Introduction
Humans’ relationships to objects begin early in development (Erkolahti and Nyström, Reference Erkolahti and Nyström2009) and become increasingly complex across the lifespan (Ekerdt et al., Reference Ekerdt, Sergeant, Dingel and Bowen2004). At all stages of development, there are myriad reasons why individuals save objects. Objects provide utility in activities of daily living, facilitate human adaption and survival, commemorate relationships with others, and symbolize traditions passed through generations. For these reasons and more, objects provide a sense of security, support, and reward, leading individuals to develop emotional attachments to objects (Richins, Reference Richins2004; Timpano and Port, Reference Timpano and Port2021).
Emotional attachment to certain belongings is normative (Norberg and Rucker, Reference Norberg and Rucker2021); yet, excessive emotional attachment to objects can have deleterious effects. Excessive emotional attachment to objects is a core feature of hoarding disorder, a chronic mental health condition characterized by pervasive object saving tendencies, excessive acquiring behaviours, and disabling amounts of clutter (American Psychiatric Association, 2013). Currently estimated to affect about 2–5% of the general population (Postlethwaite et al., Reference Postlethwaite, Kellett and Mataix-Cols2019), hoarding symptoms are dimensionally distributed along a spectrum ranging from normative acquiring and collecting to severely excessive acquiring and extreme difficulty discarding (Timpano et al., Reference Timpano, Broman-Fulks, Glaesmer, Exner, Rief, Olatunji, Keough, Riccardi, Brähler, Wilhelm and Schmidt2013). Hoarding disorder can be highly debilitating, leading to significant impairments in multiple areas of functioning, including unusable and unsafe living spaces, social isolation, interpersonal conflicts, and reduced quality of life (Chen et al., Reference Chen, McDonald, Wearne and Grisham2022; Nutley et al., Reference Nutley, Read, Martinez, Eichenbaum, Nosheny, Weiner, Mackin and Mathews2022; Tolin et al., Reference Tolin, Das, Hallion, Levy, Wootton and Stevens2019; Yap et al., Reference Yap, Timpano, Isemann, Svehla and Grisham2023).
Object attachment, defined as a strong emotional bond or connection that an individual forms with an inanimate object, has been established as a central motivator of hoarding behaviours (Frost and Hartl, Reference Frost and Hartl1996; Yap and Grisham, Reference Yap and Grisham2019). Individuals with heightened object attachment exhibit hyper-sentimentality to items, viewing their possessions as extensions of the self and relying on them for comfort when experiencing strong negative emotions (David et al., Reference David, Blonner, Forbes and Norberg2021). Higher levels of object attachment are associated with greater hoarding symptoms in both clinical (Grisham et al., Reference Grisham, Frost, Steketee, Kim, Tarkoff and Hood2009; Kyrios et al., Reference Kyrios, Mogan, Moulding, Frost, Yap and Fassnacht2018) and non-clinical (Frost et al., Reference Frost, Hartl, Christian and Williams1995; Yap and Grisham, Reference Yap and Grisham2019) samples. This association remains significant even after controlling for depression, anxiety, and non-sentimental hoarding beliefs (Yap and Grisham, Reference Yap and Grisham2019). A growing body of research aims to better understand drivers of heightened object attachment in order to inform prevention and intervention efforts for hoarding disorder.
One factor thought to underlie object attachment is psychological essentialism (Gelman, Reference Gelman2013). Essentialism is a reasoning heuristic that can impart meaning on objects beyond their physical characteristics, imbuing them with additional, unobserved essences (Bloom, Reference Bloom2011). Critically, the unobserved properties of an object can take precedence over its perceptual or observed properties (Gelman, Reference Gelman2004). Essentialist tendencies allow individuals to make inferences based on both observed and unobserved properties of an object and are informative for inductive reasoning (Gelman and Markman, Reference Gelman and Markman1987; Tarlowski, Reference Tarlowski2018). Essentialism can help explain why, for one person, an empty bottle of sunscreen is perceived as trash, while, for someone else, an empty bottle of sunscreen evokes memories of a meaningful vacation with loved ones. The bottle’s ‘essence’ shifts to encapsulate a cherished memory, and the sentimental value placed on the otherwise trivial object confers upon it an irreplaceable, invisible worth.
