive anthropology is not uniformly dualistic but instead reflects diverse and culturally influenced models of intuitive ontology (Barlev and Shtulman 2021). Empirical research has identified not only intuitive dualism but also intuitive physicalism, in which the soul is treated as a metaphor for psychological functioning (Czaja 2006), and intuitive trialism, in which body, mind, and soul are differentiated as distinct components (Gut et al. 2021; Lindeman et al. 2015; Richert and Lesage 2019).

Within this diversified landscape of beliefs, the near-complete omission of the hylomorphic perspective in empirical studies is particularly striking. On the one hand, this absence is surprising, given that hylomorphism constitutes the anthropological backbone of Catholic doctrine. On the other hand, it is understandable from a methodological standpoint: hylomorphic beliefs are difficult to capture empirically, as they do not fit neatly into the dominant dualist–monist distinction. The soul is understood neither as a separate substance nor as a mere metaphor of corporeality, but as the form that organizes and actualizes the living body. Consequently, such beliefs may be systematically misclassified in both questionnaire-based and qualitative research as either dualist (emphasizing separation) or monist (emphasizing unity).

This limitation has important consequences for empirical research, as it hinders the reconstruction of the full spectrum of beliefs underlying commonsense thinking about human nature and may lead to erroneous conclusions, particularly in culturally homogeneous populations such as Catholic societies, whose internal diversity has not yet been precisely mapped.

1.2. The Missing Perspective: Hylomorphism

To clarify and adequately capture this underrepresented form of anthropological intuition, it is necessary to introduce the conceptual framework of hylomorphism. In the present study, this framework is not treated as a philosophical doctrine to be mastered by respondents, but as a theoretical reference point for identifying lay beliefs that may be cognitively consistent with some of its core assumptions.

Hylomorphism (from Greek hýlē—matter, morphḗ—form) is a philosophical theory of the ontic structure and functioning of natural bodies, that is, entities that are neither human-made nor mathematical objects (Shields 1988). Its central claim was formulated by Aristotle (2011) in De Anima: natural substances are composed of prime matter (pure potentiality) and substantial form. These two principles do not exist independently but constitute a being only in mutual unity. Form functions as the principle of actualization: it determines what a given entity is, providing qualitative specification, essential properties (e.g., humanity), unity, and absolute identity that is independent of external factors. Matter, in turn, is what receives this specification—it is the substrate understood as the principle of potentiality underlying all things and all natural change (see Ainsworth 2024; Bostock 2006; Piwowarczyk 2020).

The paradigmatic example originally used to illustrate this composition was the living organism. In this case, substantial form is identified with the soul, while matter corresponds to the body. The soul is the principle by which the body becomes a living, organic being composed of diverse parts organized into a dynamic and functional unity. Because the human being is a natural and living entity, hylomorphism was adopted to explain human structure and functioning: the soul ensures that the human body is truly a body and serves as the source of intellectual and volitional capacities. This understanding of the soul–body relationship provides a conceptual basis for distinguishing hylomorphic intuitions from both dualist and reductive physicalist interpretations in empirical research.

Hylomorphism entered Christian anthropological reflection primarily through Albert the Great and Thomas Aquinas. From the thirteenth century onward, it gradually displaced earlier Platonic–Augustinian models and became the dominant philosophical–theological framework in the Latin West (MacDonald 2003). Following Aristotle, Christian thinkers affirmed that soul and body are inseparably united: the soul is not a body, does not pre-exist the body, and is not a self-subsistent principle of motion independent of the organism. A body without a soul is dead, and a soul without bodily organs cannot perform its functions. As a non-autonomous principle that organizes matter for life, the soul encompasses vegetative, sensory, and rational capacities: the anima vegetativa governs growth, reproduction, and nutrition; the anima sensitiva underlies perception, memory, imagination, emotion, instinct, and locomotion; and the anima rationalis enables thinking and has a spiritual character. Strictly speaking, the human being possesses one soul that simultaneously performs all three sets of functions.

Thomas Aquinas further developed Aristotle’s account by defining the soul as the substantial form of the body—the first and intrinsic act of its existence (Aquinas 1951). The soul not only organizes matter but also confers a specific mode of being. As the principle of vegetative, sensory, and intellectual life, it has powers distinct from its essence (e.g., intellect and will) and is capable of immaterial operations such as knowing, choosing, and loving. Despite its intimate unity with the body, in Thomistic thought, the soul is considered immortal: its source is not the body but the Creator, who brings it into existence. After bodily death, the soul continues to exist but remains ontologically incomplete, oriented toward reunion with the body for the full realization of its nature.

Hylomorphism, long embedded in Christian anthropology, is often contrasted with various forms of dualism—Platonic, Cartesian, and others (Fine 1994; Loux 2005; Swinburne 1997; Rozemond 1998)—even though it is sometimes itself labeled “dualist” (Moreland 2014, p. 21). The crucial difference lies in its strong emphasis on the intrinsic unity of the human being. This unity is not external or accidental, nor does it arise from combining two independent substances; rather, it is internal, as the soul organizes matter and thereby constitutes the body as a living whole. Accordingly, the body is not treated as a lower or inferior component, nor is the soul understood as imprisoned within it. Instead, the soul organizes the entirety of bodily life, from physiological processes to intellectual activity.

Because the soul is directly united with the body, it does not occupy a specific spatial location but is present throughout the organism and all its parts. Union with the body is therefore more natural to the soul than disembodied existence. After death, the soul ceases to perform vegetative and sensory functions but retains its orientation toward the body. For this reason, Aquinas emphasized the doctrine of the resurrection of the body, understood as the restoration of full psychophysical unity.

This framework was incorporated into Christian theology and widely disseminated through religious education, liturgy, catechesis, and Church documents. For decades, Thomistic anthropology was presented in theological handbooks through propositions such as: “The human being is composed of body and spiritual soul; the soul is united to the body as form to matter; the soul is immortal; and it is created directly by God” (Ravasi 2003, p. 194). A well-known compendium of doctrinal statements (Neuner and Dupuis 2004) similarly summarizes Church teaching with entries such as: “The human being is created with soul and body … the soul is directly created by God … the soul is the ‘form’ of the body” (p. 166).

Although the Catholic Church does not bind itself to a single anthropological model, hylomorphism remains an influential framework despite twentieth-century critiques of Thomism (Ganoczy 1976). This persistence may reflect the absence of an alternative account offering a comparably coherent understanding of human nature. As Vaccaro (2001) observes, “if neo-scholastic anthropology is still used (or at least studied), it is not out of affection for theological archaeology but because modern culture has failed to provide an equally well-constructed anthropology” (p. 44).