Essentialist reasoning is a common cognitive tendency, yet research suggests that this tendency can be influenced by a variety of individual (Foster and Keil, Reference Foster and Keil2009), cultural (Chiu et al., Reference Chiu, Morris, Hong and Menon2000), and contextual factors (Morton et al., Reference Morton, Hornsey and Postmes2009). Given the role of individual differences in shaping essentialist reasoning, we theorize that essentialism may be more pronounced in individuals with hoarding tendencies than those without hoarding tendencies and may further interact with other hoarding risk factors to maintain hoarding symptoms. Frost and Hartl (Reference Frost and Hartl1996) noted that individuals with hoarding disorder often describe their possessions as meaningful symbols of past events and memories. Individuals with hoarding disorder also tend to report lower confidence in their memory, which may elicit an urge to keep possessions in view and concerns about the consequences of forgetting (Hallion et al., Reference Hallion, Diefenbach and Tolin2015; Hartl et al., Reference Hartl, Frost, Allen, Deckersbach, Steketee, Duffany and Savage2004; Kyrios et al., Reference Kyrios, Mogan, Moulding, Frost, Yap and Fassnacht2018). Memories are central to the development of an individual’s self-concept, and individuals with hoarding disorder may view their possessions as a mechanism through which to maintain both their memories and also their sense of self (Yap and Grisham, Reference Yap and Grisham2019). The act of discarding an object, therefore, poses a threat to the memory itself, the emotions associated with the memory, and the aspects of the self that the memory represents (Cherrier and Ponnor, Reference Cherrier and Ponnor2010; Yap and Grisham, Reference Yap and Grisham2019).
A Proustian memory is a specific type of memory that may prompt essentialist reasoning related to objects and may be relevant to understanding object attachment and difficulties discarding in hoarding. The term Proustian memory refers to Marcel Proust’s novel (‘Remembrance of Things Past’) wherein the taste of a madeleine tea cake instantly awakened the narrator’s senses and triggered a powerful childhood memory (Proust, Reference Proust1981). Proustian memories are vivid memories, often triggered by a sight or smell. They evoke intense emotions, sensory experiences, nostalgia, and social-connectiveness (Green et al., Reference Green, Reid, Kneuer and Hedgebeth2023) and prompt the individual to relive past feelings as vivid as when they were first experienced. Proustian memories triggered by objects may thereby strengthen an individual’s urge to save the object by highlighting both the observed memory features (e.g. sunscreen bottle triggers memory of vacation that the individual wants to remember) and the unobserved, emotionally salient properties of the object that increase its perceived value (e.g. sunscreen bottle triggers feelings of calmness as experienced during vacation). The ability of an object to trigger strong memories is a frequently reported memory-associated belief that drives saving behaviours in hoarding disorder (Frost and Hartl, Reference Frost and Hartl1996). The strong emotional and sensory experiences characteristic of the Proust effect may explain, at least in part, why individuals with hoarding symptoms imbue memory-associated objects with sentimental value and exhibit heightened attachment to and difficulties discarding objects regardless of the actual value of the object.
To date, there has been only one unpublished pilot study that provided preliminary evidence of the role of Proustian memories in object attachment and hoarding symptoms (Frost et al., Reference Frost, Wraga, Eldevik, Woolley and Going2019). In this study, college students were asked to recall a memory associated with ten ‘personal’ objects (e.g. blanket, book, clothing) and ten ‘non-personal’ objects (e.g. plate, scissors) and rate the vividness of each memory. Participants described memories of ‘personal’ objects as more vivid and emotional (i.e. a Proust-like effect) than memories of ‘non-personal’ objects, and greater memory vividness was associated with greater hoarding symptoms. While these findings provide initial evidence of the relationship between Proustian-like memories and hoarding symptoms, the study was limited by the reliance on a young adult, college sample. The study was also not able to address whether hoarding symptoms are associated with greater vividness of all types of memories or specifically memories associated with objects. Moreover, objects were standardized across participants, which did not allow for an assessment of idiographic differences in personal relevance or subjective value attributed to the object. Given how the subjective, rather than objective, value of objects has been shown to relate to hoarding behaviours (Grisham et al., Reference Grisham, Frost, Steketee, Kim, Tarkoff and Hood2009), it is crucial to consider subjective value to more accurately assess the relationships among Proustian-like memory, object attachment, and hoarding symptoms.
The current study expanded on the previously conducted pilot study and examined the effect of Proustian-like memories on hoarding symptoms in a large community sample. Specifically, we investigated the relationships between hoarding symptom severity and memory vividness, which is the core feature of Proustian memories (Green et al., Reference Green, Reid, Kneuer and Hedgebeth2023). Participants completed a novel task in which they were asked to recall and rate the intensity of a memory associated with a cherished belonging, a mundane belonging, and a recent vacation, which served as a non-object control memory. This allowed for a direct comparison of the vividness of memories of objects and non-objects, as well as between belongings with higher and lower perceived values. Our first aim was to examine the association between memory vividness and hoarding symptoms. Based on previous findings of the role of memory in object attachment (e.g. Hallion et al., Reference Hallion, Diefenbach and Tolin2015; Yap and Grisham, Reference Yap and Grisham2019), we hypothesized that greater vividness of object memories, but not vividness of non-object memories, would be associated with greater hoarding symptoms. Furthermore, given that a feature of hoarding symptoms is the desire to collect even objects of little actual value or function (Frost and Hartl, Reference Frost and Hartl1996; Yap and Grisham, Reference Yap and Grisham2019), we hypothesized that vividness of memories of mundane objects would be more strongly associated with hoarding symptoms than those of cherished objects. In other words, we expected that the relationship between memory vividness and hoarding symptoms would be moderated by the type of memory in recall (i.e. memory of a cherished belonging, memory of a mundane belonging, or a non-object memory). Our second aim was to examine whether memory vividness impacted the urge to save the object in recall. We expected that individuals would report a stronger urge to save cherished compared with mundane objects. We hypothesized that greater memory vividness would be associated with a greater urge to save the object in recall. We further hypothesized that the relationship between memory vividness and urge to save would be moderated by object type (cherished versus mundane), such that memory vividness would have a greater effect on the urge to save mundane belongings compared with cherished belongings. Our final aim was to examine the moderating effect of hoarding symptoms on the relationship between memory vividness and urge to save objects. We hypothesized that the effect of memory vividness on urge to save objects would be stronger among individuals with greater hoarding symptoms compared with those with fewer hoarding symptoms.
Method
Participants and procedures
Participants (N=508) were recruited from Prolific.com, an online crowd-sourcing platform known for its reliability in collecting high-quality data from a diverse sample of adults. Research has demonstrated that Prolific consistently produces high data quality in online human-subjects research (Douglas et al., Reference Douglas, Ewell and Brauer2023). Online crowd-sourcing platforms also offer a cost-effective approach to investigating vulnerability for different mental illnesses, given higher rates of psychiatric symptom endorsement compared with the general population (Arditte et al., Reference Arditte, Çek, Shaw and Timpano2016). As hoarding behaviours exist on a continuum, ranging from non-clinical to impairing (Timpano et al., Reference Timpano, Broman-Fulks, Glaesmer, Exner, Rief, Olatunji, Keough, Riccardi, Brähler, Wilhelm and Schmidt2013), we anticipated that the associations observed in our study would be present, and potentially more pronounced, in clinical samples.
This study was performed in line with the principles of the Declaration of Helsinki. All procedures were approved by the University of Miami Institutional Review Board (#20130686). All participants provided informed consent. After providing informed consent, participants completed the Proustian Memory Task and a battery of self-report questionnaires. The survey was estimated to take approximately 30 minutes to complete, and participants were compensated in accordance with the standard rate of US$12.00 per hour in Prolific studies. On average, participants were compensated US$4.20 for their participation.