Today, the primary source of religious education is the Catechism of the Catholic Church (CCC), promulgated in 1992. Although written in accessible language, the CCC presents a systematic account of Catholic doctrine grounded in its philosophical tradition. Anthropological teaching appears especially in “Section 2”, “The Profession of the Christian Faith,” under the chapter Corpore et anima unus (“one in body and soul”). In this doctrinal framework, the soul is described as a substantial, immaterial, and immortal principle created directly by God, the center of moral agency, and the “f” of the body. The hylomorphic basis of this view is expressed in the following passage: “The unity of soul and body is so profound that one has to consider the soul to be the ‘form’ of the body; that is, it is because of its spiritual soul that the body made of matter is a living, human body… their union forms a single nature” (Catechism of the Catholic Church 2023).

These doctrinal formulations provide a culturally transmitted framework that may shape lay representations of the soul–body relationship, even when these representations are not explicitly articulated in theoretical terms. Accordingly, such beliefs should be understood as emerging from processes of religious education, language use, and socialization rather than as direct expressions of universal cognitive structures.

Given the historical development of Christian doctrine and its reliance on Aristotelian–Thomistic conceptual tools, one might expect hylomorphic views of the soul–body relationship to be present within Christian populations. However, their prevalence, internal structure, and psychological functions remain empirically unexplored. This gap is consequential, as it limits the reconstruction of the full spectrum of lay anthropological beliefs and may lead to misleading conclusions, particularly in culturally homogeneous populations such as Catholic societies.

From an empirical perspective, this omission also introduces the risk of systematic misclassification, as beliefs consistent with hylomorphism may be interpreted within the dominant dualist–monist framework, depending on how they are operationalized. Addressing this limitation requires conceptual and methodological tools capable of capturing underrepresented belief structures. The present study responds to this need by developing and validating the Hylomorphism Inventory (HI), a psychometric instrument designed to identify beliefs that are cognitively consistent with key assumptions of hylomorphism and to distinguish them from dualist and monist belief systems.

Importantly, the analysis does not assume or test the ontological validity of any of the above claims. Hylomorphism serves here as a theoretical reference point for identifying patterns of lay beliefs that may be cognitively congruent with selected aspects of this framework. The construct under investigation is therefore epistemically “thin” and descriptive, rather than metaphysically committal. Moreover, no claims are made regarding the innate or universal nature of these intuitions. The focus is on belief patterns within a culturally and religiously shaped population, in which anthropological representations are likely to reflect processes of socialization and cultural transmission. Addressing questions concerning cognitive “default settings” would require developmental, cross-cultural, and implicit-measure approaches that fall beyond the scope of this work.

2. The Study Overview

The development of the HI followed a two-stage process: construction and validation. Therefore, the study: (1) generated a set of items, (2) revealed the factor structure, (3) tested the factorial composition of the generated items by using Exploratory Factor Analysis (EFA, training subsample, n = 157), and (4) confirmed the proposed construct structure (CFA, test subsample, n = 250). In both cases, the sample sizes met the requirements for scale development procedures (Hoe 2008; Pett et al. 2003; Singh et al. 2016). The procedures, data collection, and results are described in detail in the next subchapters. Based on the analysis, the one-dimensional HI scale was developed. Each stage of the research was conducted in Poland.

All statistical analyses were conducted using the R program (4.5.2) and RStudio (2023.06.1+524), harnessing various packages for comprehensive analysis. The tools included (in alphabetical order): corrplot for correlation visualization (Wei and Simko 2024), dplyr for data manipulation (Wickham et al. 2026), haven for data import/export (Wickham et al. 2025), Hmisc for advanced statistical procedures (Harrell 2026), lavaan for Confirmatory Factor Analysis (Rosseel et al. 2025), mvnTest for multivariate normality testing (Pya et al. 2016), PerformanceAnalytics for performance analysis (Peterson and Carl 2024), psych for psychometric analysis (Revelle 2026), RColorBrewer for color schemes (Neuwirth 2022), and semTools for structural equation modeling tools (Jorgensen et al. 2026).

3. Materials and Methods

3.1. Scale Development

The Hylomorphism Inventory (HI) was developed using a deductive, theoretically grounded approach to item construction (Burisch 1984). Its aim was to operationalize beliefs that are congruent with a hylomorphic understanding of the human person. The construct targeted by the scale is therefore best conceptualized as a domain of psychological attitudes and beliefs aligned with key themes of hylomorphic anthropology, rather than as an assessment of philosophical knowledge or doctrinal mastery. Accordingly, the HI should not be interpreted as a direct measure of hylomorphism as a philosophical system, but as an instrument designed to capture lay beliefs that are compatible with selected aspects of a hylomorphic perspective, without presupposing familiarity with, or endorsement of, its full theoretical framework.

An initial pool of 15 items in Polish was generated on the basis of the conceptual framework and subsequently evaluated by expert judges, including psychologists, cognitive scientists, and philosophers. Items deemed insufficiently representative or too narrowly tied to specific formulations of the construct were removed, resulting in a preliminary 10-item version of the scale. Three items (Q3, Q6, Q8) were reverse-scored to reduce response bias, particularly acquiescence, defined as the tendency to agree with statements regardless of their content.

3.2. Participants and Procedure

Participants (N = 407) were adult residents of Poland recruited through the Ariadna online research panel. Data were collected in Polish in July and August 2025 using an online questionnaire. Before participation, respondents were informed about the scientific purpose of the study, the voluntary nature of participation, confidentiality and anonymity protections, and their right to withdraw at any time without penalty. Informed consent was obtained prior to participation. All respondents were complete on the HI items and on the criterion variables used in the present analyses.

In the analytic sample, 45.2% were women (n = 184) and 54.8% were men (n = 223). The mean age was 43.71 years (SD = 10.74). Educational attainment was 0.5% primary or lower secondary (n = 2), 8.6% basic vocational (n = 35), 27.0% secondary (n = 110), 8.4% postsecondary or vocational college (n = 34), 14.3% bachelor’s degree (n = 58), and 41.3% master’s degree or equivalent (n = 168). Place of residence was 20.6% village (n = 84), 9.8% small town up to 20,000 inhabitants (n = 40), 23.3% town of 20,000–99,000 inhabitants (n = 95), 22.4% city of 100,000–500,000 inhabitants (n = 91), and 23.8% metropolis above 500,000 inhabitants (n = 97). With respect to religion, 77.1% identified as Roman Catholic (n = 314), whereas the remaining respondents reported a range of smaller affiliations or selected Other (17.9%, n = 73).