Participants were excluded from data analysis if they started but did not complete the study (n=42), completed the survey in fewer than 600 seconds (i.e. indicating lack of thoughtful responding to study questions; n=15), failed to identify objects and associated memories for the Proustian Memory Task (n=6), failed an attention check (n=1), or were recorded as having an IP address outside of the United States (n=1). Qualitative responses to the Proustian Memory Task were inspected for non-sensical responses (i.e. responses in which answers were not consistent with questions asked), of which none was identified.
The final sample included 443 individuals aged 19 to 83 (M=38.80 years old) who resided in the United States. Approximately half of the sample identified as cisgender female (52.60%), approximately half identified as cisgender male (44.70%), and a few participants identified as non-binary or transgender (2.70%). A majority of participants were White (73.30%), while smaller numbers of participants were Black or African American (12.60%), Asian (6.90%), Native American/Pacific Islander (0.90%), Middle Eastern (0.50%), or multi-racial (5.70%). Most participants were non-Hispanic (87.58%). Most common educational status was having received a degree from a four-year college (37.70%). A considerable minority of participants (32.05%) reported having received a prior psychiatric diagnosis, and approximately half of the sample (48.76%) reported having received prior counseling for mental health concerns.
Measures
Proustian Memory Task
A novel task measure was developed to assess Proustian memories associated with objects and non-objects. The task used in the current study improved upon the version developed in the aforementioned pilot study (Frost et al., 2019) in two ways. First, while the task developed in the pilot study only assessed vividness of memories associated with objects, the current task included an item assessing vividness of a memory associated with a non-object. This allowed us to examine whether differences in memory vividness in individuals with and without hoarding symptoms were specific to object memories or represented a more general memory-related trait. Second, while the task developed in the pilot study assessed vividness of memories associated with a standardized list of objects (e.g. blanket, book, scissors), the current task asked participants to provide idiographic descriptions of objects that they possess. Participants were asked to recall, describe in written text, and rate the vividness and emotionality of three different types of memories: (1) a memory associated with their most valued belonging (i.e. cherished object), (2) a memory associated with a random possession selected from their kitchen ‘junk’ drawer (i.e. mundane object), and (3) a memory associated with a past vacation (i.e. non-object). Memory vividness was captured using six items (e.g. ‘How vivid or powerful is the memory of this event?’, ‘How strong is the feeling of being brought back in time to the occurrence of the event?’), each rated on a Likert-type scale from 1 (‘not at all’) to 9 (‘very much’). An overall memory vividness score for each condition was calculated as the mean of responses to these six items. For each object (cherished and mundane), participants also rated the strength of their urge to save the object. See Supplementary material (1) for the specific memory prompts and items used in the Proustian Memory Task.
Saving Inventory Revised (SI-R; Frost et al., Reference Frost, Steketee and Grisham2004)
The SI-R is a 23-item self-report measure of hoarding symptoms across three subscales: acquisition, difficulty discarding, and clutter. Participants rate items on a 5-point Likert scale from 0 (‘not at all/no distress’) to 4 (‘always or almost always/extreme distress’). The total score is a measure of total hoarding symptom severity, with higher scores indicating greater severity of hoarding behaviors. The SI-R has demonstrated good internal reliability, and test–retest reliability in both clinical (Frost et al., Reference Frost, Steketee and Grisham2004) and normative samples (Coles et al., Reference Coles, Frost, Heimberg and Steketee2003). The current study used a shortened, 12-item version of the original SI-R. This shortened version (SIR-12) was derived from the 23-item version using an item response theory framework in a large, independent sample of individuals with hoarding disorder and matched controls. Additional information on this 12-item version is available from the corresponding author. The internal consistency reliability of the SIR-12 in the present sample was excellent (α=.93).