3.3. Measures

Hylomorphism Inventory. The preliminary HI comprised 10 items rated on a 7-point Likert-type scale ranging from 1 (strongly disagree) to 7 (strongly agree), with 4 indicating neither agree nor disagree (Appendix A). Items were administered in Polish. Reverse-worded items (3, 6, and 8) were recoded so that higher values consistently indicated stronger endorsement of beliefs congruent with hylomorphic anthropology.

Anthropological indicators for validity analyses. Convergent and discriminant validity were examined using indicators derived from the Intuitive Ontology Questionnaire (IOQ) developed in this study and conceptually informed by Fortuna et al. (2023a), which includes Likert-type items assessing beliefs about the constitutive role of the body, mind, and soul, together with pictorial overlap items based on methods described by Richert and Harris (2008) and Burgmer and Forstmann (2018). Two related but theoretically distinguishable domains were operationalized. The constitution domain assessed the extent to which participants endorsed body, mind, and soul as constitutive components of the human being. Constitution was indexed by three indicators (Human-Body, Human-Mind, and Human-Soul) rated on a 7-point Likert-type scale ranging from 1 (strongly disagree) to 7 (strongly agree), with 4 indicating neither agree nor disagree. The linkage domain assessed perceived integration or overlap among these components. Linkage was measured by perceived overlap between body and mind, body and soul, and mind and soul. Each pair was assessed on a pictorial scale presenting 11 ordered configurations from complete independence to total overlap. Responses were coded from 1 to 11, with higher values indicating stronger perceived integration.

3.4. Analytic Strategy

The analytic sequence was designed to examine dimensionality and item functioning, test competing structural specifications under genuine cross-validation, and evaluate validity evidence aligned with the revised construct definition. The analytic sample (N = 407) was randomly split, using a fixed seed, into a training subsample for exploratory analyses (n = 157) and an independent test subsample for confirmatory analyses (n = 250). The split was stratified by sex-by-age strata to preserve the overall sample composition.

First, dimensionality was explored in the training subsample using exploratory factor analysis (EFA) with minimum residual extraction based on Pearson correlations. One- and two-factor solutions were compared, and a parallel analysis was conducted.

Second, the item set was examined using Mokken scale analysis (van der Ark 2012) to assess nonparametric scalability, monotonicity, and invariant item ordering, with focused comparisons among four variants: the full 10-item version, a 9-item version excluding item 8, a 9-item version excluding item 10, and an 8-item version excluding both items 8 and 10. These analyses were conducted in the full sample and replicated in the training and test subsamples to evaluate cross-sample stability.

Third, confirmatory factor analysis (CFA) was conducted in the independent test subsample using robust maximum likelihood estimation (MLR). For each item variant, two structural specifications were compared: a baseline one-factor model and a one-factor model augmented with an orthogonal residual wording factor for reverse-worded items. The primary confirmatory model was selected on the basis of robust fit, with information criteria used as tie-breakers, and the selected model was then refit in the full analytic sample.

Fourth, convergent and discriminant validity were examined on the full sample by estimating separate latent profile analyses (LPAs) for the constitution and linkage indicators and testing whether profile membership differentiated HI scores. Follow-up analyses included Welch tests, additive regression models, and a constitution x linkage interaction model.

4. Results

4.1. Sample and Item Descriptives

Analyses were based on the analytic sample of 407 respondents. The exploratory stage used the training subsample (n = 157), whereas the confirmatory stage used the independent test subsample (n = 250). Across the preliminary 10 HI items in the full analytic sample, all seven response categories were used. Item means ranged from 3.13 to 5.10, standard deviations ranged from 1.74 to 1.96, and corrected item-total correlations ranged from 0.32 to 0.83. The lowest corrected item-total correlation was observed for item 8, followed by item 10, which foreshadowed the subsequent item-sensitivity analyses.

4.2. Exploratory Factor Analysis

EFA was conducted in the training subsample using minimum residual extraction on the Pearson correlation matrix. Parallel analysis suggested a two-factor solution. However, interpretation of the two-factor solution was tempered by auxiliary warnings indicating that the estimated factor-score weights were probably incorrect and that an ultra-Heywood case had been detected. These warnings suggested that the improved fit of the two-factor solution should be interpreted cautiously rather than as straightforward evidence for a stable second substantive factor.

In the one-factor solution, most items loaded moderately to strongly on the general factor, but item 8 was distinctly weak (loading = 0.27, h2 = 0.08), and item 10 was also weaker than the core indicators (loading = 0.50, h2 = 0.25). The one-factor model showed poor global fit, chi-square(35) = 132.07, TLI = 0.845, RMSEA = 0.133, 90% CI [0.110, 0.158], SRMR = 0.08, BIC = −44.90. The two-factor oblimin solution improved fit, chi-square(26) = 53.35, TLI = 0.941, RMSEA = 0.082, 90% CI [0.050, 0.113], SRMR = 0.03, BIC = −78.11, with a modest factor correlation of r = 0.31. Nevertheless, the second factor was dominated by item 8 (loading on Factor 2 = 0.75) and, to a lesser extent, by reverse-worded variance in items 3 and 6. This pattern did not support interpreting the second factor as a clean substantive dimension of hylomorphism-related belief.

Omega-based diagnostics converged on the same interpretation. In the training subsample, alpha = 0.89, omega total = 0.91, omega hierarchical = 0.37, and explained common variance (ECV) = 0.34. In the full analytic sample, alpha remained 0.89, omega total = 0.91, omega hierarchical increased modestly to 0.42, and ECV = 0.40. Thus, the preliminary 10-item set clearly contained a strong common core, but the dominance of a single general factor was not sufficient to treat the original version as unproblematically unidimensional. Taken together, the EFA stage suggested that the apparent multidimensionality was driven primarily by item 8 and, secondarily, by residual reverse-wording variance rather than by a stable second substantive construct (see

Table 1. Exploratory factor model fit and omega diagnostics.

Model	Chi-Square (df)	TLI	RMSEA [90% CI]	SRMR	Key Value
1-factor (TRAIN)	132.07 (35)	0.845	0.133 [0.110, 0.158]	0.08	BIC = −44.90
2-factor (TRAIN)	53.35 (26)	0.941	0.082 [0.050, 0.113]	0.03	BIC = −78.11
Omega diagnostics (TRAIN)	-	-	-	-	alpha = 0.89; omega total = 0.91; omega h = 0.37; ECV = 0.34
Omega diagnostics (FULL)	-	-	-	-	alpha = 0.89; omega total = 0.91; omega h = 0.42; ECV = 0.40

Note. EFA was estimated in the training subsample (n = 157) using minimum residual extraction. The two-factor solution fit better statistically, but its interpretability was limited by concentration of the second factor on item 8 and reverse-wording variance. Omega diagnostics refer to the preliminary 10-item form.

and

Table 2. Training-sample EFA pattern loadings for the preliminary 10-item HI.