Data analytic approach
Analyses were conducted in RStudio (R Core Team, 2023). Given that incomplete survey responses were excluded from this study, there were no missing data on the measures included in analysis. Pearson correlations were conducted among all study variables. Multi-level modelling was used for data analysis as multiple responses (i.e. vividness and urge to save ratings for multiple object types) were nested within participants. To examine aim 1, we conducted a multi-level model of the effect of memory type (cherished object, mundane object, or non-object) and hoarding symptoms on memory vividness. We used planned orthogonal contrasts to compare the relationship between hoarding symptoms and memory vividness between (1) objects and non-objects and (2) cherished objects and mundane objects. To examine aim 2, we conducted a multi-level model of the effect of memory vividness, object type (cherished or mundane), and their interaction on the urge to save the object. Memory vividness and urge to save ratings were recoded to subtract 1 such that scores ranged from 0 to 8, with 0 representing no salient memory. To examine aim 3, we added hoarding symptoms to the multi-level model conducted in aim 1 and examined the interaction between memory vividness, object type, and hoarding symptoms on urge to save the identified object. Cohen’s d was used to characterize effect sizes for all analyses, with values of 0–0.2 indicating very small (often trivial) effects, 0.2–0.5 indicating small to medium effects, 0.5–0.8 indicating medium to large effects, and above 0.8 indicating large effects.
Results
There was wide variability in participants’ descriptions of memories of cherished and mundane objects. Examples of cherished objects and associated memories included ‘A horse figurine; my late husband bought it for me when we were first dating’, ‘My guitar; touring Europe with my band’, and ‘Books; sitting back, relaxing, and reading’. Examples of mundane objects and associated memories included ‘Scissors; opening an Amazon box’, ‘Rusty can opener; being annoyed I couldn’t open a can easily’, and ‘Flashlight; I got it last year when we lost power’. See Supplementary material (2) for density plots of memory vividness by each type of memory. On average, collapsing across cherished and mundane objects, participants reported greater vividness of memories associated with non-objects (M=4.56 out of 8 possible points; SD=1.18) compared with those associated with objects (M=4.25 out of 8 possible points; SD=1.14; t 442=5.08, p<.001, d=0.24). Participants reported greater vividness of memories associated with cherished (M=6.45 out of 8 possible points; SD=1.33) compared with mundane objects (M=2.05 out of 8 possible points; SD=1.60; t 442=49.74, p<.001, d=2.36).
Aim 1: Relationship between memory vividness and hoarding symptoms
See Supplementary material (3) for descriptive statistics and bivariate correlations among all variables included in the study. Hoarding symptoms were associated with greater vividness of mundane object memories (r=.13, p=.008) but were not associated with vividness of cherished object memories (r=.05, p=.304), or with vividness of non-object memories (r=–.02, p=.708). As hypothesized, memory type (cherished, mundane, or non-object) moderated the relationship between memory vividness and hoarding symptoms (Fig. 1). Individuals with greater hoarding symptoms did not differ from those with lower hoarding symptoms on vividness of memories associated with cherished objects (B=0.01, SE=0.01, p=.322), nor with non-objects (B=–0.00, SE=0.01, p=.747) but reported higher vividness of memories associated with mundane objects compared with individuals with lower hoarding symptoms (B=0.02, SE=0.01, p=.002). Collapsing across cherished and mundane objects, individuals with greater hoarding symptoms exhibited a smaller discrepancy in vividness ratings between object memories versus non-object memories compared with individuals with lower hoarding symptoms (B=–0.02, SE=0.01, p=.025). See Table 1 for full model results.
Figure 1. Moderating effect of memory type (object memories versus non-object memory) on the relationship between memory vividness and hoarding symptoms.
Table 1. Multi-level model of the impact of memory type (cherished object, mundane object, or non-object) and hoarding symptoms on memory vividness
Planned orthogonal contrasts were (1) Object versus non-object and (2) Object type: cherished object versus mundane object. Hoarding symptoms were mean-centered. *p <.05; ***p<.001.
Aim 2: Impact of memory vividness on urge to save the object
As expected, participants reported a greater urge to save cherished objects (M=7.21 out of 8 possible points; SD=1.30) compared with mundane objects (M=2.07 out of 8 possible points; SD=2.02; t 442=48.12, p<.001). Collapsing across cherished and mundane objects, greater memory vividness was associated with a greater urge to save the object (B=0.53, SE=0.04, p<.001). Object type moderated the relationship between memory vividness and urge to save (B=–0.26, SE=0.07, p<.001), such that the relationship was stronger for mundane (B=0.66, SE=0.04, p<.001) compared with cherished objects (B=0.41, SE=0.05, p<.001). See Table 2 for full model results and see Fig. 2 for a plot of this moderation effect.
Table 2. Multi-level model of the impact of object type (cherished object or mundane object) and object memory vividness on urge to save the identified object
Memory vividness is mean-centered. Object type (cherished versus mundane object) is coded as 0.5=cherished object and –0.5=mundane object. ***p<.001.
Figure 2. Moderating effect of object type (cherished versus mundane) on the relationship between object memory vividness and urge to save the object.
Aim 3: Moderating effect of hoarding symptoms on relationship between memory vividness, object type, and urge to save
Contrary to our hypothesis, hoarding symptoms did not moderate the relationship between memory vividness, object type, and urge to save the identified object (B=0.01, SE=0.01, p=.273). The relationship between memory vividness and urge to save cherished and mundane objects was not impacted by hoarding symptom severity. See Table 3 for full model results and Supplementary material (4) for a plot of this analysis.
Table 3. Multi-level model of the impact of memory vividness, object type, and hoarding symptoms on the urge to save the identified object
Object type (cherished versus mundane object) is coded as 0.5=cherished object and –0.5=mundane object. Memory vividness and hoarding symptoms are mean-centered. ***p<.001.
Discussion
The current study piloted a novel Proustian Memory Task with a large community sample to examine whether self-reported vividness of memories associated with cherished objects, mundane objects, and a recent vacation (i.e. non-object control) was associated with hoarding symptoms and urge to save the belongings. Overall, findings provide preliminary evidence that the experience of Proustian-like memories, particularly those related to mundane objects, is implicated in hoarding symptoms.
Regardless of hoarding symptoms, participants rated cherished object memories as the most vivid, followed by memories of a recent vacation and finally memories of mundane objects. Consistent with our hypothesis, hoarding symptoms were specifically associated with greater vividness of mundane object memories, although the size of this effect was small. This finding aligns with previous research showing that individuals with and without hoarding symptoms exhibit similar emotional attachment to cherished possessions (Yap and Grisham, Reference Yap and Grisham2020). This suggests that it is not just the experience of Proustian memories broadly, but specifically the experience of Proustian memories related to mundane possessions, that may be implicated in hoarding symptoms.
A similar pattern of findings emerged for the urge to save objects. Greater vividness of memories associated with objects was associated with a greater urge to save both cherished and mundane objects, providing further evidence of the role of Proustian-like memories in features associated with hoarding symptoms. However, the relationship between memory vividness and urge to save was stronger for mundane compared with cherished objects. In other words, the experience of a Proustian-like memory associated with a cherished possession may increase urge to save the possession, but individuals likely have an urge to save their cherished possessions even in the absence of Proustian memories associated with them. Interestingly, and contrary to our hypothesis, the relationship between memory vividness and urge to save cherished and mundane objects was not impacted by hoarding symptoms. Taken together, these findings are consistent with existing literature suggesting that the saving behaviours characteristic of individuals with hoarding disorder are not necessarily functionally distinct from those exhibited by individuals without hoarding disorder. Instead, it is the range of objects to which these motivations apply – specifically mundane objects – and the intensity of the associated emotional experience that differentiates individuals with and without hoarding symptoms (Frost et al., Reference Frost, Steketee, Tolin, Sinopoli and Ruby2015).