Item	1-Factor	2-Factor F1	2-Factor F2
1	0.83	0.89	−0.08
2	0.63	0.80	−0.28
3 (R)	0.65	0.48	0.38
4	0.89	0.82	0.15
5	0.81	0.77	0.11
6 (R)	0.71	0.50	0.51
7	0.62	0.58	0.11
8 (R)	0.27	−0.05	0.75
9	0.78	0.81	−0.03
10	0.50	0.45	0.12

Note. The one-factor solution was estimated without rotation; the two-factor solution used oblimin rotation. Bold values indicate the primary factor loading for each item. Item 8 was the clearest outlying indicator, whereas item 10 was weaker than the retained core items but did not define the apparent second factor.

).

4.3. Mokken Scale Analysis and Item Sensitivity

Mokken scale analysis was used to evaluate nonparametric item scalability and to test whether the emerging item-retention conclusions were stable across alternative deletion rules and across samples. Four variants were compared: the full 10-item form, the 9-item form excluding item 8, the 9-item form excluding item 10, and the 8-item form excluding both items 8 and 10.

In the full 10-item version, overall scalability was medium (H = 0.482), but the item-level pattern pointed to a single central problem. The minimum item scalability coefficient was min Hi = 0.249, and one item fell below the conventional warning threshold of Hi < 0.30: item 8. Item 8 also showed the weakest item-rest correlation (r with rest = 0.318), whereas item 10 was weaker than most other items but less problematic (r with rest = 0.408). Monotonicity and invariant item ordering diagnostics converged on the same interpretation. In the full version, Item 8r showed 14 monotonicity violations (crit = 78) and the highest IIO criticism value (crit = 616), whereas item 10 showed 8 monotonicity violations (crit = 82) but a much smaller IIO criticism value (crit = 41). The IIO accuracy coefficient for the full 10-item version was HT = 0.286.

Removing item 8 produced the best overall balance of improvement and content retention. In the full sample, the item-8 deletion increased H to 0.543, raised the weakest retained item coefficient to 0.364, eliminated items below Hi < 0.30, and improved HT to 0.341. By contrast, removing item 10 alone did not resolve the central problem because item 8 remained weak (min Hi = 0.260) and HT declined to 0.196. Removing both items 8 and 10 produced the highest H (0.586) but not the best invariant-ordering performance (HT = 0.236).

This pattern replicated across the training and test subsamples. In the training subsample, Item 8r again showed the lowest ItemH (0.21) and the largest IIO criticism value (crit = 823), and its problems became even more pronounced when Item 10 was removed while retaining Item 8r (crit = 937; HT = 0.144). In the test subsample, Item 8r again fell below Hi = 0.30 in the full version (Hi = 0.28), had the highest IIO criticism value (crit = 598), and remained problematic after deleting Item 10 (Hi = 0.30; crit = 683; HT = 0.217). By contrast, the Item-8r deletion eliminated low-Hi items in both subsamples and produced the strongest HT values among the practically defensible reduced variants, especially in the test subsample (HT = 0.351). Across the full, training, and test datasets, Mokken evidence therefore converged on a clear conclusion: Item 8r was the primary psychometric problem, whereas Item 10 was comparatively weak but not the central source of misfit (see

Table 3. Key Mokken indices across full-sample variants and train-test replications.

Variant	Full H	Full Min Hi	Full HT	Train HT	Test HT	Low-Hi Items Full/Train/Test
Full	0.482	0.249	0.286	0.249	0.313	1/1/1
Drop item 8	0.543	0.364	0.341	0.243	0.351	0/0/0
Drop item 10	0.513	0.260	0.196	0.144	0.217	1/1/1
Drop items 8 and 10	0.586	0.528	0.236	0.144	0.256	0/0/0

Note. H = scale scalability coefficient; Hi = item scalability coefficient; HT = invariant item ordering coefficient. ‘Low-Hi items’ gives the number of items below Hi < 0.30 in the full, training, and test datasets, respectively. The 9-item variant excluding Item 8r provided the most favorable balance between improved scalability, acceptable minimum Hi, and preserved invariant ordering.

).

4.4. Confirmatory Factor Analysis

CFA was conducted in the independent test subsample (n = 250) to cross-validate the exploratory and nonparametric findings. For each item set, two confirmatory models were estimated: a baseline one-factor model and a one-factor model with an orthogonal residual wording factor for reverse-worded items. Using the pre-specified selection rule applied in the test subsample, defined as meeting at least 3 out of 4 model fit improvement criteria (CFI, TLI, RMSEA, SRMR), along with a lower AIC in cases where the change in AIC was statistically significant, the selected primary solution was the base one-factor model (9 items, item 8 removed) without a wording factor.

In the test subsample, the selected model achieved robust CFI = 0.949, robust TLI = 0.930, robust RMSEA = 0.097, and SRMR = 0.038, with AIC = 8055.37 and BIC = 8150.45. Global model fit was acceptable based on CFI, TLI, and SRMR, although RMSEA (~0.10) indicated potential misfit. Given the simplicity of the model and strong evidence of unidimensionality, RMSEA was interpreted with caution and not treated as decisive, as it is known to over-reject simple models with few degrees of freedom and should therefore be considered alongside other fit indices (Kenny et al. 2015). All substantive loadings were statistically significant and ranged from 0.388 (item 10) to 0.901 (item 4), indicating that item 10 remained the weakest retained indicator but still functioned meaningfully. However, although the one-factor model (9 items, item 8 removed) with a wording factor achieved a slightly lower AIC than the base one-factor model, the wording-factor loadings were small (0.224 for item 3 and 0.229 for item 6). Composite reliability was high for the substantive factor (CR = 0.904) and very low for the wording factor (CR = 0.067). Variance-partitioning indices supported dominant unidimensional saturation by the substantive factor (ECV = 0.979, mean IECV = 0.978, min IECV = 0.897). A nested likelihood-ratio test did not indicate a statistically significant improvement from adding the wording factor within the 9-item drop-item-8 variant, ΔΧ² (1) = 0.97, p = 0.325, suggesting that reverse-wording covariance was modest once item 8 had been removed. Therefore, we retained a more parsimonious one-factor model without a method factor, as the inclusion of the wording factor was not supported by the data and did not reflect a substantively meaningful latent dimension.

The selected model was then refit in the full analytic sample (N = 407). Fit in the full sample was mixed but acceptable with caution: robust CFI = 0.946, robust TLI = 0.927, robust RMSEA = 0.099, and SRMR = 0.037. Thus, SRMR and the comparative fit indices were consistent with acceptable fit, whereas RMSEA remained elevated and should not be overinterpreted as evidence of excellent global fit. Standardized loadings on the substantive factor were all significant and ranged from 0.431 (item 10) to 0.897 (item 4), with especially strong indicators for Items 1, 4, 5, and 9 (see

Table 4. Standardized items loadings for the one-factor model (9 items, item 8 removed) refitted in full sample (N = 407).

Item	Standardized Loadings	SE
1	0.811	0.027
2	0.667	0.043
3 (R)	0.649	0.034
4	0.897	0.013
5	0.802	0.028
6 (R)	0.710	0.038
7	0.691	0.038
9	0.769	0.030
10	0.431	0.045

).

The retained 9-item form was overwhelmingly saturated by the substantive factor (ECV = 0.974, mean IECV = 0.972, min IECV = 0.868). Taken together, the CFA stage cross-validated the same solution suggested by EFA and Mokken analyses and supported the final retained form as a primarily unidimensional 9-item scale excluding item 8 (see

Table 5. Robust fit indices for candidate CFA models in the test subsample (n = 250) and refit of the selected model in the full sample (N = 407).

Variant	Model	Robust CFI	Robust TLI	Robust RMSEA	SRMR	AIC	BIC
10-item	1-factor	0.928	0.907	0.104	0.048	9070.73	9176.38
10-item	1-factor + wording factor	0.944	0.925	0.093	0.042	9050.26	9159.43
9-item drop item 8	1-factor	0.949	0.930	0.097	0.038	8055.37	8150.45
9-item drop item 8	1-factor + wording factor	0.949	0.929	0.099	0.038	8055.35	8153.95
9-item drop item 10	1-factor	0.929	0.904	0.115	0.049	8119.80	8214.88
9-item drop item 10	1-factor + wording factor	0.945	0.926	0.103	0.041	8098.69	8197.29
8-item drop items 8 and 10	1-factor	0.949	0.928	0.110	0.037	7104.64	7189.16
8-item drop items 8 and 10	1-factor + wording factor	0.950	0.928	0.113	0.037	7104.29	7192.33
Selected model refit (N = 407)	1-factor	0.946	0.927	0.099	0.037	13,222.44	13,330.68

Note. All models were estimated with robust maximum likelihood. Bold values indicate the best-fitting model based on the overall pattern of fit indices (CFI, TLI, RMSEA, SRMR, AIC, and BIC). The selected primary model from the test subsample was the base one-factor model (9 items, item 8 removed). The final row reports refit of that selected model in the full analytic sample (N = 407).

).

4.5. Convergent and Discriminant Validity

Validity evidence focused on whether HI scores were more strongly associated with endorsements of the components of human nature (constitution) or with their perceived integration or unity (linkage). All downstream validity analyses used the final 9-item HI mean score excluding item 8.

Separate LPAs were estimated for the constitution and linkage indicator blocks in the full analytic sample. In both domains, a two-class Model 1 solution was preferred on the basis of BIC (see

Table 6. Latent profile characteristics for constitution and linkage domains.

Domain	Selected Solution	BIC	Profiles	Profile Means
Constitution	2-class Model 1	3151.50	High: n = 347 (85.3%), entropy = 0.946;	High means:
				body 6.77,
				mind 6.81,
				soul 5.59;
			Low: n = 60 (14.7%)	Low means:
				body 4.57,
				mind 5.28,
				soul 4.72
Linkage	2-class Model 1	3049.20	High: n = 287 (70.5%), entropy = 0.851;	High means:
				body-mind 9.52,
				body-soul 9.00,
				mind-soul 9.08;
			Low: n = 120 (29.5%)	Low means:
				body-mind 7.00,
				body-soul 3.05,
				mind-soul 3.42

). For constitution, the preferred solution had BIC = 3151.50 and yielded a high-constitution profile (n = 347, 85.3%; entropy = 0.946) and a low-constitution profile (n = 60, 14.7%). The high-constitution profile was characterized by very strong endorsement of body and mind as constitutive of the human person, with somewhat weaker endorsement of soul. For linkage, the preferred solution had BIC = 3049.20 and yielded a high-linkage profile (n = 287, 70.5%; entropy = 0.851) and a low-linkage profile (n = 120, 29.5%). The high-linkage profile showed consistently strong perceived overlap across all three body–mind–soul pairings, whereas the low-linkage profile showed substantially reduced overlap, especially for pairings involving soul.

The two domains differed sharply in their relation to the HI criterion. Constitution profile membership did not significantly differentiate HI scores, Welch t(97.63) = 1.25, p = 0.213, d = 0.15, with only a trivial mean difference (4.25 vs. 4.05). In contrast, linkage profile membership showed a large and statistically robust association with HI scores, Welch t(166.66) = 8.77, p < 0.001, d = 1.11, with substantially higher HI scores in the high-linkage group than in the low-linkage group (4.62 vs. 3.27).

The additive regression model reinforced this pattern. Linkage was a strong positive predictor of HI scores (b = 1.34, SE = 0.13, t = 10.15, p < 0.001), whereas constitution was negligible (b = 0.06, SE = 0.17, t = 0.34, p = 0.735). The additive model explained 20.5% of the variance in HI scores (R² = 0.205). In the interaction model, the constitution x linkage interaction was significant (b = 1.00, SE = 0.35, t = 2.86, p = 0.004), increasing explained variance to R² = 0.221. The corresponding omnibus test indicated a negligible main effect of constitution, F(1, 403) = 1.37, p = 0.242, ${\eta }_p^2$ = 0.003, a large main effect of linkage, F(1, 403) = 104.92, p < 0.001, ${\eta }_p^2$ = 0.207, and a small but significant interaction, F(1, 403) = 8.18, p = 0.004, ${\eta }_p^2$ = 0.020.