Findings can be used to inform existing theoretical models and treatments for hoarding disorder. The cognitive behavioural model of hoarding disorder (Frost and Hartl, Reference Frost and Hartl1996) posits that hoarding symptoms stem in part from erroneous or maladaptive beliefs about possessions, such as the over-estimation of the value of an object. Our findings suggest that individuals with hoarding symptoms experience strong, emotional memories associated with mundane objects. These Proustian-like memories may confer upon the object a symbolic meaning related to the memory itself and/or the experience that the memory provokes. Individuals with hoarding symptoms may use essentialist reasoning heuristics (Gelman, Reference Gelman2004) to determine that the unobserved properties of the object (i.e. the Proustian memory that the object elicits) take precedence over its observed properties (i.e. functional or monetary value). Individuals with hoarding symptoms also often report low confidence in their memory and concerns about the consequences of forgetting memories (Hallion et al., Reference Hallion, Diefenbach and Tolin2015; Hartl et al., Reference Hartl, Frost, Allen, Deckersbach, Steketee, Duffany and Savage2004), which may be heightened for Proustian memories as these can evoke strong sensory experiences, nostalgia, and feelings of social connectedness (Green et al., Reference Green, Reid, Kneuer and Hedgebeth2023). Accordingly, the urge to save and avoidance of discarding mundane objects may be driven by a fear of losing not just the memory but also the sense of connection to the self and others that the memory provides (Cherrier and Ponnor, Reference Cherrier and Ponnor2010; Yap and Grisham, Reference Yap and Grisham2019). Interventions aiming to improve hoarding symptoms should therefore aim to address maladaptive beliefs about the importance of Proustian memories, increase tolerance for uncertainty related to forgetting, bolster social connectedness through means other than Proustian memories, and reduce avoidance of emotion-provoking discarding tasks.
Our conclusions should be interpreted in the context of study limitations. First, the Proust effect is thought to occur in response to a sensation, such as sight or smell (e.g. Chu and Downes, Reference Chu and Downes2002; Herz and Schooler, Reference Herz and Schooler2002). In this study, we asked participants to recall memories associated with cherished and mundane objects, but the objects were not in sight during recall. Recalling such memories at the sight of a possession may elicit greater memory vividness. Relatedly, the Proust effect has not yet been studied within the context of hoarding symptoms, and the Proustian Memory Task used to assess this construct has yet to be validated. The absence of a relationship between hoarding symptoms and vividness of cherished object memories may have resulted from a ceiling effect of memory vividness ratings as participants generally rated these memories as highly vivid. We used a shortened 12-item version of the longer 23-item Saving Inventory-Revised (Frost et al., Reference Frost, Steketee and Grisham2004), and the psychometric properties of this shortened version are not published but may be made available upon request from the corresponding author. A study that was recently published by a different group examined the psychometric properties of two alternative shortened versions: a 9-item version and 3-item version (Sarvestani et al., Reference Sarvestani, David, Nedeljkovic, Norberg and Moulding2025). The items included in the 9-item and 3-item versions largely overlap with those included in our 12-item version; the distinctions between our 12-item version and the published 9- and 3-item versions relate to minor differences in items used to capture control of and interference due to clutter. Additionally, participants in this study were predominantly White, non-Hispanic, and college educated. This sample is therefore not representative of the general United States population and is likely not representative of individuals seeking treatment for hoarding disorder. Specifically, existing meta-analytic studies characterizing individuals seeking treatment for hoarding disorder indicate that these individuals are often older in age and predominantly female (Rodgers et al., Reference Rodgers, McDonald and Wootton2021). Our reliance on a crowd-sourcing platform and a shortened version of the measure of hoarding symptoms did not allow us to characterize the percentage of individuals in our sample who exhibited clinically elevated hoarding symptoms; however, previous research conducted with similar crowd-sourcing platforms has found higher rates of clinically elevated hoarding symptoms (7%) compared with the national 12-month hoarding prevalence rate (approximately 2%; Arditte Hall et al., Reference Arditte Hall, Stamatis, Shaw and Timpano2019). The cross-sectional design of this study also limits the ability to make inferences about the directionality of the associations observed. For example, while it is theorized that the experience of Proustian memories related to objects is the driver of urge to save the object (i.e. in a causal sequence), it is also possible that the urge to save a given possession leads individuals to reminisce and experience Proustian memories, or simply that the association observed is due shared method variance.