Cell means clarified the interaction. HI scores were highest among respondents high in both constitution and linkage (M = 4.67), followed by those low in constitution but high in linkage (M = 4.25), those low in both domains (M = 3.73), and those high in constitution but low in linkage (M = 3.16). Tukey-adjusted comparisons indicated that the high-constitution/high-linkage group scored higher than the low-constitution/low-linkage group (p = 0.002) and higher than the high-constitution/low-linkage group (p < 0.001). In addition, the low-constitution/high-linkage group scored higher than the high-constitution/low-linkage group (p < 0.001). The remaining three cell contrasts were not significant after adjustment. In simple-effects analyses, linkage differentiated HI scores within the high-constitution stratum (p < 0.001) but not within the low-constitution stratum (p = 0.111). Constitution differed within both linkage strata: when linkage was low, the high-constitution group scored lower on the HI than the low-constitution group (p = 0.039), whereas when linkage was high, the high-constitution group scored higher than the low-constitution group (p = 0.046).

Overall, the validity pattern was theoretically coherent. Endorsing body, mind, and soul as components of the person was not sufficient to produce higher HI scores. Rather, the strongest evidence of convergent validity emerged for the linkage domain, that is, for perceiving body, mind, and soul as mutually integrated within a unified whole. At the same time, the weak constitution main effect and the small but significant constitution x linkage interaction suggest that constitution captures related content, but that its alignment with the HI depends on whether those components are also perceived as genuinely unified.

5. Discussion

This study set out to operationalize hylomorphism-congruent beliefs—understood not as a full philosophical or theological doctrine, but as a psychologically meaningful pattern of lay beliefs reflecting selected aspects of this framework.

Across complementary psychometric procedures (cross-validated EFA/CFA, Mokken scaling, and item sensitivity checks), the evidence converged on a primarily unidimensional 9-item Hylomorphism Inventory (HI), with one reversed item (Item 8) emerging as the central source of misfit and being removed. The resulting scale retained strong saturation by a substantive common factor, while residual wording covariance among remaining reverse-keyed items appeared modest and did not warrant a method-factor solution as the primary model.

Most importantly for construct validity, HI scores were not meaningfully differentiated by simply endorsing body, mind, and soul as components of the human person (constitution). Instead, HI scores were strongly associated with perceiving these aspects as integrated (linkage), and the observed constitution × linkage pattern suggests that component-endorsing “trialist” intuitions only align with hylomorphic belief when they are coupled with a robust unity intuition. This validity pattern directly resonates with hylomorphism’s core claim that the soul–body relation is an intrinsic unity rather than a contingent conjunction of separable substances.

The present findings should be read against the article’s wider research program: psychology can no longer treat “soul” as a dispensable relic without forfeiting a category that remains active in ordinary experience, language, and cultural meaning-making (Kugelmann 2013, 2023; Wierzbicka 1989). Empirically, belief in the soul is widespread across contexts and persists across development, including the spontaneous differentiation of soul from mind (Gut et al. 2021; Halman et al. 2008; Harris and Corriveau 2020; Richert and Harris 2008).

Within cognitive approaches to folk anthropology, beliefs about body, mind, and soul function as part of an “anthropological mindset” that supports category-based reasoning about what a human being is (Cohen and Barrett 2011; Weisman 2022). Yet those mindsets show substantial variation: people may express intuitive physicalism, intuitive dualism, or intuitive trialism, often without explicit justification and with change across the life course (Barlev and Shtulman 2021; Bloom 2004; Czaja 2006; Lindeman et al. 2015; Richert and Lesage 2019).

The distinctive gap addressed here is that hylomorphism—despite its centrality in Aristotelian–Thomistic anthropology and its longstanding diffusion through Catholic doctrine—has rarely been operationalized as a psychologically measurable belief pattern. Hylomorphism is neither a separable-substances dualism nor a reductive monism; it treats the soul as the substantial form that actualizes matter into a living human body (Ainsworth 2024; Aristotle 2011; Aquinas 1951; Bostock 2006; Piwowarczyk 2020; Shields 1988). It is historically integral to Christian anthropology in the Latin West (MacDonald 2003) and is explicitly represented in catechetical formulations emphasizing profound unity (Catechism of the Catholic Church 2023).

From a measurement standpoint, the challenge is clear: if a questionnaire merely asks whether one “has a soul,” it risks collapsing hylomorphic belief into broader dualist/trialist response patterns. Conversely, if it operationalizes “unity” without theological and philosophical specificity, it risks losing the hylomorphic signature. The present scale-development strategy directly targeted this challenge by constructing items deduced from the hylomorphic conceptual core while explicitly avoiding any implication that the instrument measures formal philosophical training.

Three strands of evidence converged on the same conclusion: a strong common core underlies responses to the HI, but the preliminary 10-item form was not psychometrically “clean” because one reverse-worded item behaved anomalously. In the training-sample EFA, a statistically better-fitting two-factor solution was flagged as interpretively unstable: model warnings (including an ultra-Heywood case) and the concentration of the second factor on Item 8 suggested artifactual multidimensionality rather than a replicable second substantive dimension.

This interpretation was strengthened by nonparametric Mokken scaling. Item 8 carried the weakest item scalability coefficient, violated monotonicity most strongly, and dominated criticism indices for invariant item ordering; removing it produced a marked improvement in scalability while preserving content breadth. Importantly, this pattern replicated across the training and test splits, indicating that the conclusion was not an idiosyncrasy of one subsample.

In confirmatory analyses, the 9-item solution (dropping Item 8) provided the most defensible balance: comparative fit indices and SRMR were acceptable, and—crucially for construct interpretation—variance-partitioning indices indicated that responses were overwhelmingly saturated by the substantive factor. Although RMSEA remained elevated (≈0.10), the authors’ own analytic rationale treated RMSEA cautiously given model simplicity and the broader convergence on unidimensional saturation.

Reverse-keyed items are often used to reduce acquiescence, but in belief domains that are (a) culturally loaded, (b) metaphysically complex, and (c) linguistically delicate, reverse wording can inadvertently introduce interpretive heterogeneity. Here, that risk likely increases because hylomorphism’s central affirmation is not merely “soul exists,” but “soul exists as the body’s form and therefore is intrinsically united to it.” Slight shifts in negation, abstraction, or metaphor may push respondents to read a reverse-keyed statement through a dualist or materialist lens rather than a hylomorphic one.

The CFA comparison between a baseline one-factor model and a one-factor-plus-wording-factor specification is informative in this respect. Once Item 8 was removed, remaining reverse-worded covariance was small, wording-factor reliability was negligible, and the nested test did not support the method factor as a substantively meaningful latent dimension. Practically, this implies that the final 9-item HI is not simply a “pro-soul” scale with a hidden reversal artifact; it behaves as a coherent belief measure whose psychometric integrity depends primarily on avoiding one particularly problematic reversed expression.

The most theoretically diagnostic finding is that HI scores were strongly related to the linkage domain—perceived integration among body, mind, and soul—while the constitution domain—endorsing body/mind/soul as components—did not meaningfully differentiate HI scores.

This pattern is consistent with the interpretation that the HI indexes a hylomorphism-congruent stance rather than generic soul affirmation. Hylomorphism is defined by an internal, non-accidental unity: the soul is the principle that makes the human body a living human body, not a second substance “alongside” the body (Aristotle 2011; Aquinas 1951; Shields 1988). Accordingly, merely saying “humans have body, mind, and soul” is compatible with positions that hylomorphism explicitly rejects or reframes: Cartesian-style separation, folk dualism, or a loosely additive trialism (Barlev and Shtulman 2021; Bloom 2004; Lindeman et al. 2015).

Conceptually, constitution items can be endorsed for multiple reasons: a dualist might affirm “soul” as a separable essence; a physicalist might interpret “soul” metaphorically (Czaja 2006); a trialist might separate mind and soul into different immaterial “centers” (Richert and Lesage 2019). In contrast, linkage items—especially those directly involving the soul—force respondents to make an integration judgment that aligns more tightly with the hylomorphic claim that soul is “form,” not detachable occupant (Catechism of the Catholic Church 2023).

The constitution × linkage interaction supplies an additional interpretive layer: the highest HI scores occurred among those high in both constitution and linkage, but linkage was the primary differentiator within the high-constitution group. This suggests that constitution may function as a necessary-but-not-sufficient background condition for hylomorphic belief endorsement in this population. That is, respondents may need to accept the vocabulary of “body/mind/soul” as meaningfully applicable to human nature before the specifically hylomorphic integration commitment becomes psychologically operative.

At the same time, the lowest HI scores among the high-constitution/low-linkage profile highlight exactly the misclassification problem named in the Introduction: without measuring unity explicitly, a survey might mistakenly treat such respondents as “high in soul belief” or “strongly spiritual,” when their implicit anthropology may be closer to separative dualism than to hylomorphism.

Interpreted through the framework of folk anthropologies, the interaction can be read as evidence that “belief in soul” is not a monolithic dimension. Rather, it is structured by ontological presuppositions about relations among body, mind, and soul—presuppositions that are shaped by development, culture, and religious formation (Boyer 2000; de Cruz and de Smedt 2007, 2015; Forstmann and Burgmer 2015). The HI appears to be particularly sensitive to the unity-relevant aspect of that structure.

In this respect, the data suggest that intuitive anthropology is not uniformly dualistic and instead support more flexible, culturally shaped models of intuitive ontology (Barlev and Shtulman 2021). The present validity evidence offers a refinement that is especially important for religious populations influenced by Aristotelian–Thomistic catechesis: an individual can (a) endorse soul language, (b) even differentiate soul from mind, yet (c) still diverge sharply in whether the soul is imagined as integrated with the body or externally paired with it.

This refinement also aligns with the Introduction’s point that hylomorphism has been deeply disseminated through religious education and authoritative texts, yet its prevalence and functions remain empirically unknown. The linkage-dominant association suggests that the HI is positioned to answer a more specific question than “Do people believe in the soul?”—namely, “Do people hold an integrative soul–body model congruent with hylomorphism?”

6. Limitations

The study relied on an adult Polish sample recruited from an online research panel, with data collected in mid-2025 and a majority identifying as Roman Catholic. This context is substantively appropriate for an initial test because the Introduction argues that hylomorphism is a central component of Catholic anthropological doctrine. At the same time, this sampling frame limits generalization in several ways.

First, online panels can approximate population structure only to the extent that recruitment, participation, and incentive structures do not systematically bias results. Second, even within Poland, the relationship between Catholic identification and the internalization of doctrinal anthropological concepts (including hylomorphism) may vary substantially by cohort, education, and local religious practice. Third, because the HI was first administered in Polish, some of its success (and the failure of the removed item) may partly reflect language-specific semantic nuances in how negation and metaphysical vocabulary are processed.

The construct definition used here is intentionally practical: it targets beliefs congruent with hylomorphism’s key themes without requiring explicit philosophical comprehension. This is a strength for psychological research, but it creates a continuing tension between coverage and purity. A scale can be “hylomorphic” in orientation while still being endorsed for non-hylomorphic reasons (e.g., generic religiosity, cultural conformity, or a non-Thomistic spiritual worldview). Conversely, respondents who are hylomorphically inclined but unfamiliar with some doctrinal vocabulary might under-endorse certain items.

This is partly why the linkage-driven validity result is so important: it provides external evidence that HI scores are not reducible to component endorsement alone. Nonetheless, future work should continue to test whether the HI distinguishes hylomorphism specifically from (a) separative dualism (Fine 1994; Rozemond 1998), (b) more contemporary “soul = mind” intuitions, and (c) metaphorical uses of “soul.”

Finally, the HI captures explicitly articulated beliefs and may not fully reflect implicit or pre-reflective anthropological intuitions, which would require alternative methodological approaches.

7. Future Directions

The Introduction highlights several applied domains in which “soul” discourse plays a role, including moral decision-making (e.g., embryo research), moral status attribution, and contexts of dehumanization such as war, migration, illness, and addiction (Fortuna et al. 2023a, 2023b; Landry et al. 2024; Markowitz and Slovic 2021; Richert and Harris 2008). The HI provides a concrete tool for examining whether beliefs congruent with hylomorphism have distinctive psychological correlates beyond general endorsement of the soul. Building on the present findings, future research may be organized into several interrelated directions.

First, replication with independent samples and explicit tests of measurement invariance is essential. Given that reverse-worded items may introduce interpretive ambiguity, future scale development should prioritize semantically transparent formulations. Where reverse-keyed items are retained, they should be subjected to cognitive interviewing to ensure that they do not inadvertently invite dualist reinterpretations.

Second, cross-linguistic and cross-cultural validation represents a key next step. As hylomorphism has been historically disseminated within Catholic contexts across different cultures, translations of the HI should be evaluated not only in terms of back-translation equivalence but also with respect to the semantic and cultural connotations of “soul” terminology in local linguistic environments (Wierzbicka 1989).

Third, the observed asymmetry between constitution and linkage suggests a theoretically informative distinction that warrants further investigation. Future research should examine the difference between endorsing anthropological components (body, mind, soul) and endorsing their integration, as this distinction appears central to the construct captured by the HI. This distinction may support a refined typology of lay anthropological representations, including: (a) component endorsement without integration (potentially dualist or separative), (b) integrated representations with component endorsement (hylomorphism-congruent), and (c) integration judgments without strong component endorsement (potentially metaphorical or non-substantialist). Such a framework could be linked to broader debates on whether intuitive anthropology is predominantly dualistic or shaped by cultural and educational influences (Barlev and Shtulman 2021; Bloom 2004).