Future studies of the Proust effect would benefit from supplementing self-reports with behavioral indicators (e.g. physiological arousal during memory recall) to assess method variance. As the current study design was cross-sectional in nature, further experimental and/or longitudinal research is needed to establish the directionality of the noted relationships. Analysis of text responses provided to memory recall prompts (e.g. via qualitative coding or natural language processing) could provide information as to whether the specific objects identified and memories recalled differ qualitatively between individuals with and without hoarding symptoms. Future research should also consider the associations between Proustian memories and hoarding symptoms in conjunction with other cognitive and memory-related processes. Existing evidence indicates that individuals with hoarding disorder report lower confidence in their memory and greater fear of forgetting important memories than individuals without hoarding disorder (Hallion et al., Reference Hallion, Diefenbach and Tolin2015; Hartl et al., Reference Hartl, Frost, Allen, Deckersbach, Steketee, Duffany and Savage2004; Kyrios et al., Reference Kyrios, Mogan, Moulding, Frost, Yap and Fassnacht2018). It would be interesting to assess whether the experience of Proustian memories related to objects impacts memory confidence. While it may appear as though greater memory vividness should be associated with greater confidence in memory, it could also be the case that individuals experiencing Proustian memories feel immersed in the sensory-related aspects of the memory and less confident about the actual events that occurred. Assessing memory-related cognitions (e.g. via the Saving Cognitions Inventory; Steketee et al., Reference Steketee, Frost and Kyrios2003) in addition to memory vividness and memory confidence could yield insight into whether individuals with hoarding disorder are fearful of losing not only the memory itself but also the emotions and sensory experiences that the memory invokes (see Cherrier and Ponnor, Reference Cherrier and Ponnor2010; Yap and Grisham, Reference Yap and Grisham2019).
Our study provides preliminary support of a Proust effect that may influence object attachment and hoarding symptoms. While Proustian memories are generally considered to be pleasant emotional and sensory experiences (and were operationalized as such for the purpose of this study), it is possible that some individuals experience Proustian memories as distressing (e.g. a vivid, sensory-inducing memory of a traumatic event; Stewart et al., Reference Stewart, Brewin and Gregory2020) or experience both positive and negative emotions as a result of these memories. Future research examining the affective qualities of object memories for individuals with hoarding disorder and potential mediators of the relationship between Proustian memories and hoarding symptoms could inform strategies to intervene on memory processes more directly (e.g. using rescripting, imaginal exposure, and/or cognitive restructuring; see Sabel et al., Reference Sabel, Yap and Grisham2024) to hopefully alleviate distress and impairment associated with object attachment in hoarding disorder.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S135246582510115X
Data availability statement
The data may be made available upon request to the corresponding author.
Acknowledgements
None.
Author contributions
Lauren Milgram: Formal analysis (lead), Writing - original draft (lead), Writing - review & editing (equal); Junjia Xu: Writing - original draft (equal), Writing - review & editing (supporting); Randy O. Frost: Conceptualization (lead), Methodology (equal), Writing - review & editing (equal); Elizabeth A. Offermann: Conceptualization (equal), Writing - review & editing (supporting); Kiara R. Timpano: Conceptualization (equal), Funding acquisition (lead), Methodology (lead), Project administration (lead), Writing - original draft (equal), Writing - review & editing (equal).
Financial support
This work was funded by the National Institute of Mental Health (NIMH 1R21MH116131-01A1).
Competing interests
The authors report no competing interests.
Ethical standards
This study was performed in line with the principles of the Declaration of Helsinki. All procedures were approved by the University of Miami Institutional Review Board (#20130686). All participants provided informed consent.
Open access
Comments
No Comments have been published for this article.