Fourth, future research should examine the predictive role of hylomorphism-congruent beliefs across applied domains. Prior work suggests that soul-related beliefs are associated with moral judgments, bioethical positions, and preferences for practices related to well-being (Fortuna et al. 2023a, 2023b; McGhee et al. 2024; Richert and Harris 2008). The present findings suggest a more specific hypothesis: integration-focused beliefs may be associated with distinct patterns of moral reasoning and behavioral preferences compared to separation-focused (dualist) beliefs, even when both involve endorsement of the soul.

For example, a hylomorphic orientation—emphasizing the intrinsic unity of body and soul and the value of embodied existence—may be associated with stronger endorsement of embodied forms of religious practice (e.g., ritual participation, positively framed fasting practices, or care for bodily health as a religious obligation), as well as with distinctive positions in bioethical dilemmas where bodily continuity and personhood are central. These predictions are theoretically grounded in contrasts with dualist traditions, which often conceptualize the body as secondary or instrumental relative to the soul (Fine 1994; Moreland 2014; Swinburne 1997).

Beyond these domain-specific predictions, the re-emergence of “soul” language in contexts of dignity violation points toward a promising applied research trajectory. Future studies may examine whether anthropological beliefs congruent with hylomorphism function as a psychological buffer against dehumanization by anchoring human dignity in an integrated conception of the person rather than in cognitive performance or disembodied mental attributes (Landry et al. 2024; Markowitz and Slovic 2021). If supported, such findings would position the HI not only as a descriptive tool for studying religiosity, but also as a potentially relevant instrument in contexts where anthropological beliefs inform practices of care, recovery, and social reintegration.

The relevance of this perspective becomes particularly evident in the domain of well-being. Research on intuitive anthropological beliefs suggests that such beliefs may be associated with key dimensions of human flourishing, including health-related behaviors (Burgmer and Forstmann 2018; Forstmann et al. 2012). Empirical findings indicate that individuals endorsing more materialist or embodiment-oriented intuitions may be more likely to engage in health-promoting behaviors and to hold more favorable attitudes toward them, whereas dualist intuitions may weaken the perceived connection between bodily states and psychological well-being.

On this basis, it can be hypothesized that individual differences in engagement with practices aimed at self-improvement and well-being may be partly rooted in implicit anthropological beliefs. From this perspective, the selection, maintenance, and discontinuation of such practices need not be understood solely as a function of situational preferences, but may also be co-shaped by underlying assumptions about the nature of the human person. Anthropological intuitions may thus be conceptualized as a potential “compass” guiding both the perceived appropriateness and the expected effectiveness of these practices.

Accordingly, individuals may differ not only in their preferences for particular forms of practice, but also in their beliefs about whether and how such practices may influence their well-being. It is also plausible that they may be more inclined to initiate and sustain engagement in practices perceived as consistent with their implicit assumptions about the relationship between body, mind, and soul, while being less inclined to engage in practices experienced as incompatible with these assumptions. These hypotheses find indirect support in cross-cultural observations indicating that interpretations of illness and healing may be associated with preferred forms of intervention. For example, in cultural contexts where illness is understood as a disturbance of spiritual order, practices aimed at restoring the integrity of the person through spiritual means—such as rituals performed by intermediaries between material and non-material domains—are more commonly observed (Winkelman 2010).

Within this framework, a particularly promising direction for future research is the examination of beliefs congruent with hylomorphism and their potential role in shaping both preferences for and the effectiveness of well-being–oriented practices. This line of inquiry may be especially relevant in the field of mental health, where materialist models, functional forms of dualism or trialism (e.g., biopsychosocial approaches), and secularized forms of spirituality continue to dominate. In this context, special attention should be given to contemplative practices and mindfulness training, which often represent adaptations of meditative traditions detached from their original soteriological and doctrinal frameworks and reoriented toward psychological goals (Chachignon et al. 2024; Goniszewski 2024). As a result, mindfulness may function as a form of “secular spirituality,” accessible across diverse worldviews, yet transformed in terms of its aims, mechanisms, and underlying anthropological assumptions. This raises the question of the extent to which the effectiveness of such practices may depend on their alignment with individuals’ implicit beliefs about the nature of the human person.

Finally, an important direction for future research concerns the role of hylomorphism-congruent beliefs in shaping everyday practices related to well-being and mental health. Building on prior findings linking anthropological intuitions to health-related behavior, future studies may examine whether the alignment between implicit beliefs about the body–mind–soul relationship and specific forms of practice—such as contemplative practices, mindfulness-based interventions, or religious and pastoral activities—modulates their adoption, persistence, and perceived effectiveness. In contemporary mental health contexts, where materialist and biopsychosocial models coexist with secularized forms of spirituality, this line of inquiry may help clarify whether different anthropological frameworks are associated with systematically different patterns of engagement with practices aimed at psychological and existential functioning. Importantly, these questions should be approached empirically, without presupposing normative superiority of any given model.

8. Conclusions

The development and validation of the HI address a gap in empirical psychology: while existing research offers numerous tools for measuring general belief in the soul, comparatively little attention has been devoted to distinguishing the underlying anthropological assumptions that such beliefs imply—especially in populations shaped by Catholic doctrine and Aristotelian–Thomistic conceptual frameworks.

The psychometric results support a pragmatic conclusion: a concise 9-item HI behaves as a largely unidimensional measure of hylomorphism-congruent belief, provided that item formulations avoid specific reverse-wording pitfalls that can introduce artifactual structure. The validity findings, in turn, support a more substantive theoretical interpretation: what appears to distinguish hylomorphism psychologically is not simply affirming “body, mind, soul” as components, but affirming their deep unity, consistent with the hylomorphic claim that the soul is the form that makes the body living and human (Aristotle 2011; Aquinas 1951; Catechism of the Catholic Church 2023).

In that sense, the HI does more than add another index to the measurement toolbox. It opens a theoretically grounded and empirically testable path for examining whether hylomorphic anthropology—long central in theological discourse and culturally disseminated through catechesis—has distinct empirical correlates in moral cognition, spiritual practice, well-being pursuits, and responses to dignity violations (Burgmer and Forstmann 2018; Fortuna et al. 2023a, 2023b; Landry et al. 2024; Markowitz and Slovic 2021; McGhee et al. 2024).